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.)
840 * On Reading this Text::
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.
1099 * Names & Definitions::
1100 * Lisp Interpreter::
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
1137 * Whitespace in 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.
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::
1649 * Evaluating Inner 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::
1782 @node fill-column Example, Void Function, Variables, Variables
1784 @unnumberedsubsec @code{fill-column}, an Example Variable
1787 @findex fill-column, @r{an example variable}
1788 @cindex Example variable, @code{fill-column}
1789 @cindex Variable, example of, @code{fill-column}
1790 The variable @code{fill-column} illustrates a symbol with a value
1791 attached to it: in every GNU Emacs buffer, this symbol is set to some
1792 value, usually 72 or 70, but sometimes to some other value. To find the
1793 value of this symbol, evaluate it by itself. If you are reading this in
1794 Info inside of GNU Emacs, you can do this by putting the cursor after
1795 the symbol and typing @kbd{C-x C-e}:
1802 After I typed @kbd{C-x C-e}, Emacs printed the number 72 in my echo
1803 area. This is the value for which @code{fill-column} is set for me as I
1804 write this. It may be different for you in your Info buffer. Notice
1805 that the value returned as a variable is printed in exactly the same way
1806 as the value returned by a function carrying out its instructions. From
1807 the point of view of the Lisp interpreter, a value returned is a value
1808 returned. What kind of expression it came from ceases to matter once
1811 A symbol can have any value attached to it or, to use the jargon, we can
1812 @dfn{bind} the variable to a value: to a number, such as 72; to a
1813 string, @code{"such as this"}; to a list, such as @code{(spruce pine
1814 oak)}; we can even bind a variable to a function definition.
1816 A symbol can be bound to a value in several ways. @xref{set & setq, ,
1817 Setting the Value of a Variable}, for information about one way to do
1820 @node Void Function, Void Variable, fill-column Example, Variables
1821 @comment node-name, next, previous, up
1822 @subsection Error Message for a Symbol Without a Function
1823 @cindex Symbol without function error
1824 @cindex Error for symbol without function
1826 When we evaluated @code{fill-column} to find its value as a variable,
1827 we did not place parentheses around the word. This is because we did
1828 not intend to use it as a function name.
1830 If @code{fill-column} were the first or only element of a list, the
1831 Lisp interpreter would attempt to find the function definition
1832 attached to it. But @code{fill-column} has no function definition.
1833 Try evaluating this:
1841 In GNU Emacs version 22, you will create a @file{*Backtrace*} buffer
1846 ---------- Buffer: *Backtrace* ----------
1847 Debugger entered--Lisp error: (void-function fill-column)
1850 eval-last-sexp-1(nil)
1852 call-interactively(eval-last-sexp)
1853 ---------- Buffer: *Backtrace* ----------
1858 (Remember, to quit the debugger and make the debugger window go away,
1859 type @kbd{q} in the @file{*Backtrace*} buffer.)
1863 In GNU Emacs 20 and before, you will produce an error message that says:
1866 Symbol's function definition is void:@: fill-column
1870 (The message will go away as soon as you move the cursor or type
1874 @node Void Variable, , Void Function, Variables
1875 @comment node-name, next, previous, up
1876 @subsection Error Message for a Symbol Without a Value
1877 @cindex Symbol without value error
1878 @cindex Error for symbol without value
1880 If you attempt to evaluate a symbol that does not have a value bound to
1881 it, you will receive an error message. You can see this by
1882 experimenting with our 2 plus 2 addition. In the following expression,
1883 put your cursor right after the @code{+}, before the first number 2,
1892 In GNU Emacs 22, you will create a @file{*Backtrace*} buffer that
1897 ---------- Buffer: *Backtrace* ----------
1898 Debugger entered--Lisp error: (void-variable +)
1900 eval-last-sexp-1(nil)
1902 call-interactively(eval-last-sexp)
1903 ---------- Buffer: *Backtrace* ----------
1908 (As with the other times we entered the debugger, you can quit by
1909 typing @kbd{q} in the @file{*Backtrace*} buffer.)
