1 \input ../man/texinfo @c -*-texinfo-*-
2 @c \input texinfo @c -*-texinfo-*-
3 @c change to \input texinfo if building on own.
4 @comment %**start of header
5 @setfilename ../info/eintr
6 @c setfilename emacs-lisp-intro.info
7 @c sethtmlfilename emacs-lisp-intro.html
8 @settitle Programming in Emacs Lisp
11 @setchapternewpage odd
15 @c <<<< For hard copy printing, this file is now
16 @c set for smallbook, which works for all sizes
17 @c of paper, and with Postscript figures >>>>
20 @set print-postscript-figures
22 @c clear print-postscript-figures
25 @comment %**end of header
27 @set edition-number 3.07
28 @set update-date 9 November 2006
31 ## Summary of shell commands to create various output formats:
33 pushd /usr/local/src/emacs/lispintro/
37 makeinfo --paragraph-indent=0 --verbose emacs-lisp-intro.texi
39 ## ;; (progn (when (bufferp (get-buffer "*info*")) (kill-buffer "*info*")) (info "/usr/local/src/emacs/info/eintr"))
42 texi2dvi emacs-lisp-intro.texi
44 ## xdvi -margins 24pt -topmargin 4pt -offsets 24pt -geometry 760x1140 -s 5 -useTeXpages -mousemode 1 emacs-lisp-intro.dvi &
47 makeinfo --html --no-split --verbose emacs-lisp-intro.texi
49 ## galeon emacs-lisp-intro.html
52 makeinfo --fill-column=70 --no-split --paragraph-indent=0 \
53 --verbose --no-headers --output=emacs-lisp-intro.txt emacs-lisp-intro.texi
59 mtusb # mount -v -t ext3 /dev/sda /mnt
60 cp -v /u/intro/emacs-lisp-intro.texi /mnt/backup/intro/emacs-lisp-intro.texi
61 umtusb # umount -v /mnt
64 ## Other shell commands
66 pushd /usr/local/src/emacs/lispintro/
70 texi2dvi --pdf emacs-lisp-intro.texi
71 # xpdf emacs-lisp-intro.pdf &
73 ## DocBook -- note file extension
74 makeinfo --docbook --no-split --paragraph-indent=0 \
75 --verbose --output=emacs-lisp-intro.docbook emacs-lisp-intro.texi
77 ## XML with a Texinfo DTD -- note file extension
78 makeinfo --xml --no-split --paragraph-indent=0 \
79 --verbose --output=emacs-lisp-intro.texinfoxml emacs-lisp-intro.texi
81 ## PostScript (needs DVI)
82 # gv emacs-lisp-intro.ps &
83 # Create DVI if we lack it
84 # texi2dvi emacs-lisp-intro.texi
85 dvips emacs-lisp-intro.dvi -o emacs-lisp-intro.ps
88 # Use OpenOffice to view RTF
89 # Create HTML if we lack it
90 # makeinfo --no-split --html emacs-lisp-intro.texi
91 /usr/local/src/html2rtf.pl emacs-lisp-intro.html
94 /usr/bin/rtf2latex emacs-lisp-intro.rtf
100 @c ================ Included Figures ================
102 @c Set print-postscript-figures if you print PostScript figures.
103 @c If you clear this, the ten figures will be printed as ASCII diagrams.
104 @c (This is not relevant to Info, since Info only handles ASCII.)
105 @c Your site may require editing changes to print PostScript; in this
106 @c case, search for `print-postscript-figures' and make appropriate changes.
108 @c ================ How to Create an Info file ================
110 @c If you have `makeinfo' installed, run the following command
112 @c makeinfo emacs-lisp-intro.texi
114 @c or, if you want a single, large Info file, and no paragraph indents:
115 @c makeinfo --no-split --paragraph-indent=0 --verbose emacs-lisp-intro.texi
117 @c After creating the Info file, edit your Info `dir' file, if the
118 @c `dircategory' section below does not enable your system to
119 @c install the manual automatically.
120 @c (The `dir' file is often in the `/usr/local/info/' directory.)
122 @c ================ How to Create an HTML file ================
124 @c To convert to HTML format
125 @c makeinfo --html --no-split --verbose emacs-lisp-intro.texi
127 @c ================ How to Print a Book in Various Sizes ================
129 @c This book can be printed in any of three different sizes.
130 @c In the above header, set @-commands appropriately.
140 @c European A4 size paper:
145 @c ================ How to Typeset and Print ================
147 @c If you do not include PostScript figures, run either of the
148 @c following command sequences, or similar commands suited to your
151 @c texi2dvi emacs-lisp-intro.texi
152 @c lpr -d emacs-lisp-intro.dvi
156 @c tex emacs-lisp-intro.texi
157 @c texindex emacs-lisp-intro.??
158 @c tex emacs-lisp-intro.texi
159 @c lpr -d emacs-lisp-intro.dvi
161 @c If you include the PostScript figures, and you have old software,
162 @c you may need to convert the .dvi file to a .ps file before
163 @c printing. Run either of the following command sequences, or one
166 @c dvips -f < emacs-lisp-intro.dvi > emacs-lisp-intro.ps
170 @c postscript -p < emacs-lisp-intro.dvi > emacs-lisp-intro.ps
173 @c (Note: if you edit the book so as to change the length of the
174 @c table of contents, you may have to change the value of `pageno' below.)
176 @c ================ End of Formatting Sections ================
178 @c For next or subsequent edition:
179 @c create function using with-output-to-temp-buffer
180 @c create a major mode, with keymaps
181 @c run an asynchronous process, like grep or diff
183 @c For 8.5 by 11 inch format: do not use such a small amount of
184 @c whitespace between paragraphs as smallbook format
187 \global\parskip 6pt plus 1pt
191 @c For all sized formats: print within-book cross
192 @c reference with ``...'' rather than [...]
194 @c This works with the texinfo.tex file, version 2003-05-04.08,
195 @c in the Texinfo version 4.6 of the 2003 Jun 13 distribution.
198 \if \xrefprintnodename
199 \global\def\xrefprintnodename#1{\unskip, ``#1''}
201 \global\def\xrefprintnodename#1{ ``#1''}
203 % \global\def\xrefprintnodename#1{, ``#1''}
206 @c ----------------------------------------------------
210 * Emacs Lisp Intro: (eintr).
211 A simple introduction to Emacs Lisp programming.
215 This is an @cite{Introduction to Programming in Emacs Lisp}, for
216 people who are not programmers.
218 Edition @value{edition-number}, @value{update-date}
220 Copyright @copyright{} 1990, 1991, 1992, 1993, 1994, 1995, 1997, 2001,
221 2002, 2003, 2004, 2005, 2006 Free Software Foundation, Inc.
227 GNU Press, @hfill @uref{http://www.gnupress.org}@*
228 a division of the @hfill General: @email{press@@gnu.org}@*
229 Free Software Foundation, Inc. @hfill Orders:@w{ } @email{sales@@gnu.org}@*
230 51 Franklin Street, Fifth Floor @hfill Tel: +1 (617) 542-5942@*
231 Boston, MA 02110-1301 USA @hfill Fax: +1 (617) 542-2652@*
238 GNU Press, Website: http://www.gnupress.org
239 a division of the General: press@@gnu.org
240 Free Software Foundation, Inc. Orders: sales@@gnu.org
241 51 Franklin Street, Fifth Floor Tel: +1 (617) 542-5942
242 Boston, MA 02110-1301 USA Fax: +1 (617) 542-2652
247 @c Printed copies are available for $30 each.@*
250 Permission is granted to copy, distribute and/or modify this document
251 under the terms of the GNU Free Documentation License, Version 1.2 or
252 any later version published by the Free Software Foundation; there
253 being no Invariant Section, with the Front-Cover Texts being ``A GNU
254 Manual'', and with the Back-Cover Texts as in (a) below. A copy of
255 the license is included in the section entitled ``GNU Free
256 Documentation License''.
258 (a) The FSF's Back-Cover Text is: ``You have freedom to copy and
259 modify this GNU Manual, like GNU software. Copies published by the
260 Free Software Foundation raise funds for GNU development.''
263 @c half title; two lines here, so do not use `shorttitlepage'
266 \hbox{}\vskip 1.5in \chaprm \centerline{An Introduction to}%
268 {\begingroup\hbox{}\vskip 0.25in \chaprm%
269 \centerline{Programming in Emacs Lisp}%
270 \endgroup\page\hbox{}\page}
275 @center @titlefont{An Introduction to}
277 @center @titlefont{Programming in Emacs Lisp}
279 @center Revised Third Edition
281 @center by Robert J. Chassell
284 @vskip 0pt plus 1filll
290 @evenheading @thispage @| @| @thischapter
291 @oddheading @thissection @| @| @thispage
295 @c Keep T.O.C. short by tightening up for largebook
298 \global\parskip 2pt plus 1pt
299 \global\advance\baselineskip by -1pt
308 @node Top, Preface, (dir), (dir)
309 @top An Introduction to Programming in Emacs Lisp
313 This master menu first lists each chapter and index; then it lists
314 every node in every chapter.
317 @c >>>> Set pageno appropriately <<<<
319 @c The first page of the Preface is a roman numeral; it is the first
320 @c right handed page after the Table of Contents; hence the following
321 @c setting must be for an odd negative number.
328 * Preface:: What to look for.
329 * List Processing:: What is Lisp?
330 * Practicing Evaluation:: Running several programs.
331 * Writing Defuns:: How to write function definitions.
332 * Buffer Walk Through:: Exploring a few buffer-related functions.
333 * More Complex:: A few, even more complex functions.
334 * Narrowing & Widening:: Restricting your and Emacs attention to
336 * car cdr & cons:: Fundamental functions in Lisp.
337 * Cutting & Storing Text:: Removing text and saving it.
338 * List Implementation:: How lists are implemented in the computer.
339 * Yanking:: Pasting stored text.
340 * Loops & Recursion:: How to repeat a process.
341 * Regexp Search:: Regular expression searches.
342 * Counting Words:: A review of repetition and regexps.
343 * Words in a defun:: Counting words in a @code{defun}.
344 * Readying a Graph:: A prototype graph printing function.
345 * Emacs Initialization:: How to write a @file{.emacs} file.
346 * Debugging:: How to run the Emacs Lisp debuggers.
347 * Conclusion:: Now you have the basics.
348 * the-the:: An appendix: how to find reduplicated words.
349 * Kill Ring:: An appendix: how the kill ring works.
350 * Full Graph:: How to create a graph with labelled axes.
351 * Free Software and Free Manuals::
352 * GNU Free Documentation License::
357 --- The Detailed Node Listing ---
361 * Why:: Why learn Emacs Lisp?
362 * On Reading this Text:: Read, gain familiarity, pick up habits....
363 * Who You Are:: For whom this is written.
365 * Note for Novices:: You can read this as a novice.
370 * Lisp Lists:: What are lists?
371 * Run a Program:: Any list in Lisp is a program ready to run.
372 * Making Errors:: Generating an error message.
373 * Names & Definitions:: Names of symbols and function definitions.
374 * Lisp Interpreter:: What the Lisp interpreter does.
375 * Evaluation:: Running a program.
376 * Variables:: Returning a value from a variable.
377 * Arguments:: Passing information to a function.
378 * set & setq:: Setting the value of a variable.
379 * Summary:: The major points.
380 * Error Message Exercises::
384 * Numbers Lists:: List have numbers, other lists, in them.
385 * Lisp Atoms:: Elemental entities.
386 * Whitespace in Lists:: Formatting lists to be readable.
387 * Typing Lists:: How GNU Emacs helps you type lists.
391 * Complications:: Variables, Special forms, Lists within.
392 * Byte Compiling:: Specially processing code for speed.
396 * How the Interpreter Acts:: Returns and Side Effects...
397 * Evaluating Inner Lists:: Lists within lists...
401 * fill-column Example::
402 * Void Function:: The error message for a symbol
404 * Void Variable:: The error message for a symbol without a value.
408 * Data types:: Types of data passed to a function.
409 * Args as Variable or List:: An argument can be the value
410 of a variable or list.
411 * Variable Number of Arguments:: Some functions may take a
412 variable number of arguments.
413 * Wrong Type of Argument:: Passing an argument of the wrong type
415 * message:: A useful function for sending messages.
417 Setting the Value of a Variable
419 * Using set:: Setting values.
420 * Using setq:: Setting a quoted value.
421 * Counting:: Using @code{setq} to count.
423 Practicing Evaluation
425 * How to Evaluate:: Typing editing commands or @kbd{C-x C-e}
427 * Buffer Names:: Buffers and files are different.
428 * Getting Buffers:: Getting a buffer itself, not merely its name.
429 * Switching Buffers:: How to change to another buffer.
430 * Buffer Size & Locations:: Where point is located and the size of
432 * Evaluation Exercise::
434 How To Write Function Definitions
436 * Primitive Functions::
437 * defun:: The @code{defun} special form.
438 * Install:: Install a function definition.
439 * Interactive:: Making a function interactive.
440 * Interactive Options:: Different options for @code{interactive}.
441 * Permanent Installation:: Installing code permanently.
442 * let:: Creating and initializing local variables.
444 * else:: If--then--else expressions.
445 * Truth & Falsehood:: What Lisp considers false and true.
446 * save-excursion:: Keeping track of point, mark, and buffer.
450 Install a Function Definition
452 * Effect of installation::
453 * Change a defun:: How to change a function definition.
455 Make a Function Interactive
457 * Interactive multiply-by-seven:: An overview.
458 * multiply-by-seven in detail:: The interactive version.
462 * Prevent confusion::
463 * Parts of let Expression::
464 * Sample let Expression::
465 * Uninitialized let Variables::
467 The @code{if} Special Form
469 * if in more detail::
470 * type-of-animal in detail:: An example of an @code{if} expression.
472 Truth and Falsehood in Emacs Lisp
474 * nil explained:: @code{nil} has two meanings.
476 @code{save-excursion}
478 * Point and mark:: A review of various locations.
479 * Template for save-excursion::
481 A Few Buffer--Related Functions
483 * Finding More:: How to find more information.
484 * simplified-beginning-of-buffer:: Shows @code{goto-char},
485 @code{point-min}, and @code{push-mark}.
486 * mark-whole-buffer:: Almost the same as @code{beginning-of-buffer}.
487 * append-to-buffer:: Uses @code{save-excursion} and
488 @code{insert-buffer-substring}.
489 * Buffer Related Review:: Review.
492 The Definition of @code{mark-whole-buffer}
494 * mark-whole-buffer overview::
495 * Body of mark-whole-buffer:: Only three lines of code.
497 The Definition of @code{append-to-buffer}
499 * append-to-buffer overview::
500 * append interactive:: A two part interactive expression.
501 * append-to-buffer body:: Incorporates a @code{let} expression.
502 * append save-excursion:: How the @code{save-excursion} works.
504 A Few More Complex Functions
506 * copy-to-buffer:: With @code{set-buffer}, @code{get-buffer-create}.
507 * insert-buffer:: Read-only, and with @code{or}.
508 * beginning-of-buffer:: Shows @code{goto-char},
509 @code{point-min}, and @code{push-mark}.
510 * Second Buffer Related Review::
511 * optional Exercise::
513 The Definition of @code{insert-buffer}
515 * insert-buffer code::
516 * insert-buffer interactive:: When you can read, but not write.
517 * insert-buffer body:: The body has an @code{or} and a @code{let}.
518 * if & or:: Using an @code{if} instead of an @code{or}.
519 * Insert or:: How the @code{or} expression works.
520 * Insert let:: Two @code{save-excursion} expressions.
521 * New insert-buffer::
523 The Interactive Expression in @code{insert-buffer}
525 * Read-only buffer:: When a buffer cannot be modified.
526 * b for interactive:: An existing buffer or else its name.
528 Complete Definition of @code{beginning-of-buffer}
530 * Optional Arguments::
531 * beginning-of-buffer opt arg:: Example with optional argument.
532 * beginning-of-buffer complete::
534 @code{beginning-of-buffer} with an Argument
536 * Disentangle beginning-of-buffer::
537 * Large buffer case::
538 * Small buffer case::
540 Narrowing and Widening
542 * Narrowing advantages:: The advantages of narrowing
543 * save-restriction:: The @code{save-restriction} special form.
544 * what-line:: The number of the line that point is on.
547 @code{car}, @code{cdr}, @code{cons}: Fundamental Functions
549 * Strange Names:: An historical aside: why the strange names?
550 * car & cdr:: Functions for extracting part of a list.
551 * cons:: Constructing a list.
552 * nthcdr:: Calling @code{cdr} repeatedly.
554 * setcar:: Changing the first element of a list.
555 * setcdr:: Changing the rest of a list.
561 * length:: How to find the length of a list.
563 Cutting and Storing Text
565 * Storing Text:: Text is stored in a list.
566 * zap-to-char:: Cutting out text up to a character.
567 * kill-region:: Cutting text out of a region.
568 * copy-region-as-kill:: A definition for copying text.
569 * Digression into C:: Minor note on C programming language macros.
570 * defvar:: How to give a variable an initial value.
571 * cons & search-fwd Review::
576 * Complete zap-to-char:: The complete implementation.
577 * zap-to-char interactive:: A three part interactive expression.
578 * zap-to-char body:: A short overview.
579 * search-forward:: How to search for a string.
580 * progn:: The @code{progn} special form.
581 * Summing up zap-to-char:: Using @code{point} and @code{search-forward}.
585 * Complete kill-region:: The function definition.
586 * condition-case:: Dealing with a problem.
589 @code{copy-region-as-kill}
591 * Complete copy-region-as-kill:: The complete function definition.
592 * copy-region-as-kill body:: The body of @code{copy-region-as-kill}.
594 The Body of @code{copy-region-as-kill}
596 * last-command & this-command::
597 * kill-append function::
598 * kill-new function::
600 Initializing a Variable with @code{defvar}
602 * See variable current value::
603 * defvar and asterisk::
605 How Lists are Implemented
608 * Symbols as Chest:: Exploring a powerful metaphor.
613 * Kill Ring Overview::
614 * kill-ring-yank-pointer:: The kill ring is a list.
615 * yank nthcdr Exercises:: The @code{kill-ring-yank-pointer} variable.
619 * while:: Causing a stretch of code to repeat.
621 * Recursion:: Causing a function to call itself.
626 * Looping with while:: Repeat so long as test returns true.
627 * Loop Example:: A @code{while} loop that uses a list.
628 * print-elements-of-list:: Uses @code{while}, @code{car}, @code{cdr}.
629 * Incrementing Loop:: A loop with an incrementing counter.
630 * Incrementing Loop Details::
631 * Decrementing Loop:: A loop with a decrementing counter.
633 Details of an Incrementing Loop
635 * Incrementing Example:: Counting pebbles in a triangle.
636 * Inc Example parts:: The parts of the function definition.
637 * Inc Example altogether:: Putting the function definition together.
639 Loop with a Decrementing Counter
641 * Decrementing Example:: More pebbles on the beach.
642 * Dec Example parts:: The parts of the function definition.
643 * Dec Example altogether:: Putting the function definition together.
645 Save your time: @code{dolist} and @code{dotimes}
652 * Building Robots:: Same model, different serial number ...
653 * Recursive Definition Parts:: Walk until you stop ...
654 * Recursion with list:: Using a list as the test whether to recurse.
655 * Recursive triangle function::
656 * Recursion with cond::
657 * Recursive Patterns:: Often used templates.
658 * No Deferment:: Don't store up work ...
659 * No deferment solution::
661 Recursion in Place of a Counter
663 * Recursive Example arg of 1 or 2::
664 * Recursive Example arg of 3 or 4::
672 Regular Expression Searches
674 * sentence-end:: The regular expression for @code{sentence-end}.
675 * re-search-forward:: Very similar to @code{search-forward}.
676 * forward-sentence:: A straightforward example of regexp search.
677 * forward-paragraph:: A somewhat complex example.
678 * etags:: How to create your own @file{TAGS} table.
680 * re-search Exercises::
682 @code{forward-sentence}
684 * Complete forward-sentence::
685 * fwd-sentence while loops:: Two @code{while} loops.
686 * fwd-sentence re-search:: A regular expression search.
688 @code{forward-paragraph}: a Goldmine of Functions
690 * forward-paragraph in brief:: Key parts of the function definition.
691 * fwd-para let:: The @code{let*} expression.
692 * fwd-para while:: The forward motion @code{while} loop.
694 Counting: Repetition and Regexps
697 * count-words-region:: Use a regexp, but find a problem.
698 * recursive-count-words:: Start with case of no words in region.
699 * Counting Exercise::
701 The @code{count-words-region} Function
703 * Design count-words-region:: The definition using a @code{while} loop.
704 * Whitespace Bug:: The Whitespace Bug in @code{count-words-region}.
706 Counting Words in a @code{defun}
708 * Divide and Conquer::
709 * Words and Symbols:: What to count?
710 * Syntax:: What constitutes a word or symbol?
711 * count-words-in-defun:: Very like @code{count-words}.
712 * Several defuns:: Counting several defuns in a file.
713 * Find a File:: Do you want to look at a file?
714 * lengths-list-file:: A list of the lengths of many definitions.
715 * Several files:: Counting in definitions in different files.
716 * Several files recursively:: Recursively counting in different files.
717 * Prepare the data:: Prepare the data for display in a graph.
719 Count Words in @code{defuns} in Different Files
721 * lengths-list-many-files:: Return a list of the lengths of defuns.
722 * append:: Attach one list to another.
724 Prepare the Data for Display in a Graph
726 * Data for Display in Detail::
727 * Sorting:: Sorting lists.
728 * Files List:: Making a list of files.
729 * Counting function definitions::
733 * Columns of a graph::
734 * graph-body-print:: How to print the body of a graph.
735 * recursive-graph-body-print::
737 * Line Graph Exercise::
739 Your @file{.emacs} File
741 * Default Configuration::
742 * Site-wide Init:: You can write site-wide init files.
743 * defcustom:: Emacs will write code for you.
744 * Beginning a .emacs File:: How to write a @code{.emacs file}.
745 * Text and Auto-fill:: Automatically wrap lines.
746 * Mail Aliases:: Use abbreviations for email addresses.
747 * Indent Tabs Mode:: Don't use tabs with @TeX{}
748 * Keybindings:: Create some personal keybindings.
749 * Keymaps:: More about key binding.
750 * Loading Files:: Load (i.e., evaluate) files automatically.
751 * Autoload:: Make functions available.
752 * Simple Extension:: Define a function; bind it to a key.
753 * X11 Colors:: Colors in X.
755 * Mode Line:: How to customize your mode line.
759 * debug:: How to use the built-in debugger.
760 * debug-on-entry:: Start debugging when you call a function.
761 * debug-on-quit:: Start debugging when you quit with @kbd{C-g}.
762 * edebug:: How to use Edebug, a source level debugger.
763 * Debugging Exercises::
765 Handling the Kill Ring
767 * What the Kill Ring Does::
769 * yank:: Paste a copy of a clipped element.
770 * yank-pop:: Insert element pointed to.
773 The @code{current-kill} Function
775 * Understanding current-kill::
777 @code{current-kill} in Outline
779 * Body of current-kill::
780 * Digression concerning error:: How to mislead humans, but not computers.
781 * Determining the Element::
783 A Graph with Labelled Axes
786 * print-graph Varlist:: @code{let} expression in @code{print-graph}.
787 * print-Y-axis:: Print a label for the vertical axis.
788 * print-X-axis:: Print a horizontal label.
789 * Print Whole Graph:: The function to print a complete graph.
791 The @code{print-Y-axis} Function
793 * print-Y-axis in Detail::
794 * Height of label:: What height for the Y axis?
795 * Compute a Remainder:: How to compute the remainder of a division.
796 * Y Axis Element:: Construct a line for the Y axis.
797 * Y-axis-column:: Generate a list of Y axis labels.
798 * print-Y-axis Penultimate:: A not quite final version.
800 The @code{print-X-axis} Function
802 * Similarities differences:: Much like @code{print-Y-axis}, but not exactly.
803 * X Axis Tic Marks:: Create tic marks for the horizontal axis.
805 Printing the Whole Graph
807 * The final version:: A few changes.
808 * Test print-graph:: Run a short test.
809 * Graphing words in defuns:: Executing the final code.
810 * lambda:: How to write an anonymous function.
811 * mapcar:: Apply a function to elements of a list.
812 * Another Bug:: Yet another bug @dots{} most insidious.
813 * Final printed graph:: The graph itself!
818 @node Preface, List Processing, Top, Top
819 @comment node-name, next, previous, up
822 Most of the GNU Emacs integrated environment is written in the programming
823 language called Emacs Lisp. The code written in this programming
824 language is the software---the sets of instructions---that tell the
825 computer what to do when you give it commands. Emacs is designed so
826 that you can write new code in Emacs Lisp and easily install it as an
827 extension to the editor.
829 (GNU Emacs is sometimes called an ``extensible editor'', but it does
830 much more than provide editing capabilities. It is better to refer to
831 Emacs as an ``extensible computing environment''. However, that
832 phrase is quite a mouthful. It is easier to refer to Emacs simply as
833 an editor. Moreover, everything you do in Emacs---find the Mayan date
834 and phases of the moon, simplify polynomials, debug code, manage
835 files, read letters, write books---all these activities are kinds of
836 editing in the most general sense of the word.)
839 * Why:: Why learn Emacs Lisp?
840 * On Reading this Text:: Read, gain familiarity, pick up habits....
841 * Who You Are:: For whom this is written.
843 * Note for Novices:: You can read this as a novice.
847 @node Why, On Reading this Text, Preface, Preface
849 @unnumberedsec Why Study Emacs Lisp?
852 Although Emacs Lisp is usually thought of in association only with Emacs,
853 it is a full computer programming language. You can use Emacs Lisp as
854 you would any other programming language.
856 Perhaps you want to understand programming; perhaps you want to extend
857 Emacs; or perhaps you want to become a programmer. This introduction to
858 Emacs Lisp is designed to get you started: to guide you in learning the
859 fundamentals of programming, and more importantly, to show you how you
860 can teach yourself to go further.
862 @node On Reading this Text, Who You Are, Why, Preface
863 @comment node-name, next, previous, up
864 @unnumberedsec On Reading this Text
866 All through this document, you will see little sample programs you can
867 run inside of Emacs. If you read this document in Info inside of GNU
868 Emacs, you can run the programs as they appear. (This is easy to do and
869 is explained when the examples are presented.) Alternatively, you can
870 read this introduction as a printed book while sitting beside a computer
871 running Emacs. (This is what I like to do; I like printed books.) If
872 you don't have a running Emacs beside you, you can still read this book,
873 but in this case, it is best to treat it as a novel or as a travel guide
874 to a country not yet visited: interesting, but not the same as being
877 Much of this introduction is dedicated to walk-throughs or guided tours
878 of code used in GNU Emacs. These tours are designed for two purposes:
879 first, to give you familiarity with real, working code (code you use
880 every day); and, second, to give you familiarity with the way Emacs
881 works. It is interesting to see how a working environment is
884 hope that you will pick up the habit of browsing through source code.
885 You can learn from it and mine it for ideas. Having GNU Emacs is like
886 having a dragon's cave of treasures.
888 In addition to learning about Emacs as an editor and Emacs Lisp as a
889 programming language, the examples and guided tours will give you an
890 opportunity to get acquainted with Emacs as a Lisp programming
891 environment. GNU Emacs supports programming and provides tools that
892 you will want to become comfortable using, such as @kbd{M-.} (the key
893 which invokes the @code{find-tag} command). You will also learn about
894 buffers and other objects that are part of the environment.
895 Learning about these features of Emacs is like learning new routes
896 around your home town.
899 In addition, I have written several programs as extended examples.
900 Although these are examples, the programs are real. I use them.
901 Other people use them. You may use them. Beyond the fragments of
902 programs used for illustrations, there is very little in here that is
903 `just for teaching purposes'; what you see is used. This is a great
904 advantage of Emacs Lisp: it is easy to learn to use it for work.
907 Finally, I hope to convey some of the skills for using Emacs to
908 learn aspects of programming that you don't know. You can often use
909 Emacs to help you understand what puzzles you or to find out how to do
910 something new. This self-reliance is not only a pleasure, but an
913 @node Who You Are, Lisp History, On Reading this Text, Preface
914 @comment node-name, next, previous, up
915 @unnumberedsec For Whom This is Written
917 This text is written as an elementary introduction for people who are
918 not programmers. If you are a programmer, you may not be satisfied with
919 this primer. The reason is that you may have become expert at reading
920 reference manuals and be put off by the way this text is organized.
922 An expert programmer who reviewed this text said to me:
925 @i{I prefer to learn from reference manuals. I ``dive into'' each
926 paragraph, and ``come up for air'' between paragraphs.}
928 @i{When I get to the end of a paragraph, I assume that that subject is
929 done, finished, that I know everything I need (with the
930 possible exception of the case when the next paragraph starts talking
931 about it in more detail). I expect that a well written reference manual
932 will not have a lot of redundancy, and that it will have excellent
933 pointers to the (one) place where the information I want is.}
936 This introduction is not written for this person!
938 Firstly, I try to say everything at least three times: first, to
939 introduce it; second, to show it in context; and third, to show it in a
940 different context, or to review it.
942 Secondly, I hardly ever put all the information about a subject in one
943 place, much less in one paragraph. To my way of thinking, that imposes
944 too heavy a burden on the reader. Instead I try to explain only what
945 you need to know at the time. (Sometimes I include a little extra
946 information so you won't be surprised later when the additional
947 information is formally introduced.)
949 When you read this text, you are not expected to learn everything the
950 first time. Frequently, you need only make, as it were, a `nodding
951 acquaintance' with some of the items mentioned. My hope is that I have
952 structured the text and given you enough hints that you will be alert to
953 what is important, and concentrate on it.
955 You will need to ``dive into'' some paragraphs; there is no other way
956 to read them. But I have tried to keep down the number of such
957 paragraphs. This book is intended as an approachable hill, rather than
958 as a daunting mountain.
960 This introduction to @cite{Programming in Emacs Lisp} has a companion
963 @cite{The GNU Emacs Lisp Reference Manual}.
966 @ref{Top, , The GNU Emacs Lisp Reference Manual, elisp, The GNU
967 Emacs Lisp Reference Manual}.
969 The reference manual has more detail than this introduction. In the
970 reference manual, all the information about one topic is concentrated
971 in one place. You should turn to it if you are like the programmer
972 quoted above. And, of course, after you have read this
973 @cite{Introduction}, you will find the @cite{Reference Manual} useful
974 when you are writing your own programs.
976 @node Lisp History, Note for Novices, Who You Are, Preface
977 @unnumberedsec Lisp History
980 Lisp was first developed in the late 1950s at the Massachusetts
981 Institute of Technology for research in artificial intelligence. The
982 great power of the Lisp language makes it superior for other purposes as
983 well, such as writing editor commands and integrated environments.
987 GNU Emacs Lisp is largely inspired by Maclisp, which was written at MIT
988 in the 1960s. It is somewhat inspired by Common Lisp, which became a
989 standard in the 1980s. However, Emacs Lisp is much simpler than Common
990 Lisp. (The standard Emacs distribution contains an optional extensions
991 file, @file{cl.el}, that adds many Common Lisp features to Emacs Lisp.)
993 @node Note for Novices, Thank You, Lisp History, Preface
994 @comment node-name, next, previous, up
995 @unnumberedsec A Note for Novices
997 If you don't know GNU Emacs, you can still read this document
998 profitably. However, I recommend you learn Emacs, if only to learn to
999 move around your computer screen. You can teach yourself how to use
1000 Emacs with the on-line tutorial. To use it, type @kbd{C-h t}. (This
1001 means you press and release the @key{CTRL} key and the @kbd{h} at the
1002 same time, and then press and release @kbd{t}.)
1004 Also, I often refer to one of Emacs' standard commands by listing the
1005 keys which you press to invoke the command and then giving the name of
1006 the command in parentheses, like this: @kbd{M-C-\}
1007 (@code{indent-region}). What this means is that the
1008 @code{indent-region} command is customarily invoked by typing
1009 @kbd{M-C-\}. (You can, if you wish, change the keys that are typed to
1010 invoke the command; this is called @dfn{rebinding}. @xref{Keymaps, ,
1011 Keymaps}.) The abbreviation @kbd{M-C-\} means that you type your
1012 @key{META} key, @key{CTRL} key and @key{\} key all at the same time.
1013 (On many modern keyboards the @key{META} key is labelled
1015 Sometimes a combination like this is called a keychord, since it is
1016 similar to the way you play a chord on a piano. If your keyboard does
1017 not have a @key{META} key, the @key{ESC} key prefix is used in place
1018 of it. In this case, @kbd{M-C-\} means that you press and release your
1019 @key{ESC} key and then type the @key{CTRL} key and the @key{\} key at
1020 the same time. But usually @kbd{M-C-\} means press the @key{CTRL} key
1021 along with the key that is labelled @key{ALT} and, at the same time,
1022 press the @key{\} key.
1024 In addition to typing a lone keychord, you can prefix what you type
1025 with @kbd{C-u}, which is called the `universal argument'. The
1026 @kbd{C-u} keychord passes an argument to the subsequent command.
1027 Thus, to indent a region of plain text by 6 spaces, mark the region,
1028 and then type @w{@kbd{C-u 6 M-C-\}}. (If you do not specify a number,
1029 Emacs either passes the number 4 to the command or otherwise runs the
1030 command differently than it would otherwise.) @xref{Arguments, ,
1031 Numeric Arguments, emacs, The GNU Emacs Manual}.
1033 If you are reading this in Info using GNU Emacs, you can read through
1034 this whole document just by pressing the space bar, @key{SPC}.
1035 (To learn about Info, type @kbd{C-h i} and then select Info.)
1037 A note on terminology: when I use the word Lisp alone, I often am
1038 referring to the various dialects of Lisp in general, but when I speak
1039 of Emacs Lisp, I am referring to GNU Emacs Lisp in particular.
1041 @node Thank You, , Note for Novices, Preface
1042 @comment node-name, next, previous, up
1043 @unnumberedsec Thank You
1045 My thanks to all who helped me with this book. My especial thanks to
1046 @r{Jim Blandy}, @r{Noah Friedman}, @w{Jim Kingdon}, @r{Roland
1047 McGrath}, @w{Frank Ritter}, @w{Randy Smith}, @w{Richard M.@:
1048 Stallman}, and @w{Melissa Weisshaus}. My thanks also go to both
1049 @w{Philip Johnson} and @w{David Stampe} for their patient
1050 encouragement. My mistakes are my own.
1056 @c ================ Beginning of main text ================
1058 @c Start main text on right-hand (verso) page
1061 \par\vfill\supereject
1064 \par\vfill\supereject
1066 \par\vfill\supereject
1068 \par\vfill\supereject
1074 @evenheading @thispage @| @| @thischapter
1075 @oddheading @thissection @| @| @thispage
1079 @node List Processing, Practicing Evaluation, Preface, Top
1080 @comment node-name, next, previous, up
1081 @chapter List Processing
1083 To the untutored eye, Lisp is a strange programming language. In Lisp
1084 code there are parentheses everywhere. Some people even claim that
1085 the name stands for `Lots of Isolated Silly Parentheses'. But the
1086 claim is unwarranted. Lisp stands for LISt Processing, and the
1087 programming language handles @emph{lists} (and lists of lists) by
1088 putting them between parentheses. The parentheses mark the boundaries
1089 of the list. Sometimes a list is preceded by a single apostrophe or
1090 quotation mark, @samp{'}@footnote{The single apostrophe or quotation
1091 mark is an abbreviation for the function @code{quote}; you need not
1092 think about functions now; functions are defined in @ref{Making
1093 Errors, , Generate an Error Message}.} Lists are the basis of Lisp.
1096 * Lisp Lists:: What are lists?
1097 * Run a Program:: Any list in Lisp is a program ready to run.
1098 * Making Errors:: Generating an error message.
1099 * Names & Definitions:: Names of symbols and function definitions.
1100 * Lisp Interpreter:: What the Lisp interpreter does.
1101 * Evaluation:: Running a program.
1102 * Variables:: Returning a value from a variable.
1103 * Arguments:: Passing information to a function.
1104 * set & setq:: Setting the value of a variable.
1105 * Summary:: The major points.
1106 * Error Message Exercises::
1109 @node Lisp Lists, Run a Program, List Processing, List Processing
1110 @comment node-name, next, previous, up
1114 In Lisp, a list looks like this: @code{'(rose violet daisy buttercup)}.
1115 This list is preceded by a single apostrophe. It could just as well be
1116 written as follows, which looks more like the kind of list you are likely
1117 to be familiar with:
1129 The elements of this list are the names of the four different flowers,
1130 separated from each other by whitespace and surrounded by parentheses,
1131 like flowers in a field with a stone wall around them.
1132 @cindex Flowers in a field
1135 * Numbers Lists:: List have numbers, other lists, in them.
1136 * Lisp Atoms:: Elemental entities.
1137 * Whitespace in Lists:: Formatting lists to be readable.
1138 * Typing Lists:: How GNU Emacs helps you type lists.
1141 @node Numbers Lists, Lisp Atoms, Lisp Lists, Lisp Lists
1143 @unnumberedsubsec Numbers, Lists inside of Lists
1146 Lists can also have numbers in them, as in this list: @code{(+ 2 2)}.
1147 This list has a plus-sign, @samp{+}, followed by two @samp{2}s, each
1148 separated by whitespace.
1150 In Lisp, both data and programs are represented the same way; that is,
1151 they are both lists of words, numbers, or other lists, separated by
1152 whitespace and surrounded by parentheses. (Since a program looks like
1153 data, one program may easily serve as data for another; this is a very
1154 powerful feature of Lisp.) (Incidentally, these two parenthetical
1155 remarks are @emph{not} Lisp lists, because they contain @samp{;} and
1156 @samp{.} as punctuation marks.)
1159 Here is another list, this time with a list inside of it:
1162 '(this list has (a list inside of it))
1165 The components of this list are the words @samp{this}, @samp{list},
1166 @samp{has}, and the list @samp{(a list inside of it)}. The interior
1167 list is made up of the words @samp{a}, @samp{list}, @samp{inside},
1168 @samp{of}, @samp{it}.
1170 @node Lisp Atoms, Whitespace in Lists, Numbers Lists, Lisp Lists
1171 @comment node-name, next, previous, up
1172 @subsection Lisp Atoms
1175 In Lisp, what we have been calling words are called @dfn{atoms}. This
1176 term comes from the historical meaning of the word atom, which means
1177 `indivisible'. As far as Lisp is concerned, the words we have been
1178 using in the lists cannot be divided into any smaller parts and still
1179 mean the same thing as part of a program; likewise with numbers and
1180 single character symbols like @samp{+}. On the other hand, unlike an
1181 ancient atom, a list can be split into parts. (@xref{car cdr & cons,
1182 , @code{car} @code{cdr} & @code{cons} Fundamental Functions}.)
1184 In a list, atoms are separated from each other by whitespace. They can be
1185 right next to a parenthesis.
1187 @cindex @samp{empty list} defined
1188 Technically speaking, a list in Lisp consists of parentheses surrounding
1189 atoms separated by whitespace or surrounding other lists or surrounding
1190 both atoms and other lists. A list can have just one atom in it or
1191 have nothing in it at all. A list with nothing in it looks like this:
1192 @code{()}, and is called the @dfn{empty list}. Unlike anything else, an
1193 empty list is considered both an atom and a list at the same time.
1195 @cindex Symbolic expressions, introduced
1196 @cindex @samp{expression} defined
1197 @cindex @samp{form} defined
1198 The printed representation of both atoms and lists are called
1199 @dfn{symbolic expressions} or, more concisely, @dfn{s-expressions}.
1200 The word @dfn{expression} by itself can refer to either the printed
1201 representation, or to the atom or list as it is held internally in the
1202 computer. Often, people use the term @dfn{expression}
1203 indiscriminately. (Also, in many texts, the word @dfn{form} is used
1204 as a synonym for expression.)
1206 Incidentally, the atoms that make up our universe were named such when
1207 they were thought to be indivisible; but it has been found that physical
1208 atoms are not indivisible. Parts can split off an atom or it can
1209 fission into two parts of roughly equal size. Physical atoms were named
1210 prematurely, before their truer nature was found. In Lisp, certain
1211 kinds of atom, such as an array, can be separated into parts; but the
1212 mechanism for doing this is different from the mechanism for splitting a
1213 list. As far as list operations are concerned, the atoms of a list are
1216 As in English, the meanings of the component letters of a Lisp atom
1217 are different from the meaning the letters make as a word. For
1218 example, the word for the South American sloth, the @samp{ai}, is
1219 completely different from the two words, @samp{a}, and @samp{i}.
1221 There are many kinds of atom in nature but only a few in Lisp: for
1222 example, @dfn{numbers}, such as 37, 511, or 1729, and @dfn{symbols}, such
1223 as @samp{+}, @samp{foo}, or @samp{forward-line}. The words we have
1224 listed in the examples above are all symbols. In everyday Lisp
1225 conversation, the word ``atom'' is not often used, because programmers
1226 usually try to be more specific about what kind of atom they are dealing
1227 with. Lisp programming is mostly about symbols (and sometimes numbers)
1228 within lists. (Incidentally, the preceding three word parenthetical
1229 remark is a proper list in Lisp, since it consists of atoms, which in
1230 this case are symbols, separated by whitespace and enclosed by
1231 parentheses, without any non-Lisp punctuation.)
1234 In addition, text between double quotation marks---even sentences or
1235 paragraphs---is an atom. Here is an example:
1236 @cindex Text between double quotation marks
1239 '(this list includes "text between quotation marks.")
1242 @cindex @samp{string} defined
1244 In Lisp, all of the quoted text including the punctuation mark and the
1245 blank spaces is a single atom. This kind of atom is called a
1246 @dfn{string} (for `string of characters') and is the sort of thing that
1247 is used for messages that a computer can print for a human to read.
1248 Strings are a different kind of atom than numbers or symbols and are
1251 @node Whitespace in Lists, Typing Lists, Lisp Atoms, Lisp Lists
1252 @comment node-name, next, previous, up
1253 @subsection Whitespace in Lists
1254 @cindex Whitespace in lists
1257 The amount of whitespace in a list does not matter. From the point of view
1258 of the Lisp language,
1269 is exactly the same as this:
1272 '(this list looks like this)
1275 Both examples show what to Lisp is the same list, the list made up of
1276 the symbols @samp{this}, @samp{list}, @samp{looks}, @samp{like}, and
1277 @samp{this} in that order.
1279 Extra whitespace and newlines are designed to make a list more readable
1280 by humans. When Lisp reads the expression, it gets rid of all the extra
1281 whitespace (but it needs to have at least one space between atoms in
1282 order to tell them apart.)
1284 Odd as it seems, the examples we have seen cover almost all of what Lisp
1285 lists look like! Every other list in Lisp looks more or less like one
1286 of these examples, except that the list may be longer and more complex.
1287 In brief, a list is between parentheses, a string is between quotation
1288 marks, a symbol looks like a word, and a number looks like a number.
1289 (For certain situations, square brackets, dots and a few other special
1290 characters may be used; however, we will go quite far without them.)
1292 @node Typing Lists, , Whitespace in Lists, Lisp Lists
1293 @comment node-name, next, previous, up
1294 @subsection GNU Emacs Helps You Type Lists
1295 @cindex Help typing lists
1296 @cindex Formatting help
1298 When you type a Lisp expression in GNU Emacs using either Lisp
1299 Interaction mode or Emacs Lisp mode, you have available to you several
1300 commands to format the Lisp expression so it is easy to read. For
1301 example, pressing the @key{TAB} key automatically indents the line the
1302 cursor is on by the right amount. A command to properly indent the
1303 code in a region is customarily bound to @kbd{M-C-\}. Indentation is
1304 designed so that you can see which elements of a list belong to which
1305 list---elements of a sub-list are indented more than the elements of
1308 In addition, when you type a closing parenthesis, Emacs momentarily
1309 jumps the cursor back to the matching opening parenthesis, so you can
1310 see which one it is. This is very useful, since every list you type
1311 in Lisp must have its closing parenthesis match its opening
1312 parenthesis. (@xref{Major Modes, , Major Modes, emacs, The GNU Emacs
1313 Manual}, for more information about Emacs' modes.)
1315 @node Run a Program, Making Errors, Lisp Lists, List Processing
1316 @comment node-name, next, previous, up
1317 @section Run a Program
1318 @cindex Run a program
1319 @cindex Program, running one
1321 @cindex @samp{evaluate} defined
1322 A list in Lisp---any list---is a program ready to run. If you run it
1323 (for which the Lisp jargon is @dfn{evaluate}), the computer will do one
1324 of three things: do nothing except return to you the list itself; send
1325 you an error message; or, treat the first symbol in the list as a
1326 command to do something. (Usually, of course, it is the last of these
1327 three things that you really want!)
1329 @c use code for the single apostrophe, not samp.
1330 The single apostrophe, @code{'}, that I put in front of some of the
1331 example lists in preceding sections is called a @dfn{quote}; when it
1332 precedes a list, it tells Lisp to do nothing with the list, other than
1333 take it as it is written. But if there is no quote preceding a list,
1334 the first item of the list is special: it is a command for the computer
1335 to obey. (In Lisp, these commands are called @emph{functions}.) The list
1336 @code{(+ 2 2)} shown above did not have a quote in front of it, so Lisp
1337 understands that the @code{+} is an instruction to do something with the
1338 rest of the list: add the numbers that follow.
1341 If you are reading this inside of GNU Emacs in Info, here is how you can
1342 evaluate such a list: place your cursor immediately after the right
1343 hand parenthesis of the following list and then type @kbd{C-x C-e}:
1349 @c use code for the number four, not samp.
1351 You will see the number @code{4} appear in the echo area. (In the
1352 jargon, what you have just done is ``evaluate the list.'' The echo area
1353 is the line at the bottom of the screen that displays or ``echoes''
1354 text.) Now try the same thing with a quoted list: place the cursor
1355 right after the following list and type @kbd{C-x C-e}:
1358 '(this is a quoted list)
1362 You will see @code{(this is a quoted list)} appear in the echo area.
1364 @cindex Lisp interpreter, explained
1365 @cindex Interpreter, Lisp, explained
1366 In both cases, what you are doing is giving a command to the program
1367 inside of GNU Emacs called the @dfn{Lisp interpreter}---giving the
1368 interpreter a command to evaluate the expression. The name of the Lisp
1369 interpreter comes from the word for the task done by a human who comes
1370 up with the meaning of an expression---who ``interprets'' it.
1372 You can also evaluate an atom that is not part of a list---one that is
1373 not surrounded by parentheses; again, the Lisp interpreter translates
1374 from the humanly readable expression to the language of the computer.
1375 But before discussing this (@pxref{Variables}), we will discuss what the
1376 Lisp interpreter does when you make an error.
1378 @node Making Errors, Names & Definitions, Run a Program, List Processing
1379 @comment node-name, next, previous, up
1380 @section Generate an Error Message
1381 @cindex Generate an error message
1382 @cindex Error message generation
1384 Partly so you won't worry if you do it accidentally, we will now give
1385 a command to the Lisp interpreter that generates an error message.
1386 This is a harmless activity; and indeed, we will often try to generate
1387 error messages intentionally. Once you understand the jargon, error
1388 messages can be informative. Instead of being called ``error''
1389 messages, they should be called ``help'' messages. They are like
1390 signposts to a traveller in a strange country; deciphering them can be
1391 hard, but once understood, they can point the way.
1393 The error message is generated by a built-in GNU Emacs debugger. We
1394 will `enter the debugger'. You get out of the debugger by typing @code{q}.
1396 What we will do is evaluate a list that is not quoted and does not
1397 have a meaningful command as its first element. Here is a list almost
1398 exactly the same as the one we just used, but without the single-quote
1399 in front of it. Position the cursor right after it and type @kbd{C-x
1403 (this is an unquoted list)
1407 What you see depends on which version of Emacs you are running. GNU
1408 Emacs version 22 provides more information than version 20 and before.
1409 First, the more recent result of generating an error; then the
1410 earlier, version 20 result.
1414 In GNU Emacs version 22, a @file{*Backtrace*} window will open up and
1415 you will see the following in it:
1419 ---------- Buffer: *Backtrace* ----------
1420 Debugger entered--Lisp error: (void-function this)
1421 (this is an unquoted list)
1422 eval((this is an unquoted list))
1423 eval-last-sexp-1(nil)
1425 call-interactively(eval-last-sexp)
1426 ---------- Buffer: *Backtrace* ----------
1432 Your cursor will be in this window (you may have to wait a few seconds
1433 before it becomes visible). To quit the debugger and make the
1434 debugger window go away, type:
1441 Please type @kbd{q} right now, so you become confident that you can
1442 get out of the debugger. Then, type @kbd{C-x C-e} again to re-enter
1445 @cindex @samp{function} defined
1446 Based on what we already know, we can almost read this error message.
1448 You read the @file{*Backtrace*} buffer from the bottom up; it tells
1449 you what Emacs did. When you typed @kbd{C-x C-e}, you made an
1450 interactive call to the command @code{eval-last-sexp}. @code{eval} is
1451 an abbreviation for `evaluate' and @code{sexp} is an abbreviation for
1452 `symbolic expression'. The command means `evaluate last symbolic
1453 expression', which is the expression just before your cursor.
1455 Each line above tells you what the Lisp interpreter evaluated next.
1456 The most recent action is at the top. The buffer is called the
1457 @file{*Backtrace*} buffer because it enables you to track Emacs
1461 At the top of the @file{*Backtrace*} buffer, you see the line:
1464 Debugger entered--Lisp error: (void-function this)
1468 The Lisp interpreter tried to evaluate the first atom of the list, the
1469 word @samp{this}. It is this action that generated the error message
1470 @samp{void-function this}.
1472 The message contains the words @samp{void-function} and @samp{this}.
1474 @cindex @samp{function} defined
1475 The word @samp{function} was mentioned once before. It is a very
1476 important word. For our purposes, we can define it by saying that a
1477 @dfn{function} is a set of instructions to the computer that tell the
1478 computer to do something.
1480 Now we can begin to understand the error message: @samp{void-function
1481 this}. The function (that is, the word @samp{this}) does not have a
1482 definition of any set of instructions for the computer to carry out.
1484 The slightly odd word, @samp{void-function}, is designed to cover the
1485 way Emacs Lisp is implemented, which is that when a symbol does not
1486 have a function definition attached to it, the place that should
1487 contain the instructions is `void'.
1489 On the other hand, since we were able to add 2 plus 2 successfully, by
1490 evaluating @code{(+ 2 2)}, we can infer that the symbol @code{+} must
1491 have a set of instructions for the computer to obey and those
1492 instructions must be to add the numbers that follow the @code{+}.
1495 In GNU Emacs version 20, and in earlier versions, you will see only
1496 one line of error message; it will appear in the echo area and look
1500 Symbol's function definition is void:@: this
1504 (Also, your terminal may beep at you---some do, some don't; and others
1505 blink. This is just a device to get your attention.) The message goes
1506 away as soon as you type another key, even just to move the cursor.
1508 We know the meaning of the word @samp{Symbol}. It refers to the first
1509 atom of the list, the word @samp{this}. The word @samp{function}
1510 refers to the instructions that tell the computer what to do.
1511 (Technically, the symbol tells the computer where to find the
1512 instructions, but this is a complication we can ignore for the
1515 The error message can be understood: @samp{Symbol's function
1516 definition is void:@: this}. The symbol (that is, the word
1517 @samp{this}) lacks instructions for the computer to carry out.
1519 @node Names & Definitions, Lisp Interpreter, Making Errors, List Processing
1520 @comment node-name, next, previous, up
1521 @section Symbol Names and Function Definitions
1522 @cindex Symbol names
1524 We can articulate another characteristic of Lisp based on what we have
1525 discussed so far---an important characteristic: a symbol, like
1526 @code{+}, is not itself the set of instructions for the computer to
1527 carry out. Instead, the symbol is used, perhaps temporarily, as a way
1528 of locating the definition or set of instructions. What we see is the
1529 name through which the instructions can be found. Names of people
1530 work the same way. I can be referred to as @samp{Bob}; however, I am
1531 not the letters @samp{B}, @samp{o}, @samp{b} but am, or was, the
1532 consciousness consistently associated with a particular life-form.
1533 The name is not me, but it can be used to refer to me.
1535 In Lisp, one set of instructions can be attached to several names.
1536 For example, the computer instructions for adding numbers can be
1537 linked to the symbol @code{plus} as well as to the symbol @code{+}
1538 (and are in some dialects of Lisp). Among humans, I can be referred
1539 to as @samp{Robert} as well as @samp{Bob} and by other words as well.
1541 On the other hand, a symbol can have only one function definition
1542 attached to it at a time. Otherwise, the computer would be confused as
1543 to which definition to use. If this were the case among people, only
1544 one person in the world could be named @samp{Bob}. However, the function
1545 definition to which the name refers can be changed readily.
1546 (@xref{Install, , Install a Function Definition}.)
1548 Since Emacs Lisp is large, it is customary to name symbols in a way
1549 that identifies the part of Emacs to which the function belongs.
1550 Thus, all the names for functions that deal with Texinfo start with
1551 @samp{texinfo-} and those for functions that deal with reading mail
1552 start with @samp{rmail-}.
1554 @node Lisp Interpreter, Evaluation, Names & Definitions, List Processing
1555 @comment node-name, next, previous, up
1556 @section The Lisp Interpreter
1557 @cindex Lisp interpreter, what it does
1558 @cindex Interpreter, what it does
1560 Based on what we have seen, we can now start to figure out what the
1561 Lisp interpreter does when we command it to evaluate a list.
1562 First, it looks to see whether there is a quote before the list; if
1563 there is, the interpreter just gives us the list. On the other
1564 hand, if there is no quote, the interpreter looks at the first element
1565 in the list and sees whether it has a function definition. If it does,
1566 the interpreter carries out the instructions in the function definition.
1567 Otherwise, the interpreter prints an error message.
1569 This is how Lisp works. Simple. There are added complications which we
1570 will get to in a minute, but these are the fundamentals. Of course, to
1571 write Lisp programs, you need to know how to write function definitions
1572 and attach them to names, and how to do this without confusing either
1573 yourself or the computer.
1576 * Complications:: Variables, Special forms, Lists within.
1577 * Byte Compiling:: Specially processing code for speed.
1580 @node Complications, Byte Compiling, Lisp Interpreter, Lisp Interpreter
1582 @unnumberedsubsec Complications
1585 Now, for the first complication. In addition to lists, the Lisp
1586 interpreter can evaluate a symbol that is not quoted and does not have
1587 parentheses around it. The Lisp interpreter will attempt to determine
1588 the symbol's value as a @dfn{variable}. This situation is described
1589 in the section on variables. (@xref{Variables}.)
1591 @cindex Special form
1592 The second complication occurs because some functions are unusual and do
1593 not work in the usual manner. Those that don't are called @dfn{special
1594 forms}. They are used for special jobs, like defining a function, and
1595 there are not many of them. In the next few chapters, you will be
1596 introduced to several of the more important special forms.
1598 The third and final complication is this: if the function that the
1599 Lisp interpreter is looking at is not a special form, and if it is part
1600 of a list, the Lisp interpreter looks to see whether the list has a list
1601 inside of it. If there is an inner list, the Lisp interpreter first
1602 figures out what it should do with the inside list, and then it works on
1603 the outside list. If there is yet another list embedded inside the
1604 inner list, it works on that one first, and so on. It always works on
1605 the innermost list first. The interpreter works on the innermost list
1606 first, to evaluate the result of that list. The result may be
1607 used by the enclosing expression.
1609 Otherwise, the interpreter works left to right, from one expression to
1612 @node Byte Compiling, , Complications, Lisp Interpreter
1613 @subsection Byte Compiling
1614 @cindex Byte compiling
1616 One other aspect of interpreting: the Lisp interpreter is able to
1617 interpret two kinds of entity: humanly readable code, on which we will
1618 focus exclusively, and specially processed code, called @dfn{byte
1619 compiled} code, which is not humanly readable. Byte compiled code
1620 runs faster than humanly readable code.
1622 You can transform humanly readable code into byte compiled code by
1623 running one of the compile commands such as @code{byte-compile-file}.
1624 Byte compiled code is usually stored in a file that ends with a
1625 @file{.elc} extension rather than a @file{.el} extension. You will
1626 see both kinds of file in the @file{emacs/lisp} directory; the files
1627 to read are those with @file{.el} extensions.
1629 As a practical matter, for most things you might do to customize or
1630 extend Emacs, you do not need to byte compile; and I will not discuss
1631 the topic here. @xref{Byte Compilation, , Byte Compilation, elisp,
1632 The GNU Emacs Lisp Reference Manual}, for a full description of byte
1635 @node Evaluation, Variables, Lisp Interpreter, List Processing
1636 @comment node-name, next, previous, up
1640 When the Lisp interpreter works on an expression, the term for the
1641 activity is called @dfn{evaluation}. We say that the interpreter
1642 `evaluates the expression'. I've used this term several times before.
1643 The word comes from its use in everyday language, `to ascertain the
1644 value or amount of; to appraise', according to @cite{Webster's New
1645 Collegiate Dictionary}.
1648 * How the Interpreter Acts:: Returns and Side Effects...
1649 * Evaluating Inner Lists:: Lists within lists...
1652 @node How the Interpreter Acts, Evaluating Inner Lists, Evaluation, Evaluation
1654 @unnumberedsubsec How the Lisp Interpreter Acts
1657 @cindex @samp{returned value} explained
1658 After evaluating an expression, the Lisp interpreter will most likely
1659 @dfn{return} the value that the computer produces by carrying out the
1660 instructions it found in the function definition, or perhaps it will
1661 give up on that function and produce an error message. (The interpreter
1662 may also find itself tossed, so to speak, to a different function or it
1663 may attempt to repeat continually what it is doing for ever and ever in
1664 what is called an `infinite loop'. These actions are less common; and
1665 we can ignore them.) Most frequently, the interpreter returns a value.
1667 @cindex @samp{side effect} defined
1668 At the same time the interpreter returns a value, it may do something
1669 else as well, such as move a cursor or copy a file; this other kind of
1670 action is called a @dfn{side effect}. Actions that we humans think are
1671 important, such as printing results, are often ``side effects'' to the
1672 Lisp interpreter. The jargon can sound peculiar, but it turns out that
1673 it is fairly easy to learn to use side effects.
1675 In summary, evaluating a symbolic expression most commonly causes the
1676 Lisp interpreter to return a value and perhaps carry out a side effect;
1677 or else produce an error.
1679 @node Evaluating Inner Lists, , How the Interpreter Acts, Evaluation
1680 @comment node-name, next, previous, up
1681 @subsection Evaluating Inner Lists
1682 @cindex Inner list evaluation
1683 @cindex Evaluating inner lists
1685 If evaluation applies to a list that is inside another list, the outer
1686 list may use the value returned by the first evaluation as information
1687 when the outer list is evaluated. This explains why inner expressions
1688 are evaluated first: the values they return are used by the outer
1692 We can investigate this process by evaluating another addition example.
1693 Place your cursor after the following expression and type @kbd{C-x C-e}:
1700 The number 8 will appear in the echo area.
1702 What happens is that the Lisp interpreter first evaluates the inner
1703 expression, @code{(+ 3 3)}, for which the value 6 is returned; then it
1704 evaluates the outer expression as if it were written @code{(+ 2 6)}, which
1705 returns the value 8. Since there are no more enclosing expressions to
1706 evaluate, the interpreter prints that value in the echo area.
1708 Now it is easy to understand the name of the command invoked by the
1709 keystrokes @kbd{C-x C-e}: the name is @code{eval-last-sexp}. The
1710 letters @code{sexp} are an abbreviation for `symbolic expression', and
1711 @code{eval} is an abbreviation for `evaluate'. The command means
1712 `evaluate last symbolic expression'.
1714 As an experiment, you can try evaluating the expression by putting the
1715 cursor at the beginning of the next line immediately following the
1716 expression, or inside the expression.
1719 Here is another copy of the expression:
1726 If you place the cursor at the beginning of the blank line that
1727 immediately follows the expression and type @kbd{C-x C-e}, you will
1728 still get the value 8 printed in the echo area. Now try putting the
1729 cursor inside the expression. If you put it right after the next to
1730 last parenthesis (so it appears to sit on top of the last parenthesis),
1731 you will get a 6 printed in the echo area! This is because the command
1732 evaluates the expression @code{(+ 3 3)}.
1734 Now put the cursor immediately after a number. Type @kbd{C-x C-e} and
1735 you will get the number itself. In Lisp, if you evaluate a number, you
1736 get the number itself---this is how numbers differ from symbols. If you
1737 evaluate a list starting with a symbol like @code{+}, you will get a
1738 value returned that is the result of the computer carrying out the
1739 instructions in the function definition attached to that name. If a
1740 symbol by itself is evaluated, something different happens, as we will
1741 see in the next section.
1743 @node Variables, Arguments, Evaluation, List Processing
1744 @comment node-name, next, previous, up
1748 In Emacs Lisp, a symbol can have a value attached to it just as it can
1749 have a function definition attached to it. The two are different.
1750 The function definition is a set of instructions that a computer will
1751 obey. A value, on the other hand, is something, such as number or a
1752 name, that can vary (which is why such a symbol is called a variable).
1753 The value of a symbol can be any expression in Lisp, such as a symbol,
1754 number, list, or string. A symbol that has a value is often called a
1757 A symbol can have both a function definition and a value attached to
1758 it at the same time. Or it can have just one or the other.
1759 The two are separate. This is somewhat similar
1760 to the way the name Cambridge can refer to the city in Massachusetts
1761 and have some information attached to the name as well, such as
1762 ``great programming center''.
1765 (Incidentally, in Emacs Lisp, a symbol can have two
1766 other things attached to it, too: a property list and a documentation
1767 string; these are discussed later.)
1770 Another way to think about this is to imagine a symbol as being a chest
1771 of drawers. The function definition is put in one drawer, the value in
1772 another, and so on. What is put in the drawer holding the value can be
1773 changed without affecting the contents of the drawer holding the
1774 function definition, and vice-verse.
1777 * fill-column Example::
1778 * Void Function:: The error message for a symbol
1780 * Void Variable:: The error message for a symbol without a value.
1783 @node fill-column Example, Void Function, Variables, Variables
1785 @unnumberedsubsec @code{fill-column}, an Example Variable
1788 @findex fill-column, @r{an example variable}
1789 @cindex Example variable, @code{fill-column}
1790 @cindex Variable, example of, @code{fill-column}
1791 The variable @code{fill-column} illustrates a symbol with a value
1792 attached to it: in every GNU Emacs buffer, this symbol is set to some
1793 value, usually 72 or 70, but sometimes to some other value. To find the
1794 value of this symbol, evaluate it by itself. If you are reading this in
1795 Info inside of GNU Emacs, you can do this by putting the cursor after
1796 the symbol and typing @kbd{C-x C-e}:
1803 After I typed @kbd{C-x C-e}, Emacs printed the number 72 in my echo
1804 area. This is the value for which @code{fill-column} is set for me as I
1805 write this. It may be different for you in your Info buffer. Notice
1806 that the value returned as a variable is printed in exactly the same way
1807 as the value returned by a function carrying out its instructions. From
1808 the point of view of the Lisp interpreter, a value returned is a value
1809 returned. What kind of expression it came from ceases to matter once
1812 A symbol can have any value attached to it or, to use the jargon, we can
1813 @dfn{bind} the variable to a value: to a number, such as 72; to a
1814 string, @code{"such as this"}; to a list, such as @code{(spruce pine
1815 oak)}; we can even bind a variable to a function definition.
1817 A symbol can be bound to a value in several ways. @xref{set & setq, ,
1818 Setting the Value of a Variable}, for information about one way to do
1821 @node Void Function, Void Variable, fill-column Example, Variables
1822 @comment node-name, next, previous, up
1823 @subsection Error Message for a Symbol Without a Function
1824 @cindex Symbol without function error
1825 @cindex Error for symbol without function
1827 When we evaluated @code{fill-column} to find its value as a variable,
1828 we did not place parentheses around the word. This is because we did
1829 not intend to use it as a function name.
1831 If @code{fill-column} were the first or only element of a list, the
1832 Lisp interpreter would attempt to find the function definition
1833 attached to it. But @code{fill-column} has no function definition.
1834 Try evaluating this:
1842 In GNU Emacs version 22, you will create a @file{*Backtrace*} buffer
1847 ---------- Buffer: *Backtrace* ----------
1848 Debugger entered--Lisp error: (void-function fill-column)
1851 eval-last-sexp-1(nil)
1853 call-interactively(eval-last-sexp)
1854 ---------- Buffer: *Backtrace* ----------
1859 (Remember, to quit the debugger and make the debugger window go away,
1860 type @kbd{q} in the @file{*Backtrace*} buffer.)
1864 In GNU Emacs 20 and before, you will produce an error message that says:
1867 Symbol's function definition is void:@: fill-column
1871 (The message will go away as soon as you move the cursor or type
1875 @node Void Variable, , Void Function, Variables
1876 @comment node-name, next, previous, up
1877 @subsection Error Message for a Symbol Without a Value
1878 @cindex Symbol without value error
1879 @cindex Error for symbol without value
1881 If you attempt to evaluate a symbol that does not have a value bound to
1882 it, you will receive an error message. You can see this by
1883 experimenting with our 2 plus 2 addition. In the following expression,
1884 put your cursor right after the @code{+}, before the first number 2,
1893 In GNU Emacs 22, you will create a @file{*Backtrace*} buffer that
1898 ---------- Buffer: *Backtrace* ----------
1899 Debugger entered--Lisp error: (void-variable +)
1901 eval-last-sexp-1(nil)
1903 call-interactively(eval-last-sexp)
1904 ---------- Buffer: *Backtrace* ----------
1909 (As with the other times we entered the debugger, you can quit by
1910 typing @kbd{q} in the @file{*Backtrace*} buffer.)
1912 This backtrace is different from the very first error message we saw,
1913 which said, @samp{Debugger entered--Lisp error: (void-function this)}.
1914 In this case, the function does not have a value as a variable; while
1915 in the other error message, the function (the word `this') did not
1918 In this experiment with the @code{+}, what we did was cause the Lisp
1919 interpreter to evaluate the @code{+} and look for the value of the
1920 variable instead of the function definition. We did this by placing the
1921 cursor right after the symbol rather than after the parenthesis of the
1922 enclosing list as we did before. As a consequence, the Lisp interpreter
1923 evaluated the preceding s-expression, which in this case was the
1926 Since @code{+} does not have a value bound to it, just the function
1927 definition, the error message reported that the symbol's value as a
1932 In GNU Emacs version 20 and before, your error message will say:
1935 Symbol's value as variable is void:@: +
1939 The meaning is the same as in GNU Emacs 22.
1942 @node Arguments, set & setq, Variables, List Processing
1943 @comment node-name, next, previous, up
1946 @cindex Passing information to functions
1948 To see how information is passed to functions, let's look again at
1949 our old standby, the addition of two plus two. In Lisp, this is written
1956 If you evaluate this expression, the number 4 will appear in your echo
1957 area. What the Lisp interpreter does is add the numbers that follow
1960 @cindex @samp{argument} defined
1961 The numbers added by @code{+} are called the @dfn{arguments} of the
1962 function @code{+}. These numbers are the information that is given to
1963 or @dfn{passed} to the function.
1965 The word `argument' comes from the way it is used in mathematics and
1966 does not refer to a disputation between two people; instead it refers to
1967 the information presented to the function, in this case, to the
1968 @code{+}. In Lisp, the arguments to a function are the atoms or lists
1969 that follow the function. The values returned by the evaluation of
1970 these atoms or lists are passed to the function. Different functions
1971 require different numbers of arguments; some functions require none at
1972 all.@footnote{It is curious to track the path by which the word `argument'
1973 came to have two different meanings, one in mathematics and the other in
1974 everyday English. According to the @cite{Oxford English Dictionary},
1975 the word derives from the Latin for @samp{to make clear, prove}; thus it
1976 came to mean, by one thread of derivation, `the evidence offered as
1977 proof', which is to say, `the information offered', which led to its
1978 meaning in Lisp. But in the other thread of derivation, it came to mean
1979 `to assert in a manner against which others may make counter
1980 assertions', which led to the meaning of the word as a disputation.
1981 (Note here that the English word has two different definitions attached
1982 to it at the same time. By contrast, in Emacs Lisp, a symbol cannot
1983 have two different function definitions at the same time.)}
1986 * Data types:: Types of data passed to a function.
1987 * Args as Variable or List:: An argument can be the value
1988 of a variable or list.
1989 * Variable Number of Arguments:: Some functions may take a
1990 variable number of arguments.
1991 * Wrong Type of Argument:: Passing an argument of the wrong type
1993 * message:: A useful function for sending messages.
1996 @node Data types, Args as Variable or List, Arguments, Arguments
1997 @comment node-name, next, previous, up
1998 @subsection Arguments' Data Types
2000 @cindex Types of data
2001 @cindex Arguments' data types
2003 The type of data that should be passed to a function depends on what
2004 kind of information it uses. The arguments to a function such as
2005 @code{+} must have values that are numbers, since @code{+} adds numbers.
2006 Other functions use different kinds of data for their arguments.
2010 For example, the @code{concat} function links together or unites two or
2011 more strings of text to produce a string. The arguments are strings.
2012 Concatenating the two character strings @code{abc}, @code{def} produces
2013 the single string @code{abcdef}. This can be seen by evaluating the
2017 (concat "abc" "def")
2021 The value produced by evaluating this expression is @code{"abcdef"}.
2023 A function such as @code{substring} uses both a string and numbers as
2024 arguments. The function returns a part of the string, a substring of
2025 the first argument. This function takes three arguments. Its first
2026 argument is the string of characters, the second and third arguments are
2027 numbers that indicate the beginning and end of the substring. The
2028 numbers are a count of the number of characters (including spaces and
2029 punctuations) from the beginning of the string.
2032 For example, if you evaluate the following:
2035 (substring "The quick brown fox jumped." 16 19)
2039 you will see @code{"fox"} appear in the echo area. The arguments are the
2040 string and the two numbers.
2042 Note that the string passed to @code{substring} is a single atom even
2043 though it is made up of several words separated by spaces. Lisp counts
2044 everything between the two quotation marks as part of the string,
2045 including the spaces. You can think of the @code{substring} function as
2046 a kind of `atom smasher' since it takes an otherwise indivisible atom
2047 and extracts a part. However, @code{substring} is only able to extract
2048 a substring from an argument that is a string, not from another type of
2049 atom such as a number or symbol.
2051 @node Args as Variable or List, Variable Number of Arguments, Data types, Arguments
2052 @comment node-name, next, previous, up
2053 @subsection An Argument as the Value of a Variable or List
2055 An argument can be a symbol that returns a value when it is evaluated.
2056 For example, when the symbol @code{fill-column} by itself is evaluated,
2057 it returns a number. This number can be used in an addition.
2060 Position the cursor after the following expression and type @kbd{C-x
2068 The value will be a number two more than what you get by evaluating
2069 @code{fill-column} alone. For me, this is 74, because my value of
2070 @code{fill-column} is 72.
2072 As we have just seen, an argument can be a symbol that returns a value
2073 when evaluated. In addition, an argument can be a list that returns a
2074 value when it is evaluated. For example, in the following expression,
2075 the arguments to the function @code{concat} are the strings
2076 @w{@code{"The "}} and @w{@code{" red foxes."}} and the list
2077 @code{(number-to-string (+ 2 fill-column))}.
2079 @c For GNU Emacs 22, need number-to-string
2081 (concat "The " (number-to-string (+ 2 fill-column)) " red foxes.")
2085 If you evaluate this expression---and if, as with my Emacs,
2086 @code{fill-column} evaluates to 72---@code{"The 74 red foxes."} will
2087 appear in the echo area. (Note that you must put spaces after the
2088 word @samp{The} and before the word @samp{red} so they will appear in
2089 the final string. The function @code{number-to-string} converts the
2090 integer that the addition function returns to a string.
2091 @code{number-to-string} is also known as @code{int-to-string}.)
2093 @node Variable Number of Arguments, Wrong Type of Argument, Args as Variable or List, Arguments
2094 @comment node-name, next, previous, up
2095 @subsection Variable Number of Arguments
2096 @cindex Variable number of arguments
2097 @cindex Arguments, variable number of
2099 Some functions, such as @code{concat}, @code{+} or @code{*}, take any
2100 number of arguments. (The @code{*} is the symbol for multiplication.)
2101 This can be seen by evaluating each of the following expressions in
2102 the usual way. What you will see in the echo area is printed in this
2103 text after @samp{@result{}}, which you may read as `evaluates to'.
2106 In the first set, the functions have no arguments:
2117 In this set, the functions have one argument each:
2128 In this set, the functions have three arguments each:
2132 (+ 3 4 5) @result{} 12
2134 (* 3 4 5) @result{} 60
2138 @node Wrong Type of Argument, message, Variable Number of Arguments, Arguments
2139 @comment node-name, next, previous, up
2140 @subsection Using the Wrong Type Object as an Argument
2141 @cindex Wrong type of argument
2142 @cindex Argument, wrong type of
2144 When a function is passed an argument of the wrong type, the Lisp
2145 interpreter produces an error message. For example, the @code{+}
2146 function expects the values of its arguments to be numbers. As an
2147 experiment we can pass it the quoted symbol @code{hello} instead of a
2148 number. Position the cursor after the following expression and type
2156 When you do this you will generate an error message. What has happened
2157 is that @code{+} has tried to add the 2 to the value returned by
2158 @code{'hello}, but the value returned by @code{'hello} is the symbol
2159 @code{hello}, not a number. Only numbers can be added. So @code{+}
2160 could not carry out its addition.
2163 In GNU Emacs version 22, you will create and enter a
2164 @file{*Backtrace*} buffer that says:
2169 ---------- Buffer: *Backtrace* ----------
2170 Debugger entered--Lisp error:
2171 (wrong-type-argument number-or-marker-p hello)
2173 eval((+ 2 (quote hello)))
2174 eval-last-sexp-1(nil)
2176 call-interactively(eval-last-sexp)
2177 ---------- Buffer: *Backtrace* ----------
2182 As usual, the error message tries to be helpful and makes sense after you
2183 learn how to read it.@footnote{@code{(quote hello)} is an expansion of
2184 the abbreviation @code{'hello}.}
2186 The first part of the error message is straightforward; it says
2187 @samp{wrong type argument}. Next comes the mysterious jargon word
2188 @w{@samp{number-or-marker-p}}. This word is trying to tell you what
2189 kind of argument the @code{+} expected.
2191 The symbol @code{number-or-marker-p} says that the Lisp interpreter is
2192 trying to determine whether the information presented it (the value of
2193 the argument) is a number or a marker (a special object representing a
2194 buffer position). What it does is test to see whether the @code{+} is
2195 being given numbers to add. It also tests to see whether the
2196 argument is something called a marker, which is a specific feature of
2197 Emacs Lisp. (In Emacs, locations in a buffer are recorded as markers.
2198 When the mark is set with the @kbd{C-@@} or @kbd{C-@key{SPC}} command,
2199 its position is kept as a marker. The mark can be considered a
2200 number---the number of characters the location is from the beginning
2201 of the buffer.) In Emacs Lisp, @code{+} can be used to add the
2202 numeric value of marker positions as numbers.
2204 The @samp{p} of @code{number-or-marker-p} is the embodiment of a
2205 practice started in the early days of Lisp programming. The @samp{p}
2206 stands for `predicate'. In the jargon used by the early Lisp
2207 researchers, a predicate refers to a function to determine whether some
2208 property is true or false. So the @samp{p} tells us that
2209 @code{number-or-marker-p} is the name of a function that determines
2210 whether it is true or false that the argument supplied is a number or
2211 a marker. Other Lisp symbols that end in @samp{p} include @code{zerop},
2212 a function that tests whether its argument has the value of zero, and
2213 @code{listp}, a function that tests whether its argument is a list.
2215 Finally, the last part of the error message is the symbol @code{hello}.
2216 This is the value of the argument that was passed to @code{+}. If the
2217 addition had been passed the correct type of object, the value passed
2218 would have been a number, such as 37, rather than a symbol like
2219 @code{hello}. But then you would not have got the error message.
2223 In GNU Emacs version 20 and before, the echo area displays an error
2227 Wrong type argument:@: number-or-marker-p, hello
2230 This says, in different words, the same as the top line of the
2231 @file{*Backtrace*} buffer.
2234 @node message, , Wrong Type of Argument, Arguments
2235 @comment node-name, next, previous, up
2236 @subsection The @code{message} Function
2239 Like @code{+}, the @code{message} function takes a variable number of
2240 arguments. It is used to send messages to the user and is so useful
2241 that we will describe it here.
2244 A message is printed in the echo area. For example, you can print a
2245 message in your echo area by evaluating the following list:
2248 (message "This message appears in the echo area!")
2251 The whole string between double quotation marks is a single argument
2252 and is printed @i{in toto}. (Note that in this example, the message
2253 itself will appear in the echo area within double quotes; that is
2254 because you see the value returned by the @code{message} function. In
2255 most uses of @code{message} in programs that you write, the text will
2256 be printed in the echo area as a side-effect, without the quotes.
2257 @xref{multiply-by-seven in detail, , @code{multiply-by-seven} in
2258 detail}, for an example of this.)
2260 However, if there is a @samp{%s} in the quoted string of characters, the
2261 @code{message} function does not print the @samp{%s} as such, but looks
2262 to the argument that follows the string. It evaluates the second
2263 argument and prints the value at the location in the string where the
2267 You can see this by positioning the cursor after the following
2268 expression and typing @kbd{C-x C-e}:
2271 (message "The name of this buffer is: %s." (buffer-name))
2275 In Info, @code{"The name of this buffer is: *info*."} will appear in the
2276 echo area. The function @code{buffer-name} returns the name of the
2277 buffer as a string, which the @code{message} function inserts in place
2280 To print a value as an integer, use @samp{%d} in the same way as
2281 @samp{%s}. For example, to print a message in the echo area that
2282 states the value of the @code{fill-column}, evaluate the following:
2285 (message "The value of fill-column is %d." fill-column)
2289 On my system, when I evaluate this list, @code{"The value of
2290 fill-column is 72."} appears in my echo area@footnote{Actually, you
2291 can use @code{%s} to print a number. It is non-specific. @code{%d}
2292 prints only the part of a number left of a decimal point, and not
2293 anything that is not a number.}.
2295 If there is more than one @samp{%s} in the quoted string, the value of
2296 the first argument following the quoted string is printed at the
2297 location of the first @samp{%s} and the value of the second argument is
2298 printed at the location of the second @samp{%s}, and so on.
2301 For example, if you evaluate the following,
2305 (message "There are %d %s in the office!"
2306 (- fill-column 14) "pink elephants")
2311 a rather whimsical message will appear in your echo area. On my system
2312 it says, @code{"There are 58 pink elephants in the office!"}.
2314 The expression @code{(- fill-column 14)} is evaluated and the resulting
2315 number is inserted in place of the @samp{%d}; and the string in double
2316 quotes, @code{"pink elephants"}, is treated as a single argument and
2317 inserted in place of the @samp{%s}. (That is to say, a string between
2318 double quotes evaluates to itself, like a number.)
2320 Finally, here is a somewhat complex example that not only illustrates
2321 the computation of a number, but also shows how you can use an
2322 expression within an expression to generate the text that is substituted
2327 (message "He saw %d %s"
2331 "The quick brown foxes jumped." 16 21)
2336 In this example, @code{message} has three arguments: the string,
2337 @code{"He saw %d %s"}, the expression, @code{(- fill-column 32)}, and
2338 the expression beginning with the function @code{concat}. The value
2339 resulting from the evaluation of @code{(- fill-column 32)} is inserted
2340 in place of the @samp{%d}; and the value returned by the expression
2341 beginning with @code{concat} is inserted in place of the @samp{%s}.
2343 When your fill column is 70 and you evaluate the expression, the
2344 message @code{"He saw 38 red foxes leaping."} appears in your echo
2347 @node set & setq, Summary, Arguments, List Processing
2348 @comment node-name, next, previous, up
2349 @section Setting the Value of a Variable
2350 @cindex Variable, setting value
2351 @cindex Setting value of variable
2353 @cindex @samp{bind} defined
2354 There are several ways by which a variable can be given a value. One of
2355 the ways is to use either the function @code{set} or the function
2356 @code{setq}. Another way is to use @code{let} (@pxref{let}). (The
2357 jargon for this process is to @dfn{bind} a variable to a value.)
2359 The following sections not only describe how @code{set} and @code{setq}
2360 work but also illustrate how arguments are passed.
2363 * Using set:: Setting values.
2364 * Using setq:: Setting a quoted value.
2365 * Counting:: Using @code{setq} to count.
2368 @node Using set, Using setq, set & setq, set & setq
2369 @comment node-name, next, previous, up
2370 @subsection Using @code{set}
2373 To set the value of the symbol @code{flowers} to the list @code{'(rose
2374 violet daisy buttercup)}, evaluate the following expression by
2375 positioning the cursor after the expression and typing @kbd{C-x C-e}.
2378 (set 'flowers '(rose violet daisy buttercup))
2382 The list @code{(rose violet daisy buttercup)} will appear in the echo
2383 area. This is what is @emph{returned} by the @code{set} function. As a
2384 side effect, the symbol @code{flowers} is bound to the list; that is,
2385 the symbol @code{flowers}, which can be viewed as a variable, is given
2386 the list as its value. (This process, by the way, illustrates how a
2387 side effect to the Lisp interpreter, setting the value, can be the
2388 primary effect that we humans are interested in. This is because every
2389 Lisp function must return a value if it does not get an error, but it
2390 will only have a side effect if it is designed to have one.)
2392 After evaluating the @code{set} expression, you can evaluate the symbol
2393 @code{flowers} and it will return the value you just set. Here is the
2394 symbol. Place your cursor after it and type @kbd{C-x C-e}.
2401 When you evaluate @code{flowers}, the list
2402 @code{(rose violet daisy buttercup)} appears in the echo area.
2404 Incidentally, if you evaluate @code{'flowers}, the variable with a quote
2405 in front of it, what you will see in the echo area is the symbol itself,
2406 @code{flowers}. Here is the quoted symbol, so you can try this:
2412 Note also, that when you use @code{set}, you need to quote both
2413 arguments to @code{set}, unless you want them evaluated. Since we do
2414 not want either argument evaluated, neither the variable
2415 @code{flowers} nor the list @code{(rose violet daisy buttercup)}, both
2416 are quoted. (When you use @code{set} without quoting its first
2417 argument, the first argument is evaluated before anything else is
2418 done. If you did this and @code{flowers} did not have a value
2419 already, you would get an error message that the @samp{Symbol's value
2420 as variable is void}; on the other hand, if @code{flowers} did return
2421 a value after it was evaluated, the @code{set} would attempt to set
2422 the value that was returned. There are situations where this is the
2423 right thing for the function to do; but such situations are rare.)
2425 @node Using setq, Counting, Using set, set & setq
2426 @comment node-name, next, previous, up
2427 @subsection Using @code{setq}
2430 As a practical matter, you almost always quote the first argument to
2431 @code{set}. The combination of @code{set} and a quoted first argument
2432 is so common that it has its own name: the special form @code{setq}.
2433 This special form is just like @code{set} except that the first argument
2434 is quoted automatically, so you don't need to type the quote mark
2435 yourself. Also, as an added convenience, @code{setq} permits you to set
2436 several different variables to different values, all in one expression.
2438 To set the value of the variable @code{carnivores} to the list
2439 @code{'(lion tiger leopard)} using @code{setq}, the following expression
2443 (setq carnivores '(lion tiger leopard))
2447 This is exactly the same as using @code{set} except the first argument
2448 is automatically quoted by @code{setq}. (The @samp{q} in @code{setq}
2449 means @code{quote}.)
2452 With @code{set}, the expression would look like this:
2455 (set 'carnivores '(lion tiger leopard))
2458 Also, @code{setq} can be used to assign different values to
2459 different variables. The first argument is bound to the value
2460 of the second argument, the third argument is bound to the value of the
2461 fourth argument, and so on. For example, you could use the following to
2462 assign a list of trees to the symbol @code{trees} and a list of herbivores
2463 to the symbol @code{herbivores}:
2467 (setq trees '(pine fir oak maple)
2468 herbivores '(gazelle antelope zebra))
2473 (The expression could just as well have been on one line, but it might
2474 not have fit on a page; and humans find it easier to read nicely
2477 Although I have been using the term `assign', there is another way of
2478 thinking about the workings of @code{set} and @code{setq}; and that is to
2479 say that @code{set} and @code{setq} make the symbol @emph{point} to the
2480 list. This latter way of thinking is very common and in forthcoming
2481 chapters we shall come upon at least one symbol that has `pointer' as
2482 part of its name. The name is chosen because the symbol has a value,
2483 specifically a list, attached to it; or, expressed another way,
2484 the symbol is set to ``point'' to the list.
2486 @node Counting, , Using setq, set & setq
2487 @comment node-name, next, previous, up
2488 @subsection Counting
2491 Here is an example that shows how to use @code{setq} in a counter. You
2492 might use this to count how many times a part of your program repeats
2493 itself. First set a variable to zero; then add one to the number each
2494 time the program repeats itself. To do this, you need a variable that
2495 serves as a counter, and two expressions: an initial @code{setq}
2496 expression that sets the counter variable to zero; and a second
2497 @code{setq} expression that increments the counter each time it is
2502 (setq counter 0) ; @r{Let's call this the initializer.}
2504 (setq counter (+ counter 1)) ; @r{This is the incrementer.}
2506 counter ; @r{This is the counter.}
2511 (The text following the @samp{;} are comments. @xref{Change a
2512 defun, , Change a Function Definition}.)
2514 If you evaluate the first of these expressions, the initializer,
2515 @code{(setq counter 0)}, and then evaluate the third expression,
2516 @code{counter}, the number @code{0} will appear in the echo area. If
2517 you then evaluate the second expression, the incrementer, @code{(setq
2518 counter (+ counter 1))}, the counter will get the value 1. So if you
2519 again evaluate @code{counter}, the number @code{1} will appear in the
2520 echo area. Each time you evaluate the second expression, the value of
2521 the counter will be incremented.
2523 When you evaluate the incrementer, @code{(setq counter (+ counter 1))},
2524 the Lisp interpreter first evaluates the innermost list; this is the
2525 addition. In order to evaluate this list, it must evaluate the variable
2526 @code{counter} and the number @code{1}. When it evaluates the variable
2527 @code{counter}, it receives its current value. It passes this value and
2528 the number @code{1} to the @code{+} which adds them together. The sum
2529 is then returned as the value of the inner list and passed to the
2530 @code{setq} which sets the variable @code{counter} to this new value.
2531 Thus, the value of the variable, @code{counter}, is changed.
2533 @node Summary, Error Message Exercises, set & setq, List Processing
2534 @comment node-name, next, previous, up
2537 Learning Lisp is like climbing a hill in which the first part is the
2538 steepest. You have now climbed the most difficult part; what remains
2539 becomes easier as you progress onwards.
2547 Lisp programs are made up of expressions, which are lists or single atoms.
2550 Lists are made up of zero or more atoms or inner lists, separated by whitespace and
2551 surrounded by parentheses. A list can be empty.
2554 Atoms are multi-character symbols, like @code{forward-paragraph}, single
2555 character symbols like @code{+}, strings of characters between double
2556 quotation marks, or numbers.
2559 A number evaluates to itself.
2562 A string between double quotes also evaluates to itself.
2565 When you evaluate a symbol by itself, its value is returned.
2568 When you evaluate a list, the Lisp interpreter looks at the first symbol
2569 in the list and then at the function definition bound to that symbol.
2570 Then the instructions in the function definition are carried out.
2573 A single quotation mark,
2580 , tells the Lisp interpreter that it should
2581 return the following expression as written, and not evaluate it as it
2582 would if the quote were not there.
2585 Arguments are the information passed to a function. The arguments to a
2586 function are computed by evaluating the rest of the elements of the list
2587 of which the function is the first element.
2590 A function always returns a value when it is evaluated (unless it gets
2591 an error); in addition, it may also carry out some action called a
2592 ``side effect''. In many cases, a function's primary purpose is to
2593 create a side effect.
2596 @node Error Message Exercises, , Summary, List Processing
2597 @comment node-name, next, previous, up
2600 A few simple exercises:
2604 Generate an error message by evaluating an appropriate symbol that is
2605 not within parentheses.