1911 This backtrace is different from the very first error message we saw,
1912 which said, @samp{Debugger entered--Lisp error: (void-function this)}.
1913 In this case, the function does not have a value as a variable; while
1914 in the other error message, the function (the word `this') did not
1917 In this experiment with the @code{+}, what we did was cause the Lisp
1918 interpreter to evaluate the @code{+} and look for the value of the
1919 variable instead of the function definition. We did this by placing the
1920 cursor right after the symbol rather than after the parenthesis of the
1921 enclosing list as we did before. As a consequence, the Lisp interpreter
1922 evaluated the preceding s-expression, which in this case was the
1925 Since @code{+} does not have a value bound to it, just the function
1926 definition, the error message reported that the symbol's value as a
1931 In GNU Emacs version 20 and before, your error message will say:
1934 Symbol's value as variable is void:@: +
1938 The meaning is the same as in GNU Emacs 22.
1941 @node Arguments, set & setq, Variables, List Processing
1942 @comment node-name, next, previous, up
1945 @cindex Passing information to functions
1947 To see how information is passed to functions, let's look again at
1948 our old standby, the addition of two plus two. In Lisp, this is written
1955 If you evaluate this expression, the number 4 will appear in your echo
1956 area. What the Lisp interpreter does is add the numbers that follow
1959 @cindex @samp{argument} defined
1960 The numbers added by @code{+} are called the @dfn{arguments} of the
1961 function @code{+}. These numbers are the information that is given to
1962 or @dfn{passed} to the function.
1964 The word `argument' comes from the way it is used in mathematics and
1965 does not refer to a disputation between two people; instead it refers to
1966 the information presented to the function, in this case, to the
1967 @code{+}. In Lisp, the arguments to a function are the atoms or lists
1968 that follow the function. The values returned by the evaluation of
1969 these atoms or lists are passed to the function. Different functions
1970 require different numbers of arguments; some functions require none at
1971 all.@footnote{It is curious to track the path by which the word `argument'
1972 came to have two different meanings, one in mathematics and the other in
1973 everyday English. According to the @cite{Oxford English Dictionary},
1974 the word derives from the Latin for @samp{to make clear, prove}; thus it
1975 came to mean, by one thread of derivation, `the evidence offered as
1976 proof', which is to say, `the information offered', which led to its
1977 meaning in Lisp. But in the other thread of derivation, it came to mean
1978 `to assert in a manner against which others may make counter
1979 assertions', which led to the meaning of the word as a disputation.
1980 (Note here that the English word has two different definitions attached
1981 to it at the same time. By contrast, in Emacs Lisp, a symbol cannot
1982 have two different function definitions at the same time.)}
1986 * Args as Variable or List::
1987 * Variable Number of Arguments::
1988 * Wrong Type of Argument::
1992 @node Data types, Args as Variable or List, Arguments, Arguments
1993 @comment node-name, next, previous, up
1994 @subsection Arguments' Data Types
1996 @cindex Types of data
1997 @cindex Arguments' data types
1999 The type of data that should be passed to a function depends on what
2000 kind of information it uses. The arguments to a function such as
2001 @code{+} must have values that are numbers, since @code{+} adds numbers.
2002 Other functions use different kinds of data for their arguments.
2006 For example, the @code{concat} function links together or unites two or
2007 more strings of text to produce a string. The arguments are strings.
2008 Concatenating the two character strings @code{abc}, @code{def} produces
2009 the single string @code{abcdef}. This can be seen by evaluating the
2013 (concat "abc" "def")
2017 The value produced by evaluating this expression is @code{"abcdef"}.
2019 A function such as @code{substring} uses both a string and numbers as
2020 arguments. The function returns a part of the string, a substring of
2021 the first argument. This function takes three arguments. Its first
2022 argument is the string of characters, the second and third arguments are
2023 numbers that indicate the beginning and end of the substring. The
2024 numbers are a count of the number of characters (including spaces and
2025 punctuations) from the beginning of the string.