2608 Generate an error message by evaluating an appropriate symbol that is
2609 between parentheses.
2612 Create a counter that increments by two rather than one.
2615 Write an expression that prints a message in the echo area when
2619 @node Practicing Evaluation, Writing Defuns, List Processing, Top
2620 @comment node-name, next, previous, up
2621 @chapter Practicing Evaluation
2622 @cindex Practicing evaluation
2623 @cindex Evaluation practice
2625 Before learning how to write a function definition in Emacs Lisp, it is
2626 useful to spend a little time evaluating various expressions that have
2627 already been written. These expressions will be lists with the
2628 functions as their first (and often only) element. Since some of the
2629 functions associated with buffers are both simple and interesting, we
2630 will start with those. In this section, we will evaluate a few of
2631 these. In another section, we will study the code of several other
2632 buffer-related functions, to see how they were written.
2635 * How to Evaluate:: Typing editing commands or @kbd{C-x C-e}
2637 * Buffer Names:: Buffers and files are different.
2638 * Getting Buffers:: Getting a buffer itself, not merely its name.
2639 * Switching Buffers:: How to change to another buffer.
2640 * Buffer Size & Locations:: Where point is located and the size of
2642 * Evaluation Exercise::
2645 @node How to Evaluate, Buffer Names, Practicing Evaluation, Practicing Evaluation
2647 @unnumberedsec How to Evaluate
2650 @i{Whenever you give an editing command} to Emacs Lisp, such as the
2651 command to move the cursor or to scroll the screen, @i{you are evaluating
2652 an expression,} the first element of which is a function. @i{This is
2655 @cindex @samp{interactive function} defined
2656 @cindex @samp{command} defined
2657 When you type keys, you cause the Lisp interpreter to evaluate an
2658 expression and that is how you get your results. Even typing plain text
2659 involves evaluating an Emacs Lisp function, in this case, one that uses
2660 @code{self-insert-command}, which simply inserts the character you
2661 typed. The functions you evaluate by typing keystrokes are called
2662 @dfn{interactive} functions, or @dfn{commands}; how you make a function
2663 interactive will be illustrated in the chapter on how to write function
2664 definitions. @xref{Interactive, , Making a Function Interactive}.
2666 In addition to typing keyboard commands, we have seen a second way to
2667 evaluate an expression: by positioning the cursor after a list and
2668 typing @kbd{C-x C-e}. This is what we will do in the rest of this
2669 section. There are other ways to evaluate an expression as well; these
2670 will be described as we come to them.
2672 Besides being used for practicing evaluation, the functions shown in the
2673 next few sections are important in their own right. A study of these
2674 functions makes clear the distinction between buffers and files, how to
2675 switch to a buffer, and how to determine a location within it.
2677 @node Buffer Names, Getting Buffers, How to Evaluate, Practicing Evaluation
2678 @comment node-name, next, previous, up
2679 @section Buffer Names
2681 @findex buffer-file-name
2683 The two functions, @code{buffer-name} and @code{buffer-file-name}, show
2684 the difference between a file and a buffer. When you evaluate the
2685 following expression, @code{(buffer-name)}, the name of the buffer
2686 appears in the echo area. When you evaluate @code{(buffer-file-name)},
2687 the name of the file to which the buffer refers appears in the echo
2688 area. Usually, the name returned by @code{(buffer-name)} is the same as
2689 the name of the file to which it refers, and the name returned by
2690 @code{(buffer-file-name)} is the full path-name of the file.
2692 A file and a buffer are two different entities. A file is information
2693 recorded permanently in the computer (unless you delete it). A buffer,
2694 on the other hand, is information inside of Emacs that will vanish at
2695 the end of the editing session (or when you kill the buffer). Usually,
2696 a buffer contains information that you have copied from a file; we say
2697 the buffer is @dfn{visiting} that file. This copy is what you work on
2698 and modify. Changes to the buffer do not change the file, until you
2699 save the buffer. When you save the buffer, the buffer is copied to the file
2700 and is thus saved permanently.
2703 If you are reading this in Info inside of GNU Emacs, you can evaluate
2704 each of the following expressions by positioning the cursor after it and
2705 typing @kbd{C-x C-e}.
2716 When I do this in Info, the value returned by evaluating
2717 @code{(buffer-name)} is @file{"*info*"}, and the value returned by
2718 evaluating @code{(buffer-file-name)} is @file{nil}.
2720 On the other hand, while I am writing this Introduction, the value
2721 returned by evaluating @code{(buffer-name)} is
2722 @file{"introduction.texinfo"}, and the value returned by evaluating
2723 @code{(buffer-file-name)} is
2724 @file{"/gnu/work/intro/introduction.texinfo"}.
2726 @cindex @code{nil}, history of word
2727 The former is the name of the buffer and the latter is the name of the
2728 file. In Info, the buffer name is @file{"*info*"}. Info does not
2729 point to any file, so the result of evaluating
2730 @code{(buffer-file-name)} is @file{nil}. The symbol @code{nil} is
2731 from the Latin word for `nothing'; in this case, it means that the
2732 buffer is not associated with any file. (In Lisp, @code{nil} is also
2733 used to mean `false' and is a synonym for the empty list, @code{()}.)
2735 When I am writing, the name of my buffer is
2736 @file{"introduction.texinfo"}. The name of the file to which it
2737 points is @file{"/gnu/work/intro/introduction.texinfo"}.
2739 (In the expressions, the parentheses tell the Lisp interpreter to
2740 treat @w{@code{buffer-name}} and @w{@code{buffer-file-name}} as
2741 functions; without the parentheses, the interpreter would attempt to
2742 evaluate the symbols as variables. @xref{Variables}.)
2744 In spite of the distinction between files and buffers, you will often
2745 find that people refer to a file when they mean a buffer and vice-verse.
2746 Indeed, most people say, ``I am editing a file,'' rather than saying,
2747 ``I am editing a buffer which I will soon save to a file.'' It is
2748 almost always clear from context what people mean. When dealing with
2749 computer programs, however, it is important to keep the distinction in mind,
2750 since the computer is not as smart as a person.
2752 @cindex Buffer, history of word
2753 The word `buffer', by the way, comes from the meaning of the word as a
2754 cushion that deadens the force of a collision. In early computers, a
2755 buffer cushioned the interaction between files and the computer's
2756 central processing unit. The drums or tapes that held a file and the
2757 central processing unit were pieces of equipment that were very
2758 different from each other, working at their own speeds, in spurts. The
2759 buffer made it possible for them to work together effectively.
2760 Eventually, the buffer grew from being an intermediary, a temporary
2761 holding place, to being the place where work is done. This
2762 transformation is rather like that of a small seaport that grew into a
2763 great city: once it was merely the place where cargo was warehoused
2764 temporarily before being loaded onto ships; then it became a business
2765 and cultural center in its own right.
2767 Not all buffers are associated with files. For example, a
2768 @file{*scratch*} buffer does not visit any file. Similarly, a
2769 @file{*Help*} buffer is not associated with any file.
2771 In the old days, when you lacked a @file{~/.emacs} file and started an
2772 Emacs session by typing the command @code{emacs} alone, without naming
2773 any files, Emacs started with the @file{*scratch*} buffer visible.
2774 Nowadays, you will see a splash screen. You can follow one of the
2775 commands suggested on the splash screen, visit a file, or press the
2776 spacebar to reach the @file{*scratch*} buffer.
2778 If you switch to the @file{*scratch*} buffer, type
2779 @code{(buffer-name)}, position the cursor after it, and then type
2780 @kbd{C-x C-e} to evaluate the expression. The name @code{"*scratch*"}
2781 will be returned and will appear in the echo area. @code{"*scratch*"}
2782 is the name of the buffer. When you type @code{(buffer-file-name)} in
2783 the @file{*scratch*} buffer and evaluate that, @code{nil} will appear
2784 in the echo area, just as it does when you evaluate
2785 @code{(buffer-file-name)} in Info.
2787 Incidentally, if you are in the @file{*scratch*} buffer and want the
2788 value returned by an expression to appear in the @file{*scratch*}
2789 buffer itself rather than in the echo area, type @kbd{C-u C-x C-e}
2790 instead of @kbd{C-x C-e}. This causes the value returned to appear
2791 after the expression. The buffer will look like this:
2794 (buffer-name)"*scratch*"
2798 You cannot do this in Info since Info is read-only and it will not allow
2799 you to change the contents of the buffer. But you can do this in any
2800 buffer you can edit; and when you write code or documentation (such as
2801 this book), this feature is very useful.
2803 @node Getting Buffers, Switching Buffers, Buffer Names, Practicing Evaluation
2804 @comment node-name, next, previous, up
2805 @section Getting Buffers
2806 @findex current-buffer
2807 @findex other-buffer
2808 @cindex Getting a buffer
2810 The @code{buffer-name} function returns the @emph{name} of the buffer;
2811 to get the buffer @emph{itself}, a different function is needed: the
2812 @code{current-buffer} function. If you use this function in code, what
2813 you get is the buffer itself.
2815 A name and the object or entity to which the name refers are different
2816 from each other. You are not your name. You are a person to whom
2817 others refer by name. If you ask to speak to George and someone hands you
2818 a card with the letters @samp{G}, @samp{e}, @samp{o}, @samp{r},
2819 @samp{g}, and @samp{e} written on it, you might be amused, but you would
2820 not be satisfied. You do not want to speak to the name, but to the
2821 person to whom the name refers. A buffer is similar: the name of the
2822 scratch buffer is @file{*scratch*}, but the name is not the buffer. To
2823 get a buffer itself, you need to use a function such as
2824 @code{current-buffer}.
2826 However, there is a slight complication: if you evaluate
2827 @code{current-buffer} in an expression on its own, as we will do here,
2828 what you see is a printed representation of the name of the buffer
2829 without the contents of the buffer. Emacs works this way for two
2830 reasons: the buffer may be thousands of lines long---too long to be
2831 conveniently displayed; and, another buffer may have the same contents
2832 but a different name, and it is important to distinguish between them.
2835 Here is an expression containing the function:
2842 If you evaluate this expression in Info in Emacs in the usual way,
2843 @file{#<buffer *info*>} will appear in the echo area. The special
2844 format indicates that the buffer itself is being returned, rather than
2847 Incidentally, while you can type a number or symbol into a program, you
2848 cannot do that with the printed representation of a buffer: the only way
2849 to get a buffer itself is with a function such as @code{current-buffer}.
2851 A related function is @code{other-buffer}. This returns the most
2852 recently selected buffer other than the one you are in currently, not
2853 a printed representation of its name. If you have recently switched
2854 back and forth from the @file{*scratch*} buffer, @code{other-buffer}
2855 will return that buffer.
2858 You can see this by evaluating the expression:
2865 You should see @file{#<buffer *scratch*>} appear in the echo area, or
2866 the name of whatever other buffer you switched back from most
2867 recently@footnote{Actually, by default, if the buffer from which you
2868 just switched is visible to you in another window, @code{other-buffer}
2869 will choose the most recent buffer that you cannot see; this is a
2870 subtlety that I often forget.}.
2872 @node Switching Buffers, Buffer Size & Locations, Getting Buffers, Practicing Evaluation
2873 @comment node-name, next, previous, up
2874 @section Switching Buffers
2875 @findex switch-to-buffer
2877 @cindex Switching to a buffer
2879 The @code{other-buffer} function actually provides a buffer when it is
2880 used as an argument to a function that requires one. We can see this
2881 by using @code{other-buffer} and @code{switch-to-buffer} to switch to a
2884 But first, a brief introduction to the @code{switch-to-buffer}
2885 function. When you switched back and forth from Info to the
2886 @file{*scratch*} buffer to evaluate @code{(buffer-name)}, you most
2887 likely typed @kbd{C-x b} and then typed @file{*scratch*}@footnote{Or
2888 rather, to save typing, you probably only typed @kbd{RET} if the
2889 default buffer was @file{*scratch*}, or if it was different, then you
2890 typed just part of the name, such as @code{*sc}, pressed your
2891 @kbd{TAB} key to cause it to expand to the full name, and then typed
2892 your @kbd{RET} key.} when prompted in the minibuffer for the name of
2893 the buffer to which you wanted to switch. The keystrokes, @kbd{C-x
2894 b}, cause the Lisp interpreter to evaluate the interactive function
2895 @code{switch-to-buffer}. As we said before, this is how Emacs works:
2896 different keystrokes call or run different functions. For example,
2897 @kbd{C-f} calls @code{forward-char}, @kbd{M-e} calls
2898 @code{forward-sentence}, and so on.
2900 By writing @code{switch-to-buffer} in an expression, and giving it a
2901 buffer to switch to, we can switch buffers just the way @kbd{C-x b}
2905 Here is the Lisp expression:
2908 (switch-to-buffer (other-buffer))
2912 The symbol @code{switch-to-buffer} is the first element of the list,
2913 so the Lisp interpreter will treat it as a function and carry out the
2914 instructions that are attached to it. But before doing that, the
2915 interpreter will note that @code{other-buffer} is inside parentheses
2916 and work on that symbol first. @code{other-buffer} is the first (and
2917 in this case, the only) element of this list, so the Lisp interpreter
2918 calls or runs the function. It returns another buffer. Next, the
2919 interpreter runs @code{switch-to-buffer}, passing to it, as an
2920 argument, the other buffer, which is what Emacs will switch to. If
2921 you are reading this in Info, try this now. Evaluate the expression.
2922 (To get back, type @kbd{C-x b @key{RET}}.)@footnote{Remember, this
2923 expression will move you to your most recent other buffer that you
2924 cannot see. If you really want to go to your most recently selected
2925 buffer, even if you can still see it, you need to evaluate the
2926 following more complex expression:
2929 (switch-to-buffer (other-buffer (current-buffer) t))
2933 In this case, the first argument to @code{other-buffer} tells it which
2934 buffer to skip---the current one---and the second argument tells
2935 @code{other-buffer} it is OK to switch to a visible buffer.
2936 In regular use, @code{switch-to-buffer} takes you to an invisible
2937 window since you would most likely use @kbd{C-x o} (@code{other-window})
2938 to go to another visible buffer.}
2940 In the programming examples in later sections of this document, you will
2941 see the function @code{set-buffer} more often than
2942 @code{switch-to-buffer}. This is because of a difference between
2943 computer programs and humans: humans have eyes and expect to see the
2944 buffer on which they are working on their computer terminals. This is
2945 so obvious, it almost goes without saying. However, programs do not
2946 have eyes. When a computer program works on a buffer, that buffer does
2947 not need to be visible on the screen.
2949 @code{switch-to-buffer} is designed for humans and does two different
2950 things: it switches the buffer to which Emacs' attention is directed; and
2951 it switches the buffer displayed in the window to the new buffer.
2952 @code{set-buffer}, on the other hand, does only one thing: it switches
2953 the attention of the computer program to a different buffer. The buffer
2954 on the screen remains unchanged (of course, normally nothing happens
2955 there until the command finishes running).
2957 @cindex @samp{call} defined
2958 Also, we have just introduced another jargon term, the word @dfn{call}.
2959 When you evaluate a list in which the first symbol is a function, you
2960 are calling that function. The use of the term comes from the notion of
2961 the function as an entity that can do something for you if you `call'
2962 it---just as a plumber is an entity who can fix a leak if you call him
2965 @node Buffer Size & Locations, Evaluation Exercise, Switching Buffers, Practicing Evaluation
2966 @comment node-name, next, previous, up
2967 @section Buffer Size and the Location of Point
2968 @cindex Size of buffer
2970 @cindex Point location
2971 @cindex Location of point
2973 Finally, let's look at several rather simple functions,
2974 @code{buffer-size}, @code{point}, @code{point-min}, and
2975 @code{point-max}. These give information about the size of a buffer and
2976 the location of point within it.
2978 The function @code{buffer-size} tells you the size of the current
2979 buffer; that is, the function returns a count of the number of
2980 characters in the buffer.
2987 You can evaluate this in the usual way, by positioning the
2988 cursor after the expression and typing @kbd{C-x C-e}.
2990 @cindex @samp{point} defined
2991 In Emacs, the current position of the cursor is called @dfn{point}.
2992 The expression @code{(point)} returns a number that tells you where the
2993 cursor is located as a count of the number of characters from the
2994 beginning of the buffer up to point.
2997 You can see the character count for point in this buffer by evaluating
2998 the following expression in the usual way:
3005 As I write this, the value of @code{point} is 65724. The @code{point}
3006 function is frequently used in some of the examples later in this
3010 The value of point depends, of course, on its location within the
3011 buffer. If you evaluate point in this spot, the number will be larger:
3018 For me, the value of point in this location is 66043, which means that
3019 there are 319 characters (including spaces) between the two
3020 expressions. (Doubtless, you will see different numbers, since I will
3021 have edited this since I first evaluated point.)
3023 @cindex @samp{narrowing} defined
3024 The function @code{point-min} is somewhat similar to @code{point}, but
3025 it returns the value of the minimum permissible value of point in the
3026 current buffer. This is the number 1 unless @dfn{narrowing} is in
3027 effect. (Narrowing is a mechanism whereby you can restrict yourself,
3028 or a program, to operations on just a part of a buffer.
3029 @xref{Narrowing & Widening, , Narrowing and Widening}.) Likewise, the
3030 function @code{point-max} returns the value of the maximum permissible
3031 value of point in the current buffer.
3033 @node Evaluation Exercise, , Buffer Size & Locations, Practicing Evaluation
3036 Find a file with which you are working and move towards its middle.
3037 Find its buffer name, file name, length, and your position in the file.
3039 @node Writing Defuns, Buffer Walk Through, Practicing Evaluation, Top
3040 @comment node-name, next, previous, up
3041 @chapter How To Write Function Definitions
3042 @cindex Definition writing
3043 @cindex Function definition writing
3044 @cindex Writing a function definition
3046 When the Lisp interpreter evaluates a list, it looks to see whether the
3047 first symbol on the list has a function definition attached to it; or,
3048 put another way, whether the symbol points to a function definition. If
3049 it does, the computer carries out the instructions in the definition. A
3050 symbol that has a function definition is called, simply, a function
3051 (although, properly speaking, the definition is the function and the
3052 symbol refers to it.)
3055 * Primitive Functions::
3056 * defun:: The @code{defun} special form.
3057 * Install:: Install a function definition.
3058 * Interactive:: Making a function interactive.
3059 * Interactive Options:: Different options for @code{interactive}.
3060 * Permanent Installation:: Installing code permanently.
3061 * let:: Creating and initializing local variables.
3063 * else:: If--then--else expressions.
3064 * Truth & Falsehood:: What Lisp considers false and true.
3065 * save-excursion:: Keeping track of point, mark, and buffer.
3070 @node Primitive Functions, defun, Writing Defuns, Writing Defuns
3072 @unnumberedsec An Aside about Primitive Functions
3074 @cindex Primitive functions
3075 @cindex Functions, primitive
3077 @cindex C language primitives
3078 @cindex Primitives written in C
3079 All functions are defined in terms of other functions, except for a few
3080 @dfn{primitive} functions that are written in the C programming
3081 language. When you write functions' definitions, you will write them in
3082 Emacs Lisp and use other functions as your building blocks. Some of the
3083 functions you will use will themselves be written in Emacs Lisp (perhaps
3084 by you) and some will be primitives written in C. The primitive
3085 functions are used exactly like those written in Emacs Lisp and behave
3086 like them. They are written in C so we can easily run GNU Emacs on any
3087 computer that has sufficient power and can run C.
3089 Let me re-emphasize this: when you write code in Emacs Lisp, you do not
3090 distinguish between the use of functions written in C and the use of
3091 functions written in Emacs Lisp. The difference is irrelevant. I
3092 mention the distinction only because it is interesting to know. Indeed,
3093 unless you investigate, you won't know whether an already-written
3094 function is written in Emacs Lisp or C.
3096 @node defun, Install, Primitive Functions, Writing Defuns
3097 @comment node-name, next, previous, up
3098 @section The @code{defun} Special Form
3100 @cindex Special form of @code{defun}
3102 @cindex @samp{function definition} defined
3103 In Lisp, a symbol such as @code{mark-whole-buffer} has code attached to
3104 it that tells the computer what to do when the function is called.
3105 This code is called the @dfn{function definition} and is created by
3106 evaluating a Lisp expression that starts with the symbol @code{defun}
3107 (which is an abbreviation for @emph{define function}). Because
3108 @code{defun} does not evaluate its arguments in the usual way, it is
3109 called a @dfn{special form}.
3111 In subsequent sections, we will look at function definitions from the
3112 Emacs source code, such as @code{mark-whole-buffer}. In this section,
3113 we will describe a simple function definition so you can see how it
3114 looks. This function definition uses arithmetic because it makes for a
3115 simple example. Some people dislike examples using arithmetic; however,
3116 if you are such a person, do not despair. Hardly any of the code we
3117 will study in the remainder of this introduction involves arithmetic or
3118 mathematics. The examples mostly involve text in one way or another.
3120 A function definition has up to five parts following the word
3125 The name of the symbol to which the function definition should be
3129 A list of the arguments that will be passed to the function. If no
3130 arguments will be passed to the function, this is an empty list,
3134 Documentation describing the function. (Technically optional, but
3135 strongly recommended.)
3138 Optionally, an expression to make the function interactive so you can
3139 use it by typing @kbd{M-x} and then the name of the function; or by
3140 typing an appropriate key or keychord.
3142 @cindex @samp{body} defined
3144 The code that instructs the computer what to do: the @dfn{body} of the
3145 function definition.
3148 It is helpful to think of the five parts of a function definition as
3149 being organized in a template, with slots for each part:
3153 (defun @var{function-name} (@var{arguments}@dots{})
3154 "@var{optional-documentation}@dots{}"
3155 (interactive @var{argument-passing-info}) ; @r{optional}
3160 As an example, here is the code for a function that multiplies its
3161 argument by 7. (This example is not interactive. @xref{Interactive,
3162 , Making a Function Interactive}, for that information.)
3166 (defun multiply-by-seven (number)
3167 "Multiply NUMBER by seven."
3172 This definition begins with a parenthesis and the symbol @code{defun},
3173 followed by the name of the function.
3175 @cindex @samp{argument list} defined
3176 The name of the function is followed by a list that contains the
3177 arguments that will be passed to the function. This list is called
3178 the @dfn{argument list}. In this example, the list has only one
3179 element, the symbol, @code{number}. When the function is used, the
3180 symbol will be bound to the value that is used as the argument to the
3183 Instead of choosing the word @code{number} for the name of the argument,
3184 I could have picked any other name. For example, I could have chosen
3185 the word @code{multiplicand}. I picked the word `number' because it
3186 tells what kind of value is intended for this slot; but I could just as
3187 well have chosen the word `multiplicand' to indicate the role that the
3188 value placed in this slot will play in the workings of the function. I
3189 could have called it @code{foogle}, but that would have been a bad
3190 choice because it would not tell humans what it means. The choice of
3191 name is up to the programmer and should be chosen to make the meaning of
3194 Indeed, you can choose any name you wish for a symbol in an argument
3195 list, even the name of a symbol used in some other function: the name
3196 you use in an argument list is private to that particular definition.
3197 In that definition, the name refers to a different entity than any use
3198 of the same name outside the function definition. Suppose you have a
3199 nick-name `Shorty' in your family; when your family members refer to
3200 `Shorty', they mean you. But outside your family, in a movie, for
3201 example, the name `Shorty' refers to someone else. Because a name in an
3202 argument list is private to the function definition, you can change the
3203 value of such a symbol inside the body of a function without changing
3204 its value outside the function. The effect is similar to that produced
3205 by a @code{let} expression. (@xref{let, , @code{let}}.)
3208 Note also that we discuss the word `number' in two different ways: as a
3209 symbol that appears in the code, and as the name of something that will
3210 be replaced by a something else during the evaluation of the function.
3211 In the first case, @code{number} is a symbol, not a number; it happens
3212 that within the function, it is a variable who value is the number in
3213 question, but our primary interest in it is as a symbol. On the other
3214 hand, when we are talking about the function, our interest is that we
3215 will substitute a number for the word @var{number}. To keep this
3216 distinction clear, we use different typography for the two
3217 circumstances. When we talk about this function, or about how it works,
3218 we refer to this number by writing @var{number}. In the function
3219 itself, we refer to it by writing @code{number}.
3222 The argument list is followed by the documentation string that
3223 describes the function. This is what you see when you type
3224 @w{@kbd{C-h f}} and the name of a function. Incidentally, when you
3225 write a documentation string like this, you should make the first line
3226 a complete sentence since some commands, such as @code{apropos}, print
3227 only the first line of a multi-line documentation string. Also, you
3228 should not indent the second line of a documentation string, if you
3229 have one, because that looks odd when you use @kbd{C-h f}
3230 (@code{describe-function}). The documentation string is optional, but
3231 it is so useful, it should be included in almost every function you
3234 @findex * @r{(multiplication)}
3235 The third line of the example consists of the body of the function
3236 definition. (Most functions' definitions, of course, are longer than
3237 this.) In this function, the body is the list, @code{(* 7 number)}, which
3238 says to multiply the value of @var{number} by 7. (In Emacs Lisp,
3239 @code{*} is the function for multiplication, just as @code{+} is the
3240 function for addition.)
3242 When you use the @code{multiply-by-seven} function, the argument
3243 @code{number} evaluates to the actual number you want used. Here is an
3244 example that shows how @code{multiply-by-seven} is used; but don't try
3245 to evaluate this yet!
3248 (multiply-by-seven 3)
3252 The symbol @code{number}, specified in the function definition in the
3253 next section, is given or ``bound to'' the value 3 in the actual use of
3254 the function. Note that although @code{number} was inside parentheses
3255 in the function definition, the argument passed to the
3256 @code{multiply-by-seven} function is not in parentheses. The
3257 parentheses are written in the function definition so the computer can
3258 figure out where the argument list ends and the rest of the function
3261 If you evaluate this example, you are likely to get an error message.
3262 (Go ahead, try it!) This is because we have written the function
3263 definition, but not yet told the computer about the definition---we have
3264 not yet installed (or `loaded') the function definition in Emacs.
3265 Installing a function is the process that tells the Lisp interpreter the
3266 definition of the function. Installation is described in the next
3269 @node Install, Interactive, defun, Writing Defuns
3270 @comment node-name, next, previous, up
3271 @section Install a Function Definition
3272 @cindex Install a Function Definition
3273 @cindex Definition installation
3274 @cindex Function definition installation
3276 If you are reading this inside of Info in Emacs, you can try out the
3277 @code{multiply-by-seven} function by first evaluating the function
3278 definition and then evaluating @code{(multiply-by-seven 3)}. A copy of
3279 the function definition follows. Place the cursor after the last
3280 parenthesis of the function definition and type @kbd{C-x C-e}. When you
3281 do this, @code{multiply-by-seven} will appear in the echo area. (What
3282 this means is that when a function definition is evaluated, the value it
3283 returns is the name of the defined function.) At the same time, this
3284 action installs the function definition.
3288 (defun multiply-by-seven (number)
3289 "Multiply NUMBER by seven."
3295 By evaluating this @code{defun}, you have just installed
3296 @code{multiply-by-seven} in Emacs. The function is now just as much a
3297 part of Emacs as @code{forward-word} or any other editing function you
3298 use. (@code{multiply-by-seven} will stay installed until you quit
3299 Emacs. To reload code automatically whenever you start Emacs, see
3300 @ref{Permanent Installation, , Installing Code Permanently}.)
3303 * Effect of installation::
3304 * Change a defun:: How to change a function definition.
3307 @node Effect of installation, Change a defun, Install, Install
3309 @unnumberedsubsec The effect of installation
3312 You can see the effect of installing @code{multiply-by-seven} by
3313 evaluating the following sample. Place the cursor after the following
3314 expression and type @kbd{C-x C-e}. The number 21 will appear in the
3318 (multiply-by-seven 3)
3321 If you wish, you can read the documentation for the function by typing
3322 @kbd{C-h f} (@code{describe-function}) and then the name of the
3323 function, @code{multiply-by-seven}. When you do this, a
3324 @file{*Help*} window will appear on your screen that says:
3328 multiply-by-seven is a Lisp function.
3329 (multiply-by-seven NUMBER)
3331 Multiply NUMBER by seven.
3336 (To return to a single window on your screen, type @kbd{C-x 1}.)
3338 @node Change a defun, , Effect of installation, Install
3339 @comment node-name, next, previous, up
3340 @subsection Change a Function Definition
3341 @cindex Changing a function definition
3342 @cindex Function definition, how to change
3343 @cindex Definition, how to change
3345 If you want to change the code in @code{multiply-by-seven}, just rewrite
3346 it. To install the new version in place of the old one, evaluate the
3347 function definition again. This is how you modify code in Emacs. It is
3350 As an example, you can change the @code{multiply-by-seven} function to
3351 add the number to itself seven times instead of multiplying the number
3352 by seven. It produces the same answer, but by a different path. At
3353 the same time, we will add a comment to the code; a comment is text
3354 that the Lisp interpreter ignores, but that a human reader may find
3355 useful or enlightening. The comment is that this is the ``second
3360 (defun multiply-by-seven (number) ; @r{Second version.}
3361 "Multiply NUMBER by seven."
3362 (+ number number number number number number number))
3366 @cindex Comments in Lisp code
3367 The comment follows a semicolon, @samp{;}. In Lisp, everything on a
3368 line that follows a semicolon is a comment. The end of the line is the
3369 end of the comment. To stretch a comment over two or more lines, begin
3370 each line with a semicolon.
3372 @xref{Beginning a .emacs File, , Beginning a @file{.emacs}
3373 File}, and @ref{Comments, , Comments, elisp, The GNU Emacs Lisp
3374 Reference Manual}, for more about comments.
3376 You can install this version of the @code{multiply-by-seven} function by
3377 evaluating it in the same way you evaluated the first function: place
3378 the cursor after the last parenthesis and type @kbd{C-x C-e}.
3380 In summary, this is how you write code in Emacs Lisp: you write a
3381 function; install it; test it; and then make fixes or enhancements and
3384 @node Interactive, Interactive Options, Install, Writing Defuns
3385 @comment node-name, next, previous, up
3386 @section Make a Function Interactive
3387 @cindex Interactive functions
3390 You make a function interactive by placing a list that begins with
3391 the special form @code{interactive} immediately after the
3392 documentation. A user can invoke an interactive function by typing
3393 @kbd{M-x} and then the name of the function; or by typing the keys to
3394 which it is bound, for example, by typing @kbd{C-n} for
3395 @code{next-line} or @kbd{C-x h} for @code{mark-whole-buffer}.
3397 Interestingly, when you call an interactive function interactively,
3398 the value returned is not automatically displayed in the echo area.
3399 This is because you often call an interactive function for its side
3400 effects, such as moving forward by a word or line, and not for the
3401 value returned. If the returned value were displayed in the echo area
3402 each time you typed a key, it would be very distracting.
3405 * Interactive multiply-by-seven:: An overview.
3406 * multiply-by-seven in detail:: The interactive version.
3409 @node Interactive multiply-by-seven, multiply-by-seven in detail, Interactive, Interactive
3411 @unnumberedsubsec An Interactive @code{multiply-by-seven}, An Overview
3414 Both the use of the special form @code{interactive} and one way to
3415 display a value in the echo area can be illustrated by creating an
3416 interactive version of @code{multiply-by-seven}.
3423 (defun multiply-by-seven (number) ; @r{Interactive version.}
3424 "Multiply NUMBER by seven."
3426 (message "The result is %d" (* 7 number)))
3431 You can install this code by placing your cursor after it and typing
3432 @kbd{C-x C-e}. The name of the function will appear in your echo area.
3433 Then, you can use this code by typing @kbd{C-u} and a number and then
3434 typing @kbd{M-x multiply-by-seven} and pressing @key{RET}. The phrase
3435 @samp{The result is @dots{}} followed by the product will appear in the
3438 Speaking more generally, you invoke a function like this in either of two
3443 By typing a prefix argument that contains the number to be passed, and
3444 then typing @kbd{M-x} and the name of the function, as with
3445 @kbd{C-u 3 M-x forward-sentence}; or,
3448 By typing whatever key or keychord the function is bound to, as with
3453 Both the examples just mentioned work identically to move point forward
3454 three sentences. (Since @code{multiply-by-seven} is not bound to a key,
3455 it could not be used as an example of key binding.)
3457 (@xref{Keybindings, , Some Keybindings}, to learn how to bind a command
3460 A prefix argument is passed to an interactive function by typing the
3461 @key{META} key followed by a number, for example, @kbd{M-3 M-e}, or by
3462 typing @kbd{C-u} and then a number, for example, @kbd{C-u 3 M-e} (if you
3463 type @kbd{C-u} without a number, it defaults to 4).
3465 @node multiply-by-seven in detail, , Interactive multiply-by-seven, Interactive
3466 @comment node-name, next, previous, up
3467 @subsection An Interactive @code{multiply-by-seven}
3469 Let's look at the use of the special form @code{interactive} and then at
3470 the function @code{message} in the interactive version of
3471 @code{multiply-by-seven}. You will recall that the function definition
3476 (defun multiply-by-seven (number) ; @r{Interactive version.}
3477 "Multiply NUMBER by seven."
3479 (message "The result is %d" (* 7 number)))
3483 In this function, the expression, @code{(interactive "p")}, is a list of
3484 two elements. The @code{"p"} tells Emacs to pass the prefix argument to
3485 the function and use its value for the argument of the function.
3488 The argument will be a number. This means that the symbol
3489 @code{number} will be bound to a number in the line:
3492 (message "The result is %d" (* 7 number))
3497 For example, if your prefix argument is 5, the Lisp interpreter will
3498 evaluate the line as if it were:
3501 (message "The result is %d" (* 7 5))
3505 (If you are reading this in GNU Emacs, you can evaluate this expression
3506 yourself.) First, the interpreter will evaluate the inner list, which
3507 is @code{(* 7 5)}. This returns a value of 35. Next, it
3508 will evaluate the outer list, passing the values of the second and
3509 subsequent elements of the list to the function @code{message}.
3511 As we have seen, @code{message} is an Emacs Lisp function especially
3512 designed for sending a one line message to a user. (@xref{message, ,
3513 The @code{message} function}.) In summary, the @code{message}
3514 function prints its first argument in the echo area as is, except for
3515 occurrences of @samp{%d} or @samp{%s} (and various other %-sequences
3516 which we have not mentioned). When it sees a control sequence, the
3517 function looks to the second or subsequent arguments and prints the
3518 value of the argument in the location in the string where the control
3519 sequence is located.
3521 In the interactive @code{multiply-by-seven} function, the control string
3522 is @samp{%d}, which requires a number, and the value returned by
3523 evaluating @code{(* 7 5)} is the number 35. Consequently, the number 35
3524 is printed in place of the @samp{%d} and the message is @samp{The result
3527 (Note that when you call the function @code{multiply-by-seven}, the
3528 message is printed without quotes, but when you call @code{message}, the
3529 text is printed in double quotes. This is because the value returned by
3530 @code{message} is what appears in the echo area when you evaluate an
3531 expression whose first element is @code{message}; but when embedded in a
3532 function, @code{message} prints the text as a side effect without
3535 @node Interactive Options, Permanent Installation, Interactive, Writing Defuns
3536 @comment node-name, next, previous, up
3537 @section Different Options for @code{interactive}
3538 @cindex Options for @code{interactive}
3539 @cindex Interactive options
3541 In the example, @code{multiply-by-seven} used @code{"p"} as the
3542 argument to @code{interactive}. This argument told Emacs to interpret
3543 your typing either @kbd{C-u} followed by a number or @key{META}
3544 followed by a number as a command to pass that number to the function
3545 as its argument. Emacs has more than twenty characters predefined for
3546 use with @code{interactive}. In almost every case, one of these
3547 options will enable you to pass the right information interactively to
3548 a function. (@xref{Interactive Codes, , Code Characters for
3549 @code{interactive}, elisp, The GNU Emacs Lisp Reference Manual}.)
3552 Consider the function @code{zap-to-char}. Its interactive expression
3556 (interactive "p\ncZap to char: ")
3559 The first part of the argument to @code{interactive} is @samp{p}, with
3560 which you are already familiar. This argument tells Emacs to
3561 interpret a `prefix', as a number to be passed to the function. You
3562 can specify a prefix either by typing @kbd{C-u} followed by a number
3563 or by typing @key{META} followed by a number. The prefix is the
3564 number of specified characters. Thus, if your prefix is three and the
3565 specified character is @samp{x}, then you will delete all the text up
3566 to and including the third next @samp{x}. If you do not set a prefix,
3567 then you delete all the text up to and including the specified
3568 character, but no more.
3570 The @samp{c} tells the function the name of the character to which to delete.
3572 More formally, a function with two or more arguments can have
3573 information passed to each argument by adding parts to the string that
3574 follows @code{interactive}. When you do this, the information is
3575 passed to each argument in the same order it is specified in the
3576 @code{interactive} list. In the string, each part is separated from
3577 the next part by a @samp{\n}, which is a newline. For example, you
3578 can follow @samp{p} with a @samp{\n} and an @samp{cZap to char:@: }.
3579 This causes Emacs to pass the value of the prefix argument (if there
3580 is one) and the character.
3582 In this case, the function definition looks like the following, where
3583 @code{arg} and @code{char} are the symbols to which @code{interactive}
3584 binds the prefix argument and the specified character:
3588 (defun @var{name-of-function} (arg char)
3589 "@var{documentation}@dots{}"
3590 (interactive "p\ncZap to char: ")
3591 @var{body-of-function}@dots{})
3596 (The space after the colon in the prompt makes it look better when you
3597 are prompted. @xref{copy-to-buffer, , The Definition of
3598 @code{copy-to-buffer}}, for an example.)
3600 When a function does not take arguments, @code{interactive} does not
3601 require any. Such a function contains the simple expression
3602 @code{(interactive)}. The @code{mark-whole-buffer} function is like
3605 Alternatively, if the special letter-codes are not right for your
3606 application, you can pass your own arguments to @code{interactive} as
3609 @xref{append-to-buffer, , The Definition of @code{append-to-buffer}},
3610 for an example. @xref{Using Interactive, , Using @code{Interactive},
3611 elisp, The GNU Emacs Lisp Reference Manual}, for a more complete
3612 explanation about this technique.
3614 @node Permanent Installation, let, Interactive Options, Writing Defuns
3615 @comment node-name, next, previous, up
3616 @section Install Code Permanently
3617 @cindex Install code permanently
3618 @cindex Permanent code installation
3619 @cindex Code installation
3621 When you install a function definition by evaluating it, it will stay
3622 installed until you quit Emacs. The next time you start a new session
3623 of Emacs, the function will not be installed unless you evaluate the
3624 function definition again.
3626 At some point, you may want to have code installed automatically
3627 whenever you start a new session of Emacs. There are several ways of
3632 If you have code that is just for yourself, you can put the code for the
3633 function definition in your @file{.emacs} initialization file. When you
3634 start Emacs, your @file{.emacs} file is automatically evaluated and all
3635 the function definitions within it are installed.
3636 @xref{Emacs Initialization, , Your @file{.emacs} File}.
3639 Alternatively, you can put the function definitions that you want
3640 installed in one or more files of their own and use the @code{load}
3641 function to cause Emacs to evaluate and thereby install each of the
3642 functions in the files.
3643 @xref{Loading Files, , Loading Files}.
3646 Thirdly, if you have code that your whole site will use, it is usual
3647 to put it in a file called @file{site-init.el} that is loaded when
3648 Emacs is built. This makes the code available to everyone who uses
3649 your machine. (See the @file{INSTALL} file that is part of the Emacs
3653 Finally, if you have code that everyone who uses Emacs may want, you
3654 can post it on a computer network or send a copy to the Free Software
3655 Foundation. (When you do this, please license the code and its
3656 documentation under a license that permits other people to run, copy,
3657 study, modify, and redistribute the code and which protects you from
3658 having your work taken from you.) If you send a copy of your code to
3659 the Free Software Foundation, and properly protect yourself and
3660 others, it may be included in the next release of Emacs. In large
3661 part, this is how Emacs has grown over the past years, by donations.
3663 @node let, if, Permanent Installation, Writing Defuns
3664 @comment node-name, next, previous, up
3668 The @code{let} expression is a special form in Lisp that you will need
3669 to use in most function definitions.
3671 @code{let} is used to attach or bind a symbol to a value in such a way
3672 that the Lisp interpreter will not confuse the variable with a
3673 variable of the same name that is not part of the function.
3675 To understand why the @code{let} special form is necessary, consider
3676 the situation in which you own a home that you generally refer to as
3677 `the house', as in the sentence, ``The house needs painting.'' If you
3678 are visiting a friend and your host refers to `the house', he is
3679 likely to be referring to @emph{his} house, not yours, that is, to a
3682 If your friend is referring to his house and you think he is referring
3683 to your house, you may be in for some confusion. The same thing could
3684 happen in Lisp if a variable that is used inside of one function has
3685 the same name as a variable that is used inside of another function,
3686 and the two are not intended to refer to the same value. The
3687 @code{let} special form prevents this kind of confusion.
3690 * Prevent confusion::
3691 * Parts of let Expression::
3692 * Sample let Expression::
3693 * Uninitialized let Variables::
3696 @node Prevent confusion, Parts of let Expression, let, let
3698 @unnumberedsubsec @code{let} Prevents Confusion
3701 @cindex @samp{local variable} defined
3702 @cindex @samp{variable, local}, defined
3703 The @code{let} special form prevents confusion. @code{let} creates a
3704 name for a @dfn{local variable} that overshadows any use of the same
3705 name outside the @code{let} expression. This is like understanding
3706 that whenever your host refers to `the house', he means his house, not
3707 yours. (Symbols used in argument lists work the same way.
3708 @xref{defun, , The @code{defun} Special Form}.)
3710 Local variables created by a @code{let} expression retain their value
3711 @emph{only} within the @code{let} expression itself (and within
3712 expressions called within the @code{let} expression); the local
3713 variables have no effect outside the @code{let} expression.
3715 Another way to think about @code{let} is that it is like a @code{setq}
3716 that is temporary and local. The values set by @code{let} are
3717 automatically undone when the @code{let} is finished. The setting
3718 only affects expressions that are inside the bounds of the @code{let}
3719 expression. In computer science jargon, we would say ``the binding of
3720 a symbol is visible only in functions called in the @code{let} form;
3721 in Emacs Lisp, scoping is dynamic, not lexical.''
3723 @code{let} can create more than one variable at once. Also,
3724 @code{let} gives each variable it creates an initial value, either a
3725 value specified by you, or @code{nil}. (In the jargon, this is called
3726 `binding the variable to the value'.) After @code{let} has created
3727 and bound the variables, it executes the code in the body of the
3728 @code{let}, and returns the value of the last expression in the body,
3729 as the value of the whole @code{let} expression. (`Execute' is a jargon
3730 term that means to evaluate a list; it comes from the use of the word
3731 meaning `to give practical effect to' (@cite{Oxford English
3732 Dictionary}). Since you evaluate an expression to perform an action,
3733 `execute' has evolved as a synonym to `evaluate'.)
3735 @node Parts of let Expression, Sample let Expression, Prevent confusion, let
3736 @comment node-name, next, previous, up
3737 @subsection The Parts of a @code{let} Expression
3738 @cindex @code{let} expression, parts of
3739 @cindex Parts of @code{let} expression
3741 @cindex @samp{varlist} defined
3742 A @code{let} expression is a list of three parts. The first part is
3743 the symbol @code{let}. The second part is a list, called a
3744 @dfn{varlist}, each element of which is either a symbol by itself or a
3745 two-element list, the first element of which is a symbol. The third
3746 part of the @code{let} expression is the body of the @code{let}. The
3747 body usually consists of one or more lists.
3750 A template for a @code{let} expression looks like this:
3753 (let @var{varlist} @var{body}@dots{})
3757 The symbols in the varlist are the variables that are given initial
3758 values by the @code{let} special form. Symbols by themselves are given
3759 the initial value of @code{nil}; and each symbol that is the first
3760 element of a two-element list is bound to the value that is returned
3761 when the Lisp interpreter evaluates the second element.
3763 Thus, a varlist might look like this: @code{(thread (needles 3))}. In
3764 this case, in a @code{let} expression, Emacs binds the symbol
3765 @code{thread} to an initial value of @code{nil}, and binds the symbol
3766 @code{needles} to an initial value of 3.
3768 When you write a @code{let} expression, what you do is put the
3769 appropriate expressions in the slots of the @code{let} expression
3772 If the varlist is composed of two-element lists, as is often the case,
3773 the template for the @code{let} expression looks like this:
3777 (let ((@var{variable} @var{value})
3778 (@var{variable} @var{value})
3784 @node Sample let Expression, Uninitialized let Variables, Parts of let Expression, let
3785 @comment node-name, next, previous, up
3786 @subsection Sample @code{let} Expression
3787 @cindex Sample @code{let} expression
3788 @cindex @code{let} expression sample
3790 The following expression creates and gives initial values
3791 to the two variables @code{zebra} and @code{tiger}. The body of the
3792 @code{let} expression is a list which calls the @code{message} function.
3796 (let ((zebra 'stripes)
3798 (message "One kind of animal has %s and another is %s."
3803 Here, the varlist is @code{((zebra 'stripes) (tiger 'fierce))}.
3805 The two variables are @code{zebra} and @code{tiger}. Each variable is
3806 the first element of a two-element list and each value is the second
3807 element of its two-element list. In the varlist, Emacs binds the
3808 variable @code{zebra} to the value @code{stripes}@footnote{According
3809 to Jared Diamond in @cite{Guns, Germs, and Steel}, ``@dots{} zebras
3810 become impossibly dangerous as they grow older'' but the claim here is
3811 that they do not become fierce like a tiger. (1997, W. W. Norton and
3812 Co., ISBN 0-393-03894-2, page 171)}, and binds the
3813 variable @code{tiger} to the value @code{fierce}. In this example,
3814 both values are symbols preceded by a quote. The values could just as
3815 well have been another list or a string. The body of the @code{let}
3816 follows after the list holding the variables. In this example, the
3817 body is a list that uses the @code{message} function to print a string
3821 You may evaluate the example in the usual fashion, by placing the
3822 cursor after the last parenthesis and typing @kbd{C-x C-e}. When you do
3823 this, the following will appear in the echo area:
3826 "One kind of animal has stripes and another is fierce."
3829 As we have seen before, the @code{message} function prints its first
3830 argument, except for @samp{%s}. In this example, the value of the variable
3831 @code{zebra} is printed at the location of the first @samp{%s} and the
3832 value of the variable @code{tiger} is printed at the location of the
3835 @node Uninitialized let Variables, , Sample let Expression, let
3836 @comment node-name, next, previous, up
3837 @subsection Uninitialized Variables in a @code{let} Statement
3838 @cindex Uninitialized @code{let} variables
3839 @cindex @code{let} variables uninitialized
3841 If you do not bind the variables in a @code{let} statement to specific
3842 initial values, they will automatically be bound to an initial value of
3843 @code{nil}, as in the following expression:
3852 "Here are %d variables with %s, %s, and %s value."
3853 birch pine fir oak))
3858 Here, the varlist is @code{((birch 3) pine fir (oak 'some))}.
3861 If you evaluate this expression in the usual way, the following will
3862 appear in your echo area:
3865 "Here are 3 variables with nil, nil, and some value."
3869 In this example, Emacs binds the symbol @code{birch} to the number 3,
3870 binds the symbols @code{pine} and @code{fir} to @code{nil}, and binds
3871 the symbol @code{oak} to the value @code{some}.
3873 Note that in the first part of the @code{let}, the variables @code{pine}
3874 and @code{fir} stand alone as atoms that are not surrounded by
3875 parentheses; this is because they are being bound to @code{nil}, the
3876 empty list. But @code{oak} is bound to @code{some} and so is a part of
3877 the list @code{(oak 'some)}. Similarly, @code{birch} is bound to the
3878 number 3 and so is in a list with that number. (Since a number
3879 evaluates to itself, the number does not need to be quoted. Also, the
3880 number is printed in the message using a @samp{%d} rather than a
3881 @samp{%s}.) The four variables as a group are put into a list to
3882 delimit them from the body of the @code{let}.
3884 @node if, else, let, Writing Defuns
3885 @comment node-name, next, previous, up
3886 @section The @code{if} Special Form
3888 @cindex Conditional with @code{if}
3890 A third special form, in addition to @code{defun} and @code{let}, is the
3891 conditional @code{if}. This form is used to instruct the computer to
3892 make decisions. You can write function definitions without using
3893 @code{if}, but it is used often enough, and is important enough, to be
3894 included here. It is used, for example, in the code for the
3895 function @code{beginning-of-buffer}.
3897 The basic idea behind an @code{if}, is that ``@emph{if} a test is true,
3898 @emph{then} an expression is evaluated.'' If the test is not true, the
3899 expression is not evaluated. For example, you might make a decision
3900 such as, ``if it is warm and sunny, then go to the beach!''
3903 * if in more detail::
3904 * type-of-animal in detail:: An example of an @code{if} expression.
3907 @node if in more detail, type-of-animal in detail, if, if
3909 @unnumberedsubsec @code{if} in more detail
3912 @cindex @samp{if-part} defined
3913 @cindex @samp{then-part} defined
3914 An @code{if} expression written in Lisp does not use the word `then';
3915 the test and the action are the second and third elements of the list
3916 whose first element is @code{if}. Nonetheless, the test part of an
3917 @code{if} expression is often called the @dfn{if-part} and the second
3918 argument is often called the @dfn{then-part}.
3920 Also, when an @code{if} expression is written, the true-or-false-test
3921 is usually written on the same line as the symbol @code{if}, but the
3922 action to carry out if the test is true, the ``then-part'', is written
3923 on the second and subsequent lines. This makes the @code{if}
3924 expression easier to read.
3928 (if @var{true-or-false-test}
3929 @var{action-to-carry-out-if-test-is-true})
3934 The true-or-false-test will be an expression that
3935 is evaluated by the Lisp interpreter.
3937 Here is an example that you can evaluate in the usual manner. The test
3938 is whether the number 5 is greater than the number 4. Since it is, the
3939 message @samp{5 is greater than 4!} will be printed.
3943 (if (> 5 4) ; @r{if-part}
3944 (message "5 is greater than 4!")) ; @r{then-part}
3949 (The function @code{>} tests whether its first argument is greater than
3950 its second argument and returns true if it is.)
3951 @findex > (greater than)
3953 Of course, in actual use, the test in an @code{if} expression will not
3954 be fixed for all time as it is by the expression @code{(> 5 4)}.
3955 Instead, at least one of the variables used in the test will be bound to
3956 a value that is not known ahead of time. (If the value were known ahead
3957 of time, we would not need to run the test!)
3959 For example, the value may be bound to an argument of a function
3960 definition. In the following function definition, the character of the
3961 animal is a value that is passed to the function. If the value bound to
3962 @code{characteristic} is @code{fierce}, then the message, @samp{It's a
3963 tiger!} will be printed; otherwise, @code{nil} will be returned.
3967 (defun type-of-animal (characteristic)
3968 "Print message in echo area depending on CHARACTERISTIC.
3969 If the CHARACTERISTIC is the symbol `fierce',
3970 then warn of a tiger."
3971 (if (equal characteristic 'fierce)
3972 (message "It's a tiger!")))
3978 If you are reading this inside of GNU Emacs, you can evaluate the
3979 function definition in the usual way to install it in Emacs, and then you
3980 can evaluate the following two expressions to see the results:
3984 (type-of-animal 'fierce)
3986 (type-of-animal 'zebra)
3991 @c Following sentences rewritten to prevent overfull hbox.
3993 When you evaluate @code{(type-of-animal 'fierce)}, you will see the
3994 following message printed in the echo area: @code{"It's a tiger!"}; and
3995 when you evaluate @code{(type-of-animal 'zebra)} you will see @code{nil}
3996 printed in the echo area.
3998 @node type-of-animal in detail, , if in more detail, if
3999 @comment node-name, next, previous, up
4000 @subsection The @code{type-of-animal} Function in Detail
4002 Let's look at the @code{type-of-animal} function in detail.
4004 The function definition for @code{type-of-animal} was written by filling
4005 the slots of two templates, one for a function definition as a whole, and
4006 a second for an @code{if} expression.
4009 The template for every function that is not interactive is:
4013 (defun @var{name-of-function} (@var{argument-list})
4014 "@var{documentation}@dots{}"
4020 The parts of the function that match this template look like this:
4024 (defun type-of-animal (characteristic)
4025 "Print message in echo area depending on CHARACTERISTIC.
4026 If the CHARACTERISTIC is the symbol `fierce',
4027 then warn of a tiger."
4028 @var{body: the} @code{if} @var{expression})
4032 The name of function is @code{type-of-animal}; it is passed the value
4033 of one argument. The argument list is followed by a multi-line
4034 documentation string. The documentation string is included in the
4035 example because it is a good habit to write documentation string for
4036 every function definition. The body of the function definition
4037 consists of the @code{if} expression.
4040 The template for an @code{if} expression looks like this:
4044 (if @var{true-or-false-test}
4045 @var{action-to-carry-out-if-the-test-returns-true})
4050 In the @code{type-of-animal} function, the code for the @code{if}
4055 (if (equal characteristic 'fierce)
4056 (message "It's a tiger!")))
4061 Here, the true-or-false-test is the expression:
4064 (equal characteristic 'fierce)
4068 In Lisp, @code{equal} is a function that determines whether its first
4069 argument is equal to its second argument. The second argument is the
4070 quoted symbol @code{'fierce} and the first argument is the value of the
4071 symbol @code{characteristic}---in other words, the argument passed to
4074 In the first exercise of @code{type-of-animal}, the argument
4075 @code{fierce} is passed to @code{type-of-animal}. Since @code{fierce}
4076 is equal to @code{fierce}, the expression, @code{(equal characteristic
4077 'fierce)}, returns a value of true. When this happens, the @code{if}
4078 evaluates the second argument or then-part of the @code{if}:
4079 @code{(message "It's tiger!")}.
4081 On the other hand, in the second exercise of @code{type-of-animal}, the
4082 argument @code{zebra} is passed to @code{type-of-animal}. @code{zebra}
4083 is not equal to @code{fierce}, so the then-part is not evaluated and
4084 @code{nil} is returned by the @code{if} expression.
4086 @node else, Truth & Falsehood, if, Writing Defuns
4087 @comment node-name, next, previous, up
4088 @section If--then--else Expressions
4091 An @code{if} expression may have an optional third argument, called
4092 the @dfn{else-part}, for the case when the true-or-false-test returns
4093 false. When this happens, the second argument or then-part of the
4094 overall @code{if} expression is @emph{not} evaluated, but the third or
4095 else-part @emph{is} evaluated. You might think of this as the cloudy
4096 day alternative for the decision ``if it is warm and sunny, then go to
4097 the beach, else read a book!''.
4099 The word ``else'' is not written in the Lisp code; the else-part of an
4100 @code{if} expression comes after the then-part. In the written Lisp, the
4101 else-part is usually written to start on a line of its own and is
4102 indented less than the then-part:
4106 (if @var{true-or-false-test}
4107 @var{action-to-carry-out-if-the-test-returns-true}
4108 @var{action-to-carry-out-if-the-test-returns-false})
4112 For example, the following @code{if} expression prints the message @samp{4
4113 is not greater than 5!} when you evaluate it in the usual way:
4117 (if (> 4 5) ; @r{if-part}
4118 (message "5 is greater than 4!") ; @r{then-part}
4119 (message "4 is not greater than 5!")) ; @r{else-part}
4124 Note that the different levels of indentation make it easy to
4125 distinguish the then-part from the else-part. (GNU Emacs has several
4126 commands that automatically indent @code{if} expressions correctly.
4127 @xref{Typing Lists, , GNU Emacs Helps You Type Lists}.)
4129 We can extend the @code{type-of-animal} function to include an
4130 else-part by simply incorporating an additional part to the @code{if}
4134 You can see the consequences of doing this if you evaluate the following
4135 version of the @code{type-of-animal} function definition to install it
4136 and then evaluate the two subsequent expressions to pass different
4137 arguments to the function.
4141 (defun type-of-animal (characteristic) ; @r{Second version.}
4142 "Print message in echo area depending on CHARACTERISTIC.
4143 If the CHARACTERISTIC is the symbol `fierce',
4144 then warn of a tiger;
4145 else say it's not fierce."
4146 (if (equal characteristic 'fierce)
4147 (message "It's a tiger!")
4148 (message "It's not fierce!")))
4155 (type-of-animal 'fierce)
4157 (type-of-animal 'zebra)
4162 @c Following sentence rewritten to prevent overfull hbox.
4164 When you evaluate @code{(type-of-animal 'fierce)}, you will see the
4165 following message printed in the echo area: @code{"It's a tiger!"}; but
4166 when you evaluate @code{(type-of-animal 'zebra)}, you will see
4167 @code{"It's not fierce!"}.
4169 (Of course, if the @var{characteristic} were @code{ferocious}, the
4170 message @code{"It's not fierce!"} would be printed; and it would be
4171 misleading! When you write code, you need to take into account the
4172 possibility that some such argument will be tested by the @code{if}
4173 and write your program accordingly.)
4175 @node Truth & Falsehood, save-excursion, else, Writing Defuns
4176 @comment node-name, next, previous, up
4177 @section Truth and Falsehood in Emacs Lisp
4178 @cindex Truth and falsehood in Emacs Lisp
4179 @cindex Falsehood and truth in Emacs Lisp
4182 There is an important aspect to the truth test in an @code{if}
4183 expression. So far, we have spoken of `true' and `false' as values of
4184 predicates as if they were new kinds of Emacs Lisp objects. In fact,
4185 `false' is just our old friend @code{nil}. Anything else---anything
4188 The expression that tests for truth is interpreted as @dfn{true}
4189 if the result of evaluating it is a value that is not @code{nil}. In
4190 other words, the result of the test is considered true if the value
4191 returned is a number such as 47, a string such as @code{"hello"}, or a
4192 symbol (other than @code{nil}) such as @code{flowers}, or a list (so
4193 long as it is not empty), or even a buffer!
4196 * nil explained:: @code{nil} has two meanings.
4199 @node nil explained, , Truth & Falsehood, Truth & Falsehood
4201 @unnumberedsubsec An explanation of @code{nil}
4204 Before illustrating a test for truth, we need an explanation of @code{nil}.
4206 In Emacs Lisp, the symbol @code{nil} has two meanings. First, it means the
4207 empty list. Second, it means false and is the value returned when a
4208 true-or-false-test tests false. @code{nil} can be written as an empty
4209 list, @code{()}, or as @code{nil}. As far as the Lisp interpreter is
4210 concerned, @code{()} and @code{nil} are the same. Humans, however, tend
4211 to use @code{nil} for false and @code{()} for the empty list.
4213 In Emacs Lisp, any value that is not @code{nil}---is not the empty
4214 list---is considered true. This means that if an evaluation returns
4215 something that is not an empty list, an @code{if} expression will test
4216 true. For example, if a number is put in the slot for the test, it
4217 will be evaluated and will return itself, since that is what numbers
4218 do when evaluated. In this conditional, the @code{if} expression will
4219 test true. The expression tests false only when @code{nil}, an empty
4220 list, is returned by evaluating the expression.
4222 You can see this by evaluating the two expressions in the following examples.
4224 In the first example, the number 4 is evaluated as the test in the
4225 @code{if} expression and returns itself; consequently, the then-part
4226 of the expression is evaluated and returned: @samp{true} appears in
4227 the echo area. In the second example, the @code{nil} indicates false;
4228 consequently, the else-part of the expression is evaluated and
4229 returned: @samp{false} appears in the echo area.
4246 Incidentally, if some other useful value is not available for a test that
4247 returns true, then the Lisp interpreter will return the symbol @code{t}
4248 for true. For example, the expression @code{(> 5 4)} returns @code{t}
4249 when evaluated, as you can see by evaluating it in the usual way:
4257 On the other hand, this function returns @code{nil} if the test is false.
4263 @node save-excursion, Review, Truth & Falsehood, Writing Defuns
4264 @comment node-name, next, previous, up
4265 @section @code{save-excursion}
4266 @findex save-excursion
4267 @cindex Region, what it is
4268 @cindex Preserving point, mark, and buffer
4269 @cindex Point, mark, buffer preservation
4273 The @code{save-excursion} function is the fourth and final special form
4274 that we will discuss in this chapter.
4276 In Emacs Lisp programs used for editing, the @code{save-excursion}
4277 function is very common. It saves the location of point and mark,
4278 executes the body of the function, and then restores point and mark to
4279 their previous positions if their locations were changed. Its primary
4280 purpose is to keep the user from being surprised and disturbed by
4281 unexpected movement of point or mark.
4284 * Point and mark:: A review of various locations.
4285 * Template for save-excursion::
4288 @node Point and mark, Template for save-excursion, save-excursion, save-excursion
4290 @unnumberedsubsec Point and Mark
4293 Before discussing @code{save-excursion}, however, it may be useful
4294 first to review what point and mark are in GNU Emacs. @dfn{Point} is
4295 the current location of the cursor. Wherever the cursor
4296 is, that is point. More precisely, on terminals where the cursor
4297 appears to be on top of a character, point is immediately before the
4298 character. In Emacs Lisp, point is an integer. The first character in
4299 a buffer is number one, the second is number two, and so on. The
4300 function @code{point} returns the current position of the cursor as a
4301 number. Each buffer has its own value for point.
4303 The @dfn{mark} is another position in the buffer; its value can be set
4304 with a command such as @kbd{C-@key{SPC}} (@code{set-mark-command}). If
4305 a mark has been set, you can use the command @kbd{C-x C-x}
4306 (@code{exchange-point-and-mark}) to cause the cursor to jump to the mark
4307 and set the mark to be the previous position of point. In addition, if
4308 you set another mark, the position of the previous mark is saved in the
4309 mark ring. Many mark positions can be saved this way. You can jump the
4310 cursor to a saved mark by typing @kbd{C-u C-@key{SPC}} one or more
4313 The part of the buffer between point and mark is called @dfn{the
4314 region}. Numerous commands work on the region, including
4315 @code{center-region}, @code{count-lines-region}, @code{kill-region}, and
4316 @code{print-region}.
4318 The @code{save-excursion} special form saves the locations of point and
4319 mark and restores those positions after the code within the body of the
4320 special form is evaluated by the Lisp interpreter. Thus, if point were
4321 in the beginning of a piece of text and some code moved point to the end
4322 of the buffer, the @code{save-excursion} would put point back to where
4323 it was before, after the expressions in the body of the function were
4326 In Emacs, a function frequently moves point as part of its internal
4327 workings even though a user would not expect this. For example,
4328 @code{count-lines-region} moves point. To prevent the user from being
4329 bothered by jumps that are both unexpected and (from the user's point of
4330 view) unnecessary, @code{save-excursion} is often used to keep point and
4331 mark in the location expected by the user. The use of
4332 @code{save-excursion} is good housekeeping.
4334 To make sure the house stays clean, @code{save-excursion} restores the
4335 values of point and mark even if something goes wrong in the code inside
4336 of it (or, to be more precise and to use the proper jargon, ``in case of
4337 abnormal exit''). This feature is very helpful.
4339 In addition to recording the values of point and mark,
4340 @code{save-excursion} keeps track of the current buffer, and restores
4341 it, too. This means you can write code that will change the buffer and
4342 have @code{save-excursion} switch you back to the original buffer.
4343 This is how @code{save-excursion} is used in @code{append-to-buffer}.
4344 (@xref{append-to-buffer, , The Definition of @code{append-to-buffer}}.)
4346 @node Template for save-excursion, , Point and mark, save-excursion
4347 @comment node-name, next, previous, up
4348 @subsection Template for a @code{save-excursion} Expression
4351 The template for code using @code{save-excursion} is simple:
4361 The body of the function is one or more expressions that will be
4362 evaluated in sequence by the Lisp interpreter. If there is more than
4363 one expression in the body, the value of the last one will be returned
4364 as the value of the @code{save-excursion} function. The other
4365 expressions in the body are evaluated only for their side effects; and
4366 @code{save-excursion} itself is used only for its side effect (which
4367 is restoring the positions of point and mark).
4370 In more detail, the template for a @code{save-excursion} expression
4376 @var{first-expression-in-body}
4377 @var{second-expression-in-body}
4378 @var{third-expression-in-body}
4380 @var{last-expression-in-body})
4385 An expression, of course, may be a symbol on its own or a list.
4387 In Emacs Lisp code, a @code{save-excursion} expression often occurs
4388 within the body of a @code{let} expression. It looks like this:
4398 @node Review, defun Exercises, save-excursion, Writing Defuns
4399 @comment node-name, next, previous, up
4402 In the last few chapters we have introduced a fair number of functions
4403 and special forms. Here they are described in brief, along with a few
4404 similar functions that have not been mentioned yet.
4407 @item eval-last-sexp
4408 Evaluate the last symbolic expression before the current location of
4409 point. The value is printed in the echo area unless the function is
4410 invoked with an argument; in that case, the output is printed in the
4411 current buffer. This command is normally bound to @kbd{C-x C-e}.
4414 Define function. This special form has up to five parts: the name,
4415 a template for the arguments that will be passed to the function,
4416 documentation, an optional interactive declaration, and the body of the
4420 For example, in an early version of Emacs, the function definition was
4421 as follows. (It is slightly more complex now that it seeks the first
4422 non-whitespace character rather than the first visible character.)
4426 (defun back-to-indentation ()
4427 "Move point to first visible character on line."
4429 (beginning-of-line 1)
4430 (skip-chars-forward " \t"))
4437 (defun backward-to-indentation (&optional arg)
4438 "Move backward ARG lines and position at first nonblank character."
4440 (forward-line (- (or arg 1)))
4441 (skip-chars-forward " \t"))
4443 (defun back-to-indentation ()
4444 "Move point to the first non-whitespace character on this line."
4446 (beginning-of-line 1)
4447 (skip-syntax-forward " " (line-end-position))
4448 ;; Move back over chars that have whitespace syntax but have the p flag.
4449 (backward-prefix-chars))
4453 Declare to the interpreter that the function can be used
4454 interactively. This special form may be followed by a string with one
4455 or more parts that pass the information to the arguments of the
4456 function, in sequence. These parts may also tell the interpreter to
4457 prompt for information. Parts of the string are separated by
4458 newlines, @samp{\n}.
4461 Common code characters are:
4465 The name of an existing buffer.
4468 The name of an existing file.
4471 The numeric prefix argument. (Note that this `p' is lower case.)
4474 Point and the mark, as two numeric arguments, smallest first. This
4475 is the only code letter that specifies two successive arguments
4479 @xref{Interactive Codes, , Code Characters for @samp{interactive},
4480 elisp, The GNU Emacs Lisp Reference Manual}, for a complete list of
4484 Declare that a list of variables is for use within the body of the
4485 @code{let} and give them an initial value, either @code{nil} or a
4486 specified value; then evaluate the rest of the expressions in the body
4487 of the @code{let} and return the value of the last one. Inside the
4488 body of the @code{let}, the Lisp interpreter does not see the values of
4489 the variables of the same names that are bound outside of the
4497 (let ((foo (buffer-name))
4498 (bar (buffer-size)))
4500 "This buffer is %s and has %d characters."
4505 @item save-excursion
4506 Record the values of point and mark and the current buffer before
4507 evaluating the body of this special form. Restore the values of point
4508 and mark and buffer afterward.
4515 (message "We are %d characters into this buffer."
4518 (goto-char (point-min)) (point))))
4523 Evaluate the first argument to the function; if it is true, evaluate
4524 the second argument; else evaluate the third argument, if there is one.
4526 The @code{if} special form is called a @dfn{conditional}. There are
4527 other conditionals in Emacs Lisp, but @code{if} is perhaps the most
4535 (if (= 22 emacs-major-version)
4536 (message "This is version 22 Emacs")
4537 (message "This is not version 22 Emacs"))
4546 The @code{<} function tests whether its first argument is smaller than
4547 its second argument. A corresponding function, @code{>}, tests whether
4548 the first argument is greater than the second. Likewise, @code{<=}
4549 tests whether the first argument is less than or equal to the second and
4550 @code{>=} tests whether the first argument is greater than or equal to
4551 the second. In all cases, both arguments must be numbers or markers
4552 (markers indicate positions in buffers).
4556 The @code{=} function tests whether two arguments, both numbers or
4562 Test whether two objects are the same. @code{equal} uses one meaning
4563 of the word `same' and @code{eq} uses another: @code{equal} returns
4564 true if the two objects have a similar structure and contents, such as
4565 two copies of the same book. On the other hand, @code{eq}, returns
4566 true if both arguments are actually the same object.
4575 The @code{string-lessp} function tests whether its first argument is
4576 smaller than the second argument. A shorter, alternative name for the
4577 same function (a @code{defalias}) is @code{string<}.
4579 The arguments to @code{string-lessp} must be strings or symbols; the
4580 ordering is lexicographic, so case is significant. The print names of
4581 symbols are used instead of the symbols themselves.
4583 @cindex @samp{empty string} defined
4584 An empty string, @samp{""}, a string with no characters in it, is
4585 smaller than any string of characters.
4587 @code{string-equal} provides the corresponding test for equality. Its
4588 shorter, alternative name is @code{string=}. There are no string test
4589 functions that correspond to @var{>}, @code{>=}, or @code{<=}.
4592 Print a message in the echo area. The first argument is a string that
4593 can contain @samp{%s}, @samp{%d}, or @samp{%c} to print the value of
4594 arguments that follow the string. The argument used by @samp{%s} must
4595 be a string or a symbol; the argument used by @samp{%d} must be a
4596 number. The argument used by @samp{%c} must be an @sc{ascii} code
4597 number; it will be printed as the character with that @sc{ascii} code.
4598 (Various other %-sequences have not been mentioned.)
4602 The @code{setq} function sets the value of its first argument to the
4603 value of the second argument. The first argument is automatically
4604 quoted by @code{setq}. It does the same for succeeding pairs of
4605 arguments. Another function, @code{set}, takes only two arguments and
4606 evaluates both of them before setting the value returned by its first
4607 argument to the value returned by its second argument.
4610 Without an argument, return the name of the buffer, as a string.
4612 @itemx buffer-file-name
4613 Without an argument, return the name of the file the buffer is
4616 @item current-buffer
4617 Return the buffer in which Emacs is active; it may not be
4618 the buffer that is visible on the screen.
4621 Return the most recently selected buffer (other than the buffer passed
4622 to @code{other-buffer} as an argument and other than the current
4625 @item switch-to-buffer
4626 Select a buffer for Emacs to be active in and display it in the current
4627 window so users can look at it. Usually bound to @kbd{C-x b}.
4630 Switch Emacs' attention to a buffer on which programs will run. Don't
4631 alter what the window is showing.
4634 Return the number of characters in the current buffer.
4637 Return the value of the current position of the cursor, as an
4638 integer counting the number of characters from the beginning of the
4642 Return the minimum permissible value of point in
4643 the current buffer. This is 1, unless narrowing is in effect.
4646 Return the value of the maximum permissible value of point in the
4647 current buffer. This is the end of the buffer, unless narrowing is in
4652 @node defun Exercises, , Review, Writing Defuns
4657 Write a non-interactive function that doubles the value of its
4658 argument, a number. Make that function interactive.
4661 Write a function that tests whether the current value of
4662 @code{fill-column} is greater than the argument passed to the function,
4663 and if so, prints an appropriate message.
4666 @node Buffer Walk Through, More Complex, Writing Defuns, Top
4667 @comment node-name, next, previous, up
4668 @chapter A Few Buffer--Related Functions
4670 In this chapter we study in detail several of the functions used in GNU
4671 Emacs. This is called a ``walk-through''. These functions are used as
4672 examples of Lisp code, but are not imaginary examples; with the
4673 exception of the first, simplified function definition, these functions
4674 show the actual code used in GNU Emacs. You can learn a great deal from
4675 these definitions. The functions described here are all related to
4676 buffers. Later, we will study other functions.
4679 * Finding More:: How to find more information.
4680 * simplified-beginning-of-buffer:: Shows @code{goto-char},
4681 @code{point-min}, and @code{push-mark}.
4682 * mark-whole-buffer:: Almost the same as @code{beginning-of-buffer}.
4683 * append-to-buffer:: Uses @code{save-excursion} and
4684 @code{insert-buffer-substring}.
4685 * Buffer Related Review:: Review.
4686 * Buffer Exercises::
4689 @node Finding More, simplified-beginning-of-buffer, Buffer Walk Through, Buffer Walk Through
4690 @section Finding More Information
4692 @findex describe-function, @r{introduced}
4693 @cindex Find function documentation
4694 In this walk-through, I will describe each new function as we come to
4695 it, sometimes in detail and sometimes briefly. If you are interested,
4696 you can get the full documentation of any Emacs Lisp function at any
4697 time by typing @kbd{C-h f} and then the name of the function (and then
4698 @key{RET}). Similarly, you can get the full documentation for a
4699 variable by typing @kbd{C-h v} and then the name of the variable (and
4702 @cindex Find source of function
4703 @c In version 22, tells location both of C and of Emacs Lisp
4704 Also, @code{describe-function} will tell you the location of the
4705 function definition.
4707 Put point into the name of the file that contains the function and
4708 press the @key{RET} key. In this case, @key{RET} means
4709 @code{push-button} rather than `return' or `enter'. Emacs will take
4710 you directly to the function definition.
4715 If you move point over the file name and press
4716 the @key{RET} key, which in this case means @code{help-follow} rather
4717 than `return' or `enter', Emacs will take you directly to the function
4721 More generally, if you want to see a function in its original source
4722 file, you can use the @code{find-tags} function to jump to it.
4723 @code{find-tags} works with a wide variety of languages, not just
4724 Lisp, and C, and it works with non-programming text as well. For
4725 example, @code{find-tags} will jump to the various nodes in the
4726 Texinfo source file of this document.
4727 The @code{find-tags} function depends on `tags tables' that record
4728 the locations of the functions, variables, and other items to which
4729 @code{find-tags} jumps.
4731 To use the @code{find-tags} command, type @kbd{M-.} (i.e., press the
4732 period key while holding down the @key{META} key, or else type the
4733 @key{ESC} key and then type the period key), and then, at the prompt,
4734 type in the name of the function whose source code you want to see,
4735 such as @code{mark-whole-buffer}, and then type @key{RET}. Emacs will
4736 switch buffers and display the source code for the function on your
4737 screen. To switch back to your current buffer, type @kbd{C-x b
4738 @key{RET}}. (On some keyboards, the @key{META} key is labelled
4741 @c !!! 22.1.1 tags table location in this paragraph
4742 @cindex TAGS table, specifying
4744 Depending on how the initial default values of your copy of Emacs are
4745 set, you may also need to specify the location of your `tags table',
4746 which is a file called @file{TAGS}. For example, if you are
4747 interested in Emacs sources, the tags table you will most likely want,
4748 if it has already been created for you, will be in a subdirectory of
4749 the @file{/usr/local/share/emacs/} directory; thus you would use the
4750 @code{M-x visit-tags-table} command and specify a pathname such as
4751 @file{/usr/local/share/emacs/22.1.1/lisp/TAGS}. If the tags table
4752 has not already been created, you will have to create it yourself. It
4753 will in a file such as @file{/usr/local/src/emacs/src/TAGS}.
4756 To create a @file{TAGS} file in a specific directory, switch to that
4757 directory in Emacs using @kbd{M-x cd} command, or list the directory
4758 with @kbd{C-x d} (@code{dired}). Then run the compile command, with
4759 @w{@code{etags *.el}} as the command to execute:
4762 M-x compile RET etags *.el RET
4765 For more information, see @ref{etags, , Create Your Own @file{TAGS} File}.
4767 After you become more familiar with Emacs Lisp, you will find that you will
4768 frequently use @code{find-tags} to navigate your way around source code;
4769 and you will create your own @file{TAGS} tables.
4771 @cindex Library, as term for `file'
4772 Incidentally, the files that contain Lisp code are conventionally
4773 called @dfn{libraries}. The metaphor is derived from that of a
4774 specialized library, such as a law library or an engineering library,
4775 rather than a general library. Each library, or file, contains
4776 functions that relate to a particular topic or activity, such as
4777 @file{abbrev.el} for handling abbreviations and other typing
4778 shortcuts, and @file{help.el} for on-line help. (Sometimes several
4779 libraries provide code for a single activity, as the various
4780 @file{rmail@dots{}} files provide code for reading electronic mail.)
4781 In @cite{The GNU Emacs Manual}, you will see sentences such as ``The
4782 @kbd{C-h p} command lets you search the standard Emacs Lisp libraries
4783 by topic keywords.''
4785 @node simplified-beginning-of-buffer, mark-whole-buffer, Finding More, Buffer Walk Through
4786 @comment node-name, next, previous, up
4787 @section A Simplified @code{beginning-of-buffer} Definition
4788 @findex simplified-beginning-of-buffer
4790 The @code{beginning-of-buffer} command is a good function to start with
4791 since you are likely to be familiar with it and it is easy to
4792 understand. Used as an interactive command, @code{beginning-of-buffer}
4793 moves the cursor to the beginning of the buffer, leaving the mark at the
4794 previous position. It is generally bound to @kbd{M-<}.
4796 In this section, we will discuss a shortened version of the function
4797 that shows how it is most frequently used. This shortened function
4798 works as written, but it does not contain the code for a complex option.
4799 In another section, we will describe the entire function.
4800 (@xref{beginning-of-buffer, , Complete Definition of
4801 @code{beginning-of-buffer}}.)
4803 Before looking at the code, let's consider what the function
4804 definition has to contain: it must include an expression that makes
4805 the function interactive so it can be called by typing @kbd{M-x
4806 beginning-of-buffer} or by typing a keychord such as @kbd{M-<}; it
4807 must include code to leave a mark at the original position in the
4808 buffer; and it must include code to move the cursor to the beginning
4812 Here is the complete text of the shortened version of the function:
4816 (defun simplified-beginning-of-buffer ()
4817 "Move point to the beginning of the buffer;
4818 leave mark at previous position."
4821 (goto-char (point-min)))
4825 Like all function definitions, this definition has five parts following
4826 the special form @code{defun}:
4830 The name: in this example, @code{simplified-beginning-of-buffer}.
4833 A list of the arguments: in this example, an empty list, @code{()},
4836 The documentation string.
4839 The interactive expression.
4846 In this function definition, the argument list is empty; this means that
4847 this function does not require any arguments. (When we look at the
4848 definition for the complete function, we will see that it may be passed
4849 an optional argument.)
4851 The interactive expression tells Emacs that the function is intended to
4852 be used interactively. In this example, @code{interactive} does not have
4853 an argument because @code{simplified-beginning-of-buffer} does not
4857 The body of the function consists of the two lines:
4862 (goto-char (point-min))
4866 The first of these lines is the expression, @code{(push-mark)}. When
4867 this expression is evaluated by the Lisp interpreter, it sets a mark at
4868 the current position of the cursor, wherever that may be. The position
4869 of this mark is saved in the mark ring.
4871 The next line is @code{(goto-char (point-min))}. This expression
4872 jumps the cursor to the minimum point in the buffer, that is, to the
4873 beginning of the buffer (or to the beginning of the accessible portion
4874 of the buffer if it is narrowed. @xref{Narrowing & Widening, ,
4875 Narrowing and Widening}.)
4877 The @code{push-mark} command sets a mark at the place where the cursor
4878 was located before it was moved to the beginning of the buffer by the
4879 @code{(goto-char (point-min))} expression. Consequently, you can, if
4880 you wish, go back to where you were originally by typing @kbd{C-x C-x}.
4882 That is all there is to the function definition!
4884 @findex describe-function
4885 When you are reading code such as this and come upon an unfamiliar
4886 function, such as @code{goto-char}, you can find out what it does by
4887 using the @code{describe-function} command. To use this command, type
4888 @kbd{C-h f} and then type in the name of the function and press
4889 @key{RET}. The @code{describe-function} command will print the
4890 function's documentation string in a @file{*Help*} window. For
4891 example, the documentation for @code{goto-char} is:
4895 Set point to POSITION, a number or marker.
4896 Beginning of buffer is position (point-min), end is (point-max).
4901 The function's one argument is the desired position.
4904 (The prompt for @code{describe-function} will offer you the symbol
4905 under or preceding the cursor, so you can save typing by positioning
4906 the cursor right over or after the function and then typing @kbd{C-h f
4909 The @code{end-of-buffer} function definition is written in the same way as
4910 the @code{beginning-of-buffer} definition except that the body of the
4911 function contains the expression @code{(goto-char (point-max))} in place
4912 of @code{(goto-char (point-min))}.
4914 @node mark-whole-buffer, append-to-buffer, simplified-beginning-of-buffer, Buffer Walk Through
4915 @comment node-name, next, previous, up
4916 @section The Definition of @code{mark-whole-buffer}
4917 @findex mark-whole-buffer
4919 The @code{mark-whole-buffer} function is no harder to understand than the
4920 @code{simplified-beginning-of-buffer} function. In this case, however,
4921 we will look at the complete function, not a shortened version.
4923 The @code{mark-whole-buffer} function is not as commonly used as the
4924 @code{beginning-of-buffer} function, but is useful nonetheless: it
4925 marks a whole buffer as a region by putting point at the beginning and
4926 a mark at the end of the buffer. It is generally bound to @kbd{C-x
4930 * mark-whole-buffer overview::
4931 * Body of mark-whole-buffer:: Only three lines of code.
4934 @node mark-whole-buffer overview, Body of mark-whole-buffer, mark-whole-buffer, mark-whole-buffer
4936 @unnumberedsubsec An overview of @code{mark-whole-buffer}
4940 In GNU Emacs 22, the code for the complete function looks like this:
4944 (defun mark-whole-buffer ()
4945 "Put point at beginning and mark at end of buffer.
4946 You probably should not use this function in Lisp programs;
4947 it is usually a mistake for a Lisp function to use any subroutine
4948 that uses or sets the mark."
4951 (push-mark (point-max) nil t)
4952 (goto-char (point-min)))
4957 Like all other functions, the @code{mark-whole-buffer} function fits
4958 into the template for a function definition. The template looks like
4963 (defun @var{name-of-function} (@var{argument-list})
4964 "@var{documentation}@dots{}"
4965 (@var{interactive-expression}@dots{})
4970 Here is how the function works: the name of the function is
4971 @code{mark-whole-buffer}; it is followed by an empty argument list,
4972 @samp{()}, which means that the function does not require arguments.
4973 The documentation comes next.
4975 The next line is an @code{(interactive)} expression that tells Emacs
4976 that the function will be used interactively. These details are similar
4977 to the @code{simplified-beginning-of-buffer} function described in the
4981 @node Body of mark-whole-buffer, , mark-whole-buffer overview, mark-whole-buffer
4982 @comment node-name, next, previous, up
4983 @subsection Body of @code{mark-whole-buffer}
4985 The body of the @code{mark-whole-buffer} function consists of three
4992 (push-mark (point-max) nil t)
4993 (goto-char (point-min))
4997 The first of these lines is the expression, @code{(push-mark (point))}.
4999 This line does exactly the same job as the first line of the body of
5000 the @code{simplified-beginning-of-buffer} function, which is written
5001 @code{(push-mark)}. In both cases, the Lisp interpreter sets a mark
5002 at the current position of the cursor.
5004 I don't know why the expression in @code{mark-whole-buffer} is written
5005 @code{(push-mark (point))} and the expression in
5006 @code{beginning-of-buffer} is written @code{(push-mark)}. Perhaps
5007 whoever wrote the code did not know that the arguments for
5008 @code{push-mark} are optional and that if @code{push-mark} is not
5009 passed an argument, the function automatically sets mark at the
5010 location of point by default. Or perhaps the expression was written
5011 so as to parallel the structure of the next line. In any case, the
5012 line causes Emacs to determine the position of point and set a mark
5015 In earlier versions of GNU Emacs, the next line of
5016 @code{mark-whole-buffer} was @code{(push-mark (point-max))}. This
5017 expression sets a mark at the point in the buffer that has the highest
5018 number. This will be the end of the buffer (or, if the buffer is
5019 narrowed, the end of the accessible portion of the buffer.
5020 @xref{Narrowing & Widening, , Narrowing and Widening}, for more about
5021 narrowing.) After this mark has been set, the previous mark, the one
5022 set at point, is no longer set, but Emacs remembers its position, just
5023 as all other recent marks are always remembered. This means that you
5024 can, if you wish, go back to that position by typing @kbd{C-u
5028 In GNU Emacs 22, the @code{(point-max)} is slightly more complicated.
5032 (push-mark (point-max) nil t)
5036 The expression works nearly the same as before. It sets a mark at the
5037 highest numbered place in the buffer that it can. However, in this
5038 version, @code{push-mark} has two additional arguments. The second
5039 argument to @code{push-mark} is @code{nil}. This tells the function
5040 it @emph{should} display a message that says `Mark set' when it pushes
5041 the mark. The third argument is @code{t}. This tells
5042 @code{push-mark} to activate the mark when Transient Mark mode is
5043 turned on. Transient Mark mode highlights the currently active
5044 region. It is often turned off.
5046 Finally, the last line of the function is @code{(goto-char
5047 (point-min)))}. This is written exactly the same way as it is written
5048 in @code{beginning-of-buffer}. The expression moves the cursor to
5049 the minimum point in the buffer, that is, to the beginning of the buffer
5050 (or to the beginning of the accessible portion of the buffer). As a
5051 result of this, point is placed at the beginning of the buffer and mark
5052 is set at the end of the buffer. The whole buffer is, therefore, the
5055 @node append-to-buffer, Buffer Related Review, mark-whole-buffer, Buffer Walk Through
5056 @comment node-name, next, previous, up
5057 @section The Definition of @code{append-to-buffer}
5058 @findex append-to-buffer
5060 The @code{append-to-buffer} command is more complex than the
5061 @code{mark-whole-buffer} command. What it does is copy the region
5062 (that is, the part of the buffer between point and mark) from the
5063 current buffer to a specified buffer.
5066 * append-to-buffer overview::
5067 * append interactive:: A two part interactive expression.
5068 * append-to-buffer body:: Incorporates a @code{let} expression.
5069 * append save-excursion:: How the @code{save-excursion} works.
5072 @node append-to-buffer overview, append interactive, append-to-buffer, append-to-buffer
5074 @unnumberedsubsec An Overview of @code{append-to-buffer}
5077 @findex insert-buffer-substring
5078 The @code{append-to-buffer} command uses the
5079 @code{insert-buffer-substring} function to copy the region.
5080 @code{insert-buffer-substring} is described by its name: it takes a
5081 string of characters from part of a buffer, a ``substring'', and
5082 inserts them into another buffer.
5084 Most of @code{append-to-buffer} is
5085 concerned with setting up the conditions for
5086 @code{insert-buffer-substring} to work: the code must specify both the
5087 buffer to which the text will go, the window it comes from and goes
5088 to, and the region that will be copied.
5091 Here is the complete text of the function:
5095 (defun append-to-buffer (buffer start end)
5096 "Append to specified buffer the text of the region.
5097 It is inserted into that buffer before its point.
5101 When calling from a program, give three arguments:
5102 BUFFER (or buffer name), START and END.
5103 START and END specify the portion of the current buffer to be copied."
5105 (list (read-buffer "Append to buffer: " (other-buffer
5106 (current-buffer) t))
5107 (region-beginning) (region-end)))
5110 (let ((oldbuf (current-buffer)))
5112 (let* ((append-to (get-buffer-create buffer))
5113 (windows (get-buffer-window-list append-to t t))
5115 (set-buffer append-to)
5116 (setq point (point))
5117 (barf-if-buffer-read-only)
5118 (insert-buffer-substring oldbuf start end)
5119 (dolist (window windows)
5120 (when (= (window-point window) point)
5121 (set-window-point window (point))))))))
5125 The function can be understood by looking at it as a series of
5126 filled-in templates.
5128 The outermost template is for the function definition. In this
5129 function, it looks like this (with several slots filled in):
5133 (defun append-to-buffer (buffer start end)
5134 "@var{documentation}@dots{}"
5135 (interactive @dots{})
5140 The first line of the function includes its name and three arguments.
5141 The arguments are the @code{buffer} to which the text will be copied, and
5142 the @code{start} and @code{end} of the region in the current buffer that
5145 The next part of the function is the documentation, which is clear and
5146 complete. As is conventional, the three arguments are written in
5147 upper case so you will notice them easily. Even better, they are
5148 described in the same order as in the argument list.
5150 Note that the documentation distinguishes between a buffer and its
5151 name. (The function can handle either.)