2028 For example, if you evaluate the following:
2031 (substring "The quick brown fox jumped." 16 19)
2035 you will see @code{"fox"} appear in the echo area. The arguments are the
2036 string and the two numbers.
2038 Note that the string passed to @code{substring} is a single atom even
2039 though it is made up of several words separated by spaces. Lisp counts
2040 everything between the two quotation marks as part of the string,
2041 including the spaces. You can think of the @code{substring} function as
2042 a kind of `atom smasher' since it takes an otherwise indivisible atom
2043 and extracts a part. However, @code{substring} is only able to extract
2044 a substring from an argument that is a string, not from another type of
2045 atom such as a number or symbol.
2047 @node Args as Variable or List, Variable Number of Arguments, Data types, Arguments
2048 @comment node-name, next, previous, up
2049 @subsection An Argument as the Value of a Variable or List
2051 An argument can be a symbol that returns a value when it is evaluated.
2052 For example, when the symbol @code{fill-column} by itself is evaluated,
2053 it returns a number. This number can be used in an addition.
2056 Position the cursor after the following expression and type @kbd{C-x
2064 The value will be a number two more than what you get by evaluating
2065 @code{fill-column} alone. For me, this is 74, because my value of
2066 @code{fill-column} is 72.
2068 As we have just seen, an argument can be a symbol that returns a value
2069 when evaluated. In addition, an argument can be a list that returns a
2070 value when it is evaluated. For example, in the following expression,
2071 the arguments to the function @code{concat} are the strings
2072 @w{@code{"The "}} and @w{@code{" red foxes."}} and the list
2073 @code{(number-to-string (+ 2 fill-column))}.
2075 @c For GNU Emacs 22, need number-to-string
2077 (concat "The " (number-to-string (+ 2 fill-column)) " red foxes.")
2081 If you evaluate this expression---and if, as with my Emacs,
2082 @code{fill-column} evaluates to 72---@code{"The 74 red foxes."} will
2083 appear in the echo area. (Note that you must put spaces after the
2084 word @samp{The} and before the word @samp{red} so they will appear in
2085 the final string. The function @code{number-to-string} converts the
2086 integer that the addition function returns to a string.
2087 @code{number-to-string} is also known as @code{int-to-string}.)
2089 @node Variable Number of Arguments, Wrong Type of Argument, Args as Variable or List, Arguments
2090 @comment node-name, next, previous, up
2091 @subsection Variable Number of Arguments
2092 @cindex Variable number of arguments
2093 @cindex Arguments, variable number of
2095 Some functions, such as @code{concat}, @code{+} or @code{*}, take any
2096 number of arguments. (The @code{*} is the symbol for multiplication.)
2097 This can be seen by evaluating each of the following expressions in
2098 the usual way. What you will see in the echo area is printed in this
2099 text after @samp{@result{}}, which you may read as `evaluates to'.
2102 In the first set, the functions have no arguments:
2113 In this set, the functions have one argument each:
2124 In this set, the functions have three arguments each:
2128 (+ 3 4 5) @result{} 12
2130 (* 3 4 5) @result{} 60
2134 @node Wrong Type of Argument, message, Variable Number of Arguments, Arguments
2135 @comment node-name, next, previous, up
2136 @subsection Using the Wrong Type Object as an Argument
2137 @cindex Wrong type of argument
2138 @cindex Argument, wrong type of
2140 When a function is passed an argument of the wrong type, the Lisp
2141 interpreter produces an error message. For example, the @code{+}
2142 function expects the values of its arguments to be numbers. As an
2143 experiment we can pass it the quoted symbol @code{hello} instead of a
2144 number. Position the cursor after the following expression and type
2152 When you do this you will generate an error message. What has happened
2153 is that @code{+} has tried to add the 2 to the value returned by
2154 @code{'hello}, but the value returned by @code{'hello} is the symbol
2155 @code{hello}, not a number. Only numbers can be added. So @code{+}
2156 could not carry out its addition.