5153 @node append interactive, append-to-buffer body, append-to-buffer overview, append-to-buffer
5154 @comment node-name, next, previous, up
5155 @subsection The @code{append-to-buffer} Interactive Expression
5157 Since the @code{append-to-buffer} function will be used interactively,
5158 the function must have an @code{interactive} expression. (For a
5159 review of @code{interactive}, see @ref{Interactive, , Making a
5160 Function Interactive}.) The expression reads as follows:
5166 "Append to buffer: "
5167 (other-buffer (current-buffer) t))
5174 This expression is not one with letters standing for parts, as
5175 described earlier. Instead, it starts a list with thee parts.
5177 The first part of the list is an expression to read the name of a
5178 buffer and return it as a string. That is @code{read-buffer}. The
5179 function requires a prompt as its first argument, @samp{"Append to
5180 buffer: "}. Its second argument tells the command what value to
5181 provide if you don't specify anything.
5183 In this case that second argument is an expression containing the
5184 function @code{other-buffer}, an exception, and a @samp{t}, standing
5187 The first argument to @code{other-buffer}, the exception, is yet
5188 another function, @code{current-buffer}. That is not going to be
5189 returned. The second argument is the symbol for true, @code{t}. that
5190 tells @code{other-buffer} that it may show visible buffers (except in
5191 this case, it will not show the current buffer, which makes sense).
5194 The expression looks like this:
5197 (other-buffer (current-buffer) t)
5200 The second and third arguments to the @code{list} expression are
5201 @code{(region-beginning)} and @code{(region-end)}. These two
5202 functions specify the beginning and end of the text to be appended.
5205 Originally, the command used the letters @samp{B} and @samp{r}.
5206 The whole @code{interactive} expression looked like this:
5209 (interactive "BAppend to buffer:@: \nr")
5213 But when that was done, the default value of the buffer switched to
5214 was invisible. That was not wanted.
5216 (The prompt was separated from the second argument with a newline,
5217 @samp{\n}. It was followed by an @samp{r} that told Emacs to bind the
5218 two arguments that follow the symbol @code{buffer} in the function's
5219 argument list (that is, @code{start} and @code{end}) to the values of
5220 point and mark. That argument worked fine.)
5222 @node append-to-buffer body, append save-excursion, append interactive, append-to-buffer
5223 @comment node-name, next, previous, up
5224 @subsection The Body of @code{append-to-buffer}
5227 in GNU Emacs 22 in /usr/local/src/emacs/lisp/simple.el
5229 (defun append-to-buffer (buffer start end)
5230 "Append to specified buffer the text of the region.
5231 It is inserted into that buffer before its point.
5233 When calling from a program, give three arguments:
5234 BUFFER (or buffer name), START and END.
5235 START and END specify the portion of the current buffer to be copied."
5237 (list (read-buffer "Append to buffer: " (other-buffer (current-buffer) t))
5238 (region-beginning) (region-end)))
5239 (let ((oldbuf (current-buffer)))
5241 (let* ((append-to (get-buffer-create buffer))
5242 (windows (get-buffer-window-list append-to t t))
5244 (set-buffer append-to)
5245 (setq point (point))
5246 (barf-if-buffer-read-only)
5247 (insert-buffer-substring oldbuf start end)
5248 (dolist (window windows)
5249 (when (= (window-point window) point)
5250 (set-window-point window (point))))))))
5253 The body of the @code{append-to-buffer} function begins with @code{let}.
5255 As we have seen before (@pxref{let, , @code{let}}), the purpose of a
5256 @code{let} expression is to create and give initial values to one or
5257 more variables that will only be used within the body of the
5258 @code{let}. This means that such a variable will not be confused with
5259 any variable of the same name outside the @code{let} expression.
5261 We can see how the @code{let} expression fits into the function as a
5262 whole by showing a template for @code{append-to-buffer} with the
5263 @code{let} expression in outline:
5267 (defun append-to-buffer (buffer start end)
5268 "@var{documentation}@dots{}"
5269 (interactive @dots{})
5270 (let ((@var{variable} @var{value}))
5275 The @code{let} expression has three elements:
5279 The symbol @code{let};
5282 A varlist containing, in this case, a single two-element list,
5283 @code{(@var{variable} @var{value})};
5286 The body of the @code{let} expression.
5290 In the @code{append-to-buffer} function, the varlist looks like this:
5293 (oldbuf (current-buffer))
5297 In this part of the @code{let} expression, the one variable,
5298 @code{oldbuf}, is bound to the value returned by the
5299 @code{(current-buffer)} expression. The variable, @code{oldbuf}, is
5300 used to keep track of the buffer in which you are working and from
5301 which you will copy.
5303 The element or elements of a varlist are surrounded by a set of
5304 parentheses so the Lisp interpreter can distinguish the varlist from
5305 the body of the @code{let}. As a consequence, the two-element list
5306 within the varlist is surrounded by a circumscribing set of parentheses.
5307 The line looks like this:
5311 (let ((oldbuf (current-buffer)))
5317 The two parentheses before @code{oldbuf} might surprise you if you did
5318 not realize that the first parenthesis before @code{oldbuf} marks the
5319 boundary of the varlist and the second parenthesis marks the beginning
5320 of the two-element list, @code{(oldbuf (current-buffer))}.
5322 @node append save-excursion, , append-to-buffer body, append-to-buffer
5323 @comment node-name, next, previous, up
5324 @subsection @code{save-excursion} in @code{append-to-buffer}
5326 The body of the @code{let} expression in @code{append-to-buffer}
5327 consists of a @code{save-excursion} expression.
5329 The @code{save-excursion} function saves the locations of point and
5330 mark, and restores them to those positions after the expressions in the
5331 body of the @code{save-excursion} complete execution. In addition,
5332 @code{save-excursion} keeps track of the original buffer, and
5333 restores it. This is how @code{save-excursion} is used in
5334 @code{append-to-buffer}.
5337 @cindex Indentation for formatting
5338 @cindex Formatting convention
5339 Incidentally, it is worth noting here that a Lisp function is normally
5340 formatted so that everything that is enclosed in a multi-line spread is
5341 indented more to the right than the first symbol. In this function
5342 definition, the @code{let} is indented more than the @code{defun}, and
5343 the @code{save-excursion} is indented more than the @code{let}, like
5359 This formatting convention makes it easy to see that the lines in
5360 the body of the @code{save-excursion} are enclosed by the parentheses
5361 associated with @code{save-excursion}, just as the
5362 @code{save-excursion} itself is enclosed by the parentheses associated
5363 with the @code{let}:
5367 (let ((oldbuf (current-buffer)))
5370 (set-buffer @dots{})
5371 (insert-buffer-substring oldbuf start end)
5377 The use of the @code{save-excursion} function can be viewed as a process
5378 of filling in the slots of a template:
5383 @var{first-expression-in-body}
5384 @var{second-expression-in-body}
5386 @var{last-expression-in-body})
5392 In this function, the body of the @code{save-excursion} contains only
5393 one expression, the @code{let*} expression. You know about a
5394 @code{let} function. The @code{let*} function is different. It has a
5395 @samp{*} in its name. It enables Emacs to set each variable in its
5396 varlist in sequence, one after another.
5398 Its critical feature is that variables later in the varlist can make
5399 use of the values to which Emacs set variables earlier in the varlist.
5400 @xref{fwd-para let, , The @code{let*} expression}.
5402 We will skip functions like @code{let*} and focus on two: the
5403 @code{set-buffer} function and the @code{insert-buffer-substring}
5407 In the old days, the @code{set-buffer} expression was simply
5410 (set-buffer (get-buffer-create buffer))
5418 (set-buffer append-to)
5422 @code{append-to} is bound to @code{(get-buffer-create buffer)} earlier
5423 on in the @code{let*} expression. That extra binding would not be
5424 necessary except for that @code{append-to} is used later in the
5425 varlist as an argument to @code{get-buffer-window-list}.
5430 (let ((oldbuf (current-buffer)))
5432 (let* ((append-to (get-buffer-create buffer))
5433 (windows (get-buffer-window-list append-to t t))
5435 (set-buffer append-to)
5436 (setq point (point))
5437 (barf-if-buffer-read-only)
5438 (insert-buffer-substring oldbuf start end)
5439 (dolist (window windows)
5440 (when (= (window-point window) point)
5441 (set-window-point window (point))))))))
5444 The @code{append-to-buffer} function definition inserts text from the
5445 buffer in which you are currently to a named buffer. It happens that
5446 @code{insert-buffer-substring} copies text from another buffer to the
5447 current buffer, just the reverse---that is why the
5448 @code{append-to-buffer} definition starts out with a @code{let} that
5449 binds the local symbol @code{oldbuf} to the value returned by
5450 @code{current-buffer}.
5453 The @code{insert-buffer-substring} expression looks like this:
5456 (insert-buffer-substring oldbuf start end)
5460 The @code{insert-buffer-substring} function copies a string
5461 @emph{from} the buffer specified as its first argument and inserts the
5462 string into the present buffer. In this case, the argument to
5463 @code{insert-buffer-substring} is the value of the variable created
5464 and bound by the @code{let}, namely the value of @code{oldbuf}, which
5465 was the current buffer when you gave the @code{append-to-buffer}
5468 After @code{insert-buffer-substring} has done its work,
5469 @code{save-excursion} will restore the action to the original buffer
5470 and @code{append-to-buffer} will have done its job.
5473 Written in skeletal form, the workings of the body look like this:
5477 (let (@var{bind-}@code{oldbuf}@var{-to-value-of-}@code{current-buffer})
5478 (save-excursion ; @r{Keep track of buffer.}
5480 @var{insert-substring-from-}@code{oldbuf}@var{-into-buffer})
5482 @var{change-back-to-original-buffer-when-finished}
5483 @var{let-the-local-meaning-of-}@code{oldbuf}@var{-disappear-when-finished}
5487 In summary, @code{append-to-buffer} works as follows: it saves the
5488 value of the current buffer in the variable called @code{oldbuf}. It
5489 gets the new buffer (creating one if need be) and switches Emacs'
5490 attention to it. Using the value of @code{oldbuf}, it inserts the
5491 region of text from the old buffer into the new buffer; and then using
5492 @code{save-excursion}, it brings you back to your original buffer.
5494 In looking at @code{append-to-buffer}, you have explored a fairly
5495 complex function. It shows how to use @code{let} and
5496 @code{save-excursion}, and how to change to and come back from another
5497 buffer. Many function definitions use @code{let},
5498 @code{save-excursion}, and @code{set-buffer} this way.
5500 @node Buffer Related Review, Buffer Exercises, append-to-buffer, Buffer Walk Through
5501 @comment node-name, next, previous, up
5504 Here is a brief summary of the various functions discussed in this chapter.
5507 @item describe-function
5508 @itemx describe-variable
5509 Print the documentation for a function or variable.
5510 Conventionally bound to @kbd{C-h f} and @kbd{C-h v}.
5513 Find the file containing the source for a function or variable and
5514 switch buffers to it, positioning point at the beginning of the item.
5515 Conventionally bound to @kbd{M-.} (that's a period following the
5518 @item save-excursion
5519 Save the location of point and mark and restore their values after the
5520 arguments to @code{save-excursion} have been evaluated. Also, remember
5521 the current buffer and return to it.
5524 Set mark at a location and record the value of the previous mark on the
5525 mark ring. The mark is a location in the buffer that will keep its
5526 relative position even if text is added to or removed from the buffer.
5529 Set point to the location specified by the value of the argument, which
5530 can be a number, a marker, or an expression that returns the number of
5531 a position, such as @code{(point-min)}.
5533 @item insert-buffer-substring
5534 Copy a region of text from a buffer that is passed to the function as
5535 an argument and insert the region into the current buffer.
5537 @item mark-whole-buffer
5538 Mark the whole buffer as a region. Normally bound to @kbd{C-x h}.
5541 Switch the attention of Emacs to another buffer, but do not change the
5542 window being displayed. Used when the program rather than a human is
5543 to work on a different buffer.
5545 @item get-buffer-create
5547 Find a named buffer or create one if a buffer of that name does not
5548 exist. The @code{get-buffer} function returns @code{nil} if the named
5549 buffer does not exist.
5553 @node Buffer Exercises, , Buffer Related Review, Buffer Walk Through
5558 Write your own @code{simplified-end-of-buffer} function definition;
5559 then test it to see whether it works.
5562 Use @code{if} and @code{get-buffer} to write a function that prints a
5563 message telling you whether a buffer exists.
5566 Using @code{find-tag}, find the source for the @code{copy-to-buffer}
5570 @node More Complex, Narrowing & Widening, Buffer Walk Through, Top
5571 @comment node-name, next, previous, up
5572 @chapter A Few More Complex Functions
5574 In this chapter, we build on what we have learned in previous chapters
5575 by looking at more complex functions. The @code{copy-to-buffer}
5576 function illustrates use of two @code{save-excursion} expressions in
5577 one definition, while the @code{insert-buffer} function illustrates
5578 use of an asterisk in an @code{interactive} expression, use of
5579 @code{or}, and the important distinction between a name and the object
5580 to which the name refers.
5583 * copy-to-buffer:: With @code{set-buffer}, @code{get-buffer-create}.
5584 * insert-buffer:: Read-only, and with @code{or}.
5585 * beginning-of-buffer:: Shows @code{goto-char},
5586 @code{point-min}, and @code{push-mark}.
5587 * Second Buffer Related Review::
5588 * optional Exercise::
5591 @node copy-to-buffer, insert-buffer, More Complex, More Complex
5592 @comment node-name, next, previous, up
5593 @section The Definition of @code{copy-to-buffer}
5594 @findex copy-to-buffer
5596 After understanding how @code{append-to-buffer} works, it is easy to
5597 understand @code{copy-to-buffer}. This function copies text into a
5598 buffer, but instead of adding to the second buffer, it replaces all the
5599 previous text in the second buffer.
5602 The body of @code{copy-to-buffer} looks like this,
5607 (interactive "BCopy to buffer: \nr")
5608 (let ((oldbuf (current-buffer)))
5609 (with-current-buffer (get-buffer-create buffer)
5610 (barf-if-buffer-read-only)
5613 (insert-buffer-substring oldbuf start end)))))
5617 The @code{copy-to-buffer} function has a simpler @code{interactive}
5618 expression than @code{append-to-buffer}.
5621 The definition then says
5624 (with-current-buffer (get-buffer-create buffer) @dots{}
5627 First, look at the earliest inner expression; that is evaluated first.
5628 That expression starts with @code{get-buffer-create buffer}. The
5629 function tells the computer to use the buffer with the name specified
5630 as the one to which you are copying, or if such a buffer does not
5631 exist, to create it. Then, the @code{with-current-buffer} function
5632 evaluates its body with that buffer temporarily current.
5634 (This demonstrates another way to shift the computer's attention but
5635 not the user's. The @code{append-to-buffer} function showed how to do
5636 the same with @code{save-excursion} and @code{set-buffer}.
5637 @code{with-current-buffer} is a newer, and arguably easier,
5640 The @code{barf-if-buffer-read-only} function sends you an error
5641 message saying the buffer is read-only if you cannot modify it.
5643 The next line has the @code{erase-buffer} function as its sole
5644 contents. That function erases the buffer.
5646 Finally, the last two lines contain the @code{save-excursion}
5647 expression with @code{insert-buffer-substring} as its body.
5648 The @code{insert-buffer-substring} expression copies the text from
5649 the buffer you are in (and you have not seen the computer shift its
5650 attention, so you don't know that that buffer is now called
5653 Incidentally, this is what is meant by `replacement'. To replace text,
5654 Emacs erases the previous text and then inserts new text.
5657 In outline, the body of @code{copy-to-buffer} looks like this:
5661 (let (@var{bind-}@code{oldbuf}@var{-to-value-of-}@code{current-buffer})
5662 (@var{with-the-buffer-you-are-copying-to}
5663 (@var{but-do-not-erase-or-copy-to-a-read-only-buffer})
5666 @var{insert-substring-from-}@code{oldbuf}@var{-into-buffer})))
5670 @node insert-buffer, beginning-of-buffer, copy-to-buffer, More Complex
5671 @comment node-name, next, previous, up
5672 @section The Definition of @code{insert-buffer}
5673 @findex insert-buffer
5675 @code{insert-buffer} is yet another buffer-related function. This
5676 command copies another buffer @emph{into} the current buffer. It is the
5677 reverse of @code{append-to-buffer} or @code{copy-to-buffer}, since they
5678 copy a region of text @emph{from} the current buffer to another buffer.
5680 Here is a discussion based on the original code. The code was
5681 simplified in 2003 and is harder to understand.
5683 (@xref{New insert-buffer, , New Body for @code{insert-buffer}}, to see
5684 a discussion of the new body.)
5686 In addition, this code illustrates the use of @code{interactive} with a
5687 buffer that might be @dfn{read-only} and the important distinction
5688 between the name of an object and the object actually referred to.
5691 * insert-buffer code::
5692 * insert-buffer interactive:: When you can read, but not write.
5693 * insert-buffer body:: The body has an @code{or} and a @code{let}.
5694 * if & or:: Using an @code{if} instead of an @code{or}.
5695 * Insert or:: How the @code{or} expression works.
5696 * Insert let:: Two @code{save-excursion} expressions.
5697 * New insert-buffer::
5700 @node insert-buffer code, insert-buffer interactive, insert-buffer, insert-buffer
5702 @unnumberedsubsec The Code for @code{insert-buffer}
5706 Here is the earlier code:
5710 (defun insert-buffer (buffer)
5711 "Insert after point the contents of BUFFER.
5712 Puts mark after the inserted text.
5713 BUFFER may be a buffer or a buffer name."
5714 (interactive "*bInsert buffer:@: ")
5717 (or (bufferp buffer)
5718 (setq buffer (get-buffer buffer)))
5719 (let (start end newmark)
5723 (setq start (point-min) end (point-max)))
5726 (insert-buffer-substring buffer start end)
5727 (setq newmark (point)))
5728 (push-mark newmark)))
5733 As with other function definitions, you can use a template to see an
5734 outline of the function:
5738 (defun insert-buffer (buffer)
5739 "@var{documentation}@dots{}"
5740 (interactive "*bInsert buffer:@: ")
5745 @node insert-buffer interactive, insert-buffer body, insert-buffer code, insert-buffer
5746 @comment node-name, next, previous, up
5747 @subsection The Interactive Expression in @code{insert-buffer}
5748 @findex interactive, @r{example use of}
5750 In @code{insert-buffer}, the argument to the @code{interactive}
5751 declaration has two parts, an asterisk, @samp{*}, and @samp{bInsert
5755 * Read-only buffer:: When a buffer cannot be modified.
5756 * b for interactive:: An existing buffer or else its name.
5759 @node Read-only buffer, b for interactive, insert-buffer interactive, insert-buffer interactive
5760 @comment node-name, next, previous, up
5761 @unnumberedsubsubsec A Read-only Buffer
5762 @cindex Read-only buffer
5763 @cindex Asterisk for read-only buffer
5764 @findex * @r{for read-only buffer}
5766 The asterisk is for the situation when the current buffer is a
5767 read-only buffer---a buffer that cannot be modified. If
5768 @code{insert-buffer} is called when the current buffer is read-only, a
5769 message to this effect is printed in the echo area and the terminal
5770 may beep or blink at you; you will not be permitted to insert anything
5771 into current buffer. The asterisk does not need to be followed by a
5772 newline to separate it from the next argument.
5774 @node b for interactive, , Read-only buffer, insert-buffer interactive
5775 @comment node-name, next, previous, up
5776 @unnumberedsubsubsec @samp{b} in an Interactive Expression
5778 The next argument in the interactive expression starts with a lower
5779 case @samp{b}. (This is different from the code for
5780 @code{append-to-buffer}, which uses an upper-case @samp{B}.
5781 @xref{append-to-buffer, , The Definition of @code{append-to-buffer}}.)
5782 The lower-case @samp{b} tells the Lisp interpreter that the argument
5783 for @code{insert-buffer} should be an existing buffer or else its
5784 name. (The upper-case @samp{B} option provides for the possibility
5785 that the buffer does not exist.) Emacs will prompt you for the name
5786 of the buffer, offering you a default buffer, with name completion
5787 enabled. If the buffer does not exist, you receive a message that
5788 says ``No match''; your terminal may beep at you as well.
5790 The new and simplified code generates a list for @code{interactive}.
5791 It uses the @code{barf-if-buffer-read-only} and @code{read-buffer}
5792 functions with which we are already familiar and the @code{progn}
5793 special form with which we are not. (It will be described later.)
5795 @node insert-buffer body, if & or, insert-buffer interactive, insert-buffer
5796 @comment node-name, next, previous, up
5797 @subsection The Body of the @code{insert-buffer} Function
5799 The body of the @code{insert-buffer} function has two major parts: an
5800 @code{or} expression and a @code{let} expression. The purpose of the
5801 @code{or} expression is to ensure that the argument @code{buffer} is
5802 bound to a buffer and not just the name of a buffer. The body of the
5803 @code{let} expression contains the code which copies the other buffer
5804 into the current buffer.
5807 In outline, the two expressions fit into the @code{insert-buffer}
5812 (defun insert-buffer (buffer)
5813 "@var{documentation}@dots{}"
5814 (interactive "*bInsert buffer:@: ")
5819 (let (@var{varlist})
5820 @var{body-of-}@code{let}@dots{} )
5824 To understand how the @code{or} expression ensures that the argument
5825 @code{buffer} is bound to a buffer and not to the name of a buffer, it
5826 is first necessary to understand the @code{or} function.
5828 Before doing this, let me rewrite this part of the function using
5829 @code{if} so that you can see what is done in a manner that will be familiar.
5831 @node if & or, Insert or, insert-buffer body, insert-buffer
5832 @comment node-name, next, previous, up
5833 @subsection @code{insert-buffer} With an @code{if} Instead of an @code{or}
5835 The job to be done is to make sure the value of @code{buffer} is a
5836 buffer itself and not the name of a buffer. If the value is the name,
5837 then the buffer itself must be got.
5839 You can imagine yourself at a conference where an usher is wandering
5840 around holding a list with your name on it and looking for you: the
5841 usher is ``bound'' to your name, not to you; but when the usher finds
5842 you and takes your arm, the usher becomes ``bound'' to you.
5845 In Lisp, you might describe this situation like this:
5849 (if (not (holding-on-to-guest))
5850 (find-and-take-arm-of-guest))
5854 We want to do the same thing with a buffer---if we do not have the
5855 buffer itself, we want to get it.
5858 Using a predicate called @code{bufferp} that tells us whether we have a
5859 buffer (rather than its name), we can write the code like this:
5863 (if (not (bufferp buffer)) ; @r{if-part}
5864 (setq buffer (get-buffer buffer))) ; @r{then-part}
5869 Here, the true-or-false-test of the @code{if} expression is
5870 @w{@code{(not (bufferp buffer))}}; and the then-part is the expression
5871 @w{@code{(setq buffer (get-buffer buffer))}}.
5873 In the test, the function @code{bufferp} returns true if its argument is
5874 a buffer---but false if its argument is the name of the buffer. (The
5875 last character of the function name @code{bufferp} is the character
5876 @samp{p}; as we saw earlier, such use of @samp{p} is a convention that
5877 indicates that the function is a predicate, which is a term that means
5878 that the function will determine whether some property is true or false.
5879 @xref{Wrong Type of Argument, , Using the Wrong Type Object as an
5883 The function @code{not} precedes the expression @code{(bufferp buffer)},
5884 so the true-or-false-test looks like this:
5887 (not (bufferp buffer))
5891 @code{not} is a function that returns true if its argument is false
5892 and false if its argument is true. So if @code{(bufferp buffer)}
5893 returns true, the @code{not} expression returns false and vice-verse:
5894 what is ``not true'' is false and what is ``not false'' is true.
5896 Using this test, the @code{if} expression works as follows: when the
5897 value of the variable @code{buffer} is actually a buffer rather than
5898 its name, the true-or-false-test returns false and the @code{if}
5899 expression does not evaluate the then-part. This is fine, since we do
5900 not need to do anything to the variable @code{buffer} if it really is
5903 On the other hand, when the value of @code{buffer} is not a buffer
5904 itself, but the name of a buffer, the true-or-false-test returns true
5905 and the then-part of the expression is evaluated. In this case, the
5906 then-part is @code{(setq buffer (get-buffer buffer))}. This
5907 expression uses the @code{get-buffer} function to return an actual
5908 buffer itself, given its name. The @code{setq} then sets the variable
5909 @code{buffer} to the value of the buffer itself, replacing its previous
5910 value (which was the name of the buffer).
5912 @node Insert or, Insert let, if & or, insert-buffer
5913 @comment node-name, next, previous, up
5914 @subsection The @code{or} in the Body
5916 The purpose of the @code{or} expression in the @code{insert-buffer}
5917 function is to ensure that the argument @code{buffer} is bound to a
5918 buffer and not just to the name of a buffer. The previous section shows
5919 how the job could have been done using an @code{if} expression.
5920 However, the @code{insert-buffer} function actually uses @code{or}.
5921 To understand this, it is necessary to understand how @code{or} works.
5924 An @code{or} function can have any number of arguments. It evaluates
5925 each argument in turn and returns the value of the first of its
5926 arguments that is not @code{nil}. Also, and this is a crucial feature
5927 of @code{or}, it does not evaluate any subsequent arguments after
5928 returning the first non-@code{nil} value.
5931 The @code{or} expression looks like this:
5935 (or (bufferp buffer)
5936 (setq buffer (get-buffer buffer)))
5941 The first argument to @code{or} is the expression @code{(bufferp buffer)}.
5942 This expression returns true (a non-@code{nil} value) if the buffer is
5943 actually a buffer, and not just the name of a buffer. In the @code{or}
5944 expression, if this is the case, the @code{or} expression returns this
5945 true value and does not evaluate the next expression---and this is fine
5946 with us, since we do not want to do anything to the value of
5947 @code{buffer} if it really is a buffer.
5949 On the other hand, if the value of @code{(bufferp buffer)} is @code{nil},
5950 which it will be if the value of @code{buffer} is the name of a buffer,
5951 the Lisp interpreter evaluates the next element of the @code{or}
5952 expression. This is the expression @code{(setq buffer (get-buffer
5953 buffer))}. This expression returns a non-@code{nil} value, which
5954 is the value to which it sets the variable @code{buffer}---and this
5955 value is a buffer itself, not the name of a buffer.
5957 The result of all this is that the symbol @code{buffer} is always
5958 bound to a buffer itself rather than to the name of a buffer. All
5959 this is necessary because the @code{set-buffer} function in a
5960 following line only works with a buffer itself, not with the name to a
5964 Incidentally, using @code{or}, the situation with the usher would be
5968 (or (holding-on-to-guest) (find-and-take-arm-of-guest))
5971 @node Insert let, New insert-buffer, Insert or, insert-buffer
5972 @comment node-name, next, previous, up
5973 @subsection The @code{let} Expression in @code{insert-buffer}
5975 After ensuring that the variable @code{buffer} refers to a buffer itself
5976 and not just to the name of a buffer, the @code{insert-buffer function}
5977 continues with a @code{let} expression. This specifies three local
5978 variables, @code{start}, @code{end}, and @code{newmark} and binds them
5979 to the initial value @code{nil}. These variables are used inside the
5980 remainder of the @code{let} and temporarily hide any other occurrence of
5981 variables of the same name in Emacs until the end of the @code{let}.
5984 The body of the @code{let} contains two @code{save-excursion}
5985 expressions. First, we will look at the inner @code{save-excursion}
5986 expression in detail. The expression looks like this:
5992 (setq start (point-min) end (point-max)))
5997 The expression @code{(set-buffer buffer)} changes Emacs' attention
5998 from the current buffer to the one from which the text will copied.
5999 In that buffer, the variables @code{start} and @code{end} are set to
6000 the beginning and end of the buffer, using the commands
6001 @code{point-min} and @code{point-max}. Note that we have here an
6002 illustration of how @code{setq} is able to set two variables in the
6003 same expression. The first argument of @code{setq} is set to the
6004 value of its second, and its third argument is set to the value of its
6007 After the body of the inner @code{save-excursion} is evaluated, the
6008 @code{save-excursion} restores the original buffer, but @code{start} and
6009 @code{end} remain set to the values of the beginning and end of the
6010 buffer from which the text will be copied.
6013 The outer @code{save-excursion} expression looks like this:
6018 (@var{inner-}@code{save-excursion}@var{-expression}
6019 (@var{go-to-new-buffer-and-set-}@code{start}@var{-and-}@code{end})
6020 (insert-buffer-substring buffer start end)
6021 (setq newmark (point)))
6026 The @code{insert-buffer-substring} function copies the text
6027 @emph{into} the current buffer @emph{from} the region indicated by
6028 @code{start} and @code{end} in @code{buffer}. Since the whole of the
6029 second buffer lies between @code{start} and @code{end}, the whole of
6030 the second buffer is copied into the buffer you are editing. Next,
6031 the value of point, which will be at the end of the inserted text, is
6032 recorded in the variable @code{newmark}.
6034 After the body of the outer @code{save-excursion} is evaluated, point
6035 and mark are relocated to their original places.
6037 However, it is convenient to locate a mark at the end of the newly
6038 inserted text and locate point at its beginning. The @code{newmark}
6039 variable records the end of the inserted text. In the last line of
6040 the @code{let} expression, the @code{(push-mark newmark)} expression
6041 function sets a mark to this location. (The previous location of the
6042 mark is still accessible; it is recorded on the mark ring and you can
6043 go back to it with @kbd{C-u C-@key{SPC}}.) Meanwhile, point is
6044 located at the beginning of the inserted text, which is where it was
6045 before you called the insert function, the position of which was saved
6046 by the first @code{save-excursion}.
6049 The whole @code{let} expression looks like this:
6053 (let (start end newmark)
6057 (setq start (point-min) end (point-max)))
6058 (insert-buffer-substring buffer start end)
6059 (setq newmark (point)))
6060 (push-mark newmark))
6064 Like the @code{append-to-buffer} function, the @code{insert-buffer}
6065 function uses @code{let}, @code{save-excursion}, and
6066 @code{set-buffer}. In addition, the function illustrates one way to
6067 use @code{or}. All these functions are building blocks that we will
6068 find and use again and again.
6070 @node New insert-buffer, , Insert let, insert-buffer
6071 @comment node-name, next, previous, up
6072 @subsection New Body for @code{insert-buffer}
6073 @findex insert-buffer, new version body
6074 @findex new version body for insert-buffer
6076 The body in the GNU Emacs 22 version is more confusing than the original.
6079 It consists of two expressions,
6085 (insert-buffer-substring (get-buffer buffer))
6093 except, and this is what confuses novices, very important work is done
6094 inside the @code{push-mark} expression.
6096 The @code{get-buffer} function returns a buffer with the name
6097 provided. You will note that the function is @emph{not} called
6098 @code{get-buffer-create}; it does not create a buffer if one does not
6099 already exist. The buffer returned by @code{get-buffer}, an existing
6100 buffer, is passed to @code{insert-buffer-substring}, which inserts the
6101 whole of the buffer (since you did not specify anything else).
6103 The location into which the buffer is inserted is recorded by
6104 @code{push-mark}. Then the function returns @code{nil}, the value of
6105 its last command. Put another way, the @code{insert-buffer} function
6106 exists only to produce a side effect, inserting another buffer, not to
6109 @node beginning-of-buffer, Second Buffer Related Review, insert-buffer, More Complex
6110 @comment node-name, next, previous, up
6111 @section Complete Definition of @code{beginning-of-buffer}
6112 @findex beginning-of-buffer
6114 The basic structure of the @code{beginning-of-buffer} function has
6115 already been discussed. (@xref{simplified-beginning-of-buffer, , A
6116 Simplified @code{beginning-of-buffer} Definition}.)
6117 This section describes the complex part of the definition.
6119 As previously described, when invoked without an argument,
6120 @code{beginning-of-buffer} moves the cursor to the beginning of the
6121 buffer (in truth, the beginning of the accessible portion of the
6122 buffer), leaving the mark at the previous position. However, when the
6123 command is invoked with a number between one and ten, the function
6124 considers that number to be a fraction of the length of the buffer,
6125 measured in tenths, and Emacs moves the cursor that fraction of the
6126 way from the beginning of the buffer. Thus, you can either call this
6127 function with the key command @kbd{M-<}, which will move the cursor to
6128 the beginning of the buffer, or with a key command such as @kbd{C-u 7
6129 M-<} which will move the cursor to a point 70% of the way through the
6130 buffer. If a number bigger than ten is used for the argument, it
6131 moves to the end of the buffer.
6133 The @code{beginning-of-buffer} function can be called with or without an
6134 argument. The use of the argument is optional.
6137 * Optional Arguments::
6138 * beginning-of-buffer opt arg:: Example with optional argument.
6139 * beginning-of-buffer complete::
6142 @node Optional Arguments, beginning-of-buffer opt arg, beginning-of-buffer, beginning-of-buffer
6143 @subsection Optional Arguments
6145 Unless told otherwise, Lisp expects that a function with an argument in
6146 its function definition will be called with a value for that argument.
6147 If that does not happen, you get an error and a message that says
6148 @samp{Wrong number of arguments}.
6150 @cindex Optional arguments
6153 However, optional arguments are a feature of Lisp: a particular
6154 @dfn{keyword} is used to tell the Lisp interpreter that an argument is
6155 optional. The keyword is @code{&optional}. (The @samp{&} in front of
6156 @samp{optional} is part of the keyword.) In a function definition, if
6157 an argument follows the keyword @code{&optional}, no value need be
6158 passed to that argument when the function is called.
6161 The first line of the function definition of @code{beginning-of-buffer}
6162 therefore looks like this:
6165 (defun beginning-of-buffer (&optional arg)
6169 In outline, the whole function looks like this:
6173 (defun beginning-of-buffer (&optional arg)
6174 "@var{documentation}@dots{}"
6176 (or (@var{is-the-argument-a-cons-cell} arg)
6177 (and @var{are-both-transient-mark-mode-and-mark-active-true})
6179 (let (@var{determine-size-and-set-it})
6181 (@var{if-there-is-an-argument}
6182 @var{figure-out-where-to-go}
6189 The function is similar to the @code{simplified-beginning-of-buffer}
6190 function except that the @code{interactive} expression has @code{"P"}
6191 as an argument and the @code{goto-char} function is followed by an
6192 if-then-else expression that figures out where to put the cursor if
6193 there is an argument that is not a cons cell.
6195 (Since I do not explain a cons cell for many more chapters, please
6196 consider ignoring the function @code{consp}. @xref{List
6197 Implementation, , How Lists are Implemented}, and @ref{Cons Cell Type,
6198 , Cons Cell and List Types, elisp, The GNU Emacs Lisp Reference
6201 The @code{"P"} in the @code{interactive} expression tells Emacs to
6202 pass a prefix argument, if there is one, to the function in raw form.
6203 A prefix argument is made by typing the @key{META} key followed by a
6204 number, or by typing @kbd{C-u} and then a number. (If you don't type
6205 a number, @kbd{C-u} defaults to a cons cell with a 4. A lowercase
6206 @code{"p"} in the @code{interactive} expression causes the function to
6207 convert a prefix arg to a number.)
6209 The true-or-false-test of the @code{if} expression looks complex, but
6210 it is not: it checks whether @code{arg} has a value that is not
6211 @code{nil} and whether it is a cons cell. (That is what @code{consp}
6212 does; it checks whether its argument is a cons cell.) If @code{arg}
6213 has a value that is not @code{nil} (and is not a cons cell), which
6214 will be the case if @code{beginning-of-buffer} is called with a
6215 numeric argument, then this true-or-false-test will return true and
6216 the then-part of the @code{if} expression will be evaluated. On the
6217 other hand, if @code{beginning-of-buffer} is not called with an
6218 argument, the value of @code{arg} will be @code{nil} and the else-part
6219 of the @code{if} expression will be evaluated. The else-part is
6220 simply @code{point-min}, and when this is the outcome, the whole
6221 @code{goto-char} expression is @code{(goto-char (point-min))}, which
6222 is how we saw the @code{beginning-of-buffer} function in its
6225 @node beginning-of-buffer opt arg, beginning-of-buffer complete, Optional Arguments, beginning-of-buffer
6226 @subsection @code{beginning-of-buffer} with an Argument
6228 When @code{beginning-of-buffer} is called with an argument, an
6229 expression is evaluated which calculates what value to pass to
6230 @code{goto-char}. This expression is rather complicated at first sight.
6231 It includes an inner @code{if} expression and much arithmetic. It looks
6236 (if (> (buffer-size) 10000)
6237 ;; @r{Avoid overflow for large buffer sizes!}
6238 (* (prefix-numeric-value arg)
6243 size (prefix-numeric-value arg))) 10)))
6248 * Disentangle beginning-of-buffer::
6249 * Large buffer case::
6250 * Small buffer case::
6253 @node Disentangle beginning-of-buffer, Large buffer case, beginning-of-buffer opt arg, beginning-of-buffer opt arg
6255 @unnumberedsubsubsec Disentangle @code{beginning-of-buffer}
6258 Like other complex-looking expressions, the conditional expression
6259 within @code{beginning-of-buffer} can be disentangled by looking at it
6260 as parts of a template, in this case, the template for an if-then-else
6261 expression. In skeletal form, the expression looks like this:
6265 (if (@var{buffer-is-large}
6266 @var{divide-buffer-size-by-10-and-multiply-by-arg}
6267 @var{else-use-alternate-calculation}
6271 The true-or-false-test of this inner @code{if} expression checks the
6272 size of the buffer. The reason for this is that the old version 18
6273 Emacs used numbers that are no bigger than eight million or so and in
6274 the computation that followed, the programmer feared that Emacs might
6275 try to use over-large numbers if the buffer were large. The term
6276 `overflow', mentioned in the comment, means numbers that are over
6277 large. More recent versions of Emacs use larger numbers, but this
6278 code has not been touched, if only because people now look at buffers
6279 that are far, far larger than ever before.
6281 There are two cases: if the buffer is large and if it is not.
6283 @node Large buffer case, Small buffer case, Disentangle beginning-of-buffer, beginning-of-buffer opt arg
6284 @comment node-name, next, previous, up
6285 @unnumberedsubsubsec What happens in a large buffer
6287 In @code{beginning-of-buffer}, the inner @code{if} expression tests
6288 whether the size of the buffer is greater than 10,000 characters. To do
6289 this, it uses the @code{>} function and the computation of @code{size}
6290 that comes from the let expression.
6292 In the old days, the function @code{buffer-size} was used. Not only
6293 was that function called several times, it gave the size of the whole
6294 buffer, not the accessible part. The computation makes much more
6295 sense when it handles just the accessible part. (@xref{Narrowing &
6296 Widening, , Narrowing and Widening}, for more information on focusing
6297 attention to an `accessible' part.)
6300 The line looks like this:
6308 When the buffer is large, the then-part of the @code{if} expression is
6309 evaluated. It reads like this (after formatting for easy reading):
6314 (prefix-numeric-value arg)
6320 This expression is a multiplication, with two arguments to the function
6323 The first argument is @code{(prefix-numeric-value arg)}. When
6324 @code{"P"} is used as the argument for @code{interactive}, the value
6325 passed to the function as its argument is passed a ``raw prefix
6326 argument'', and not a number. (It is a number in a list.) To perform
6327 the arithmetic, a conversion is necessary, and
6328 @code{prefix-numeric-value} does the job.
6330 @findex / @r{(division)}
6332 The second argument is @code{(/ size 10)}. This expression divides
6333 the numeric value by ten --- the numeric value of the size of the
6334 accessible portion of the buffer. This produces a number that tells
6335 how many characters make up one tenth of the buffer size. (In Lisp,
6336 @code{/} is used for division, just as @code{*} is used for
6340 In the multiplication expression as a whole, this amount is multiplied
6341 by the value of the prefix argument---the multiplication looks like this:
6345 (* @var{numeric-value-of-prefix-arg}
6346 @var{number-of-characters-in-one-tenth-of-the-accessible-buffer})
6351 If, for example, the prefix argument is @samp{7}, the one-tenth value
6352 will be multiplied by 7 to give a position 70% of the way through.
6355 The result of all this is that if the accessible portion of the buffer
6356 is large, the @code{goto-char} expression reads like this:
6360 (goto-char (* (prefix-numeric-value arg)
6365 This puts the cursor where we want it.
6367 @node Small buffer case, , Large buffer case, beginning-of-buffer opt arg
6368 @comment node-name, next, previous, up
6369 @unnumberedsubsubsec What happens in a small buffer
6371 If the buffer contains fewer than 10,000 characters, a slightly
6372 different computation is performed. You might think this is not
6373 necessary, since the first computation could do the job. However, in
6374 a small buffer, the first method may not put the cursor on exactly the
6375 desired line; the second method does a better job.
6378 The code looks like this:
6380 @c Keep this on one line.
6382 (/ (+ 10 (* size (prefix-numeric-value arg))) 10))
6387 This is code in which you figure out what happens by discovering how the
6388 functions are embedded in parentheses. It is easier to read if you
6389 reformat it with each expression indented more deeply than its
6390 enclosing expression:
6398 (prefix-numeric-value arg)))
6405 Looking at parentheses, we see that the innermost operation is
6406 @code{(prefix-numeric-value arg)}, which converts the raw argument to
6407 a number. In the following expression, this number is multiplied by
6408 the size of the accessible portion of the buffer:
6411 (* size (prefix-numeric-value arg))
6415 This multiplication creates a number that may be larger than the size of
6416 the buffer---seven times larger if the argument is 7, for example. Ten
6417 is then added to this number and finally the large number is divided by
6418 ten to provide a value that is one character larger than the percentage
6419 position in the buffer.
6421 The number that results from all this is passed to @code{goto-char} and
6422 the cursor is moved to that point.
6425 @node beginning-of-buffer complete, , beginning-of-buffer opt arg, beginning-of-buffer
6426 @comment node-name, next, previous, up
6427 @subsection The Complete @code{beginning-of-buffer}
6430 Here is the complete text of the @code{beginning-of-buffer} function:
6436 (defun beginning-of-buffer (&optional arg)
6437 "Move point to the beginning of the buffer;
6438 leave mark at previous position.
6439 With \\[universal-argument] prefix,
6440 do not set mark at previous position.
6442 put point N/10 of the way from the beginning.
6444 If the buffer is narrowed,
6445 this command uses the beginning and size
6446 of the accessible part of the buffer.
6450 Don't use this command in Lisp programs!
6451 \(goto-char (point-min)) is faster
6452 and avoids clobbering the mark."
6455 (and transient-mark-mode mark-active)
6459 (let ((size (- (point-max) (point-min))))
6460 (goto-char (if (and arg (not (consp arg)))
6463 ;; Avoid overflow for large buffer sizes!
6464 (* (prefix-numeric-value arg)
6466 (/ (+ 10 (* size (prefix-numeric-value arg))) 10)))
6468 (if arg (forward-line 1)))
6473 From before GNU Emacs 22
6476 (defun beginning-of-buffer (&optional arg)
6477 "Move point to the beginning of the buffer;
6478 leave mark at previous position.
6479 With arg N, put point N/10 of the way
6480 from the true beginning.
6483 Don't use this in Lisp programs!
6484 \(goto-char (point-min)) is faster
6485 and does not set the mark."
6492 (if (> (buffer-size) 10000)
6493 ;; @r{Avoid overflow for large buffer sizes!}
6494 (* (prefix-numeric-value arg)
6495 (/ (buffer-size) 10))
6498 (/ (+ 10 (* (buffer-size)
6499 (prefix-numeric-value arg)))
6502 (if arg (forward-line 1)))
6508 Except for two small points, the previous discussion shows how this
6509 function works. The first point deals with a detail in the
6510 documentation string, and the second point concerns the last line of
6514 In the documentation string, there is reference to an expression:
6517 \\[universal-argument]
6521 A @samp{\\} is used before the first square bracket of this
6522 expression. This @samp{\\} tells the Lisp interpreter to substitute
6523 whatever key is currently bound to the @samp{[@dots{}]}. In the case
6524 of @code{universal-argument}, that is usually @kbd{C-u}, but it might
6525 be different. (@xref{Documentation Tips, , Tips for Documentation
6526 Strings, elisp, The GNU Emacs Lisp Reference Manual}, for more
6530 Finally, the last line of the @code{beginning-of-buffer} command says
6531 to move point to the beginning of the next line if the command is
6532 invoked with an argument:
6535 (if arg (forward-line 1)))
6539 This puts the cursor at the beginning of the first line after the
6540 appropriate tenths position in the buffer. This is a flourish that
6541 means that the cursor is always located @emph{at least} the requested
6542 tenths of the way through the buffer, which is a nicety that is,
6543 perhaps, not necessary, but which, if it did not occur, would be sure
6546 On the other hand, it also means that if you specify the command with
6547 a @kbd{C-u}, but without a number, that is to say, if the `raw prefix
6548 argument' is simply a cons cell, then the command puts you at the
6549 beginning of the second line @dots{} I don't know whether this is
6550 intended or whether no one has dealt with the code to avoid this
6553 @node Second Buffer Related Review, optional Exercise, beginning-of-buffer, More Complex
6554 @comment node-name, next, previous, up
6557 Here is a brief summary of some of the topics covered in this chapter.
6561 Evaluate each argument in sequence, and return the value of the first
6562 argument that is not @code{nil}; if none return a value that is not
6563 @code{nil}, return @code{nil}. In brief, return the first true value
6564 of the arguments; return a true value if one @emph{or} any of the
6568 Evaluate each argument in sequence, and if any are @code{nil}, return
6569 @code{nil}; if none are @code{nil}, return the value of the last
6570 argument. In brief, return a true value only if all the arguments are
6571 true; return a true value if one @emph{and} each of the others is
6575 A keyword used to indicate that an argument to a function definition
6576 is optional; this means that the function can be evaluated without the
6577 argument, if desired.
6579 @item prefix-numeric-value
6580 Convert the `raw prefix argument' produced by @code{(interactive
6581 "P")} to a numeric value.
6584 Move point forward to the beginning of the next line, or if the argument
6585 is greater than one, forward that many lines. If it can't move as far
6586 forward as it is supposed to, @code{forward-line} goes forward as far as
6587 it can and then returns a count of the number of additional lines it was
6588 supposed to move but couldn't.
6591 Delete the entire contents of the current buffer.
6594 Return @code{t} if its argument is a buffer; otherwise return @code{nil}.
6597 @node optional Exercise, , Second Buffer Related Review, More Complex
6598 @section @code{optional} Argument Exercise
6600 Write an interactive function with an optional argument that tests
6601 whether its argument, a number, is greater than or equal to, or else,
6602 less than the value of @code{fill-column}, and tells you which, in a
6603 message. However, if you do not pass an argument to the function, use
6604 56 as a default value.
6606 @node Narrowing & Widening, car cdr & cons, More Complex, Top
6607 @comment node-name, next, previous, up
6608 @chapter Narrowing and Widening
6609 @cindex Focusing attention (narrowing)
6613 Narrowing is a feature of Emacs that makes it possible for you to focus
6614 on a specific part of a buffer, and work without accidentally changing
6615 other parts. Narrowing is normally disabled since it can confuse
6619 * Narrowing advantages:: The advantages of narrowing
6620 * save-restriction:: The @code{save-restriction} special form.
6621 * what-line:: The number of the line that point is on.
6625 @node Narrowing advantages, save-restriction, Narrowing & Widening, Narrowing & Widening
6627 @unnumberedsec The Advantages of Narrowing
6630 With narrowing, the rest of a buffer is made invisible, as if it weren't
6631 there. This is an advantage if, for example, you want to replace a word
6632 in one part of a buffer but not in another: you narrow to the part you want
6633 and the replacement is carried out only in that section, not in the rest
6634 of the buffer. Searches will only work within a narrowed region, not
6635 outside of one, so if you are fixing a part of a document, you can keep
6636 yourself from accidentally finding parts you do not need to fix by
6637 narrowing just to the region you want.
6638 (The key binding for @code{narrow-to-region} is @kbd{C-x n n}.)
6640 However, narrowing does make the rest of the buffer invisible, which
6641 can scare people who inadvertently invoke narrowing and think they
6642 have deleted a part of their file. Moreover, the @code{undo} command
6643 (which is usually bound to @kbd{C-x u}) does not turn off narrowing
6644 (nor should it), so people can become quite desperate if they do not
6645 know that they can return the rest of a buffer to visibility with the
6646 @code{widen} command.
6647 (The key binding for @code{widen} is @kbd{C-x n w}.)
6649 Narrowing is just as useful to the Lisp interpreter as to a human.
6650 Often, an Emacs Lisp function is designed to work on just part of a
6651 buffer; or conversely, an Emacs Lisp function needs to work on all of a
6652 buffer that has been narrowed. The @code{what-line} function, for
6653 example, removes the narrowing from a buffer, if it has any narrowing
6654 and when it has finished its job, restores the narrowing to what it was.
6655 On the other hand, the @code{count-lines} function, which is called by
6656 @code{what-line}, uses narrowing to restrict itself to just that portion
6657 of the buffer in which it is interested and then restores the previous
6660 @node save-restriction, what-line, Narrowing advantages, Narrowing & Widening
6661 @comment node-name, next, previous, up
6662 @section The @code{save-restriction} Special Form
6663 @findex save-restriction
6665 In Emacs Lisp, you can use the @code{save-restriction} special form to
6666 keep track of whatever narrowing is in effect, if any. When the Lisp
6667 interpreter meets with @code{save-restriction}, it executes the code
6668 in the body of the @code{save-restriction} expression, and then undoes
6669 any changes to narrowing that the code caused. If, for example, the
6670 buffer is narrowed and the code that follows @code{save-restriction}
6671 gets rid of the narrowing, @code{save-restriction} returns the buffer
6672 to its narrowed region afterwards. In the @code{what-line} command,
6673 any narrowing the buffer may have is undone by the @code{widen}
6674 command that immediately follows the @code{save-restriction} command.
6675 Any original narrowing is restored just before the completion of the
6679 The template for a @code{save-restriction} expression is simple:
6689 The body of the @code{save-restriction} is one or more expressions that
6690 will be evaluated in sequence by the Lisp interpreter.
6692 Finally, a point to note: when you use both @code{save-excursion} and
6693 @code{save-restriction}, one right after the other, you should use
6694 @code{save-excursion} outermost. If you write them in reverse order,
6695 you may fail to record narrowing in the buffer to which Emacs switches
6696 after calling @code{save-excursion}. Thus, when written together,
6697 @code{save-excursion} and @code{save-restriction} should be written
6708 In other circumstances, when not written together, the
6709 @code{save-excursion} and @code{save-restriction} special forms must
6710 be written in the order appropriate to the function.
6726 /usr/local/src/emacs/lisp/simple.el
6729 "Print the current buffer line number and narrowed line number of point."
6731 (let ((start (point-min))
6732 (n (line-number-at-pos)))
6734 (message "Line %d" n)
6738 (message "line %d (narrowed line %d)"
6739 (+ n (line-number-at-pos start) -1) n))))))
6741 (defun line-number-at-pos (&optional pos)
6742 "Return (narrowed) buffer line number at position POS.
6743 If POS is nil, use current buffer location.
6744 Counting starts at (point-min), so the value refers
6745 to the contents of the accessible portion of the buffer."
6746 (let ((opoint (or pos (point))) start)
6748 (goto-char (point-min))
6749 (setq start (point))
6752 (1+ (count-lines start (point))))))
6754 (defun count-lines (start end)
6755 "Return number of lines between START and END.
6756 This is usually the number of newlines between them,
6757 but can be one more if START is not equal to END
6758 and the greater of them is not at the start of a line."
6761 (narrow-to-region start end)
6762 (goto-char (point-min))
6763 (if (eq selective-display t)
6766 (while (re-search-forward "[\n\C-m]" nil t 40)
6767 (setq done (+ 40 done)))
6768 (while (re-search-forward "[\n\C-m]" nil t 1)
6769 (setq done (+ 1 done)))
6770 (goto-char (point-max))
6771 (if (and (/= start end)
6775 (- (buffer-size) (forward-line (buffer-size)))))))
6778 @node what-line, narrow Exercise, save-restriction, Narrowing & Widening
6779 @comment node-name, next, previous, up
6780 @section @code{what-line}
6782 @cindex Widening, example of
6784 The @code{what-line} command tells you the number of the line in which
6785 the cursor is located. The function illustrates the use of the
6786 @code{save-restriction} and @code{save-excursion} commands. Here is the
6787 original text of the function:
6792 "Print the current line number (in the buffer) of point."
6799 (1+ (count-lines 1 (point)))))))
6803 (In recent versions of GNU Emacs, the @code{what-line} function has
6804 been expanded to tell you your line number in a narrowed buffer as
6805 well as your line number in a widened buffer. The recent version is
6806 more complex than the version shown here. If you feel adventurous,
6807 you might want to look at it after figuring out how this version
6808 works. You will probably need to use @kbd{C-h f}
6809 (@code{describe-function}). The newer version uses a conditional to
6810 determine whether the buffer has been narrowed.
6812 (Also, it uses @code{line-number-at-pos}, which among other simple
6813 expressions, such as @code{(goto-char (point-min))}, moves point to
6814 the beginning of the current line with @code{(forward-line 0)} rather
6815 than @code{beginning-of-line}.)
6817 The @code{what-line} function as shown here has a documentation line
6818 and is interactive, as you would expect. The next two lines use the
6819 functions @code{save-restriction} and @code{widen}.
6821 The @code{save-restriction} special form notes whatever narrowing is in
6822 effect, if any, in the current buffer and restores that narrowing after
6823 the code in the body of the @code{save-restriction} has been evaluated.
6825 The @code{save-restriction} special form is followed by @code{widen}.
6826 This function undoes any narrowing the current buffer may have had
6827 when @code{what-line} was called. (The narrowing that was there is
6828 the narrowing that @code{save-restriction} remembers.) This widening
6829 makes it possible for the line counting commands to count from the
6830 beginning of the buffer. Otherwise, they would have been limited to
6831 counting within the accessible region. Any original narrowing is
6832 restored just before the completion of the function by the
6833 @code{save-restriction} special form.
6835 The call to @code{widen} is followed by @code{save-excursion}, which
6836 saves the location of the cursor (i.e., of point) and of the mark, and
6837 restores them after the code in the body of the @code{save-excursion}
6838 uses the @code{beginning-of-line} function to move point.
6840 (Note that the @code{(widen)} expression comes between the
6841 @code{save-restriction} and @code{save-excursion} special forms. When
6842 you write the two @code{save- @dots{}} expressions in sequence, write
6843 @code{save-excursion} outermost.)
6846 The last two lines of the @code{what-line} function are functions to
6847 count the number of lines in the buffer and then print the number in the
6853 (1+ (count-lines 1 (point)))))))
6857 The @code{message} function prints a one-line message at the bottom of
6858 the Emacs screen. The first argument is inside of quotation marks and
6859 is printed as a string of characters. However, it may contain a
6860 @samp{%d} expression to print a following argument. @samp{%d} prints
6861 the argument as a decimal, so the message will say something such as
6865 The number that is printed in place of the @samp{%d} is computed by the
6866 last line of the function:
6869 (1+ (count-lines 1 (point)))
6875 (defun count-lines (start end)
6876 "Return number of lines between START and END.
6877 This is usually the number of newlines between them,
6878 but can be one more if START is not equal to END
6879 and the greater of them is not at the start of a line."
6882 (narrow-to-region start end)
6883 (goto-char (point-min))
6884 (if (eq selective-display t)
6887 (while (re-search-forward "[\n\C-m]" nil t 40)
6888 (setq done (+ 40 done)))
6889 (while (re-search-forward "[\n\C-m]" nil t 1)
6890 (setq done (+ 1 done)))
6891 (goto-char (point-max))
6892 (if (and (/= start end)
6896 (- (buffer-size) (forward-line (buffer-size)))))))
6900 What this does is count the lines from the first position of the
6901 buffer, indicated by the @code{1}, up to @code{(point)}, and then add
6902 one to that number. (The @code{1+} function adds one to its
6903 argument.) We add one to it because line 2 has only one line before
6904 it, and @code{count-lines} counts only the lines @emph{before} the
6907 After @code{count-lines} has done its job, and the message has been
6908 printed in the echo area, the @code{save-excursion} restores point and
6909 mark to their original positions; and @code{save-restriction} restores
6910 the original narrowing, if any.
6912 @node narrow Exercise, , what-line, Narrowing & Widening
6913 @section Exercise with Narrowing
6915 Write a function that will display the first 60 characters of the
6916 current buffer, even if you have narrowed the buffer to its latter
6917 half so that the first line is inaccessible. Restore point, mark, and
6918 narrowing. For this exercise, you need to use a whole potpourri of
6919 functions, including @code{save-restriction}, @code{widen},
6920 @code{goto-char}, @code{point-min}, @code{message}, and
6921 @code{buffer-substring}.
6923 @cindex Properties, mention of @code{buffer-substring-no-properties}
6924 (@code{buffer-substring} is a previously unmentioned function you will
6925 have to investigate yourself; or perhaps you will have to use
6926 @code{buffer-substring-no-properties} or
6927 @code{filter-buffer-substring} @dots{}, yet other functions. Text
6928 properties are a feature otherwise not discussed here. @xref{Text
6929 Properties, , Text Properties, elisp, The GNU Emacs Lisp Reference
6932 Additionally, do you really need @code{goto-char} or @code{point-min}?
6933 Or can you write the function without them?
6935 @node car cdr & cons, Cutting & Storing Text, Narrowing & Widening, Top
6936 @comment node-name, next, previous, up
6937 @chapter @code{car}, @code{cdr}, @code{cons}: Fundamental Functions
6938 @findex car, @r{introduced}
6939 @findex cdr, @r{introduced}
6941 In Lisp, @code{car}, @code{cdr}, and @code{cons} are fundamental
6942 functions. The @code{cons} function is used to construct lists, and
6943 the @code{car} and @code{cdr} functions are used to take them apart.
6945 In the walk through of the @code{copy-region-as-kill} function, we
6946 will see @code{cons} as well as two variants on @code{cdr},
6947 namely, @code{setcdr} and @code{nthcdr}. (@xref{copy-region-as-kill}.)
6950 * Strange Names:: An historical aside: why the strange names?
6951 * car & cdr:: Functions for extracting part of a list.
6952 * cons:: Constructing a list.
6953 * nthcdr:: Calling @code{cdr} repeatedly.
6955 * setcar:: Changing the first element of a list.
6956 * setcdr:: Changing the rest of a list.
6960 @node Strange Names, car & cdr, car cdr & cons, car cdr & cons
6962 @unnumberedsec Strange Names
6965 The name of the @code{cons} function is not unreasonable: it is an
6966 abbreviation of the word `construct'. The origins of the names for
6967 @code{car} and @code{cdr}, on the other hand, are esoteric: @code{car}
6968 is an acronym from the phrase `Contents of the Address part of the
6969 Register'; and @code{cdr} (pronounced `could-er') is an acronym from
6970 the phrase `Contents of the Decrement part of the Register'. These
6971 phrases refer to specific pieces of hardware on the very early
6972 computer on which the original Lisp was developed. Besides being
6973 obsolete, the phrases have been completely irrelevant for more than 25
6974 years to anyone thinking about Lisp. Nonetheless, although a few
6975 brave scholars have begun to use more reasonable names for these
6976 functions, the old terms are still in use. In particular, since the
6977 terms are used in the Emacs Lisp source code, we will use them in this
6980 @node car & cdr, cons, Strange Names, car cdr & cons
6981 @comment node-name, next, previous, up
6982 @section @code{car} and @code{cdr}
6984 The @sc{car} of a list is, quite simply, the first item in the list.
6985 Thus the @sc{car} of the list @code{(rose violet daisy buttercup)} is
6989 If you are reading this in Info in GNU Emacs, you can see this by
6990 evaluating the following:
6993 (car '(rose violet daisy buttercup))
6997 After evaluating the expression, @code{rose} will appear in the echo
7000 Clearly, a more reasonable name for the @code{car} function would be
7001 @code{first} and this is often suggested.
7003 @code{car} does not remove the first item from the list; it only reports
7004 what it is. After @code{car} has been applied to a list, the list is
7005 still the same as it was. In the jargon, @code{car} is
7006 `non-destructive'. This feature turns out to be important.
7008 The @sc{cdr} of a list is the rest of the list, that is, the
7009 @code{cdr} function returns the part of the list that follows the
7010 first item. Thus, while the @sc{car} of the list @code{'(rose violet
7011 daisy buttercup)} is @code{rose}, the rest of the list, the value
7012 returned by the @code{cdr} function, is @code{(violet daisy
7016 You can see this by evaluating the following in the usual way:
7019 (cdr '(rose violet daisy buttercup))
7023 When you evaluate this, @code{(violet daisy buttercup)} will appear in
7026 Like @code{car}, @code{cdr} does not remove any elements from the
7027 list---it just returns a report of what the second and subsequent
7030 Incidentally, in the example, the list of flowers is quoted. If it were
7031 not, the Lisp interpreter would try to evaluate the list by calling
7032 @code{rose} as a function. In this example, we do not want to do that.
7034 Clearly, a more reasonable name for @code{cdr} would be @code{rest}.
7036 (There is a lesson here: when you name new functions, consider very
7037 carefully what you are doing, since you may be stuck with the names
7038 for far longer than you expect. The reason this document perpetuates
7039 these names is that the Emacs Lisp source code uses them, and if I did
7040 not use them, you would have a hard time reading the code; but do,
7041 please, try to avoid using these terms yourself. The people who come
7042 after you will be grateful to you.)
7044 When @code{car} and @code{cdr} are applied to a list made up of symbols,
7045 such as the list @code{(pine fir oak maple)}, the element of the list
7046 returned by the function @code{car} is the symbol @code{pine} without
7047 any parentheses around it. @code{pine} is the first element in the
7048 list. However, the @sc{cdr} of the list is a list itself, @code{(fir
7049 oak maple)}, as you can see by evaluating the following expressions in
7054 (car '(pine fir oak maple))
7056 (cdr '(pine fir oak maple))
7060 On the other hand, in a list of lists, the first element is itself a
7061 list. @code{car} returns this first element as a list. For example,
7062 the following list contains three sub-lists, a list of carnivores, a
7063 list of herbivores and a list of sea mammals:
7067 (car '((lion tiger cheetah)
7068 (gazelle antelope zebra)
7069 (whale dolphin seal)))
7074 In this example, the first element or @sc{car} of the list is the list of
7075 carnivores, @code{(lion tiger cheetah)}, and the rest of the list is
7076 @code{((gazelle antelope zebra) (whale dolphin seal))}.
7080 (cdr '((lion tiger cheetah)
7081 (gazelle antelope zebra)
7082 (whale dolphin seal)))
7086 It is worth saying again that @code{car} and @code{cdr} are
7087 non-destructive---that is, they do not modify or change lists to which
7088 they are applied. This is very important for how they are used.
7090 Also, in the first chapter, in the discussion about atoms, I said that
7091 in Lisp, ``certain kinds of atom, such as an array, can be separated
7092 into parts; but the mechanism for doing this is different from the
7093 mechanism for splitting a list. As far as Lisp is concerned, the
7094 atoms of a list are unsplittable.'' (@xref{Lisp Atoms}.) The
7095 @code{car} and @code{cdr} functions are used for splitting lists and
7096 are considered fundamental to Lisp. Since they cannot split or gain
7097 access to the parts of an array, an array is considered an atom.
7098 Conversely, the other fundamental function, @code{cons}, can put
7099 together or construct a list, but not an array. (Arrays are handled
7100 by array-specific functions. @xref{Arrays, , Arrays, elisp, The GNU
7101 Emacs Lisp Reference Manual}.)
7103 @node cons, nthcdr, car & cdr, car cdr & cons
7104 @comment node-name, next, previous, up
7105 @section @code{cons}
7106 @findex cons, @r{introduced}
7108 The @code{cons} function constructs lists; it is the inverse of
7109 @code{car} and @code{cdr}. For example, @code{cons} can be used to make
7110 a four element list from the three element list, @code{(fir oak maple)}:
7113 (cons 'pine '(fir oak maple))
7118 After evaluating this list, you will see
7121 (pine fir oak maple)
7125 appear in the echo area. @code{cons} causes the creation of a new
7126 list in which the element is followed by the elements of the original
7129 We often say that `@code{cons} puts a new element at the beginning of
7130 a list; it attaches or pushes elements onto the list', but this
7131 phrasing can be misleading, since @code{cons} does not change an
7132 existing list, but creates a new one.
7134 Like @code{car} and @code{cdr}, @code{cons} is non-destructive.
7138 * length:: How to find the length of a list.
7141 @node Build a list, length, cons, cons
7143 @unnumberedsubsec Build a list
7146 @code{cons} must have a list to attach to.@footnote{Actually, you can
7147 @code{cons} an element to an atom to produce a dotted pair. Dotted
7148 pairs are not discussed here; see @ref{Dotted Pair Notation, , Dotted
7149 Pair Notation, elisp, The GNU Emacs Lisp Reference Manual}.} You
7150 cannot start from absolutely nothing. If you are building a list, you
7151 need to provide at least an empty list at the beginning. Here is a
7152 series of @code{cons} expressions that build up a list of flowers. If
7153 you are reading this in Info in GNU Emacs, you can evaluate each of
7154 the expressions in the usual way; the value is printed in this text
7155 after @samp{@result{}}, which you may read as `evaluates to'.
7159 (cons 'buttercup ())
7160 @result{} (buttercup)
7164 (cons 'daisy '(buttercup))
7165 @result{} (daisy buttercup)
7169 (cons 'violet '(daisy buttercup))
7170 @result{} (violet daisy buttercup)
7174 (cons 'rose '(violet daisy buttercup))
7175 @result{} (rose violet daisy buttercup)
7180 In the first example, the empty list is shown as @code{()} and a list
7181 made up of @code{buttercup} followed by the empty list is constructed.
7182 As you can see, the empty list is not shown in the list that was
7183 constructed. All that you see is @code{(buttercup)}. The empty list is
7184 not counted as an element of a list because there is nothing in an empty
7185 list. Generally speaking, an empty list is invisible.
7187 The second example, @code{(cons 'daisy '(buttercup))} constructs a new,
7188 two element list by putting @code{daisy} in front of @code{buttercup};
7189 and the third example constructs a three element list by putting
7190 @code{violet} in front of @code{daisy} and @code{buttercup}.
7192 @node length, , Build a list, cons
7193 @comment node-name, next, previous, up
7194 @subsection Find the Length of a List: @code{length}
7197 You can find out how many elements there are in a list by using the Lisp
7198 function @code{length}, as in the following examples:
7202 (length '(buttercup))
7207 (length '(daisy buttercup))
7212 (length (cons 'violet '(daisy buttercup)))
7218 In the third example, the @code{cons} function is used to construct a
7219 three element list which is then passed to the @code{length} function as
7223 We can also use @code{length} to count the number of elements in an
7234 As you would expect, the number of elements in an empty list is zero.
7236 An interesting experiment is to find out what happens if you try to find
7237 the length of no list at all; that is, if you try to call @code{length}
7238 without giving it an argument, not even an empty list:
7246 What you see, if you evaluate this, is the error message
7249 Lisp error: (wrong-number-of-arguments length 0)
7253 This means that the function receives the wrong number of
7254 arguments, zero, when it expects some other number of arguments. In
7255 this case, one argument is expected, the argument being a list whose
7256 length the function is measuring. (Note that @emph{one} list is
7257 @emph{one} argument, even if the list has many elements inside it.)
7259 The part of the error message that says @samp{length} is the name of
7263 @code{length} is still a subroutine, but you need C-h f to discover that.
7265 In an earlier version:
7266 This is written with a special notation, @samp{#<subr},
7267 that indicates that the function @code{length} is one of the primitive
7268 functions written in C rather than in Emacs Lisp. (@samp{subr} is an
7269 abbreviation for `subroutine'.) @xref{What Is a Function, , What Is a
7270 Function?, elisp , The GNU Emacs Lisp Reference Manual}, for more
7274 @node nthcdr, nth, cons, car cdr & cons
7275 @comment node-name, next, previous, up
7276 @section @code{nthcdr}
7279 The @code{nthcdr} function is associated with the @code{cdr} function.
7280 What it does is take the @sc{cdr} of a list repeatedly.
7282 If you take the @sc{cdr} of the list @code{(pine fir
7283 oak maple)}, you will be returned the list @code{(fir oak maple)}. If you
7284 repeat this on what was returned, you will be returned the list
7285 @code{(oak maple)}. (Of course, repeated @sc{cdr}ing on the original
7286 list will just give you the original @sc{cdr} since the function does
7287 not change the list. You need to evaluate the @sc{cdr} of the
7288 @sc{cdr} and so on.) If you continue this, eventually you will be
7289 returned an empty list, which in this case, instead of being shown as
7290 @code{()} is shown as @code{nil}.
7293 For review, here is a series of repeated @sc{cdr}s, the text following
7294 the @samp{@result{}} shows what is returned.
7298 (cdr '(pine fir oak maple))
7299 @result{}(fir oak maple)
7303 (cdr '(fir oak maple))
7304 @result{} (oak maple)
7329 You can also do several @sc{cdr}s without printing the values in
7334 (cdr (cdr '(pine fir oak maple)))
7335 @result{} (oak maple)
7340 In this example, the Lisp interpreter evaluates the innermost list first.
7341 The innermost list is quoted, so it just passes the list as it is to the
7342 innermost @code{cdr}. This @code{cdr} passes a list made up of the
7343 second and subsequent elements of the list to the outermost @code{cdr},
7344 which produces a list composed of the third and subsequent elements of
7345 the original list. In this example, the @code{cdr} function is repeated
7346 and returns a list that consists of the original list without its
7349 The @code{nthcdr} function does the same as repeating the call to
7350 @code{cdr}. In the following example, the argument 2 is passed to the
7351 function @code{nthcdr}, along with the list, and the value returned is
7352 the list without its first two items, which is exactly the same
7353 as repeating @code{cdr} twice on the list:
7357 (nthcdr 2 '(pine fir oak maple))
7358 @result{} (oak maple)
7363 Using the original four element list, we can see what happens when
7364 various numeric arguments are passed to @code{nthcdr}, including 0, 1,
7369 ;; @r{Leave the list as it was.}
7370 (nthcdr 0 '(pine fir oak maple))
7371 @result{} (pine fir oak maple)
7375 ;; @r{Return a copy without the first element.}
7376 (nthcdr 1 '(pine fir oak maple))
7377 @result{} (fir oak maple)
7381 ;; @r{Return a copy of the list without three elements.}
7382 (nthcdr 3 '(pine fir oak maple))
7387 ;; @r{Return a copy lacking all four elements.}
7388 (nthcdr 4 '(pine fir oak maple))
7393 ;; @r{Return a copy lacking all elements.}
7394 (nthcdr 5 '(pine fir oak maple))
7399 @node nth, setcar, nthcdr, car cdr & cons
7400 @comment node-name, next, previous, up
7404 The @code{nthcdr} function takes the @sc{cdr} of a list repeatedly.
7405 The @code{nth} function takes the @sc{car} of the result returned by
7406 @code{nthcdr}. It returns the Nth element of the list.
7409 Thus, if it were not defined in C for speed, the definition of
7410 @code{nth} would be:
7415 "Returns the Nth element of LIST.
7416 N counts from zero. If LIST is not that long, nil is returned."
7417 (car (nthcdr n list)))
7422 (Originally, @code{nth} was defined in Emacs Lisp in @file{subr.el},
7423 but its definition was redone in C in the 1980s.)
7425 The @code{nth} function returns a single element of a list.
7426 This can be very convenient.
7428 Note that the elements are numbered from zero, not one. That is to
7429 say, the first element of a list, its @sc{car} is the zeroth element.
7430 This is called `zero-based' counting and often bothers people who
7431 are accustomed to the first element in a list being number one, which
7439 (nth 0 '("one" "two" "three"))
7442 (nth 1 '("one" "two" "three"))
7447 It is worth mentioning that @code{nth}, like @code{nthcdr} and
7448 @code{cdr}, does not change the original list---the function is
7449 non-destructive. This is in sharp contrast to the @code{setcar} and
7450 @code{setcdr} functions.
7452 @node setcar, setcdr, nth, car cdr & cons
7453 @comment node-name, next, previous, up
7454 @section @code{setcar}
7457 As you might guess from their names, the @code{setcar} and @code{setcdr}
7458 functions set the @sc{car} or the @sc{cdr} of a list to a new value.
7459 They actually change the original list, unlike @code{car} and @code{cdr}
7460 which leave the original list as it was. One way to find out how this
7461 works is to experiment. We will start with the @code{setcar} function.
7464 First, we can make a list and then set the value of a variable to the
7465 list, using the @code{setq} function. Here is a list of animals:
7468 (setq animals '(antelope giraffe lion tiger))
7472 If you are reading this in Info inside of GNU Emacs, you can evaluate
7473 this expression in the usual fashion, by positioning the cursor after
7474 the expression and typing @kbd{C-x C-e}. (I'm doing this right here
7475 as I write this. This is one of the advantages of having the
7476 interpreter built into the computing environment. Incidentally, when
7477 there is nothing on the line after the final parentheses, such as a
7478 comment, point can be on the next line. Thus, if your cursor is in
7479 the first column of the next line, you do not need to move it.
7480 Indeed, Emacs permits any amount of white space after the final
7484 When we evaluate the variable @code{animals}, we see that it is bound to
7485 the list @code{(antelope giraffe lion tiger)}:
7490 @result{} (antelope giraffe lion tiger)
7495 Put another way, the variable @code{animals} points to the list
7496 @code{(antelope giraffe lion tiger)}.
7498 Next, evaluate the function @code{setcar} while passing it two
7499 arguments, the variable @code{animals} and the quoted symbol
7500 @code{hippopotamus}; this is done by writing the three element list
7501 @code{(setcar animals 'hippopotamus)} and then evaluating it in the
7505 (setcar animals 'hippopotamus)
7510 After evaluating this expression, evaluate the variable @code{animals}
7511 again. You will see that the list of animals has changed:
7516 @result{} (hippopotamus giraffe lion tiger)
7521 The first element on the list, @code{antelope} is replaced by
7522 @code{hippopotamus}.
7524 So we can see that @code{setcar} did not add a new element to the list
7525 as @code{cons} would have; it replaced @code{antelope} with
7526 @code{hippopotamus}; it @emph{changed} the list.
7528 @node setcdr, cons Exercise, setcar, car cdr & cons
7529 @comment node-name, next, previous, up
7530 @section @code{setcdr}
7533 The @code{setcdr} function is similar to the @code{setcar} function,
7534 except that the function replaces the second and subsequent elements of
7535 a list rather than the first element.
7537 (To see how to change the last element of a list, look ahead to
7538 @ref{kill-new function, , The @code{kill-new} function}, which uses
7539 the @code{nthcdr} and @code{setcdr} functions.)
7542 To see how this works, set the value of the variable to a list of
7543 domesticated animals by evaluating the following expression:
7546 (setq domesticated-animals '(horse cow sheep goat))
7551 If you now evaluate the list, you will be returned the list
7552 @code{(horse cow sheep goat)}:
7556 domesticated-animals
7557 @result{} (horse cow sheep goat)
7562 Next, evaluate @code{setcdr} with two arguments, the name of the
7563 variable which has a list as its value, and the list to which the
7564 @sc{cdr} of the first list will be set;
7567 (setcdr domesticated-animals '(cat dog))
7571 If you evaluate this expression, the list @code{(cat dog)} will appear
7572 in the echo area. This is the value returned by the function. The
7573 result we are interested in is the ``side effect'', which we can see by
7574 evaluating the variable @code{domesticated-animals}:
7578 domesticated-animals
7579 @result{} (horse cat dog)
7584 Indeed, the list is changed from @code{(horse cow sheep goat)} to
7585 @code{(horse cat dog)}. The @sc{cdr} of the list is changed from
7586 @code{(cow sheep goat)} to @code{(cat dog)}.
7588 @node cons Exercise, , setcdr, car cdr & cons
7591 Construct a list of four birds by evaluating several expressions with
7592 @code{cons}. Find out what happens when you @code{cons} a list onto
7593 itself. Replace the first element of the list of four birds with a
7594 fish. Replace the rest of that list with a list of other fish.
7596 @node Cutting & Storing Text, List Implementation, car cdr & cons, Top
7597 @comment node-name, next, previous, up
7598 @chapter Cutting and Storing Text
7599 @cindex Cutting and storing text
7600 @cindex Storing and cutting text
7601 @cindex Killing text
7602 @cindex Clipping text
7603 @cindex Erasing text
7604 @cindex Deleting text
7606 Whenever you cut or clip text out of a buffer with a `kill' command in
7607 GNU Emacs, it is stored in a list and you can bring it back with a
7610 (The use of the word `kill' in Emacs for processes which specifically
7611 @emph{do not} destroy the values of the entities is an unfortunate
7612 historical accident. A much more appropriate word would be `clip' since
7613 that is what the kill commands do; they clip text out of a buffer and
7614 put it into storage from which it can be brought back. I have often
7615 been tempted to replace globally all occurrences of `kill' in the Emacs
7616 sources with `clip' and all occurrences of `killed' with `clipped'.)
7619 * Storing Text:: Text is stored in a list.
7620 * zap-to-char:: Cutting out text up to a character.
7621 * kill-region:: Cutting text out of a region.
7622 * copy-region-as-kill:: A definition for copying text.
7623 * Digression into C:: Minor note on C programming language macros.
7624 * defvar:: How to give a variable an initial value.
7625 * cons & search-fwd Review::
7626 * search Exercises::
7629 @node Storing Text, zap-to-char, Cutting & Storing Text, Cutting & Storing Text
7631 @unnumberedsec Storing Text in a List
7634 When text is cut out of a buffer, it is stored on a list. Successive
7635 pieces of text are stored on the list successively, so the list might
7639 ("a piece of text" "previous piece")
7644 The function @code{cons} can be used to create a new list from a piece
7645 of text (an `atom', to use the jargon) and an existing list, like
7650 (cons "another piece"
7651 '("a piece of text" "previous piece"))
7657 If you evaluate this expression, a list of three elements will appear in
7661 ("another piece" "a piece of text" "previous piece")
7664 With the @code{car} and @code{nthcdr} functions, you can retrieve
7665 whichever piece of text you want. For example, in the following code,
7666 @code{nthcdr 1 @dots{}} returns the list with the first item removed;
7667 and the @code{car} returns the first element of that remainder---the
7668 second element of the original list:
7672 (car (nthcdr 1 '("another piece"
7675 @result{} "a piece of text"
7679 The actual functions in Emacs are more complex than this, of course.
7680 The code for cutting and retrieving text has to be written so that
7681 Emacs can figure out which element in the list you want---the first,
7682 second, third, or whatever. In addition, when you get to the end of
7683 the list, Emacs should give you the first element of the list, rather
7684 than nothing at all.
7686 The list that holds the pieces of text is called the @dfn{kill ring}.
7687 This chapter leads up to a description of the kill ring and how it is
7688 used by first tracing how the @code{zap-to-char} function works. This
7689 function uses (or `calls') a function that invokes a function that
7690 manipulates the kill ring. Thus, before reaching the mountains, we
7691 climb the foothills.
7693 A subsequent chapter describes how text that is cut from the buffer is
7694 retrieved. @xref{Yanking, , Yanking Text Back}.
7696 @node zap-to-char, kill-region, Storing Text, Cutting & Storing Text
7697 @comment node-name, next, previous, up
7698 @section @code{zap-to-char}
7701 The @code{zap-to-char} function changed little between GNU Emacs
7702 version 19 and GNU Emacs version 22. However, @code{zap-to-char}
7703 calls another function, @code{kill-region}, which enjoyed a major
7706 The @code{kill-region} function in Emacs 19 is complex, but does not
7707 use code that is important at this time. We will skip it.
7709 The @code{kill-region} function in Emacs 22 is easier to read than the
7710 same function in Emacs 19 and introduces a very important concept,
7711 that of error handling. We will walk through the function.
7713 But first, let us look at the interactive @code{zap-to-char} function.
7716 * Complete zap-to-char:: The complete implementation.
7717 * zap-to-char interactive:: A three part interactive expression.
7718 * zap-to-char body:: A short overview.
7719 * search-forward:: How to search for a string.
7720 * progn:: The @code{progn} special form.
7721 * Summing up zap-to-char:: Using @code{point} and @code{search-forward}.
7724 @node Complete zap-to-char, zap-to-char interactive, zap-to-char, zap-to-char
7726 @unnumberedsubsec The Complete @code{zap-to-char} Implementation
7729 The @code{zap-to-char} function removes the text in the region between
7730 the location of the cursor (i.e., of point) up to and including the
7731 next occurrence of a specified character. The text that
7732 @code{zap-to-char} removes is put in the kill ring; and it can be
7733 retrieved from the kill ring by typing @kbd{C-y} (@code{yank}). If
7734 the command is given an argument, it removes text through that number
7735 of occurrences. Thus, if the cursor were at the beginning of this
7736 sentence and the character were @samp{s}, @samp{Thus} would be
7737 removed. If the argument were two, @samp{Thus, if the curs} would be
7738 removed, up to and including the @samp{s} in @samp{cursor}.
7740 If the specified character is not found, @code{zap-to-char} will say
7741 ``Search failed'', tell you the character you typed, and not remove
7744 In order to determine how much text to remove, @code{zap-to-char} uses
7745 a search function. Searches are used extensively in code that
7746 manipulates text, and we will focus attention on them as well as on the
7750 @c GNU Emacs version 19
7751 (defun zap-to-char (arg char) ; version 19 implementation
7752 "Kill up to and including ARG'th occurrence of CHAR.
7753 Goes backward if ARG is negative; error if CHAR not found."
7754 (interactive "*p\ncZap to char: ")
7755 (kill-region (point)
7758 (char-to-string char) nil nil arg)
7763 Here is the complete text of the version 22 implementation of the function:
7768 (defun zap-to-char (arg char)
7769 "Kill up to and including ARG'th occurrence of CHAR.
7770 Case is ignored if `case-fold-search' is non-nil in the current buffer.
7771 Goes backward if ARG is negative; error if CHAR not found."
7772 (interactive "p\ncZap to char: ")
7773 (if (char-table-p translation-table-for-input)
7774 (setq char (or (aref translation-table-for-input char) char)))
7775 (kill-region (point) (progn
7776 (search-forward (char-to-string char) nil nil arg)
7781 The documentation is thorough. You do need to know the jargon meaning
7784 @node zap-to-char interactive, zap-to-char body, Complete zap-to-char, zap-to-char
7785 @comment node-name, next, previous, up
7786 @subsection The @code{interactive} Expression
7789 The interactive expression in the @code{zap-to-char} command looks like
7793 (interactive "p\ncZap to char: ")
7796 The part within quotation marks, @code{"p\ncZap to char:@: "}, specifies
7797 two different things. First, and most simply, is the @samp{p}.
7798 This part is separated from the next part by a newline, @samp{\n}.
7799 The @samp{p} means that the first argument to the function will be
7800 passed the value of a `processed prefix'. The prefix argument is
7801 passed by typing @kbd{C-u} and a number, or @kbd{M-} and a number. If
7802 the function is called interactively without a prefix, 1 is passed to
7805 The second part of @code{"p\ncZap to char:@: "} is
7806 @samp{cZap to char:@: }. In this part, the lower case @samp{c}
7807 indicates that @code{interactive} expects a prompt and that the
7808 argument will be a character. The prompt follows the @samp{c} and is
7809 the string @samp{Zap to char:@: } (with a space after the colon to
7812 What all this does is prepare the arguments to @code{zap-to-char} so they
7813 are of the right type, and give the user a prompt.
7815 In a read-only buffer, the @code{zap-to-char} function copies the text
7816 to the kill ring, but does not remove it. The echo area displays a
7817 message saying that the buffer is read-only. Also, the terminal may
7818 beep or blink at you.
7820 @node zap-to-char body, search-forward, zap-to-char interactive, zap-to-char
7821 @comment node-name, next, previous, up
7822 @subsection The Body of @code{zap-to-char}
7824 The body of the @code{zap-to-char} function contains the code that
7825 kills (that is, removes) the text in the region from the current
7826 position of the cursor up to and including the specified character.
7828 The first part of the code looks like this:
7831 (if (char-table-p translation-table-for-input)
7832 (setq char (or (aref translation-table-for-input char) char)))
7833 (kill-region (point) (progn
7834 (search-forward (char-to-string char) nil nil arg)
7839 @code{char-table-p} is an hitherto unseen function. It determines
7840 whether its argument is a character table. When it is, it sets the
7841 character passed to @code{zap-to-char} to one of them, if that
7842 character exists, or to the character itself. (This becomes important
7843 for certain characters in non-European languages. The @code{aref}
7844 function extracts an element from an array. It is an array-specific
7845 function that is not described in this document. @xref{Arrays, ,
7846 Arrays, elisp, The GNU Emacs Lisp Reference Manual}.)
7849 @code{(point)} is the current position of the cursor.
7851 The next part of the code is an expression using @code{progn}. The body
7852 of the @code{progn} consists of calls to @code{search-forward} and
7855 It is easier to understand how @code{progn} works after learning about
7856 @code{search-forward}, so we will look at @code{search-forward} and
7857 then at @code{progn}.
7859 @node search-forward, progn, zap-to-char body, zap-to-char
7860 @comment node-name, next, previous, up
7861 @subsection The @code{search-forward} Function
7862 @findex search-forward
7864 The @code{search-forward} function is used to locate the
7865 zapped-for-character in @code{zap-to-char}. If the search is
7866 successful, @code{search-forward} leaves point immediately after the
7867 last character in the target string. (In @code{zap-to-char}, the
7868 target string is just one character long. @code{zap-to-char} uses the
7869 function @code{char-to-string} to ensure that the computer treats that
7870 character as a string.) If the search is backwards,
7871 @code{search-forward} leaves point just before the first character in
7872 the target. Also, @code{search-forward} returns @code{t} for true.
7873 (Moving point is therefore a `side effect'.)
7876 In @code{zap-to-char}, the @code{search-forward} function looks like this:
7879 (search-forward (char-to-string char) nil nil arg)
7882 The @code{search-forward} function takes four arguments:
7886 The first argument is the target, what is searched for. This must be a
7887 string, such as @samp{"z"}.
7889 As it happens, the argument passed to @code{zap-to-char} is a single
7890 character. Because of the way computers are built, the Lisp
7891 interpreter may treat a single character as being different from a
7892 string of characters. Inside the computer, a single character has a
7893 different electronic format than a string of one character. (A single
7894 character can often be recorded in the computer using exactly one
7895 byte; but a string may be longer, and the computer needs to be ready
7896 for this.) Since the @code{search-forward} function searches for a
7897 string, the character that the @code{zap-to-char} function receives as
7898 its argument must be converted inside the computer from one format to
7899 the other; otherwise the @code{search-forward} function will fail.
7900 The @code{char-to-string} function is used to make this conversion.
7903 The second argument bounds the search; it is specified as a position in
7904 the buffer. In this case, the search can go to the end of the buffer,
7905 so no bound is set and the second argument is @code{nil}.
7908 The third argument tells the function what it should do if the search
7909 fails---it can signal an error (and print a message) or it can return
7910 @code{nil}. A @code{nil} as the third argument causes the function to
7911 signal an error when the search fails.
7914 The fourth argument to @code{search-forward} is the repeat count---how
7915 many occurrences of the string to look for. This argument is optional
7916 and if the function is called without a repeat count, this argument is
7917 passed the value 1. If this argument is negative, the search goes
7922 In template form, a @code{search-forward} expression looks like this:
7926 (search-forward "@var{target-string}"
7927 @var{limit-of-search}
7928 @var{what-to-do-if-search-fails}
7933 We will look at @code{progn} next.
7935 @node progn, Summing up zap-to-char, search-forward, zap-to-char
7936 @comment node-name, next, previous, up
7937 @subsection The @code{progn} Special Form
7940 @code{progn} is a special form that causes each of its arguments to be
7941 evaluated in sequence and then returns the value of the last one. The
7942 preceding expressions are evaluated only for the side effects they
7943 perform. The values produced by them are discarded.
7946 The template for a @code{progn} expression is very simple:
7955 In @code{zap-to-char}, the @code{progn} expression has to do two things:
7956 put point in exactly the right position; and return the location of
7957 point so that @code{kill-region} will know how far to kill to.
7959 The first argument to the @code{progn} is @code{search-forward}. When
7960 @code{search-forward} finds the string, the function leaves point
7961 immediately after the last character in the target string. (In this
7962 case the target string is just one character long.) If the search is
7963 backwards, @code{search-forward} leaves point just before the first
7964 character in the target. The movement of point is a side effect.