2159 In GNU Emacs version 22, you will create and enter a
2160 @file{*Backtrace*} buffer that says:
2165 ---------- Buffer: *Backtrace* ----------
2166 Debugger entered--Lisp error:
2167 (wrong-type-argument number-or-marker-p hello)
2169 eval((+ 2 (quote hello)))
2170 eval-last-sexp-1(nil)
2172 call-interactively(eval-last-sexp)
2173 ---------- Buffer: *Backtrace* ----------
2178 As usual, the error message tries to be helpful and makes sense after you
2179 learn how to read it.@footnote{@code{(quote hello)} is an expansion of
2180 the abbreviation @code{'hello}.}
2182 The first part of the error message is straightforward; it says
2183 @samp{wrong type argument}. Next comes the mysterious jargon word
2184 @w{@samp{number-or-marker-p}}. This word is trying to tell you what
2185 kind of argument the @code{+} expected.
2187 The symbol @code{number-or-marker-p} says that the Lisp interpreter is
2188 trying to determine whether the information presented it (the value of
2189 the argument) is a number or a marker (a special object representing a
2190 buffer position). What it does is test to see whether the @code{+} is
2191 being given numbers to add. It also tests to see whether the
2192 argument is something called a marker, which is a specific feature of
2193 Emacs Lisp. (In Emacs, locations in a buffer are recorded as markers.
2194 When the mark is set with the @kbd{C-@@} or @kbd{C-@key{SPC}} command,
2195 its position is kept as a marker. The mark can be considered a
2196 number---the number of characters the location is from the beginning
2197 of the buffer.) In Emacs Lisp, @code{+} can be used to add the
2198 numeric value of marker positions as numbers.
2200 The @samp{p} of @code{number-or-marker-p} is the embodiment of a
2201 practice started in the early days of Lisp programming. The @samp{p}
2202 stands for `predicate'. In the jargon used by the early Lisp
2203 researchers, a predicate refers to a function to determine whether some
2204 property is true or false. So the @samp{p} tells us that
2205 @code{number-or-marker-p} is the name of a function that determines
2206 whether it is true or false that the argument supplied is a number or
2207 a marker. Other Lisp symbols that end in @samp{p} include @code{zerop},
2208 a function that tests whether its argument has the value of zero, and
2209 @code{listp}, a function that tests whether its argument is a list.
2211 Finally, the last part of the error message is the symbol @code{hello}.
2212 This is the value of the argument that was passed to @code{+}. If the
2213 addition had been passed the correct type of object, the value passed
2214 would have been a number, such as 37, rather than a symbol like
2215 @code{hello}. But then you would not have got the error message.
2219 In GNU Emacs version 20 and before, the echo area displays an error
2223 Wrong type argument:@: number-or-marker-p, hello
2226 This says, in different words, the same as the top line of the
2227 @file{*Backtrace*} buffer.
2230 @node message, , Wrong Type of Argument, Arguments
2231 @comment node-name, next, previous, up
2232 @subsection The @code{message} Function
2235 Like @code{+}, the @code{message} function takes a variable number of
2236 arguments. It is used to send messages to the user and is so useful
2237 that we will describe it here.
2240 A message is printed in the echo area. For example, you can print a
2241 message in your echo area by evaluating the following list:
2244 (message "This message appears in the echo area!")
2247 The whole string between double quotation marks is a single argument
2248 and is printed @i{in toto}. (Note that in this example, the message
2249 itself will appear in the echo area within double quotes; that is
2250 because you see the value returned by the @code{message} function. In
2251 most uses of @code{message} in programs that you write, the text will
2252 be printed in the echo area as a side-effect, without the quotes.