7966 The second and last argument to @code{progn} is the expression
7967 @code{(point)}. This expression returns the value of point, which in
7968 this case will be the location to which it has been moved by
7969 @code{search-forward}. (In the source, a line that tells the function
7970 to go to the previous character, if it is going forward, was commented
7971 out in 1999; I don't remember whether that feature or mis-feature was
7972 ever a part of the distributed source.) The value of @code{point} is
7973 returned by the @code{progn} expression and is passed to
7974 @code{kill-region} as @code{kill-region}'s second argument.
7976 @node Summing up zap-to-char, , progn, zap-to-char
7977 @comment node-name, next, previous, up
7978 @subsection Summing up @code{zap-to-char}
7980 Now that we have seen how @code{search-forward} and @code{progn} work,
7981 we can see how the @code{zap-to-char} function works as a whole.
7983 The first argument to @code{kill-region} is the position of the cursor
7984 when the @code{zap-to-char} command is given---the value of point at
7985 that time. Within the @code{progn}, the search function then moves
7986 point to just after the zapped-to-character and @code{point} returns the
7987 value of this location. The @code{kill-region} function puts together
7988 these two values of point, the first one as the beginning of the region
7989 and the second one as the end of the region, and removes the region.
7991 The @code{progn} special form is necessary because the
7992 @code{kill-region} command takes two arguments; and it would fail if
7993 @code{search-forward} and @code{point} expressions were written in
7994 sequence as two additional arguments. The @code{progn} expression is
7995 a single argument to @code{kill-region} and returns the one value that
7996 @code{kill-region} needs for its second argument.
7998 @node kill-region, copy-region-as-kill, zap-to-char, Cutting & Storing Text
7999 @comment node-name, next, previous, up
8000 @section @code{kill-region}
8003 The @code{zap-to-char} function uses the @code{kill-region} function.
8004 This function clips text from a region and copies that text to
8005 the kill ring, from which it may be retrieved.
8010 (defun kill-region (beg end &optional yank-handler)
8011 "Kill (\"cut\") text between point and mark.
8012 This deletes the text from the buffer and saves it in the kill ring.
8013 The command \\[yank] can retrieve it from there.
8014 \(If you want to kill and then yank immediately, use \\[kill-ring-save].)
8016 If you want to append the killed region to the last killed text,
8017 use \\[append-next-kill] before \\[kill-region].
8019 If the buffer is read-only, Emacs will beep and refrain from deleting
8020 the text, but put the text in the kill ring anyway. This means that
8021 you can use the killing commands to copy text from a read-only buffer.
8023 This is the primitive for programs to kill text (as opposed to deleting it).
8024 Supply two arguments, character positions indicating the stretch of text
8026 Any command that calls this function is a \"kill command\".
8027 If the previous command was also a kill command,
8028 the text killed this time appends to the text killed last time
8029 to make one entry in the kill ring.
8031 In Lisp code, optional third arg YANK-HANDLER, if non-nil,
8032 specifies the yank-handler text property to be set on the killed
8033 text. See `insert-for-yank'."
8034 ;; Pass point first, then mark, because the order matters
8035 ;; when calling kill-append.
8036 (interactive (list (point) (mark)))
8037 (unless (and beg end)
8038 (error "The mark is not set now, so there is no region"))
8040 (let ((string (filter-buffer-substring beg end t)))
8041 (when string ;STRING is nil if BEG = END
8042 ;; Add that string to the kill ring, one way or another.
8043 (if (eq last-command 'kill-region)
8044 (kill-append string (< end beg) yank-handler)
8045 (kill-new string nil yank-handler)))
8046 (when (or string (eq last-command 'kill-region))
8047 (setq this-command 'kill-region))
8049 ((buffer-read-only text-read-only)
8050 ;; The code above failed because the buffer, or some of the characters
8051 ;; in the region, are read-only.
8052 ;; We should beep, in case the user just isn't aware of this.
8053 ;; However, there's no harm in putting
8054 ;; the region's text in the kill ring, anyway.
8055 (copy-region-as-kill beg end)
8056 ;; Set this-command now, so it will be set even if we get an error.
8057 (setq this-command 'kill-region)
8058 ;; This should barf, if appropriate, and give us the correct error.
8059 (if kill-read-only-ok
8060 (progn (message "Read only text copied to kill ring") nil)
8061 ;; Signal an error if the buffer is read-only.
8062 (barf-if-buffer-read-only)
8063 ;; If the buffer isn't read-only, the text is.
8064 (signal 'text-read-only (list (current-buffer)))))))
8067 The Emacs 22 version of that function uses @code{condition-case} and
8068 @code{copy-region-as-kill}, both of which we will explain.
8069 @code{condition-case} is an important special form.
8071 In essence, the @code{kill-region} function calls
8072 @code{condition-case}, which takes three arguments. In this function,
8073 the first argument does nothing. The second argument contains the
8074 code that does the work when all goes well. The third argument
8075 contains the code that is called in the event of an error.
8078 * Complete kill-region:: The function definition.
8079 * condition-case:: Dealing with a problem.
8083 @node Complete kill-region, condition-case, kill-region, kill-region
8085 @unnumberedsubsec The Complete @code{kill-region} Definition
8089 We will go through the @code{condition-case} code in a moment. First,
8090 let us look at the definition of @code{kill-region}, with comments
8096 (defun kill-region (beg end)
8097 "Kill (\"cut\") text between point and mark.
8098 This deletes the text from the buffer and saves it in the kill ring.
8099 The command \\[yank] can retrieve it from there. @dots{} "
8103 ;; @bullet{} Since order matters, pass point first.
8104 (interactive (list (point) (mark)))
8105 ;; @bullet{} And tell us if we cannot cut the text.
8106 ;; `unless' is an `if' without a then-part.
8107 (unless (and beg end)
8108 (error "The mark is not set now, so there is no region"))
8112 ;; @bullet{} `condition-case' takes three arguments.
8113 ;; If the first argument is nil, as it is here,
8114 ;; information about the error signal is not
8115 ;; stored for use by another function.
8120 ;; @bullet{} The second argument to `condition-case' tells the
8121 ;; Lisp interpreter what to do when all goes well.
8125 ;; It starts with a `let' function that extracts the string
8126 ;; and tests whether it exists. If so (that is what the
8127 ;; `when' checks), it calls an `if' function that determines
8128 ;; whether the previous command was another call to
8129 ;; `kill-region'; if it was, then the new text is appended to
8130 ;; the previous text; if not, then a different function,
8131 ;; `kill-new', is called.
8135 ;; The `kill-append' function concatenates the new string and
8136 ;; the old. The `kill-new' function inserts text into a new
8137 ;; item in the kill ring.
8141 ;; `when' is an `if' without an else-part. The second `when'
8142 ;; again checks whether the current string exists; in
8143 ;; addition, it checks whether the previous command was
8144 ;; another call to `kill-region'. If one or the other
8145 ;; condition is true, then it sets the current command to
8146 ;; be `kill-region'.
8149 (let ((string (filter-buffer-substring beg end t)))
8150 (when string ;STRING is nil if BEG = END
8151 ;; Add that string to the kill ring, one way or another.
8152 (if (eq last-command 'kill-region)
8155 ;; @minus{} `yank-handler' is an optional argument to
8156 ;; `kill-region' that tells the `kill-append' and
8157 ;; `kill-new' functions how deal with properties
8158 ;; added to the text, such as `bold' or `italics'.
8159 (kill-append string (< end beg) yank-handler)
8160 (kill-new string nil yank-handler)))
8161 (when (or string (eq last-command 'kill-region))
8162 (setq this-command 'kill-region))
8167 ;; @bullet{} The third argument to `condition-case' tells the interpreter
8168 ;; what to do with an error.
8171 ;; The third argument has a conditions part and a body part.
8172 ;; If the conditions are met (in this case,
8173 ;; if text or buffer are read-only)
8174 ;; then the body is executed.
8177 ;; The first part of the third argument is the following:
8178 ((buffer-read-only text-read-only) ;; the if-part
8179 ;; @dots{} the then-part
8180 (copy-region-as-kill beg end)
8183 ;; Next, also as part of the then-part, set this-command, so
8184 ;; it will be set in an error
8185 (setq this-command 'kill-region)
8186 ;; Finally, in the then-part, send a message if you may copy
8187 ;; the text to the kill ring without signally an error, but
8188 ;; don't if you may not.
8191 (if kill-read-only-ok
8192 (progn (message "Read only text copied to kill ring") nil)
8193 (barf-if-buffer-read-only)
8194 ;; If the buffer isn't read-only, the text is.
8195 (signal 'text-read-only (list (current-buffer)))))
8203 (defun kill-region (beg end)
8204 "Kill between point and mark.
8205 The text is deleted but saved in the kill ring."
8210 ;; 1. `condition-case' takes three arguments.
8211 ;; If the first argument is nil, as it is here,
8212 ;; information about the error signal is not
8213 ;; stored for use by another function.
8218 ;; 2. The second argument to `condition-case'
8219 ;; tells the Lisp interpreter what to do when all goes well.
8223 ;; The `delete-and-extract-region' function usually does the
8224 ;; work. If the beginning and ending of the region are both
8225 ;; the same, then the variable `string' will be empty, or nil
8226 (let ((string (delete-and-extract-region beg end)))
8230 ;; `when' is an `if' clause that cannot take an `else-part'.
8231 ;; Emacs normally sets the value of `last-command' to the
8232 ;; previous command.
8235 ;; `kill-append' concatenates the new string and the old.
8236 ;; `kill-new' inserts text into a new item in the kill ring.
8238 (if (eq last-command 'kill-region)
8239 ;; if true, prepend string
8240 (kill-append string (< end beg))
8242 (setq this-command 'kill-region))
8246 ;; 3. The third argument to `condition-case' tells the interpreter
8247 ;; what to do with an error.
8250 ;; The third argument has a conditions part and a body part.
8251 ;; If the conditions are met (in this case,
8252 ;; if text or buffer are read-only)
8253 ;; then the body is executed.
8256 ((buffer-read-only text-read-only) ;; this is the if-part
8258 (copy-region-as-kill beg end)
8261 (if kill-read-only-ok ;; usually this variable is nil
8262 (message "Read only text copied to kill ring")
8263 ;; or else, signal an error if the buffer is read-only;
8264 (barf-if-buffer-read-only)
8265 ;; and, in any case, signal that the text is read-only.
8266 (signal 'text-read-only (list (current-buffer)))))))
8271 @node condition-case, Lisp macro, Complete kill-region, kill-region
8272 @comment node-name, next, previous, up
8273 @subsection @code{condition-case}
8274 @findex condition-case
8276 As we have seen earlier (@pxref{Making Errors, , Generate an Error
8277 Message}), when the Emacs Lisp interpreter has trouble evaluating an
8278 expression, it provides you with help; in the jargon, this is called
8279 ``signaling an error''. Usually, the computer stops the program and
8280 shows you a message.
8282 However, some programs undertake complicated actions. They should not
8283 simply stop on an error. In the @code{kill-region} function, the most
8284 likely error is that you will try to kill text that is read-only and
8285 cannot be removed. So the @code{kill-region} function contains code
8286 to handle this circumstance. This code, which makes up the body of
8287 the @code{kill-region} function, is inside of a @code{condition-case}
8291 The template for @code{condition-case} looks like this:
8298 @var{error-handler}@dots{})
8302 The second argument, @var{bodyform}, is straightforward. The
8303 @code{condition-case} special form causes the Lisp interpreter to
8304 evaluate the code in @var{bodyform}. If no error occurs, the special
8305 form returns the code's value and produces the side-effects, if any.
8307 In short, the @var{bodyform} part of a @code{condition-case}
8308 expression determines what should happen when everything works
8311 However, if an error occurs, among its other actions, the function
8312 generating the error signal will define one or more error condition
8315 An error handler is the third argument to @code{condition case}.
8316 An error handler has two parts, a @var{condition-name} and a
8317 @var{body}. If the @var{condition-name} part of an error handler
8318 matches a condition name generated by an error, then the @var{body}
8319 part of the error handler is run.
8321 As you will expect, the @var{condition-name} part of an error handler
8322 may be either a single condition name or a list of condition names.
8324 Also, a complete @code{condition-case} expression may contain more
8325 than one error handler. When an error occurs, the first applicable
8328 Lastly, the first argument to the @code{condition-case} expression,
8329 the @var{var} argument, is sometimes bound to a variable that
8330 contains information about the error. However, if that argument is
8331 nil, as is the case in @code{kill-region}, that information is
8335 In brief, in the @code{kill-region} function, the code
8336 @code{condition-case} works like this:
8340 @var{If no errors}, @var{run only this code}
8341 @var{but}, @var{if errors}, @var{run this other code}.
8348 copy-region-as-kill is short, 12 lines, and uses
8349 filter-buffer-substring, which is longer, 39 lines
8350 and has delete-and-extract-region in it.
8351 delete-and-extract-region is written in C.
8353 see Initializing a Variable with @code{defvar}
8355 Initializing a Variable with @code{defvar} includes line 8350
8358 @node Lisp macro, , condition-case, kill-region
8359 @comment node-name, next, previous, up
8360 @subsection Lisp macro
8364 The part of the @code{condition-case} expression that is evaluated in
8365 the expectation that all goes well has a @code{when}. The code uses
8366 @code{when} to determine whether the @code{string} variable points to
8369 A @code{when} expression is simply a programmers' convenience. It is
8370 an @code{if} without the possibility of an else clause. In your mind,
8371 you can replace @code{when} with @code{if} and understand what goes
8372 on. That is what the Lisp interpreter does.
8374 Technically speaking, @code{when} is a Lisp macro. A Lisp @dfn{macro}
8375 enables you to define new control constructs and other language
8376 features. It tells the interpreter how to compute another Lisp
8377 expression which will in turn compute the value. In this case, the
8378 `other expression' is an @code{if} expression.
8380 The @code{kill-region} function definition also has an @code{unless}
8381 macro; it is the converse of @code{when}. The @code{unless} macro is
8382 an @code{if} without a then clause
8384 For more about Lisp macros, see @ref{Macros, , Macros, elisp, The GNU
8385 Emacs Lisp Reference Manual}. The C programming language also
8386 provides macros. These are different, but also useful.
8389 We will briefly look at C macros in
8390 @ref{Digression into C}.
8394 Regarding the @code{when} macro, in the @code{condition-case}
8395 expression, when the string has content, then another conditional
8396 expression is executed. This is an @code{if} with both a then-part
8401 (if (eq last-command 'kill-region)
8402 (kill-append string (< end beg) yank-handler)
8403 (kill-new string nil yank-handler))
8407 The then-part is evaluated if the previous command was another call to
8408 @code{kill-region}; if not, the else-part is evaluated.
8410 @code{yank-handler} is an optional argument to @code{kill-region} that
8411 tells the @code{kill-append} and @code{kill-new} functions how deal
8412 with properties added to the text, such as `bold' or `italics'.
8414 @code{last-command} is a variable that comes with Emacs that we have
8415 not seen before. Normally, whenever a function is executed, Emacs
8416 sets the value of @code{last-command} to the previous command.
8419 In this segment of the definition, the @code{if} expression checks
8420 whether the previous command was @code{kill-region}. If it was,
8423 (kill-append string (< end beg) yank-handler)
8427 concatenates a copy of the newly clipped text to the just previously
8428 clipped text in the kill ring.
8430 @node copy-region-as-kill, Digression into C, kill-region, Cutting & Storing Text
8431 @comment node-name, next, previous, up
8432 @section @code{copy-region-as-kill}
8433 @findex copy-region-as-kill
8436 The @code{copy-region-as-kill} function copies a region of text from a
8437 buffer and (via either @code{kill-append} or @code{kill-new}) saves it
8438 in the @code{kill-ring}.
8440 If you call @code{copy-region-as-kill} immediately after a
8441 @code{kill-region} command, Emacs appends the newly copied text to the
8442 previously copied text. This means that if you yank back the text, you
8443 get it all, from both this and the previous operation. On the other
8444 hand, if some other command precedes the @code{copy-region-as-kill},
8445 the function copies the text into a separate entry in the kill ring.
8448 * Complete copy-region-as-kill:: The complete function definition.
8449 * copy-region-as-kill body:: The body of @code{copy-region-as-kill}.
8452 @node Complete copy-region-as-kill, copy-region-as-kill body, copy-region-as-kill, copy-region-as-kill
8454 @unnumberedsubsec The complete @code{copy-region-as-kill} function definition
8458 Here is the complete text of the version 22 @code{copy-region-as-kill}
8463 (defun copy-region-as-kill (beg end)
8464 "Save the region as if killed, but don't kill it.
8465 In Transient Mark mode, deactivate the mark.
8466 If `interprogram-cut-function' is non-nil, also save the text for a window
8467 system cut and paste."
8471 (if (eq last-command 'kill-region)
8472 (kill-append (filter-buffer-substring beg end) (< end beg))
8473 (kill-new (filter-buffer-substring beg end)))
8476 (if transient-mark-mode
8477 (setq deactivate-mark t))
8483 As usual, this function can be divided into its component parts:
8487 (defun copy-region-as-kill (@var{argument-list})
8488 "@var{documentation}@dots{}"
8494 The arguments are @code{beg} and @code{end} and the function is
8495 interactive with @code{"r"}, so the two arguments must refer to the
8496 beginning and end of the region. If you have been reading though this
8497 document from the beginning, understanding these parts of a function is
8498 almost becoming routine.
8500 The documentation is somewhat confusing unless you remember that the
8501 word `kill' has a meaning different from usual. The `Transient Mark'
8502 and @code{interprogram-cut-function} comments explain certain
8505 After you once set a mark, a buffer always contains a region. If you
8506 wish, you can use Transient Mark mode to highlight the region
8507 temporarily. (No one wants to highlight the region all the time, so
8508 Transient Mark mode highlights it only at appropriate times. Many
8509 people turn off Transient Mark mode, so the region is never
8512 Also, a windowing system allows you to copy, cut, and paste among
8513 different programs. In the X windowing system, for example, the
8514 @code{interprogram-cut-function} function is @code{x-select-text},
8515 which works with the windowing system's equivalent of the Emacs kill
8518 The body of the @code{copy-region-as-kill} function starts with an
8519 @code{if} clause. What this clause does is distinguish between two
8520 different situations: whether or not this command is executed
8521 immediately after a previous @code{kill-region} command. In the first
8522 case, the new region is appended to the previously copied text.
8523 Otherwise, it is inserted into the beginning of the kill ring as a
8524 separate piece of text from the previous piece.
8526 The last two lines of the function prevent the region from lighting up
8527 if Transient Mark mode is turned on.
8529 The body of @code{copy-region-as-kill} merits discussion in detail.
8531 @node copy-region-as-kill body, , Complete copy-region-as-kill, copy-region-as-kill
8532 @comment node-name, next, previous, up
8533 @subsection The Body of @code{copy-region-as-kill}
8535 The @code{copy-region-as-kill} function works in much the same way as
8536 the @code{kill-region} function. Both are written so that two or more
8537 kills in a row combine their text into a single entry. If you yank
8538 back the text from the kill ring, you get it all in one piece.
8539 Moreover, kills that kill forward from the current position of the
8540 cursor are added to the end of the previously copied text and commands
8541 that copy text backwards add it to the beginning of the previously
8542 copied text. This way, the words in the text stay in the proper
8545 Like @code{kill-region}, the @code{copy-region-as-kill} function makes
8546 use of the @code{last-command} variable that keeps track of the
8547 previous Emacs command.
8550 * last-command & this-command::
8551 * kill-append function::
8552 * kill-new function::
8555 @node last-command & this-command, kill-append function, copy-region-as-kill body, copy-region-as-kill body
8557 @unnumberedsubsubsec @code{last-command} and @code{this-command}
8560 Normally, whenever a function is executed, Emacs sets the value of
8561 @code{this-command} to the function being executed (which in this case
8562 would be @code{copy-region-as-kill}). At the same time, Emacs sets
8563 the value of @code{last-command} to the previous value of
8564 @code{this-command}.
8566 In the first part of the body of the @code{copy-region-as-kill}
8567 function, an @code{if} expression determines whether the value of
8568 @code{last-command} is @code{kill-region}. If so, the then-part of
8569 the @code{if} expression is evaluated; it uses the @code{kill-append}
8570 function to concatenate the text copied at this call to the function
8571 with the text already in the first element (the @sc{car}) of the kill
8572 ring. On the other hand, if the value of @code{last-command} is not
8573 @code{kill-region}, then the @code{copy-region-as-kill} function
8574 attaches a new element to the kill ring using the @code{kill-new}
8578 The @code{if} expression reads as follows; it uses @code{eq}:
8582 (if (eq last-command 'kill-region)
8584 (kill-append (filter-buffer-substring beg end) (< end beg))
8586 (kill-new (filter-buffer-substring beg end)))
8590 @findex filter-buffer-substring
8591 (The @code{filter-buffer-substring} function returns a filtered
8592 substring of the buffer, if any. Optionally---the arguments are not
8593 here, so neither is done---the function may delete the initial text or
8594 return the text without its properties; this function is a replacement
8595 for the older @code{buffer-substring} function, which came before text
8596 properties were implemented.)
8598 @findex eq @r{(example of use)}
8600 The @code{eq} function tests whether its first argument is the same Lisp
8601 object as its second argument. The @code{eq} function is similar to the
8602 @code{equal} function in that it is used to test for equality, but
8603 differs in that it determines whether two representations are actually
8604 the same object inside the computer, but with different names.
8605 @code{equal} determines whether the structure and contents of two
8606 expressions are the same.
8608 If the previous command was @code{kill-region}, then the Emacs Lisp
8609 interpreter calls the @code{kill-append} function
8611 @node kill-append function, kill-new function, last-command & this-command, copy-region-as-kill body
8612 @unnumberedsubsubsec The @code{kill-append} function
8616 The @code{kill-append} function looks like this:
8621 (defun kill-append (string before-p &optional yank-handler)
8622 "Append STRING to the end of the latest kill in the kill ring.
8623 If BEFORE-P is non-nil, prepend STRING to the kill.
8625 (let* ((cur (car kill-ring)))
8626 (kill-new (if before-p (concat string cur) (concat cur string))
8627 (or (= (length cur) 0)
8629 (get-text-property 0 'yank-handler cur)))
8636 (defun kill-append (string before-p)
8637 "Append STRING to the end of the latest kill in the kill ring.
8638 If BEFORE-P is non-nil, prepend STRING to the kill.
8639 If `interprogram-cut-function' is set, pass the resulting kill to
8641 (kill-new (if before-p
8642 (concat string (car kill-ring))
8643 (concat (car kill-ring) string))
8648 The @code{kill-append} function is fairly straightforward. It uses
8649 the @code{kill-new} function, which we will discuss in more detail in
8652 (Also, the function provides an optional argument called
8653 @code{yank-handler}; when invoked, this argument tells the function
8654 how to deal with properties added to the text, such as `bold' or
8657 @c !!! bug in GNU Emacs 22 version of kill-append ?
8658 It has a @code{let*} function to set the value of the first element of
8659 the kill ring to @code{cur}. (I do not know why the function does not
8660 use @code{let} instead; only one value is set in the expression.
8661 Perhaps this is a bug that produces no problems?)
8663 Consider the conditional that is one of the two arguments to
8664 @code{kill-new}. It uses @code{concat} to concatenate the new text to
8665 the @sc{car} of the kill ring. Whether it prepends or appends the
8666 text depends on the results of an @code{if} expression:
8670 (if before-p ; @r{if-part}
8671 (concat string cur) ; @r{then-part}
8672 (concat cur string)) ; @r{else-part}
8677 If the region being killed is before the region that was killed in the
8678 last command, then it should be prepended before the material that was
8679 saved in the previous kill; and conversely, if the killed text follows
8680 what was just killed, it should be appended after the previous text.
8681 The @code{if} expression depends on the predicate @code{before-p} to
8682 decide whether the newly saved text should be put before or after the
8683 previously saved text.
8685 The symbol @code{before-p} is the name of one of the arguments to
8686 @code{kill-append}. When the @code{kill-append} function is
8687 evaluated, it is bound to the value returned by evaluating the actual
8688 argument. In this case, this is the expression @code{(< end beg)}.
8689 This expression does not directly determine whether the killed text in
8690 this command is located before or after the kill text of the last
8691 command; what it does is determine whether the value of the variable
8692 @code{end} is less than the value of the variable @code{beg}. If it
8693 is, it means that the user is most likely heading towards the
8694 beginning of the buffer. Also, the result of evaluating the predicate
8695 expression, @code{(< end beg)}, will be true and the text will be
8696 prepended before the previous text. On the other hand, if the value of
8697 the variable @code{end} is greater than the value of the variable
8698 @code{beg}, the text will be appended after the previous text.
8701 When the newly saved text will be prepended, then the string with the new
8702 text will be concatenated before the old text:
8710 But if the text will be appended, it will be concatenated
8714 (concat cur string))
8717 To understand how this works, we first need to review the
8718 @code{concat} function. The @code{concat} function links together or
8719 unites two strings of text. The result is a string. For example:
8723 (concat "abc" "def")
8729 (car '("first element" "second element")))
8730 @result{} "new first element"
8733 '("first element" "second element")) " modified")
8734 @result{} "first element modified"
8738 We can now make sense of @code{kill-append}: it modifies the contents
8739 of the kill ring. The kill ring is a list, each element of which is
8740 saved text. The @code{kill-append} function uses the @code{kill-new}
8741 function which in turn uses the @code{setcar} function.
8743 @node kill-new function, , kill-append function, copy-region-as-kill body
8744 @unnumberedsubsubsec The @code{kill-new} function
8747 @c in GNU Emacs 22, additional documentation to kill-new:
8749 Optional third arguments YANK-HANDLER controls how the STRING is later
8750 inserted into a buffer; see `insert-for-yank' for details.
8751 When a yank handler is specified, STRING must be non-empty (the yank
8752 handler, if non-nil, is stored as a `yank-handler' text property on STRING).
8754 When the yank handler has a non-nil PARAM element, the original STRING
8755 argument is not used by `insert-for-yank'. However, since Lisp code
8756 may access and use elements from the kill ring directly, the STRING
8757 argument should still be a \"useful\" string for such uses."
8760 The @code{kill-new} function looks like this:
8764 (defun kill-new (string &optional replace yank-handler)
8765 "Make STRING the latest kill in the kill ring.
8766 Set `kill-ring-yank-pointer' to point to it.
8768 If `interprogram-cut-function' is non-nil, apply it to STRING.
8769 Optional second argument REPLACE non-nil means that STRING will replace
8770 the front of the kill ring, rather than being added to the list.
8774 (if (> (length string) 0)
8776 (put-text-property 0 (length string)
8777 'yank-handler yank-handler string))
8779 (signal 'args-out-of-range
8780 (list string "yank-handler specified for empty string"))))
8783 (if (fboundp 'menu-bar-update-yank-menu)
8784 (menu-bar-update-yank-menu string (and replace (car kill-ring))))
8787 (if (and replace kill-ring)
8788 (setcar kill-ring string)
8789 (push string kill-ring)
8790 (if (> (length kill-ring) kill-ring-max)
8791 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)))
8794 (setq kill-ring-yank-pointer kill-ring)
8795 (if interprogram-cut-function
8796 (funcall interprogram-cut-function string (not replace))))
8801 (defun kill-new (string &optional replace)
8802 "Make STRING the latest kill in the kill ring.
8803 Set the kill-ring-yank pointer to point to it.
8804 If `interprogram-cut-function' is non-nil, apply it to STRING.
8805 Optional second argument REPLACE non-nil means that STRING will replace
8806 the front of the kill ring, rather than being added to the list."
8807 (and (fboundp 'menu-bar-update-yank-menu)
8808 (menu-bar-update-yank-menu string (and replace (car kill-ring))))
8809 (if (and replace kill-ring)
8810 (setcar kill-ring string)
8811 (setq kill-ring (cons string kill-ring))
8812 (if (> (length kill-ring) kill-ring-max)
8813 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)))
8814 (setq kill-ring-yank-pointer kill-ring)
8815 (if interprogram-cut-function
8816 (funcall interprogram-cut-function string (not replace))))
8819 (Notice that the function is not interactive.)
8821 As usual, we can look at this function in parts.
8823 The function definition has an optional @code{yank-handler} argument,
8824 which when invoked tells the function how to deal with properties
8825 added to the text, such as `bold' or `italics'. We will skip that.
8828 The first line of the documentation makes sense:
8831 Make STRING the latest kill in the kill ring.
8835 Let's skip over the rest of the documentation for the moment.
8838 Also, let's skip over the initial @code{if} expression and those lines
8839 of code involving @code{menu-bar-update-yank-menu}. We will explain
8843 The critical lines are these:
8847 (if (and replace kill-ring)
8849 (setcar kill-ring string)
8853 (push string kill-ring)
8856 (setq kill-ring (cons string kill-ring))
8857 (if (> (length kill-ring) kill-ring-max)
8858 ;; @r{avoid overly long kill ring}
8859 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)))
8862 (setq kill-ring-yank-pointer kill-ring)
8863 (if interprogram-cut-function
8864 (funcall interprogram-cut-function string (not replace))))
8868 The conditional test is @w{@code{(and replace kill-ring)}}.
8869 This will be true when two conditions are met: the kill ring has
8870 something in it, and the @code{replace} variable is true.
8873 When the @code{kill-append} function sets @code{replace} to be true
8874 and when the kill ring has at least one item in it, the @code{setcar}
8875 expression is executed:
8878 (setcar kill-ring string)
8881 The @code{setcar} function actually changes the first element of the
8882 @code{kill-ring} list to the value of @code{string}. It replaces the
8886 On the other hand, if the kill ring is empty, or replace is false, the
8887 else-part of the condition is executed:
8890 (push string kill-ring)
8895 @code{push} puts its first argument onto the second. It is similar to
8899 (setq kill-ring (cons string kill-ring))
8907 (add-to-list kill-ring string)
8911 When it is false, the expression first constructs a new version of the
8912 kill ring by prepending @code{string} to the existing kill ring as a
8913 new element (that is what the @code{push} does). Then it executes a
8914 second @code{if} clause. This second @code{if} clause keeps the kill
8915 ring from growing too long.
8917 Let's look at these two expressions in order.
8919 The @code{push} line of the else-part sets the new value of the kill
8920 ring to what results from adding the string being killed to the old
8923 We can see how this works with an example.
8929 (setq example-list '("here is a clause" "another clause"))
8934 After evaluating this expression with @kbd{C-x C-e}, you can evaluate
8935 @code{example-list} and see what it returns:
8940 @result{} ("here is a clause" "another clause")
8946 Now, we can add a new element on to this list by evaluating the
8947 following expression:
8948 @findex push, @r{example}
8951 (push "a third clause" example-list)
8956 When we evaluate @code{example-list}, we find its value is:
8961 @result{} ("a third clause" "here is a clause" "another clause")
8966 Thus, the third clause is added to the list by @code{push}.
8969 Now for the second part of the @code{if} clause. This expression
8970 keeps the kill ring from growing too long. It looks like this:
8974 (if (> (length kill-ring) kill-ring-max)
8975 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil))
8979 The code checks whether the length of the kill ring is greater than
8980 the maximum permitted length. This is the value of
8981 @code{kill-ring-max} (which is 60, by default). If the length of the
8982 kill ring is too long, then this code sets the last element of the
8983 kill ring to @code{nil}. It does this by using two functions,
8984 @code{nthcdr} and @code{setcdr}.
8986 We looked at @code{setcdr} earlier (@pxref{setcdr, , @code{setcdr}}).
8987 It sets the @sc{cdr} of a list, just as @code{setcar} sets the
8988 @sc{car} of a list. In this case, however, @code{setcdr} will not be
8989 setting the @sc{cdr} of the whole kill ring; the @code{nthcdr}
8990 function is used to cause it to set the @sc{cdr} of the next to last
8991 element of the kill ring---this means that since the @sc{cdr} of the
8992 next to last element is the last element of the kill ring, it will set
8993 the last element of the kill ring.
8995 @findex nthcdr, @r{example}
8996 The @code{nthcdr} function works by repeatedly taking the @sc{cdr} of a
8997 list---it takes the @sc{cdr} of the @sc{cdr} of the @sc{cdr}
8998 @dots{} It does this @var{N} times and returns the results.
8999 (@xref{nthcdr, , @code{nthcdr}}.)
9001 @findex setcdr, @r{example}
9002 Thus, if we had a four element list that was supposed to be three
9003 elements long, we could set the @sc{cdr} of the next to last element
9004 to @code{nil}, and thereby shorten the list. (If you set the last
9005 element to some other value than @code{nil}, which you could do, then
9006 you would not have shortened the list. @xref{setcdr, ,
9009 You can see shortening by evaluating the following three expressions
9010 in turn. First set the value of @code{trees} to @code{(maple oak pine
9011 birch)}, then set the @sc{cdr} of its second @sc{cdr} to @code{nil}
9012 and then find the value of @code{trees}:
9016 (setq trees '(maple oak pine birch))
9017 @result{} (maple oak pine birch)
9021 (setcdr (nthcdr 2 trees) nil)
9025 @result{} (maple oak pine)
9030 (The value returned by the @code{setcdr} expression is @code{nil} since
9031 that is what the @sc{cdr} is set to.)
9033 To repeat, in @code{kill-new}, the @code{nthcdr} function takes the
9034 @sc{cdr} a number of times that is one less than the maximum permitted
9035 size of the kill ring and @code{setcdr} sets the @sc{cdr} of that
9036 element (which will be the rest of the elements in the kill ring) to
9037 @code{nil}. This prevents the kill ring from growing too long.
9040 The next to last expression in the @code{kill-new} function is
9043 (setq kill-ring-yank-pointer kill-ring)
9046 The @code{kill-ring-yank-pointer} is a global variable that is set to be
9047 the @code{kill-ring}.
9049 Even though the @code{kill-ring-yank-pointer} is called a
9050 @samp{pointer}, it is a variable just like the kill ring. However, the
9051 name has been chosen to help humans understand how the variable is used.
9054 Now, to return to an early expression in the body of the function:
9058 (if (fboundp 'menu-bar-update-yank-menu)
9059 (menu-bar-update-yank-menu string (and replace (car kill-ring))))
9064 It starts with an @code{if} expression
9066 In this case, the expression tests first to see whether
9067 @code{menu-bar-update-yank-menu} exists as a function, and if so,
9068 calls it. The @code{fboundp} function returns true if the symbol it
9069 is testing has a function definition that `is not void'. If the
9070 symbol's function definition were void, we would receive an error
9071 message, as we did when we created errors intentionally (@pxref{Making
9072 Errors, , Generate an Error Message}).
9075 The then-part contains an expression whose first element is the
9076 function @code{and}.
9079 The @code{and} special form evaluates each of its arguments until one
9080 of the arguments returns a value of @code{nil}, in which case the
9081 @code{and} expression returns @code{nil}; however, if none of the
9082 arguments returns a value of @code{nil}, the value resulting from
9083 evaluating the last argument is returned. (Since such a value is not
9084 @code{nil}, it is considered true in Emacs Lisp.) In other words, an
9085 @code{and} expression returns a true value only if all its arguments
9086 are true. (@xref{Second Buffer Related Review}.)
9088 The expression determines whether the second argument to
9089 @code{menu-bar-update-yank-menu} is true or not.
9091 ;; If we're supposed to be extending an existing string, and that
9092 ;; string really is at the front of the menu, then update it in place.
9095 @code{menu-bar-update-yank-menu} is one of the functions that make it
9096 possible to use the `Select and Paste' menu in the Edit item of a menu
9097 bar; using a mouse, you can look at the various pieces of text you
9098 have saved and select one piece to paste.
9100 The last expression in the @code{kill-new} function adds the newly
9101 copied string to whatever facility exists for copying and pasting
9102 among different programs running in a windowing system. In the X
9103 Windowing system, for example, the @code{x-select-text} function takes
9104 the string and stores it in memory operated by X. You can paste the
9105 string in another program, such as an Xterm.
9108 The expression looks like this:
9112 (if interprogram-cut-function
9113 (funcall interprogram-cut-function string (not replace))))
9117 If an @code{interprogram-cut-function} exists, then Emacs executes
9118 @code{funcall}, which in turn calls its first argument as a function
9119 and passes the remaining arguments to it. (Incidentally, as far as I
9120 can see, this @code{if} expression could be replaced by an @code{and}
9121 expression similar to the one in the first part of the function.)
9123 We are not going to discuss windowing systems and other programs
9124 further, but merely note that this is a mechanism that enables GNU
9125 Emacs to work easily and well with other programs.
9127 This code for placing text in the kill ring, either concatenated with
9128 an existing element or as a new element, leads us to the code for
9129 bringing back text that has been cut out of the buffer---the yank
9130 commands. However, before discussing the yank commands, it is better
9131 to learn how lists are implemented in a computer. This will make
9132 clear such mysteries as the use of the term `pointer'. But before
9133 that, we will digress into C.
9136 @c is this true in Emacs 22? Does not seems to be
9138 (If the @w{@code{(< end beg))}}
9139 expression is true, @code{kill-append} prepends the string to the just
9140 previously clipped text. For a detailed discussion, see
9141 @ref{kill-append function, , The @code{kill-append} function}.)
9143 If you then yank back the text, i.e., `paste' it, you get both
9144 pieces of text at once. That way, if you delete two words in a row,
9145 and then yank them back, you get both words, in their proper order,
9146 with one yank. (The @w{@code{(< end beg))}} expression makes sure the
9149 On the other hand, if the previous command is not @code{kill-region},
9150 then the @code{kill-new} function is called, which adds the text to
9151 the kill ring as the latest item, and sets the
9152 @code{kill-ring-yank-pointer} variable to point to it.
9156 @c Evidently, changed for Emacs 22. The zap-to-char command does not
9157 @c use the delete-and-extract-region function
9159 2006 Oct 26, the Digression into C is now OK but should come after
9160 copy-region-as-kill and filter-buffer-substring
9164 copy-region-as-kill is short, 12 lines, and uses
9165 filter-buffer-substring, which is longer, 39 lines
9166 and has delete-and-extract-region in it.
9167 delete-and-extract-region is written in C.
9169 see Initializing a Variable with @code{defvar}
9172 @node Digression into C, defvar, copy-region-as-kill, Cutting & Storing Text
9173 @comment node-name, next, previous, up
9174 @section Digression into C
9175 @findex delete-and-extract-region
9176 @cindex C, a digression into
9177 @cindex Digression into C
9179 The @code{copy-region-as-kill} function (@pxref{copy-region-as-kill, ,
9180 @code{copy-region-as-kill}}) uses the @code{filter-buffer-substring}
9181 function, which in turn uses the @code{delete-and-extract-region}
9182 function. It removes the contents of a region and you cannot get them
9185 Unlike the other code discussed here, the
9186 @code{delete-and-extract-region} function is not written in Emacs
9187 Lisp; it is written in C and is one of the primitives of the GNU Emacs
9188 system. Since it is very simple, I will digress briefly from Lisp and
9191 @c GNU Emacs 22 in /usr/local/src/emacs/src/editfns.c
9192 @c the DEFUN for buffer-substring-no-properties
9195 Like many of the other Emacs primitives,
9196 @code{delete-and-extract-region} is written as an instance of a C
9197 macro, a macro being a template for code. The complete macro looks
9202 DEFUN ("buffer-substring-no-properties", Fbuffer_substring_no_properties,
9203 Sbuffer_substring_no_properties, 2, 2, 0,
9204 doc: /* Return the characters of part of the buffer,
9205 without the text properties.
9206 The two arguments START and END are character positions;
9207 they can be in either order. */)
9209 Lisp_Object start, end;
9213 validate_region (&start, &end);
9217 return make_buffer_string (b, e, 0);
9222 Without going into the details of the macro writing process, let me
9223 point out that this macro starts with the word @code{DEFUN}. The word
9224 @code{DEFUN} was chosen since the code serves the same purpose as
9225 @code{defun} does in Lisp. (The @code{DEFUN} C macro is defined in
9226 @file{emacs/src/lisp.h}.)
9228 The word @code{DEFUN} is followed by seven parts inside of
9233 The first part is the name given to the function in Lisp,
9234 @code{delete-and-extract-region}.
9237 The second part is the name of the function in C,
9238 @code{Fdelete_and_extract_region}. By convention, it starts with
9239 @samp{F}. Since C does not use hyphens in names, underscores are used
9243 The third part is the name for the C constant structure that records
9244 information on this function for internal use. It is the name of the
9245 function in C but begins with an @samp{S} instead of an @samp{F}.
9248 The fourth and fifth parts specify the minimum and maximum number of
9249 arguments the function can have. This function demands exactly 2
9253 The sixth part is nearly like the argument that follows the
9254 @code{interactive} declaration in a function written in Lisp: a letter
9255 followed, perhaps, by a prompt. The only difference from the Lisp is
9256 when the macro is called with no arguments. Then you write a @code{0}
9257 (which is a `null string'), as in this macro.
9259 If you were to specify arguments, you would place them between
9260 quotation marks. The C macro for @code{goto-char} includes
9261 @code{"NGoto char: "} in this position to indicate that the function
9262 expects a raw prefix, in this case, a numerical location in a buffer,
9263 and provides a prompt.
9266 The seventh part is a documentation string, just like the one for a
9267 function written in Emacs Lisp, except that every newline must be
9268 written explicitly as @samp{\n} followed by a backslash and carriage
9272 Thus, the first two lines of documentation for @code{goto-char} are
9277 "Set point to POSITION, a number or marker.\n\
9278 Beginning of buffer is position (point-min), end is (point-max)."
9284 In a C macro, the formal parameters come next, with a statement of
9285 what kind of object they are, followed by what might be called the `body'
9286 of the macro. For @code{delete-and-extract-region} the `body'
9287 consists of the following four lines:
9291 validate_region (&start, &end);
9292 if (XINT (start) == XINT (end))
9293 return build_string ("");
9294 return del_range_1 (XINT (start), XINT (end), 1, 1);
9298 The @code{validate_region} function checks whether the values
9299 passed as the beginning and end of the region are the proper type and
9300 are within range. If the beginning and end positions are the same,
9301 then return and empty string.
9303 The @code{del_range_1} function actually deletes the text. It is a
9304 complex function we will not look into. It updates the buffer and
9305 does other things. However, it is worth looking at the two arguments
9306 passed to @code{del_range}. These are @w{@code{XINT (start)}} and
9307 @w{@code{XINT (end)}}.
9309 As far as the C language is concerned, @code{start} and @code{end} are
9310 two integers that mark the beginning and end of the region to be
9311 deleted@footnote{More precisely, and requiring more expert knowledge
9312 to understand, the two integers are of type `Lisp_Object', which can
9313 also be a C union instead of an integer type.}.
9315 In early versions of Emacs, these two numbers were thirty-two bits
9316 long, but the code is slowly being generalized to handle other
9317 lengths. Three of the available bits are used to specify the type of
9318 information; the remaining bits are used as `content'.
9320 @samp{XINT} is a C macro that extracts the relevant number from the
9321 longer collection of bits; the three other bits are discarded.
9324 The command in @code{delete-and-extract-region} looks like this:
9327 del_range_1 (XINT (start), XINT (end), 1, 1);
9331 It deletes the region between the beginning position, @code{start},
9332 and the ending position, @code{end}.
9334 From the point of view of the person writing Lisp, Emacs is all very
9335 simple; but hidden underneath is a great deal of complexity to make it
9338 @node defvar, cons & search-fwd Review, Digression into C, Cutting & Storing Text
9339 @comment node-name, next, previous, up
9340 @section Initializing a Variable with @code{defvar}
9342 @cindex Initializing a variable
9343 @cindex Variable initialization
9348 copy-region-as-kill is short, 12 lines, and uses
9349 filter-buffer-substring, which is longer, 39 lines
9350 and has delete-and-extract-region in it.
9351 delete-and-extract-region is written in C.
9353 see Initializing a Variable with @code{defvar}
9357 The @code{copy-region-as-kill} function is written in Emacs Lisp. Two
9358 functions within it, @code{kill-append} and @code{kill-new}, copy a
9359 region in a buffer and save it in a variable called the
9360 @code{kill-ring}. This section describes how the @code{kill-ring}
9361 variable is created and initialized using the @code{defvar} special
9364 (Again we note that the term @code{kill-ring} is a misnomer. The text
9365 that is clipped out of the buffer can be brought back; it is not a ring
9366 of corpses, but a ring of resurrectable text.)
9368 In Emacs Lisp, a variable such as the @code{kill-ring} is created and
9369 given an initial value by using the @code{defvar} special form. The
9370 name comes from ``define variable''.
9372 The @code{defvar} special form is similar to @code{setq} in that it sets
9373 the value of a variable. It is unlike @code{setq} in two ways: first,
9374 it only sets the value of the variable if the variable does not already
9375 have a value. If the variable already has a value, @code{defvar} does
9376 not override the existing value. Second, @code{defvar} has a
9377 documentation string.
9379 (Another special form, @code{defcustom}, is designed for variables
9380 that people customize. It has more features than @code{defvar}.
9381 (@xref{defcustom, , Setting Variables with @code{defcustom}}.)
9384 * See variable current value::
9385 * defvar and asterisk::
9388 @node See variable current value, defvar and asterisk, defvar, defvar
9390 @unnumberedsubsec Seeing the Current Value of a Variable
9393 You can see the current value of a variable, any variable, by using
9394 the @code{describe-variable} function, which is usually invoked by
9395 typing @kbd{C-h v}. If you type @kbd{C-h v} and then @code{kill-ring}
9396 (followed by @key{RET}) when prompted, you will see what is in your
9397 current kill ring---this may be quite a lot! Conversely, if you have
9398 been doing nothing this Emacs session except read this document, you
9399 may have nothing in it. Also, you will see the documentation for
9405 List of killed text sequences.
9406 Since the kill ring is supposed to interact nicely with cut-and-paste
9407 facilities offered by window systems, use of this variable should
9410 interact nicely with `interprogram-cut-function' and
9411 `interprogram-paste-function'. The functions `kill-new',
9412 `kill-append', and `current-kill' are supposed to implement this
9413 interaction; you may want to use them instead of manipulating the kill
9419 The kill ring is defined by a @code{defvar} in the following way:
9423 (defvar kill-ring nil
9424 "List of killed text sequences.
9430 In this variable definition, the variable is given an initial value of
9431 @code{nil}, which makes sense, since if you have saved nothing, you want
9432 nothing back if you give a @code{yank} command. The documentation
9433 string is written just like the documentation string of a @code{defun}.
9434 As with the documentation string of the @code{defun}, the first line of
9435 the documentation should be a complete sentence, since some commands,
9436 like @code{apropos}, print only the first line of documentation.
9437 Succeeding lines should not be indented; otherwise they look odd when
9438 you use @kbd{C-h v} (@code{describe-variable}).
9440 @node defvar and asterisk, , See variable current value, defvar
9441 @subsection @code{defvar} and an asterisk
9442 @findex defvar @r{for a user customizable variable}
9443 @findex defvar @r{with an asterisk}
9445 In the past, Emacs used the @code{defvar} special form both for
9446 internal variables that you would not expect a user to change and for
9447 variables that you do expect a user to change. Although you can still
9448 use @code{defvar} for user customizable variables, please use
9449 @code{defcustom} instead, since that special form provides a path into
9450 the Customization commands. (@xref{defcustom, , Specifying Variables
9451 using @code{defcustom}}.)
9453 When you specified a variable using the @code{defvar} special form,
9454 you could distinguish a readily settable variable from others by
9455 typing an asterisk, @samp{*}, in the first column of its documentation
9456 string. For example:
9460 (defvar shell-command-default-error-buffer nil
9461 "*Buffer name for `shell-command' @dots{} error output.
9466 @findex set-variable
9468 You could (and still can) use the @code{set-variable} command to
9469 change the value of @code{shell-command-default-error-buffer}
9470 temporarily. However, options set using @code{set-variable} are set
9471 only for the duration of your editing session. The new values are not
9472 saved between sessions. Each time Emacs starts, it reads the original
9473 value, unless you change the value within your @file{.emacs} file,
9474 either by setting it manually or by using @code{customize}.
9475 @xref{Emacs Initialization, , Your @file{.emacs} File}.
9477 For me, the major use of the @code{set-variable} command is to suggest
9478 variables that I might want to set in my @file{.emacs} file. There
9479 are now more than 700 such variables --- far too many to remember
9480 readily. Fortunately, you can press @key{TAB} after calling the
9481 @code{M-x set-variable} command to see the list of variables.
9482 (@xref{Examining, , Examining and Setting Variables, emacs,
9483 The GNU Emacs Manual}.)
9486 @node cons & search-fwd Review, search Exercises, defvar, Cutting & Storing Text
9487 @comment node-name, next, previous, up
9490 Here is a brief summary of some recently introduced functions.
9495 @code{car} returns the first element of a list; @code{cdr} returns the
9496 second and subsequent elements of a list.
9503 (car '(1 2 3 4 5 6 7))
9505 (cdr '(1 2 3 4 5 6 7))
9506 @result{} (2 3 4 5 6 7)
9511 @code{cons} constructs a list by prepending its first argument to its
9525 @code{funcall} evaluates its first argument as a function. It passes
9526 its remaining arguments to its first argument.
9529 Return the result of taking @sc{cdr} `n' times on a list.
9537 The `rest of the rest', as it were.
9544 (nthcdr 3 '(1 2 3 4 5 6 7))
9551 @code{setcar} changes the first element of a list; @code{setcdr}
9552 changes the second and subsequent elements of a list.
9559 (setq triple '(1 2 3))
9566 (setcdr triple '("foo" "bar"))
9569 @result{} (37 "foo" "bar")
9574 Evaluate each argument in sequence and then return the value of the
9587 @item save-restriction
9588 Record whatever narrowing is in effect in the current buffer, if any,
9589 and restore that narrowing after evaluating the arguments.
9591 @item search-forward
9592 Search for a string, and if the string is found, move point. With a
9593 regular expression, use the similar @code{re-search-forward}.
9594 (@xref{Regexp Search, , Regular Expression Searches}, for an
9595 explanation of regular expression patterns and searches.)
9599 @code{search-forward} and @code{re-search-forward} take four
9604 The string or regular expression to search for.
9607 Optionally, the limit of the search.
9610 Optionally, what to do if the search fails, return @code{nil} or an
9614 Optionally, how many times to repeat the search; if negative, the
9615 search goes backwards.
9619 @itemx delete-and-extract-region
9620 @itemx copy-region-as-kill
9622 @code{kill-region} cuts the text between point and mark from the
9623 buffer and stores that text in the kill ring, so you can get it back
9626 @code{copy-region-as-kill} copies the text between point and mark into
9627 the kill ring, from which you can get it by yanking. The function
9628 does not cut or remove the text from the buffer.
9631 @code{delete-and-extract-region} removes the text between point and
9632 mark from the buffer and throws it away. You cannot get it back.
9633 (This is not an interactive command.)
9636 @node search Exercises, , cons & search-fwd Review, Cutting & Storing Text
9637 @section Searching Exercises
9641 Write an interactive function that searches for a string. If the
9642 search finds the string, leave point after it and display a message
9643 that says ``Found!''. (Do not use @code{search-forward} for the name
9644 of this function; if you do, you will overwrite the existing version of
9645 @code{search-forward} that comes with Emacs. Use a name such as
9646 @code{test-search} instead.)
9649 Write a function that prints the third element of the kill ring in the
9650 echo area, if any; if the kill ring does not contain a third element,
9651 print an appropriate message.
9654 @node List Implementation, Yanking, Cutting & Storing Text, Top
9655 @comment node-name, next, previous, up
9656 @chapter How Lists are Implemented
9657 @cindex Lists in a computer
9659 In Lisp, atoms are recorded in a straightforward fashion; if the
9660 implementation is not straightforward in practice, it is, nonetheless,
9661 straightforward in theory. The atom @samp{rose}, for example, is
9662 recorded as the four contiguous letters @samp{r}, @samp{o}, @samp{s},
9663 @samp{e}. A list, on the other hand, is kept differently. The mechanism
9664 is equally simple, but it takes a moment to get used to the idea. A
9665 list is kept using a series of pairs of pointers. In the series, the
9666 first pointer in each pair points to an atom or to another list, and the
9667 second pointer in each pair points to the next pair, or to the symbol
9668 @code{nil}, which marks the end of the list.
9670 A pointer itself is quite simply the electronic address of what is
9671 pointed to. Hence, a list is kept as a series of electronic addresses.
9674 * Lists diagrammed::
9675 * Symbols as Chest:: Exploring a powerful metaphor.
9679 @node Lists diagrammed, Symbols as Chest, List Implementation, List Implementation
9681 @unnumberedsec Lists diagrammed
9684 For example, the list @code{(rose violet buttercup)} has three elements,
9685 @samp{rose}, @samp{violet}, and @samp{buttercup}. In the computer, the
9686 electronic address of @samp{rose} is recorded in a segment of computer
9687 memory along with the address that gives the electronic address of where
9688 the atom @samp{violet} is located; and that address (the one that tells
9689 where @samp{violet} is located) is kept along with an address that tells
9690 where the address for the atom @samp{buttercup} is located.
9693 This sounds more complicated than it is and is easier seen in a diagram:
9695 @c clear print-postscript-figures
9696 @c !!! cons-cell-diagram #1
9700 ___ ___ ___ ___ ___ ___
9701 |___|___|--> |___|___|--> |___|___|--> nil
9704 --> rose --> violet --> buttercup
9708 @ifset print-postscript-figures
9711 @center @image{cons-1}
9712 %%%% old method of including an image
9713 % \input /usr/local/lib/tex/inputs/psfig.tex
9714 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-1.eps}}
9719 @ifclear print-postscript-figures
9723 ___ ___ ___ ___ ___ ___
9724 |___|___|--> |___|___|--> |___|___|--> nil
9727 --> rose --> violet --> buttercup
9734 In the diagram, each box represents a word of computer memory that
9735 holds a Lisp object, usually in the form of a memory address. The boxes,
9736 i.e.@: the addresses, are in pairs. Each arrow points to what the address
9737 is the address of, either an atom or another pair of addresses. The
9738 first box is the electronic address of @samp{rose} and the arrow points
9739 to @samp{rose}; the second box is the address of the next pair of boxes,
9740 the first part of which is the address of @samp{violet} and the second
9741 part of which is the address of the next pair. The very last box
9742 points to the symbol @code{nil}, which marks the end of the list.
9745 When a variable is set to a list with a function such as @code{setq},
9746 it stores the address of the first box in the variable. Thus,
9747 evaluation of the expression
9750 (setq bouquet '(rose violet buttercup))
9755 creates a situation like this:
9757 @c cons-cell-diagram #2
9763 | ___ ___ ___ ___ ___ ___
9764 --> |___|___|--> |___|___|--> |___|___|--> nil
9767 --> rose --> violet --> buttercup
9771 @ifset print-postscript-figures
9774 @center @image{cons-2}
9775 %%%% old method of including an image
9776 % \input /usr/local/lib/tex/inputs/psfig.tex
9777 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-2.eps}}
9782 @ifclear print-postscript-figures
9788 | ___ ___ ___ ___ ___ ___
9789 --> |___|___|--> |___|___|--> |___|___|--> nil
9792 --> rose --> violet --> buttercup
9799 In this example, the symbol @code{bouquet} holds the address of the first
9803 This same list can be illustrated in a different sort of box notation
9806 @c cons-cell-diagram #2a
9812 | -------------- --------------- ----------------
9813 | | car | cdr | | car | cdr | | car | cdr |
9814 -->| rose | o------->| violet | o------->| butter- | nil |
9815 | | | | | | | cup | |
9816 -------------- --------------- ----------------
9820 @ifset print-postscript-figures
9823 @center @image{cons-2a}
9824 %%%% old method of including an image
9825 % \input /usr/local/lib/tex/inputs/psfig.tex
9826 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-2a.eps}}
9831 @ifclear print-postscript-figures
9837 | -------------- --------------- ----------------
9838 | | car | cdr | | car | cdr | | car | cdr |
9839 -->| rose | o------->| violet | o------->| butter- | nil |
9840 | | | | | | | cup | |
9841 -------------- --------------- ----------------
9847 (Symbols consist of more than pairs of addresses, but the structure of
9848 a symbol is made up of addresses. Indeed, the symbol @code{bouquet}
9849 consists of a group of address-boxes, one of which is the address of
9850 the printed word @samp{bouquet}, a second of which is the address of a
9851 function definition attached to the symbol, if any, a third of which
9852 is the address of the first pair of address-boxes for the list
9853 @code{(rose violet buttercup)}, and so on. Here we are showing that
9854 the symbol's third address-box points to the first pair of
9855 address-boxes for the list.)
9857 If a symbol is set to the @sc{cdr} of a list, the list itself is not
9858 changed; the symbol simply has an address further down the list. (In
9859 the jargon, @sc{car} and @sc{cdr} are `non-destructive'.) Thus,
9860 evaluation of the following expression
9863 (setq flowers (cdr bouquet))
9870 @c cons-cell-diagram #3
9877 | ___ ___ | ___ ___ ___ ___
9878 --> | | | --> | | | | | |
9879 |___|___|----> |___|___|--> |___|___|--> nil
9882 --> rose --> violet --> buttercup
9887 @ifset print-postscript-figures
9890 @center @image{cons-3}
9891 %%%% old method of including an image
9892 % \input /usr/local/lib/tex/inputs/psfig.tex
9893 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-3.eps}}
9898 @ifclear print-postscript-figures
9905 | ___ ___ | ___ ___ ___ ___
9906 --> | | | --> | | | | | |
9907 |___|___|----> |___|___|--> |___|___|--> nil
9910 --> rose --> violet --> buttercup
9918 The value of @code{flowers} is @code{(violet buttercup)}, which is
9919 to say, the symbol @code{flowers} holds the address of the pair of
9920 address-boxes, the first of which holds the address of @code{violet},
9921 and the second of which holds the address of @code{buttercup}.
9923 A pair of address-boxes is called a @dfn{cons cell} or @dfn{dotted
9924 pair}. @xref{Cons Cell Type, , Cons Cell and List Types, elisp, The GNU Emacs Lisp
9925 Reference Manual}, and @ref{Dotted Pair Notation, , Dotted Pair
9926 Notation, elisp, The GNU Emacs Lisp Reference Manual}, for more
9927 information about cons cells and dotted pairs.
9930 The function @code{cons} adds a new pair of addresses to the front of
9931 a series of addresses like that shown above. For example, evaluating
9935 (setq bouquet (cons 'lily bouquet))
9942 @c cons-cell-diagram #4
9949 | ___ ___ ___ ___ | ___ ___ ___ ___
9950 --> | | | | | | --> | | | | | |
9951 |___|___|----> |___|___|----> |___|___|---->|___|___|--> nil
9954 --> lily --> rose --> violet --> buttercup
9959 @ifset print-postscript-figures
9962 @center @image{cons-4}
9963 %%%% old method of including an image
9964 % \input /usr/local/lib/tex/inputs/psfig.tex
9965 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-4.eps}}
9970 @ifclear print-postscript-figures
9977 | ___ ___ ___ ___ | ___ ___ ___ ___
9978 --> | | | | | | --> | | | | | |
9979 |___|___|----> |___|___|----> |___|___|---->|___|___|--> nil
9982 --> lily --> rose --> violet --> buttercup
9991 However, this does not change the value of the symbol
9992 @code{flowers}, as you can see by evaluating the following,
9995 (eq (cdr (cdr bouquet)) flowers)
9999 which returns @code{t} for true.
10001 Until it is reset, @code{flowers} still has the value
10002 @code{(violet buttercup)}; that is, it has the address of the cons
10003 cell whose first address is of @code{violet}. Also, this does not
10004 alter any of the pre-existing cons cells; they are all still there.
10006 Thus, in Lisp, to get the @sc{cdr} of a list, you just get the address
10007 of the next cons cell in the series; to get the @sc{car} of a list,
10008 you get the address of the first element of the list; to @code{cons} a
10009 new element on a list, you add a new cons cell to the front of the list.
10010 That is all there is to it! The underlying structure of Lisp is
10011 brilliantly simple!
10013 And what does the last address in a series of cons cells refer to? It
10014 is the address of the empty list, of @code{nil}.
10016 In summary, when a Lisp variable is set to a value, it is provided with
10017 the address of the list to which the variable refers.
10019 @node Symbols as Chest, List Exercise, Lists diagrammed, List Implementation
10020 @section Symbols as a Chest of Drawers
10021 @cindex Symbols as a Chest of Drawers
10022 @cindex Chest of Drawers, metaphor for a symbol
10023 @cindex Drawers, Chest of, metaphor for a symbol
10025 In an earlier section, I suggested that you might imagine a symbol as
10026 being a chest of drawers. The function definition is put in one
10027 drawer, the value in another, and so on. What is put in the drawer
10028 holding the value can be changed without affecting the contents of the
10029 drawer holding the function definition, and vice-verse.
10031 Actually, what is put in each drawer is the address of the value or
10032 function definition. It is as if you found an old chest in the attic,
10033 and in one of its drawers you found a map giving you directions to
10034 where the buried treasure lies.
10036 (In addition to its name, symbol definition, and variable value, a
10037 symbol has a `drawer' for a @dfn{property list} which can be used to
10038 record other information. Property lists are not discussed here; see
10039 @ref{Property Lists, , Property Lists, elisp, The GNU Emacs Lisp
10040 Reference Manual}.)
10043 Here is a fanciful representation:
10045 @c chest-of-drawers diagram
10050 Chest of Drawers Contents of Drawers
10054 ---------------------
10055 | directions to | [map to]
10056 | symbol name | bouquet
10058 +---------------------+
10060 | symbol definition | [none]
10062 +---------------------+
10063 | directions to | [map to]
10064 | variable value | (rose violet buttercup)
10066 +---------------------+
10068 | property list | [not described here]
10070 +---------------------+
10076 @ifset print-postscript-figures
10079 @center @image{drawers}
10080 %%%% old method of including an image
10081 % \input /usr/local/lib/tex/inputs/psfig.tex
10082 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/drawers.eps}}
10087 @ifclear print-postscript-figures
10092 Chest of Drawers Contents of Drawers
10096 ---------------------
10097 | directions to | [map to]
10098 | symbol name | bouquet
10100 +---------------------+
10102 | symbol definition | [none]
10104 +---------------------+
10105 | directions to | [map to]
10106 | variable value | (rose violet buttercup)
10108 +---------------------+
10110 | property list | [not described here]
10112 +---------------------+
10120 @node List Exercise, , Symbols as Chest, List Implementation
10123 Set @code{flowers} to @code{violet} and @code{buttercup}. Cons two
10124 more flowers on to this list and set this new list to
10125 @code{more-flowers}. Set the @sc{car} of @code{flowers} to a fish.
10126 What does the @code{more-flowers} list now contain?
10128 @node Yanking, Loops & Recursion, List Implementation, Top
10129 @comment node-name, next, previous, up
10130 @chapter Yanking Text Back
10132 @cindex Text retrieval
10133 @cindex Retrieving text
10134 @cindex Pasting text
10136 Whenever you cut text out of a buffer with a `kill' command in GNU Emacs,
10137 you can bring it back with a `yank' command. The text that is cut out of
10138 the buffer is put in the kill ring and the yank commands insert the
10139 appropriate contents of the kill ring back into a buffer (not necessarily
10140 the original buffer).
10142 A simple @kbd{C-y} (@code{yank}) command inserts the first item from
10143 the kill ring into the current buffer. If the @kbd{C-y} command is
10144 followed immediately by @kbd{M-y}, the first element is replaced by
10145 the second element. Successive @kbd{M-y} commands replace the second
10146 element with the third, fourth, or fifth element, and so on. When the
10147 last element in the kill ring is reached, it is replaced by the first
10148 element and the cycle is repeated. (Thus the kill ring is called a
10149 `ring' rather than just a `list'. However, the actual data structure
10150 that holds the text is a list.
10151 @xref{Kill Ring, , Handling the Kill Ring}, for the details of how the
10152 list is handled as a ring.)
10155 * Kill Ring Overview::
10156 * kill-ring-yank-pointer:: The kill ring is a list.
10157 * yank nthcdr Exercises:: The @code{kill-ring-yank-pointer} variable.
10160 @node Kill Ring Overview, kill-ring-yank-pointer, Yanking, Yanking
10161 @comment node-name, next, previous, up
10162 @section Kill Ring Overview
10163 @cindex Kill ring overview
10165 The kill ring is a list of textual strings. This is what it looks like:
10168 ("some text" "a different piece of text" "yet more text")
10171 If this were the contents of my kill ring and I pressed @kbd{C-y}, the
10172 string of characters saying @samp{some text} would be inserted in this
10173 buffer where my cursor is located.
10175 The @code{yank} command is also used for duplicating text by copying it.
10176 The copied text is not cut from the buffer, but a copy of it is put on the
10177 kill ring and is inserted by yanking it back.
10179 Three functions are used for bringing text back from the kill ring:
10180 @code{yank}, which is usually bound to @kbd{C-y}; @code{yank-pop},
10181 which is usually bound to @kbd{M-y}; and @code{rotate-yank-pointer},
10182 which is used by the two other functions.
10184 These functions refer to the kill ring through a variable called the
10185 @code{kill-ring-yank-pointer}. Indeed, the insertion code for both the
10186 @code{yank} and @code{yank-pop} functions is:
10189 (insert (car kill-ring-yank-pointer))
10193 (Well, no more. In GNU Emacs 22, the function has been replaced by
10194 @code{insert-for-yank} which calls @code{insert-for-yank-1}
10195 repetitively for each @code{yank-handler} segment. In turn,
10196 @code{insert-for-yank-1} strips text properties from the inserted text
10197 according to @code{yank-excluded-properties}. Otherwise, it is just
10198 like @code{insert}. We will stick with plain @code{insert} since it
10199 is easier to understand.)
10201 To begin to understand how @code{yank} and @code{yank-pop} work, it is
10202 first necessary to look at the @code{kill-ring-yank-pointer} variable.
10204 @node kill-ring-yank-pointer, yank nthcdr Exercises, Kill Ring Overview, Yanking
10205 @comment node-name, next, previous, up
10206 @section The @code{kill-ring-yank-pointer} Variable
10208 @code{kill-ring-yank-pointer} is a variable, just as @code{kill-ring} is
10209 a variable. It points to something by being bound to the value of what
10210 it points to, like any other Lisp variable.
10213 Thus, if the value of the kill ring is:
10216 ("some text" "a different piece of text" "yet more text")
10221 and the @code{kill-ring-yank-pointer} points to the second clause, the
10222 value of @code{kill-ring-yank-pointer} is:
10225 ("a different piece of text" "yet more text")
10228 As explained in the previous chapter (@pxref{List Implementation}), the
10229 computer does not keep two different copies of the text being pointed to
10230 by both the @code{kill-ring} and the @code{kill-ring-yank-pointer}. The
10231 words ``a different piece of text'' and ``yet more text'' are not
10232 duplicated. Instead, the two Lisp variables point to the same pieces of
10233 text. Here is a diagram:
10235 @c cons-cell-diagram #5
10239 kill-ring kill-ring-yank-pointer
10241 | ___ ___ | ___ ___ ___ ___
10242 ---> | | | --> | | | | | |
10243 |___|___|----> |___|___|--> |___|___|--> nil
10246 | | --> "yet more text"
10248 | --> "a different piece of text"
10255 @ifset print-postscript-figures
10258 @center @image{cons-5}
10259 %%%% old method of including an image
10260 % \input /usr/local/lib/tex/inputs/psfig.tex
10261 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-5.eps}}
10266 @ifclear print-postscript-figures
10270 kill-ring kill-ring-yank-pointer
10272 | ___ ___ | ___ ___ ___ ___
10273 ---> | | | --> | | | | | |
10274 |___|___|----> |___|___|--> |___|___|--> nil
10277 | | --> "yet more text"
10279 | --> "a different piece of text
10288 Both the variable @code{kill-ring} and the variable
10289 @code{kill-ring-yank-pointer} are pointers. But the kill ring itself is
10290 usually described as if it were actually what it is composed of. The
10291 @code{kill-ring} is spoken of as if it were the list rather than that it
10292 points to the list. Conversely, the @code{kill-ring-yank-pointer} is
10293 spoken of as pointing to a list.
10295 These two ways of talking about the same thing sound confusing at first but
10296 make sense on reflection. The kill ring is generally thought of as the
10297 complete structure of data that holds the information of what has recently
10298 been cut out of the Emacs buffers. The @code{kill-ring-yank-pointer}
10299 on the other hand, serves to indicate---that is, to `point to'---that part
10300 of the kill ring of which the first element (the @sc{car}) will be
10304 In GNU Emacs 22, the @code{kill-new} function calls
10306 @code{(setq kill-ring-yank-pointer kill-ring)}
10308 (defun rotate-yank-pointer (arg)
10309 "Rotate the yanking point in the kill ring.
10310 With argument, rotate that many kills forward (or backward, if negative)."
10312 (current-kill arg))
10314 (defun current-kill (n &optional do-not-move)
10315 "Rotate the yanking point by N places, and then return that kill.
10316 If N is zero, `interprogram-paste-function' is set, and calling it
10317 returns a string, then that string is added to the front of the
10318 kill ring and returned as the latest kill.
10319 If optional arg DO-NOT-MOVE is non-nil, then don't actually move the
10320 yanking point; just return the Nth kill forward."
10321 (let ((interprogram-paste (and (= n 0)
10322 interprogram-paste-function
10323 (funcall interprogram-paste-function))))
10324 (if interprogram-paste
10326 ;; Disable the interprogram cut function when we add the new
10327 ;; text to the kill ring, so Emacs doesn't try to own the
10328 ;; selection, with identical text.
10329 (let ((interprogram-cut-function nil))
10330 (kill-new interprogram-paste))
10331 interprogram-paste)
10332 (or kill-ring (error "Kill ring is empty"))
10333 (let ((ARGth-kill-element
10334 (nthcdr (mod (- n (length kill-ring-yank-pointer))
10335 (length kill-ring))
10338 (setq kill-ring-yank-pointer ARGth-kill-element))
10339 (car ARGth-kill-element)))))
10344 @node yank nthcdr Exercises, , kill-ring-yank-pointer, Yanking
10345 @section Exercises with @code{yank} and @code{nthcdr}
10349 Using @kbd{C-h v} (@code{describe-variable}), look at the value of
10350 your kill ring. Add several items to your kill ring; look at its
10351 value again. Using @kbd{M-y} (@code{yank-pop)}, move all the way
10352 around the kill ring. How many items were in your kill ring? Find
10353 the value of @code{kill-ring-max}. Was your kill ring full, or could
10354 you have kept more blocks of text within it?
10357 Using @code{nthcdr} and @code{car}, construct a series of expressions
10358 to return the first, second, third, and fourth elements of a list.
10361 @node Loops & Recursion, Regexp Search, Yanking, Top
10362 @comment node-name, next, previous, up
10363 @chapter Loops and Recursion
10364 @cindex Loops and recursion
10365 @cindex Recursion and loops
10366 @cindex Repetition (loops)
10368 Emacs Lisp has two primary ways to cause an expression, or a series of
10369 expressions, to be evaluated repeatedly: one uses a @code{while}
10370 loop, and the other uses @dfn{recursion}.
10372 Repetition can be very valuable. For example, to move forward four
10373 sentences, you need only write a program that will move forward one
10374 sentence and then repeat the process four times. Since a computer does
10375 not get bored or tired, such repetitive action does not have the
10376 deleterious effects that excessive or the wrong kinds of repetition can
10379 People mostly write Emacs Lisp functions using @code{while} loops and
10380 their kin; but you can use recursion, which provides a very powerful
10381 way to think about and then to solve problems@footnote{You can write
10382 recursive functions to be frugal or wasteful of mental or computer
10383 resources; as it happens, methods that people find easy---that are
10384 frugal of `mental resources'---sometimes use considerable computer
10385 resources. Emacs was designed to run on machines that we now consider
10386 limited and its default settings are conservative. You may want to
10387 increase the values of @code{max-specpdl-size} and
10388 @code{max-lisp-eval-depth}. In my @file{.emacs} file, I set them to
10389 15 and 30 times their default value.}.
10392 * while:: Causing a stretch of code to repeat.
10394 * Recursion:: Causing a function to call itself.
10395 * Looping exercise::
10398 @node while, dolist dotimes, Loops & Recursion, Loops & Recursion
10399 @comment node-name, next, previous, up
10400 @section @code{while}
10404 The @code{while} special form tests whether the value returned by
10405 evaluating its first argument is true or false. This is similar to what
10406 the Lisp interpreter does with an @code{if}; what the interpreter does
10407 next, however, is different.
10409 In a @code{while} expression, if the value returned by evaluating the
10410 first argument is false, the Lisp interpreter skips the rest of the
10411 expression (the @dfn{body} of the expression) and does not evaluate it.
10412 However, if the value is true, the Lisp interpreter evaluates the body
10413 of the expression and then again tests whether the first argument to
10414 @code{while} is true or false. If the value returned by evaluating the
10415 first argument is again true, the Lisp interpreter again evaluates the
10416 body of the expression.
10419 The template for a @code{while} expression looks like this:
10423 (while @var{true-or-false-test}
10429 * Looping with while:: Repeat so long as test returns true.
10430 * Loop Example:: A @code{while} loop that uses a list.
10431 * print-elements-of-list:: Uses @code{while}, @code{car}, @code{cdr}.
10432 * Incrementing Loop:: A loop with an incrementing counter.
10433 * Incrementing Loop Details::
10434 * Decrementing Loop:: A loop with a decrementing counter.
10437 @node Looping with while, Loop Example, while, while
10439 @unnumberedsubsec Looping with @code{while}
10442 So long as the true-or-false-test of the @code{while} expression
10443 returns a true value when it is evaluated, the body is repeatedly
10444 evaluated. This process is called a loop since the Lisp interpreter
10445 repeats the same thing again and again, like an airplane doing a loop.
10446 When the result of evaluating the true-or-false-test is false, the
10447 Lisp interpreter does not evaluate the rest of the @code{while}
10448 expression and `exits the loop'.
10450 Clearly, if the value returned by evaluating the first argument to
10451 @code{while} is always true, the body following will be evaluated
10452 again and again @dots{} and again @dots{} forever. Conversely, if the
10453 value returned is never true, the expressions in the body will never
10454 be evaluated. The craft of writing a @code{while} loop consists of
10455 choosing a mechanism such that the true-or-false-test returns true
10456 just the number of times that you want the subsequent expressions to
10457 be evaluated, and then have the test return false.
10459 The value returned by evaluating a @code{while} is the value of the
10460 true-or-false-test. An interesting consequence of this is that a
10461 @code{while} loop that evaluates without error will return @code{nil}
10462 or false regardless of whether it has looped 1 or 100 times or none at
10463 all. A @code{while} expression that evaluates successfully never
10464 returns a true value! What this means is that @code{while} is always
10465 evaluated for its side effects, which is to say, the consequences of
10466 evaluating the expressions within the body of the @code{while} loop.
10467 This makes sense. It is not the mere act of looping that is desired,
10468 but the consequences of what happens when the expressions in the loop
10469 are repeatedly evaluated.
10471 @node Loop Example, print-elements-of-list, Looping with while, while
10472 @comment node-name, next, previous, up
10473 @subsection A @code{while} Loop and a List
10475 A common way to control a @code{while} loop is to test whether a list
10476 has any elements. If it does, the loop is repeated; but if it does not,
10477 the repetition is ended. Since this is an important technique, we will
10478 create a short example to illustrate it.
10480 A simple way to test whether a list has elements is to evaluate the
10481 list: if it has no elements, it is an empty list and will return the
10482 empty list, @code{()}, which is a synonym for @code{nil} or false. On
10483 the other hand, a list with elements will return those elements when it
10484 is evaluated. Since Emacs Lisp considers as true any value that is not
10485 @code{nil}, a list that returns elements will test true in a
10489 For example, you can set the variable @code{empty-list} to @code{nil} by
10490 evaluating the following @code{setq} expression:
10493 (setq empty-list ())
10497 After evaluating the @code{setq} expression, you can evaluate the
10498 variable @code{empty-list} in the usual way, by placing the cursor after
10499 the symbol and typing @kbd{C-x C-e}; @code{nil} will appear in your
10506 On the other hand, if you set a variable to be a list with elements, the
10507 list will appear when you evaluate the variable, as you can see by
10508 evaluating the following two expressions:
10512 (setq animals '(gazelle giraffe lion tiger))
10518 Thus, to create a @code{while} loop that tests whether there are any
10519 items in the list @code{animals}, the first part of the loop will be
10530 When the @code{while} tests its first argument, the variable
10531 @code{animals} is evaluated. It returns a list. So long as the list
10532 has elements, the @code{while} considers the results of the test to be
10533 true; but when the list is empty, it considers the results of the test
10536 To prevent the @code{while} loop from running forever, some mechanism
10537 needs to be provided to empty the list eventually. An oft-used
10538 technique is to have one of the subsequent forms in the @code{while}
10539 expression set the value of the list to be the @sc{cdr} of the list.
10540 Each time the @code{cdr} function is evaluated, the list will be made
10541 shorter, until eventually only the empty list will be left. At this
10542 point, the test of the @code{while} loop will return false, and the
10543 arguments to the @code{while} will no longer be evaluated.
10545 For example, the list of animals bound to the variable @code{animals}
10546 can be set to be the @sc{cdr} of the original list with the
10547 following expression:
10550 (setq animals (cdr animals))
10554 If you have evaluated the previous expressions and then evaluate this
10555 expression, you will see @code{(giraffe lion tiger)} appear in the echo
10556 area. If you evaluate the expression again, @code{(lion tiger)} will
10557 appear in the echo area. If you evaluate it again and yet again,
10558 @code{(tiger)} appears and then the empty list, shown by @code{nil}.
10560 A template for a @code{while} loop that uses the @code{cdr} function
10561 repeatedly to cause the true-or-false-test eventually to test false
10566 (while @var{test-whether-list-is-empty}
10568 @var{set-list-to-cdr-of-list})
10572 This test and use of @code{cdr} can be put together in a function that
10573 goes through a list and prints each element of the list on a line of its
10576 @node print-elements-of-list, Incrementing Loop, Loop Example, while
10577 @subsection An Example: @code{print-elements-of-list}
10578 @findex print-elements-of-list
10580 The @code{print-elements-of-list} function illustrates a @code{while}
10583 @cindex @file{*scratch*} buffer
10584 The function requires several lines for its output. If you are
10585 reading this in a recent instance of GNU Emacs,
10586 @c GNU Emacs 21, GNU Emacs 22, or a later version,
10587 you can evaluate the following expression inside of Info, as usual.
10589 If you are using an earlier version of Emacs, you need to copy the
10590 necessary expressions to your @file{*scratch*} buffer and evaluate
10591 them there. This is because the echo area had only one line in the
10594 You can copy the expressions by marking the beginning of the region
10595 with @kbd{C-@key{SPC}} (@code{set-mark-command}), moving the cursor to
10596 the end of the region and then copying the region using @kbd{M-w}
10597 (@code{kill-ring-save}, which calls @code{copy-region-as-kill} and
10598 then provides visual feedback). In the @file{*scratch*}
10599 buffer, you can yank the expressions back by typing @kbd{C-y}
10602 After you have copied the expressions to the @file{*scratch*} buffer,
10603 evaluate each expression in turn. Be sure to evaluate the last
10604 expression, @code{(print-elements-of-list animals)}, by typing
10605 @kbd{C-u C-x C-e}, that is, by giving an argument to
10606 @code{eval-last-sexp}. This will cause the result of the evaluation
10607 to be printed in the @file{*scratch*} buffer instead of being printed
10608 in the echo area. (Otherwise you will see something like this in your
10609 echo area: @code{^Jgazelle^J^Jgiraffe^J^Jlion^J^Jtiger^Jnil}, in which
10610 each @samp{^J} stands for a `newline'.)
10613 In a recent instance of GNU Emacs, you can evaluate these expressions
10614 directly in the Info buffer, and the echo area will grow to show the
10619 (setq animals '(gazelle giraffe lion tiger))
10621 (defun print-elements-of-list (list)
10622 "Print each element of LIST on a line of its own."
10625 (setq list (cdr list))))
10627 (print-elements-of-list animals)
10633 When you evaluate the three expressions in sequence, you will see
10649 Each element of the list is printed on a line of its own (that is what
10650 the function @code{print} does) and then the value returned by the
10651 function is printed. Since the last expression in the function is the
10652 @code{while} loop, and since @code{while} loops always return
10653 @code{nil}, a @code{nil} is printed after the last element of the list.
10655 @node Incrementing Loop, Incrementing Loop Details, print-elements-of-list, while
10656 @comment node-name, next, previous, up
10657 @subsection A Loop with an Incrementing Counter
10659 A loop is not useful unless it stops when it ought. Besides
10660 controlling a loop with a list, a common way of stopping a loop is to
10661 write the first argument as a test that returns false when the correct
10662 number of repetitions are complete. This means that the loop must
10663 have a counter---an expression that counts how many times the loop
10666 @node Incrementing Loop Details, Decrementing Loop, Incrementing Loop, while
10668 @unnumberedsubsec Details of an Incrementing Loop
10671 The test for a loop with an incrementing counter can be an expression
10672 such as @code{(< count desired-number)} which returns @code{t} for
10673 true if the value of @code{count} is less than the
10674 @code{desired-number} of repetitions and @code{nil} for false if the
10675 value of @code{count} is equal to or is greater than the
10676 @code{desired-number}. The expression that increments the count can
10677 be a simple @code{setq} such as @code{(setq count (1+ count))}, where
10678 @code{1+} is a built-in function in Emacs Lisp that adds 1 to its
10679 argument. (The expression @w{@code{(1+ count)}} has the same result
10680 as @w{@code{(+ count 1)}}, but is easier for a human to read.)
10683 The template for a @code{while} loop controlled by an incrementing
10684 counter looks like this:
10688 @var{set-count-to-initial-value}
10689 (while (< count desired-number) ; @r{true-or-false-test}
10691 (setq count (1+ count))) ; @r{incrementer}
10696 Note that you need to set the initial value of @code{count}; usually it
10700 * Incrementing Example:: Counting pebbles in a triangle.
10701 * Inc Example parts:: The parts of the function definition.
10702 * Inc Example altogether:: Putting the function definition together.
10705 @node Incrementing Example, Inc Example parts, Incrementing Loop Details, Incrementing Loop Details
10706 @unnumberedsubsubsec Example with incrementing counter
10708 Suppose you are playing on the beach and decide to make a triangle of
10709 pebbles, putting one pebble in the first row, two in the second row,
10710 three in the third row and so on, like this:
10728 @bullet{} @bullet{}
10729 @bullet{} @bullet{} @bullet{}
10730 @bullet{} @bullet{} @bullet{} @bullet{}
10737 (About 2500 years ago, Pythagoras and others developed the beginnings of
10738 number theory by considering questions such as this.)
10740 Suppose you want to know how many pebbles you will need to make a
10741 triangle with 7 rows?
10743 Clearly, what you need to do is add up the numbers from 1 to 7. There
10744 are two ways to do this; start with the smallest number, one, and add up
10745 the list in sequence, 1, 2, 3, 4 and so on; or start with the largest
10746 number and add the list going down: 7, 6, 5, 4 and so on. Because both
10747 mechanisms illustrate common ways of writing @code{while} loops, we will
10748 create two examples, one counting up and the other counting down. In
10749 this first example, we will start with 1 and add 2, 3, 4 and so on.
10751 If you are just adding up a short list of numbers, the easiest way to do
10752 it is to add up all the numbers at once. However, if you do not know
10753 ahead of time how many numbers your list will have, or if you want to be
10754 prepared for a very long list, then you need to design your addition so
10755 that what you do is repeat a simple process many times instead of doing
10756 a more complex process once.
10758 For example, instead of adding up all the pebbles all at once, what you
10759 can do is add the number of pebbles in the first row, 1, to the number
10760 in the second row, 2, and then add the total of those two rows to the
10761 third row, 3. Then you can add the number in the fourth row, 4, to the
10762 total of the first three rows; and so on.
10764 The critical characteristic of the process is that each repetitive
10765 action is simple. In this case, at each step we add only two numbers,
10766 the number of pebbles in the row and the total already found. This
10767 process of adding two numbers is repeated again and again until the last
10768 row has been added to the total of all the preceding rows. In a more
10769 complex loop the repetitive action might not be so simple, but it will
10770 be simpler than doing everything all at once.
10772 @node Inc Example parts, Inc Example altogether, Incrementing Example, Incrementing Loop Details
10773 @unnumberedsubsubsec The parts of the function definition
10775 The preceding analysis gives us the bones of our function definition:
10776 first, we will need a variable that we can call @code{total} that will
10777 be the total number of pebbles. This will be the value returned by
10780 Second, we know that the function will require an argument: this
10781 argument will be the total number of rows in the triangle. It can be
10782 called @code{number-of-rows}.
10784 Finally, we need a variable to use as a counter. We could call this
10785 variable @code{counter}, but a better name is @code{row-number}. That
10786 is because what the counter does in this function is count rows, and a
10787 program should be written to be as understandable as possible.
10789 When the Lisp interpreter first starts evaluating the expressions in the
10790 function, the value of @code{total} should be set to zero, since we have
10791 not added anything to it. Then the function should add the number of
10792 pebbles in the first row to the total, and then add the number of
10793 pebbles in the second to the total, and then add the number of
10794 pebbles in the third row to the total, and so on, until there are no
10795 more rows left to add.
10797 Both @code{total} and @code{row-number} are used only inside the
10798 function, so they can be declared as local variables with @code{let}
10799 and given initial values. Clearly, the initial value for @code{total}
10800 should be 0. The initial value of @code{row-number} should be 1,
10801 since we start with the first row. This means that the @code{let}
10802 statement will look like this:
10812 After the internal variables are declared and bound to their initial
10813 values, we can begin the @code{while} loop. The expression that serves
10814 as the test should return a value of @code{t} for true so long as the
10815 @code{row-number} is less than or equal to the @code{number-of-rows}.
10816 (If the expression tests true only so long as the row number is less
10817 than the number of rows in the triangle, the last row will never be
10818 added to the total; hence the row number has to be either less than or
10819 equal to the number of rows.)
10822 @findex <= @r{(less than or equal)}
10823 Lisp provides the @code{<=} function that returns true if the value of
10824 its first argument is less than or equal to the value of its second
10825 argument and false otherwise. So the expression that the @code{while}
10826 will evaluate as its test should look like this:
10829 (<= row-number number-of-rows)
10832 The total number of pebbles can be found by repeatedly adding the number
10833 of pebbles in a row to the total already found. Since the number of
10834 pebbles in the row is equal to the row number, the total can be found by
10835 adding the row number to the total. (Clearly, in a more complex
10836 situation, the number of pebbles in the row might be related to the row
10837 number in a more complicated way; if this were the case, the row number
10838 would be replaced by the appropriate expression.)
10841 (setq total (+ total row-number))
10845 What this does is set the new value of @code{total} to be equal to the
10846 sum of adding the number of pebbles in the row to the previous total.
10848 After setting the value of @code{total}, the conditions need to be
10849 established for the next repetition of the loop, if there is one. This
10850 is done by incrementing the value of the @code{row-number} variable,
10851 which serves as a counter. After the @code{row-number} variable has
10852 been incremented, the true-or-false-test at the beginning of the
10853 @code{while} loop tests whether its value is still less than or equal to
10854 the value of the @code{number-of-rows} and if it is, adds the new value
10855 of the @code{row-number} variable to the @code{total} of the previous
10856 repetition of the loop.
10859 The built-in Emacs Lisp function @code{1+} adds 1 to a number, so the
10860 @code{row-number} variable can be incremented with this expression:
10863 (setq row-number (1+ row-number))
10866 @node Inc Example altogether, , Inc Example parts, Incrementing Loop Details
10867 @unnumberedsubsubsec Putting the function definition together
10869 We have created the parts for the function definition; now we need to
10873 First, the contents of the @code{while} expression:
10877 (while (<= row-number number-of-rows) ; @r{true-or-false-test}
10878 (setq total (+ total row-number))
10879 (setq row-number (1+ row-number))) ; @r{incrementer}
10883 Along with the @code{let} expression varlist, this very nearly
10884 completes the body of the function definition. However, it requires
10885 one final element, the need for which is somewhat subtle.
10887 The final touch is to place the variable @code{total} on a line by
10888 itself after the @code{while} expression. Otherwise, the value returned
10889 by the whole function is the value of the last expression that is
10890 evaluated in the body of the @code{let}, and this is the value
10891 returned by the @code{while}, which is always @code{nil}.
10893 This may not be evident at first sight. It almost looks as if the
10894 incrementing expression is the last expression of the whole function.
10895 But that expression is part of the body of the @code{while}; it is the
10896 last element of the list that starts with the symbol @code{while}.
10897 Moreover, the whole of the @code{while} loop is a list within the body
10901 In outline, the function will look like this:
10905 (defun @var{name-of-function} (@var{argument-list})
10906 "@var{documentation}@dots{}"
10907 (let (@var{varlist})
10908 (while (@var{true-or-false-test})
10909 @var{body-of-while}@dots{} )
10910 @dots{} )) ; @r{Need final expression here.}
10914 The result of evaluating the @code{let} is what is going to be returned
10915 by the @code{defun} since the @code{let} is not embedded within any
10916 containing list, except for the @code{defun} as a whole. However, if
10917 the @code{while} is the last element of the @code{let} expression, the
10918 function will always return @code{nil}. This is not what we want!
10919 Instead, what we want is the value of the variable @code{total}. This
10920 is returned by simply placing the symbol as the last element of the list
10921 starting with @code{let}. It gets evaluated after the preceding
10922 elements of the list are evaluated, which means it gets evaluated after
10923 it has been assigned the correct value for the total.
10925 It may be easier to see this by printing the list starting with
10926 @code{let} all on one line. This format makes it evident that the
10927 @var{varlist} and @code{while} expressions are the second and third
10928 elements of the list starting with @code{let}, and the @code{total} is
10933 (let (@var{varlist}) (while (@var{true-or-false-test}) @var{body-of-while}@dots{} ) total)
10938 Putting everything together, the @code{triangle} function definition
10943 (defun triangle (number-of-rows) ; @r{Version with}
10944 ; @r{ incrementing counter.}
10945 "Add up the number of pebbles in a triangle.
10946 The first row has one pebble, the second row two pebbles,
10947 the third row three pebbles, and so on.
10948 The argument is NUMBER-OF-ROWS."
10953 (while (<= row-number number-of-rows)
10954 (setq total (+ total row-number))
10955 (setq row-number (1+ row-number)))
10961 After you have installed @code{triangle} by evaluating the function, you
10962 can try it out. Here are two examples:
10973 The sum of the first four numbers is 10 and the sum of the first seven
10976 @node Decrementing Loop, , Incrementing Loop Details, while
10977 @comment node-name, next, previous, up
10978 @subsection Loop with a Decrementing Counter
10980 Another common way to write a @code{while} loop is to write the test
10981 so that it determines whether a counter is greater than zero. So long
10982 as the counter is greater than zero, the loop is repeated. But when
10983 the counter is equal to or less than zero, the loop is stopped. For
10984 this to work, the counter has to start out greater than zero and then
10985 be made smaller and smaller by a form that is evaluated
10988 The test will be an expression such as @code{(> counter 0)} which
10989 returns @code{t} for true if the value of @code{counter} is greater
10990 than zero, and @code{nil} for false if the value of @code{counter} is
10991 equal to or less than zero. The expression that makes the number
10992 smaller and smaller can be a simple @code{setq} such as @code{(setq
10993 counter (1- counter))}, where @code{1-} is a built-in function in
10994 Emacs Lisp that subtracts 1 from its argument.
10997 The template for a decrementing @code{while} loop looks like this:
11001 (while (> counter 0) ; @r{true-or-false-test}
11003 (setq counter (1- counter))) ; @r{decrementer}
11008 * Decrementing Example:: More pebbles on the beach.
11009 * Dec Example parts:: The parts of the function definition.
11010 * Dec Example altogether:: Putting the function definition together.
11013 @node Decrementing Example, Dec Example parts, Decrementing Loop, Decrementing Loop
11014 @unnumberedsubsubsec Example with decrementing counter
11016 To illustrate a loop with a decrementing counter, we will rewrite the
11017 @code{triangle} function so the counter decreases to zero.
11019 This is the reverse of the earlier version of the function. In this
11020 case, to find out how many pebbles are needed to make a triangle with
11021 3 rows, add the number of pebbles in the third row, 3, to the number
11022 in the preceding row, 2, and then add the total of those two rows to
11023 the row that precedes them, which is 1.
11025 Likewise, to find the number of pebbles in a triangle with 7 rows, add
11026 the number of pebbles in the seventh row, 7, to the number in the
11027 preceding row, which is 6, and then add the total of those two rows to
11028 the row that precedes them, which is 5, and so on. As in the previous
11029 example, each addition only involves adding two numbers, the total of
11030 the rows already added up and the number of pebbles in the row that is
11031 being added to the total. This process of adding two numbers is
11032 repeated again and again until there are no more pebbles to add.
11034 We know how many pebbles to start with: the number of pebbles in the
11035 last row is equal to the number of rows. If the triangle has seven
11036 rows, the number of pebbles in the last row is 7. Likewise, we know how
11037 many pebbles are in the preceding row: it is one less than the number in
11040 @node Dec Example parts, Dec Example altogether, Decrementing Example, Decrementing Loop
11041 @unnumberedsubsubsec The parts of the function definition
11043 We start with three variables: the total number of rows in the
11044 triangle; the number of pebbles in a row; and the total number of
11045 pebbles, which is what we want to calculate. These variables can be
11046 named @code{number-of-rows}, @code{number-of-pebbles-in-row}, and
11047 @code{total}, respectively.
11049 Both @code{total} and @code{number-of-pebbles-in-row} are used only
11050 inside the function and are declared with @code{let}. The initial
11051 value of @code{total} should, of course, be zero. However, the
11052 initial value of @code{number-of-pebbles-in-row} should be equal to
11053 the number of rows in the triangle, since the addition will start with
11057 This means that the beginning of the @code{let} expression will look
11063 (number-of-pebbles-in-row number-of-rows))
11068 The total number of pebbles can be found by repeatedly adding the number
11069 of pebbles in a row to the total already found, that is, by repeatedly
11070 evaluating the following expression:
11073 (setq total (+ total number-of-pebbles-in-row))
11077 After the @code{number-of-pebbles-in-row} is added to the @code{total},
11078 the @code{number-of-pebbles-in-row} should be decremented by one, since
11079 the next time the loop repeats, the preceding row will be
11080 added to the total.
11082 The number of pebbles in a preceding row is one less than the number of
11083 pebbles in a row, so the built-in Emacs Lisp function @code{1-} can be
11084 used to compute the number of pebbles in the preceding row. This can be
11085 done with the following expression:
11089 (setq number-of-pebbles-in-row
11090 (1- number-of-pebbles-in-row))
11094 Finally, we know that the @code{while} loop should stop making repeated
11095 additions when there are no pebbles in a row. So the test for
11096 the @code{while} loop is simply:
11099 (while (> number-of-pebbles-in-row 0)
11102 @node Dec Example altogether, , Dec Example parts, Decrementing Loop
11103 @unnumberedsubsubsec Putting the function definition together
11105 We can put these expressions together to create a function definition
11106 that works. However, on examination, we find that one of the local
11107 variables is unneeded!
11110 The function definition looks like this:
11114 ;;; @r{First subtractive version.}
11115 (defun triangle (number-of-rows)
11116 "Add up the number of pebbles in a triangle."
11118 (number-of-pebbles-in-row number-of-rows))
11119 (while (> number-of-pebbles-in-row 0)
11120 (setq total (+ total number-of-pebbles-in-row))
11121 (setq number-of-pebbles-in-row
11122 (1- number-of-pebbles-in-row)))
11127 As written, this function works.
11129 However, we do not need @code{number-of-pebbles-in-row}.
11131 @cindex Argument as local variable
11132 When the @code{triangle} function is evaluated, the symbol
11133 @code{number-of-rows} will be bound to a number, giving it an initial
11134 value. That number can be changed in the body of the function as if
11135 it were a local variable, without any fear that such a change will
11136 effect the value of the variable outside of the function. This is a
11137 very useful characteristic of Lisp; it means that the variable
11138 @code{number-of-rows} can be used anywhere in the function where
11139 @code{number-of-pebbles-in-row} is used.
11142 Here is a second version of the function written a bit more cleanly:
11146 (defun triangle (number) ; @r{Second version.}
11147 "Return sum of numbers 1 through NUMBER inclusive."
11149 (while (> number 0)
11150 (setq total (+ total number))
11151 (setq number (1- number)))
11156 In brief, a properly written @code{while} loop will consist of three parts:
11160 A test that will return false after the loop has repeated itself the
11161 correct number of times.
11164 An expression the evaluation of which will return the value desired
11165 after being repeatedly evaluated.
11168 An expression to change the value passed to the true-or-false-test so
11169 that the test returns false after the loop has repeated itself the right
11173 @node dolist dotimes, Recursion, while, Loops & Recursion
11174 @comment node-name, next, previous, up
11175 @section Save your time: @code{dolist} and @code{dotimes}
11177 In addition to @code{while}, both @code{dolist} and @code{dotimes}
11178 provide for looping. Sometimes these are quicker to write than the
11179 equivalent @code{while} loop. Both are Lisp macros. (@xref{Macros, ,
11180 Macros, elisp, The GNU Emacs Lisp Reference Manual}. )
11182 @code{dolist} works like a @code{while} loop that `@sc{cdr}s down a
11183 list': @code{dolist} automatically shortens the list each time it
11184 loops---takes the @sc{cdr} of the list---and binds the @sc{car} of
11185 each shorter version of the list to the first of its arguments.
11187 @code{dotimes} loops a specific number of times: you specify the number.
11194 @node dolist, dotimes, dolist dotimes, dolist dotimes
11195 @unnumberedsubsubsec The @code{dolist} Macro
11198 Suppose, for example, you want to reverse a list, so that
11199 ``first'' ``second'' ``third'' becomes ``third'' ``second'' ``first''.
11202 In practice, you would use the @code{reverse} function, like this:
11206 (setq animals '(gazelle giraffe lion tiger))
11214 Here is how you could reverse the list using a @code{while} loop:
11218 (setq animals '(gazelle giraffe lion tiger))
11220 (defun reverse-list-with-while (list)
11221 "Using while, reverse the order of LIST."
11222 (let (value) ; make sure list starts empty
11224 (setq value (cons (car list) value))
11225 (setq list (cdr list)))
11228 (reverse-list-with-while animals)
11234 And here is how you could use the @code{dolist} macro:
11238 (setq animals '(gazelle giraffe lion tiger))
11240 (defun reverse-list-with-dolist (list)
11241 "Using dolist, reverse the order of LIST."
11242 (let (value) ; make sure list starts empty
11243 (dolist (element list value)
11244 (setq value (cons element value)))))
11246 (reverse-list-with-dolist animals)
11252 In Info, you can place your cursor after the closing parenthesis of
11253 each expression and type @kbd{C-x C-e}; in each case, you should see
11256 (tiger lion giraffe gazelle)
11262 For this example, the existing @code{reverse} function is obviously best.
11263 The @code{while} loop is just like our first example (@pxref{Loop
11264 Example, , A @code{while} Loop and a List}). The @code{while} first
11265 checks whether the list has elements; if so, it constructs a new list
11266 by adding the first element of the list to the existing list (which in
11267 the first iteration of the loop is @code{nil}). Since the second
11268 element is prepended in front of the first element, and the third
11269 element is prepended in front of the second element, the list is reversed.
11271 In the expression using a @code{while} loop,
11272 the @w{@code{(setq list (cdr list))}}
11273 expression shortens the list, so the @code{while} loop eventually
11274 stops. In addition, it provides the @code{cons} expression with a new
11275 first element by creating a new and shorter list at each repetition of
11278 The @code{dolist} expression does very much the same as the
11279 @code{while} expression, except that the @code{dolist} macro does some
11280 of the work you have to do when writing a @code{while} expression.
11282 Like a @code{while} loop, a @code{dolist} loops. What is different is
11283 that it automatically shortens the list each time it loops --- it
11284 `@sc{cdr}s down the list' on its own --- and it automatically binds
11285 the @sc{car} of each shorter version of the list to the first of its
11288 In the example, the @sc{car} of each shorter version of the list is
11289 referred to using the symbol @samp{element}, the list itself is called
11290 @samp{list}, and the value returned is called @samp{value}. The
11291 remainder of the @code{dolist} expression is the body.
11293 The @code{dolist} expression binds the @sc{car} of each shorter
11294 version of the list to @code{element} and then evaluates the body of
11295 the expression; and repeats the loop. The result is returned in
11298 @node dotimes, , dolist, dolist dotimes
11299 @unnumberedsubsubsec The @code{dotimes} Macro
11302 The @code{dotimes} macro is similar to @code{dolist}, except that it
11303 loops a specific number of times.
11305 The first argument to @code{dotimes} is assigned the numbers 0, 1, 2
11306 and so forth each time around the loop, and the value of the third
11307 argument is returned. You need to provide the value of the second
11308 argument, which is how many times the macro loops.
11311 For example, the following binds the numbers from 0 up to, but not
11312 including, the number 3 to the first argument, @var{number}, and then
11313 constructs a list of the three numbers. (The first number is 0, the
11314 second number is 1, and the third number is 2; this makes a total of
11315 three numbers in all, starting with zero as the first number.)
11319 (let (value) ; otherwise a value is a void variable
11320 (dotimes (number 3 value)
11321 (setq value (cons number value))))
11328 @code{dotimes} returns @code{value}, so the way to use
11329 @code{dotimes} is to operate on some expression @var{number} number of
11330 times and then return the result, either as a list or an atom.
11333 Here is an example of a @code{defun} that uses @code{dotimes} to add
11334 up the number of pebbles in a triangle.
11338 (defun triangle-using-dotimes (number-of-rows)
11339 "Using dotimes, add up the number of pebbles in a triangle."
11340 (let ((total 0)) ; otherwise a total is a void variable
11341 (dotimes (number number-of-rows total)
11342 (setq total (+ total (1+ number))))))
11344 (triangle-using-dotimes 4)
11348 @node Recursion, Looping exercise, dolist dotimes, Loops & Recursion
11349 @comment node-name, next, previous, up
11353 A recursive function contains code that tells the Lisp interpreter to
11354 call a program that runs exactly like itself, but with slightly
11355 different arguments. The code runs exactly the same because it has
11356 the same name. However, even though the program has the same name, it
11357 is not the same entity. It is different. In the jargon, it is a
11358 different `instance'.
11360 Eventually, if the program is written correctly, the `slightly
11361 different arguments' will become sufficiently different from the first
11362 arguments that the final instance will stop.
11365 * Building Robots:: Same model, different serial number ...
11366 * Recursive Definition Parts:: Walk until you stop ...
11367 * Recursion with list:: Using a list as the test whether to recurse.
11368 * Recursive triangle function::
11369 * Recursion with cond::
11370 * Recursive Patterns:: Often used templates.
11371 * No Deferment:: Don't store up work ...
11372 * No deferment solution::
11375 @node Building Robots, Recursive Definition Parts, Recursion, Recursion
11376 @comment node-name, next, previous, up
11377 @subsection Building Robots: Extending the Metaphor
11378 @cindex Building robots
11379 @cindex Robots, building
11381 It is sometimes helpful to think of a running program as a robot that
11382 does a job. In doing its job, a recursive function calls on a second
11383 robot to help it. The second robot is identical to the first in every
11384 way, except that the second robot helps the first and has been
11385 passed different arguments than the first.
11387 In a recursive function, the second robot may call a third; and the
11388 third may call a fourth, and so on. Each of these is a different
11389 entity; but all are clones.
11391 Since each robot has slightly different instructions---the arguments
11392 will differ from one robot to the next---the last robot should know
11395 Let's expand on the metaphor in which a computer program is a robot.
11397 A function definition provides the blueprints for a robot. When you
11398 install a function definition, that is, when you evaluate a
11399 @code{defun} special form, you install the necessary equipment to
11400 build robots. It is as if you were in a factory, setting up an
11401 assembly line. Robots with the same name are built according to the
11402 same blueprints. So they have, as it were, the same `model number',
11403 but a different `serial number'.
11405 We often say that a recursive function `calls itself'. What we mean
11406 is that the instructions in a recursive function cause the Lisp
11407 interpreter to run a different function that has the same name and
11408 does the same job as the first, but with different arguments.
11410 It is important that the arguments differ from one instance to the
11411 next; otherwise, the process will never stop.
11413 @node Recursive Definition Parts, Recursion with list, Building Robots, Recursion
11414 @comment node-name, next, previous, up
11415 @subsection The Parts of a Recursive Definition
11416 @cindex Parts of a Recursive Definition
11417 @cindex Recursive Definition Parts
11419 A recursive function typically contains a conditional expression which
11424 A true-or-false-test that determines whether the function is called
11425 again, here called the @dfn{do-again-test}.
11428 The name of the function. When this name is called, a new instance of
11429 the function---a new robot, as it were---is created and told what to do.
11432 An expression that returns a different value each time the function is
11433 called, here called the @dfn{next-step-expression}. Consequently, the
11434 argument (or arguments) passed to the new instance of the function
11435 will be different from that passed to the previous instance. This
11436 causes the conditional expression, the @dfn{do-again-test}, to test
11437 false after the correct number of repetitions.
11440 Recursive functions can be much simpler than any other kind of
11441 function. Indeed, when people first start to use them, they often look
11442 so mysteriously simple as to be incomprehensible. Like riding a
11443 bicycle, reading a recursive function definition takes a certain knack
11444 which is hard at first but then seems simple.
11447 There are several different common recursive patterns. A very simple
11448 pattern looks like this:
11452 (defun @var{name-of-recursive-function} (@var{argument-list})
11453 "@var{documentation}@dots{}"
11454 (if @var{do-again-test}
11456 (@var{name-of-recursive-function}
11457 @var{next-step-expression})))
11461 Each time a recursive function is evaluated, a new instance of it is
11462 created and told what to do. The arguments tell the instance what to do.
11464 An argument is bound to the value of the next-step-expression. Each
11465 instance runs with a different value of the next-step-expression.
11467 The value in the next-step-expression is used in the do-again-test.
11469 The value returned by the next-step-expression is passed to the new
11470 instance of the function, which evaluates it (or some
11471 transmogrification of it) to determine whether to continue or stop.
11472 The next-step-expression is designed so that the do-again-test returns
11473 false when the function should no longer be repeated.
11475 The do-again-test is sometimes called the @dfn{stop condition},
11476 since it stops the repetitions when it tests false.
11478 @node Recursion with list, Recursive triangle function, Recursive Definition Parts, Recursion
11479 @comment node-name, next, previous, up
11480 @subsection Recursion with a List
11482 The example of a @code{while} loop that printed the elements of a list
11483 of numbers can be written recursively. Here is the code, including
11484 an expression to set the value of the variable @code{animals} to a list.
11486 If you are using GNU Emacs 20 or before, this example must be copied
11487 to the @file{*scratch*} buffer and each expression must be evaluated
11488 there. Use @kbd{C-u C-x C-e} to evaluate the
11489 @code{(print-elements-recursively animals)} expression so that the
11490 results are printed in the buffer; otherwise the Lisp interpreter will
11491 try to squeeze the results into the one line of the echo area.
11493 Also, place your cursor immediately after the last closing parenthesis
11494 of the @code{print-elements-recursively} function, before the comment.
11495 Otherwise, the Lisp interpreter will try to evaluate the comment.
11497 If you are using a more recent version of Emacs, you can evaluate this
11498 expression directly in Info.
11500 @findex print-elements-recursively
11503 (setq animals '(gazelle giraffe lion tiger))
11505 (defun print-elements-recursively (list)
11506 "Print each element of LIST on a line of its own.
11508 (when list ; @r{do-again-test}
11509 (print (car list)) ; @r{body}
11510 (print-elements-recursively ; @r{recursive call}
11511 (cdr list)))) ; @r{next-step-expression}
11513 (print-elements-recursively animals)
11517 The @code{print-elements-recursively} function first tests whether
11518 there is any content in the list; if there is, the function prints the
11519 first element of the list, the @sc{car} of the list. Then the
11520 function `invokes itself', but gives itself as its argument, not the
11521 whole list, but the second and subsequent elements of the list, the
11522 @sc{cdr} of the list.
11524 Put another way, if the list is not empty, the function invokes
11525 another instance of code that is similar to the initial code, but is a
11526 different thread of execution, with different arguments than the first
11529 Put in yet another way, if the list is not empty, the first robot
11530 assemblies a second robot and tells it what to do; the second robot is
11531 a different individual from the first, but is the same model.
11533 When the second evaluation occurs, the @code{when} expression is
11534 evaluated and if true, prints the first element of the list it
11535 receives as its argument (which is the second element of the original
11536 list). Then the function `calls itself' with the @sc{cdr} of the list
11537 it is invoked with, which (the second time around) is the @sc{cdr} of
11538 the @sc{cdr} of the original list.
11540 Note that although we say that the function `calls itself', what we
11541 mean is that the Lisp interpreter assembles and instructs a new
11542 instance of the program. The new instance is a clone of the first,
11543 but is a separate individual.
11545 Each time the function `invokes itself', it invokes itself on a
11546 shorter version of the original list. It creates a new instance that
11547 works on a shorter list.
11549 Eventually, the function invokes itself on an empty list. It creates
11550 a new instance whose argument is @code{nil}. The conditional expression
11551 tests the value of @code{list}. Since the value of @code{list} is
11552 @code{nil}, the @code{when} expression tests false so the then-part is
11553 not evaluated. The function as a whole then returns @code{nil}.
11556 When you evaluate @code{(print-elements-recursively animals)} in the
11557 @file{*scratch*} buffer, you see this result:
11573 @node Recursive triangle function, Recursion with cond, Recursion with list, Recursion
11574 @comment node-name, next, previous, up
11575 @subsection Recursion in Place of a Counter
11576 @findex triangle-recursively
11579 The @code{triangle} function described in a previous section can also
11580 be written recursively. It looks like this:
11584 (defun triangle-recursively (number)
11585 "Return the sum of the numbers 1 through NUMBER inclusive.
11587 (if (= number 1) ; @r{do-again-test}
11589 (+ number ; @r{else-part}
11590 (triangle-recursively ; @r{recursive call}
11591 (1- number))))) ; @r{next-step-expression}
11593 (triangle-recursively 7)
11598 You can install this function by evaluating it and then try it by
11599 evaluating @code{(triangle-recursively 7)}. (Remember to put your
11600 cursor immediately after the last parenthesis of the function
11601 definition, before the comment.) The function evaluates to 28.
11603 To understand how this function works, let's consider what happens in the
11604 various cases when the function is passed 1, 2, 3, or 4 as the value of
11608 * Recursive Example arg of 1 or 2::
11609 * Recursive Example arg of 3 or 4::
11612 @node Recursive Example arg of 1 or 2, Recursive Example arg of 3 or 4, Recursive triangle function, Recursive triangle function
11614 @unnumberedsubsubsec An argument of 1 or 2
11617 First, what happens if the value of the argument is 1?
11619 The function has an @code{if} expression after the documentation
11620 string. It tests whether the value of @code{number} is equal to 1; if
11621 so, Emacs evaluates the then-part of the @code{if} expression, which
11622 returns the number 1 as the value of the function. (A triangle with
11623 one row has one pebble in it.)
11625 Suppose, however, that the value of the argument is 2. In this case,
11626 Emacs evaluates the else-part of the @code{if} expression.
11629 The else-part consists of an addition, the recursive call to
11630 @code{triangle-recursively} and a decrementing action; and it looks like
11634 (+ number (triangle-recursively (1- number)))
11637 When Emacs evaluates this expression, the innermost expression is
11638 evaluated first; then the other parts in sequence. Here are the steps
11642 @item Step 1 @w{ } Evaluate the innermost expression.
11644 The innermost expression is @code{(1- number)} so Emacs decrements the
11645 value of @code{number} from 2 to 1.
11647 @item Step 2 @w{ } Evaluate the @code{triangle-recursively} function.
11649 The Lisp interpreter creates an individual instance of
11650 @code{triangle-recursively}. It does not matter that this function is
11651 contained within itself. Emacs passes the result Step 1 as the
11652 argument used by this instance of the @code{triangle-recursively}
11655 In this case, Emacs evaluates @code{triangle-recursively} with an
11656 argument of 1. This means that this evaluation of
11657 @code{triangle-recursively} returns 1.
11659 @item Step 3 @w{ } Evaluate the value of @code{number}.
11661 The variable @code{number} is the second element of the list that
11662 starts with @code{+}; its value is 2.
11664 @item Step 4 @w{ } Evaluate the @code{+} expression.
11666 The @code{+} expression receives two arguments, the first
11667 from the evaluation of @code{number} (Step 3) and the second from the
11668 evaluation of @code{triangle-recursively} (Step 2).
11670 The result of the addition is the sum of 2 plus 1, and the number 3 is
11671 returned, which is correct. A triangle with two rows has three
11675 @node Recursive Example arg of 3 or 4, , Recursive Example arg of 1 or 2, Recursive triangle function
11676 @unnumberedsubsubsec An argument of 3 or 4
11678 Suppose that @code{triangle-recursively} is called with an argument of
11682 @item Step 1 @w{ } Evaluate the do-again-test.
11684 The @code{if} expression is evaluated first. This is the do-again
11685 test and returns false, so the else-part of the @code{if} expression
11686 is evaluated. (Note that in this example, the do-again-test causes
11687 the function to call itself when it tests false, not when it tests
11690 @item Step 2 @w{ } Evaluate the innermost expression of the else-part.
11692 The innermost expression of the else-part is evaluated, which decrements
11693 3 to 2. This is the next-step-expression.
11695 @item Step 3 @w{ } Evaluate the @code{triangle-recursively} function.
11697 The number 2 is passed to the @code{triangle-recursively} function.
11699 We know what happens when Emacs evaluates @code{triangle-recursively} with
11700 an argument of 2. After going through the sequence of actions described
11701 earlier, it returns a value of 3. So that is what will happen here.
11703 @item Step 4 @w{ } Evaluate the addition.
11705 3 will be passed as an argument to the addition and will be added to the
11706 number with which the function was called, which is 3.
11710 The value returned by the function as a whole will be 6.
11712 Now that we know what will happen when @code{triangle-recursively} is
11713 called with an argument of 3, it is evident what will happen if it is
11714 called with an argument of 4:
11718 In the recursive call, the evaluation of
11721 (triangle-recursively (1- 4))
11726 will return the value of evaluating
11729 (triangle-recursively 3)
11733 which is 6 and this value will be added to 4 by the addition in the
11738 The value returned by the function as a whole will be 10.
11740 Each time @code{triangle-recursively} is evaluated, it evaluates a
11741 version of itself---a different instance of itself---with a smaller
11742 argument, until the argument is small enough so that it does not
11745 Note that this particular design for a recursive function
11746 requires that operations be deferred.
11748 Before @code{(triangle-recursively 7)} can calculate its answer, it
11749 must call @code{(triangle-recursively 6)}; and before
11750 @code{(triangle-recursively 6)} can calculate its answer, it must call
11751 @code{(triangle-recursively 5)}; and so on. That is to say, the
11752 calculation that @code{(triangle-recursively 7)} makes must be
11753 deferred until @code{(triangle-recursively 6)} makes its calculation;
11754 and @code{(triangle-recursively 6)} must defer until
11755 @code{(triangle-recursively 5)} completes; and so on.
11757 If each of these instances of @code{triangle-recursively} are thought
11758 of as different robots, the first robot must wait for the second to
11759 complete its job, which must wait until the third completes, and so
11762 There is a way around this kind of waiting, which we will discuss in
11763 @ref{No Deferment, , Recursion without Deferments}.
11765 @node Recursion with cond, Recursive Patterns, Recursive triangle function, Recursion
11766 @comment node-name, next, previous, up
11767 @subsection Recursion Example Using @code{cond}
11770 The version of @code{triangle-recursively} described earlier is written
11771 with the @code{if} special form. It can also be written using another
11772 special form called @code{cond}. The name of the special form
11773 @code{cond} is an abbreviation of the word @samp{conditional}.
11775 Although the @code{cond} special form is not used as often in the
11776 Emacs Lisp sources as @code{if}, it is used often enough to justify
11780 The template for a @code{cond} expression looks like this:
11790 where the @var{body} is a series of lists.
11793 Written out more fully, the template looks like this:
11798 (@var{first-true-or-false-test} @var{first-consequent})
11799 (@var{second-true-or-false-test} @var{second-consequent})
11800 (@var{third-true-or-false-test} @var{third-consequent})
11805 When the Lisp interpreter evaluates the @code{cond} expression, it
11806 evaluates the first element (the @sc{car} or true-or-false-test) of
11807 the first expression in a series of expressions within the body of the
11810 If the true-or-false-test returns @code{nil} the rest of that
11811 expression, the consequent, is skipped and the true-or-false-test of the
11812 next expression is evaluated. When an expression is found whose
11813 true-or-false-test returns a value that is not @code{nil}, the
11814 consequent of that expression is evaluated. The consequent can be one
11815 or more expressions. If the consequent consists of more than one
11816 expression, the expressions are evaluated in sequence and the value of
11817 the last one is returned. If the expression does not have a consequent,
11818 the value of the true-or-false-test is returned.
11820 If none of the true-or-false-tests test true, the @code{cond} expression
11821 returns @code{nil}.
11824 Written using @code{cond}, the @code{triangle} function looks like this:
11828 (defun triangle-using-cond (number)
11829 (cond ((<= number 0) 0)
11832 (+ number (triangle-using-cond (1- number))))))
11837 In this example, the @code{cond} returns 0 if the number is less than or
11838 equal to 0, it returns 1 if the number is 1 and it evaluates @code{(+
11839 number (triangle-using-cond (1- number)))} if the number is greater than
11842 @node Recursive Patterns, No Deferment, Recursion with cond, Recursion
11843 @comment node-name, next, previous, up
11844 @subsection Recursive Patterns
11845 @cindex Recursive Patterns
11847 Here are three common recursive patterns. Each involves a list.
11848 Recursion does not need to involve lists, but Lisp is designed for lists
11849 and this provides a sense of its primal capabilities.
11857 @node Every, Accumulate, Recursive Patterns, Recursive Patterns
11858 @comment node-name, next, previous, up
11859 @unnumberedsubsubsec Recursive Pattern: @emph{every}
11860 @cindex Every, type of recursive pattern
11861 @cindex Recursive pattern: every
11863 In the @code{every} recursive pattern, an action is performed on every
11867 The basic pattern is:
11871 If a list be empty, return @code{nil}.
11873 Else, act on the beginning of the list (the @sc{car} of the list)
11876 through a recursive call by the function on the rest (the
11877 @sc{cdr}) of the list,
11879 and, optionally, combine the acted-on element, using @code{cons},
11880 with the results of acting on the rest.
11889 (defun square-each (numbers-list)
11890 "Square each of a NUMBERS LIST, recursively."
11891 (if (not numbers-list) ; do-again-test
11894 (* (car numbers-list) (car numbers-list))
11895 (square-each (cdr numbers-list))))) ; next-step-expression
11899 (square-each '(1 2 3))
11906 If @code{numbers-list} is empty, do nothing. But if it has content,
11907 construct a list combining the square of the first number in the list
11908 with the result of the recursive call.
11910 (The example follows the pattern exactly: @code{nil} is returned if
11911 the numbers' list is empty. In practice, you would write the
11912 conditional so it carries out the action when the numbers' list is not
11915 The @code{print-elements-recursively} function (@pxref{Recursion with
11916 list, , Recursion with a List}) is another example of an @code{every}
11917 pattern, except in this case, rather than bring the results together
11918 using @code{cons}, we print each element of output.
11921 The @code{print-elements-recursively} function looks like this:
11925 (setq animals '(gazelle giraffe lion tiger))
11929 (defun print-elements-recursively (list)
11930 "Print each element of LIST on a line of its own.
11932 (when list ; @r{do-again-test}
11933 (print (car list)) ; @r{body}
11934 (print-elements-recursively ; @r{recursive call}
11935 (cdr list)))) ; @r{next-step-expression}
11937 (print-elements-recursively animals)
11942 The pattern for @code{print-elements-recursively} is:
11946 When the list is empty, do nothing.
11948 But when the list has at least one element,
11951 act on the beginning of the list (the @sc{car} of the list),
11953 and make a recursive call on the rest (the @sc{cdr}) of the list.
11957 @node Accumulate, Keep, Every, Recursive Patterns
11958 @comment node-name, next, previous, up
11959 @unnumberedsubsubsec Recursive Pattern: @emph{accumulate}
11960 @cindex Accumulate, type of recursive pattern
11961 @cindex Recursive pattern: accumulate
11963 Another recursive pattern is called the @code{accumulate} pattern. In
11964 the @code{accumulate} recursive pattern, an action is performed on
11965 every element of a list and the result of that action is accumulated
11966 with the results of performing the action on the other elements.
11968 This is very like the `every' pattern using @code{cons}, except that
11969 @code{cons} is not used, but some other combiner.
11976 If a list be empty, return zero or some other constant.
11978 Else, act on the beginning of the list (the @sc{car} of the list),
11981 and combine that acted-on element, using @code{+} or
11982 some other combining function, with
11984 a recursive call by the function on the rest (the @sc{cdr}) of the list.
11989 Here is an example:
11993 (defun add-elements (numbers-list)
11994 "Add the elements of NUMBERS-LIST together."
11995 (if (not numbers-list)
11997 (+ (car numbers-list) (add-elements (cdr numbers-list)))))
12001 (add-elements '(1 2 3 4))
12006 @xref{Files List, , Making a List of Files}, for an example of the
12007 accumulate pattern.
12009 @node Keep, , Accumulate, Recursive Patterns
12010 @comment node-name, next, previous, up
12011 @unnumberedsubsubsec Recursive Pattern: @emph{keep}
12012 @cindex Keep, type of recursive pattern
12013 @cindex Recursive pattern: keep
12015 A third recursive pattern is called the @code{keep} pattern.
12016 In the @code{keep} recursive pattern, each element of a list is tested;
12017 the element is acted on and the results are kept only if the element
12020 Again, this is very like the `every' pattern, except the element is
12021 skipped unless it meets a criterion.
12024 The pattern has three parts:
12028 If a list be empty, return @code{nil}.
12030 Else, if the beginning of the list (the @sc{car} of the list) passes
12034 act on that element and combine it, using @code{cons} with
12036 a recursive call by the function on the rest (the @sc{cdr}) of the list.
12039 Otherwise, if the beginning of the list (the @sc{car} of the list) fails
12043 skip on that element,
12045 and, recursively call the function on the rest (the @sc{cdr}) of the list.
12050 Here is an example that uses @code{cond}:
12054 (defun keep-three-letter-words (word-list)
12055 "Keep three letter words in WORD-LIST."
12057 ;; First do-again-test: stop-condition
12058 ((not word-list) nil)
12060 ;; Second do-again-test: when to act
12061 ((eq 3 (length (symbol-name (car word-list))))
12062 ;; combine acted-on element with recursive call on shorter list
12063 (cons (car word-list) (keep-three-letter-words (cdr word-list))))
12065 ;; Third do-again-test: when to skip element;
12066 ;; recursively call shorter list with next-step expression
12067 (t (keep-three-letter-words (cdr word-list)))))
12071 (keep-three-letter-words '(one two three four five six))
12072 @result{} (one two six)
12076 It goes without saying that you need not use @code{nil} as the test for
12077 when to stop; and you can, of course, combine these patterns.
12079 @node No Deferment, No deferment solution, Recursive Patterns, Recursion
12080 @subsection Recursion without Deferments
12081 @cindex Deferment in recursion
12082 @cindex Recursion without Deferments
12084 Let's consider again what happens with the @code{triangle-recursively}
12085 function. We will find that the intermediate calculations are
12086 deferred until all can be done.
12089 Here is the function definition:
12093 (defun triangle-recursively (number)
12094 "Return the sum of the numbers 1 through NUMBER inclusive.
12096 (if (= number 1) ; @r{do-again-test}
12098 (+ number ; @r{else-part}
12099 (triangle-recursively ; @r{recursive call}
12100 (1- number))))) ; @r{next-step-expression}
12104 What happens when we call this function with a argument of 7?
12106 The first instance of the @code{triangle-recursively} function adds
12107 the number 7 to the value returned by a second instance of
12108 @code{triangle-recursively}, an instance that has been passed an
12109 argument of 6. That is to say, the first calculation is:
12112 (+ 7 (triangle-recursively 6))
12116 The first instance of @code{triangle-recursively}---you may want to
12117 think of it as a little robot---cannot complete its job. It must hand
12118 off the calculation for @code{(triangle-recursively 6)} to a second
12119 instance of the program, to a second robot. This second individual is
12120 completely different from the first one; it is, in the jargon, a
12121 `different instantiation'. Or, put another way, it is a different
12122 robot. It is the same model as the first; it calculates triangle
12123 numbers recursively; but it has a different serial number.
12125 And what does @code{(triangle-recursively 6)} return? It returns the
12126 number 6 added to the value returned by evaluating
12127 @code{triangle-recursively} with an argument of 5. Using the robot
12128 metaphor, it asks yet another robot to help it.
12134 (+ 7 6 (triangle-recursively 5))
12138 And what happens next?
12141 (+ 7 6 5 (triangle-recursively 4))
12144 Each time @code{triangle-recursively} is called, except for the last
12145 time, it creates another instance of the program---another robot---and
12146 asks it to make a calculation.
12149 Eventually, the full addition is set up and performed:
12155 This design for the function defers the calculation of the first step
12156 until the second can be done, and defers that until the third can be
12157 done, and so on. Each deferment means the computer must remember what
12158 is being waited on. This is not a problem when there are only a few
12159 steps, as in this example. But it can be a problem when there are
12162 @node No deferment solution, , No Deferment, Recursion
12163 @subsection No Deferment Solution
12164 @cindex No deferment solution
12165 @cindex Defermentless solution
12166 @cindex Solution without deferment
12168 The solution to the problem of deferred operations is to write in a
12169 manner that does not defer operations@footnote{The phrase @dfn{tail
12170 recursive} is used to describe such a process, one that uses
12171 `constant space'.}. This requires
12172 writing to a different pattern, often one that involves writing two
12173 function definitions, an `initialization' function and a `helper'
12176 The `initialization' function sets up the job; the `helper' function
12180 Here are the two function definitions for adding up numbers. They are
12181 so simple, I find them hard to understand.
12185 (defun triangle-initialization (number)
12186 "Return the sum of the numbers 1 through NUMBER inclusive.
12187 This is the `initialization' component of a two function
12188 duo that uses recursion."
12189 (triangle-recursive-helper 0 0 number))
12195 (defun triangle-recursive-helper (sum counter number)
12196 "Return SUM, using COUNTER, through NUMBER inclusive.
12197 This is the `helper' component of a two function duo
12198 that uses recursion."
12199 (if (> counter number)
12201 (triangle-recursive-helper (+ sum counter) ; @r{sum}
12202 (1+ counter) ; @r{counter}
12203 number))) ; @r{number}
12208 Install both function definitions by evaluating them, then call
12209 @code{triangle-initialization} with 2 rows:
12213 (triangle-initialization 2)
12218 The `initialization' function calls the first instance of the `helper'
12219 function with three arguments: zero, zero, and a number which is the
12220 number of rows in the triangle.
12222 The first two arguments passed to the `helper' function are
12223 initialization values. These values are changed when
12224 @code{triangle-recursive-helper} invokes new instances.@footnote{The
12225 jargon is mildly confusing: @code{triangle-recursive-helper} uses a
12226 process that is iterative in a procedure that is recursive. The
12227 process is called iterative because the computer need only record the
12228 three values, @code{sum}, @code{counter}, and @code{number}; the
12229 procedure is recursive because the function `calls itself'. On the
12230 other hand, both the process and the procedure used by
12231 @code{triangle-recursively} are called recursive. The word
12232 `recursive' has different meanings in the two contexts.}
12234 Let's see what happens when we have a triangle that has one row. (This
12235 triangle will have one pebble in it!)
12238 @code{triangle-initialization} will call its helper with
12239 the arguments @w{@code{0 0 1}}. That function will run the conditional
12240 test whether @code{(> counter number)}:
12248 and find that the result is false, so it will invoke
12249 the else-part of the @code{if} clause:
12253 (triangle-recursive-helper
12254 (+ sum counter) ; @r{sum plus counter} @result{} @r{sum}
12255 (1+ counter) ; @r{increment counter} @result{} @r{counter}
12256 number) ; @r{number stays the same}
12262 which will first compute:
12266 (triangle-recursive-helper (+ 0 0) ; @r{sum}
12267 (1+ 0) ; @r{counter}
12271 (triangle-recursive-helper 0 1 1)
12275 Again, @code{(> counter number)} will be false, so again, the Lisp
12276 interpreter will evaluate @code{triangle-recursive-helper}, creating a
12277 new instance with new arguments.
12280 This new instance will be;
12284 (triangle-recursive-helper
12285 (+ sum counter) ; @r{sum plus counter} @result{} @r{sum}
12286 (1+ counter) ; @r{increment counter} @result{} @r{counter}
12287 number) ; @r{number stays the same}
12291 (triangle-recursive-helper 1 2 1)
12295 In this case, the @code{(> counter number)} test will be true! So the
12296 instance will return the value of the sum, which will be 1, as
12299 Now, let's pass @code{triangle-initialization} an argument
12300 of 2, to find out how many pebbles there are in a triangle with two rows.
12302 That function calls @code{(triangle-recursive-helper 0 0 2)}.
12305 In stages, the instances called will be:
12309 @r{sum counter number}
12310 (triangle-recursive-helper 0 1 2)
12312 (triangle-recursive-helper 1 2 2)
12314 (triangle-recursive-helper 3 3 2)
12318 When the last instance is called, the @code{(> counter number)} test
12319 will be true, so the instance will return the value of @code{sum},
12322 This kind of pattern helps when you are writing functions that can use
12323 many resources in a computer.
12326 @node Looping exercise, , Recursion, Loops & Recursion
12327 @section Looping Exercise
12331 Write a function similar to @code{triangle} in which each row has a
12332 value which is the square of the row number. Use a @code{while} loop.
12335 Write a function similar to @code{triangle} that multiplies instead of
12339 Rewrite these two functions recursively. Rewrite these functions
12342 @c comma in printed title causes problem in Info cross reference
12344 Write a function for Texinfo mode that creates an index entry at the
12345 beginning of a paragraph for every @samp{@@dfn} within the paragraph.
12346 (In a Texinfo file, @samp{@@dfn} marks a definition. This book is
12347 written in Texinfo.)
12349 Many of the functions you will need are described in two of the
12350 previous chapters, @ref{Cutting & Storing Text, , Cutting and Storing
12351 Text}, and @ref{Yanking, , Yanking Text Back}. If you use
12352 @code{forward-paragraph} to put the index entry at the beginning of
12353 the paragraph, you will have to use @w{@kbd{C-h f}}
12354 (@code{describe-function}) to find out how to make the command go
12357 For more information, see
12359 @ref{Indicating, , Indicating Definitions, texinfo}.
12362 @ref{Indicating, , Indicating, texinfo, Texinfo Manual}, which goes to
12363 a Texinfo manual in the current directory. Or, if you are on the
12365 @uref{http://www.gnu.org/software/texinfo/manual/texinfo/}
12368 ``Indicating Definitions, Commands, etc.'' in @cite{Texinfo, The GNU
12369 Documentation Format}.
12373 @node Regexp Search, Counting Words, Loops & Recursion, Top
12374 @comment node-name, next, previous, up
12375 @chapter Regular Expression Searches
12376 @cindex Searches, illustrating
12377 @cindex Regular expression searches
12378 @cindex Patterns, searching for
12379 @cindex Motion by sentence and paragraph
12380 @cindex Sentences, movement by
12381 @cindex Paragraphs, movement by
12383 Regular expression searches are used extensively in GNU Emacs. The
12384 two functions, @code{forward-sentence} and @code{forward-paragraph},
12385 illustrate these searches well. They use regular expressions to find
12386 where to move point. The phrase `regular expression' is often written
12389 Regular expression searches are described in @ref{Regexp Search, ,
12390 Regular Expression Search, emacs, The GNU Emacs Manual}, as well as in
12391 @ref{Regular Expressions, , , elisp, The GNU Emacs Lisp Reference
12392 Manual}. In writing this chapter, I am presuming that you have at
12393 least a mild acquaintance with them. The major point to remember is
12394 that regular expressions permit you to search for patterns as well as
12395 for literal strings of characters. For example, the code in
12396 @code{forward-sentence} searches for the pattern of possible
12397 characters that could mark the end of a sentence, and moves point to
12400 Before looking at the code for the @code{forward-sentence} function, it
12401 is worth considering what the pattern that marks the end of a sentence
12402 must be. The pattern is discussed in the next section; following that
12403 is a description of the regular expression search function,
12404 @code{re-search-forward}. The @code{forward-sentence} function
12405 is described in the section following. Finally, the
12406 @code{forward-paragraph} function is described in the last section of
12407 this chapter. @code{forward-paragraph} is a complex function that
12408 introduces several new features.
12411 * sentence-end:: The regular expression for @code{sentence-end}.
12412 * re-search-forward:: Very similar to @code{search-forward}.
12413 * forward-sentence:: A straightforward example of regexp search.
12414 * forward-paragraph:: A somewhat complex example.
12415 * etags:: How to create your own @file{TAGS} table.
12417 * re-search Exercises::
12420 @node sentence-end, re-search-forward, Regexp Search, Regexp Search
12421 @comment node-name, next, previous, up
12422 @section The Regular Expression for @code{sentence-end}
12423 @findex sentence-end
12425 The symbol @code{sentence-end} is bound to the pattern that marks the
12426 end of a sentence. What should this regular expression be?
12428 Clearly, a sentence may be ended by a period, a question mark, or an
12429 exclamation mark. Indeed, in English, only clauses that end with one
12430 of those three characters should be considered the end of a sentence.
12431 This means that the pattern should include the character set:
12437 However, we do not want @code{forward-sentence} merely to jump to a
12438 period, a question mark, or an exclamation mark, because such a character
12439 might be used in the middle of a sentence. A period, for example, is
12440 used after abbreviations. So other information is needed.
12442 According to convention, you type two spaces after every sentence, but
12443 only one space after a period, a question mark, or an exclamation mark in
12444 the body of a sentence. So a period, a question mark, or an exclamation
12445 mark followed by two spaces is a good indicator of an end of sentence.
12446 However, in a file, the two spaces may instead be a tab or the end of a
12447 line. This means that the regular expression should include these three
12448 items as alternatives.
12451 This group of alternatives will look like this:
12462 Here, @samp{$} indicates the end of the line, and I have pointed out
12463 where the tab and two spaces are inserted in the expression. Both are
12464 inserted by putting the actual characters into the expression.
12466 Two backslashes, @samp{\\}, are required before the parentheses and
12467 vertical bars: the first backslash quotes the following backslash in
12468 Emacs; and the second indicates that the following character, the
12469 parenthesis or the vertical bar, is special.
12472 Also, a sentence may be followed by one or more carriage returns, like
12483 Like tabs and spaces, a carriage return is inserted into a regular
12484 expression by inserting it literally. The asterisk indicates that the
12485 @key{RET} is repeated zero or more times.
12487 But a sentence end does not consist only of a period, a question mark or
12488 an exclamation mark followed by appropriate space: a closing quotation
12489 mark or a closing brace of some kind may precede the space. Indeed more
12490 than one such mark or brace may precede the space. These require a
12491 expression that looks like this:
12497 In this expression, the first @samp{]} is the first character in the
12498 expression; the second character is @samp{"}, which is preceded by a
12499 @samp{\} to tell Emacs the @samp{"} is @emph{not} special. The last
12500 three characters are @samp{'}, @samp{)}, and @samp{@}}.
12502 All this suggests what the regular expression pattern for matching the
12503 end of a sentence should be; and, indeed, if we evaluate
12504 @code{sentence-end} we find that it returns the following value:
12509 @result{} "[.?!][]\"')@}]*\\($\\| \\| \\)[
12515 (Well, not in GNU Emacs 22; that is because of an effort to make the
12516 process simpler and to handle more glyphs and languages. When the
12517 value of @code{sentence-end} is @code{nil}, then use the value defined
12518 by the function @code{sentence-end}. (Here is a use of the difference
12519 between a value and a function in Emacs Lisp.) The function returns a
12520 value constructed from the variables @code{sentence-end-base},
12521 @code{sentence-end-double-space}, @code{sentence-end-without-period},
12522 and @code{sentence-end-without-space}. The critical variable is
12523 @code{sentence-end-base}; its global value is similar to the one
12524 described above but it also contains two additional quotation marks.
12525 These have differing degrees of curliness. The
12526 @code{sentence-end-without-period} variable, when true, tells Emacs
12527 that a sentence may end without a period, such as text in Thai.)
12531 (Note that here the @key{TAB}, two spaces, and @key{RET} are shown
12532 literally in the pattern.)
12534 This regular expression can be deciphered as follows:
12538 The first part of the pattern is the three characters, a period, a question
12539 mark and an exclamation mark, within square brackets. The pattern must
12540 begin with one or other of these characters.
12543 The second part of the pattern is the group of closing braces and
12544 quotation marks, which can appear zero or more times. These may follow
12545 the period, question mark or exclamation mark. In a regular expression,
12546 the backslash, @samp{\}, followed by the double quotation mark,
12547 @samp{"}, indicates the class of string-quote characters. Usually, the
12548 double quotation mark is the only character in this class. The
12549 asterisk, @samp{*}, indicates that the items in the previous group (the
12550 group surrounded by square brackets, @samp{[]}) may be repeated zero or
12553 @item \\($\\| \\| \\)
12554 The third part of the pattern is one or other of: either the end of a
12555 line, or two blank spaces, or a tab. The double back-slashes are used
12556 to prevent Emacs from reading the parentheses and vertical bars as part
12557 of the search pattern; the parentheses are used to mark the group and
12558 the vertical bars are used to indicated that the patterns to either side
12559 of them are alternatives. The dollar sign is used to indicate the end
12560 of a line and both the two spaces and the tab are each inserted as is to
12561 indicate what they are.
12564 Finally, the last part of the pattern indicates that the end of the line
12565 or the whitespace following the period, question mark or exclamation
12566 mark may, but need not, be followed by one or more carriage returns. In
12567 the pattern, the carriage return is inserted as an actual carriage
12568 return between square brackets but here it is shown as @key{RET}.
12572 @node re-search-forward, forward-sentence, sentence-end, Regexp Search
12573 @comment node-name, next, previous, up
12574 @section The @code{re-search-forward} Function
12575 @findex re-search-forward
12577 The @code{re-search-forward} function is very like the
12578 @code{search-forward} function. (@xref{search-forward, , The
12579 @code{search-forward} Function}.)
12581 @code{re-search-forward} searches for a regular expression. If the
12582 search is successful, it leaves point immediately after the last
12583 character in the target. If the search is backwards, it leaves point
12584 just before the first character in the target. You may tell
12585 @code{re-search-forward} to return @code{t} for true. (Moving point
12586 is therefore a `side effect'.)
12588 Like @code{search-forward}, the @code{re-search-forward} function takes
12593 The first argument is the regular expression that the function searches
12594 for. The regular expression will be a string between quotations marks.
12597 The optional second argument limits how far the function will search; it is a
12598 bound, which is specified as a position in the buffer.
12601 The optional third argument specifies how the function responds to
12602 failure: @code{nil} as the third argument causes the function to
12603 signal an error (and print a message) when the search fails; any other
12604 value causes it to return @code{nil} if the search fails and @code{t}
12605 if the search succeeds.
12608 The optional fourth argument is the repeat count. A negative repeat
12609 count causes @code{re-search-forward} to search backwards.
12613 The template for @code{re-search-forward} looks like this:
12617 (re-search-forward "@var{regular-expression}"
12618 @var{limit-of-search}
12619 @var{what-to-do-if-search-fails}
12620 @var{repeat-count})
12624 The second, third, and fourth arguments are optional. However, if you
12625 want to pass a value to either or both of the last two arguments, you
12626 must also pass a value to all the preceding arguments. Otherwise, the
12627 Lisp interpreter will mistake which argument you are passing the value
12631 In the @code{forward-sentence} function, the regular expression will be
12632 the value of the variable @code{sentence-end}. In simple form, that is:
12636 "[.?!][]\"')@}]*\\($\\| \\| \\)[
12642 The limit of the search will be the end of the paragraph (since a
12643 sentence cannot go beyond a paragraph). If the search fails, the
12644 function will return @code{nil}; and the repeat count will be provided
12645 by the argument to the @code{forward-sentence} function.
12647 @node forward-sentence, forward-paragraph, re-search-forward, Regexp Search
12648 @comment node-name, next, previous, up
12649 @section @code{forward-sentence}
12650 @findex forward-sentence
12652 The command to move the cursor forward a sentence is a straightforward
12653 illustration of how to use regular expression searches in Emacs Lisp.
12654 Indeed, the function looks longer and more complicated than it is; this
12655 is because the function is designed to go backwards as well as forwards;
12656 and, optionally, over more than one sentence. The function is usually
12657 bound to the key command @kbd{M-e}.
12660 * Complete forward-sentence::
12661 * fwd-sentence while loops:: Two @code{while} loops.
12662 * fwd-sentence re-search:: A regular expression search.
12665 @node Complete forward-sentence, fwd-sentence while loops, forward-sentence, forward-sentence
12667 @unnumberedsubsec Complete @code{forward-sentence} function definition
12671 Here is the code for @code{forward-sentence}:
12676 (defun forward-sentence (&optional arg)
12677 "Move forward to next `sentence-end'. With argument, repeat.
12678 With negative argument, move backward repeatedly to `sentence-beginning'.
12680 The variable `sentence-end' is a regular expression that matches ends of
12681 sentences. Also, every paragraph boundary terminates sentences as well."
12685 (or arg (setq arg 1))
12686 (let ((opoint (point))
12687 (sentence-end (sentence-end)))
12689 (let ((pos (point))
12690 (par-beg (save-excursion (start-of-paragraph-text) (point))))
12691 (if (and (re-search-backward sentence-end par-beg t)
12692 (or (< (match-end 0) pos)
12693 (re-search-backward sentence-end par-beg t)))
12694 (goto-char (match-end 0))
12695 (goto-char par-beg)))
12696 (setq arg (1+ arg)))
12700 (let ((par-end (save-excursion (end-of-paragraph-text) (point))))
12701 (if (re-search-forward sentence-end par-end t)
12702 (skip-chars-backward " \t\n")
12703 (goto-char par-end)))
12704 (setq arg (1- arg)))
12705 (constrain-to-field nil opoint t)))
12713 (defun forward-sentence (&optional arg)
12714 "Move forward to next sentence-end. With argument, repeat.
12715 With negative argument, move backward repeatedly to sentence-beginning.
12716 Sentence ends are identified by the value of sentence-end
12717 treated as a regular expression. Also, every paragraph boundary
12718 terminates sentences as well."
12722 (or arg (setq arg 1))
12725 (save-excursion (start-of-paragraph-text) (point))))
12726 (if (re-search-backward
12727 (concat sentence-end "[^ \t\n]") par-beg t)
12728 (goto-char (1- (match-end 0)))
12729 (goto-char par-beg)))
12730 (setq arg (1+ arg)))
12733 (save-excursion (end-of-paragraph-text) (point))))
12734 (if (re-search-forward sentence-end par-end t)
12735 (skip-chars-backward " \t\n")
12736 (goto-char par-end)))
12737 (setq arg (1- arg))))
12742 The function looks long at first sight and it is best to look at its
12743 skeleton first, and then its muscle. The way to see the skeleton is to
12744 look at the expressions that start in the left-most columns:
12748 (defun forward-sentence (&optional arg)
12749 "@var{documentation}@dots{}"
12751 (or arg (setq arg 1))
12752 (let ((opoint (point)) (sentence-end (sentence-end)))
12754 (let ((pos (point))
12755 (par-beg (save-excursion (start-of-paragraph-text) (point))))
12756 @var{rest-of-body-of-while-loop-when-going-backwards}
12758 (let ((par-end (save-excursion (end-of-paragraph-text) (point))))
12759 @var{rest-of-body-of-while-loop-when-going-forwards}
12760 @var{handle-forms-and-equivalent}
12764 This looks much simpler! The function definition consists of
12765 documentation, an @code{interactive} expression, an @code{or}
12766 expression, a @code{let} expression, and @code{while} loops.
12768 Let's look at each of these parts in turn.
12770 We note that the documentation is thorough and understandable.
12772 The function has an @code{interactive "p"} declaration. This means
12773 that the processed prefix argument, if any, is passed to the
12774 function as its argument. (This will be a number.) If the function
12775 is not passed an argument (it is optional) then the argument
12776 @code{arg} will be bound to 1.
12778 When @code{forward-sentence} is called non-interactively without an
12779 argument, @code{arg} is bound to @code{nil}. The @code{or} expression
12780 handles this. What it does is either leave the value of @code{arg} as
12781 it is, but only if @code{arg} is bound to a value; or it sets the
12782 value of @code{arg} to 1, in the case when @code{arg} is bound to
12785 Next is a @code{let}. That specifies the values of two local
12786 variables, @code{point} and @code{sentence-end}. The local value of
12787 point, from before the search, is used in the
12788 @code{constrain-to-field} function which handles forms and
12789 equivalents. The @code{sentence-end} variable is set by the
12790 @code{sentence-end} function.
12792 @node fwd-sentence while loops, fwd-sentence re-search, Complete forward-sentence, forward-sentence
12793 @unnumberedsubsec The @code{while} loops
12795 Two @code{while} loops follow. The first @code{while} has a
12796 true-or-false-test that tests true if the prefix argument for
12797 @code{forward-sentence} is a negative number. This is for going
12798 backwards. The body of this loop is similar to the body of the second
12799 @code{while} clause, but it is not exactly the same. We will skip
12800 this @code{while} loop and concentrate on the second @code{while}
12804 The second @code{while} loop is for moving point forward. Its skeleton
12809 (while (> arg 0) ; @r{true-or-false-test}
12811 (if (@var{true-or-false-test})
12814 (setq arg (1- arg)))) ; @code{while} @r{loop decrementer}
12818 The @code{while} loop is of the decrementing kind.
12819 (@xref{Decrementing Loop, , A Loop with a Decrementing Counter}.) It
12820 has a true-or-false-test that tests true so long as the counter (in
12821 this case, the variable @code{arg}) is greater than zero; and it has a
12822 decrementer that subtracts 1 from the value of the counter every time
12825 If no prefix argument is given to @code{forward-sentence}, which is
12826 the most common way the command is used, this @code{while} loop will
12827 run once, since the value of @code{arg} will be 1.
12829 The body of the @code{while} loop consists of a @code{let} expression,
12830 which creates and binds a local variable, and has, as its body, an
12831 @code{if} expression.
12834 The body of the @code{while} loop looks like this:
12839 (save-excursion (end-of-paragraph-text) (point))))
12840 (if (re-search-forward sentence-end par-end t)
12841 (skip-chars-backward " \t\n")
12842 (goto-char par-end)))
12846 The @code{let} expression creates and binds the local variable
12847 @code{par-end}. As we shall see, this local variable is designed to
12848 provide a bound or limit to the regular expression search. If the
12849 search fails to find a proper sentence ending in the paragraph, it will
12850 stop on reaching the end of the paragraph.
12852 But first, let us examine how @code{par-end} is bound to the value of
12853 the end of the paragraph. What happens is that the @code{let} sets the
12854 value of @code{par-end} to the value returned when the Lisp interpreter
12855 evaluates the expression
12859 (save-excursion (end-of-paragraph-text) (point))
12864 In this expression, @code{(end-of-paragraph-text)} moves point to the
12865 end of the paragraph, @code{(point)} returns the value of point, and then
12866 @code{save-excursion} restores point to its original position. Thus,
12867 the @code{let} binds @code{par-end} to the value returned by the
12868 @code{save-excursion} expression, which is the position of the end of
12869 the paragraph. (The @code{end-of-paragraph-text} function uses
12870 @code{forward-paragraph}, which we will discuss shortly.)
12873 Emacs next evaluates the body of the @code{let}, which is an @code{if}
12874 expression that looks like this:
12878 (if (re-search-forward sentence-end par-end t) ; @r{if-part}
12879 (skip-chars-backward " \t\n") ; @r{then-part}
12880 (goto-char par-end))) ; @r{else-part}
12884 The @code{if} tests whether its first argument is true and if so,
12885 evaluates its then-part; otherwise, the Emacs Lisp interpreter
12886 evaluates the else-part. The true-or-false-test of the @code{if}
12887 expression is the regular expression search.
12889 It may seem odd to have what looks like the `real work' of
12890 the @code{forward-sentence} function buried here, but this is a common
12891 way this kind of operation is carried out in Lisp.
12893 @node fwd-sentence re-search, , fwd-sentence while loops, forward-sentence
12894 @unnumberedsubsec The regular expression search
12896 The @code{re-search-forward} function searches for the end of the
12897 sentence, that is, for the pattern defined by the @code{sentence-end}
12898 regular expression. If the pattern is found---if the end of the sentence is
12899 found---then the @code{re-search-forward} function does two things:
12903 The @code{re-search-forward} function carries out a side effect, which
12904 is to move point to the end of the occurrence found.
12907 The @code{re-search-forward} function returns a value of true. This is
12908 the value received by the @code{if}, and means that the search was
12913 The side effect, the movement of point, is completed before the
12914 @code{if} function is handed the value returned by the successful
12915 conclusion of the search.
12917 When the @code{if} function receives the value of true from a successful
12918 call to @code{re-search-forward}, the @code{if} evaluates the then-part,
12919 which is the expression @code{(skip-chars-backward " \t\n")}. This
12920 expression moves backwards over any blank spaces, tabs or carriage
12921 returns until a printed character is found and then leaves point after
12922 the character. Since point has already been moved to the end of the
12923 pattern that marks the end of the sentence, this action leaves point
12924 right after the closing printed character of the sentence, which is
12927 On the other hand, if the @code{re-search-forward} function fails to
12928 find a pattern marking the end of the sentence, the function returns
12929 false. The false then causes the @code{if} to evaluate its third
12930 argument, which is @code{(goto-char par-end)}: it moves point to the
12931 end of the paragraph.
12933 (And if the text is in a form or equivalent, and point may not move
12934 fully, then the @code{constrain-to-field} function comes into play.)
12936 Regular expression searches are exceptionally useful and the pattern
12937 illustrated by @code{re-search-forward}, in which the search is the
12938 test of an @code{if} expression, is handy. You will see or write code
12939 incorporating this pattern often.
12941 @node forward-paragraph, etags, forward-sentence, Regexp Search
12942 @comment node-name, next, previous, up
12943 @section @code{forward-paragraph}: a Goldmine of Functions
12944 @findex forward-paragraph
12948 (defun forward-paragraph (&optional arg)
12949 "Move forward to end of paragraph.
12950 With argument ARG, do it ARG times;
12951 a negative argument ARG = -N means move backward N paragraphs.
12953 A line which `paragraph-start' matches either separates paragraphs
12954 \(if `paragraph-separate' matches it also) or is the first line of a paragraph.
12955 A paragraph end is the beginning of a line which is not part of the paragraph
12956 to which the end of the previous line belongs, or the end of the buffer.
12957 Returns the count of paragraphs left to move."
12959 (or arg (setq arg 1))
12960 (let* ((opoint (point))
12961 (fill-prefix-regexp
12962 (and fill-prefix (not (equal fill-prefix ""))
12963 (not paragraph-ignore-fill-prefix)
12964 (regexp-quote fill-prefix)))
12965 ;; Remove ^ from paragraph-start and paragraph-sep if they are there.
12966 ;; These regexps shouldn't be anchored, because we look for them
12967 ;; starting at the left-margin. This allows paragraph commands to
12968 ;; work normally with indented text.
12969 ;; This hack will not find problem cases like "whatever\\|^something".
12970 (parstart (if (and (not (equal "" paragraph-start))
12971 (equal ?^ (aref paragraph-start 0)))
12972 (substring paragraph-start 1)
12974 (parsep (if (and (not (equal "" paragraph-separate))
12975 (equal ?^ (aref paragraph-separate 0)))
12976 (substring paragraph-separate 1)
12977 paragraph-separate))
12979 (if fill-prefix-regexp
12980 (concat parsep "\\|"
12981 fill-prefix-regexp "[ \t]*$")
12983 ;; This is used for searching.
12984 (sp-parstart (concat "^[ \t]*\\(?:" parstart "\\|" parsep "\\)"))
12986 (while (and (< arg 0) (not (bobp)))
12987 (if (and (not (looking-at parsep))
12988 (re-search-backward "^\n" (max (1- (point)) (point-min)) t)
12989 (looking-at parsep))
12990 (setq arg (1+ arg))
12991 (setq start (point))
12992 ;; Move back over paragraph-separating lines.
12993 (forward-char -1) (beginning-of-line)
12994 (while (and (not (bobp))
12995 (progn (move-to-left-margin)
12996 (looking-at parsep)))
13000 (setq arg (1+ arg))
13001 ;; Go to end of the previous (non-separating) line.
13003 ;; Search back for line that starts or separates paragraphs.
13004 (if (if fill-prefix-regexp
13005 ;; There is a fill prefix; it overrides parstart.
13006 (let (multiple-lines)
13007 (while (and (progn (beginning-of-line) (not (bobp)))
13008 (progn (move-to-left-margin)
13009 (not (looking-at parsep)))
13010 (looking-at fill-prefix-regexp))
13011 (unless (= (point) start)
13012 (setq multiple-lines t))
13014 (move-to-left-margin)
13015 ;; This deleted code caused a long hanging-indent line
13016 ;; not to be filled together with the following lines.
13017 ;; ;; Don't move back over a line before the paragraph
13018 ;; ;; which doesn't start with fill-prefix
13019 ;; ;; unless that is the only line we've moved over.
13020 ;; (and (not (looking-at fill-prefix-regexp))
13022 ;; (forward-line 1))
13024 (while (and (re-search-backward sp-parstart nil 1)
13025 (setq found-start t)
13026 ;; Found a candidate, but need to check if it is a
13028 (progn (setq start (point))
13029 (move-to-left-margin)
13030 (not (looking-at parsep)))
13031 (not (and (looking-at parstart)
13032 (or (not use-hard-newlines)
13035 (1- start) 'hard)))))
13036 (setq found-start nil)
13041 ;; Move forward over paragraph separators.
13042 ;; We know this cannot reach the place we started
13043 ;; because we know we moved back over a non-separator.
13044 (while (and (not (eobp))
13045 (progn (move-to-left-margin)
13046 (looking-at parsep)))
13048 ;; If line before paragraph is just margin, back up to there.
13050 (if (> (current-column) (current-left-margin))
13052 (skip-chars-backward " \t")
13054 (forward-line 1))))
13055 ;; No starter or separator line => use buffer beg.
13056 (goto-char (point-min))))))
13058 (while (and (> arg 0) (not (eobp)))
13059 ;; Move forward over separator lines...
13060 (while (and (not (eobp))
13061 (progn (move-to-left-margin) (not (eobp)))
13062 (looking-at parsep))
13064 (unless (eobp) (setq arg (1- arg)))
13065 ;; ... and one more line.
13067 (if fill-prefix-regexp
13068 ;; There is a fill prefix; it overrides parstart.
13069 (while (and (not (eobp))
13070 (progn (move-to-left-margin) (not (eobp)))
13071 (not (looking-at parsep))
13072 (looking-at fill-prefix-regexp))
13074 (while (and (re-search-forward sp-parstart nil 1)
13075 (progn (setq start (match-beginning 0))
13078 (progn (move-to-left-margin)
13079 (not (looking-at parsep)))
13080 (or (not (looking-at parstart))
13081 (and use-hard-newlines
13082 (not (get-text-property (1- start) 'hard)))))
13084 (if (< (point) (point-max))
13085 (goto-char start))))
13086 (constrain-to-field nil opoint t)
13087 ;; Return the number of steps that could not be done.
13091 The @code{forward-paragraph} function moves point forward to the end
13092 of the paragraph. It is usually bound to @kbd{M-@}} and makes use of a
13093 number of functions that are important in themselves, including
13094 @code{let*}, @code{match-beginning}, and @code{looking-at}.
13096 The function definition for @code{forward-paragraph} is considerably
13097 longer than the function definition for @code{forward-sentence}
13098 because it works with a paragraph, each line of which may begin with a
13101 A fill prefix consists of a string of characters that are repeated at
13102 the beginning of each line. For example, in Lisp code, it is a
13103 convention to start each line of a paragraph-long comment with
13104 @samp{;;; }. In Text mode, four blank spaces make up another common
13105 fill prefix, creating an indented paragraph. (@xref{Fill Prefix, , ,
13106 emacs, The GNU Emacs Manual}, for more information about fill
13109 The existence of a fill prefix means that in addition to being able to
13110 find the end of a paragraph whose lines begin on the left-most
13111 column, the @code{forward-paragraph} function must be able to find the
13112 end of a paragraph when all or many of the lines in the buffer begin
13113 with the fill prefix.
13115 Moreover, it is sometimes practical to ignore a fill prefix that
13116 exists, especially when blank lines separate paragraphs.
13117 This is an added complication.
13120 * forward-paragraph in brief:: Key parts of the function definition.
13121 * fwd-para let:: The @code{let*} expression.
13122 * fwd-para while:: The forward motion @code{while} loop.
13125 @node forward-paragraph in brief, fwd-para let, forward-paragraph, forward-paragraph
13127 @unnumberedsubsec Shortened @code{forward-paragraph} function definition
13130 Rather than print all of the @code{forward-paragraph} function, we
13131 will only print parts of it. Read without preparation, the function
13135 In outline, the function looks like this:
13139 (defun forward-paragraph (&optional arg)
13140 "@var{documentation}@dots{}"
13142 (or arg (setq arg 1))
13145 (while (and (< arg 0) (not (bobp))) ; @r{backward-moving-code}
13147 (while (and (> arg 0) (not (eobp))) ; @r{forward-moving-code}
13152 The first parts of the function are routine: the function's argument
13153 list consists of one optional argument. Documentation follows.
13155 The lower case @samp{p} in the @code{interactive} declaration means
13156 that the processed prefix argument, if any, is passed to the function.
13157 This will be a number, and is the repeat count of how many paragraphs
13158 point will move. The @code{or} expression in the next line handles
13159 the common case when no argument is passed to the function, which occurs
13160 if the function is called from other code rather than interactively.
13161 This case was described earlier. (@xref{forward-sentence, The
13162 @code{forward-sentence} function}.) Now we reach the end of the
13163 familiar part of this function.
13165 @node fwd-para let, fwd-para while, forward-paragraph in brief, forward-paragraph
13166 @unnumberedsubsec The @code{let*} expression
13168 The next line of the @code{forward-paragraph} function begins a
13169 @code{let*} expression. This is a different than @code{let}. The
13170 symbol is @code{let*} not @code{let}.
13172 The @code{let*} special form is like @code{let} except that Emacs sets
13173 each variable in sequence, one after another, and variables in the
13174 latter part of the varlist can make use of the values to which Emacs
13175 set variables in the earlier part of the varlist.
13178 ( refappend save-excursion, , code save-excursion in code append-to-buffer .)
13181 (@ref{append save-excursion, , @code{save-excursion} in @code{append-to-buffer}}.)
13183 In the @code{let*} expression in this function, Emacs binds a total of
13184 seven variables: @code{opoint}, @code{fill-prefix-regexp},
13185 @code{parstart}, @code{parsep}, @code{sp-parstart}, @code{start}, and
13186 @code{found-start}.
13188 The variable @code{parsep} appears twice, first, to remove instances
13189 of @samp{^}, and second, to handle fill prefixes.
13191 The variable @code{opoint} is just the value of @code{point}. As you
13192 can guess, it is used in a @code{constrain-to-field} expression, just
13193 as in @code{forward-sentence}.
13195 The variable @code{fill-prefix-regexp} is set to the value returned by
13196 evaluating the following list:
13201 (not (equal fill-prefix ""))
13202 (not paragraph-ignore-fill-prefix)
13203 (regexp-quote fill-prefix))
13208 This is an expression whose first element is the @code{and} special form.
13210 As we learned earlier (@pxref{kill-new function, , The @code{kill-new}
13211 function}), the @code{and} special form evaluates each of its
13212 arguments until one of the arguments returns a value of @code{nil}, in
13213 which case the @code{and} expression returns @code{nil}; however, if
13214 none of the arguments returns a value of @code{nil}, the value
13215 resulting from evaluating the last argument is returned. (Since such
13216 a value is not @code{nil}, it is considered true in Lisp.) In other
13217 words, an @code{and} expression returns a true value only if all its
13218 arguments are true.
13221 In this case, the variable @code{fill-prefix-regexp} is bound to a
13222 non-@code{nil} value only if the following four expressions produce a
13223 true (i.e., a non-@code{nil}) value when they are evaluated; otherwise,
13224 @code{fill-prefix-regexp} is bound to @code{nil}.
13228 When this variable is evaluated, the value of the fill prefix, if any,
13229 is returned. If there is no fill prefix, this variable returns
13232 @item (not (equal fill-prefix "")
13233 This expression checks whether an existing fill prefix is an empty
13234 string, that is, a string with no characters in it. An empty string is
13235 not a useful fill prefix.
13237 @item (not paragraph-ignore-fill-prefix)
13238 This expression returns @code{nil} if the variable
13239 @code{paragraph-ignore-fill-prefix} has been turned on by being set to a
13240 true value such as @code{t}.
13242 @item (regexp-quote fill-prefix)
13243 This is the last argument to the @code{and} special form. If all the
13244 arguments to the @code{and} are true, the value resulting from
13245 evaluating this expression will be returned by the @code{and} expression
13246 and bound to the variable @code{fill-prefix-regexp},
13249 @findex regexp-quote
13251 The result of evaluating this @code{and} expression successfully is that
13252 @code{fill-prefix-regexp} will be bound to the value of
13253 @code{fill-prefix} as modified by the @code{regexp-quote} function.
13254 What @code{regexp-quote} does is read a string and return a regular
13255 expression that will exactly match the string and match nothing else.
13256 This means that @code{fill-prefix-regexp} will be set to a value that
13257 will exactly match the fill prefix if the fill prefix exists.
13258 Otherwise, the variable will be set to @code{nil}.
13260 The next two local variables in the @code{let*} expression are
13261 designed to remove instances of @samp{^} from @code{parstart} and
13262 @code{parsep}, the local variables which indicate the paragraph start
13263 and the paragraph separator. The next expression sets @code{parsep}
13264 again. That is to handle fill prefixes.
13266 This is the setting that requires the definition call @code{let*}
13267 rather than @code{let}. The true-or-false-test for the @code{if}
13268 depends on whether the variable @code{fill-prefix-regexp} evaluates to
13269 @code{nil} or some other value.
13271 If @code{fill-prefix-regexp} does not have a value, Emacs evaluates
13272 the else-part of the @code{if} expression and binds @code{parsep} to
13273 its local value. (@code{parsep} is a regular expression that matches
13274 what separates paragraphs.)
13276 But if @code{fill-prefix-regexp} does have a value, Emacs evaluates
13277 the then-part of the @code{if} expression and binds @code{parsep} to a
13278 regular expression that includes the @code{fill-prefix-regexp} as part
13281 Specifically, @code{parsep} is set to the original value of the
13282 paragraph separate regular expression concatenated with an alternative
13283 expression that consists of the @code{fill-prefix-regexp} followed by
13284 optional whitespace to the end of the line. The whitespace is defined
13285 by @w{@code{"[ \t]*$"}}.) The @samp{\\|} defines this portion of the
13286 regexp as an alternative to @code{parsep}.
13288 According to a comment in the code, the next local variable,
13289 @code{sp-parstart}, is used for searching, and then the final two,
13290 @code{start} and @code{found-start}, are set to @code{nil}.
13292 Now we get into the body of the @code{let*}. The first part of the body
13293 of the @code{let*} deals with the case when the function is given a
13294 negative argument and is therefore moving backwards. We will skip this
13297 @node fwd-para while, , fwd-para let, forward-paragraph
13298 @unnumberedsubsec The forward motion @code{while} loop
13300 The second part of the body of the @code{let*} deals with forward
13301 motion. It is a @code{while} loop that repeats itself so long as the
13302 value of @code{arg} is greater than zero. In the most common use of
13303 the function, the value of the argument is 1, so the body of the
13304 @code{while} loop is evaluated exactly once, and the cursor moves
13305 forward one paragraph.
13308 (while (and (> arg 0) (not (eobp)))
13310 ;; Move forward over separator lines...
13311 (while (and (not (eobp))
13312 (progn (move-to-left-margin) (not (eobp)))
13313 (looking-at parsep))
13315 (unless (eobp) (setq arg (1- arg)))
13316 ;; ... and one more line.
13319 (if fill-prefix-regexp
13320 ;; There is a fill prefix; it overrides parstart.
13321 (while (and (not (eobp))
13322 (progn (move-to-left-margin) (not (eobp)))
13323 (not (looking-at parsep))
13324 (looking-at fill-prefix-regexp))
13327 (while (and (re-search-forward sp-parstart nil 1)
13328 (progn (setq start (match-beginning 0))
13331 (progn (move-to-left-margin)
13332 (not (looking-at parsep)))
13333 (or (not (looking-at parstart))
13334 (and use-hard-newlines
13335 (not (get-text-property (1- start) 'hard)))))
13338 (if (< (point) (point-max))
13339 (goto-char start))))
13342 This part handles three situations: when point is between paragraphs,
13343 when there is a fill prefix and when there is no fill prefix.
13346 The @code{while} loop looks like this:
13350 ;; @r{going forwards and not at the end of the buffer}
13351 (while (and (> arg 0) (not (eobp)))
13353 ;; @r{between paragraphs}
13354 ;; Move forward over separator lines...
13355 (while (and (not (eobp))
13356 (progn (move-to-left-margin) (not (eobp)))
13357 (looking-at parsep))
13359 ;; @r{This decrements the loop}
13360 (unless (eobp) (setq arg (1- arg)))
13361 ;; ... and one more line.
13366 (if fill-prefix-regexp
13367 ;; There is a fill prefix; it overrides parstart;
13368 ;; we go forward line by line
13369 (while (and (not (eobp))
13370 (progn (move-to-left-margin) (not (eobp)))
13371 (not (looking-at parsep))
13372 (looking-at fill-prefix-regexp))
13377 ;; There is no fill prefix;
13378 ;; we go forward character by character
13379 (while (and (re-search-forward sp-parstart nil 1)
13380 (progn (setq start (match-beginning 0))
13383 (progn (move-to-left-margin)
13384 (not (looking-at parsep)))
13385 (or (not (looking-at parstart))
13386 (and use-hard-newlines
13387 (not (get-text-property (1- start) 'hard)))))
13392 ;; and if there is no fill prefix and if we are not at the end,
13393 ;; go to whatever was found in the regular expression search
13395 (if (< (point) (point-max))
13396 (goto-char start))))
13401 We can see that this is a decrementing counter @code{while} loop,
13402 using the expression @code{(setq arg (1- arg))} as the decrementer.
13403 That expression is not far from the @code{while}, but is hidden in
13404 another Lisp macro, an @code{unless} macro. Unless we are at the end
13405 of the buffer --- that is what the @code{eobp} function determines; it
13406 is an abbreviation of @samp{End Of Buffer P} --- we decrease the value
13407 of @code{arg} by one.
13409 (If we are at the end of the buffer, we cannot go forward any more and
13410 the next loop of the @code{while} expression will test false since the
13411 test is an @code{and} with @code{(not (eobp))}. The @code{not}
13412 function means exactly as you expect; it is another name for
13413 @code{null}, a function that returns true when its argument is false.)
13415 Interestingly, the loop count is not decremented until we leave the
13416 space between paragraphs, unless we come to the end of buffer or stop
13417 seeing the local value of the paragraph separator.
13419 That second @code{while} also has a @code{(move-to-left-margin)}
13420 expression. The function is self-explanatory. It is inside a
13421 @code{progn} expression and not the last element of its body, so it is
13422 only invoked for its side effect, which is to move point to the left
13423 margin of the current line.
13426 The @code{looking-at} function is also self-explanatory; it returns
13427 true if the text after point matches the regular expression given as
13430 The rest of the body of the loop looks difficult at first, but makes
13431 sense as you come to understand it.
13434 First consider what happens if there is a fill prefix:
13438 (if fill-prefix-regexp
13439 ;; There is a fill prefix; it overrides parstart;
13440 ;; we go forward line by line
13441 (while (and (not (eobp))
13442 (progn (move-to-left-margin) (not (eobp)))
13443 (not (looking-at parsep))
13444 (looking-at fill-prefix-regexp))
13450 This expression moves point forward line by line so long
13451 as four conditions are true:
13455 Point is not at the end of the buffer.
13458 We can move to the left margin of the text and are
13459 not at the end of the buffer.
13462 The text following point does not separate paragraphs.
13465 The pattern following point is the fill prefix regular expression.
13468 The last condition may be puzzling, until you remember that point was
13469 moved to the beginning of the line early in the @code{forward-paragraph}
13470 function. This means that if the text has a fill prefix, the
13471 @code{looking-at} function will see it.
13474 Consider what happens when there is no fill prefix.
13478 (while (and (re-search-forward sp-parstart nil 1)
13479 (progn (setq start (match-beginning 0))
13482 (progn (move-to-left-margin)
13483 (not (looking-at parsep)))
13484 (or (not (looking-at parstart))
13485 (and use-hard-newlines
13486 (not (get-text-property (1- start) 'hard)))))
13492 This @code{while} loop has us searching forward for
13493 @code{sp-parstart}, which is the combination of possible whitespace
13494 with a the local value of the start of a paragraph or of a paragraph
13495 separator. (The latter two are within an expression starting
13496 @code{\(?:} so that they are not referenced by the
13497 @code{match-beginning} function.)
13500 The two expressions,
13504 (setq start (match-beginning 0))
13510 mean go to the start of the text matched by the regular expression
13513 The @code{(match-beginning 0)} expression is new. It returns a number
13514 specifying the location of the start of the text that was matched by
13517 The @code{match-beginning} function is used here because of a
13518 characteristic of a forward search: a successful forward search,
13519 regardless of whether it is a plain search or a regular expression
13520 search, moves point to the end of the text that is found. In this
13521 case, a successful search moves point to the end of the pattern for
13522 @code{sp-parstart}.
13524 However, we want to put point at the end of the current paragraph, not
13525 somewhere else. Indeed, since the search possibly includes the
13526 paragraph separator, point may end up at the beginning of the next one
13527 unless we use an expression that includes @code{match-beginning}.
13529 @findex match-beginning
13530 When given an argument of 0, @code{match-beginning} returns the
13531 position that is the start of the text matched by the most recent
13532 search. In this case, the most recent search looks for
13533 @code{sp-parstart}. The @code{(match-beginning 0)} expression returns
13534 the beginning position of that pattern, rather than the end position
13537 (Incidentally, when passed a positive number as an argument, the
13538 @code{match-beginning} function returns the location of point at that
13539 parenthesized expression in the last search unless that parenthesized
13540 expression begins with @code{\(?:}. I don't know why @code{\(?:}
13541 appears here since the argument is 0.)
13544 The last expression when there is no fill prefix is
13548 (if (< (point) (point-max))
13549 (goto-char start))))
13554 This says that if there is no fill prefix and if we are not at the
13555 end, point should move to the beginning of whatever was found by the
13556 regular expression search for @code{sp-parstart}.
13558 The full definition for the @code{forward-paragraph} function not only
13559 includes code for going forwards, but also code for going backwards.
13561 If you are reading this inside of GNU Emacs and you want to see the
13562 whole function, you can type @kbd{C-h f} (@code{describe-function})
13563 and the name of the function. This gives you the function
13564 documentation and the name of the library containing the function's
13565 source. Place point over the name of the library and press the RET
13566 key; you will be taken directly to the source. (Be sure to install
13567 your sources! Without them, you are like a person who tries to drive
13568 a car with his eyes shut!)
13570 @node etags, Regexp Review, forward-paragraph, Regexp Search
13571 @section Create Your Own @file{TAGS} File
13573 @cindex @file{TAGS} file, create own
13575 Besides @kbd{C-h f} (@code{describe-function}), another way to see the
13576 source of a function is to type @kbd{M-.} (@code{find-tag}) and the
13577 name of the function when prompted for it. This is a good habit to
13578 get into. The @kbd{M-.} (@code{find-tag}) command takes you directly
13579 to the source for a function, variable, or node. The function depends
13580 on tags tables to tell it where to go.
13582 If the @code{find-tag} function first asks you for the name of a
13583 @file{TAGS} table, give it the name of a @file{TAGS} file such as
13584 @file{/usr/local/src/emacs/src/TAGS}. (The exact path to your
13585 @file{TAGS} file depends on how your copy of Emacs was installed. I
13586 just told you the location that provides both my C and my Emacs Lisp
13589 You can also create your own @file{TAGS} file for directories that
13592 You often need to build and install tags tables yourself. They are
13593 not built automatically. A tags table is called a @file{TAGS} file;
13594 the name is in upper case letters.
13596 You can create a @file{TAGS} file by calling the @code{etags} program
13597 that comes as a part of the Emacs distribution. Usually, @code{etags}
13598 is compiled and installed when Emacs is built. (@code{etags} is not
13599 an Emacs Lisp function or a part of Emacs; it is a C program.)
13602 To create a @file{TAGS} file, first switch to the directory in which
13603 you want to create the file. In Emacs you can do this with the
13604 @kbd{M-x cd} command, or by visiting a file in the directory, or by
13605 listing the directory with @kbd{C-x d} (@code{dired}). Then run the
13606 compile command, with @w{@code{etags *.el}} as the command to execute
13609 M-x compile RET etags *.el RET
13613 to create a @file{TAGS} file for Emacs Lisp.
13615 For example, if you have a large number of files in your
13616 @file{~/emacs} directory, as I do---I have 137 @file{.el} files in it,
13617 of which I load 12---you can create a @file{TAGS} file for the Emacs
13618 Lisp files in that directory.
13621 The @code{etags} program takes all the usual shell `wildcards'. For
13622 example, if you have two directories for which you want a single
13623 @file{TAGS} file, type @w{@code{etags *.el ../elisp/*.el}}, where
13624 @file{../elisp/} is the second directory:
13627 M-x compile RET etags *.el ../elisp/*.el RET
13634 M-x compile RET etags --help RET
13638 to see a list of the options accepted by @code{etags} as well as a
13639 list of supported languages.
13641 The @code{etags} program handles more than 20 languages, including
13642 Emacs Lisp, Common Lisp, Scheme, C, C++, Ada, Fortran, HTML, Java,
13643 LaTeX, Pascal, Perl, Postscript, Python, TeX, Texinfo, makefiles, and
13644 most assemblers. The program has no switches for specifying the
13645 language; it recognizes the language in an input file according to its
13646 file name and contents.
13648 @file{etags} is very helpful when you are writing code yourself and
13649 want to refer back to functions you have already written. Just run
13650 @code{etags} again at intervals as you write new functions, so they
13651 become part of the @file{TAGS} file.
13653 If you think an appropriate @file{TAGS} file already exists for what
13654 you want, but do not know where it is, you can use the @code{locate}
13655 program to attempt to find it.
13657 Type @w{@kbd{M-x locate @key{RET} TAGS @key{RET}}} and Emacs will list
13658 for you the full path names of all your @file{TAGS} files. On my
13659 system, this command lists 34 @file{TAGS} files. On the other hand, a
13660 `plain vanilla' system I recently installed did not contain any
13663 If the tags table you want has been created, you can use the @code{M-x
13664 visit-tags-table} command to specify it. Otherwise, you will need to
13665 create the tag table yourself and then use @code{M-x
13668 @subsubheading Building Tags in the Emacs sources
13669 @cindex Building Tags in the Emacs sources
13670 @cindex Tags in the Emacs sources
13673 The GNU Emacs sources come with a @file{Makefile} that contains a
13674 sophisticated @code{etags} command that creates, collects, and merges
13675 tags tables from all over the Emacs sources and puts the information
13676 into one @file{TAGS} file in the @file{src/} directory. (The
13677 @file{src/} directory is below the top level of your Emacs directory.)
13680 To build this @file{TAGS} file, go to the top level of your Emacs
13681 source directory and run the compile command @code{make tags}:
13684 M-x compile RET make tags RET
13688 (The @code{make tags} command works well with the GNU Emacs sources,
13689 as well as with some other source packages.)