2253 @xref{multiply-by-seven in detail, , @code{multiply-by-seven} in
2254 detail}, for an example of this.)
2256 However, if there is a @samp{%s} in the quoted string of characters, the
2257 @code{message} function does not print the @samp{%s} as such, but looks
2258 to the argument that follows the string. It evaluates the second
2259 argument and prints the value at the location in the string where the
2263 You can see this by positioning the cursor after the following
2264 expression and typing @kbd{C-x C-e}:
2267 (message "The name of this buffer is: %s." (buffer-name))
2271 In Info, @code{"The name of this buffer is: *info*."} will appear in the
2272 echo area. The function @code{buffer-name} returns the name of the
2273 buffer as a string, which the @code{message} function inserts in place
2276 To print a value as an integer, use @samp{%d} in the same way as
2277 @samp{%s}. For example, to print a message in the echo area that
2278 states the value of the @code{fill-column}, evaluate the following:
2281 (message "The value of fill-column is %d." fill-column)
2285 On my system, when I evaluate this list, @code{"The value of
2286 fill-column is 72."} appears in my echo area@footnote{Actually, you
2287 can use @code{%s} to print a number. It is non-specific. @code{%d}
2288 prints only the part of a number left of a decimal point, and not
2289 anything that is not a number.}.
2291 If there is more than one @samp{%s} in the quoted string, the value of
2292 the first argument following the quoted string is printed at the
2293 location of the first @samp{%s} and the value of the second argument is
2294 printed at the location of the second @samp{%s}, and so on.
2297 For example, if you evaluate the following,
2301 (message "There are %d %s in the office!"
2302 (- fill-column 14) "pink elephants")
2307 a rather whimsical message will appear in your echo area. On my system
2308 it says, @code{"There are 58 pink elephants in the office!"}.
2310 The expression @code{(- fill-column 14)} is evaluated and the resulting
2311 number is inserted in place of the @samp{%d}; and the string in double
2312 quotes, @code{"pink elephants"}, is treated as a single argument and
2313 inserted in place of the @samp{%s}. (That is to say, a string between
2314 double quotes evaluates to itself, like a number.)
2316 Finally, here is a somewhat complex example that not only illustrates
2317 the computation of a number, but also shows how you can use an
2318 expression within an expression to generate the text that is substituted
2323 (message "He saw %d %s"
2327 "The quick brown foxes jumped." 16 21)
2332 In this example, @code{message} has three arguments: the string,
2333 @code{"He saw %d %s"}, the expression, @code{(- fill-column 32)}, and
2334 the expression beginning with the function @code{concat}. The value
2335 resulting from the evaluation of @code{(- fill-column 32)} is inserted
2336 in place of the @samp{%d}; and the value returned by the expression
2337 beginning with @code{concat} is inserted in place of the @samp{%s}.
2339 When your fill column is 70 and you evaluate the expression, the
2340 message @code{"He saw 38 red foxes leaping."} appears in your echo
2343 @node set & setq, Summary, Arguments, List Processing
2344 @comment node-name, next, previous, up
2345 @section Setting the Value of a Variable
2346 @cindex Variable, setting value
2347 @cindex Setting value of variable
2349 @cindex @samp{bind} defined
2350 There are several ways by which a variable can be given a value. One of
2351 the ways is to use either the function @code{set} or the function
2352 @code{setq}. Another way is to use @code{let} (@pxref{let}). (The
2353 jargon for this process is to @dfn{bind} a variable to a value.)
2355 The following sections not only describe how @code{set} and @code{setq}
2356 work but also illustrate how arguments are passed.
2364 @node Using set, Using setq, set & setq, set & setq
2365 @comment node-name, next, previous, up
2366 @subsection Using @code{set}
2369 To set the value of the symbol @code{flowers} to the list @code{'(rose
2370 violet daisy buttercup)}, evaluate the following expression by
2371 positioning the cursor after the expression and typing @kbd{C-x C-e}.