13691 For more information, see @ref{Tags, , Tag Tables, emacs, The GNU Emacs
13694 @node Regexp Review, re-search Exercises, etags, Regexp Search
13695 @comment node-name, next, previous, up
13698 Here is a brief summary of some recently introduced functions.
13702 Repeatedly evaluate the body of the expression so long as the first
13703 element of the body tests true. Then return @code{nil}. (The
13704 expression is evaluated only for its side effects.)
13713 (insert (format "foo is %d.\n" foo))
13714 (setq foo (1- foo))))
13716 @result{} foo is 2.
13723 (The @code{insert} function inserts its arguments at point; the
13724 @code{format} function returns a string formatted from its arguments
13725 the way @code{message} formats its arguments; @code{\n} produces a new
13728 @item re-search-forward
13729 Search for a pattern, and if the pattern is found, move point to rest
13733 Takes four arguments, like @code{search-forward}:
13737 A regular expression that specifies the pattern to search for.
13738 (Remember to put quotation marks around this argument!)
13741 Optionally, the limit of the search.
13744 Optionally, what to do if the search fails, return @code{nil} or an
13748 Optionally, how many times to repeat the search; if negative, the
13749 search goes backwards.
13753 Bind some variables locally to particular values,
13754 and then evaluate the remaining arguments, returning the value of the
13755 last one. While binding the local variables, use the local values of
13756 variables bound earlier, if any.
13765 (message "`bar' is %d." bar))
13766 @result{} `bar' is 21.
13770 @item match-beginning
13771 Return the position of the start of the text found by the last regular
13775 Return @code{t} for true if the text after point matches the argument,
13776 which should be a regular expression.
13779 Return @code{t} for true if point is at the end of the accessible part
13780 of a buffer. The end of the accessible part is the end of the buffer
13781 if the buffer is not narrowed; it is the end of the narrowed part if
13782 the buffer is narrowed.
13786 @node re-search Exercises, , Regexp Review, Regexp Search
13787 @section Exercises with @code{re-search-forward}
13791 Write a function to search for a regular expression that matches two
13792 or more blank lines in sequence.
13795 Write a function to search for duplicated words, such as `the the'.
13796 @xref{Regexps, , Syntax of Regular Expressions, emacs, The GNU Emacs
13797 Manual}, for information on how to write a regexp (a regular
13798 expression) to match a string that is composed of two identical
13799 halves. You can devise several regexps; some are better than others.
13800 The function I use is described in an appendix, along with several
13801 regexps. @xref{the-the, , @code{the-the} Duplicated Words Function}.
13804 @node Counting Words, Words in a defun, Regexp Search, Top
13805 @chapter Counting: Repetition and Regexps
13806 @cindex Repetition for word counting
13807 @cindex Regular expressions for word counting
13809 Repetition and regular expression searches are powerful tools that you
13810 often use when you write code in Emacs Lisp. This chapter illustrates
13811 the use of regular expression searches through the construction of
13812 word count commands using @code{while} loops and recursion.
13815 * Why Count Words::
13816 * count-words-region:: Use a regexp, but find a problem.
13817 * recursive-count-words:: Start with case of no words in region.
13818 * Counting Exercise::
13821 @node Why Count Words, count-words-region, Counting Words, Counting Words
13823 @unnumberedsec Counting words
13826 The standard Emacs distribution contains a function for counting the
13827 number of lines within a region. However, there is no corresponding
13828 function for counting words.
13830 Certain types of writing ask you to count words. Thus, if you write
13831 an essay, you may be limited to 800 words; if you write a novel, you
13832 may discipline yourself to write 1000 words a day. It seems odd to me
13833 that Emacs lacks a word count command. Perhaps people use Emacs
13834 mostly for code or types of documentation that do not require word
13835 counts; or perhaps they restrict themselves to the operating system
13836 word count command, @code{wc}. Alternatively, people may follow
13837 the publishers' convention and compute a word count by dividing the
13838 number of characters in a document by five. In any event, here are
13839 commands to count words.
13841 @node count-words-region, recursive-count-words, Why Count Words, Counting Words
13842 @comment node-name, next, previous, up
13843 @section The @code{count-words-region} Function
13844 @findex count-words-region
13846 A word count command could count words in a line, paragraph, region,
13847 or buffer. What should the command cover? You could design the
13848 command to count the number of words in a complete buffer. However,
13849 the Emacs tradition encourages flexibility---you may want to count
13850 words in just a section, rather than all of a buffer. So it makes
13851 more sense to design the command to count the number of words in a
13852 region. Once you have a @code{count-words-region} command, you can,
13853 if you wish, count words in a whole buffer by marking it with
13854 @w{@kbd{C-x h}} (@code{mark-whole-buffer}).
13856 Clearly, counting words is a repetitive act: starting from the
13857 beginning of the region, you count the first word, then the second
13858 word, then the third word, and so on, until you reach the end of the
13859 region. This means that word counting is ideally suited to recursion
13860 or to a @code{while} loop.
13863 * Design count-words-region:: The definition using a @code{while} loop.
13864 * Whitespace Bug:: The Whitespace Bug in @code{count-words-region}.
13867 @node Design count-words-region, Whitespace Bug, count-words-region, count-words-region
13869 @unnumberedsubsec Designing @code{count-words-region}
13872 First, we will implement the word count command with a @code{while}
13873 loop, then with recursion. The command will, of course, be
13877 The template for an interactive function definition is, as always:
13881 (defun @var{name-of-function} (@var{argument-list})
13882 "@var{documentation}@dots{}"
13883 (@var{interactive-expression}@dots{})
13888 What we need to do is fill in the slots.
13890 The name of the function should be self-explanatory and similar to the
13891 existing @code{count-lines-region} name. This makes the name easier
13892 to remember. @code{count-words-region} is a good choice.
13894 The function counts words within a region. This means that the
13895 argument list must contain symbols that are bound to the two
13896 positions, the beginning and end of the region. These two positions
13897 can be called @samp{beginning} and @samp{end} respectively. The first
13898 line of the documentation should be a single sentence, since that is
13899 all that is printed as documentation by a command such as
13900 @code{apropos}. The interactive expression will be of the form
13901 @samp{(interactive "r")}, since that will cause Emacs to pass the
13902 beginning and end of the region to the function's argument list. All
13905 The body of the function needs to be written to do three tasks:
13906 first, to set up conditions under which the @code{while} loop can
13907 count words, second, to run the @code{while} loop, and third, to send
13908 a message to the user.
13910 When a user calls @code{count-words-region}, point may be at the
13911 beginning or the end of the region. However, the counting process
13912 must start at the beginning of the region. This means we will want
13913 to put point there if it is not already there. Executing
13914 @code{(goto-char beginning)} ensures this. Of course, we will want to
13915 return point to its expected position when the function finishes its
13916 work. For this reason, the body must be enclosed in a
13917 @code{save-excursion} expression.
13919 The central part of the body of the function consists of a
13920 @code{while} loop in which one expression jumps point forward word by
13921 word, and another expression counts those jumps. The true-or-false-test
13922 of the @code{while} loop should test true so long as point should jump
13923 forward, and false when point is at the end of the region.
13925 We could use @code{(forward-word 1)} as the expression for moving point
13926 forward word by word, but it is easier to see what Emacs identifies as a
13927 `word' if we use a regular expression search.
13929 A regular expression search that finds the pattern for which it is
13930 searching leaves point after the last character matched. This means
13931 that a succession of successful word searches will move point forward
13934 As a practical matter, we want the regular expression search to jump
13935 over whitespace and punctuation between words as well as over the
13936 words themselves. A regexp that refuses to jump over interword
13937 whitespace would never jump more than one word! This means that
13938 the regexp should include the whitespace and punctuation that follows
13939 a word, if any, as well as the word itself. (A word may end a buffer
13940 and not have any following whitespace or punctuation, so that part of
13941 the regexp must be optional.)
13943 Thus, what we want for the regexp is a pattern defining one or more
13944 word constituent characters followed, optionally, by one or more
13945 characters that are not word constituents. The regular expression for
13953 The buffer's syntax table determines which characters are and are not
13954 word constituents. (@xref{Syntax, , What Constitutes a Word or
13955 Symbol?}, for more about syntax. Also, see @ref{Syntax, Syntax, The
13956 Syntax Table, emacs, The GNU Emacs Manual}, and @ref{Syntax Tables, ,
13957 Syntax Tables, elisp, The GNU Emacs Lisp Reference Manual}.)
13960 The search expression looks like this:
13963 (re-search-forward "\\w+\\W*")
13967 (Note that paired backslashes precede the @samp{w} and @samp{W}. A
13968 single backslash has special meaning to the Emacs Lisp interpreter.
13969 It indicates that the following character is interpreted differently
13970 than usual. For example, the two characters, @samp{\n}, stand for
13971 @samp{newline}, rather than for a backslash followed by @samp{n}. Two
13972 backslashes in a row stand for an ordinary, `unspecial' backslash, so
13973 Emacs Lisp interpreter ends of seeing a single backslash followed by a
13974 letter. So it discovers the letter is special.)
13976 We need a counter to count how many words there are; this variable
13977 must first be set to 0 and then incremented each time Emacs goes
13978 around the @code{while} loop. The incrementing expression is simply:
13981 (setq count (1+ count))
13984 Finally, we want to tell the user how many words there are in the
13985 region. The @code{message} function is intended for presenting this
13986 kind of information to the user. The message has to be phrased so
13987 that it reads properly regardless of how many words there are in the
13988 region: we don't want to say that ``there are 1 words in the region''.
13989 The conflict between singular and plural is ungrammatical. We can
13990 solve this problem by using a conditional expression that evaluates
13991 different messages depending on the number of words in the region.
13992 There are three possibilities: no words in the region, one word in the
13993 region, and more than one word. This means that the @code{cond}
13994 special form is appropriate.
13997 All this leads to the following function definition:
14001 ;;; @r{First version; has bugs!}
14002 (defun count-words-region (beginning end)
14003 "Print number of words in the region.
14004 Words are defined as at least one word-constituent
14005 character followed by at least one character that
14006 is not a word-constituent. The buffer's syntax
14007 table determines which characters these are."
14009 (message "Counting words in region ... ")
14013 ;;; @r{1. Set up appropriate conditions.}
14015 (goto-char beginning)
14020 ;;; @r{2. Run the} while @r{loop.}
14021 (while (< (point) end)
14022 (re-search-forward "\\w+\\W*")
14023 (setq count (1+ count)))
14027 ;;; @r{3. Send a message to the user.}
14028 (cond ((zerop count)
14030 "The region does NOT have any words."))
14033 "The region has 1 word."))
14036 "The region has %d words." count))))))
14041 As written, the function works, but not in all circumstances.
14043 @node Whitespace Bug, , Design count-words-region, count-words-region
14044 @comment node-name, next, previous, up
14045 @subsection The Whitespace Bug in @code{count-words-region}
14047 The @code{count-words-region} command described in the preceding
14048 section has two bugs, or rather, one bug with two manifestations.
14049 First, if you mark a region containing only whitespace in the middle
14050 of some text, the @code{count-words-region} command tells you that the
14051 region contains one word! Second, if you mark a region containing
14052 only whitespace at the end of the buffer or the accessible portion of
14053 a narrowed buffer, the command displays an error message that looks
14057 Search failed: "\\w+\\W*"
14060 If you are reading this in Info in GNU Emacs, you can test for these
14063 First, evaluate the function in the usual manner to install it.
14065 Here is a copy of the definition. Place your cursor after the closing
14066 parenthesis and type @kbd{C-x C-e} to install it.
14070 ;; @r{First version; has bugs!}
14071 (defun count-words-region (beginning end)
14072 "Print number of words in the region.
14073 Words are defined as at least one word-constituent character followed
14074 by at least one character that is not a word-constituent. The buffer's
14075 syntax table determines which characters these are."
14079 (message "Counting words in region ... ")
14083 ;;; @r{1. Set up appropriate conditions.}
14085 (goto-char beginning)
14090 ;;; @r{2. Run the} while @r{loop.}
14091 (while (< (point) end)
14092 (re-search-forward "\\w+\\W*")
14093 (setq count (1+ count)))
14097 ;;; @r{3. Send a message to the user.}
14098 (cond ((zerop count)
14099 (message "The region does NOT have any words."))
14100 ((= 1 count) (message "The region has 1 word."))
14101 (t (message "The region has %d words." count))))))
14107 If you wish, you can also install this keybinding by evaluating it:
14110 (global-set-key "\C-c=" 'count-words-region)
14113 To conduct the first test, set mark and point to the beginning and end
14114 of the following line and then type @kbd{C-c =} (or @kbd{M-x
14115 count-words-region} if you have not bound @kbd{C-c =}):
14122 Emacs will tell you, correctly, that the region has three words.
14124 Repeat the test, but place mark at the beginning of the line and place
14125 point just @emph{before} the word @samp{one}. Again type the command
14126 @kbd{C-c =} (or @kbd{M-x count-words-region}). Emacs should tell you
14127 that the region has no words, since it is composed only of the
14128 whitespace at the beginning of the line. But instead Emacs tells you
14129 that the region has one word!
14131 For the third test, copy the sample line to the end of the
14132 @file{*scratch*} buffer and then type several spaces at the end of the
14133 line. Place mark right after the word @samp{three} and point at the
14134 end of line. (The end of the line will be the end of the buffer.)
14135 Type @kbd{C-c =} (or @kbd{M-x count-words-region}) as you did before.
14136 Again, Emacs should tell you that the region has no words, since it is
14137 composed only of the whitespace at the end of the line. Instead,
14138 Emacs displays an error message saying @samp{Search failed}.
14140 The two bugs stem from the same problem.
14142 Consider the first manifestation of the bug, in which the command
14143 tells you that the whitespace at the beginning of the line contains
14144 one word. What happens is this: The @code{M-x count-words-region}
14145 command moves point to the beginning of the region. The @code{while}
14146 tests whether the value of point is smaller than the value of
14147 @code{end}, which it is. Consequently, the regular expression search
14148 looks for and finds the first word. It leaves point after the word.
14149 @code{count} is set to one. The @code{while} loop repeats; but this
14150 time the value of point is larger than the value of @code{end}, the
14151 loop is exited; and the function displays a message saying the number
14152 of words in the region is one. In brief, the regular expression
14153 search looks for and finds the word even though it is outside
14156 In the second manifestation of the bug, the region is whitespace at
14157 the end of the buffer. Emacs says @samp{Search failed}. What happens
14158 is that the true-or-false-test in the @code{while} loop tests true, so
14159 the search expression is executed. But since there are no more words
14160 in the buffer, the search fails.
14162 In both manifestations of the bug, the search extends or attempts to
14163 extend outside of the region.
14165 The solution is to limit the search to the region---this is a fairly
14166 simple action, but as you may have come to expect, it is not quite as
14167 simple as you might think.
14169 As we have seen, the @code{re-search-forward} function takes a search
14170 pattern as its first argument. But in addition to this first,
14171 mandatory argument, it accepts three optional arguments. The optional
14172 second argument bounds the search. The optional third argument, if
14173 @code{t}, causes the function to return @code{nil} rather than signal
14174 an error if the search fails. The optional fourth argument is a
14175 repeat count. (In Emacs, you can see a function's documentation by
14176 typing @kbd{C-h f}, the name of the function, and then @key{RET}.)
14178 In the @code{count-words-region} definition, the value of the end of
14179 the region is held by the variable @code{end} which is passed as an
14180 argument to the function. Thus, we can add @code{end} as an argument
14181 to the regular expression search expression:
14184 (re-search-forward "\\w+\\W*" end)
14187 However, if you make only this change to the @code{count-words-region}
14188 definition and then test the new version of the definition on a
14189 stretch of whitespace, you will receive an error message saying
14190 @samp{Search failed}.
14192 What happens is this: the search is limited to the region, and fails
14193 as you expect because there are no word-constituent characters in the
14194 region. Since it fails, we receive an error message. But we do not
14195 want to receive an error message in this case; we want to receive the
14196 message that "The region does NOT have any words."
14198 The solution to this problem is to provide @code{re-search-forward}
14199 with a third argument of @code{t}, which causes the function to return
14200 @code{nil} rather than signal an error if the search fails.
14202 However, if you make this change and try it, you will see the message
14203 ``Counting words in region ... '' and @dots{} you will keep on seeing
14204 that message @dots{}, until you type @kbd{C-g} (@code{keyboard-quit}).
14206 Here is what happens: the search is limited to the region, as before,
14207 and it fails because there are no word-constituent characters in the
14208 region, as expected. Consequently, the @code{re-search-forward}
14209 expression returns @code{nil}. It does nothing else. In particular,
14210 it does not move point, which it does as a side effect if it finds the
14211 search target. After the @code{re-search-forward} expression returns
14212 @code{nil}, the next expression in the @code{while} loop is evaluated.
14213 This expression increments the count. Then the loop repeats. The
14214 true-or-false-test tests true because the value of point is still less
14215 than the value of end, since the @code{re-search-forward} expression
14216 did not move point. @dots{} and the cycle repeats @dots{}
14218 The @code{count-words-region} definition requires yet another
14219 modification, to cause the true-or-false-test of the @code{while} loop
14220 to test false if the search fails. Put another way, there are two
14221 conditions that must be satisfied in the true-or-false-test before the
14222 word count variable is incremented: point must still be within the
14223 region and the search expression must have found a word to count.
14225 Since both the first condition and the second condition must be true
14226 together, the two expressions, the region test and the search
14227 expression, can be joined with an @code{and} special form and embedded in
14228 the @code{while} loop as the true-or-false-test, like this:
14231 (and (< (point) end) (re-search-forward "\\w+\\W*" end t))
14234 @c colon in printed section title causes problem in Info cross reference
14235 @c also trouble with an overfull hbox
14238 (For information about @code{and}, see
14239 @ref{kill-new function, , The @code{kill-new} function}.)
14243 (@xref{kill-new function, , The @code{kill-new} function}, for
14244 information about @code{and}.)
14247 The @code{re-search-forward} expression returns @code{t} if the search
14248 succeeds and as a side effect moves point. Consequently, as words are
14249 found, point is moved through the region. When the search expression
14250 fails to find another word, or when point reaches the end of the
14251 region, the true-or-false-test tests false, the @code{while} loop
14252 exits, and the @code{count-words-region} function displays one or
14253 other of its messages.
14255 After incorporating these final changes, the @code{count-words-region}
14256 works without bugs (or at least, without bugs that I have found!).
14257 Here is what it looks like:
14261 ;;; @r{Final version:} @code{while}
14262 (defun count-words-region (beginning end)
14263 "Print number of words in the region."
14265 (message "Counting words in region ... ")
14269 ;;; @r{1. Set up appropriate conditions.}
14272 (goto-char beginning)
14276 ;;; @r{2. Run the} while @r{loop.}
14277 (while (and (< (point) end)
14278 (re-search-forward "\\w+\\W*" end t))
14279 (setq count (1+ count)))
14283 ;;; @r{3. Send a message to the user.}
14284 (cond ((zerop count)
14286 "The region does NOT have any words."))
14289 "The region has 1 word."))
14292 "The region has %d words." count))))))
14296 @node recursive-count-words, Counting Exercise, count-words-region, Counting Words
14297 @comment node-name, next, previous, up
14298 @section Count Words Recursively
14299 @cindex Count words recursively
14300 @cindex Recursively counting words
14301 @cindex Words, counted recursively
14303 You can write the function for counting words recursively as well as
14304 with a @code{while} loop. Let's see how this is done.
14306 First, we need to recognize that the @code{count-words-region}
14307 function has three jobs: it sets up the appropriate conditions for
14308 counting to occur; it counts the words in the region; and it sends a
14309 message to the user telling how many words there are.
14311 If we write a single recursive function to do everything, we will
14312 receive a message for every recursive call. If the region contains 13
14313 words, we will receive thirteen messages, one right after the other.
14314 We don't want this! Instead, we must write two functions to do the
14315 job, one of which (the recursive function) will be used inside of the
14316 other. One function will set up the conditions and display the
14317 message; the other will return the word count.
14319 Let us start with the function that causes the message to be displayed.
14320 We can continue to call this @code{count-words-region}.
14322 This is the function that the user will call. It will be interactive.
14323 Indeed, it will be similar to our previous versions of this
14324 function, except that it will call @code{recursive-count-words} to
14325 determine how many words are in the region.
14328 We can readily construct a template for this function, based on our
14333 ;; @r{Recursive version; uses regular expression search}
14334 (defun count-words-region (beginning end)
14335 "@var{documentation}@dots{}"
14336 (@var{interactive-expression}@dots{})
14340 ;;; @r{1. Set up appropriate conditions.}
14341 (@var{explanatory message})
14342 (@var{set-up functions}@dots{}
14346 ;;; @r{2. Count the words.}
14347 @var{recursive call}
14351 ;;; @r{3. Send a message to the user.}
14352 @var{message providing word count}))
14356 The definition looks straightforward, except that somehow the count
14357 returned by the recursive call must be passed to the message
14358 displaying the word count. A little thought suggests that this can be
14359 done by making use of a @code{let} expression: we can bind a variable
14360 in the varlist of a @code{let} expression to the number of words in
14361 the region, as returned by the recursive call; and then the
14362 @code{cond} expression, using binding, can display the value to the
14365 Often, one thinks of the binding within a @code{let} expression as
14366 somehow secondary to the `primary' work of a function. But in this
14367 case, what you might consider the `primary' job of the function,
14368 counting words, is done within the @code{let} expression.
14371 Using @code{let}, the function definition looks like this:
14375 (defun count-words-region (beginning end)
14376 "Print number of words in the region."
14381 ;;; @r{1. Set up appropriate conditions.}
14382 (message "Counting words in region ... ")
14384 (goto-char beginning)
14388 ;;; @r{2. Count the words.}
14389 (let ((count (recursive-count-words end)))
14393 ;;; @r{3. Send a message to the user.}
14394 (cond ((zerop count)
14396 "The region does NOT have any words."))
14399 "The region has 1 word."))
14402 "The region has %d words." count))))))
14406 Next, we need to write the recursive counting function.
14408 A recursive function has at least three parts: the `do-again-test', the
14409 `next-step-expression', and the recursive call.
14411 The do-again-test determines whether the function will or will not be
14412 called again. Since we are counting words in a region and can use a
14413 function that moves point forward for every word, the do-again-test
14414 can check whether point is still within the region. The do-again-test
14415 should find the value of point and determine whether point is before,
14416 at, or after the value of the end of the region. We can use the
14417 @code{point} function to locate point. Clearly, we must pass the
14418 value of the end of the region to the recursive counting function as an
14421 In addition, the do-again-test should also test whether the search finds a
14422 word. If it does not, the function should not call itself again.
14424 The next-step-expression changes a value so that when the recursive
14425 function is supposed to stop calling itself, it stops. More
14426 precisely, the next-step-expression changes a value so that at the
14427 right time, the do-again-test stops the recursive function from
14428 calling itself again. In this case, the next-step-expression can be
14429 the expression that moves point forward, word by word.
14431 The third part of a recursive function is the recursive call.
14433 Somewhere, also, we also need a part that does the `work' of the
14434 function, a part that does the counting. A vital part!
14437 But already, we have an outline of the recursive counting function:
14441 (defun recursive-count-words (region-end)
14442 "@var{documentation}@dots{}"
14443 @var{do-again-test}
14444 @var{next-step-expression}
14445 @var{recursive call})
14449 Now we need to fill in the slots. Let's start with the simplest cases
14450 first: if point is at or beyond the end of the region, there cannot
14451 be any words in the region, so the function should return zero.
14452 Likewise, if the search fails, there are no words to count, so the
14453 function should return zero.
14455 On the other hand, if point is within the region and the search
14456 succeeds, the function should call itself again.
14459 Thus, the do-again-test should look like this:
14463 (and (< (point) region-end)
14464 (re-search-forward "\\w+\\W*" region-end t))
14468 Note that the search expression is part of the do-again-test---the
14469 function returns @code{t} if its search succeeds and @code{nil} if it
14470 fails. (@xref{Whitespace Bug, , The Whitespace Bug in
14471 @code{count-words-region}}, for an explanation of how
14472 @code{re-search-forward} works.)
14474 The do-again-test is the true-or-false test of an @code{if} clause.
14475 Clearly, if the do-again-test succeeds, the then-part of the @code{if}
14476 clause should call the function again; but if it fails, the else-part
14477 should return zero since either point is outside the region or the
14478 search failed because there were no words to find.
14480 But before considering the recursive call, we need to consider the
14481 next-step-expression. What is it? Interestingly, it is the search
14482 part of the do-again-test.
14484 In addition to returning @code{t} or @code{nil} for the
14485 do-again-test, @code{re-search-forward} moves point forward as a side
14486 effect of a successful search. This is the action that changes the
14487 value of point so that the recursive function stops calling itself
14488 when point completes its movement through the region. Consequently,
14489 the @code{re-search-forward} expression is the next-step-expression.
14492 In outline, then, the body of the @code{recursive-count-words}
14493 function looks like this:
14497 (if @var{do-again-test-and-next-step-combined}
14499 @var{recursive-call-returning-count}
14505 How to incorporate the mechanism that counts?
14507 If you are not used to writing recursive functions, a question like
14508 this can be troublesome. But it can and should be approached
14511 We know that the counting mechanism should be associated in some way
14512 with the recursive call. Indeed, since the next-step-expression moves
14513 point forward by one word, and since a recursive call is made for
14514 each word, the counting mechanism must be an expression that adds one
14515 to the value returned by a call to @code{recursive-count-words}.
14518 Consider several cases:
14522 If there are two words in the region, the function should return
14523 a value resulting from adding one to the value returned when it counts
14524 the first word, plus the number returned when it counts the remaining
14525 words in the region, which in this case is one.
14528 If there is one word in the region, the function should return
14529 a value resulting from adding one to the value returned when it counts
14530 that word, plus the number returned when it counts the remaining
14531 words in the region, which in this case is zero.
14534 If there are no words in the region, the function should return zero.
14537 From the sketch we can see that the else-part of the @code{if} returns
14538 zero for the case of no words. This means that the then-part of the
14539 @code{if} must return a value resulting from adding one to the value
14540 returned from a count of the remaining words.
14543 The expression will look like this, where @code{1+} is a function that
14544 adds one to its argument.
14547 (1+ (recursive-count-words region-end))
14551 The whole @code{recursive-count-words} function will then look like
14556 (defun recursive-count-words (region-end)
14557 "@var{documentation}@dots{}"
14559 ;;; @r{1. do-again-test}
14560 (if (and (< (point) region-end)
14561 (re-search-forward "\\w+\\W*" region-end t))
14565 ;;; @r{2. then-part: the recursive call}
14566 (1+ (recursive-count-words region-end))
14568 ;;; @r{3. else-part}
14574 Let's examine how this works:
14576 If there are no words in the region, the else part of the @code{if}
14577 expression is evaluated and consequently the function returns zero.
14579 If there is one word in the region, the value of point is less than
14580 the value of @code{region-end} and the search succeeds. In this case,
14581 the true-or-false-test of the @code{if} expression tests true, and the
14582 then-part of the @code{if} expression is evaluated. The counting
14583 expression is evaluated. This expression returns a value (which will
14584 be the value returned by the whole function) that is the sum of one
14585 added to the value returned by a recursive call.
14587 Meanwhile, the next-step-expression has caused point to jump over the
14588 first (and in this case only) word in the region. This means that
14589 when @code{(recursive-count-words region-end)} is evaluated a second
14590 time, as a result of the recursive call, the value of point will be
14591 equal to or greater than the value of region end. So this time,
14592 @code{recursive-count-words} will return zero. The zero will be added
14593 to one, and the original evaluation of @code{recursive-count-words}
14594 will return one plus zero, which is one, which is the correct amount.
14596 Clearly, if there are two words in the region, the first call to
14597 @code{recursive-count-words} returns one added to the value returned
14598 by calling @code{recursive-count-words} on a region containing the
14599 remaining word---that is, it adds one to one, producing two, which is
14600 the correct amount.
14602 Similarly, if there are three words in the region, the first call to
14603 @code{recursive-count-words} returns one added to the value returned
14604 by calling @code{recursive-count-words} on a region containing the
14605 remaining two words---and so on and so on.
14609 With full documentation the two functions look like this:
14613 The recursive function:
14615 @findex recursive-count-words
14618 (defun recursive-count-words (region-end)
14619 "Number of words between point and REGION-END."
14623 ;;; @r{1. do-again-test}
14624 (if (and (< (point) region-end)
14625 (re-search-forward "\\w+\\W*" region-end t))
14629 ;;; @r{2. then-part: the recursive call}
14630 (1+ (recursive-count-words region-end))
14632 ;;; @r{3. else-part}
14643 ;;; @r{Recursive version}
14644 (defun count-words-region (beginning end)
14645 "Print number of words in the region.
14649 Words are defined as at least one word-constituent
14650 character followed by at least one character that is
14651 not a word-constituent. The buffer's syntax table
14652 determines which characters these are."
14656 (message "Counting words in region ... ")
14658 (goto-char beginning)
14659 (let ((count (recursive-count-words end)))
14662 (cond ((zerop count)
14664 "The region does NOT have any words."))
14668 (message "The region has 1 word."))
14671 "The region has %d words." count))))))
14675 @node Counting Exercise, , recursive-count-words, Counting Words
14676 @section Exercise: Counting Punctuation
14678 Using a @code{while} loop, write a function to count the number of
14679 punctuation marks in a region---period, comma, semicolon, colon,
14680 exclamation mark, and question mark. Do the same using recursion.
14682 @node Words in a defun, Readying a Graph, Counting Words, Top
14683 @chapter Counting Words in a @code{defun}
14684 @cindex Counting words in a @code{defun}
14685 @cindex Word counting in a @code{defun}
14687 Our next project is to count the number of words in a function
14688 definition. Clearly, this can be done using some variant of
14689 @code{count-word-region}. @xref{Counting Words, , Counting Words:
14690 Repetition and Regexps}. If we are just going to count the words in
14691 one definition, it is easy enough to mark the definition with the
14692 @kbd{C-M-h} (@code{mark-defun}) command, and then call
14693 @code{count-word-region}.
14695 However, I am more ambitious: I want to count the words and symbols in
14696 every definition in the Emacs sources and then print a graph that
14697 shows how many functions there are of each length: how many contain 40
14698 to 49 words or symbols, how many contain 50 to 59 words or symbols,
14699 and so on. I have often been curious how long a typical function is,
14700 and this will tell.
14703 * Divide and Conquer::
14704 * Words and Symbols:: What to count?
14705 * Syntax:: What constitutes a word or symbol?
14706 * count-words-in-defun:: Very like @code{count-words}.
14707 * Several defuns:: Counting several defuns in a file.
14708 * Find a File:: Do you want to look at a file?
14709 * lengths-list-file:: A list of the lengths of many definitions.
14710 * Several files:: Counting in definitions in different files.
14711 * Several files recursively:: Recursively counting in different files.
14712 * Prepare the data:: Prepare the data for display in a graph.
14715 @node Divide and Conquer, Words and Symbols, Words in a defun, Words in a defun
14717 @unnumberedsec Divide and Conquer
14720 Described in one phrase, the histogram project is daunting; but
14721 divided into numerous small steps, each of which we can take one at a
14722 time, the project becomes less fearsome. Let us consider what the
14727 First, write a function to count the words in one definition. This
14728 includes the problem of handling symbols as well as words.
14731 Second, write a function to list the numbers of words in each function
14732 in a file. This function can use the @code{count-words-in-defun}
14736 Third, write a function to list the numbers of words in each function
14737 in each of several files. This entails automatically finding the
14738 various files, switching to them, and counting the words in the
14739 definitions within them.
14742 Fourth, write a function to convert the list of numbers that we
14743 created in step three to a form that will be suitable for printing as
14747 Fifth, write a function to print the results as a graph.
14750 This is quite a project! But if we take each step slowly, it will not
14753 @node Words and Symbols, Syntax, Divide and Conquer, Words in a defun
14754 @section What to Count?
14755 @cindex Words and symbols in defun
14757 When we first start thinking about how to count the words in a
14758 function definition, the first question is (or ought to be) what are
14759 we going to count? When we speak of `words' with respect to a Lisp
14760 function definition, we are actually speaking, in large part, of
14761 `symbols'. For example, the following @code{multiply-by-seven}
14762 function contains the five symbols @code{defun},
14763 @code{multiply-by-seven}, @code{number}, @code{*}, and @code{7}. In
14764 addition, in the documentation string, it contains the four words
14765 @samp{Multiply}, @samp{NUMBER}, @samp{by}, and @samp{seven}. The
14766 symbol @samp{number} is repeated, so the definition contains a total
14767 of ten words and symbols.
14771 (defun multiply-by-seven (number)
14772 "Multiply NUMBER by seven."
14778 However, if we mark the @code{multiply-by-seven} definition with
14779 @kbd{C-M-h} (@code{mark-defun}), and then call
14780 @code{count-words-region} on it, we will find that
14781 @code{count-words-region} claims the definition has eleven words, not
14782 ten! Something is wrong!
14784 The problem is twofold: @code{count-words-region} does not count the
14785 @samp{*} as a word, and it counts the single symbol,
14786 @code{multiply-by-seven}, as containing three words. The hyphens are
14787 treated as if they were interword spaces rather than intraword
14788 connectors: @samp{multiply-by-seven} is counted as if it were written
14789 @samp{multiply by seven}.
14791 The cause of this confusion is the regular expression search within
14792 the @code{count-words-region} definition that moves point forward word
14793 by word. In the canonical version of @code{count-words-region}, the
14801 This regular expression is a pattern defining one or more word
14802 constituent characters possibly followed by one or more characters
14803 that are not word constituents. What is meant by `word constituent
14804 characters' brings us to the issue of syntax, which is worth a section
14807 @node Syntax, count-words-in-defun, Words and Symbols, Words in a defun
14808 @section What Constitutes a Word or Symbol?
14809 @cindex Syntax categories and tables
14811 Emacs treats different characters as belonging to different
14812 @dfn{syntax categories}. For example, the regular expression,
14813 @samp{\\w+}, is a pattern specifying one or more @emph{word
14814 constituent} characters. Word constituent characters are members of
14815 one syntax category. Other syntax categories include the class of
14816 punctuation characters, such as the period and the comma, and the
14817 class of whitespace characters, such as the blank space and the tab
14818 character. (For more information, see @ref{Syntax, Syntax, The Syntax
14819 Table, emacs, The GNU Emacs Manual}, and @ref{Syntax Tables, , Syntax
14820 Tables, elisp, The GNU Emacs Lisp Reference Manual}.)
14822 Syntax tables specify which characters belong to which categories.
14823 Usually, a hyphen is not specified as a `word constituent character'.
14824 Instead, it is specified as being in the `class of characters that are
14825 part of symbol names but not words.' This means that the
14826 @code{count-words-region} function treats it in the same way it treats
14827 an interword white space, which is why @code{count-words-region}
14828 counts @samp{multiply-by-seven} as three words.
14830 There are two ways to cause Emacs to count @samp{multiply-by-seven} as
14831 one symbol: modify the syntax table or modify the regular expression.
14833 We could redefine a hyphen as a word constituent character by
14834 modifying the syntax table that Emacs keeps for each mode. This
14835 action would serve our purpose, except that a hyphen is merely the
14836 most common character within symbols that is not typically a word
14837 constituent character; there are others, too.
14839 Alternatively, we can redefine the regular expression used in the
14840 @code{count-words} definition so as to include symbols. This
14841 procedure has the merit of clarity, but the task is a little tricky.
14844 The first part is simple enough: the pattern must match ``at least one
14845 character that is a word or symbol constituent''. Thus:
14848 "\\(\\w\\|\\s_\\)+"
14852 The @samp{\\(} is the first part of the grouping construct that
14853 includes the @samp{\\w} and the @samp{\\s_} as alternatives, separated
14854 by the @samp{\\|}. The @samp{\\w} matches any word-constituent
14855 character and the @samp{\\s_} matches any character that is part of a
14856 symbol name but not a word-constituent character. The @samp{+}
14857 following the group indicates that the word or symbol constituent
14858 characters must be matched at least once.
14860 However, the second part of the regexp is more difficult to design.
14861 What we want is to follow the first part with ``optionally one or more
14862 characters that are not constituents of a word or symbol''. At first,
14863 I thought I could define this with the following:
14866 "\\(\\W\\|\\S_\\)*"
14870 The upper case @samp{W} and @samp{S} match characters that are
14871 @emph{not} word or symbol constituents. Unfortunately, this
14872 expression matches any character that is either not a word constituent
14873 or not a symbol constituent. This matches any character!
14875 I then noticed that every word or symbol in my test region was
14876 followed by white space (blank space, tab, or newline). So I tried
14877 placing a pattern to match one or more blank spaces after the pattern
14878 for one or more word or symbol constituents. This failed, too. Words
14879 and symbols are often separated by whitespace, but in actual code
14880 parentheses may follow symbols and punctuation may follow words. So
14881 finally, I designed a pattern in which the word or symbol constituents
14882 are followed optionally by characters that are not white space and
14883 then followed optionally by white space.
14886 Here is the full regular expression:
14889 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*"
14892 @node count-words-in-defun, Several defuns, Syntax, Words in a defun
14893 @section The @code{count-words-in-defun} Function
14894 @cindex Counting words in a @code{defun}
14896 We have seen that there are several ways to write a
14897 @code{count-word-region} function. To write a
14898 @code{count-words-in-defun}, we need merely adapt one of these
14901 The version that uses a @code{while} loop is easy to understand, so I
14902 am going to adapt that. Because @code{count-words-in-defun} will be
14903 part of a more complex program, it need not be interactive and it need
14904 not display a message but just return the count. These considerations
14905 simplify the definition a little.
14907 On the other hand, @code{count-words-in-defun} will be used within a
14908 buffer that contains function definitions. Consequently, it is
14909 reasonable to ask that the function determine whether it is called
14910 when point is within a function definition, and if it is, to return
14911 the count for that definition. This adds complexity to the
14912 definition, but saves us from needing to pass arguments to the
14916 These considerations lead us to prepare the following template:
14920 (defun count-words-in-defun ()
14921 "@var{documentation}@dots{}"
14922 (@var{set up}@dots{}
14923 (@var{while loop}@dots{})
14924 @var{return count})
14929 As usual, our job is to fill in the slots.
14933 We are presuming that this function will be called within a buffer
14934 containing function definitions. Point will either be within a
14935 function definition or not. For @code{count-words-in-defun} to work,
14936 point must move to the beginning of the definition, a counter must
14937 start at zero, and the counting loop must stop when point reaches the
14938 end of the definition.
14940 The @code{beginning-of-defun} function searches backwards for an
14941 opening delimiter such as a @samp{(} at the beginning of a line, and
14942 moves point to that position, or else to the limit of the search. In
14943 practice, this means that @code{beginning-of-defun} moves point to the
14944 beginning of an enclosing or preceding function definition, or else to
14945 the beginning of the buffer. We can use @code{beginning-of-defun} to
14946 place point where we wish to start.
14948 The @code{while} loop requires a counter to keep track of the words or
14949 symbols being counted. A @code{let} expression can be used to create
14950 a local variable for this purpose, and bind it to an initial value of zero.
14952 The @code{end-of-defun} function works like @code{beginning-of-defun}
14953 except that it moves point to the end of the definition.
14954 @code{end-of-defun} can be used as part of an expression that
14955 determines the position of the end of the definition.
14957 The set up for @code{count-words-in-defun} takes shape rapidly: first
14958 we move point to the beginning of the definition, then we create a
14959 local variable to hold the count, and finally, we record the position
14960 of the end of the definition so the @code{while} loop will know when to stop
14964 The code looks like this:
14968 (beginning-of-defun)
14970 (end (save-excursion (end-of-defun) (point))))
14975 The code is simple. The only slight complication is likely to concern
14976 @code{end}: it is bound to the position of the end of the definition
14977 by a @code{save-excursion} expression that returns the value of point
14978 after @code{end-of-defun} temporarily moves it to the end of the
14981 The second part of the @code{count-words-in-defun}, after the set up,
14982 is the @code{while} loop.
14984 The loop must contain an expression that jumps point forward word by
14985 word and symbol by symbol, and another expression that counts the
14986 jumps. The true-or-false-test for the @code{while} loop should test
14987 true so long as point should jump forward, and false when point is at
14988 the end of the definition. We have already redefined the regular
14989 expression for this (@pxref{Syntax}), so the loop is straightforward:
14993 (while (and (< (point) end)
14995 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*" end t)
14996 (setq count (1+ count)))
15000 The third part of the function definition returns the count of words
15001 and symbols. This part is the last expression within the body of the
15002 @code{let} expression, and can be, very simply, the local variable
15003 @code{count}, which when evaluated returns the count.
15006 Put together, the @code{count-words-in-defun} definition looks like this:
15008 @findex count-words-in-defun
15011 (defun count-words-in-defun ()
15012 "Return the number of words and symbols in a defun."
15013 (beginning-of-defun)
15015 (end (save-excursion (end-of-defun) (point))))
15019 (and (< (point) end)
15021 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*"
15023 (setq count (1+ count)))
15028 How to test this? The function is not interactive, but it is easy to
15029 put a wrapper around the function to make it interactive; we can use
15030 almost the same code as for the recursive version of
15031 @code{count-words-region}:
15035 ;;; @r{Interactive version.}
15036 (defun count-words-defun ()
15037 "Number of words and symbols in a function definition."
15040 "Counting words and symbols in function definition ... ")
15043 (let ((count (count-words-in-defun)))
15047 "The definition does NOT have any words or symbols."))
15052 "The definition has 1 word or symbol."))
15055 "The definition has %d words or symbols." count)))))
15061 Let's re-use @kbd{C-c =} as a convenient keybinding:
15064 (global-set-key "\C-c=" 'count-words-defun)
15067 Now we can try out @code{count-words-defun}: install both
15068 @code{count-words-in-defun} and @code{count-words-defun}, and set the
15069 keybinding, and then place the cursor within the following definition:
15073 (defun multiply-by-seven (number)
15074 "Multiply NUMBER by seven."
15081 Success! The definition has 10 words and symbols.
15083 The next problem is to count the numbers of words and symbols in
15084 several definitions within a single file.
15086 @node Several defuns, Find a File, count-words-in-defun, Words in a defun
15087 @section Count Several @code{defuns} Within a File
15089 A file such as @file{simple.el} may have a hundred or more function
15090 definitions within it. Our long term goal is to collect statistics on
15091 many files, but as a first step, our immediate goal is to collect
15092 statistics on one file.
15094 The information will be a series of numbers, each number being the
15095 length of a function definition. We can store the numbers in a list.
15097 We know that we will want to incorporate the information regarding one
15098 file with information about many other files; this means that the
15099 function for counting definition lengths within one file need only
15100 return the list of lengths. It need not and should not display any
15103 The word count commands contain one expression to jump point forward
15104 word by word and another expression to count the jumps. The function
15105 to return the lengths of definitions can be designed to work the same
15106 way, with one expression to jump point forward definition by
15107 definition and another expression to construct the lengths' list.
15109 This statement of the problem makes it elementary to write the
15110 function definition. Clearly, we will start the count at the
15111 beginning of the file, so the first command will be @code{(goto-char
15112 (point-min))}. Next, we start the @code{while} loop; and the
15113 true-or-false test of the loop can be a regular expression search for
15114 the next function definition---so long as the search succeeds, point
15115 is moved forward and then the body of the loop is evaluated. The body
15116 needs an expression that constructs the lengths' list. @code{cons},
15117 the list construction command, can be used to create the list. That
15118 is almost all there is to it.
15121 Here is what this fragment of code looks like:
15125 (goto-char (point-min))
15126 (while (re-search-forward "^(defun" nil t)
15128 (cons (count-words-in-defun) lengths-list)))
15132 What we have left out is the mechanism for finding the file that
15133 contains the function definitions.
15135 In previous examples, we either used this, the Info file, or we
15136 switched back and forth to some other buffer, such as the
15137 @file{*scratch*} buffer.
15139 Finding a file is a new process that we have not yet discussed.
15141 @node Find a File, lengths-list-file, Several defuns, Words in a defun
15142 @comment node-name, next, previous, up
15143 @section Find a File
15144 @cindex Find a File
15146 To find a file in Emacs, you use the @kbd{C-x C-f} (@code{find-file})
15147 command. This command is almost, but not quite right for the lengths
15151 Let's look at the source for @code{find-file}:
15155 (defun find-file (filename)
15156 "Edit file FILENAME.
15157 Switch to a buffer visiting file FILENAME,
15158 creating one if none already exists."
15159 (interactive "FFind file: ")
15160 (switch-to-buffer (find-file-noselect filename)))
15165 (The most recent version of the @code{find-file} function definition
15166 permits you to specify optional wildcards to visit multiple files; that
15167 makes the definition more complex and we will not discuss it here,
15168 since it is not relevant. You can see its source using either
15169 @kbd{M-.} (@code{find-tag}) or @kbd{C-h f} (@code{describe-function}).)
15173 (defun find-file (filename &optional wildcards)
15174 "Edit file FILENAME.
15175 Switch to a buffer visiting file FILENAME,
15176 creating one if none already exists.
15177 Interactively, the default if you just type RET is the current directory,
15178 but the visited file name is available through the minibuffer history:
15179 type M-n to pull it into the minibuffer.
15181 Interactively, or if WILDCARDS is non-nil in a call from Lisp,
15182 expand wildcards (if any) and visit multiple files. You can
15183 suppress wildcard expansion by setting `find-file-wildcards' to nil.
15185 To visit a file without any kind of conversion and without
15186 automatically choosing a major mode, use \\[find-file-literally]."
15187 (interactive (find-file-read-args "Find file: " nil))
15188 (let ((value (find-file-noselect filename nil nil wildcards)))
15190 (mapcar 'switch-to-buffer (nreverse value))
15191 (switch-to-buffer value))))
15194 The definition I am showing possesses short but complete documentation
15195 and an interactive specification that prompts you for a file name when
15196 you use the command interactively. The body of the definition
15197 contains two functions, @code{find-file-noselect} and
15198 @code{switch-to-buffer}.
15200 According to its documentation as shown by @kbd{C-h f} (the
15201 @code{describe-function} command), the @code{find-file-noselect}
15202 function reads the named file into a buffer and returns the buffer.
15203 (Its most recent version includes an optional wildcards argument,
15204 too, as well as another to read a file literally and an other you
15205 suppress warning messages. These optional arguments are irrelevant.)
15207 However, the @code{find-file-noselect} function does not select the
15208 buffer in which it puts the file. Emacs does not switch its attention
15209 (or yours if you are using @code{find-file-noselect}) to the selected
15210 buffer. That is what @code{switch-to-buffer} does: it switches the
15211 buffer to which Emacs attention is directed; and it switches the
15212 buffer displayed in the window to the new buffer. We have discussed
15213 buffer switching elsewhere. (@xref{Switching Buffers}.)
15215 In this histogram project, we do not need to display each file on the
15216 screen as the program determines the length of each definition within
15217 it. Instead of employing @code{switch-to-buffer}, we can work with
15218 @code{set-buffer}, which redirects the attention of the computer
15219 program to a different buffer but does not redisplay it on the screen.
15220 So instead of calling on @code{find-file} to do the job, we must write
15221 our own expression.
15223 The task is easy: use @code{find-file-noselect} and @code{set-buffer}.
15225 @node lengths-list-file, Several files, Find a File, Words in a defun
15226 @section @code{lengths-list-file} in Detail
15228 The core of the @code{lengths-list-file} function is a @code{while}
15229 loop containing a function to move point forward `defun by defun' and
15230 a function to count the number of words and symbols in each defun.
15231 This core must be surrounded by functions that do various other tasks,
15232 including finding the file, and ensuring that point starts out at the
15233 beginning of the file. The function definition looks like this:
15234 @findex lengths-list-file
15238 (defun lengths-list-file (filename)
15239 "Return list of definitions' lengths within FILE.
15240 The returned list is a list of numbers.
15241 Each number is the number of words or
15242 symbols in one function definition."
15245 (message "Working on `%s' ... " filename)
15247 (let ((buffer (find-file-noselect filename))
15249 (set-buffer buffer)
15250 (setq buffer-read-only t)
15252 (goto-char (point-min))
15253 (while (re-search-forward "^(defun" nil t)
15255 (cons (count-words-in-defun) lengths-list)))
15256 (kill-buffer buffer)
15262 The function is passed one argument, the name of the file on which it
15263 will work. It has four lines of documentation, but no interactive
15264 specification. Since people worry that a computer is broken if they
15265 don't see anything going on, the first line of the body is a
15268 The next line contains a @code{save-excursion} that returns Emacs'
15269 attention to the current buffer when the function completes. This is
15270 useful in case you embed this function in another function that
15271 presumes point is restored to the original buffer.
15273 In the varlist of the @code{let} expression, Emacs finds the file and
15274 binds the local variable @code{buffer} to the buffer containing the
15275 file. At the same time, Emacs creates @code{lengths-list} as a local
15278 Next, Emacs switches its attention to the buffer.
15280 In the following line, Emacs makes the buffer read-only. Ideally,
15281 this line is not necessary. None of the functions for counting words
15282 and symbols in a function definition should change the buffer.
15283 Besides, the buffer is not going to be saved, even if it were changed.
15284 This line is entirely the consequence of great, perhaps excessive,
15285 caution. The reason for the caution is that this function and those
15286 it calls work on the sources for Emacs and it is inconvenient if they
15287 are inadvertently modified. It goes without saying that I did not
15288 realize a need for this line until an experiment went awry and started
15289 to modify my Emacs source files @dots{}
15291 Next comes a call to widen the buffer if it is narrowed. This
15292 function is usually not needed---Emacs creates a fresh buffer if none
15293 already exists; but if a buffer visiting the file already exists Emacs
15294 returns that one. In this case, the buffer may be narrowed and must
15295 be widened. If we wanted to be fully `user-friendly', we would
15296 arrange to save the restriction and the location of point, but we
15299 The @code{(goto-char (point-min))} expression moves point to the
15300 beginning of the buffer.
15302 Then comes a @code{while} loop in which the `work' of the function is
15303 carried out. In the loop, Emacs determines the length of each
15304 definition and constructs a lengths' list containing the information.
15306 Emacs kills the buffer after working through it. This is to save
15307 space inside of Emacs. My version of GNU Emacs 19 contained over 300
15308 source files of interest; GNU Emacs 22 contains over a thousand source
15309 files. Another function will apply @code{lengths-list-file} to each
15312 Finally, the last expression within the @code{let} expression is the
15313 @code{lengths-list} variable; its value is returned as the value of
15314 the whole function.
15316 You can try this function by installing it in the usual fashion. Then
15317 place your cursor after the following expression and type @kbd{C-x
15318 C-e} (@code{eval-last-sexp}).
15320 @c !!! 22.1.1 lisp sources location here
15323 "/usr/local/share/emacs/22.1.1/lisp/emacs-lisp/debug.el")
15327 (You may need to change the pathname of the file; the one here is for
15328 GNU Emacs version 22.1.1. To change the expression, copy it to
15329 the @file{*scratch*} buffer and edit it.
15333 (Also, to see the full length of the list, rather than a truncated
15334 version, you may have to evaluate the following:
15337 (custom-set-variables '(eval-expression-print-length nil))
15341 (@xref{defcustom, , Specifying Variables using @code{defcustom}}.
15342 Then evaluate the @code{lengths-list-file} expression.)
15345 The lengths' list for @file{debug.el} takes less than a second to
15346 produce and looks like this in GNU Emacs 22:
15349 (83 113 105 144 289 22 30 97 48 89 25 52 52 88 28 29 77 49 43 290 232 587)
15353 (Using my old machine, the version 19 lengths' list for @file{debug.el}
15354 took seven seconds to produce and looked like this:
15357 (75 41 80 62 20 45 44 68 45 12 34 235)
15360 (The newer version of @file{debug.el} contains more defuns than the
15361 earlier one; and my new machine is much faster than the old one.)
15363 Note that the length of the last definition in the file is first in
15366 @node Several files, Several files recursively, lengths-list-file, Words in a defun
15367 @section Count Words in @code{defuns} in Different Files
15369 In the previous section, we created a function that returns a list of
15370 the lengths of each definition in a file. Now, we want to define a
15371 function to return a master list of the lengths of the definitions in
15374 Working on each of a list of files is a repetitious act, so we can use
15375 either a @code{while} loop or recursion.
15378 * lengths-list-many-files:: Return a list of the lengths of defuns.
15379 * append:: Attach one list to another.
15382 @node lengths-list-many-files, append, Several files, Several files
15384 @unnumberedsubsec Determine the lengths of @code{defuns}
15387 The design using a @code{while} loop is routine. The argument passed
15388 the function is a list of files. As we saw earlier (@pxref{Loop
15389 Example}), you can write a @code{while} loop so that the body of the
15390 loop is evaluated if such a list contains elements, but to exit the
15391 loop if the list is empty. For this design to work, the body of the
15392 loop must contain an expression that shortens the list each time the
15393 body is evaluated, so that eventually the list is empty. The usual
15394 technique is to set the value of the list to the value of the @sc{cdr}
15395 of the list each time the body is evaluated.
15398 The template looks like this:
15402 (while @var{test-whether-list-is-empty}
15404 @var{set-list-to-cdr-of-list})
15408 Also, we remember that a @code{while} loop returns @code{nil} (the
15409 result of evaluating the true-or-false-test), not the result of any
15410 evaluation within its body. (The evaluations within the body of the
15411 loop are done for their side effects.) However, the expression that
15412 sets the lengths' list is part of the body---and that is the value
15413 that we want returned by the function as a whole. To do this, we
15414 enclose the @code{while} loop within a @code{let} expression, and
15415 arrange that the last element of the @code{let} expression contains
15416 the value of the lengths' list. (@xref{Incrementing Example, , Loop
15417 Example with an Incrementing Counter}.)
15419 @findex lengths-list-many-files
15421 These considerations lead us directly to the function itself:
15425 ;;; @r{Use @code{while} loop.}
15426 (defun lengths-list-many-files (list-of-files)
15427 "Return list of lengths of defuns in LIST-OF-FILES."
15430 (let (lengths-list)
15432 ;;; @r{true-or-false-test}
15433 (while list-of-files
15438 ;;; @r{Generate a lengths' list.}
15440 (expand-file-name (car list-of-files)))))
15444 ;;; @r{Make files' list shorter.}
15445 (setq list-of-files (cdr list-of-files)))
15447 ;;; @r{Return final value of lengths' list.}
15452 @code{expand-file-name} is a built-in function that converts a file
15453 name to the absolute, long, path name form. The function employs the
15454 name of the directory in which the function is called.
15456 @c !!! 22.1.1 lisp sources location here
15458 Thus, if @code{expand-file-name} is called on @code{debug.el} when
15459 Emacs is visiting the
15460 @file{/usr/local/share/emacs/22.1.1/lisp/emacs-lisp/} directory,
15470 @c !!! 22.1.1 lisp sources location here
15472 /usr/local/share/emacs/22.1.1/lisp/emacs-lisp/debug.el
15475 The only other new element of this function definition is the as yet
15476 unstudied function @code{append}, which merits a short section for
15479 @node append, , lengths-list-many-files, Several files
15480 @subsection The @code{append} Function
15483 The @code{append} function attaches one list to another. Thus,
15486 (append '(1 2 3 4) '(5 6 7 8))
15497 This is exactly how we want to attach two lengths' lists produced by
15498 @code{lengths-list-file} to each other. The results contrast with
15502 (cons '(1 2 3 4) '(5 6 7 8))
15507 which constructs a new list in which the first argument to @code{cons}
15508 becomes the first element of the new list:
15511 ((1 2 3 4) 5 6 7 8)
15514 @node Several files recursively, Prepare the data, Several files, Words in a defun
15515 @section Recursively Count Words in Different Files
15517 Besides a @code{while} loop, you can work on each of a list of files
15518 with recursion. A recursive version of @code{lengths-list-many-files}
15519 is short and simple.
15521 The recursive function has the usual parts: the `do-again-test', the
15522 `next-step-expression', and the recursive call. The `do-again-test'
15523 determines whether the function should call itself again, which it
15524 will do if the @code{list-of-files} contains any remaining elements;
15525 the `next-step-expression' resets the @code{list-of-files} to the
15526 @sc{cdr} of itself, so eventually the list will be empty; and the
15527 recursive call calls itself on the shorter list. The complete
15528 function is shorter than this description!
15529 @findex recursive-lengths-list-many-files
15533 (defun recursive-lengths-list-many-files (list-of-files)
15534 "Return list of lengths of each defun in LIST-OF-FILES."
15535 (if list-of-files ; @r{do-again-test}
15538 (expand-file-name (car list-of-files)))
15539 (recursive-lengths-list-many-files
15540 (cdr list-of-files)))))
15545 In a sentence, the function returns the lengths' list for the first of
15546 the @code{list-of-files} appended to the result of calling itself on
15547 the rest of the @code{list-of-files}.
15549 Here is a test of @code{recursive-lengths-list-many-files}, along with
15550 the results of running @code{lengths-list-file} on each of the files
15553 Install @code{recursive-lengths-list-many-files} and
15554 @code{lengths-list-file}, if necessary, and then evaluate the
15555 following expressions. You may need to change the files' pathnames;
15556 those here work when this Info file and the Emacs sources are located
15557 in their customary places. To change the expressions, copy them to
15558 the @file{*scratch*} buffer, edit them, and then evaluate them.
15560 The results are shown after the @samp{@result{}}. (These results are
15561 for files from Emacs version 22.1.1; files from other versions of
15562 Emacs may produce different results.)
15564 @c !!! 22.1.1 lisp sources location here
15567 (cd "/usr/local/share/emacs/22.1.1/")
15569 (lengths-list-file "./lisp/macros.el")
15570 @result{} (283 263 480 90)
15574 (lengths-list-file "./lisp/mail/mailalias.el")
15575 @result{} (38 32 29 95 178 180 321 218 324)
15579 (lengths-list-file "./lisp/makesum.el")
15584 (recursive-lengths-list-many-files
15585 '("./lisp/macros.el"
15586 "./lisp/mail/mailalias.el"
15587 "./lisp/makesum.el"))
15588 @result{} (283 263 480 90 38 32 29 95 178 180 321 218 324 85 181)
15592 The @code{recursive-lengths-list-many-files} function produces the
15595 The next step is to prepare the data in the list for display in a graph.
15597 @node Prepare the data, , Several files recursively, Words in a defun
15598 @section Prepare the Data for Display in a Graph
15600 The @code{recursive-lengths-list-many-files} function returns a list
15601 of numbers. Each number records the length of a function definition.
15602 What we need to do now is transform this data into a list of numbers
15603 suitable for generating a graph. The new list will tell how many
15604 functions definitions contain less than 10 words and
15605 symbols, how many contain between 10 and 19 words and symbols, how
15606 many contain between 20 and 29 words and symbols, and so on.
15608 In brief, we need to go through the lengths' list produced by the
15609 @code{recursive-lengths-list-many-files} function and count the number
15610 of defuns within each range of lengths, and produce a list of those
15614 * Data for Display in Detail::
15615 * Sorting:: Sorting lists.
15616 * Files List:: Making a list of files.
15617 * Counting function definitions::
15620 @node Data for Display in Detail, Sorting, Prepare the data, Prepare the data
15622 @unnumberedsubsec The Data for Display in Detail
15625 Based on what we have done before, we can readily foresee that it
15626 should not be too hard to write a function that `@sc{cdr}s' down the
15627 lengths' list, looks at each element, determines which length range it
15628 is in, and increments a counter for that range.
15630 However, before beginning to write such a function, we should consider
15631 the advantages of sorting the lengths' list first, so the numbers are
15632 ordered from smallest to largest. First, sorting will make it easier
15633 to count the numbers in each range, since two adjacent numbers will
15634 either be in the same length range or in adjacent ranges. Second, by
15635 inspecting a sorted list, we can discover the highest and lowest
15636 number, and thereby determine the largest and smallest length range
15639 @node Sorting, Files List, Data for Display in Detail, Prepare the data
15640 @subsection Sorting Lists
15643 Emacs contains a function to sort lists, called (as you might guess)
15644 @code{sort}. The @code{sort} function takes two arguments, the list
15645 to be sorted, and a predicate that determines whether the first of
15646 two list elements is ``less'' than the second.
15648 As we saw earlier (@pxref{Wrong Type of Argument, , Using the Wrong
15649 Type Object as an Argument}), a predicate is a function that
15650 determines whether some property is true or false. The @code{sort}
15651 function will reorder a list according to whatever property the
15652 predicate uses; this means that @code{sort} can be used to sort
15653 non-numeric lists by non-numeric criteria---it can, for example,
15654 alphabetize a list.
15657 The @code{<} function is used when sorting a numeric list. For example,
15660 (sort '(4 8 21 17 33 7 21 7) '<)
15668 (4 7 7 8 17 21 21 33)
15672 (Note that in this example, both the arguments are quoted so that the
15673 symbols are not evaluated before being passed to @code{sort} as
15676 Sorting the list returned by the
15677 @code{recursive-lengths-list-many-files} function is straightforward;
15678 it uses the @code{<} function:
15682 In GNU Emacs 22, eval
15684 (cd "/usr/local/share/emacs/22.0.50/")
15686 (recursive-lengths-list-many-files
15687 '("./lisp/macros.el"
15688 "./lisp/mail/mailalias.el"
15689 "./lisp/makesum.el"))
15697 (recursive-lengths-list-many-files
15698 '("./lisp/macros.el"
15699 "./lisp/mailalias.el"
15700 "./lisp/makesum.el"))
15710 (29 32 38 85 90 95 178 180 181 218 263 283 321 324 480)
15714 (Note that in this example, the first argument to @code{sort} is not
15715 quoted, since the expression must be evaluated so as to produce the
15716 list that is passed to @code{sort}.)
15718 @node Files List, Counting function definitions, Sorting, Prepare the data
15719 @subsection Making a List of Files
15721 The @code{recursive-lengths-list-many-files} function requires a list
15722 of files as its argument. For our test examples, we constructed such
15723 a list by hand; but the Emacs Lisp source directory is too large for
15724 us to do for that. Instead, we will write a function to do the job
15725 for us. In this function, we will use both a @code{while} loop and a
15728 @findex directory-files
15729 We did not have to write a function like this for older versions of
15730 GNU Emacs, since they placed all the @samp{.el} files in one
15731 directory. Instead, we were able to use the @code{directory-files}
15732 function, which lists the names of files that match a specified
15733 pattern within a single directory.
15735 However, recent versions of Emacs place Emacs Lisp files in
15736 sub-directories of the top level @file{lisp} directory. This
15737 re-arrangement eases navigation. For example, all the mail related
15738 files are in a @file{lisp} sub-directory called @file{mail}. But at
15739 the same time, this arrangement forces us to create a file listing
15740 function that descends into the sub-directories.
15742 @findex files-in-below-directory
15743 We can create this function, called @code{files-in-below-directory},
15744 using familiar functions such as @code{car}, @code{nthcdr}, and
15745 @code{substring} in conjunction with an existing function called
15746 @code{directory-files-and-attributes}. This latter function not only
15747 lists all the filenames in a directory, including the names
15748 of sub-directories, but also their attributes.
15750 To restate our goal: to create a function that will enable us
15751 to feed filenames to @code{recursive-lengths-list-many-files}
15752 as a list that looks like this (but with more elements):
15756 ("./lisp/macros.el"
15757 "./lisp/mail/rmail.el"
15758 "./lisp/makesum.el")
15762 The @code{directory-files-and-attributes} function returns a list of
15763 lists. Each of the lists within the main list consists of 13
15764 elements. The first element is a string that contains the name of the
15765 file -- which, in GNU/Linux, may be a `directory file', that is to
15766 say, a file with the special attributes of a directory. The second
15767 element of the list is @code{t} for a directory, a string
15768 for symbolic link (the string is the name linked to), or @code{nil}.
15770 For example, the first @samp{.el} file in the @file{lisp/} directory
15771 is @file{abbrev.el}. Its name is
15772 @file{/usr/local/share/emacs/22.1.1/lisp/abbrev.el} and it is not a
15773 directory or a symbolic link.
15776 This is how @code{directory-files-and-attributes} lists that file and
15802 On the other hand, @file{mail/} is a directory within the @file{lisp/}
15803 directory. The beginning of its listing looks like this:
15814 (To learn about the different attributes, look at the documentation of
15815 @code{file-attributes}. Bear in mind that the @code{file-attributes}
15816 function does not list the filename, so its first element is
15817 @code{directory-files-and-attributes}'s second element.)
15819 We will want our new function, @code{files-in-below-directory}, to
15820 list the @samp{.el} files in the directory it is told to check, and in
15821 any directories below that directory.
15823 This gives us a hint on how to construct
15824 @code{files-in-below-directory}: within a directory, the function
15825 should add @samp{.el} filenames to a list; and if, within a directory,
15826 the function comes upon a sub-directory, it should go into that
15827 sub-directory and repeat its actions.
15829 However, we should note that every directory contains a name that
15830 refers to itself, called @file{.}, (``dot'') and a name that refers to
15831 its parent directory, called @file{..} (``double dot''). (In
15832 @file{/}, the root directory, @file{..} refers to itself, since
15833 @file{/} has no parent.) Clearly, we do not want our
15834 @code{files-in-below-directory} function to enter those directories,
15835 since they always lead us, directly or indirectly, to the current
15838 Consequently, our @code{files-in-below-directory} function must do
15843 Check to see whether it is looking at a filename that ends in
15844 @samp{.el}; and if so, add its name to a list.
15847 Check to see whether it is looking at a filename that is the name of a
15848 directory; and if so,
15852 Check to see whether it is looking at @file{.} or @file{..}; and if
15856 Or else, go into that directory and repeat the process.
15860 Let's write a function definition to do these tasks. We will use a
15861 @code{while} loop to move from one filename to another within a
15862 directory, checking what needs to be done; and we will use a recursive
15863 call to repeat the actions on each sub-directory. The recursive
15864 pattern is `accumulate'
15865 (@pxref{Accumulate, , Recursive Pattern: @emph{accumulate}}),
15866 using @code{append} as the combiner.
15869 (directory-files "/usr/local/src/emacs/lisp/" t "\\.el$")
15870 (shell-command "find /usr/local/src/emacs/lisp/ -name '*.el'")
15872 (directory-files "/usr/local/share/emacs/22.1.1/lisp/" t "\\.el$")
15873 (shell-command "find /usr/local/share/emacs/22.1.1/lisp/ -name '*.el'")
15876 @c /usr/local/share/emacs/22.1.1/lisp/
15879 Here is the function:
15883 (defun files-in-below-directory (directory)
15884 "List the .el files in DIRECTORY and in its sub-directories."
15885 ;; Although the function will be used non-interactively,
15886 ;; it will be easier to test if we make it interactive.
15887 ;; The directory will have a name such as
15888 ;; "/usr/local/share/emacs/22.1.1/lisp/"
15889 (interactive "DDirectory name: ")
15892 (let (el-files-list
15893 (current-directory-list
15894 (directory-files-and-attributes directory t)))
15895 ;; while we are in the current directory
15896 (while current-directory-list
15900 ;; check to see whether filename ends in `.el'
15901 ;; and if so, append its name to a list.
15902 ((equal ".el" (substring (car (car current-directory-list)) -3))
15903 (setq el-files-list
15904 (cons (car (car current-directory-list)) el-files-list)))
15907 ;; check whether filename is that of a directory
15908 ((eq t (car (cdr (car current-directory-list))))
15909 ;; decide whether to skip or recurse
15912 (substring (car (car current-directory-list)) -1))
15913 ;; then do nothing since filename is that of
15914 ;; current directory or parent, "." or ".."
15918 ;; else descend into the directory and repeat the process
15919 (setq el-files-list
15921 (files-in-below-directory
15922 (car (car current-directory-list)))
15924 ;; move to the next filename in the list; this also
15925 ;; shortens the list so the while loop eventually comes to an end
15926 (setq current-directory-list (cdr current-directory-list)))
15927 ;; return the filenames
15932 @c (files-in-below-directory "/usr/local/src/emacs/lisp/")
15933 @c (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/")
15935 The @code{files-in-below-directory} @code{directory-files} function
15936 takes one argument, the name of a directory.
15939 Thus, on my system,
15941 @c (length (files-in-below-directory "/usr/local/src/emacs/lisp/"))
15943 @c !!! 22.1.1 lisp sources location here
15947 (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/"))
15952 tells me that in and below my Lisp sources directory are 1031
15955 @code{files-in-below-directory} returns a list in reverse alphabetical
15956 order. An expression to sort the list in alphabetical order looks
15962 (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/")
15969 "Test how long it takes to find lengths of all sorted elisp defuns."
15970 (insert "\n" (current-time-string) "\n")
15973 (recursive-lengths-list-many-files
15974 (files-in-below-directory "/usr/local/src/emacs/lisp/"))
15976 (insert (format "%s" (current-time-string))))
15979 @node Counting function definitions, , Files List, Prepare the data
15980 @subsection Counting function definitions
15982 Our immediate goal is to generate a list that tells us how many
15983 function definitions contain fewer than 10 words and symbols, how many
15984 contain between 10 and 19 words and symbols, how many contain between
15985 20 and 29 words and symbols, and so on.
15987 With a sorted list of numbers, this is easy: count how many elements
15988 of the list are smaller than 10, then, after moving past the numbers
15989 just counted, count how many are smaller than 20, then, after moving
15990 past the numbers just counted, count how many are smaller than 30, and
15991 so on. Each of the numbers, 10, 20, 30, 40, and the like, is one
15992 larger than the top of that range. We can call the list of such
15993 numbers the @code{top-of-ranges} list.
15996 If we wished, we could generate this list automatically, but it is
15997 simpler to write a list manually. Here it is:
15998 @vindex top-of-ranges
16002 (defvar top-of-ranges
16005 110 120 130 140 150
16006 160 170 180 190 200
16007 210 220 230 240 250
16008 260 270 280 290 300)
16009 "List specifying ranges for `defuns-per-range'.")
16013 To change the ranges, we edit this list.
16015 Next, we need to write the function that creates the list of the
16016 number of definitions within each range. Clearly, this function must
16017 take the @code{sorted-lengths} and the @code{top-of-ranges} lists
16020 The @code{defuns-per-range} function must do two things again and
16021 again: it must count the number of definitions within a range
16022 specified by the current top-of-range value; and it must shift to the
16023 next higher value in the @code{top-of-ranges} list after counting the
16024 number of definitions in the current range. Since each of these
16025 actions is repetitive, we can use @code{while} loops for the job.
16026 One loop counts the number of definitions in the range defined by the
16027 current top-of-range value, and the other loop selects each of the
16028 top-of-range values in turn.
16030 Several entries of the @code{sorted-lengths} list are counted for each
16031 range; this means that the loop for the @code{sorted-lengths} list
16032 will be inside the loop for the @code{top-of-ranges} list, like a
16033 small gear inside a big gear.
16035 The inner loop counts the number of definitions within the range. It
16036 is a simple counting loop of the type we have seen before.
16037 (@xref{Incrementing Loop, , A loop with an incrementing counter}.)
16038 The true-or-false test of the loop tests whether the value from the
16039 @code{sorted-lengths} list is smaller than the current value of the
16040 top of the range. If it is, the function increments the counter and
16041 tests the next value from the @code{sorted-lengths} list.
16044 The inner loop looks like this:
16048 (while @var{length-element-smaller-than-top-of-range}
16049 (setq number-within-range (1+ number-within-range))
16050 (setq sorted-lengths (cdr sorted-lengths)))
16054 The outer loop must start with the lowest value of the
16055 @code{top-of-ranges} list, and then be set to each of the succeeding
16056 higher values in turn. This can be done with a loop like this:
16060 (while top-of-ranges
16061 @var{body-of-loop}@dots{}
16062 (setq top-of-ranges (cdr top-of-ranges)))
16067 Put together, the two loops look like this:
16071 (while top-of-ranges
16073 ;; @r{Count the number of elements within the current range.}
16074 (while @var{length-element-smaller-than-top-of-range}
16075 (setq number-within-range (1+ number-within-range))
16076 (setq sorted-lengths (cdr sorted-lengths)))
16078 ;; @r{Move to next range.}
16079 (setq top-of-ranges (cdr top-of-ranges)))
16083 In addition, in each circuit of the outer loop, Emacs should record
16084 the number of definitions within that range (the value of
16085 @code{number-within-range}) in a list. We can use @code{cons} for
16086 this purpose. (@xref{cons, , @code{cons}}.)
16088 The @code{cons} function works fine, except that the list it
16089 constructs will contain the number of definitions for the highest
16090 range at its beginning and the number of definitions for the lowest
16091 range at its end. This is because @code{cons} attaches new elements
16092 of the list to the beginning of the list, and since the two loops are
16093 working their way through the lengths' list from the lower end first,
16094 the @code{defuns-per-range-list} will end up largest number first.
16095 But we will want to print our graph with smallest values first and the
16096 larger later. The solution is to reverse the order of the
16097 @code{defuns-per-range-list}. We can do this using the
16098 @code{nreverse} function, which reverses the order of a list.
16105 (nreverse '(1 2 3 4))
16116 Note that the @code{nreverse} function is ``destructive''---that is,
16117 it changes the list to which it is applied; this contrasts with the
16118 @code{car} and @code{cdr} functions, which are non-destructive. In
16119 this case, we do not want the original @code{defuns-per-range-list},
16120 so it does not matter that it is destroyed. (The @code{reverse}
16121 function provides a reversed copy of a list, leaving the original list
16126 Put all together, the @code{defuns-per-range} looks like this:
16130 (defun defuns-per-range (sorted-lengths top-of-ranges)
16131 "SORTED-LENGTHS defuns in each TOP-OF-RANGES range."
16132 (let ((top-of-range (car top-of-ranges))
16133 (number-within-range 0)
16134 defuns-per-range-list)
16139 (while top-of-ranges
16145 ;; @r{Need number for numeric test.}
16146 (car sorted-lengths)
16147 (< (car sorted-lengths) top-of-range))
16151 ;; @r{Count number of definitions within current range.}
16152 (setq number-within-range (1+ number-within-range))
16153 (setq sorted-lengths (cdr sorted-lengths)))
16155 ;; @r{Exit inner loop but remain within outer loop.}
16159 (setq defuns-per-range-list
16160 (cons number-within-range defuns-per-range-list))
16161 (setq number-within-range 0) ; @r{Reset count to zero.}
16165 ;; @r{Move to next range.}
16166 (setq top-of-ranges (cdr top-of-ranges))
16167 ;; @r{Specify next top of range value.}
16168 (setq top-of-range (car top-of-ranges)))
16172 ;; @r{Exit outer loop and count the number of defuns larger than}
16173 ;; @r{ the largest top-of-range value.}
16174 (setq defuns-per-range-list
16176 (length sorted-lengths)
16177 defuns-per-range-list))
16181 ;; @r{Return a list of the number of definitions within each range,}
16182 ;; @r{ smallest to largest.}
16183 (nreverse defuns-per-range-list)))
16189 The function is straightforward except for one subtle feature. The
16190 true-or-false test of the inner loop looks like this:
16194 (and (car sorted-lengths)
16195 (< (car sorted-lengths) top-of-range))
16201 instead of like this:
16204 (< (car sorted-lengths) top-of-range)
16207 The purpose of the test is to determine whether the first item in the
16208 @code{sorted-lengths} list is less than the value of the top of the
16211 The simple version of the test works fine unless the
16212 @code{sorted-lengths} list has a @code{nil} value. In that case, the
16213 @code{(car sorted-lengths)} expression function returns
16214 @code{nil}. The @code{<} function cannot compare a number to
16215 @code{nil}, which is an empty list, so Emacs signals an error and
16216 stops the function from attempting to continue to execute.
16218 The @code{sorted-lengths} list always becomes @code{nil} when the
16219 counter reaches the end of the list. This means that any attempt to
16220 use the @code{defuns-per-range} function with the simple version of
16221 the test will fail.
16223 We solve the problem by using the @code{(car sorted-lengths)}
16224 expression in conjunction with the @code{and} expression. The
16225 @code{(car sorted-lengths)} expression returns a non-@code{nil}
16226 value so long as the list has at least one number within it, but
16227 returns @code{nil} if the list is empty. The @code{and} expression
16228 first evaluates the @code{(car sorted-lengths)} expression, and
16229 if it is @code{nil}, returns false @emph{without} evaluating the
16230 @code{<} expression. But if the @code{(car sorted-lengths)}
16231 expression returns a non-@code{nil} value, the @code{and} expression
16232 evaluates the @code{<} expression, and returns that value as the value
16233 of the @code{and} expression.
16235 @c colon in printed section title causes problem in Info cross reference
16236 This way, we avoid an error.
16239 (For information about @code{and}, see
16240 @ref{kill-new function, , The @code{kill-new} function}.)
16244 (@xref{kill-new function, , The @code{kill-new} function}, for
16245 information about @code{and}.)
16248 Here is a short test of the @code{defuns-per-range} function. First,
16249 evaluate the expression that binds (a shortened)
16250 @code{top-of-ranges} list to the list of values, then evaluate the
16251 expression for binding the @code{sorted-lengths} list, and then
16252 evaluate the @code{defuns-per-range} function.
16256 ;; @r{(Shorter list than we will use later.)}
16257 (setq top-of-ranges
16258 '(110 120 130 140 150
16259 160 170 180 190 200))
16261 (setq sorted-lengths
16262 '(85 86 110 116 122 129 154 176 179 200 265 300 300))
16264 (defuns-per-range sorted-lengths top-of-ranges)
16270 The list returned looks like this:
16273 (2 2 2 0 0 1 0 2 0 0 4)
16277 Indeed, there are two elements of the @code{sorted-lengths} list
16278 smaller than 110, two elements between 110 and 119, two elements
16279 between 120 and 129, and so on. There are four elements with a value
16282 @c The next step is to turn this numbers' list into a graph.
16283 @node Readying a Graph, Emacs Initialization, Words in a defun, Top
16284 @chapter Readying a Graph
16285 @cindex Readying a graph
16286 @cindex Graph prototype
16287 @cindex Prototype graph
16288 @cindex Body of graph
16290 Our goal is to construct a graph showing the numbers of function
16291 definitions of various lengths in the Emacs lisp sources.
16293 As a practical matter, if you were creating a graph, you would
16294 probably use a program such as @code{gnuplot} to do the job.
16295 (@code{gnuplot} is nicely integrated into GNU Emacs.) In this case,
16296 however, we create one from scratch, and in the process we will
16297 re-acquaint ourselves with some of what we learned before and learn
16300 In this chapter, we will first write a simple graph printing function.
16301 This first definition will be a @dfn{prototype}, a rapidly written
16302 function that enables us to reconnoiter this unknown graph-making
16303 territory. We will discover dragons, or find that they are myth.
16304 After scouting the terrain, we will feel more confident and enhance
16305 the function to label the axes automatically.
16308 * Columns of a graph::
16309 * graph-body-print:: How to print the body of a graph.
16310 * recursive-graph-body-print::
16312 * Line Graph Exercise::
16315 @node Columns of a graph, graph-body-print, Readying a Graph, Readying a Graph
16317 @unnumberedsec Printing the Columns of a Graph
16320 Since Emacs is designed to be flexible and work with all kinds of
16321 terminals, including character-only terminals, the graph will need to
16322 be made from one of the `typewriter' symbols. An asterisk will do; as
16323 we enhance the graph-printing function, we can make the choice of
16324 symbol a user option.
16326 We can call this function @code{graph-body-print}; it will take a
16327 @code{numbers-list} as its only argument. At this stage, we will not
16328 label the graph, but only print its body.
16330 The @code{graph-body-print} function inserts a vertical column of
16331 asterisks for each element in the @code{numbers-list}. The height of
16332 each line is determined by the value of that element of the
16333 @code{numbers-list}.
16335 Inserting columns is a repetitive act; that means that this function can
16336 be written either with a @code{while} loop or recursively.
16338 Our first challenge is to discover how to print a column of asterisks.
16339 Usually, in Emacs, we print characters onto a screen horizontally,
16340 line by line, by typing. We have two routes we can follow: write our
16341 own column-insertion function or discover whether one exists in Emacs.
16343 To see whether there is one in Emacs, we can use the @kbd{M-x apropos}
16344 command. This command is like the @kbd{C-h a} (@code{command-apropos})
16345 command, except that the latter finds only those functions that are
16346 commands. The @kbd{M-x apropos} command lists all symbols that match
16347 a regular expression, including functions that are not interactive.
16350 What we want to look for is some command that prints or inserts
16351 columns. Very likely, the name of the function will contain either
16352 the word `print' or the word `insert' or the word `column'.
16353 Therefore, we can simply type @kbd{M-x apropos RET
16354 print\|insert\|column RET} and look at the result. On my system, this
16355 command once too takes quite some time, and then produced a list of 79
16356 functions and variables. Now it does not take much time at all and
16357 produces a list of 211 functions and variables. Scanning down the
16358 list, the only function that looks as if it might do the job is
16359 @code{insert-rectangle}.
16362 Indeed, this is the function we want; its documentation says:
16367 Insert text of RECTANGLE with upper left corner at point.
16368 RECTANGLE's first line is inserted at point,
16369 its second line is inserted at a point vertically under point, etc.
16370 RECTANGLE should be a list of strings.
16371 After this command, the mark is at the upper left corner
16372 and point is at the lower right corner.
16376 We can run a quick test, to make sure it does what we expect of it.
16378 Here is the result of placing the cursor after the
16379 @code{insert-rectangle} expression and typing @kbd{C-u C-x C-e}
16380 (@code{eval-last-sexp}). The function inserts the strings
16381 @samp{"first"}, @samp{"second"}, and @samp{"third"} at and below
16382 point. Also the function returns @code{nil}.
16386 (insert-rectangle '("first" "second" "third"))first
16393 Of course, we won't be inserting the text of the
16394 @code{insert-rectangle} expression itself into the buffer in which we
16395 are making the graph, but will call the function from our program. We
16396 shall, however, have to make sure that point is in the buffer at the
16397 place where the @code{insert-rectangle} function will insert its
16400 If you are reading this in Info, you can see how this works by
16401 switching to another buffer, such as the @file{*scratch*} buffer,
16402 placing point somewhere in the buffer, typing @kbd{M-:}, typing the
16403 @code{insert-rectangle} expression into the minibuffer at the prompt,
16404 and then typing @key{RET}. This causes Emacs to evaluate the
16405 expression in the minibuffer, but to use as the value of point the
16406 position of point in the @file{*scratch*} buffer. (@kbd{M-:} is the
16407 keybinding for @code{eval-expression}. Also, @code{nil} does not
16408 appear in the @file{*scratch*} buffer since the expression is
16409 evaluated in the minibuffer.)
16411 We find when we do this that point ends up at the end of the last
16412 inserted line---that is to say, this function moves point as a
16413 side-effect. If we were to repeat the command, with point at this
16414 position, the next insertion would be below and to the right of the
16415 previous insertion. We don't want this! If we are going to make a
16416 bar graph, the columns need to be beside each other.
16418 So we discover that each cycle of the column-inserting @code{while}
16419 loop must reposition point to the place we want it, and that place
16420 will be at the top, not the bottom, of the column. Moreover, we
16421 remember that when we print a graph, we do not expect all the columns
16422 to be the same height. This means that the top of each column may be
16423 at a different height from the previous one. We cannot simply
16424 reposition point to the same line each time, but moved over to the
16425 right---or perhaps we can@dots{}
16427 We are planning to make the columns of the bar graph out of asterisks.
16428 The number of asterisks in the column is the number specified by the
16429 current element of the @code{numbers-list}. We need to construct a
16430 list of asterisks of the right length for each call to
16431 @code{insert-rectangle}. If this list consists solely of the requisite
16432 number of asterisks, then we will have position point the right number
16433 of lines above the base for the graph to print correctly. This could
16436 Alternatively, if we can figure out some way to pass
16437 @code{insert-rectangle} a list of the same length each time, then we
16438 can place point on the same line each time, but move it over one
16439 column to the right for each new column. If we do this, however, some
16440 of the entries in the list passed to @code{insert-rectangle} must be
16441 blanks rather than asterisks. For example, if the maximum height of
16442 the graph is 5, but the height of the column is 3, then
16443 @code{insert-rectangle} requires an argument that looks like this:
16446 (" " " " "*" "*" "*")
16449 This last proposal is not so difficult, so long as we can determine
16450 the column height. There are two ways for us to specify the column
16451 height: we can arbitrarily state what it will be, which would work
16452 fine for graphs of that height; or we can search through the list of
16453 numbers and use the maximum height of the list as the maximum height
16454 of the graph. If the latter operation were difficult, then the former
16455 procedure would be easiest, but there is a function built into Emacs
16456 that determines the maximum of its arguments. We can use that
16457 function. The function is called @code{max} and it returns the
16458 largest of all its arguments, which must be numbers. Thus, for
16466 returns 7. (A corresponding function called @code{min} returns the
16467 smallest of all its arguments.)
16471 However, we cannot simply call @code{max} on the @code{numbers-list};
16472 the @code{max} function expects numbers as its argument, not a list of
16473 numbers. Thus, the following expression,
16476 (max '(3 4 6 5 7 3))
16481 produces the following error message;
16484 Wrong type of argument: number-or-marker-p, (3 4 6 5 7 3)
16488 We need a function that passes a list of arguments to a function.
16489 This function is @code{apply}. This function `applies' its first
16490 argument (a function) to its remaining arguments, the last of which
16497 (apply 'max 3 4 7 3 '(4 8 5))
16503 (Incidentally, I don't know how you would learn of this function
16504 without a book such as this. It is possible to discover other
16505 functions, like @code{search-forward} or @code{insert-rectangle}, by
16506 guessing at a part of their names and then using @code{apropos}. Even
16507 though its base in metaphor is clear---`apply' its first argument to
16508 the rest---I doubt a novice would come up with that particular word
16509 when using @code{apropos} or other aid. Of course, I could be wrong;
16510 after all, the function was first named by someone who had to invent
16513 The second and subsequent arguments to @code{apply} are optional, so
16514 we can use @code{apply} to call a function and pass the elements of a
16515 list to it, like this, which also returns 8:
16518 (apply 'max '(4 8 5))
16521 This latter way is how we will use @code{apply}. The
16522 @code{recursive-lengths-list-many-files} function returns a numbers'
16523 list to which we can apply @code{max} (we could also apply @code{max} to
16524 the sorted numbers' list; it does not matter whether the list is
16528 Hence, the operation for finding the maximum height of the graph is this:
16531 (setq max-graph-height (apply 'max numbers-list))
16534 Now we can return to the question of how to create a list of strings
16535 for a column of the graph. Told the maximum height of the graph
16536 and the number of asterisks that should appear in the column, the
16537 function should return a list of strings for the
16538 @code{insert-rectangle} command to insert.
16540 Each column is made up of asterisks or blanks. Since the function is
16541 passed the value of the height of the column and the number of
16542 asterisks in the column, the number of blanks can be found by
16543 subtracting the number of asterisks from the height of the column.
16544 Given the number of blanks and the number of asterisks, two
16545 @code{while} loops can be used to construct the list:
16549 ;;; @r{First version.}
16550 (defun column-of-graph (max-graph-height actual-height)
16551 "Return list of strings that is one column of a graph."
16552 (let ((insert-list nil)
16553 (number-of-top-blanks
16554 (- max-graph-height actual-height)))
16558 ;; @r{Fill in asterisks.}
16559 (while (> actual-height 0)
16560 (setq insert-list (cons "*" insert-list))
16561 (setq actual-height (1- actual-height)))
16565 ;; @r{Fill in blanks.}
16566 (while (> number-of-top-blanks 0)
16567 (setq insert-list (cons " " insert-list))
16568 (setq number-of-top-blanks
16569 (1- number-of-top-blanks)))
16573 ;; @r{Return whole list.}
16578 If you install this function and then evaluate the following
16579 expression you will see that it returns the list as desired:
16582 (column-of-graph 5 3)
16590 (" " " " "*" "*" "*")
16593 As written, @code{column-of-graph} contains a major flaw: the symbols
16594 used for the blank and for the marked entries in the column are
16595 `hard-coded' as a space and asterisk. This is fine for a prototype,
16596 but you, or another user, may wish to use other symbols. For example,
16597 in testing the graph function, you many want to use a period in place
16598 of the space, to make sure the point is being repositioned properly
16599 each time the @code{insert-rectangle} function is called; or you might
16600 want to substitute a @samp{+} sign or other symbol for the asterisk.
16601 You might even want to make a graph-column that is more than one
16602 display column wide. The program should be more flexible. The way to
16603 do that is to replace the blank and the asterisk with two variables
16604 that we can call @code{graph-blank} and @code{graph-symbol} and define
16605 those variables separately.
16607 Also, the documentation is not well written. These considerations
16608 lead us to the second version of the function:
16612 (defvar graph-symbol "*"
16613 "String used as symbol in graph, usually an asterisk.")
16617 (defvar graph-blank " "
16618 "String used as blank in graph, usually a blank space.
16619 graph-blank must be the same number of columns wide
16625 (For an explanation of @code{defvar}, see
16626 @ref{defvar, , Initializing a Variable with @code{defvar}}.)
16630 ;;; @r{Second version.}
16631 (defun column-of-graph (max-graph-height actual-height)
16632 "Return MAX-GRAPH-HEIGHT strings; ACTUAL-HEIGHT are graph-symbols.
16636 The graph-symbols are contiguous entries at the end
16638 The list will be inserted as one column of a graph.
16639 The strings are either graph-blank or graph-symbol."
16643 (let ((insert-list nil)
16644 (number-of-top-blanks
16645 (- max-graph-height actual-height)))
16649 ;; @r{Fill in @code{graph-symbols}.}
16650 (while (> actual-height 0)
16651 (setq insert-list (cons graph-symbol insert-list))
16652 (setq actual-height (1- actual-height)))
16656 ;; @r{Fill in @code{graph-blanks}.}
16657 (while (> number-of-top-blanks 0)
16658 (setq insert-list (cons graph-blank insert-list))
16659 (setq number-of-top-blanks
16660 (1- number-of-top-blanks)))
16662 ;; @r{Return whole list.}
16667 If we wished, we could rewrite @code{column-of-graph} a third time to
16668 provide optionally for a line graph as well as for a bar graph. This
16669 would not be hard to do. One way to think of a line graph is that it
16670 is no more than a bar graph in which the part of each bar that is
16671 below the top is blank. To construct a column for a line graph, the
16672 function first constructs a list of blanks that is one shorter than
16673 the value, then it uses @code{cons} to attach a graph symbol to the
16674 list; then it uses @code{cons} again to attach the `top blanks' to
16677 It is easy to see how to write such a function, but since we don't
16678 need it, we will not do it. But the job could be done, and if it were
16679 done, it would be done with @code{column-of-graph}. Even more
16680 important, it is worth noting that few changes would have to be made
16681 anywhere else. The enhancement, if we ever wish to make it, is
16684 Now, finally, we come to our first actual graph printing function.
16685 This prints the body of a graph, not the labels for the vertical and
16686 horizontal axes, so we can call this @code{graph-body-print}.
16688 @node graph-body-print, recursive-graph-body-print, Columns of a graph, Readying a Graph
16689 @section The @code{graph-body-print} Function
16690 @findex graph-body-print
16692 After our preparation in the preceding section, the
16693 @code{graph-body-print} function is straightforward. The function
16694 will print column after column of asterisks and blanks, using the
16695 elements of a numbers' list to specify the number of asterisks in each
16696 column. This is a repetitive act, which means we can use a
16697 decrementing @code{while} loop or recursive function for the job. In
16698 this section, we will write the definition using a @code{while} loop.
16700 The @code{column-of-graph} function requires the height of the graph
16701 as an argument, so we should determine and record that as a local variable.
16703 This leads us to the following template for the @code{while} loop
16704 version of this function:
16708 (defun graph-body-print (numbers-list)
16709 "@var{documentation}@dots{}"
16710 (let ((height @dots{}
16715 (while numbers-list
16716 @var{insert-columns-and-reposition-point}
16717 (setq numbers-list (cdr numbers-list)))))
16722 We need to fill in the slots of the template.
16724 Clearly, we can use the @code{(apply 'max numbers-list)} expression to
16725 determine the height of the graph.
16727 The @code{while} loop will cycle through the @code{numbers-list} one
16728 element at a time. As it is shortened by the @code{(setq numbers-list
16729 (cdr numbers-list))} expression, the @sc{car} of each instance of the
16730 list is the value of the argument for @code{column-of-graph}.
16732 At each cycle of the @code{while} loop, the @code{insert-rectangle}
16733 function inserts the list returned by @code{column-of-graph}. Since
16734 the @code{insert-rectangle} function moves point to the lower right of
16735 the inserted rectangle, we need to save the location of point at the
16736 time the rectangle is inserted, move back to that position after the
16737 rectangle is inserted, and then move horizontally to the next place
16738 from which @code{insert-rectangle} is called.
16740 If the inserted columns are one character wide, as they will be if
16741 single blanks and asterisks are used, the repositioning command is
16742 simply @code{(forward-char 1)}; however, the width of a column may be
16743 greater than one. This means that the repositioning command should be
16744 written @code{(forward-char symbol-width)}. The @code{symbol-width}
16745 itself is the length of a @code{graph-blank} and can be found using
16746 the expression @code{(length graph-blank)}. The best place to bind
16747 the @code{symbol-width} variable to the value of the width of graph
16748 column is in the varlist of the @code{let} expression.
16751 These considerations lead to the following function definition:
16755 (defun graph-body-print (numbers-list)
16756 "Print a bar graph of the NUMBERS-LIST.
16757 The numbers-list consists of the Y-axis values."
16759 (let ((height (apply 'max numbers-list))
16760 (symbol-width (length graph-blank))
16765 (while numbers-list
16766 (setq from-position (point))
16768 (column-of-graph height (car numbers-list)))
16769 (goto-char from-position)
16770 (forward-char symbol-width)
16773 ;; @r{Draw graph column by column.}
16775 (setq numbers-list (cdr numbers-list)))
16778 ;; @r{Place point for X axis labels.}
16779 (forward-line height)
16786 The one unexpected expression in this function is the
16787 @w{@code{(sit-for 0)}} expression in the @code{while} loop. This
16788 expression makes the graph printing operation more interesting to
16789 watch than it would be otherwise. The expression causes Emacs to
16790 `sit' or do nothing for a zero length of time and then redraw the
16791 screen. Placed here, it causes Emacs to redraw the screen column by
16792 column. Without it, Emacs would not redraw the screen until the
16795 We can test @code{graph-body-print} with a short list of numbers.
16799 Install @code{graph-symbol}, @code{graph-blank},
16800 @code{column-of-graph}, which are in
16802 @ref{Readying a Graph, , Readying a Graph},
16805 @ref{Columns of a graph},
16807 and @code{graph-body-print}.
16811 Copy the following expression:
16814 (graph-body-print '(1 2 3 4 6 4 3 5 7 6 5 2 3))
16818 Switch to the @file{*scratch*} buffer and place the cursor where you
16819 want the graph to start.
16822 Type @kbd{M-:} (@code{eval-expression}).
16825 Yank the @code{graph-body-print} expression into the minibuffer
16826 with @kbd{C-y} (@code{yank)}.
16829 Press @key{RET} to evaluate the @code{graph-body-print} expression.
16833 Emacs will print a graph like this:
16847 @node recursive-graph-body-print, Printed Axes, graph-body-print, Readying a Graph
16848 @section The @code{recursive-graph-body-print} Function
16849 @findex recursive-graph-body-print
16851 The @code{graph-body-print} function may also be written recursively.
16852 The recursive solution is divided into two parts: an outside `wrapper'
16853 that uses a @code{let} expression to determine the values of several
16854 variables that need only be found once, such as the maximum height of
16855 the graph, and an inside function that is called recursively to print
16859 The `wrapper' is uncomplicated:
16863 (defun recursive-graph-body-print (numbers-list)
16864 "Print a bar graph of the NUMBERS-LIST.
16865 The numbers-list consists of the Y-axis values."
16866 (let ((height (apply 'max numbers-list))
16867 (symbol-width (length graph-blank))
16869 (recursive-graph-body-print-internal
16876 The recursive function is a little more difficult. It has four parts:
16877 the `do-again-test', the printing code, the recursive call, and the
16878 `next-step-expression'. The `do-again-test' is a @code{when}
16879 expression that determines whether the @code{numbers-list} contains
16880 any remaining elements; if it does, the function prints one column of
16881 the graph using the printing code and calls itself again. The
16882 function calls itself again according to the value produced by the
16883 `next-step-expression' which causes the call to act on a shorter
16884 version of the @code{numbers-list}.
16888 (defun recursive-graph-body-print-internal
16889 (numbers-list height symbol-width)
16890 "Print a bar graph.
16891 Used within recursive-graph-body-print function."
16896 (setq from-position (point))
16898 (column-of-graph height (car numbers-list)))
16901 (goto-char from-position)
16902 (forward-char symbol-width)
16903 (sit-for 0) ; @r{Draw graph column by column.}
16904 (recursive-graph-body-print-internal
16905 (cdr numbers-list) height symbol-width)))
16910 After installation, this expression can be tested; here is a sample:
16913 (recursive-graph-body-print '(3 2 5 6 7 5 3 4 6 4 3 2 1))
16917 Here is what @code{recursive-graph-body-print} produces:
16931 Either of these two functions, @code{graph-body-print} or
16932 @code{recursive-graph-body-print}, create the body of a graph.
16934 @node Printed Axes, Line Graph Exercise, recursive-graph-body-print, Readying a Graph
16935 @section Need for Printed Axes
16937 A graph needs printed axes, so you can orient yourself. For a do-once
16938 project, it may be reasonable to draw the axes by hand using Emacs'
16939 Picture mode; but a graph drawing function may be used more than once.
16941 For this reason, I have written enhancements to the basic
16942 @code{print-graph-body} function that automatically print labels for
16943 the horizontal and vertical axes. Since the label printing functions
16944 do not contain much new material, I have placed their description in
16945 an appendix. @xref{Full Graph, , A Graph with Labelled Axes}.
16947 @node Line Graph Exercise, , Printed Axes, Readying a Graph
16950 Write a line graph version of the graph printing functions.
16952 @node Emacs Initialization, Debugging, Readying a Graph, Top
16953 @chapter Your @file{.emacs} File
16954 @cindex @file{.emacs} file
16955 @cindex Customizing your @file{.emacs} file
16956 @cindex Initialization file
16958 ``You don't have to like Emacs to like it'' -- this seemingly
16959 paradoxical statement is the secret of GNU Emacs. The plain, `out of
16960 the box' Emacs is a generic tool. Most people who use it, customize
16961 it to suit themselves.
16963 GNU Emacs is mostly written in Emacs Lisp; this means that by writing
16964 expressions in Emacs Lisp you can change or extend Emacs.
16967 * Default Configuration::
16968 * Site-wide Init:: You can write site-wide init files.
16969 * defcustom:: Emacs will write code for you.
16970 * Beginning a .emacs File:: How to write a @code{.emacs file}.
16971 * Text and Auto-fill:: Automatically wrap lines.
16972 * Mail Aliases:: Use abbreviations for email addresses.
16973 * Indent Tabs Mode:: Don't use tabs with @TeX{}
16974 * Keybindings:: Create some personal keybindings.
16975 * Keymaps:: More about key binding.
16976 * Loading Files:: Load (i.e., evaluate) files automatically.
16977 * Autoload:: Make functions available.
16978 * Simple Extension:: Define a function; bind it to a key.
16979 * X11 Colors:: Colors in X.
16981 * Mode Line:: How to customize your mode line.
16984 @node Default Configuration, Site-wide Init, Emacs Initialization, Emacs Initialization
16986 @unnumberedsec Emacs' Default Configuration
16989 There are those who appreciate Emacs' default configuration. After
16990 all, Emacs starts you in C mode when you edit a C file, starts you in
16991 Fortran mode when you edit a Fortran file, and starts you in
16992 Fundamental mode when you edit an unadorned file. This all makes
16993 sense, if you do not know who is going to use Emacs. Who knows what a
16994 person hopes to do with an unadorned file? Fundamental mode is the
16995 right default for such a file, just as C mode is the right default for
16996 editing C code. (Enough programming languages have syntaxes
16997 that enable them to share or nearly share features, so C mode is
16998 now provided by by CC mode, the `C Collection'.)
17000 But when you do know who is going to use Emacs---you,
17001 yourself---then it makes sense to customize Emacs.
17003 For example, I seldom want Fundamental mode when I edit an
17004 otherwise undistinguished file; I want Text mode. This is why I
17005 customize Emacs: so it suits me.
17007 You can customize and extend Emacs by writing or adapting a
17008 @file{~/.emacs} file. This is your personal initialization file; its
17009 contents, written in Emacs Lisp, tell Emacs what to do.@footnote{You
17010 may also add @file{.el} to @file{~/.emacs} and call it a
17011 @file{~/.emacs.el} file. In the past, you were forbidden to type the
17012 extra keystrokes that the name @file{~/.emacs.el} requires, but now
17013 you may. The new format is consistent with the Emacs Lisp file
17014 naming conventions; the old format saves typing.}
17016 A @file{~/.emacs} file contains Emacs Lisp code. You can write this
17017 code yourself; or you can use Emacs' @code{customize} feature to write
17018 the code for you. You can combine your own expressions and
17019 auto-written Customize expressions in your @file{.emacs} file.
17021 (I myself prefer to write my own expressions, except for those,
17022 particularly fonts, that I find easier to manipulate using the
17023 @code{customize} command. I combine the two methods.)
17025 Most of this chapter is about writing expressions yourself. It
17026 describes a simple @file{.emacs} file; for more information, see
17027 @ref{Init File, , The Init File, emacs, The GNU Emacs Manual}, and
17028 @ref{Init File, , The Init File, elisp, The GNU Emacs Lisp Reference
17031 @node Site-wide Init, defcustom, Default Configuration, Emacs Initialization
17032 @section Site-wide Initialization Files
17034 @cindex @file{default.el} init file
17035 @cindex @file{site-init.el} init file
17036 @cindex @file{site-load.el} init file
17037 In addition to your personal initialization file, Emacs automatically
17038 loads various site-wide initialization files, if they exist. These
17039 have the same form as your @file{.emacs} file, but are loaded by
17042 Two site-wide initialization files, @file{site-load.el} and
17043 @file{site-init.el}, are loaded into Emacs and then `dumped' if a
17044 `dumped' version of Emacs is created, as is most common. (Dumped
17045 copies of Emacs load more quickly. However, once a file is loaded and
17046 dumped, a change to it does not lead to a change in Emacs unless you
17047 load it yourself or re-dump Emacs. @xref{Building Emacs, , Building
17048 Emacs, elisp, The GNU Emacs Lisp Reference Manual}, and the
17049 @file{INSTALL} file.)
17051 Three other site-wide initialization files are loaded automatically
17052 each time you start Emacs, if they exist. These are
17053 @file{site-start.el}, which is loaded @emph{before} your @file{.emacs}
17054 file, and @file{default.el}, and the terminal type file, which are both
17055 loaded @emph{after} your @file{.emacs} file.
17057 Settings and definitions in your @file{.emacs} file will overwrite
17058 conflicting settings and definitions in a @file{site-start.el} file,
17059 if it exists; but the settings and definitions in a @file{default.el}
17060 or terminal type file will overwrite those in your @file{.emacs} file.
17061 (You can prevent interference from a terminal type file by setting
17062 @code{term-file-prefix} to @code{nil}. @xref{Simple Extension, , A
17063 Simple Extension}.)
17065 @c Rewritten to avoid overfull hbox.
17066 The @file{INSTALL} file that comes in the distribution contains
17067 descriptions of the @file{site-init.el} and @file{site-load.el} files.
17069 The @file{loadup.el}, @file{startup.el}, and @file{loaddefs.el} files
17070 control loading. These files are in the @file{lisp} directory of the
17071 Emacs distribution and are worth perusing.
17073 The @file{loaddefs.el} file contains a good many suggestions as to
17074 what to put into your own @file{.emacs} file, or into a site-wide
17075 initialization file.
17077 @node defcustom, Beginning a .emacs File, Site-wide Init, Emacs Initialization
17078 @section Specifying Variables using @code{defcustom}
17081 You can specify variables using @code{defcustom} so that you and
17082 others can then use Emacs' @code{customize} feature to set their
17083 values. (You cannot use @code{customize} to write function
17084 definitions; but you can write @code{defuns} in your @file{.emacs}
17085 file. Indeed, you can write any Lisp expression in your @file{.emacs}
17088 The @code{customize} feature depends on the @code{defcustom} special
17089 form. Although you can use @code{defvar} or @code{setq} for variables
17090 that users set, the @code{defcustom} special form is designed for the
17093 You can use your knowledge of @code{defvar} for writing the
17094 first three arguments for @code{defcustom}. The first argument to
17095 @code{defcustom} is the name of the variable. The second argument is
17096 the variable's initial value, if any; and this value is set only if
17097 the value has not already been set. The third argument is the
17100 The fourth and subsequent arguments to @code{defcustom} specify types
17101 and options; these are not featured in @code{defvar}. (These
17102 arguments are optional.)
17104 Each of these arguments consists of a keyword followed by a value.
17105 Each keyword starts with the colon character @samp{:}.
17108 For example, the customizable user option variable
17109 @code{text-mode-hook} looks like this:
17113 (defcustom text-mode-hook nil
17114 "Normal hook run when entering Text mode and many related modes."
17116 :options '(turn-on-auto-fill flyspell-mode)
17122 The name of the variable is @code{text-mode-hook}; it has no default
17123 value; and its documentation string tells you what it does.
17125 The @code{:type} keyword tells Emacs the kind of data to which
17126 @code{text-mode-hook} should be set and how to display the value in a
17127 Customization buffer.
17129 The @code{:options} keyword specifies a suggested list of values for
17130 the variable. Usually, @code{:options} applies to a hook.
17131 The list is only a suggestion; it is not exclusive; a person who sets
17132 the variable may set it to other values; the list shown following the
17133 @code{:options} keyword is intended to offer convenient choices to a
17136 Finally, the @code{:group} keyword tells the Emacs Customization
17137 command in which group the variable is located. This tells where to
17140 The @code{defcustom} function recognizes more than a dozen keywords.
17141 For more information, see @ref{Customization, , Writing Customization
17142 Definitions, elisp, The GNU Emacs Lisp Reference Manual}.
17144 Consider @code{text-mode-hook} as an example.
17146 There are two ways to customize this variable. You can use the
17147 customization command or write the appropriate expressions yourself.
17150 Using the customization command, you can type:
17157 and find that the group for editing files of data is called `data'.
17158 Enter that group. Text Mode Hook is the first member. You can click
17159 on its various options, such as @code{turn-on-auto-fill}, to set the
17160 values. After you click on the button to
17163 Save for Future Sessions
17167 Emacs will write an expression into your @file{.emacs} file.
17168 It will look like this:
17172 (custom-set-variables
17173 ;; custom-set-variables was added by Custom.
17174 ;; If you edit it by hand, you could mess it up, so be careful.
17175 ;; Your init file should contain only one such instance.
17176 ;; If there is more than one, they won't work right.
17177 '(text-mode-hook (quote (turn-on-auto-fill text-mode-hook-identify))))
17182 (The @code{text-mode-hook-identify} function tells
17183 @code{toggle-text-mode-auto-fill} which buffers are in Text mode.
17184 It comes on automatically.)
17186 The @code{custom-set-variables} function works somewhat differently
17187 than a @code{setq}. While I have never learned the differences, I
17188 modify the @code{custom-set-variables} expressions in my @file{.emacs}
17189 file by hand: I make the changes in what appears to me to be a
17190 reasonable manner and have not had any problems. Others prefer to use
17191 the Customization command and let Emacs do the work for them.
17193 Another @code{custom-set-@dots{}} function is @code{custom-set-faces}.
17194 This function sets the various font faces. Over time, I have set a
17195 considerable number of faces. Some of the time, I re-set them using
17196 @code{customize}; other times, I simply edit the
17197 @code{custom-set-faces} expression in my @file{.emacs} file itself.
17199 The second way to customize your @code{text-mode-hook} is to set it
17200 yourself in your @file{.emacs} file using code that has nothing to do
17201 with the @code{custom-set-@dots{}} functions.
17204 When you do this, and later use @code{customize}, you will see a
17208 CHANGED outside Customize; operating on it here may be unreliable.
17212 This message is only a warning. If you click on the button to
17215 Save for Future Sessions
17219 Emacs will write a @code{custom-set-@dots{}} expression near the end
17220 of your @file{.emacs} file that will be evaluated after your
17221 hand-written expression. It will, therefore, overrule your
17222 hand-written expression. No harm will be done. When you do this,
17223 however, be careful to remember which expression is active; if you
17224 forget, you may confuse yourself.
17226 So long as you remember where the values are set, you will have no
17227 trouble. In any event, the values are always set in your
17228 initialization file, which is usually called @file{.emacs}.
17230 I myself use @code{customize} for hardly anything. Mostly, I write
17231 expressions myself.
17235 Incidentally, to be more complete concerning defines: @code{defsubst}
17236 defines an inline function. The syntax is just like that of
17237 @code{defun}. @code{defconst} defines a symbol as a constant. The
17238 intent is that neither programs nor users should ever change a value
17239 set by @code{defconst}. (You can change it; the value set is a
17240 variable; but please do not.)
17242 @node Beginning a .emacs File, Text and Auto-fill, defcustom, Emacs Initialization
17243 @section Beginning a @file{.emacs} File
17244 @cindex @file{.emacs} file, beginning of
17246 When you start Emacs, it loads your @file{.emacs} file unless you tell
17247 it not to by specifying @samp{-q} on the command line. (The
17248 @code{emacs -q} command gives you a plain, out-of-the-box Emacs.)
17250 A @file{.emacs} file contains Lisp expressions. Often, these are no
17251 more than expressions to set values; sometimes they are function
17254 @xref{Init File, , The Init File @file{~/.emacs}, emacs, The GNU Emacs
17255 Manual}, for a short description of initialization files.
17257 This chapter goes over some of the same ground, but is a walk among
17258 extracts from a complete, long-used @file{.emacs} file---my own.
17260 The first part of the file consists of comments: reminders to myself.
17261 By now, of course, I remember these things, but when I started, I did
17267 ;;;; Bob's .emacs file
17268 ; Robert J. Chassell
17269 ; 26 September 1985
17274 Look at that date! I started this file a long time ago. I have been
17275 adding to it ever since.
17279 ; Each section in this file is introduced by a
17280 ; line beginning with four semicolons; and each
17281 ; entry is introduced by a line beginning with
17282 ; three semicolons.
17287 This describes the usual conventions for comments in Emacs Lisp.
17288 Everything on a line that follows a semicolon is a comment. Two,
17289 three, and four semicolons are used as subsection and section markers.
17290 (@xref{Comments, ,, elisp, The GNU Emacs Lisp Reference Manual}, for
17291 more about comments.)
17296 ; Control-h is the help key;
17297 ; after typing control-h, type a letter to
17298 ; indicate the subject about which you want help.
17299 ; For an explanation of the help facility,
17300 ; type control-h two times in a row.
17305 Just remember: type @kbd{C-h} two times for help.
17309 ; To find out about any mode, type control-h m
17310 ; while in that mode. For example, to find out
17311 ; about mail mode, enter mail mode and then type
17317 `Mode help', as I call this, is very helpful. Usually, it tells you
17318 all you need to know.
17320 Of course, you don't need to include comments like these in your
17321 @file{.emacs} file. I included them in mine because I kept forgetting
17322 about Mode help or the conventions for comments---but I was able to
17323 remember to look here to remind myself.
17325 @node Text and Auto-fill, Mail Aliases, Beginning a .emacs File, Emacs Initialization
17326 @section Text and Auto Fill Mode
17328 Now we come to the part that `turns on' Text mode and
17333 ;;; Text mode and Auto Fill mode
17334 ; The next two lines put Emacs into Text mode
17335 ; and Auto Fill mode, and are for writers who
17336 ; want to start writing prose rather than code.
17337 (setq default-major-mode 'text-mode)
17338 (add-hook 'text-mode-hook 'turn-on-auto-fill)
17342 Here is the first part of this @file{.emacs} file that does something
17343 besides remind a forgetful human!
17345 The first of the two lines in parentheses tells Emacs to turn on Text
17346 mode when you find a file, @emph{unless} that file should go into some
17347 other mode, such as C mode.
17349 @cindex Per-buffer, local variables list
17350 @cindex Local variables list, per-buffer,
17351 @cindex Automatic mode selection
17352 @cindex Mode selection, automatic
17353 When Emacs reads a file, it looks at the extension to the file name,
17354 if any. (The extension is the part that comes after a @samp{.}.) If
17355 the file ends with a @samp{.c} or @samp{.h} extension then Emacs turns
17356 on C mode. Also, Emacs looks at first nonblank line of the file; if
17357 the line says @w{@samp{-*- C -*-}}, Emacs turns on C mode. Emacs
17358 possesses a list of extensions and specifications that it uses
17359 automatically. In addition, Emacs looks near the last page for a
17360 per-buffer, ``local variables list'', if any.
17363 @xref{Choosing Modes, , How Major Modes are Chosen, emacs, The GNU
17366 @xref{File Variables, , Local Variables in Files, emacs, The GNU Emacs
17370 See sections ``How Major Modes are Chosen'' and ``Local Variables in
17371 Files'' in @cite{The GNU Emacs Manual}.
17374 Now, back to the @file{.emacs} file.
17377 Here is the line again; how does it work?
17379 @cindex Text Mode turned on
17381 (setq default-major-mode 'text-mode)
17385 This line is a short, but complete Emacs Lisp expression.
17387 We are already familiar with @code{setq}. It sets the following variable,
17388 @code{default-major-mode}, to the subsequent value, which is
17389 @code{text-mode}. The single quote mark before @code{text-mode} tells
17390 Emacs to deal directly with the @code{text-mode} variable, not with
17391 whatever it might stand for. @xref{set & setq, , Setting the Value of
17392 a Variable}, for a reminder of how @code{setq} works. The main point
17393 is that there is no difference between the procedure you use to set
17394 a value in your @file{.emacs} file and the procedure you use anywhere
17398 Here is the next line:
17400 @cindex Auto Fill mode turned on
17403 (add-hook 'text-mode-hook 'turn-on-auto-fill)
17407 In this line, the @code{add-hook} command adds
17408 @code{turn-on-auto-fill} to the variable.
17410 @code{turn-on-auto-fill} is the name of a program, that, you guessed
17411 it!, turns on Auto Fill mode.
17413 Every time Emacs turns on Text mode, Emacs runs the commands `hooked'
17414 onto Text mode. So every time Emacs turns on Text mode, Emacs also
17415 turns on Auto Fill mode.
17417 In brief, the first line causes Emacs to enter Text mode when you edit a
17418 file, unless the file name extension, a first non-blank line, or local
17419 variables to tell Emacs otherwise.
17421 Text mode among other actions, sets the syntax table to work
17422 conveniently for writers. In Text mode, Emacs considers an apostrophe
17423 as part of a word like a letter; but Emacs does not consider a period
17424 or a space as part of a word. Thus, @kbd{M-f} moves you over
17425 @samp{it's}. On the other hand, in C mode, @kbd{M-f} stops just after
17426 the @samp{t} of @samp{it's}.
17428 The second line causes Emacs to turn on Auto Fill mode when it turns
17429 on Text mode. In Auto Fill mode, Emacs automatically breaks a line
17430 that is too wide and brings the excessively wide part of the line down
17431 to the next line. Emacs breaks lines between words, not within them.
17433 When Auto Fill mode is turned off, lines continue to the right as you
17434 type them. Depending on how you set the value of
17435 @code{truncate-lines}, the words you type either disappear off the
17436 right side of the screen, or else are shown, in a rather ugly and
17437 unreadable manner, as a continuation line on the screen.
17440 In addition, in this part of my @file{.emacs} file, I tell the Emacs
17441 fill commands to insert two spaces after a colon:
17444 (setq colon-double-space t)
17447 @node Mail Aliases, Indent Tabs Mode, Text and Auto-fill, Emacs Initialization
17448 @section Mail Aliases
17450 Here is a @code{setq} that `turns on' mail aliases, along with more
17456 ; To enter mail mode, type `C-x m'
17457 ; To enter RMAIL (for reading mail),
17459 (setq mail-aliases t)
17463 @cindex Mail aliases
17465 This @code{setq} command sets the value of the variable
17466 @code{mail-aliases} to @code{t}. Since @code{t} means true, the line
17467 says, in effect, ``Yes, use mail aliases.''
17469 Mail aliases are convenient short names for long email addresses or
17470 for lists of email addresses. The file where you keep your `aliases'
17471 is @file{~/.mailrc}. You write an alias like this:
17474 alias geo george@@foobar.wiz.edu
17478 When you write a message to George, address it to @samp{geo}; the
17479 mailer will automatically expand @samp{geo} to the full address.
17481 @node Indent Tabs Mode, Keybindings, Mail Aliases, Emacs Initialization
17482 @section Indent Tabs Mode
17483 @cindex Tabs, preventing
17484 @findex indent-tabs-mode
17486 By default, Emacs inserts tabs in place of multiple spaces when it
17487 formats a region. (For example, you might indent many lines of text
17488 all at once with the @code{indent-region} command.) Tabs look fine on
17489 a terminal or with ordinary printing, but they produce badly indented
17490 output when you use @TeX{} or Texinfo since @TeX{} ignores tabs.
17493 The following turns off Indent Tabs mode:
17497 ;;; Prevent Extraneous Tabs
17498 (setq-default indent-tabs-mode nil)
17502 Note that this line uses @code{setq-default} rather than the
17503 @code{setq} command that we have seen before. The @code{setq-default}
17504 command sets values only in buffers that do not have their own local
17505 values for the variable.
17508 @xref{Just Spaces, , Tabs vs. Spaces, emacs, The GNU Emacs Manual}.
17510 @xref{File Variables, , Local Variables in Files, emacs, The GNU Emacs
17514 See sections ``Tabs vs.@: Spaces'' and ``Local Variables in
17515 Files'' in @cite{The GNU Emacs Manual}.
17519 @node Keybindings, Keymaps, Indent Tabs Mode, Emacs Initialization
17520 @section Some Keybindings
17522 Now for some personal keybindings:
17526 ;;; Compare windows
17527 (global-set-key "\C-cw" 'compare-windows)
17531 @findex compare-windows
17532 @code{compare-windows} is a nifty command that compares the text in
17533 your current window with text in the next window. It makes the
17534 comparison by starting at point in each window, moving over text in
17535 each window as far as they match. I use this command all the time.
17537 This also shows how to set a key globally, for all modes.
17539 @cindex Setting a key globally
17540 @cindex Global set key
17541 @cindex Key setting globally
17542 @findex global-set-key
17543 The command is @code{global-set-key}. It is followed by the
17544 keybinding. In a @file{.emacs} file, the keybinding is written as
17545 shown: @code{\C-c} stands for `control-c', which means `press the
17546 control key and the @key{c} key at the same time'. The @code{w} means
17547 `press the @key{w} key'. The keybinding is surrounded by double
17548 quotation marks. In documentation, you would write this as
17549 @w{@kbd{C-c w}}. (If you were binding a @key{META} key, such as
17550 @kbd{M-c}, rather than a @key{CTRL} key, you would write
17551 @w{@code{\M-c}} in your @file{.emacs} file. @xref{Init Rebinding, ,
17552 Rebinding Keys in Your Init File, emacs, The GNU Emacs Manual}, for
17555 The command invoked by the keys is @code{compare-windows}. Note that
17556 @code{compare-windows} is preceded by a single quote; otherwise, Emacs
17557 would first try to evaluate the symbol to determine its value.
17559 These three things, the double quotation marks, the backslash before
17560 the @samp{C}, and the single quote mark are necessary parts of
17561 keybinding that I tend to forget. Fortunately, I have come to
17562 remember that I should look at my existing @file{.emacs} file, and
17563 adapt what is there.
17565 As for the keybinding itself: @kbd{C-c w}. This combines the prefix
17566 key, @kbd{C-c}, with a single character, in this case, @kbd{w}. This
17567 set of keys, @kbd{C-c} followed by a single character, is strictly
17568 reserved for individuals' own use. (I call these `own' keys, since
17569 these are for my own use.) You should always be able to create such a
17570 keybinding for your own use without stomping on someone else's
17571 keybinding. If you ever write an extension to Emacs, please avoid
17572 taking any of these keys for public use. Create a key like @kbd{C-c
17573 C-w} instead. Otherwise, we will run out of `own' keys.
17576 Here is another keybinding, with a comment:
17580 ;;; Keybinding for `occur'
17581 ; I use occur a lot, so let's bind it to a key:
17582 (global-set-key "\C-co" 'occur)
17587 The @code{occur} command shows all the lines in the current buffer
17588 that contain a match for a regular expression. Matching lines are
17589 shown in a buffer called @file{*Occur*}. That buffer serves as a menu
17590 to jump to occurrences.
17592 @findex global-unset-key
17593 @cindex Unbinding key
17594 @cindex Key unbinding
17596 Here is how to unbind a key, so it does not
17602 (global-unset-key "\C-xf")
17606 There is a reason for this unbinding: I found I inadvertently typed
17607 @w{@kbd{C-x f}} when I meant to type @kbd{C-x C-f}. Rather than find a
17608 file, as I intended, I accidentally set the width for filled text,
17609 almost always to a width I did not want. Since I hardly ever reset my
17610 default width, I simply unbound the key.
17612 @findex list-buffers, @r{rebound}
17613 @findex buffer-menu, @r{bound to key}
17615 The following rebinds an existing key:
17619 ;;; Rebind `C-x C-b' for `buffer-menu'
17620 (global-set-key "\C-x\C-b" 'buffer-menu)
17624 By default, @kbd{C-x C-b} runs the
17625 @code{list-buffers} command. This command lists
17626 your buffers in @emph{another} window. Since I
17627 almost always want to do something in that
17628 window, I prefer the @code{buffer-menu}
17629 command, which not only lists the buffers,
17630 but moves point into that window.
17632 @node Keymaps, Loading Files, Keybindings, Emacs Initialization
17635 @cindex Rebinding keys
17637 Emacs uses @dfn{keymaps} to record which keys call which commands.
17638 When you use @code{global-set-key} to set the keybinding for a single
17639 command in all parts of Emacs, you are specifying the keybinding in
17640 @code{current-global-map}.
17642 Specific modes, such as C mode or Text mode, have their own keymaps;
17643 the mode-specific keymaps override the global map that is shared by
17646 The @code{global-set-key} function binds, or rebinds, the global
17647 keymap. For example, the following binds the key @kbd{C-x C-b} to the
17648 function @code{buffer-menu}:
17651 (global-set-key "\C-x\C-b" 'buffer-menu)
17654 Mode-specific keymaps are bound using the @code{define-key} function,
17655 which takes a specific keymap as an argument, as well as the key and
17656 the command. For example, my @file{.emacs} file contains the
17657 following expression to bind the @code{texinfo-insert-@@group} command
17658 to @kbd{C-c C-c g}:
17662 (define-key texinfo-mode-map "\C-c\C-cg" 'texinfo-insert-@@group)
17667 The @code{texinfo-insert-@@group} function itself is a little extension
17668 to Texinfo mode that inserts @samp{@@group} into a Texinfo file. I
17669 use this command all the time and prefer to type the three strokes
17670 @kbd{C-c C-c g} rather than the six strokes @kbd{@@ g r o u p}.
17671 (@samp{@@group} and its matching @samp{@@end group} are commands that
17672 keep all enclosed text together on one page; many multi-line examples
17673 in this book are surrounded by @samp{@@group @dots{} @@end group}.)
17676 Here is the @code{texinfo-insert-@@group} function definition:
17680 (defun texinfo-insert-@@group ()
17681 "Insert the string @@group in a Texinfo buffer."
17683 (beginning-of-line)
17684 (insert "@@group\n"))
17688 (Of course, I could have used Abbrev mode to save typing, rather than
17689 write a function to insert a word; but I prefer key strokes consistent
17690 with other Texinfo mode key bindings.)
17692 You will see numerous @code{define-key} expressions in
17693 @file{loaddefs.el} as well as in the various mode libraries, such as
17694 @file{cc-mode.el} and @file{lisp-mode.el}.
17696 @xref{Key Bindings, , Customizing Key Bindings, emacs, The GNU Emacs
17697 Manual}, and @ref{Keymaps, , Keymaps, elisp, The GNU Emacs Lisp
17698 Reference Manual}, for more information about keymaps.
17700 @node Loading Files, Autoload, Keymaps, Emacs Initialization
17701 @section Loading Files
17702 @cindex Loading files
17705 Many people in the GNU Emacs community have written extensions to
17706 Emacs. As time goes by, these extensions are often included in new
17707 releases. For example, the Calendar and Diary packages are now part
17708 of the standard GNU Emacs, as is Calc.
17710 You can use a @code{load} command to evaluate a complete file and
17711 thereby install all the functions and variables in the file into Emacs.
17714 @c (auto-compression-mode t)
17717 (load "~/emacs/slowsplit")
17720 This evaluates, i.e.@: loads, the @file{slowsplit.el} file or if it
17721 exists, the faster, byte compiled @file{slowsplit.elc} file from the
17722 @file{emacs} sub-directory of your home directory. The file contains
17723 the function @code{split-window-quietly}, which John Robinson wrote in
17726 The @code{split-window-quietly} function splits a window with the
17727 minimum of redisplay. I installed it in 1989 because it worked well
17728 with the slow 1200 baud terminals I was then using. Nowadays, I only
17729 occasionally come across such a slow connection, but I continue to use
17730 the function because I like the way it leaves the bottom half of a
17731 buffer in the lower of the new windows and the top half in the upper
17735 To replace the key binding for the default
17736 @code{split-window-vertically}, you must also unset that key and bind
17737 the keys to @code{split-window-quietly}, like this:
17741 (global-unset-key "\C-x2")
17742 (global-set-key "\C-x2" 'split-window-quietly)
17747 If you load many extensions, as I do, then instead of specifying the
17748 exact location of the extension file, as shown above, you can specify
17749 that directory as part of Emacs' @code{load-path}. Then, when Emacs
17750 loads a file, it will search that directory as well as its default
17751 list of directories. (The default list is specified in @file{paths.h}
17752 when Emacs is built.)
17755 The following command adds your @file{~/emacs} directory to the
17756 existing load path:
17760 ;;; Emacs Load Path
17761 (setq load-path (cons "~/emacs" load-path))
17765 Incidentally, @code{load-library} is an interactive interface to the
17766 @code{load} function. The complete function looks like this:
17768 @findex load-library
17771 (defun load-library (library)
17772 "Load the library named LIBRARY.
17773 This is an interface to the function `load'."
17775 (list (completing-read "Load library: "
17776 'locate-file-completion
17777 (cons load-path (get-load-suffixes)))))
17782 The name of the function, @code{load-library}, comes from the use of
17783 `library' as a conventional synonym for `file'. The source for the
17784 @code{load-library} command is in the @file{files.el} library.
17786 Another interactive command that does a slightly different job is
17787 @code{load-file}. @xref{Lisp Libraries, , Libraries of Lisp Code for
17788 Emacs, emacs, The GNU Emacs Manual}, for information on the
17789 distinction between @code{load-library} and this command.
17791 @node Autoload, Simple Extension, Loading Files, Emacs Initialization
17792 @section Autoloading
17795 Instead of installing a function by loading the file that contains it,
17796 or by evaluating the function definition, you can make the function
17797 available but not actually install it until it is first called. This
17798 is called @dfn{autoloading}.
17800 When you execute an autoloaded function, Emacs automatically evaluates
17801 the file that contains the definition, and then calls the function.
17803 Emacs starts quicker with autoloaded functions, since their libraries
17804 are not loaded right away; but you need to wait a moment when you
17805 first use such a function, while its containing file is evaluated.
17807 Rarely used functions are frequently autoloaded. The
17808 @file{loaddefs.el} library contains hundreds of autoloaded functions,
17809 from @code{bookmark-set} to @code{wordstar-mode}. Of course, you may
17810 come to use a `rare' function frequently. When you do, you should
17811 load that function's file with a @code{load} expression in your
17812 @file{.emacs} file.
17814 In my @file{.emacs} file, I load 14 libraries that contain functions
17815 that would otherwise be autoloaded. (Actually, it would have been
17816 better to include these files in my `dumped' Emacs, but I forgot.
17817 @xref{Building Emacs, , Building Emacs, elisp, The GNU Emacs Lisp
17818 Reference Manual}, and the @file{INSTALL} file for more about
17821 You may also want to include autoloaded expressions in your @file{.emacs}
17822 file. @code{autoload} is a built-in function that takes up to five
17823 arguments, the final three of which are optional. The first argument
17824 is the name of the function to be autoloaded; the second is the name
17825 of the file to be loaded. The third argument is documentation for the
17826 function, and the fourth tells whether the function can be called
17827 interactively. The fifth argument tells what type of
17828 object---@code{autoload} can handle a keymap or macro as well as a
17829 function (the default is a function).
17832 Here is a typical example:
17836 (autoload 'html-helper-mode
17837 "html-helper-mode" "Edit HTML documents" t)
17842 (@code{html-helper-mode} is an older alternative to @code{html-mode},
17843 which is a standard part of the distribution.)
17846 This expression autoloads the @code{html-helper-mode} function. It
17847 takes it from the @file{html-helper-mode.el} file (or from the byte
17848 compiled file @file{html-helper-mode.elc}, if it exists.) The file
17849 must be located in a directory specified by @code{load-path}. The
17850 documentation says that this is a mode to help you edit documents
17851 written in the HyperText Markup Language. You can call this mode
17852 interactively by typing @kbd{M-x html-helper-mode}. (You need to
17853 duplicate the function's regular documentation in the autoload
17854 expression because the regular function is not yet loaded, so its
17855 documentation is not available.)
17857 @xref{Autoload, , Autoload, elisp, The GNU Emacs Lisp Reference
17858 Manual}, for more information.
17860 @node Simple Extension, X11 Colors, Autoload, Emacs Initialization
17861 @section A Simple Extension: @code{line-to-top-of-window}
17862 @findex line-to-top-of-window
17863 @cindex Simple extension in @file{.emacs} file
17865 Here is a simple extension to Emacs that moves the line point is on to
17866 the top of the window. I use this all the time, to make text easier
17869 You can put the following code into a separate file and then load it
17870 from your @file{.emacs} file, or you can include it within your
17871 @file{.emacs} file.
17874 Here is the definition:
17878 ;;; Line to top of window;
17879 ;;; replace three keystroke sequence C-u 0 C-l
17880 (defun line-to-top-of-window ()
17881 "Move the line point is on to top of window."
17888 Now for the keybinding.
17890 Nowadays, function keys as well as mouse button events and
17891 non-@sc{ascii} characters are written within square brackets, without
17892 quotation marks. (In Emacs version 18 and before, you had to write
17893 different function key bindings for each different make of terminal.)
17895 I bind @code{line-to-top-of-window} to my @key{F6} function key like
17899 (global-set-key [f6] 'line-to-top-of-window)
17902 For more information, see @ref{Init Rebinding, , Rebinding Keys in
17903 Your Init File, emacs, The GNU Emacs Manual}.
17905 @cindex Conditional 'twixt two versions of Emacs
17906 @cindex Version of Emacs, choosing
17907 @cindex Emacs version, choosing
17908 If you run two versions of GNU Emacs, such as versions 21 and 22, and
17909 use one @file{.emacs} file, you can select which code to evaluate with
17910 the following conditional:
17915 (= 21 emacs-major-version)
17916 ;; evaluate version 21 code
17918 (= 22 emacs-major-version)
17919 ;; evaluate version 22 code
17924 For example, in contrast to version 20, more recent versions blink
17925 their cursors by default. I hate such blinking, as well as other
17926 features, so I placed the following in my @file{.emacs}
17927 file@footnote{When I start instances of Emacs that do not load my
17928 @file{.emacs} file or any site file, I also turn off blinking:
17931 emacs -q --no-site-file -eval '(blink-cursor-mode nil)'
17933 @exdent Or nowadays, using an even more sophisticated set of options,
17941 (when (or (= 21 emacs-major-version)
17942 (= 22 emacs-major-version))
17943 (blink-cursor-mode 0)
17944 ;; Insert newline when you press `C-n' (next-line)
17945 ;; at the end of the buffer
17946 (setq next-line-add-newlines t)
17949 ;; Turn on image viewing
17950 (auto-image-file-mode t)
17953 ;; Turn on menu bar (this bar has text)
17954 ;; (Use numeric argument to turn on)
17958 ;; Turn off tool bar (this bar has icons)
17959 ;; (Use numeric argument to turn on)
17960 (tool-bar-mode nil)
17963 ;; Turn off tooltip mode for tool bar
17964 ;; (This mode causes icon explanations to pop up)
17965 ;; (Use numeric argument to turn on)
17967 ;; If tooltips turned on, make tips appear promptly
17968 (setq tooltip-delay 0.1) ; default is 0.7 second
17974 Alternatively, since @code{blink-cursor-mode} has existed since Emacs
17975 version 21 and is likely to continue, you could write
17979 (when (>= emacs-major-version 21)
17980 (blink-cursor-mode 0)
17985 and add other expressions, too.
17988 @node X11 Colors, Miscellaneous, Simple Extension, Emacs Initialization
17989 @section X11 Colors
17991 You can specify colors when you use Emacs with the MIT X Windowing
17994 I dislike the default colors and specify my own.
17997 Here are the expressions in my @file{.emacs}
17998 file that set values:
18002 ;; Set cursor color
18003 (set-cursor-color "white")
18006 (set-mouse-color "white")
18008 ;; Set foreground and background
18009 (set-foreground-color "white")
18010 (set-background-color "darkblue")
18014 ;;; Set highlighting colors for isearch and drag
18015 (set-face-foreground 'highlight "white")
18016 (set-face-background 'highlight "blue")
18020 (set-face-foreground 'region "cyan")
18021 (set-face-background 'region "blue")
18025 (set-face-foreground 'secondary-selection "skyblue")
18026 (set-face-background 'secondary-selection "darkblue")
18030 ;; Set calendar highlighting colors
18031 (setq calendar-load-hook
18033 (set-face-foreground 'diary-face "skyblue")
18034 (set-face-background 'holiday-face "slate blue")
18035 (set-face-foreground 'holiday-face "white")))
18039 The various shades of blue soothe my eye and prevent me from seeing
18040 the screen flicker.
18042 Alternatively, I could have set my specifications in various X
18043 initialization files. For example, I could set the foreground,
18044 background, cursor, and pointer (i.e., mouse) colors in my
18045 @file{~/.Xresources} file like this:
18049 Emacs*foreground: white
18050 Emacs*background: darkblue
18051 Emacs*cursorColor: white
18052 Emacs*pointerColor: white
18056 In any event, since it is not part of Emacs, I set the root color of
18057 my X window in my @file{~/.xinitrc} file, like this@footnote{I also
18058 run more modern window managers, such as Enlightenment, Gnome, or KDE;
18059 in those cases, I often specify an image rather than a plain color.}:
18062 xsetroot -solid Navy -fg white &
18066 @node Miscellaneous, Mode Line, X11 Colors, Emacs Initialization
18067 @section Miscellaneous Settings for a @file{.emacs} File
18070 Here are a few miscellaneous settings:
18075 Set the shape and color of the mouse cursor:
18079 ; Cursor shapes are defined in
18080 ; `/usr/include/X11/cursorfont.h';
18081 ; for example, the `target' cursor is number 128;
18082 ; the `top_left_arrow' cursor is number 132.
18086 (let ((mpointer (x-get-resource "*mpointer"
18087 "*emacs*mpointer")))
18088 ;; If you have not set your mouse pointer
18089 ;; then set it, otherwise leave as is:
18090 (if (eq mpointer nil)
18091 (setq mpointer "132")) ; top_left_arrow
18094 (setq x-pointer-shape (string-to-int mpointer))
18095 (set-mouse-color "white"))
18100 Or you can set the values of a variety of features in an alist, like
18106 default-frame-alist
18107 '((cursor-color . "white")
18108 (mouse-color . "white")
18109 (foreground-color . "white")
18110 (background-color . "DodgerBlue4")
18111 ;; (cursor-type . bar)
18112 (cursor-type . box)
18115 (tool-bar-lines . 0)
18116 (menu-bar-lines . 1)
18120 "-Misc-Fixed-Medium-R-Normal--20-200-75-75-C-100-ISO8859-1")
18126 Convert @kbd{@key{CTRL}-h} into @key{DEL} and @key{DEL}
18127 into @kbd{@key{CTRL}-h}.@*
18128 (Some older keyboards needed this, although I have not seen the
18133 ;; Translate `C-h' to <DEL>.
18134 ; (keyboard-translate ?\C-h ?\C-?)
18136 ;; Translate <DEL> to `C-h'.
18137 (keyboard-translate ?\C-? ?\C-h)
18141 @item Turn off a blinking cursor!
18145 (if (fboundp 'blink-cursor-mode)
18146 (blink-cursor-mode -1))
18151 or start GNU Emacs with the command @code{emacs -nbc}.
18154 @item When using `grep'@*
18155 @samp{-i}@w{ } Ignore case distinctions@*
18156 @samp{-n}@w{ } Prefix each line of output with line number@*
18157 @samp{-H}@w{ } Print the filename for each match.@*
18158 @samp{-e}@w{ } Protect patterns beginning with a hyphen character, @samp{-}
18161 (setq grep-command "grep -i -nH -e ")
18165 @c Evidently, no longer needed in GNU Emacs 22
18167 item Automatically uncompress compressed files when visiting them
18170 (load "uncompress")
18175 @item Find an existing buffer, even if it has a different name@*
18176 This avoids problems with symbolic links.
18179 (setq find-file-existing-other-name t)
18182 @item Set your language environment and default input method
18186 (set-language-environment "latin-1")
18187 ;; Remember you can enable or disable multilingual text input
18188 ;; with the @code{toggle-input-method'} (@kbd{C-\}) command
18189 (setq default-input-method "latin-1-prefix")
18193 If you want to write with Chinese `GB' characters, set this instead:
18197 (set-language-environment "Chinese-GB")
18198 (setq default-input-method "chinese-tonepy")
18203 @subsubheading Fixing Unpleasant Key Bindings
18204 @cindex Key bindings, fixing
18205 @cindex Bindings, key, fixing unpleasant
18207 Some systems bind keys unpleasantly. Sometimes, for example, the
18208 @key{CTRL} key appears in an awkward spot rather than at the far left
18211 Usually, when people fix these sorts of keybindings, they do not
18212 change their @file{~/.emacs} file. Instead, they bind the proper keys
18213 on their consoles with the @code{loadkeys} or @code{install-keymap}
18214 commands in their boot script and then include @code{xmodmap} commands
18215 in their @file{.xinitrc} or @file{.Xsession} file for X Windows.
18223 loadkeys /usr/share/keymaps/i386/qwerty/emacs2.kmap.gz
18225 install-keymap emacs2
18231 For a @file{.xinitrc} or @file{.Xsession} file when the @key{Caps
18232 Lock} key is at the far left of the home row:
18236 # Bind the key labeled `Caps Lock' to `Control'
18237 # (Such a broken user interface suggests that keyboard manufacturers
18238 # think that computers are typewriters from 1885.)
18240 xmodmap -e "clear Lock"
18241 xmodmap -e "add Control = Caps_Lock"
18247 In a @file{.xinitrc} or @file{.Xsession} file, to convert an @key{ALT}
18248 key to a @key{META} key:
18252 # Some ill designed keyboards have a key labeled ALT and no Meta
18253 xmodmap -e "keysym Alt_L = Meta_L Alt_L"
18258 @node Mode Line, , Miscellaneous, Emacs Initialization
18259 @section A Modified Mode Line
18260 @vindex default-mode-line-format
18261 @cindex Mode line format
18263 Finally, a feature I really like: a modified mode line.
18265 When I work over a network, I forget which machine I am using. Also,
18266 I tend to I lose track of where I am, and which line point is on.
18268 So I reset my mode line to look like this:
18271 -:-- foo.texi rattlesnake:/home/bob/ Line 1 (Texinfo Fill) Top
18274 I am visiting a file called @file{foo.texi}, on my machine
18275 @file{rattlesnake} in my @file{/home/bob} buffer. I am on line 1, in
18276 Texinfo mode, and am at the top of the buffer.
18279 My @file{.emacs} file has a section that looks like this:
18283 ;; Set a Mode Line that tells me which machine, which directory,
18284 ;; and which line I am on, plus the other customary information.
18285 (setq default-mode-line-format
18289 "mouse-1: select window, mouse-2: delete others ..."))
18290 mode-line-mule-info
18292 mode-line-frame-identification
18296 mode-line-buffer-identification
18299 (system-name) 0 (string-match "\\..+" (system-name))))
18304 "mouse-1: select window, mouse-2: delete others ..."))
18305 (line-number-mode " Line %l ")
18311 "mouse-1: select window, mouse-2: delete others ..."))
18312 (:eval (mode-line-mode-name))
18315 #("%n" 0 2 (help-echo "mouse-2: widen" local-map (keymap ...)))
18324 Here, I redefine the default mode line. Most of the parts are from
18325 the original; but I make a few changes. I set the @emph{default} mode
18326 line format so as to permit various modes, such as Info, to override
18329 Many elements in the list are self-explanatory:
18330 @code{mode-line-modified} is a variable that tells whether the buffer
18331 has been modified, @code{mode-name} tells the name of the mode, and so
18332 on. However, the format looks complicated because of two features we
18333 have not discussed.
18335 @cindex Properties, in mode line example
18336 The first string in the mode line is a dash or hyphen, @samp{-}. In
18337 the old days, it would have been specified simply as @code{"-"}. But
18338 nowadays, Emacs can add properties to a string, such as highlighting
18339 or, as in this case, a help feature. If you place your mouse cursor
18340 over the hyphen, some help information appears (By default, you must
18341 wait seven-tenths of a second before the information appears. You can
18342 change that timing by changing the value of @code{tooltip-delay}.)
18345 The new string format has a special syntax:
18348 #("-" 0 1 (help-echo "mouse-1: select window, ..."))
18352 The @code{#(} begins a list. The first element of the list is the
18353 string itself, just one @samp{-}. The second and third
18354 elements specify the range over which the fourth element applies. A
18355 range starts @emph{after} a character, so a zero means the range
18356 starts just before the first character; a 1 means that the range ends
18357 just after the first character. The third element is the property for
18358 the range. It consists of a property list, a
18359 property name, in this case, @samp{help-echo}, followed by a value, in this
18360 case, a string. The second, third, and fourth elements of this new
18361 string format can be repeated.
18363 @xref{Text Properties, , Text Properties, elisp, The GNU Emacs Lisp
18364 Reference Manual}, and see @ref{Mode Line Format, , Mode Line Format,
18365 elisp, The GNU Emacs Lisp Reference Manual}, for more information.
18367 @code{mode-line-buffer-identification}
18368 displays the current buffer name. It is a list
18369 beginning @code{(#("%12b" 0 4 @dots{}}.
18370 The @code{#(} begins the list.
18372 The @samp{"%12b"} displays the current buffer name, using the
18373 @code{buffer-name} function with which we are familiar; the `12'
18374 specifies the maximum number of characters that will be displayed.
18375 When a name has fewer characters, whitespace is added to fill out to
18376 this number. (Buffer names can and often should be longer than 12
18377 characters; this length works well in a typical 80 column wide
18380 @code{:eval} says to evaluate the following form and use the result as
18381 a string to display. In this case, the expression displays the first
18382 component of the full system name. The end of the first component is
18383 a @samp{.} (`period'), so I use the @code{string-match} function to
18384 tell me the length of the first component. The substring from the
18385 zeroth character to that length is the name of the machine.
18388 This is the expression:
18393 (system-name) 0 (string-match "\\..+" (system-name))))
18397 @samp{%[} and @samp{%]} cause a pair of square brackets
18398 to appear for each recursive editing level. @samp{%n} says `Narrow'
18399 when narrowing is in effect. @samp{%P} tells you the percentage of
18400 the buffer that is above the bottom of the window, or `Top', `Bottom',
18401 or `All'. (A lower case @samp{p} tell you the percentage above the
18402 @emph{top} of the window.) @samp{%-} inserts enough dashes to fill
18405 Remember, ``You don't have to like Emacs to like it'' --- your own
18406 Emacs can have different colors, different commands, and different
18407 keys than a default Emacs.
18409 On the other hand, if you want to bring up a plain `out of the box'
18410 Emacs, with no customization, type:
18417 This will start an Emacs that does @emph{not} load your
18418 @file{~/.emacs} initialization file. A plain, default Emacs. Nothing
18421 @node Debugging, Conclusion, Emacs Initialization, Top
18425 GNU Emacs has two debuggers, @code{debug} and @code{edebug}. The
18426 first is built into the internals of Emacs and is always with you;
18427 the second requires that you instrument a function before you can use it.
18429 Both debuggers are described extensively in @ref{Debugging, ,
18430 Debugging Lisp Programs, elisp, The GNU Emacs Lisp Reference Manual}.
18431 In this chapter, I will walk through a short example of each.
18434 * debug:: How to use the built-in debugger.
18435 * debug-on-entry:: Start debugging when you call a function.
18436 * debug-on-quit:: Start debugging when you quit with @kbd{C-g}.
18437 * edebug:: How to use Edebug, a source level debugger.
18438 * Debugging Exercises::
18441 @node debug, debug-on-entry, Debugging, Debugging
18442 @section @code{debug}
18445 Suppose you have written a function definition that is intended to
18446 return the sum of the numbers 1 through a given number. (This is the
18447 @code{triangle} function discussed earlier. @xref{Decrementing
18448 Example, , Example with Decrementing Counter}, for a discussion.)
18449 @c xref{Decrementing Loop,, Loop with a Decrementing Counter}, for a discussion.)
18451 However, your function definition has a bug. You have mistyped
18452 @samp{1=} for @samp{1-}. Here is the broken definition:
18454 @findex triangle-bugged
18457 (defun triangle-bugged (number)
18458 "Return sum of numbers 1 through NUMBER inclusive."
18460 (while (> number 0)
18461 (setq total (+ total number))
18462 (setq number (1= number))) ; @r{Error here.}
18467 If you are reading this in Info, you can evaluate this definition in
18468 the normal fashion. You will see @code{triangle-bugged} appear in the
18472 Now evaluate the @code{triangle-bugged} function with an
18476 (triangle-bugged 4)
18480 In a recent GNU Emacs, you will create and enter a @file{*Backtrace*}
18486 ---------- Buffer: *Backtrace* ----------
18487 Debugger entered--Lisp error: (void-function 1=)
18489 (setq number (1= number))
18490 (while (> number 0) (setq total (+ total number))
18491 (setq number (1= number)))
18492 (let ((total 0)) (while (> number 0) (setq total ...)
18493 (setq number ...)) total)
18497 eval((triangle-bugged 4))
18498 eval-last-sexp-1(nil)
18499 eval-last-sexp(nil)
18500 call-interactively(eval-last-sexp)
18501 ---------- Buffer: *Backtrace* ----------
18506 (I have reformatted this example slightly; the debugger does not fold
18507 long lines. As usual, you can quit the debugger by typing @kbd{q} in
18508 the @file{*Backtrace*} buffer.)
18510 In practice, for a bug as simple as this, the `Lisp error' line will
18511 tell you what you need to know to correct the definition. The
18512 function @code{1=} is `void'.
18516 In GNU Emacs 20 and before, you will see:
18519 Symbol's function definition is void:@: 1=
18523 which has the same meaning as the @file{*Backtrace*} buffer line in
18527 However, suppose you are not quite certain what is going on?
18528 You can read the complete backtrace.
18530 In this case, you need to run a recent GNU Emacs, which automatically
18531 starts the debugger that puts you in the @file{*Backtrace*} buffer; or
18532 else, you need to start the debugger manually as described below.
18534 Read the @file{*Backtrace*} buffer from the bottom up; it tells you
18535 what Emacs did that led to the error. Emacs made an interactive call
18536 to @kbd{C-x C-e} (@code{eval-last-sexp}), which led to the evaluation
18537 of the @code{triangle-bugged} expression. Each line above tells you
18538 what the Lisp interpreter evaluated next.
18541 The third line from the top of the buffer is
18544 (setq number (1= number))
18548 Emacs tried to evaluate this expression; in order to do so, it tried
18549 to evaluate the inner expression shown on the second line from the
18558 This is where the error occurred; as the top line says:
18561 Debugger entered--Lisp error: (void-function 1=)
18565 You can correct the mistake, re-evaluate the function definition, and
18566 then run your test again.
18568 @node debug-on-entry, debug-on-quit, debug, Debugging
18569 @section @code{debug-on-entry}
18570 @findex debug-on-entry
18572 A recent GNU Emacs starts the debugger automatically when your
18573 function has an error.
18576 GNU Emacs version 20 and before did not; it simply
18577 presented you with an error message. You had to start the debugger
18581 Incidentally, you can start the debugger manually for all versions of
18582 Emacs; the advantage is that the debugger runs even if you do not have
18583 a bug in your code. Sometimes your code will be free of bugs!
18585 You can enter the debugger when you call the function by calling
18586 @code{debug-on-entry}.
18593 M-x debug-on-entry RET triangle-bugged RET
18598 Now, evaluate the following:
18601 (triangle-bugged 5)
18605 All versions of Emacs will create a @file{*Backtrace*} buffer and tell
18606 you that it is beginning to evaluate the @code{triangle-bugged}
18611 ---------- Buffer: *Backtrace* ----------
18612 Debugger entered--entering a function:
18613 * triangle-bugged(5)
18614 eval((triangle-bugged 5))
18617 eval-last-sexp-1(nil)
18618 eval-last-sexp(nil)
18619 call-interactively(eval-last-sexp)
18620 ---------- Buffer: *Backtrace* ----------
18624 In the @file{*Backtrace*} buffer, type @kbd{d}. Emacs will evaluate
18625 the first expression in @code{triangle-bugged}; the buffer will look
18630 ---------- Buffer: *Backtrace* ----------
18631 Debugger entered--beginning evaluation of function call form:
18632 * (let ((total 0)) (while (> number 0) (setq total ...)
18633 (setq number ...)) total)
18634 * triangle-bugged(5)
18635 eval((triangle-bugged 5))
18638 eval-last-sexp-1(nil)
18639 eval-last-sexp(nil)
18640 call-interactively(eval-last-sexp)
18641 ---------- Buffer: *Backtrace* ----------
18646 Now, type @kbd{d} again, eight times, slowly. Each time you type
18647 @kbd{d}, Emacs will evaluate another expression in the function
18651 Eventually, the buffer will look like this:
18655 ---------- Buffer: *Backtrace* ----------
18656 Debugger entered--beginning evaluation of function call form:
18657 * (setq number (1= number))
18658 * (while (> number 0) (setq total (+ total number))
18659 (setq number (1= number)))
18662 * (let ((total 0)) (while (> number 0) (setq total ...)
18663 (setq number ...)) total)
18664 * triangle-bugged(5)
18665 eval((triangle-bugged 5))
18668 eval-last-sexp-1(nil)
18669 eval-last-sexp(nil)
18670 call-interactively(eval-last-sexp)
18671 ---------- Buffer: *Backtrace* ----------
18677 Finally, after you type @kbd{d} two more times, Emacs will reach the
18678 error, and the top two lines of the @file{*Backtrace*} buffer will look
18683 ---------- Buffer: *Backtrace* ----------
18684 Debugger entered--Lisp error: (void-function 1=)
18687 ---------- Buffer: *Backtrace* ----------
18691 By typing @kbd{d}, you were able to step through the function.
18693 You can quit a @file{*Backtrace*} buffer by typing @kbd{q} in it; this
18694 quits the trace, but does not cancel @code{debug-on-entry}.
18696 @findex cancel-debug-on-entry
18697 To cancel the effect of @code{debug-on-entry}, call
18698 @code{cancel-debug-on-entry} and the name of the function, like this:
18701 M-x cancel-debug-on-entry RET triangle-bugged RET
18705 (If you are reading this in Info, cancel @code{debug-on-entry} now.)
18707 @node debug-on-quit, edebug, debug-on-entry, Debugging
18708 @section @code{debug-on-quit} and @code{(debug)}
18710 In addition to setting @code{debug-on-error} or calling @code{debug-on-entry},
18711 there are two other ways to start @code{debug}.
18713 @findex debug-on-quit
18714 You can start @code{debug} whenever you type @kbd{C-g}
18715 (@code{keyboard-quit}) by setting the variable @code{debug-on-quit} to
18716 @code{t}. This is useful for debugging infinite loops.
18719 @cindex @code{(debug)} in code
18720 Or, you can insert a line that says @code{(debug)} into your code
18721 where you want the debugger to start, like this:
18725 (defun triangle-bugged (number)
18726 "Return sum of numbers 1 through NUMBER inclusive."
18728 (while (> number 0)
18729 (setq total (+ total number))
18730 (debug) ; @r{Start debugger.}
18731 (setq number (1= number))) ; @r{Error here.}
18736 The @code{debug} function is described in detail in @ref{Debugger, ,
18737 The Lisp Debugger, elisp, The GNU Emacs Lisp Reference Manual}.
18739 @node edebug, Debugging Exercises, debug-on-quit, Debugging
18740 @section The @code{edebug} Source Level Debugger
18741 @cindex Source level debugger
18744 Edebug is a source level debugger. Edebug normally displays the
18745 source of the code you are debugging, with an arrow at the left that
18746 shows which line you are currently executing.
18748 You can walk through the execution of a function, line by line, or run
18749 quickly until reaching a @dfn{breakpoint} where execution stops.
18751 Edebug is described in @ref{edebug, , Edebug, elisp, The GNU Emacs
18752 Lisp Reference Manual}.
18755 Here is a bugged function definition for @code{triangle-recursively}.
18756 @xref{Recursive triangle function, , Recursion in place of a counter},
18757 for a review of it.
18761 (defun triangle-recursively-bugged (number)
18762 "Return sum of numbers 1 through NUMBER inclusive.
18767 (triangle-recursively-bugged
18768 (1= number))))) ; @r{Error here.}
18773 Normally, you would install this definition by positioning your cursor
18774 after the function's closing parenthesis and typing @kbd{C-x C-e}
18775 (@code{eval-last-sexp}) or else by positioning your cursor within the
18776 definition and typing @kbd{C-M-x} (@code{eval-defun}). (By default,
18777 the @code{eval-defun} command works only in Emacs Lisp mode or in Lisp
18781 However, to prepare this function definition for Edebug, you must
18782 first @dfn{instrument} the code using a different command. You can do
18783 this by positioning your cursor within or just after the definition
18787 M-x edebug-defun RET
18791 This will cause Emacs to load Edebug automatically if it is not
18792 already loaded, and properly instrument the function.
18794 After instrumenting the function, place your cursor after the
18795 following expression and type @kbd{C-x C-e} (@code{eval-last-sexp}):
18798 (triangle-recursively-bugged 3)
18802 You will be jumped back to the source for
18803 @code{triangle-recursively-bugged} and the cursor positioned at the
18804 beginning of the @code{if} line of the function. Also, you will see
18805 an arrowhead at the left hand side of that line. The arrowhead marks
18806 the line where the function is executing. (In the following examples,
18807 we show the arrowhead with @samp{=>}; in a windowing system, you may
18808 see the arrowhead as a solid triangle in the window `fringe'.)
18811 =>@point{}(if (= number 1)
18816 In the example, the location of point is displayed with a star,
18817 @samp{@point{}} (in Info, it is displayed as @samp{-!-}).
18820 In the example, the location of point is displayed as @samp{@point{}}
18821 (in a printed book, it is displayed with a five pointed star).
18824 If you now press @key{SPC}, point will move to the next expression to
18825 be executed; the line will look like this:
18828 =>(if @point{}(= number 1)
18832 As you continue to press @key{SPC}, point will move from expression to
18833 expression. At the same time, whenever an expression returns a value,
18834 that value will be displayed in the echo area. For example, after you
18835 move point past @code{number}, you will see the following:
18838 Result: 3 (#o3, #x3, ?\C-c)
18842 This means the value of @code{number} is 3, which is octal three,
18843 hexadecimal three, and @sc{ascii} `control-c' (the third letter of the
18844 alphabet, in case you need to know this information).
18846 You can continue moving through the code until you reach the line with
18847 the error. Before evaluation, that line looks like this:
18850 => @point{}(1= number))))) ; @r{Error here.}
18855 When you press @key{SPC} once again, you will produce an error message
18859 Symbol's function definition is void:@: 1=
18865 Press @kbd{q} to quit Edebug.
18867 To remove instrumentation from a function definition, simply
18868 re-evaluate it with a command that does not instrument it.
18869 For example, you could place your cursor after the definition's
18870 closing parenthesis and type @kbd{C-x C-e}.
18872 Edebug does a great deal more than walk with you through a function.
18873 You can set it so it races through on its own, stopping only at an
18874 error or at specified stopping points; you can cause it to display the
18875 changing values of various expressions; you can find out how many
18876 times a function is called, and more.
18878 Edebug is described in @ref{edebug, , Edebug, elisp, The GNU Emacs
18879 Lisp Reference Manual}.
18882 @node Debugging Exercises, , edebug, Debugging
18883 @section Debugging Exercises
18887 Install the @code{count-words-region} function and then cause it to
18888 enter the built-in debugger when you call it. Run the command on a
18889 region containing two words. You will need to press @kbd{d} a
18890 remarkable number of times. On your system, is a `hook' called after
18891 the command finishes? (For information on hooks, see @ref{Command
18892 Overview, , Command Loop Overview, elisp, The GNU Emacs Lisp Reference
18896 Copy @code{count-words-region} into the @file{*scratch*} buffer,
18897 instrument the function for Edebug, and walk through its execution.
18898 The function does not need to have a bug, although you can introduce
18899 one if you wish. If the function lacks a bug, the walk-through
18900 completes without problems.
18903 While running Edebug, type @kbd{?} to see a list of all the Edebug commands.
18904 (The @code{global-edebug-prefix} is usually @kbd{C-x X}, i.e.@:
18905 @kbd{@key{CTRL}-x} followed by an upper case @kbd{X}; use this prefix
18906 for commands made outside of the Edebug debugging buffer.)
18909 In the Edebug debugging buffer, use the @kbd{p}
18910 (@code{edebug-bounce-point}) command to see where in the region the
18911 @code{count-words-region} is working.
18914 Move point to some spot further down the function and then type the
18915 @kbd{h} (@code{edebug-goto-here}) command to jump to that location.
18918 Use the @kbd{t} (@code{edebug-trace-mode}) command to cause Edebug to
18919 walk through the function on its own; use an upper case @kbd{T} for
18920 @code{edebug-Trace-fast-mode}.
18923 Set a breakpoint, then run Edebug in Trace mode until it reaches the
18927 @node Conclusion, the-the, Debugging, Top
18928 @chapter Conclusion
18930 We have now reached the end of this Introduction. You have now
18931 learned enough about programming in Emacs Lisp to set values, to write
18932 simple @file{.emacs} files for yourself and your friends, and write
18933 simple customizations and extensions to Emacs.
18935 This is a place to stop. Or, if you wish, you can now go onward, and
18938 You have learned some of the basic nuts and bolts of programming. But
18939 only some. There are a great many more brackets and hinges that are
18940 easy to use that we have not touched.
18942 A path you can follow right now lies among the sources to GNU Emacs
18945 @cite{The GNU Emacs Lisp Reference Manual}.
18948 @ref{Top, , The GNU Emacs Lisp Reference Manual, elisp, The GNU
18949 Emacs Lisp Reference Manual}.
18952 The Emacs Lisp sources are an adventure. When you read the sources and
18953 come across a function or expression that is unfamiliar, you need to
18954 figure out or find out what it does.
18956 Go to the Reference Manual. It is a thorough, complete, and fairly
18957 easy-to-read description of Emacs Lisp. It is written not only for
18958 experts, but for people who know what you know. (The @cite{Reference
18959 Manual} comes with the standard GNU Emacs distribution. Like this
18960 introduction, it comes as a Texinfo source file, so you can read it
18961 on-line and as a typeset, printed book.)
18963 Go to the other on-line help that is part of GNU Emacs: the on-line
18964 documentation for all functions and variables, and @code{find-tags},
18965 the program that takes you to sources.
18967 Here is an example of how I explore the sources. Because of its name,
18968 @file{simple.el} is the file I looked at first, a long time ago. As
18969 it happens some of the functions in @file{simple.el} are complicated,
18970 or at least look complicated at first sight. The @code{open-line}
18971 function, for example, looks complicated.
18973 You may want to walk through this function slowly, as we did with the
18974 @code{forward-sentence} function. (@xref{forward-sentence, The
18975 @code{forward-sentence} function}.) Or you may want to skip that
18976 function and look at another, such as @code{split-line}. You don't
18977 need to read all the functions. According to
18978 @code{count-words-in-defun}, the @code{split-line} function contains
18979 102 words and symbols.
18981 Even though it is short, @code{split-line} contains expressions
18982 we have not studied: @code{skip-chars-forward}, @code{indent-to},
18983 @code{current-column} and @code{insert-and-inherit}.
18985 Consider the @code{skip-chars-forward} function. (It is part of the
18986 function definition for @code{back-to-indentation}, which is shown in
18987 @ref{Review, , Review}.)
18989 In GNU Emacs, you can find out more about @code{skip-chars-forward} by
18990 typing @kbd{C-h f} (@code{describe-function}) and the name of the
18991 function. This gives you the function documentation.
18993 You may be able to guess what is done by a well named function such as
18994 @code{indent-to}; or you can look it up, too. Incidentally, the
18995 @code{describe-function} function itself is in @file{help.el}; it is
18996 one of those long, but decipherable functions. You can look up
18997 @code{describe-function} using the @kbd{C-h f} command!
18999 In this instance, since the code is Lisp, the @file{*Help*} buffer
19000 contains the name of the library containing the function's source.
19001 You can put point over the name of the library and press the RET key,
19002 which in this situation is bound to @code{help-follow}, and be taken
19003 directly to the source, in the same way as @kbd{M-.}
19006 The definition for @code{describe-function} illustrates how to
19007 customize the @code{interactive} expression without using the standard
19008 character codes; and it shows how to create a temporary buffer.
19010 (The @code{indent-to} function is written in C rather than Emacs Lisp;
19011 it is a `built-in' function. @code{help-follow} takes you to its
19012 source as does @code{find-tag}, when properly set up.)
19014 You can look at a function's source using @code{find-tag}, which is
19015 bound to @kbd{M-.} Finally, you can find out what the Reference
19016 Manual has to say by visiting the manual in Info, and typing @kbd{i}
19017 (@code{Info-index}) and the name of the function, or by looking up the
19018 function in the index to a printed copy of the manual.
19020 Similarly, you can find out what is meant by
19021 @code{insert-and-inherit}.
19023 Other interesting source files include @file{paragraphs.el},
19024 @file{loaddefs.el}, and @file{loadup.el}. The @file{paragraphs.el}
19025 file includes short, easily understood functions as well as longer
19026 ones. The @file{loaddefs.el} file contains the many standard
19027 autoloads and many keymaps. I have never looked at it all; only at
19028 parts. @file{loadup.el} is the file that loads the standard parts of
19029 Emacs; it tells you a great deal about how Emacs is built.
19030 (@xref{Building Emacs, , Building Emacs, elisp, The GNU Emacs Lisp
19031 Reference Manual}, for more about building.)
19033 As I said, you have learned some nuts and bolts; however, and very
19034 importantly, we have hardly touched major aspects of programming; I
19035 have said nothing about how to sort information, except to use the
19036 predefined @code{sort} function; I have said nothing about how to store
19037 information, except to use variables and lists; I have said nothing
19038 about how to write programs that write programs. These are topics for
19039 another, and different kind of book, a different kind of learning.
19041 What you have done is learn enough for much practical work with GNU
19042 Emacs. What you have done is get started. This is the end of a
19045 @c ================ Appendix ================
19047 @node the-the, Kill Ring, Conclusion, Top
19048 @appendix The @code{the-the} Function
19050 @cindex Duplicated words function
19051 @cindex Words, duplicated
19053 Sometimes when you you write text, you duplicate words---as with ``you
19054 you'' near the beginning of this sentence. I find that most
19055 frequently, I duplicate ``the''; hence, I call the function for
19056 detecting duplicated words, @code{the-the}.
19059 As a first step, you could use the following regular expression to
19060 search for duplicates:
19063 \\(\\w+[ \t\n]+\\)\\1
19067 This regexp matches one or more word-constituent characters followed
19068 by one or more spaces, tabs, or newlines. However, it does not detect
19069 duplicated words on different lines, since the ending of the first
19070 word, the end of the line, is different from the ending of the second
19071 word, a space. (For more information about regular expressions, see
19072 @ref{Regexp Search, , Regular Expression Searches}, as well as
19073 @ref{Regexps, , Syntax of Regular Expressions, emacs, The GNU Emacs
19074 Manual}, and @ref{Regular Expressions, , Regular Expressions, elisp,
19075 The GNU Emacs Lisp Reference Manual}.)
19077 You might try searching just for duplicated word-constituent
19078 characters but that does not work since the pattern detects doubles
19079 such as the two occurrences of `th' in `with the'.
19081 Another possible regexp searches for word-constituent characters
19082 followed by non-word-constituent characters, reduplicated. Here,
19083 @w{@samp{\\w+}} matches one or more word-constituent characters and
19084 @w{@samp{\\W*}} matches zero or more non-word-constituent characters.
19087 \\(\\(\\w+\\)\\W*\\)\\1
19093 Here is the pattern that I use. It is not perfect, but good enough.
19094 @w{@samp{\\b}} matches the empty string, provided it is at the beginning
19095 or end of a word; @w{@samp{[^@@ \n\t]+}} matches one or more occurrences of
19096 any characters that are @emph{not} an @@-sign, space, newline, or tab.
19099 \\b\\([^@@ \n\t]+\\)[ \n\t]+\\1\\b
19102 One can write more complicated expressions, but I found that this
19103 expression is good enough, so I use it.
19105 Here is the @code{the-the} function, as I include it in my
19106 @file{.emacs} file, along with a handy global key binding:
19111 "Search forward for for a duplicated word."
19113 (message "Searching for for duplicated words ...")
19117 ;; This regexp is not perfect
19118 ;; but is fairly good over all:
19119 (if (re-search-forward
19120 "\\b\\([^@@ \n\t]+\\)[ \n\t]+\\1\\b" nil 'move)
19121 (message "Found duplicated word.")
19122 (message "End of buffer")))
19126 ;; Bind `the-the' to C-c \
19127 (global-set-key "\C-c\\" 'the-the)
19136 one two two three four five
19141 You can substitute the other regular expressions shown above in the
19142 function definition and try each of them on this list.
19144 @node Kill Ring, Full Graph, the-the, Top
19145 @appendix Handling the Kill Ring
19146 @cindex Kill ring handling
19147 @cindex Handling the kill ring
19148 @cindex Ring, making a list like a
19150 The kill ring is a list that is transformed into a ring by the
19151 workings of the @code{current-kill} function. The @code{yank} and
19152 @code{yank-pop} commands use the @code{current-kill} function.
19154 This appendix describes the @code{current-kill} function as well as
19155 both the @code{yank} and the @code{yank-pop} commands, but first,
19156 consider the workings of the kill ring.
19159 * What the Kill Ring Does::
19161 * yank:: Paste a copy of a clipped element.
19162 * yank-pop:: Insert element pointed to.
19166 @node What the Kill Ring Does, current-kill, Kill Ring, Kill Ring
19168 @unnumberedsec What the Kill Ring Does
19172 The kill ring has a default maximum length of sixty items; this number
19173 is too large for an explanation. Instead, set it to four. Please
19174 evaluate the following:
19178 (setq old-kill-ring-max kill-ring-max)
19179 (setq kill-ring-max 4)
19184 Then, please copy each line of the following indented example into the
19185 kill ring. You may kill each line with @kbd{C-k} or mark it and copy
19189 (In a read-only buffer, such as the @file{*info*} buffer, the kill
19190 command, @kbd{C-k} (@code{kill-line}), will not remove the text,
19191 merely copy it to the kill ring. However, your machine may beep at
19192 you. Alternatively, for silence, you may copy the region of each line
19193 with the @kbd{M-w} (@code{kill-ring-save}) command. You must mark
19194 each line for this command to succeed, but it does not matter at which
19195 end you put point or mark.)
19199 Please invoke the calls in order, so that five elements attempt to
19200 fill the kill ring:
19205 second piece of text
19207 fourth line of text
19214 Then find the value of @code{kill-ring} by evaluating
19226 ("fifth bit of text" "fourth line of text"
19227 "third line" "second piece of text")
19232 The first element, @samp{first some text}, was dropped.
19235 To return to the old value for the length of the kill ring, evaluate:
19238 (setq kill-ring-max old-kill-ring-max)
19241 @node current-kill, yank, What the Kill Ring Does, Kill Ring
19242 @comment node-name, next, previous, up
19243 @appendixsec The @code{current-kill} Function
19244 @findex current-kill
19246 The @code{current-kill} function changes the element in the kill ring
19247 to which @code{kill-ring-yank-pointer} points. (Also, the
19248 @code{kill-new} function sets @code{kill-ring-yank-pointer} to point
19249 to the latest element of the the kill ring. The @code{kill-new}
19250 function is used directly or indirectly by @code{kill-append},
19251 @code{copy-region-as-kill}, @code{kill-ring-save}, @code{kill-line},
19252 and @code{kill-region}.)
19255 * Code for current-kill::
19256 * Understanding current-kill::
19259 @node Code for current-kill, Understanding current-kill, current-kill, current-kill
19261 @unnumberedsubsec The code for @code{current-kill}
19266 The @code{current-kill} function is used by @code{yank} and by
19267 @code{yank-pop}. Here is the code for @code{current-kill}:
19271 (defun current-kill (n &optional do-not-move)
19272 "Rotate the yanking point by N places, and then return that kill.
19273 If N is zero, `interprogram-paste-function' is set, and calling it
19274 returns a string, then that string is added to the front of the
19275 kill ring and returned as the latest kill.
19278 If optional arg DO-NOT-MOVE is non-nil, then don't actually move the
19279 yanking point; just return the Nth kill forward."
19280 (let ((interprogram-paste (and (= n 0)
19281 interprogram-paste-function
19282 (funcall interprogram-paste-function))))
19285 (if interprogram-paste
19287 ;; Disable the interprogram cut function when we add the new
19288 ;; text to the kill ring, so Emacs doesn't try to own the
19289 ;; selection, with identical text.
19290 (let ((interprogram-cut-function nil))
19291 (kill-new interprogram-paste))
19292 interprogram-paste)
19295 (or kill-ring (error "Kill ring is empty"))
19296 (let ((ARGth-kill-element
19297 (nthcdr (mod (- n (length kill-ring-yank-pointer))
19298 (length kill-ring))
19301 (setq kill-ring-yank-pointer ARGth-kill-element))
19302 (car ARGth-kill-element)))))
19306 Remember also that the @code{kill-new} function sets
19307 @code{kill-ring-yank-pointer} to the latest element of the the kill
19308 ring, which means that all the functions that call it set the value
19309 indirectly: @code{kill-append}, @code{copy-region-as-kill},
19310 @code{kill-ring-save}, @code{kill-line}, and @code{kill-region}.
19313 Here is the line in @code{kill-new}, which is explained in
19314 @ref{kill-new function, , The @code{kill-new} function}.
19317 (setq kill-ring-yank-pointer kill-ring)
19320 @node Understanding current-kill, , Code for current-kill, current-kill
19322 @unnumberedsubsec @code{current-kill} in Outline
19325 The @code{current-kill} function looks complex, but as usual, it can
19326 be understood by taking it apart piece by piece. First look at it in
19331 (defun current-kill (n &optional do-not-move)
19332 "Rotate the yanking point by N places, and then return that kill."
19338 This function takes two arguments, one of which is optional. It has a
19339 documentation string. It is @emph{not} interactive.
19342 * Body of current-kill::
19343 * Digression concerning error:: How to mislead humans, but not computers.
19344 * Determining the Element::
19347 @node Body of current-kill, Digression concerning error, Understanding current-kill, Understanding current-kill
19349 @unnumberedsubsubsec The Body of @code{current-kill}
19352 The body of the function definition is a @code{let} expression, which
19353 itself has a body as well as a @var{varlist}.
19355 The @code{let} expression declares a variable that will be only usable
19356 within the bounds of this function. This variable is called
19357 @code{interprogram-paste} and is for copying to another program. It
19358 is not for copying within this instance of GNU Emacs. Most window
19359 systems provide a facility for interprogram pasting. Sadly, that
19360 facility usually provides only for the last element. Most windowing
19361 systems have not adopted a ring of many possibilities, even though
19362 Emacs has provided it for decades.
19364 The @code{if} expression has two parts, one if there exists
19365 @code{interprogram-paste} and one if not.
19368 Let us consider the `if not' or else-part of the @code{current-kill}
19369 function. (The then-part uses the the @code{kill-new} function, which
19370 we have already described. @xref{kill-new function, , The
19371 @code{kill-new} function}.)
19375 (or kill-ring (error "Kill ring is empty"))
19376 (let ((ARGth-kill-element
19377 (nthcdr (mod (- n (length kill-ring-yank-pointer))
19378 (length kill-ring))
19381 (setq kill-ring-yank-pointer ARGth-kill-element))
19382 (car ARGth-kill-element))
19387 The code first checks whether the kill ring has content; otherwise it
19391 Note that the @code{or} expression is very similar to testing length
19398 (if (zerop (length kill-ring)) ; @r{if-part}
19399 (error "Kill ring is empty")) ; @r{then-part}
19405 If there is not anything in the kill ring, its length must be zero and
19406 an error message sent to the user: @samp{Kill ring is empty}. The
19407 @code{current-kill} function uses an @code{or} expression which is
19408 simpler. But an @code{if} expression reminds us what goes on.
19410 This @code{if} expression uses the function @code{zerop} which returns
19411 true if the value it is testing is zero. When @code{zerop} tests
19412 true, the then-part of the @code{if} is evaluated. The then-part is a
19413 list starting with the function @code{error}, which is a function that
19414 is similar to the @code{message} function
19415 (@pxref{message, , The @code{message} Function}) in that
19416 it prints a one-line message in the echo area. However, in addition
19417 to printing a message, @code{error} also stops evaluation of the
19418 function within which it is embedded. This means that the rest of the
19419 function will not be evaluated if the length of the kill ring is zero.
19421 Then the @code{current-kill} function selects the element to return.
19422 The selection depends on the number of places that @code{current-kill}
19423 rotates and on where @code{kill-ring-yank-pointer} points.
19425 Next, either the optional @code{do-not-move} argument is true or the
19426 current value of @code{kill-ring-yank-pointer} is set to point to the
19427 list. Finally, another expression returns the first element of the
19428 list even if the @code{do-not-move} argument is true.
19430 @node Digression concerning error, Determining the Element, Body of current-kill, Understanding current-kill
19432 @unnumberedsubsubsec Digression about the word `error'
19435 In my opinion, it is slightly misleading, at least to humans, to use
19436 the term `error' as the name of the @code{error} function. A better
19437 term would be `cancel'. Strictly speaking, of course, you cannot
19438 point to, much less rotate a pointer to a list that has no length, so
19439 from the point of view of the computer, the word `error' is correct.
19440 But a human expects to attempt this sort of thing, if only to find out
19441 whether the kill ring is full or empty. This is an act of
19444 From the human point of view, the act of exploration and discovery is
19445 not necessarily an error, and therefore should not be labelled as one,
19446 even in the bowels of a computer. As it is, the code in Emacs implies
19447 that a human who is acting virtuously, by exploring his or her
19448 environment, is making an error. This is bad. Even though the computer
19449 takes the same steps as it does when there is an `error', a term such as
19450 `cancel' would have a clearer connotation.
19452 @node Determining the Element, , Digression concerning error, Understanding current-kill
19454 @unnumberedsubsubsec Determining the Element
19457 Among other actions, the else-part of the @code{if} expression sets
19458 the value of @code{kill-ring-yank-pointer} to
19459 @code{ARGth-kill-element} when the kill ring has something in it and
19460 the value of @code{do-not-move} is @code{nil}.
19463 The code looks like this:
19467 (nthcdr (mod (- n (length kill-ring-yank-pointer))
19468 (length kill-ring))
19473 This needs some examination. Unless it is not supposed to move the
19474 pointer, the @code{current-kill} function changes where
19475 @code{kill-ring-yank-pointer} points.