2374 (set 'flowers '(rose violet daisy buttercup))
2378 The list @code{(rose violet daisy buttercup)} will appear in the echo
2379 area. This is what is @emph{returned} by the @code{set} function. As a
2380 side effect, the symbol @code{flowers} is bound to the list; that is,
2381 the symbol @code{flowers}, which can be viewed as a variable, is given
2382 the list as its value. (This process, by the way, illustrates how a
2383 side effect to the Lisp interpreter, setting the value, can be the
2384 primary effect that we humans are interested in. This is because every
2385 Lisp function must return a value if it does not get an error, but it
2386 will only have a side effect if it is designed to have one.)
2388 After evaluating the @code{set} expression, you can evaluate the symbol
2389 @code{flowers} and it will return the value you just set. Here is the
2390 symbol. Place your cursor after it and type @kbd{C-x C-e}.
2397 When you evaluate @code{flowers}, the list
2398 @code{(rose violet daisy buttercup)} appears in the echo area.
2400 Incidentally, if you evaluate @code{'flowers}, the variable with a quote
2401 in front of it, what you will see in the echo area is the symbol itself,
2402 @code{flowers}. Here is the quoted symbol, so you can try this:
2408 Note also, that when you use @code{set}, you need to quote both
2409 arguments to @code{set}, unless you want them evaluated. Since we do
2410 not want either argument evaluated, neither the variable
2411 @code{flowers} nor the list @code{(rose violet daisy buttercup)}, both
2412 are quoted. (When you use @code{set} without quoting its first
2413 argument, the first argument is evaluated before anything else is
2414 done. If you did this and @code{flowers} did not have a value
2415 already, you would get an error message that the @samp{Symbol's value
2416 as variable is void}; on the other hand, if @code{flowers} did return
2417 a value after it was evaluated, the @code{set} would attempt to set
2418 the value that was returned. There are situations where this is the
2419 right thing for the function to do; but such situations are rare.)
2421 @node Using setq, Counting, Using set, set & setq
2422 @comment node-name, next, previous, up
2423 @subsection Using @code{setq}
2426 As a practical matter, you almost always quote the first argument to
2427 @code{set}. The combination of @code{set} and a quoted first argument
2428 is so common that it has its own name: the special form @code{setq}.
2429 This special form is just like @code{set} except that the first argument
2430 is quoted automatically, so you don't need to type the quote mark
2431 yourself. Also, as an added convenience, @code{setq} permits you to set
2432 several different variables to different values, all in one expression.
2434 To set the value of the variable @code{carnivores} to the list
2435 @code{'(lion tiger leopard)} using @code{setq}, the following expression
2439 (setq carnivores '(lion tiger leopard))
2443 This is exactly the same as using @code{set} except the first argument
2444 is automatically quoted by @code{setq}. (The @samp{q} in @code{setq}
2445 means @code{quote}.)
2448 With @code{set}, the expression would look like this:
2451 (set 'carnivores '(lion tiger leopard))
2454 Also, @code{setq} can be used to assign different values to
2455 different variables. The first argument is bound to the value
2456 of the second argument, the third argument is bound to the value of the
2457 fourth argument, and so on. For example, you could use the following to
2458 assign a list of trees to the symbol @code{trees} and a list of herbivores
2459 to the symbol @code{herbivores}:
2463 (setq trees '(pine fir oak maple)
2464 herbivores '(gazelle antelope zebra))
2469 (The expression could just as well have been on one line, but it might
2470 not have fit on a page; and humans find it easier to read nicely
2473 Although I have been using the term `assign', there is another way of
2474 thinking about the workings of @code{set} and @code{setq}; and that is to
2475 say that @code{set} and @code{setq} make the symbol @emph{point} to the
2476 list. This latter way of thinking is very common and in forthcoming
2477 chapters we shall come upon at least one symbol that has `pointer' as
2478 part of its name. The name is chosen because the symbol has a value,
2479 specifically a list, attached to it; or, expressed another way,
2480 the symbol is set to ``point'' to the list.