19477 @w{@code{(setq kill-ring-yank-pointer ARGth-kill-element))}}
19478 expression does. Also, clearly, @code{ARGth-kill-element} is being
19479 set to be equal to some @sc{cdr} of the kill ring, using the
19480 @code{nthcdr} function that is described in an earlier section.
19481 (@xref{copy-region-as-kill}.) How does it do this?
19483 As we have seen before (@pxref{nthcdr}), the @code{nthcdr} function
19484 works by repeatedly taking the @sc{cdr} of a list---it takes the
19485 @sc{cdr} of the @sc{cdr} of the @sc{cdr} @dots{}
19488 The two following expressions produce the same result:
19492 (setq kill-ring-yank-pointer (cdr kill-ring))
19494 (setq kill-ring-yank-pointer (nthcdr 1 kill-ring))
19498 However, the @code{nthcdr} expression is more complicated. It uses
19499 the @code{mod} function to determine which @sc{cdr} to select.
19501 (You will remember to look at inner functions first; indeed, we will
19502 have to go inside the @code{mod}.)
19504 The @code{mod} function returns the value of its first argument modulo
19505 the second; that is to say, it returns the remainder after dividing
19506 the first argument by the second. The value returned has the same
19507 sign as the second argument.
19515 @result{} 0 ;; @r{because there is no remainder}
19522 In this case, the first argument is often smaller than the second.
19534 We can guess what the @code{-} function does. It is like @code{+} but
19535 subtracts instead of adds; the @code{-} function subtracts its second
19536 argument from its first. Also, we already know what the @code{length}
19537 function does (@pxref{length}). It returns the length of a list.
19539 And @code{n} is the name of the required argument to the
19540 @code{current-kill} function.
19543 So when the first argument to @code{nthcdr} is zero, the @code{nthcdr}
19544 expression returns the whole list, as you can see by evaluating the
19549 ;; kill-ring-yank-pointer @r{and} kill-ring @r{have a length of four}
19550 ;; @r{and} (mod (- 0 4) 4) @result{} 0
19551 (nthcdr (mod (- 0 4) 4)
19552 '("fourth line of text"
19554 "second piece of text"
19555 "first some text"))
19560 When the first argument to the @code{current-kill} function is one,
19561 the @code{nthcdr} expression returns the list without its first
19566 (nthcdr (mod (- 1 4) 4)
19567 '("fourth line of text"
19569 "second piece of text"
19570 "first some text"))
19574 @cindex @samp{global variable} defined
19575 @cindex @samp{variable, global}, defined
19576 Incidentally, both @code{kill-ring} and @code{kill-ring-yank-pointer}
19577 are @dfn{global variables}. That means that any expression in Emacs
19578 Lisp can access them. They are not like the local variables set by
19579 @code{let} or like the symbols in an argument list.
19580 Local variables can only be accessed
19581 within the @code{let} that defines them or the function that specifies
19582 them in an argument list (and within expressions called by them).
19585 @c texi2dvi fails when the name of the section is within ifnottex ...
19586 (@xref{Prevent confusion, , @code{let} Prevents Confusion}, and
19587 @ref{defun, , The @code{defun} Special Form}.)
19590 @node yank, yank-pop, current-kill, Kill Ring
19591 @comment node-name, next, previous, up
19592 @appendixsec @code{yank}
19595 After learning about @code{current-kill}, the code for the
19596 @code{yank} function is almost easy.
19598 The @code{yank} function does not use the
19599 @code{kill-ring-yank-pointer} variable directly. It calls
19600 @code{insert-for-yank} which calls @code{current-kill} which sets the
19601 @code{kill-ring-yank-pointer} variable.
19604 The code looks like this:
19609 (defun yank (&optional arg)
19610 "Reinsert (\"paste\") the last stretch of killed text.
19611 More precisely, reinsert the stretch of killed text most recently
19612 killed OR yanked. Put point at end, and set mark at beginning.
19613 With just \\[universal-argument] as argument, same but put point at
19614 beginning (and mark at end). With argument N, reinsert the Nth most
19615 recently killed stretch of killed text.
19617 When this command inserts killed text into the buffer, it honors
19618 `yank-excluded-properties' and `yank-handler' as described in the
19619 doc string for `insert-for-yank-1', which see.
19621 See also the command \\[yank-pop]."
19625 (setq yank-window-start (window-start))
19626 ;; If we don't get all the way thru, make last-command indicate that
19627 ;; for the following command.
19628 (setq this-command t)
19629 (push-mark (point))
19632 (insert-for-yank (current-kill (cond
19637 ;; This is like exchange-point-and-mark,
19638 ;; but doesn't activate the mark.
19639 ;; It is cleaner to avoid activation, even though the command
19640 ;; loop would deactivate the mark because we inserted text.
19641 (goto-char (prog1 (mark t)
19642 (set-marker (mark-marker) (point) (current-buffer)))))
19645 ;; If we do get all the way thru, make this-command indicate that.
19646 (if (eq this-command t)
19647 (setq this-command 'yank))
19652 The key expression is @code{insert-for-yank}, which inserts the string
19653 returned by @code{current-kill}, but removes some text properties from
19656 However, before getting to that expression, the function sets the value
19657 of @code{yank-window-start} to the position returned by the
19658 @code{(window-start)} expression, the position at which the display
19659 currently starts. The @code{yank} function also sets
19660 @code{this-command} and pushes the mark.
19662 After it yanks the appropriate element, if the optional argument is a
19663 @sc{cons} rather than a number or nothing, it puts point at beginning
19664 of the yanked text and mark at its end.
19666 (The @code{prog1} function is like @code{progn} but returns the value
19667 of its first argument rather than the value of its last argument. Its
19668 first argument is forced to return the buffer's mark as an integer.
19669 You can see the documentation for these functions by placing point
19670 over them in this buffer and then typing @kbd{C-h f}
19671 (@code{describe-function}) followed by a @kbd{RET}; the default is the
19674 The last part of the function tells what to do when it succeeds.
19676 @node yank-pop, ring file, yank, Kill Ring
19677 @comment node-name, next, previous, up
19678 @appendixsec @code{yank-pop}
19681 After understanding @code{yank} and @code{current-kill}, you know how
19682 to approach the @code{yank-pop} function. Leaving out the
19683 documentation to save space, it looks like this:
19688 (defun yank-pop (&optional arg)
19691 (if (not (eq last-command 'yank))
19692 (error "Previous command was not a yank"))
19695 (setq this-command 'yank)
19696 (unless arg (setq arg 1))
19697 (let ((inhibit-read-only t)
19698 (before (< (point) (mark t))))
19702 (funcall (or yank-undo-function 'delete-region) (point) (mark t))
19703 (funcall (or yank-undo-function 'delete-region) (mark t) (point)))
19704 (setq yank-undo-function nil)
19707 (set-marker (mark-marker) (point) (current-buffer))
19708 (insert-for-yank (current-kill arg))
19709 ;; Set the window start back where it was in the yank command,
19711 (set-window-start (selected-window) yank-window-start t)
19715 ;; This is like exchange-point-and-mark,
19716 ;; but doesn't activate the mark.
19717 ;; It is cleaner to avoid activation, even though the command
19718 ;; loop would deactivate the mark because we inserted text.
19719 (goto-char (prog1 (mark t)
19720 (set-marker (mark-marker)
19722 (current-buffer))))))
19727 The function is interactive with a small @samp{p} so the prefix
19728 argument is processed and passed to the function. The command can
19729 only be used after a previous yank; otherwise an error message is
19730 sent. This check uses the variable @code{last-command} which is set
19731 by @code{yank} and is discussed elsewhere.
19732 (@xref{copy-region-as-kill}.)
19734 The @code{let} clause sets the variable @code{before} to true or false
19735 depending whether point is before or after mark and then the region
19736 between point and mark is deleted. This is the region that was just
19737 inserted by the previous yank and it is this text that will be
19740 @code{funcall} calls its first argument as a function, passing
19741 remaining arguments to it. The first argument is whatever the
19742 @code{or} expression returns. The two remaining arguments are the
19743 positions of point and mark set by the preceding @code{yank} command.
19745 There is more, but that is the hardest part.
19747 @node ring file, , yank-pop, Kill Ring
19748 @comment node-name, next, previous, up
19749 @appendixsec The @file{ring.el} File
19750 @cindex @file{ring.el} file
19752 Interestingly, GNU Emacs posses a file called @file{ring.el} that
19753 provides many of the features we just discussed. But functions such
19754 as @code{kill-ring-yank-pointer} do not use this library, possibly
19755 because they were written earlier.
19757 @node Full Graph, Free Software and Free Manuals, Kill Ring, Top
19758 @appendix A Graph with Labelled Axes
19760 Printed axes help you understand a graph. They convey scale. In an
19761 earlier chapter (@pxref{Readying a Graph, , Readying a Graph}), we
19762 wrote the code to print the body of a graph. Here we write the code
19763 for printing and labelling vertical and horizontal axes, along with the
19767 * Labelled Example::
19768 * print-graph Varlist:: @code{let} expression in @code{print-graph}.
19769 * print-Y-axis:: Print a label for the vertical axis.
19770 * print-X-axis:: Print a horizontal label.
19771 * Print Whole Graph:: The function to print a complete graph.
19774 @node Labelled Example, print-graph Varlist, Full Graph, Full Graph
19776 @unnumberedsec Labelled Example Graph
19779 Since insertions fill a buffer to the right and below point, the new
19780 graph printing function should first print the Y or vertical axis,
19781 then the body of the graph, and finally the X or horizontal axis.
19782 This sequence lays out for us the contents of the function:
19792 Print body of graph.
19799 Here is an example of how a finished graph should look:
19812 1 - ****************
19819 In this graph, both the vertical and the horizontal axes are labelled
19820 with numbers. However, in some graphs, the horizontal axis is time
19821 and would be better labelled with months, like this:
19835 Indeed, with a little thought, we can easily come up with a variety of
19836 vertical and horizontal labelling schemes. Our task could become
19837 complicated. But complications breed confusion. Rather than permit
19838 this, it is better choose a simple labelling scheme for our first
19839 effort, and to modify or replace it later.
19842 These considerations suggest the following outline for the
19843 @code{print-graph} function:
19847 (defun print-graph (numbers-list)
19848 "@var{documentation}@dots{}"
19849 (let ((height @dots{}
19853 (print-Y-axis height @dots{} )
19854 (graph-body-print numbers-list)
19855 (print-X-axis @dots{} )))
19859 We can work on each part of the @code{print-graph} function definition
19862 @node print-graph Varlist, print-Y-axis, Labelled Example, Full Graph
19863 @comment node-name, next, previous, up
19864 @appendixsec The @code{print-graph} Varlist
19865 @cindex @code{print-graph} varlist
19867 In writing the @code{print-graph} function, the first task is to write
19868 the varlist in the @code{let} expression. (We will leave aside for the
19869 moment any thoughts about making the function interactive or about the
19870 contents of its documentation string.)
19872 The varlist should set several values. Clearly, the top of the label
19873 for the vertical axis must be at least the height of the graph, which
19874 means that we must obtain this information here. Note that the
19875 @code{print-graph-body} function also requires this information. There
19876 is no reason to calculate the height of the graph in two different
19877 places, so we should change @code{print-graph-body} from the way we
19878 defined it earlier to take advantage of the calculation.
19880 Similarly, both the function for printing the X axis labels and the
19881 @code{print-graph-body} function need to learn the value of the width of
19882 each symbol. We can perform the calculation here and change the
19883 definition for @code{print-graph-body} from the way we defined it in the
19886 The length of the label for the horizontal axis must be at least as long
19887 as the graph. However, this information is used only in the function
19888 that prints the horizontal axis, so it does not need to be calculated here.
19890 These thoughts lead us directly to the following form for the varlist
19891 in the @code{let} for @code{print-graph}:
19895 (let ((height (apply 'max numbers-list)) ; @r{First version.}
19896 (symbol-width (length graph-blank)))
19901 As we shall see, this expression is not quite right.
19904 @node print-Y-axis, print-X-axis, print-graph Varlist, Full Graph
19905 @comment node-name, next, previous, up
19906 @appendixsec The @code{print-Y-axis} Function
19907 @cindex Axis, print vertical
19908 @cindex Y axis printing
19909 @cindex Vertical axis printing
19910 @cindex Print vertical axis
19912 The job of the @code{print-Y-axis} function is to print a label for
19913 the vertical axis that looks like this:
19931 The function should be passed the height of the graph, and then should
19932 construct and insert the appropriate numbers and marks.
19935 * print-Y-axis in Detail::
19936 * Height of label:: What height for the Y axis?
19937 * Compute a Remainder:: How to compute the remainder of a division.
19938 * Y Axis Element:: Construct a line for the Y axis.
19939 * Y-axis-column:: Generate a list of Y axis labels.
19940 * print-Y-axis Penultimate:: A not quite final version.
19943 @node print-Y-axis in Detail, Height of label, print-Y-axis, print-Y-axis
19945 @unnumberedsubsec The @code{print-Y-axis} Function in Detail
19948 It is easy enough to see in the figure what the Y axis label should
19949 look like; but to say in words, and then to write a function
19950 definition to do the job is another matter. It is not quite true to
19951 say that we want a number and a tic every five lines: there are only
19952 three lines between the @samp{1} and the @samp{5} (lines 2, 3, and 4),
19953 but four lines between the @samp{5} and the @samp{10} (lines 6, 7, 8,
19954 and 9). It is better to say that we want a number and a tic mark on
19955 the base line (number 1) and then that we want a number and a tic on
19956 the fifth line from the bottom and on every line that is a multiple of
19959 @node Height of label, Compute a Remainder, print-Y-axis in Detail, print-Y-axis
19961 @unnumberedsubsec What height should the label be?
19964 The next issue is what height the label should be? Suppose the maximum
19965 height of tallest column of the graph is seven. Should the highest
19966 label on the Y axis be @samp{5 -}, and should the graph stick up above
19967 the label? Or should the highest label be @samp{7 -}, and mark the peak
19968 of the graph? Or should the highest label be @code{10 -}, which is a
19969 multiple of five, and be higher than the topmost value of the graph?
19971 The latter form is preferred. Most graphs are drawn within rectangles
19972 whose sides are an integral number of steps long---5, 10, 15, and so
19973 on for a step distance of five. But as soon as we decide to use a
19974 step height for the vertical axis, we discover that the simple
19975 expression in the varlist for computing the height is wrong. The
19976 expression is @code{(apply 'max numbers-list)}. This returns the
19977 precise height, not the maximum height plus whatever is necessary to
19978 round up to the nearest multiple of five. A more complex expression
19981 As usual in cases like this, a complex problem becomes simpler if it is
19982 divided into several smaller problems.
19984 First, consider the case when the highest value of the graph is an
19985 integral multiple of five---when it is 5, 10, 15, or some higher
19986 multiple of five. We can use this value as the Y axis height.
19988 A fairly simply way to determine whether a number is a multiple of
19989 five is to divide it by five and see if the division results in a
19990 remainder. If there is no remainder, the number is a multiple of
19991 five. Thus, seven divided by five has a remainder of two, and seven
19992 is not an integral multiple of five. Put in slightly different
19993 language, more reminiscent of the classroom, five goes into seven
19994 once, with a remainder of two. However, five goes into ten twice,
19995 with no remainder: ten is an integral multiple of five.
19997 @node Compute a Remainder, Y Axis Element, Height of label, print-Y-axis
19998 @appendixsubsec Side Trip: Compute a Remainder
20000 @findex % @r{(remainder function)}
20001 @cindex Remainder function, @code{%}
20002 In Lisp, the function for computing a remainder is @code{%}. The
20003 function returns the remainder of its first argument divided by its
20004 second argument. As it happens, @code{%} is a function in Emacs Lisp
20005 that you cannot discover using @code{apropos}: you find nothing if you
20006 type @kbd{M-x apropos @key{RET} remainder @key{RET}}. The only way to
20007 learn of the existence of @code{%} is to read about it in a book such
20008 as this or in the Emacs Lisp sources.
20010 You can try the @code{%} function by evaluating the following two
20022 The first expression returns 2 and the second expression returns 0.
20024 To test whether the returned value is zero or some other number, we
20025 can use the @code{zerop} function. This function returns @code{t} if
20026 its argument, which must be a number, is zero.
20038 Thus, the following expression will return @code{t} if the height
20039 of the graph is evenly divisible by five:
20042 (zerop (% height 5))
20046 (The value of @code{height}, of course, can be found from @code{(apply
20047 'max numbers-list)}.)
20049 On the other hand, if the value of @code{height} is not a multiple of
20050 five, we want to reset the value to the next higher multiple of five.
20051 This is straightforward arithmetic using functions with which we are
20052 already familiar. First, we divide the value of @code{height} by five
20053 to determine how many times five goes into the number. Thus, five
20054 goes into twelve twice. If we add one to this quotient and multiply by
20055 five, we will obtain the value of the next multiple of five that is
20056 larger than the height. Five goes into twelve twice. Add one to two,
20057 and multiply by five; the result is fifteen, which is the next multiple
20058 of five that is higher than twelve. The Lisp expression for this is:
20061 (* (1+ (/ height 5)) 5)
20065 For example, if you evaluate the following, the result is 15:
20068 (* (1+ (/ 12 5)) 5)
20071 All through this discussion, we have been using `five' as the value
20072 for spacing labels on the Y axis; but we may want to use some other
20073 value. For generality, we should replace `five' with a variable to
20074 which we can assign a value. The best name I can think of for this
20075 variable is @code{Y-axis-label-spacing}.
20078 Using this term, and an @code{if} expression, we produce the
20083 (if (zerop (% height Y-axis-label-spacing))
20086 (* (1+ (/ height Y-axis-label-spacing))
20087 Y-axis-label-spacing))
20092 This expression returns the value of @code{height} itself if the height
20093 is an even multiple of the value of the @code{Y-axis-label-spacing} or
20094 else it computes and returns a value of @code{height} that is equal to
20095 the next higher multiple of the value of the @code{Y-axis-label-spacing}.
20097 We can now include this expression in the @code{let} expression of the
20098 @code{print-graph} function (after first setting the value of
20099 @code{Y-axis-label-spacing}):
20100 @vindex Y-axis-label-spacing
20104 (defvar Y-axis-label-spacing 5
20105 "Number of lines from one Y axis label to next.")
20110 (let* ((height (apply 'max numbers-list))
20111 (height-of-top-line
20112 (if (zerop (% height Y-axis-label-spacing))
20117 (* (1+ (/ height Y-axis-label-spacing))
20118 Y-axis-label-spacing)))
20119 (symbol-width (length graph-blank))))
20125 (Note use of the @code{let*} function: the initial value of height is
20126 computed once by the @code{(apply 'max numbers-list)} expression and
20127 then the resulting value of @code{height} is used to compute its
20128 final value. @xref{fwd-para let, , The @code{let*} expression}, for
20129 more about @code{let*}.)
20131 @node Y Axis Element, Y-axis-column, Compute a Remainder, print-Y-axis
20132 @appendixsubsec Construct a Y Axis Element
20134 When we print the vertical axis, we want to insert strings such as
20135 @w{@samp{5 -}} and @w{@samp{10 - }} every five lines.
20136 Moreover, we want the numbers and dashes to line up, so shorter
20137 numbers must be padded with leading spaces. If some of the strings
20138 use two digit numbers, the strings with single digit numbers must
20139 include a leading blank space before the number.
20141 @findex number-to-string
20142 To figure out the length of the number, the @code{length} function is
20143 used. But the @code{length} function works only with a string, not with
20144 a number. So the number has to be converted from being a number to
20145 being a string. This is done with the @code{number-to-string} function.
20150 (length (number-to-string 35))
20153 (length (number-to-string 100))
20159 (@code{number-to-string} is also called @code{int-to-string}; you will
20160 see this alternative name in various sources.)
20162 In addition, in each label, each number is followed by a string such
20163 as @w{@samp{ - }}, which we will call the @code{Y-axis-tic} marker.
20164 This variable is defined with @code{defvar}:
20169 (defvar Y-axis-tic " - "
20170 "String that follows number in a Y axis label.")
20174 The length of the Y label is the sum of the length of the Y axis tic
20175 mark and the length of the number of the top of the graph.
20178 (length (concat (number-to-string height) Y-axis-tic)))
20181 This value will be calculated by the @code{print-graph} function in
20182 its varlist as @code{full-Y-label-width} and passed on. (Note that we
20183 did not think to include this in the varlist when we first proposed it.)
20185 To make a complete vertical axis label, a tic mark is concatenated
20186 with a number; and the two together may be preceded by one or more
20187 spaces depending on how long the number is. The label consists of
20188 three parts: the (optional) leading spaces, the number, and the tic
20189 mark. The function is passed the value of the number for the specific
20190 row, and the value of the width of the top line, which is calculated
20191 (just once) by @code{print-graph}.
20195 (defun Y-axis-element (number full-Y-label-width)
20196 "Construct a NUMBERed label element.
20197 A numbered element looks like this ` 5 - ',
20198 and is padded as needed so all line up with
20199 the element for the largest number."
20202 (let* ((leading-spaces
20203 (- full-Y-label-width
20205 (concat (number-to-string number)
20210 (make-string leading-spaces ? )
20211 (number-to-string number)
20216 The @code{Y-axis-element} function concatenates together the leading
20217 spaces, if any; the number, as a string; and the tic mark.
20219 To figure out how many leading spaces the label will need, the
20220 function subtracts the actual length of the label---the length of the
20221 number plus the length of the tic mark---from the desired label width.
20223 @findex make-string
20224 Blank spaces are inserted using the @code{make-string} function. This
20225 function takes two arguments: the first tells it how long the string
20226 will be and the second is a symbol for the character to insert, in a
20227 special format. The format is a question mark followed by a blank
20228 space, like this, @samp{? }. @xref{Character Type, , Character Type,
20229 elisp, The GNU Emacs Lisp Reference Manual}, for a description of the
20230 syntax for characters. (Of course, you might want to replace the
20231 blank space by some other character @dots{} You know what to do.)
20233 The @code{number-to-string} function is used in the concatenation
20234 expression, to convert the number to a string that is concatenated
20235 with the leading spaces and the tic mark.
20237 @node Y-axis-column, print-Y-axis Penultimate, Y Axis Element, print-Y-axis
20238 @appendixsubsec Create a Y Axis Column
20240 The preceding functions provide all the tools needed to construct a
20241 function that generates a list of numbered and blank strings to insert
20242 as the label for the vertical axis:
20244 @findex Y-axis-column
20247 (defun Y-axis-column (height width-of-label)
20248 "Construct list of Y axis labels and blank strings.
20249 For HEIGHT of line above base and WIDTH-OF-LABEL."
20253 (while (> height 1)
20254 (if (zerop (% height Y-axis-label-spacing))
20255 ;; @r{Insert label.}
20258 (Y-axis-element height width-of-label)
20262 ;; @r{Else, insert blanks.}
20265 (make-string width-of-label ? )
20267 (setq height (1- height)))
20268 ;; @r{Insert base line.}
20270 (cons (Y-axis-element 1 width-of-label) Y-axis))
20271 (nreverse Y-axis)))
20275 In this function, we start with the value of @code{height} and
20276 repetitively subtract one from its value. After each subtraction, we
20277 test to see whether the value is an integral multiple of the
20278 @code{Y-axis-label-spacing}. If it is, we construct a numbered label
20279 using the @code{Y-axis-element} function; if not, we construct a
20280 blank label using the @code{make-string} function. The base line
20281 consists of the number one followed by a tic mark.
20284 @node print-Y-axis Penultimate, , Y-axis-column, print-Y-axis
20285 @appendixsubsec The Not Quite Final Version of @code{print-Y-axis}
20287 The list constructed by the @code{Y-axis-column} function is passed to
20288 the @code{print-Y-axis} function, which inserts the list as a column.
20290 @findex print-Y-axis
20293 (defun print-Y-axis (height full-Y-label-width)
20294 "Insert Y axis using HEIGHT and FULL-Y-LABEL-WIDTH.
20295 Height must be the maximum height of the graph.
20296 Full width is the width of the highest label element."
20297 ;; Value of height and full-Y-label-width
20298 ;; are passed by `print-graph'.
20301 (let ((start (point)))
20303 (Y-axis-column height full-Y-label-width))
20304 ;; @r{Place point ready for inserting graph.}
20306 ;; @r{Move point forward by value of} full-Y-label-width
20307 (forward-char full-Y-label-width)))
20311 The @code{print-Y-axis} uses the @code{insert-rectangle} function to
20312 insert the Y axis labels created by the @code{Y-axis-column} function.
20313 In addition, it places point at the correct position for printing the body of
20316 You can test @code{print-Y-axis}:
20324 Y-axis-label-spacing
20333 Copy the following expression:
20336 (print-Y-axis 12 5)
20340 Switch to the @file{*scratch*} buffer and place the cursor where you
20341 want the axis labels to start.
20344 Type @kbd{M-:} (@code{eval-expression}).
20347 Yank the @code{graph-body-print} expression into the minibuffer
20348 with @kbd{C-y} (@code{yank)}.
20351 Press @key{RET} to evaluate the expression.
20354 Emacs will print labels vertically, the top one being @w{@samp{10 -@w{
20355 }}}. (The @code{print-graph} function will pass the value of
20356 @code{height-of-top-line}, which in this case will end up as 15,
20357 thereby getting rid of what might appear as a bug.)
20360 @node print-X-axis, Print Whole Graph, print-Y-axis, Full Graph
20361 @appendixsec The @code{print-X-axis} Function
20362 @cindex Axis, print horizontal
20363 @cindex X axis printing
20364 @cindex Print horizontal axis
20365 @cindex Horizontal axis printing
20367 X axis labels are much like Y axis labels, except that the ticks are on a
20368 line above the numbers. Labels should look like this:
20377 The first tic is under the first column of the graph and is preceded by
20378 several blank spaces. These spaces provide room in rows above for the Y
20379 axis labels. The second, third, fourth, and subsequent ticks are all
20380 spaced equally, according to the value of @code{X-axis-label-spacing}.
20382 The second row of the X axis consists of numbers, preceded by several
20383 blank spaces and also separated according to the value of the variable
20384 @code{X-axis-label-spacing}.
20386 The value of the variable @code{X-axis-label-spacing} should itself be
20387 measured in units of @code{symbol-width}, since you may want to change
20388 the width of the symbols that you are using to print the body of the
20389 graph without changing the ways the graph is labelled.
20392 * Similarities differences:: Much like @code{print-Y-axis}, but not exactly.
20393 * X Axis Tic Marks:: Create tic marks for the horizontal axis.
20396 @node Similarities differences, X Axis Tic Marks, print-X-axis, print-X-axis
20398 @unnumberedsubsec Similarities and differences
20401 The @code{print-X-axis} function is constructed in more or less the
20402 same fashion as the @code{print-Y-axis} function except that it has
20403 two lines: the line of tic marks and the numbers. We will write a
20404 separate function to print each line and then combine them within the
20405 @code{print-X-axis} function.
20407 This is a three step process:
20411 Write a function to print the X axis tic marks, @code{print-X-axis-tic-line}.
20414 Write a function to print the X numbers, @code{print-X-axis-numbered-line}.
20417 Write a function to print both lines, the @code{print-X-axis} function,
20418 using @code{print-X-axis-tic-line} and
20419 @code{print-X-axis-numbered-line}.
20422 @node X Axis Tic Marks, , Similarities differences, print-X-axis
20423 @appendixsubsec X Axis Tic Marks
20425 The first function should print the X axis tic marks. We must specify
20426 the tic marks themselves and their spacing:
20430 (defvar X-axis-label-spacing
20431 (if (boundp 'graph-blank)
20432 (* 5 (length graph-blank)) 5)
20433 "Number of units from one X axis label to next.")
20438 (Note that the value of @code{graph-blank} is set by another
20439 @code{defvar}. The @code{boundp} predicate checks whether it has
20440 already been set; @code{boundp} returns @code{nil} if it has not. If
20441 @code{graph-blank} were unbound and we did not use this conditional
20442 construction, in a recent GNU Emacs, we would enter the debugger and
20443 see an error message saying @samp{@w{Debugger entered--Lisp error:}
20444 @w{(void-variable graph-blank)}}.)
20447 Here is the @code{defvar} for @code{X-axis-tic-symbol}:
20451 (defvar X-axis-tic-symbol "|"
20452 "String to insert to point to a column in X axis.")
20457 The goal is to make a line that looks like this:
20463 The first tic is indented so that it is under the first column, which is
20464 indented to provide space for the Y axis labels.
20466 A tic element consists of the blank spaces that stretch from one tic to
20467 the next plus a tic symbol. The number of blanks is determined by the
20468 width of the tic symbol and the @code{X-axis-label-spacing}.
20471 The code looks like this:
20475 ;;; X-axis-tic-element
20479 ;; @r{Make a string of blanks.}
20480 (- (* symbol-width X-axis-label-spacing)
20481 (length X-axis-tic-symbol))
20483 ;; @r{Concatenate blanks with tic symbol.}
20489 Next, we determine how many blanks are needed to indent the first tic
20490 mark to the first column of the graph. This uses the value of
20491 @code{full-Y-label-width} passed it by the @code{print-graph} function.
20494 The code to make @code{X-axis-leading-spaces}
20499 ;; X-axis-leading-spaces
20501 (make-string full-Y-label-width ? )
20506 We also need to determine the length of the horizontal axis, which is
20507 the length of the numbers list, and the number of ticks in the horizontal
20514 (length numbers-list)
20520 (* symbol-width X-axis-label-spacing)
20524 ;; number-of-X-ticks
20525 (if (zerop (% (X-length tic-width)))
20526 (/ (X-length tic-width))
20527 (1+ (/ (X-length tic-width))))
20532 All this leads us directly to the function for printing the X axis tic line:
20534 @findex print-X-axis-tic-line
20537 (defun print-X-axis-tic-line
20538 (number-of-X-tics X-axis-leading-spaces X-axis-tic-element)
20539 "Print ticks for X axis."
20540 (insert X-axis-leading-spaces)
20541 (insert X-axis-tic-symbol) ; @r{Under first column.}
20544 ;; @r{Insert second tic in the right spot.}
20547 (- (* symbol-width X-axis-label-spacing)
20548 ;; @r{Insert white space up to second tic symbol.}
20549 (* 2 (length X-axis-tic-symbol)))
20551 X-axis-tic-symbol))
20554 ;; @r{Insert remaining ticks.}
20555 (while (> number-of-X-tics 1)
20556 (insert X-axis-tic-element)
20557 (setq number-of-X-tics (1- number-of-X-tics))))
20561 The line of numbers is equally straightforward:
20564 First, we create a numbered element with blank spaces before each number:
20566 @findex X-axis-element
20569 (defun X-axis-element (number)
20570 "Construct a numbered X axis element."
20571 (let ((leading-spaces
20572 (- (* symbol-width X-axis-label-spacing)
20573 (length (number-to-string number)))))
20574 (concat (make-string leading-spaces ? )
20575 (number-to-string number))))
20579 Next, we create the function to print the numbered line, starting with
20580 the number ``1'' under the first column:
20582 @findex print-X-axis-numbered-line
20585 (defun print-X-axis-numbered-line
20586 (number-of-X-tics X-axis-leading-spaces)
20587 "Print line of X-axis numbers"
20588 (let ((number X-axis-label-spacing))
20589 (insert X-axis-leading-spaces)
20595 ;; @r{Insert white space up to next number.}
20596 (- (* symbol-width X-axis-label-spacing) 2)
20598 (number-to-string number)))
20601 ;; @r{Insert remaining numbers.}
20602 (setq number (+ number X-axis-label-spacing))
20603 (while (> number-of-X-tics 1)
20604 (insert (X-axis-element number))
20605 (setq number (+ number X-axis-label-spacing))
20606 (setq number-of-X-tics (1- number-of-X-tics)))))
20610 Finally, we need to write the @code{print-X-axis} that uses
20611 @code{print-X-axis-tic-line} and
20612 @code{print-X-axis-numbered-line}.
20614 The function must determine the local values of the variables used by both
20615 @code{print-X-axis-tic-line} and @code{print-X-axis-numbered-line}, and
20616 then it must call them. Also, it must print the carriage return that
20617 separates the two lines.
20619 The function consists of a varlist that specifies five local variables,
20620 and calls to each of the two line printing functions:
20622 @findex print-X-axis
20625 (defun print-X-axis (numbers-list)
20626 "Print X axis labels to length of NUMBERS-LIST."
20627 (let* ((leading-spaces
20628 (make-string full-Y-label-width ? ))
20631 ;; symbol-width @r{is provided by} graph-body-print
20632 (tic-width (* symbol-width X-axis-label-spacing))
20633 (X-length (length numbers-list))
20641 ;; @r{Make a string of blanks.}
20642 (- (* symbol-width X-axis-label-spacing)
20643 (length X-axis-tic-symbol))
20647 ;; @r{Concatenate blanks with tic symbol.}
20648 X-axis-tic-symbol))
20652 (if (zerop (% X-length tic-width))
20653 (/ X-length tic-width)
20654 (1+ (/ X-length tic-width)))))
20657 (print-X-axis-tic-line tic-number leading-spaces X-tic)
20659 (print-X-axis-numbered-line tic-number leading-spaces)))
20664 You can test @code{print-X-axis}:
20668 Install @code{X-axis-tic-symbol}, @code{X-axis-label-spacing},
20669 @code{print-X-axis-tic-line}, as well as @code{X-axis-element},
20670 @code{print-X-axis-numbered-line}, and @code{print-X-axis}.
20673 Copy the following expression:
20678 (let ((full-Y-label-width 5)
20681 '(1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16))))
20686 Switch to the @file{*scratch*} buffer and place the cursor where you
20687 want the axis labels to start.
20690 Type @kbd{M-:} (@code{eval-expression}).
20693 Yank the test expression into the minibuffer
20694 with @kbd{C-y} (@code{yank)}.
20697 Press @key{RET} to evaluate the expression.
20701 Emacs will print the horizontal axis like this:
20711 @node Print Whole Graph, , print-X-axis, Full Graph
20712 @appendixsec Printing the Whole Graph
20713 @cindex Printing the whole graph
20714 @cindex Whole graph printing
20715 @cindex Graph, printing all
20717 Now we are nearly ready to print the whole graph.
20719 The function to print the graph with the proper labels follows the
20720 outline we created earlier (@pxref{Full Graph, , A Graph with Labelled
20721 Axes}), but with additions.
20724 Here is the outline:
20728 (defun print-graph (numbers-list)
20729 "@var{documentation}@dots{}"
20730 (let ((height @dots{}
20734 (print-Y-axis height @dots{} )
20735 (graph-body-print numbers-list)
20736 (print-X-axis @dots{} )))
20741 * The final version:: A few changes.
20742 * Test print-graph:: Run a short test.
20743 * Graphing words in defuns:: Executing the final code.
20744 * lambda:: How to write an anonymous function.
20745 * mapcar:: Apply a function to elements of a list.
20746 * Another Bug:: Yet another bug @dots{} most insidious.
20747 * Final printed graph:: The graph itself!
20750 @node The final version, Test print-graph, Print Whole Graph, Print Whole Graph
20752 @unnumberedsubsec Changes for the Final Version
20755 The final version is different from what we planned in two ways:
20756 first, it contains additional values calculated once in the varlist;
20757 second, it carries an option to specify the labels' increment per row.
20758 This latter feature turns out to be essential; otherwise, a graph may
20759 have more rows than fit on a display or on a sheet of paper.
20762 This new feature requires a change to the @code{Y-axis-column}
20763 function, to add @code{vertical-step} to it. The function looks like
20766 @findex Y-axis-column @r{Final version.}
20769 ;;; @r{Final version.}
20770 (defun Y-axis-column
20771 (height width-of-label &optional vertical-step)
20772 "Construct list of labels for Y axis.
20773 HEIGHT is maximum height of graph.
20774 WIDTH-OF-LABEL is maximum width of label.
20775 VERTICAL-STEP, an option, is a positive integer
20776 that specifies how much a Y axis label increments
20777 for each line. For example, a step of 5 means
20778 that each line is five units of the graph."
20782 (number-per-line (or vertical-step 1)))
20783 (while (> height 1)
20784 (if (zerop (% height Y-axis-label-spacing))
20787 ;; @r{Insert label.}
20791 (* height number-per-line)
20796 ;; @r{Else, insert blanks.}
20799 (make-string width-of-label ? )
20801 (setq height (1- height)))
20804 ;; @r{Insert base line.}
20805 (setq Y-axis (cons (Y-axis-element
20806 (or vertical-step 1)
20809 (nreverse Y-axis)))
20813 The values for the maximum height of graph and the width of a symbol
20814 are computed by @code{print-graph} in its @code{let} expression; so
20815 @code{graph-body-print} must be changed to accept them.
20817 @findex graph-body-print @r{Final version.}
20820 ;;; @r{Final version.}
20821 (defun graph-body-print (numbers-list height symbol-width)
20822 "Print a bar graph of the NUMBERS-LIST.
20823 The numbers-list consists of the Y-axis values.
20824 HEIGHT is maximum height of graph.
20825 SYMBOL-WIDTH is number of each column."
20828 (let (from-position)
20829 (while numbers-list
20830 (setq from-position (point))
20832 (column-of-graph height (car numbers-list)))
20833 (goto-char from-position)
20834 (forward-char symbol-width)
20837 ;; @r{Draw graph column by column.}
20839 (setq numbers-list (cdr numbers-list)))
20840 ;; @r{Place point for X axis labels.}
20841 (forward-line height)
20847 Finally, the code for the @code{print-graph} function:
20849 @findex print-graph @r{Final version.}
20852 ;;; @r{Final version.}
20854 (numbers-list &optional vertical-step)
20855 "Print labelled bar graph of the NUMBERS-LIST.
20856 The numbers-list consists of the Y-axis values.
20860 Optionally, VERTICAL-STEP, a positive integer,
20861 specifies how much a Y axis label increments for
20862 each line. For example, a step of 5 means that
20863 each row is five units."
20866 (let* ((symbol-width (length graph-blank))
20867 ;; @code{height} @r{is both the largest number}
20868 ;; @r{and the number with the most digits.}
20869 (height (apply 'max numbers-list))
20872 (height-of-top-line
20873 (if (zerop (% height Y-axis-label-spacing))
20876 (* (1+ (/ height Y-axis-label-spacing))
20877 Y-axis-label-spacing)))
20880 (vertical-step (or vertical-step 1))
20881 (full-Y-label-width
20887 (* height-of-top-line vertical-step))
20893 height-of-top-line full-Y-label-width vertical-step)
20897 numbers-list height-of-top-line symbol-width)
20898 (print-X-axis numbers-list)))
20902 @node Test print-graph, Graphing words in defuns, The final version, Print Whole Graph
20903 @appendixsubsec Testing @code{print-graph}
20906 We can test the @code{print-graph} function with a short list of numbers:
20910 Install the final versions of @code{Y-axis-column},
20911 @code{graph-body-print}, and @code{print-graph} (in addition to the
20915 Copy the following expression:
20918 (print-graph '(3 2 5 6 7 5 3 4 6 4 3 2 1))
20922 Switch to the @file{*scratch*} buffer and place the cursor where you
20923 want the axis labels to start.
20926 Type @kbd{M-:} (@code{eval-expression}).
20929 Yank the test expression into the minibuffer
20930 with @kbd{C-y} (@code{yank)}.
20933 Press @key{RET} to evaluate the expression.
20937 Emacs will print a graph that looks like this:
20958 On the other hand, if you pass @code{print-graph} a
20959 @code{vertical-step} value of 2, by evaluating this expression:
20962 (print-graph '(3 2 5 6 7 5 3 4 6 4 3 2 1) 2)
20967 The graph looks like this:
20988 (A question: is the `2' on the bottom of the vertical axis a bug or a
20989 feature? If you think it is a bug, and should be a `1' instead, (or
20990 even a `0'), you can modify the sources.)
20992 @node Graphing words in defuns, lambda, Test print-graph, Print Whole Graph
20993 @appendixsubsec Graphing Numbers of Words and Symbols
20995 Now for the graph for which all this code was written: a graph that
20996 shows how many function definitions contain fewer than 10 words and
20997 symbols, how many contain between 10 and 19 words and symbols, how
20998 many contain between 20 and 29 words and symbols, and so on.
21000 This is a multi-step process. First make sure you have loaded all the
21004 It is a good idea to reset the value of @code{top-of-ranges} in case
21005 you have set it to some different value. You can evaluate the
21010 (setq top-of-ranges
21013 110 120 130 140 150
21014 160 170 180 190 200
21015 210 220 230 240 250
21016 260 270 280 290 300)
21021 Next create a list of the number of words and symbols in each range.
21025 Evaluate the following:
21029 (setq list-for-graph
21032 (recursive-lengths-list-many-files
21033 (directory-files "/usr/local/emacs/lisp"
21041 On my old machine, this took about an hour. It looked though 303 Lisp
21042 files in my copy of Emacs version 19.23. After all that computing,
21043 the @code{list-for-graph} had this value:
21047 (537 1027 955 785 594 483 349 292 224 199 166 120 116 99
21048 90 80 67 48 52 45 41 33 28 26 25 20 12 28 11 13 220)
21053 This means that my copy of Emacs had 537 function definitions with
21054 fewer than 10 words or symbols in them, 1,027 function definitions
21055 with 10 to 19 words or symbols in them, 955 function definitions with
21056 20 to 29 words or symbols in them, and so on.
21058 Clearly, just by looking at this list we can see that most function
21059 definitions contain ten to thirty words and symbols.
21061 Now for printing. We do @emph{not} want to print a graph that is
21062 1,030 lines high @dots{} Instead, we should print a graph that is
21063 fewer than twenty-five lines high. A graph that height can be
21064 displayed on almost any monitor, and easily printed on a sheet of paper.
21066 This means that each value in @code{list-for-graph} must be reduced to
21067 one-fiftieth its present value.
21069 Here is a short function to do just that, using two functions we have
21070 not yet seen, @code{mapcar} and @code{lambda}.
21074 (defun one-fiftieth (full-range)
21075 "Return list, each number one-fiftieth of previous."
21076 (mapcar '(lambda (arg) (/ arg 50)) full-range))
21080 @node lambda, mapcar, Graphing words in defuns, Print Whole Graph
21081 @appendixsubsec A @code{lambda} Expression: Useful Anonymity
21082 @cindex Anonymous function
21085 @code{lambda} is the symbol for an anonymous function, a function
21086 without a name. Every time you use an anonymous function, you need to
21087 include its whole body.
21094 (lambda (arg) (/ arg 50))
21098 is a function definition that says `return the value resulting from
21099 dividing whatever is passed to me as @code{arg} by 50'.
21102 Earlier, for example, we had a function @code{multiply-by-seven}; it
21103 multiplied its argument by 7. This function is similar, except it
21104 divides its argument by 50; and, it has no name. The anonymous
21105 equivalent of @code{multiply-by-seven} is:
21108 (lambda (number) (* 7 number))
21112 (@xref{defun, , The @code{defun} Special Form}.)
21116 If we want to multiply 3 by 7, we can write:
21118 @c !!! Clear print-postscript-figures if the computer formatting this
21119 @c document is too small and cannot handle all the diagrams and figures.
21120 @c clear print-postscript-figures
21121 @c set print-postscript-figures
21122 @c lambda example diagram #1
21126 (multiply-by-seven 3)
21127 \_______________/ ^
21133 @ifset print-postscript-figures
21136 @center @image{lambda-1}
21137 %%%% old method of including an image
21138 % \input /usr/local/lib/tex/inputs/psfig.tex
21139 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/lambda-1.eps}}
21144 @ifclear print-postscript-figures
21148 (multiply-by-seven 3)
21149 \_______________/ ^
21158 This expression returns 21.
21162 Similarly, we can write:
21164 @c lambda example diagram #2
21168 ((lambda (number) (* 7 number)) 3)
21169 \____________________________/ ^
21171 anonymous function argument
21175 @ifset print-postscript-figures
21178 @center @image{lambda-2}
21179 %%%% old method of including an image
21180 % \input /usr/local/lib/tex/inputs/psfig.tex
21181 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/lambda-2.eps}}
21186 @ifclear print-postscript-figures
21190 ((lambda (number) (* 7 number)) 3)
21191 \____________________________/ ^
21193 anonymous function argument
21201 If we want to divide 100 by 50, we can write:
21203 @c lambda example diagram #3
21207 ((lambda (arg) (/ arg 50)) 100)
21208 \______________________/ \_/
21210 anonymous function argument
21214 @ifset print-postscript-figures
21217 @center @image{lambda-3}
21218 %%%% old method of including an image
21219 % \input /usr/local/lib/tex/inputs/psfig.tex
21220 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/lambda-3.eps}}
21225 @ifclear print-postscript-figures
21229 ((lambda (arg) (/ arg 50)) 100)
21230 \______________________/ \_/
21232 anonymous function argument
21239 This expression returns 2. The 100 is passed to the function, which
21240 divides that number by 50.
21242 @xref{Lambda Expressions, , Lambda Expressions, elisp, The GNU Emacs
21243 Lisp Reference Manual}, for more about @code{lambda}. Lisp and lambda
21244 expressions derive from the Lambda Calculus.
21246 @node mapcar, Another Bug, lambda, Print Whole Graph
21247 @appendixsubsec The @code{mapcar} Function
21250 @code{mapcar} is a function that calls its first argument with each
21251 element of its second argument, in turn. The second argument must be
21254 The @samp{map} part of the name comes from the mathematical phrase,
21255 `mapping over a domain', meaning to apply a function to each of the
21256 elements in a domain. The mathematical phrase is based on the
21257 metaphor of a surveyor walking, one step at a time, over an area he is
21258 mapping. And @samp{car}, of course, comes from the Lisp notion of the
21267 (mapcar '1+ '(2 4 6))
21273 The function @code{1+} which adds one to its argument, is executed on
21274 @emph{each} element of the list, and a new list is returned.
21276 Contrast this with @code{apply}, which applies its first argument to
21278 (@xref{Readying a Graph, , Readying a Graph}, for a explanation of
21282 In the definition of @code{one-fiftieth}, the first argument is the
21283 anonymous function:
21286 (lambda (arg) (/ arg 50))
21290 and the second argument is @code{full-range}, which will be bound to
21291 @code{list-for-graph}.
21294 The whole expression looks like this:
21297 (mapcar '(lambda (arg) (/ arg 50)) full-range))
21300 @xref{Mapping Functions, , Mapping Functions, elisp, The GNU Emacs
21301 Lisp Reference Manual}, for more about @code{mapcar}.
21303 Using the @code{one-fiftieth} function, we can generate a list in
21304 which each element is one-fiftieth the size of the corresponding
21305 element in @code{list-for-graph}.
21309 (setq fiftieth-list-for-graph
21310 (one-fiftieth list-for-graph))
21315 The resulting list looks like this:
21319 (10 20 19 15 11 9 6 5 4 3 3 2 2
21320 1 1 1 1 0 1 0 0 0 0 0 0 0 0 0 0 0 4)
21325 This, we are almost ready to print! (We also notice the loss of
21326 information: many of the higher ranges are 0, meaning that fewer than
21327 50 defuns had that many words or symbols---but not necessarily meaning
21328 that none had that many words or symbols.)
21330 @node Another Bug, Final printed graph, mapcar, Print Whole Graph
21331 @appendixsubsec Another Bug @dots{} Most Insidious
21332 @cindex Bug, most insidious type
21333 @cindex Insidious type of bug
21335 I said `almost ready to print'! Of course, there is a bug in the
21336 @code{print-graph} function @dots{} It has a @code{vertical-step}
21337 option, but not a @code{horizontal-step} option. The
21338 @code{top-of-range} scale goes from 10 to 300 by tens. But the
21339 @code{print-graph} function will print only by ones.
21341 This is a classic example of what some consider the most insidious
21342 type of bug, the bug of omission. This is not the kind of bug you can
21343 find by studying the code, for it is not in the code; it is an omitted
21344 feature. Your best actions are to try your program early and often;
21345 and try to arrange, as much as you can, to write code that is easy to
21346 understand and easy to change. Try to be aware, whenever you can,
21347 that whatever you have written, @emph{will} be rewritten, if not soon,
21348 eventually. A hard maxim to follow.
21350 It is the @code{print-X-axis-numbered-line} function that needs the
21351 work; and then the @code{print-X-axis} and the @code{print-graph}
21352 functions need to be adapted. Not much needs to be done; there is one
21353 nicety: the numbers ought to line up under the tic marks. This takes
21357 Here is the corrected @code{print-X-axis-numbered-line}:
21361 (defun print-X-axis-numbered-line
21362 (number-of-X-tics X-axis-leading-spaces
21363 &optional horizontal-step)
21364 "Print line of X-axis numbers"
21365 (let ((number X-axis-label-spacing)
21366 (horizontal-step (or horizontal-step 1)))
21369 (insert X-axis-leading-spaces)
21370 ;; @r{Delete extra leading spaces.}
21373 (length (number-to-string horizontal-step)))))
21378 ;; @r{Insert white space.}
21380 X-axis-label-spacing)
21383 (number-to-string horizontal-step)))
21387 (* number horizontal-step))))
21390 ;; @r{Insert remaining numbers.}
21391 (setq number (+ number X-axis-label-spacing))
21392 (while (> number-of-X-tics 1)
21393 (insert (X-axis-element
21394 (* number horizontal-step)))
21395 (setq number (+ number X-axis-label-spacing))
21396 (setq number-of-X-tics (1- number-of-X-tics)))))
21401 If you are reading this in Info, you can see the new versions of
21402 @code{print-X-axis} @code{print-graph} and evaluate them. If you are
21403 reading this in a printed book, you can see the changed lines here
21404 (the full text is too much to print).
21409 (defun print-X-axis (numbers-list horizontal-step)
21411 (print-X-axis-numbered-line
21412 tic-number leading-spaces horizontal-step))
21420 &optional vertical-step horizontal-step)
21422 (print-X-axis numbers-list horizontal-step))
21430 (defun print-X-axis (numbers-list horizontal-step)
21431 "Print X axis labels to length of NUMBERS-LIST.
21432 Optionally, HORIZONTAL-STEP, a positive integer,
21433 specifies how much an X axis label increments for
21437 ;; Value of symbol-width and full-Y-label-width
21438 ;; are passed by `print-graph'.
21439 (let* ((leading-spaces
21440 (make-string full-Y-label-width ? ))
21441 ;; symbol-width @r{is provided by} graph-body-print
21442 (tic-width (* symbol-width X-axis-label-spacing))
21443 (X-length (length numbers-list))
21449 ;; @r{Make a string of blanks.}
21450 (- (* symbol-width X-axis-label-spacing)
21451 (length X-axis-tic-symbol))
21455 ;; @r{Concatenate blanks with tic symbol.}
21456 X-axis-tic-symbol))
21458 (if (zerop (% X-length tic-width))
21459 (/ X-length tic-width)
21460 (1+ (/ X-length tic-width)))))
21464 (print-X-axis-tic-line
21465 tic-number leading-spaces X-tic)
21467 (print-X-axis-numbered-line
21468 tic-number leading-spaces horizontal-step)))
21475 (numbers-list &optional vertical-step horizontal-step)
21476 "Print labelled bar graph of the NUMBERS-LIST.
21477 The numbers-list consists of the Y-axis values.
21481 Optionally, VERTICAL-STEP, a positive integer,
21482 specifies how much a Y axis label increments for
21483 each line. For example, a step of 5 means that
21484 each row is five units.
21488 Optionally, HORIZONTAL-STEP, a positive integer,
21489 specifies how much an X axis label increments for
21491 (let* ((symbol-width (length graph-blank))
21492 ;; @code{height} @r{is both the largest number}
21493 ;; @r{and the number with the most digits.}
21494 (height (apply 'max numbers-list))
21497 (height-of-top-line
21498 (if (zerop (% height Y-axis-label-spacing))
21501 (* (1+ (/ height Y-axis-label-spacing))
21502 Y-axis-label-spacing)))
21505 (vertical-step (or vertical-step 1))
21506 (full-Y-label-width
21510 (* height-of-top-line vertical-step))
21515 height-of-top-line full-Y-label-width vertical-step)
21517 numbers-list height-of-top-line symbol-width)
21518 (print-X-axis numbers-list horizontal-step)))
21525 Graphing Definitions Re-listed
21528 Here are all the graphing definitions in their final form:
21532 (defvar top-of-ranges
21535 110 120 130 140 150
21536 160 170 180 190 200
21537 210 220 230 240 250)
21538 "List specifying ranges for `defuns-per-range'.")
21542 (defvar graph-symbol "*"
21543 "String used as symbol in graph, usually an asterisk.")
21547 (defvar graph-blank " "
21548 "String used as blank in graph, usually a blank space.
21549 graph-blank must be the same number of columns wide
21554 (defvar Y-axis-tic " - "
21555 "String that follows number in a Y axis label.")
21559 (defvar Y-axis-label-spacing 5
21560 "Number of lines from one Y axis label to next.")
21564 (defvar X-axis-tic-symbol "|"
21565 "String to insert to point to a column in X axis.")
21569 (defvar X-axis-label-spacing
21570 (if (boundp 'graph-blank)
21571 (* 5 (length graph-blank)) 5)
21572 "Number of units from one X axis label to next.")
21578 (defun count-words-in-defun ()
21579 "Return the number of words and symbols in a defun."
21580 (beginning-of-defun)
21582 (end (save-excursion (end-of-defun) (point))))
21587 (and (< (point) end)
21589 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*"
21591 (setq count (1+ count)))
21598 (defun lengths-list-file (filename)
21599 "Return list of definitions' lengths within FILE.
21600 The returned list is a list of numbers.
21601 Each number is the number of words or
21602 symbols in one function definition."
21606 (message "Working on `%s' ... " filename)
21608 (let ((buffer (find-file-noselect filename))
21610 (set-buffer buffer)
21611 (setq buffer-read-only t)
21613 (goto-char (point-min))
21617 (while (re-search-forward "^(defun" nil t)
21619 (cons (count-words-in-defun) lengths-list)))
21620 (kill-buffer buffer)
21627 (defun lengths-list-many-files (list-of-files)
21628 "Return list of lengths of defuns in LIST-OF-FILES."
21629 (let (lengths-list)
21630 ;;; @r{true-or-false-test}
21631 (while list-of-files
21637 ;;; @r{Generate a lengths' list.}
21639 (expand-file-name (car list-of-files)))))
21640 ;;; @r{Make files' list shorter.}
21641 (setq list-of-files (cdr list-of-files)))
21642 ;;; @r{Return final value of lengths' list.}
21649 (defun defuns-per-range (sorted-lengths top-of-ranges)
21650 "SORTED-LENGTHS defuns in each TOP-OF-RANGES range."
21651 (let ((top-of-range (car top-of-ranges))
21652 (number-within-range 0)
21653 defuns-per-range-list)
21658 (while top-of-ranges
21662 ;; @r{Need number for numeric test.}
21663 (car sorted-lengths)
21664 (< (car sorted-lengths) top-of-range))
21666 ;; @r{Count number of definitions within current range.}
21667 (setq number-within-range (1+ number-within-range))
21668 (setq sorted-lengths (cdr sorted-lengths)))
21672 ;; @r{Exit inner loop but remain within outer loop.}
21674 (setq defuns-per-range-list
21675 (cons number-within-range defuns-per-range-list))
21676 (setq number-within-range 0) ; @r{Reset count to zero.}
21678 ;; @r{Move to next range.}
21679 (setq top-of-ranges (cdr top-of-ranges))
21680 ;; @r{Specify next top of range value.}
21681 (setq top-of-range (car top-of-ranges)))
21685 ;; @r{Exit outer loop and count the number of defuns larger than}
21686 ;; @r{ the largest top-of-range value.}
21687 (setq defuns-per-range-list
21689 (length sorted-lengths)
21690 defuns-per-range-list))
21692 ;; @r{Return a list of the number of definitions within each range,}
21693 ;; @r{ smallest to largest.}
21694 (nreverse defuns-per-range-list)))
21700 (defun column-of-graph (max-graph-height actual-height)
21701 "Return list of MAX-GRAPH-HEIGHT strings;
21702 ACTUAL-HEIGHT are graph-symbols.
21703 The graph-symbols are contiguous entries at the end
21705 The list will be inserted as one column of a graph.
21706 The strings are either graph-blank or graph-symbol."
21710 (let ((insert-list nil)
21711 (number-of-top-blanks
21712 (- max-graph-height actual-height)))
21714 ;; @r{Fill in @code{graph-symbols}.}
21715 (while (> actual-height 0)
21716 (setq insert-list (cons graph-symbol insert-list))
21717 (setq actual-height (1- actual-height)))
21721 ;; @r{Fill in @code{graph-blanks}.}
21722 (while (> number-of-top-blanks 0)
21723 (setq insert-list (cons graph-blank insert-list))
21724 (setq number-of-top-blanks
21725 (1- number-of-top-blanks)))
21727 ;; @r{Return whole list.}
21734 (defun Y-axis-element (number full-Y-label-width)
21735 "Construct a NUMBERed label element.
21736 A numbered element looks like this ` 5 - ',
21737 and is padded as needed so all line up with
21738 the element for the largest number."
21741 (let* ((leading-spaces
21742 (- full-Y-label-width
21744 (concat (number-to-string number)
21749 (make-string leading-spaces ? )
21750 (number-to-string number)
21757 (defun print-Y-axis
21758 (height full-Y-label-width &optional vertical-step)
21759 "Insert Y axis by HEIGHT and FULL-Y-LABEL-WIDTH.
21760 Height must be the maximum height of the graph.
21761 Full width is the width of the highest label element.
21762 Optionally, print according to VERTICAL-STEP."
21765 ;; Value of height and full-Y-label-width
21766 ;; are passed by `print-graph'.
21767 (let ((start (point)))
21769 (Y-axis-column height full-Y-label-width vertical-step))
21772 ;; @r{Place point ready for inserting graph.}
21774 ;; @r{Move point forward by value of} full-Y-label-width
21775 (forward-char full-Y-label-width)))
21781 (defun print-X-axis-tic-line
21782 (number-of-X-tics X-axis-leading-spaces X-axis-tic-element)
21783 "Print ticks for X axis."
21784 (insert X-axis-leading-spaces)
21785 (insert X-axis-tic-symbol) ; @r{Under first column.}
21788 ;; @r{Insert second tic in the right spot.}
21791 (- (* symbol-width X-axis-label-spacing)
21792 ;; @r{Insert white space up to second tic symbol.}
21793 (* 2 (length X-axis-tic-symbol)))
21795 X-axis-tic-symbol))
21798 ;; @r{Insert remaining ticks.}
21799 (while (> number-of-X-tics 1)
21800 (insert X-axis-tic-element)
21801 (setq number-of-X-tics (1- number-of-X-tics))))
21807 (defun X-axis-element (number)
21808 "Construct a numbered X axis element."
21809 (let ((leading-spaces
21810 (- (* symbol-width X-axis-label-spacing)
21811 (length (number-to-string number)))))
21812 (concat (make-string leading-spaces ? )
21813 (number-to-string number))))
21819 (defun graph-body-print (numbers-list height symbol-width)
21820 "Print a bar graph of the NUMBERS-LIST.
21821 The numbers-list consists of the Y-axis values.
21822 HEIGHT is maximum height of graph.
21823 SYMBOL-WIDTH is number of each column."
21826 (let (from-position)
21827 (while numbers-list
21828 (setq from-position (point))
21830 (column-of-graph height (car numbers-list)))
21831 (goto-char from-position)
21832 (forward-char symbol-width)
21835 ;; @r{Draw graph column by column.}
21837 (setq numbers-list (cdr numbers-list)))
21838 ;; @r{Place point for X axis labels.}
21839 (forward-line height)
21846 (defun Y-axis-column
21847 (height width-of-label &optional vertical-step)
21848 "Construct list of labels for Y axis.
21849 HEIGHT is maximum height of graph.
21850 WIDTH-OF-LABEL is maximum width of label.
21853 VERTICAL-STEP, an option, is a positive integer
21854 that specifies how much a Y axis label increments
21855 for each line. For example, a step of 5 means
21856 that each line is five units of the graph."
21858 (number-per-line (or vertical-step 1)))
21861 (while (> height 1)
21862 (if (zerop (% height Y-axis-label-spacing))
21863 ;; @r{Insert label.}
21867 (* height number-per-line)
21872 ;; @r{Else, insert blanks.}
21875 (make-string width-of-label ? )
21877 (setq height (1- height)))
21880 ;; @r{Insert base line.}
21881 (setq Y-axis (cons (Y-axis-element
21882 (or vertical-step 1)
21885 (nreverse Y-axis)))
21891 (defun print-X-axis-numbered-line
21892 (number-of-X-tics X-axis-leading-spaces
21893 &optional horizontal-step)
21894 "Print line of X-axis numbers"
21895 (let ((number X-axis-label-spacing)
21896 (horizontal-step (or horizontal-step 1)))
21899 (insert X-axis-leading-spaces)
21901 (delete-char (- (1- (length (number-to-string horizontal-step)))))
21904 ;; @r{Insert white space up to next number.}
21905 (- (* symbol-width X-axis-label-spacing)
21906 (1- (length (number-to-string horizontal-step)))
21909 (number-to-string (* number horizontal-step))))
21912 ;; @r{Insert remaining numbers.}
21913 (setq number (+ number X-axis-label-spacing))
21914 (while (> number-of-X-tics 1)
21915 (insert (X-axis-element (* number horizontal-step)))
21916 (setq number (+ number X-axis-label-spacing))
21917 (setq number-of-X-tics (1- number-of-X-tics)))))
21923 (defun print-X-axis (numbers-list horizontal-step)
21924 "Print X axis labels to length of NUMBERS-LIST.
21925 Optionally, HORIZONTAL-STEP, a positive integer,
21926 specifies how much an X axis label increments for
21930 ;; Value of symbol-width and full-Y-label-width
21931 ;; are passed by `print-graph'.
21932 (let* ((leading-spaces
21933 (make-string full-Y-label-width ? ))
21934 ;; symbol-width @r{is provided by} graph-body-print
21935 (tic-width (* symbol-width X-axis-label-spacing))
21936 (X-length (length numbers-list))
21942 ;; @r{Make a string of blanks.}
21943 (- (* symbol-width X-axis-label-spacing)
21944 (length X-axis-tic-symbol))
21948 ;; @r{Concatenate blanks with tic symbol.}
21949 X-axis-tic-symbol))
21951 (if (zerop (% X-length tic-width))
21952 (/ X-length tic-width)
21953 (1+ (/ X-length tic-width)))))
21957 (print-X-axis-tic-line
21958 tic-number leading-spaces X-tic)
21960 (print-X-axis-numbered-line
21961 tic-number leading-spaces horizontal-step)))
21967 (defun one-fiftieth (full-range)
21968 "Return list, each number of which is 1/50th previous."
21969 (mapcar '(lambda (arg) (/ arg 50)) full-range))
21976 (numbers-list &optional vertical-step horizontal-step)
21977 "Print labelled bar graph of the NUMBERS-LIST.
21978 The numbers-list consists of the Y-axis values.
21982 Optionally, VERTICAL-STEP, a positive integer,
21983 specifies how much a Y axis label increments for
21984 each line. For example, a step of 5 means that
21985 each row is five units.
21989 Optionally, HORIZONTAL-STEP, a positive integer,
21990 specifies how much an X axis label increments for
21992 (let* ((symbol-width (length graph-blank))
21993 ;; @code{height} @r{is both the largest number}
21994 ;; @r{and the number with the most digits.}
21995 (height (apply 'max numbers-list))
21998 (height-of-top-line
21999 (if (zerop (% height Y-axis-label-spacing))
22002 (* (1+ (/ height Y-axis-label-spacing))
22003 Y-axis-label-spacing)))
22006 (vertical-step (or vertical-step 1))
22007 (full-Y-label-width
22011 (* height-of-top-line vertical-step))
22017 height-of-top-line full-Y-label-width vertical-step)
22019 numbers-list height-of-top-line symbol-width)
22020 (print-X-axis numbers-list horizontal-step)))
22027 @node Final printed graph, , Another Bug, Print Whole Graph
22028 @appendixsubsec The Printed Graph
22030 When made and installed, you can call the @code{print-graph} command
22036 (print-graph fiftieth-list-for-graph 50 10)
22066 50 - ***************** * *
22068 10 50 100 150 200 250 300 350
22075 The largest group of functions contain 10 -- 19 words and symbols each.
22077 @node Free Software and Free Manuals, GNU Free Documentation License, Full Graph, Top
22078 @appendix Free Software and Free Manuals
22080 @strong{by Richard M. Stallman}
22083 The biggest deficiency in free operating systems is not in the
22084 software---it is the lack of good free manuals that we can include in
22085 these systems. Many of our most important programs do not come with
22086 full manuals. Documentation is an essential part of any software
22087 package; when an important free software package does not come with a
22088 free manual, that is a major gap. We have many such gaps today.
22090 Once upon a time, many years ago, I thought I would learn Perl. I got
22091 a copy of a free manual, but I found it hard to read. When I asked
22092 Perl users about alternatives, they told me that there were better
22093 introductory manuals---but those were not free.
22095 Why was this? The authors of the good manuals had written them for
22096 O'Reilly Associates, which published them with restrictive terms---no
22097 copying, no modification, source files not available---which exclude
22098 them from the free software community.
22100 That wasn't the first time this sort of thing has happened, and (to
22101 our community's great loss) it was far from the last. Proprietary
22102 manual publishers have enticed a great many authors to restrict their
22103 manuals since then. Many times I have heard a GNU user eagerly tell me
22104 about a manual that he is writing, with which he expects to help the
22105 GNU project---and then had my hopes dashed, as he proceeded to explain
22106 that he had signed a contract with a publisher that would restrict it
22107 so that we cannot use it.
22109 Given that writing good English is a rare skill among programmers, we
22110 can ill afford to lose manuals this way.
22113 (The Free Software Foundation
22114 @uref{http://www.gnu.org/doc/doc.html#DescriptionsOfGNUDocumentation, ,
22115 sells printed copies} of free @uref{http://www.gnu.org/doc/doc.html,
22116 GNU manuals}, too.)
22118 Free documentation, like free software, is a matter of freedom, not
22119 price. The problem with these manuals was not that O'Reilly Associates
22120 charged a price for printed copies---that in itself is fine. (The Free
22121 Software Foundation sells printed copies of free GNU manuals, too.)
22122 But GNU manuals are available in source code form, while these manuals
22123 are available only on paper. GNU manuals come with permission to copy
22124 and modify; the Perl manuals do not. These restrictions are the
22127 The criterion for a free manual is pretty much the same as for free
22128 software: it is a matter of giving all users certain
22129 freedoms. Redistribution (including commercial redistribution) must be
22130 permitted, so that the manual can accompany every copy of the program,
22131 on-line or on paper. Permission for modification is crucial too.
22133 As a general rule, I don't believe that it is essential for people to
22134 have permission to modify all sorts of articles and books. The issues
22135 for writings are not necessarily the same as those for software. For
22136 example, I don't think you or I are obliged to give permission to
22137 modify articles like this one, which describe our actions and our
22140 But there is a particular reason why the freedom to modify is crucial
22141 for documentation for free software. When people exercise their right
22142 to modify the software, and add or change its features, if they are
22143 conscientious they will change the manual too---so they can provide
22144 accurate and usable documentation with the modified program. A manual
22145 which forbids programmers to be conscientious and finish the job, or
22146 more precisely requires them to write a new manual from scratch if
22147 they change the program, does not fill our community's needs.
22149 While a blanket prohibition on modification is unacceptable, some
22150 kinds of limits on the method of modification pose no problem. For
22151 example, requirements to preserve the original author's copyright
22152 notice, the distribution terms, or the list of authors, are ok. It is
22153 also no problem to require modified versions to include notice that
22154 they were modified, even to have entire sections that may not be
22155 deleted or changed, as long as these sections deal with nontechnical
22156 topics. (Some GNU manuals have them.)
22158 These kinds of restrictions are not a problem because, as a practical
22159 matter, they don't stop the conscientious programmer from adapting the
22160 manual to fit the modified program. In other words, they don't block
22161 the free software community from making full use of the manual.
22163 However, it must be possible to modify all the technical content of
22164 the manual, and then distribute the result in all the usual media,
22165 through all the usual channels; otherwise, the restrictions do block
22166 the community, the manual is not free, and so we need another manual.
22168 Unfortunately, it is often hard to find someone to write another
22169 manual when a proprietary manual exists. The obstacle is that many
22170 users think that a proprietary manual is good enough---so they don't
22171 see the need to write a free manual. They do not see that the free
22172 operating system has a gap that needs filling.
22174 Why do users think that proprietary manuals are good enough? Some have
22175 not considered the issue. I hope this article will do something to
22178 Other users consider proprietary manuals acceptable for the same
22179 reason so many people consider proprietary software acceptable: they
22180 judge in purely practical terms, not using freedom as a
22181 criterion. These people are entitled to their opinions, but since
22182 those opinions spring from values which do not include freedom, they
22183 are no guide for those of us who do value freedom.
22185 Please spread the word about this issue. We continue to lose manuals
22186 to proprietary publishing. If we spread the word that proprietary
22187 manuals are not sufficient, perhaps the next person who wants to help
22188 GNU by writing documentation will realize, before it is too late, that
22189 he must above all make it free.
22191 We can also encourage commercial publishers to sell free, copylefted
22192 manuals instead of proprietary ones. One way you can help this is to
22193 check the distribution terms of a manual before you buy it, and prefer
22194 copylefted manuals to non-copylefted ones.
22198 Note: The Free Software Foundation maintains a page on its Web site
22199 that lists free books available from other publishers:@*
22200 @uref{http://www.gnu.org/doc/other-free-books.html}
22202 @node GNU Free Documentation License, Index, Free Software and Free Manuals, Top
22203 @appendix GNU Free Documentation License
22205 @cindex FDL, GNU Free Documentation License
22206 @center Version 1.2, November 2002
22209 Copyright @copyright{} 2000,2001,2002 Free Software Foundation, Inc.
22210 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA
22212 Everyone is permitted to copy and distribute verbatim copies
22213 of this license document, but changing it is not allowed.
22220 The purpose of this License is to make a manual, textbook, or other
22221 functional and useful document @dfn{free} in the sense of freedom: to
22222 assure everyone the effective freedom to copy and redistribute it,
22223 with or without modifying it, either commercially or noncommercially.
22224 Secondarily, this License preserves for the author and publisher a way
22225 to get credit for their work, while not being considered responsible
22226 for modifications made by others.
22228 This License is a kind of ``copyleft'', which means that derivative
22229 works of the document must themselves be free in the same sense. It
22230 complements the GNU General Public License, which is a copyleft
22231 license designed for free software.
22233 We have designed this License in order to use it for manuals for free
22234 software, because free software needs free documentation: a free
22235 program should come with manuals providing the same freedoms that the
22236 software does. But this License is not limited to software manuals;
22237 it can be used for any textual work, regardless of subject matter or
22238 whether it is published as a printed book. We recommend this License
22239 principally for works whose purpose is instruction or reference.
22242 APPLICABILITY AND DEFINITIONS
22244 This License applies to any manual or other work, in any medium, that
22245 contains a notice placed by the copyright holder saying it can be
22246 distributed under the terms of this License. Such a notice grants a
22247 world-wide, royalty-free license, unlimited in duration, to use that
22248 work under the conditions stated herein. The ``Document'', below,
22249 refers to any such manual or work. Any member of the public is a
22250 licensee, and is addressed as ``you''. You accept the license if you
22251 copy, modify or distribute the work in a way requiring permission
22252 under copyright law.
22254 A ``Modified Version'' of the Document means any work containing the
22255 Document or a portion of it, either copied verbatim, or with
22256 modifications and/or translated into another language.
22258 A ``Secondary Section'' is a named appendix or a front-matter section
22259 of the Document that deals exclusively with the relationship of the
22260 publishers or authors of the Document to the Document's overall
22261 subject (or to related matters) and contains nothing that could fall
22262 directly within that overall subject. (Thus, if the Document is in
22263 part a textbook of mathematics, a Secondary Section may not explain
22264 any mathematics.) The relationship could be a matter of historical
22265 connection with the subject or with related matters, or of legal,
22266 commercial, philosophical, ethical or political position regarding
22269 The ``Invariant Sections'' are certain Secondary Sections whose titles
22270 are designated, as being those of Invariant Sections, in the notice
22271 that says that the Document is released under this License. If a
22272 section does not fit the above definition of Secondary then it is not
22273 allowed to be designated as Invariant. The Document may contain zero
22274 Invariant Sections. If the Document does not identify any Invariant
22275 Sections then there are none.
22277 The ``Cover Texts'' are certain short passages of text that are listed,
22278 as Front-Cover Texts or Back-Cover Texts, in the notice that says that
22279 the Document is released under this License. A Front-Cover Text may
22280 be at most 5 words, and a Back-Cover Text may be at most 25 words.
22282 A ``Transparent'' copy of the Document means a machine-readable copy,
22283 represented in a format whose specification is available to the
22284 general public, that is suitable for revising the document
22285 straightforwardly with generic text editors or (for images composed of
22286 pixels) generic paint programs or (for drawings) some widely available
22287 drawing editor, and that is suitable for input to text formatters or
22288 for automatic translation to a variety of formats suitable for input
22289 to text formatters. A copy made in an otherwise Transparent file
22290 format whose markup, or absence of markup, has been arranged to thwart
22291 or discourage subsequent modification by readers is not Transparent.
22292 An image format is not Transparent if used for any substantial amount
22293 of text. A copy that is not ``Transparent'' is called ``Opaque''.
22295 Examples of suitable formats for Transparent copies include plain
22296 @sc{ascii} without markup, Texinfo input format, La@TeX{} input
22297 format, @acronym{SGML} or @acronym{XML} using a publicly available
22298 @acronym{DTD}, and standard-conforming simple @acronym{HTML},
22299 PostScript or @acronym{PDF} designed for human modification. Examples
22300 of transparent image formats include @acronym{PNG}, @acronym{XCF} and
22301 @acronym{JPG}. Opaque formats include proprietary formats that can be
22302 read and edited only by proprietary word processors, @acronym{SGML} or
22303 @acronym{XML} for which the @acronym{DTD} and/or processing tools are
22304 not generally available, and the machine-generated @acronym{HTML},
22305 PostScript or @acronym{PDF} produced by some word processors for
22306 output purposes only.
22308 The ``Title Page'' means, for a printed book, the title page itself,
22309 plus such following pages as are needed to hold, legibly, the material
22310 this License requires to appear in the title page. For works in
22311 formats which do not have any title page as such, ``Title Page'' means
22312 the text near the most prominent appearance of the work's title,
22313 preceding the beginning of the body of the text.
22315 A section ``Entitled XYZ'' means a named subunit of the Document whose
22316 title either is precisely XYZ or contains XYZ in parentheses following
22317 text that translates XYZ in another language. (Here XYZ stands for a
22318 specific section name mentioned below, such as ``Acknowledgements'',
22319 ``Dedications'', ``Endorsements'', or ``History''.) To ``Preserve the Title''
22320 of such a section when you modify the Document means that it remains a
22321 section ``Entitled XYZ'' according to this definition.
22323 The Document may include Warranty Disclaimers next to the notice which
22324 states that this License applies to the Document. These Warranty
22325 Disclaimers are considered to be included by reference in this
22326 License, but only as regards disclaiming warranties: any other
22327 implication that these Warranty Disclaimers may have is void and has
22328 no effect on the meaning of this License.
22333 You may copy and distribute the Document in any medium, either
22334 commercially or noncommercially, provided that this License, the
22335 copyright notices, and the license notice saying this License applies
22336 to the Document are reproduced in all copies, and that you add no other
22337 conditions whatsoever to those of this License. You may not use
22338 technical measures to obstruct or control the reading or further
22339 copying of the copies you make or distribute. However, you may accept
22340 compensation in exchange for copies. If you distribute a large enough
22341 number of copies you must also follow the conditions in section 3.
22343 You may also lend copies, under the same conditions stated above, and
22344 you may publicly display copies.
22347 COPYING IN QUANTITY
22349 If you publish printed copies (or copies in media that commonly have
22350 printed covers) of the Document, numbering more than 100, and the
22351 Document's license notice requires Cover Texts, you must enclose the
22352 copies in covers that carry, clearly and legibly, all these Cover
22353 Texts: Front-Cover Texts on the front cover, and Back-Cover Texts on
22354 the back cover. Both covers must also clearly and legibly identify
22355 you as the publisher of these copies. The front cover must present
22356 the full title with all words of the title equally prominent and
22357 visible. You may add other material on the covers in addition.
22358 Copying with changes limited to the covers, as long as they preserve
22359 the title of the Document and satisfy these conditions, can be treated
22360 as verbatim copying in other respects.
22362 If the required texts for either cover are too voluminous to fit
22363 legibly, you should put the first ones listed (as many as fit
22364 reasonably) on the actual cover, and continue the rest onto adjacent
22367 If you publish or distribute Opaque copies of the Document numbering
22368 more than 100, you must either include a machine-readable Transparent
22369 copy along with each Opaque copy, or state in or with each Opaque copy
22370 a computer-network location from which the general network-using
22371 public has access to download using public-standard network protocols
22372 a complete Transparent copy of the Document, free of added material.
22373 If you use the latter option, you must take reasonably prudent steps,
22374 when you begin distribution of Opaque copies in quantity, to ensure
22375 that this Transparent copy will remain thus accessible at the stated
22376 location until at least one year after the last time you distribute an
22377 Opaque copy (directly or through your agents or retailers) of that
22378 edition to the public.
22380 It is requested, but not required, that you contact the authors of the
22381 Document well before redistributing any large number of copies, to give
22382 them a chance to provide you with an updated version of the Document.
22387 You may copy and distribute a Modified Version of the Document under
22388 the conditions of sections 2 and 3 above, provided that you release
22389 the Modified Version under precisely this License, with the Modified
22390 Version filling the role of the Document, thus licensing distribution
22391 and modification of the Modified Version to whoever possesses a copy
22392 of it. In addition, you must do these things in the Modified Version:
22396 Use in the Title Page (and on the covers, if any) a title distinct
22397 from that of the Document, and from those of previous versions
22398 (which should, if there were any, be listed in the History section
22399 of the Document). You may use the same title as a previous version
22400 if the original publisher of that version gives permission.
22403 List on the Title Page, as authors, one or more persons or entities
22404 responsible for authorship of the modifications in the Modified
22405 Version, together with at least five of the principal authors of the
22406 Document (all of its principal authors, if it has fewer than five),
22407 unless they release you from this requirement.
22410 State on the Title page the name of the publisher of the
22411 Modified Version, as the publisher.
22414 Preserve all the copyright notices of the Document.
22417 Add an appropriate copyright notice for your modifications
22418 adjacent to the other copyright notices.
22421 Include, immediately after the copyright notices, a license notice
22422 giving the public permission to use the Modified Version under the
22423 terms of this License, in the form shown in the Addendum below.
22426 Preserve in that license notice the full lists of Invariant Sections
22427 and required Cover Texts given in the Document's license notice.
22430 Include an unaltered copy of this License.
22433 Preserve the section Entitled ``History'', Preserve its Title, and add
22434 to it an item stating at least the title, year, new authors, and
22435 publisher of the Modified Version as given on the Title Page. If
22436 there is no section Entitled ``History'' in the Document, create one
22437 stating the title, year, authors, and publisher of the Document as
22438 given on its Title Page, then add an item describing the Modified
22439 Version as stated in the previous sentence.
22442 Preserve the network location, if any, given in the Document for
22443 public access to a Transparent copy of the Document, and likewise
22444 the network locations given in the Document for previous versions
22445 it was based on. These may be placed in the ``History'' section.
22446 You may omit a network location for a work that was published at
22447 least four years before the Document itself, or if the original
22448 publisher of the version it refers to gives permission.
22451 For any section Entitled ``Acknowledgements'' or ``Dedications'', Preserve
22452 the Title of the section, and preserve in the section all the
22453 substance and tone of each of the contributor acknowledgements and/or
22454 dedications given therein.
22457 Preserve all the Invariant Sections of the Document,
22458 unaltered in their text and in their titles. Section numbers
22459 or the equivalent are not considered part of the section titles.
22462 Delete any section Entitled ``Endorsements''. Such a section
22463 may not be included in the Modified Version.
22466 Do not retitle any existing section to be Entitled ``Endorsements'' or
22467 to conflict in title with any Invariant Section.
22470 Preserve any Warranty Disclaimers.
22473 If the Modified Version includes new front-matter sections or
22474 appendices that qualify as Secondary Sections and contain no material
22475 copied from the Document, you may at your option designate some or all
22476 of these sections as invariant. To do this, add their titles to the
22477 list of Invariant Sections in the Modified Version's license notice.
22478 These titles must be distinct from any other section titles.
22480 You may add a section Entitled ``Endorsements'', provided it contains
22481 nothing but endorsements of your Modified Version by various
22482 parties---for example, statements of peer review or that the text has
22483 been approved by an organization as the authoritative definition of a
22486 You may add a passage of up to five words as a Front-Cover Text, and a
22487 passage of up to 25 words as a Back-Cover Text, to the end of the list
22488 of Cover Texts in the Modified Version. Only one passage of
22489 Front-Cover Text and one of Back-Cover Text may be added by (or
22490 through arrangements made by) any one entity. If the Document already
22491 includes a cover text for the same cover, previously added by you or
22492 by arrangement made by the same entity you are acting on behalf of,
22493 you may not add another; but you may replace the old one, on explicit
22494 permission from the previous publisher that added the old one.
22496 The author(s) and publisher(s) of the Document do not by this License
22497 give permission to use their names for publicity for or to assert or
22498 imply endorsement of any Modified Version.
22501 COMBINING DOCUMENTS
22503 You may combine the Document with other documents released under this
22504 License, under the terms defined in section 4 above for modified
22505 versions, provided that you include in the combination all of the
22506 Invariant Sections of all of the original documents, unmodified, and
22507 list them all as Invariant Sections of your combined work in its
22508 license notice, and that you preserve all their Warranty Disclaimers.
22510 The combined work need only contain one copy of this License, and
22511 multiple identical Invariant Sections may be replaced with a single
22512 copy. If there are multiple Invariant Sections with the same name but
22513 different contents, make the title of each such section unique by
22514 adding at the end of it, in parentheses, the name of the original
22515 author or publisher of that section if known, or else a unique number.
22516 Make the same adjustment to the section titles in the list of
22517 Invariant Sections in the license notice of the combined work.
22519 In the combination, you must combine any sections Entitled ``History''
22520 in the various original documents, forming one section Entitled
22521 ``History''; likewise combine any sections Entitled ``Acknowledgements'',
22522 and any sections Entitled ``Dedications''. You must delete all
22523 sections Entitled ``Endorsements.''
22526 COLLECTIONS OF DOCUMENTS
22528 You may make a collection consisting of the Document and other documents
22529 released under this License, and replace the individual copies of this
22530 License in the various documents with a single copy that is included in
22531 the collection, provided that you follow the rules of this License for
22532 verbatim copying of each of the documents in all other respects.
22534 You may extract a single document from such a collection, and distribute
22535 it individually under this License, provided you insert a copy of this
22536 License into the extracted document, and follow this License in all
22537 other respects regarding verbatim copying of that document.
22540 AGGREGATION WITH INDEPENDENT WORKS
22542 A compilation of the Document or its derivatives with other separate
22543 and independent documents or works, in or on a volume of a storage or
22544 distribution medium, is called an ``aggregate'' if the copyright
22545 resulting from the compilation is not used to limit the legal rights
22546 of the compilation's users beyond what the individual works permit.
22547 When the Document is included in an aggregate, this License does not
22548 apply to the other works in the aggregate which are not themselves
22549 derivative works of the Document.
22551 If the Cover Text requirement of section 3 is applicable to these
22552 copies of the Document, then if the Document is less than one half of
22553 the entire aggregate, the Document's Cover Texts may be placed on
22554 covers that bracket the Document within the aggregate, or the
22555 electronic equivalent of covers if the Document is in electronic form.
22556 Otherwise they must appear on printed covers that bracket the whole
22562 Translation is considered a kind of modification, so you may
22563 distribute translations of the Document under the terms of section 4.
22564 Replacing Invariant Sections with translations requires special
22565 permission from their copyright holders, but you may include
22566 translations of some or all Invariant Sections in addition to the
22567 original versions of these Invariant Sections. You may include a
22568 translation of this License, and all the license notices in the
22569 Document, and any Warranty Disclaimers, provided that you also include
22570 the original English version of this License and the original versions
22571 of those notices and disclaimers. In case of a disagreement between
22572 the translation and the original version of this License or a notice
22573 or disclaimer, the original version will prevail.
22575 If a section in the Document is Entitled ``Acknowledgements'',
22576 ``Dedications'', or ``History'', the requirement (section 4) to Preserve
22577 its Title (section 1) will typically require changing the actual
22583 You may not copy, modify, sublicense, or distribute the Document except
22584 as expressly provided for under this License. Any other attempt to
22585 copy, modify, sublicense or distribute the Document is void, and will
22586 automatically terminate your rights under this License. However,
22587 parties who have received copies, or rights, from you under this
22588 License will not have their licenses terminated so long as such
22589 parties remain in full compliance.
22592 FUTURE REVISIONS OF THIS LICENSE
22594 The Free Software Foundation may publish new, revised versions
22595 of the GNU Free Documentation License from time to time. Such new
22596 versions will be similar in spirit to the present version, but may
22597 differ in detail to address new problems or concerns. See
22598 @uref{http://www.gnu.org/copyleft/}.
22600 Each version of the License is given a distinguishing version number.
22601 If the Document specifies that a particular numbered version of this
22602 License ``or any later version'' applies to it, you have the option of
22603 following the terms and conditions either of that specified version or
22604 of any later version that has been published (not as a draft) by the
22605 Free Software Foundation. If the Document does not specify a version
22606 number of this License, you may choose any version ever published (not
22607 as a draft) by the Free Software Foundation.
22611 @appendixsubsec ADDENDUM: How to use this License for your documents
22613 To use this License in a document you have written, include a copy of
22614 the License in the document and put the following copyright and
22615 license notices just after the title page:
22619 Copyright (C) @var{year} @var{your name}.
22620 Permission is granted to copy, distribute and/or modify this document
22621 under the terms of the GNU Free Documentation License, Version 1.2
22622 or any later version published by the Free Software Foundation;
22623 with no Invariant Sections, no Front-Cover Texts, and no Back-Cover Texts.
22624 A copy of the license is included in the section entitled ``GNU
22625 Free Documentation License''.
22629 If you have Invariant Sections, Front-Cover Texts and Back-Cover Texts,
22630 replace the ``with...Texts.'' line with this:
22634 with the Invariant Sections being @var{list their titles}, with
22635 the Front-Cover Texts being @var{list}, and with the Back-Cover Texts
22640 If you have Invariant Sections without Cover Texts, or some other
22641 combination of the three, merge those two alternatives to suit the
22644 If your document contains nontrivial examples of program code, we
22645 recommend releasing these examples in parallel under your choice of
22646 free software license, such as the GNU General Public License,
22647 to permit their use in free software.
22649 @node Index, About the Author, GNU Free Documentation License, Top
22650 @comment node-name, next, previous, up
22654 MENU ENTRY: NODE NAME.
22660 @c Place biographical information on right-hand (verso) page
22664 \par\vfill\supereject
22665 \global\evenheadline={\hfil} \global\evenfootline={\hfil}
22666 \global\oddheadline={\hfil} \global\oddfootline={\hfil}
22669 \par\vfill\supereject
22670 \par\vfill\supereject
22671 \global\evenheadline={\hfil} \global\evenfootline={\hfil}
22672 \global\oddheadline={\hfil} \global\oddfootline={\hfil}
22681 @c ================ Biographical information ================
22685 @center About the Author
22690 @node About the Author, , Index, Top
22691 @unnumbered About the Author
22695 Robert J. Chassell has worked with GNU Emacs since 1985. He writes
22696 and edits, teaches Emacs and Emacs Lisp, and speaks throughout the
22697 world on software freedom. Chassell was a founding Director and
22698 Treasurer of the Free Software Foundation, Inc. He is co-author of
22699 the @cite{Texinfo} manual, and has edited more than a dozen other
22700 books. He graduated from Cambridge University, in England. He has an
22701 abiding interest in social and economic history and flies his own
22708 @c Prevent page number on blank verso, so eject it first.
22710 \par\vfill\supereject
22715 @evenheading @thispage @| @| @thistitle
22716 @oddheading @| @| @thispage
22722 arch-tag: da1a2154-531f-43a8-8e33-fc7faad10acf