2482 @node Counting, , Using setq, set & setq
2483 @comment node-name, next, previous, up
2484 @subsection Counting
2487 Here is an example that shows how to use @code{setq} in a counter. You
2488 might use this to count how many times a part of your program repeats
2489 itself. First set a variable to zero; then add one to the number each
2490 time the program repeats itself. To do this, you need a variable that
2491 serves as a counter, and two expressions: an initial @code{setq}
2492 expression that sets the counter variable to zero; and a second
2493 @code{setq} expression that increments the counter each time it is
2498 (setq counter 0) ; @r{Let's call this the initializer.}
2500 (setq counter (+ counter 1)) ; @r{This is the incrementer.}
2502 counter ; @r{This is the counter.}
2507 (The text following the @samp{;} are comments. @xref{Change a
2508 defun, , Change a Function Definition}.)
2510 If you evaluate the first of these expressions, the initializer,
2511 @code{(setq counter 0)}, and then evaluate the third expression,
2512 @code{counter}, the number @code{0} will appear in the echo area. If
2513 you then evaluate the second expression, the incrementer, @code{(setq
2514 counter (+ counter 1))}, the counter will get the value 1. So if you
2515 again evaluate @code{counter}, the number @code{1} will appear in the
2516 echo area. Each time you evaluate the second expression, the value of
2517 the counter will be incremented.
2519 When you evaluate the incrementer, @code{(setq counter (+ counter 1))},
2520 the Lisp interpreter first evaluates the innermost list; this is the
2521 addition. In order to evaluate this list, it must evaluate the variable
2522 @code{counter} and the number @code{1}. When it evaluates the variable
2523 @code{counter}, it receives its current value. It passes this value and
2524 the number @code{1} to the @code{+} which adds them together. The sum
2525 is then returned as the value of the inner list and passed to the
2526 @code{setq} which sets the variable @code{counter} to this new value.
2527 Thus, the value of the variable, @code{counter}, is changed.
2529 @node Summary, Error Message Exercises, set & setq, List Processing
2530 @comment node-name, next, previous, up
2533 Learning Lisp is like climbing a hill in which the first part is the
2534 steepest. You have now climbed the most difficult part; what remains
2535 becomes easier as you progress onwards.
2543 Lisp programs are made up of expressions, which are lists or single atoms.
2546 Lists are made up of zero or more atoms or inner lists, separated by whitespace and
2547 surrounded by parentheses. A list can be empty.
2550 Atoms are multi-character symbols, like @code{forward-paragraph}, single
2551 character symbols like @code{+}, strings of characters between double
2552 quotation marks, or numbers.
2555 A number evaluates to itself.
2558 A string between double quotes also evaluates to itself.
2561 When you evaluate a symbol by itself, its value is returned.
2564 When you evaluate a list, the Lisp interpreter looks at the first symbol
2565 in the list and then at the function definition bound to that symbol.
2566 Then the instructions in the function definition are carried out.
2569 A single quotation mark,
2576 , tells the Lisp interpreter that it should
2577 return the following expression as written, and not evaluate it as it
2578 would if the quote were not there.
2581 Arguments are the information passed to a function. The arguments to a
2582 function are computed by evaluating the rest of the elements of the list
2583 of which the function is the first element.
2586 A function always returns a value when it is evaluated (unless it gets
2587 an error); in addition, it may also carry out some action called a
2588 ``side effect''. In many cases, a function's primary purpose is to
2589 create a side effect.
2592 @node Error Message Exercises, , Summary, List Processing
2593 @comment node-name, next, previous, up
2596 A few simple exercises:
2600 Generate an error message by evaluating an appropriate symbol that is
2601 not within parentheses.
2604 Generate an error message by evaluating an appropriate symbol that is
2605 between parentheses.
2608 Create a counter that increments by two rather than one.
2611 Write an expression that prints a message in the echo area when
2615 @node Practicing Evaluation, Writing Defuns, List Processing, Top
2616 @comment node-name, next, previous, up
2617 @chapter Practicing Evaluation
2618 @cindex Practicing evaluation
2619 @cindex Evaluation practice
2621 Before learning how to write a function definition in Emacs Lisp, it is
2622 useful to spend a little time evaluating various expressions that have
2623 already been written. These expressions will be lists with the
2624 functions as their first (and often only) element. Since some of the
2625 functions associated with buffers are both simple and interesting, we
2626 will start with those. In this section, we will evaluate a few of
2627 these. In another section, we will study the code of several other
2628 buffer-related functions, to see how they were written.
2634 * Switching Buffers::
2635 * Buffer Size & Locations::
2636 * Evaluation Exercise::
2639 @node How to Evaluate, Buffer Names, Practicing Evaluation, Practicing Evaluation
2641 @unnumberedsec How to Evaluate
2644 @i{Whenever you give an editing command} to Emacs Lisp, such as the
2645 command to move the cursor or to scroll the screen, @i{you are evaluating
2646 an expression,} the first element of which is a function. @i{This is
2649 @cindex @samp{interactive function} defined
2650 @cindex @samp{command} defined
2651 When you type keys, you cause the Lisp interpreter to evaluate an
2652 expression and that is how you get your results. Even typing plain text
2653 involves evaluating an Emacs Lisp function, in this case, one that uses
2654 @code{self-insert-command}, which simply inserts the character you
2655 typed. The functions you evaluate by typing keystrokes are called
2656 @dfn{interactive} functions, or @dfn{commands}; how you make a function
2657 interactive will be illustrated in the chapter on how to write function
2658 definitions. @xref{Interactive, , Making a Function Interactive}.
2660 In addition to typing keyboard commands, we have seen a second way to
2661 evaluate an expression: by positioning the cursor after a list and
2662 typing @kbd{C-x C-e}. This is what we will do in the rest of this
2663 section. There are other ways to evaluate an expression as well; these
2664 will be described as we come to them.
2666 Besides being used for practicing evaluation, the functions shown in the
2667 next few sections are important in their own right. A study of these
2668 functions makes clear the distinction between buffers and files, how to
2669 switch to a buffer, and how to determine a location within it.
2671 @node Buffer Names, Getting Buffers, How to Evaluate, Practicing Evaluation
2672 @comment node-name, next, previous, up
2673 @section Buffer Names
2675 @findex buffer-file-name
2677 The two functions, @code{buffer-name} and @code{buffer-file-name}, show
2678 the difference between a file and a buffer. When you evaluate the
2679 following expression, @code{(buffer-name)}, the name of the buffer
2680 appears in the echo area. When you evaluate @code{(buffer-file-name)},
2681 the name of the file to which the buffer refers appears in the echo
2682 area. Usually, the name returned by @code{(buffer-name)} is the same as
2683 the name of the file to which it refers, and the name returned by
2684 @code{(buffer-file-name)} is the full path-name of the file.
2686 A file and a buffer are two different entities. A file is information
2687 recorded permanently in the computer (unless you delete it). A buffer,
2688 on the other hand, is information inside of Emacs that will vanish at
2689 the end of the editing session (or when you kill the buffer). Usually,
2690 a buffer contains information that you have copied from a file; we say
2691 the buffer is @dfn{visiting} that file. This copy is what you work on
2692 and modify. Changes to the buffer do not change the file, until you
2693 save the buffer. When you save the buffer, the buffer is copied to the file
2694 and is thus saved permanently.
2697 If you are reading this in Info inside of GNU Emacs, you can evaluate
2698 each of the following expressions by positioning the cursor after it and
2699 typing @kbd{C-x C-e}.
2710 When I do this in Info, the value returned by evaluating
2711 @code{(buffer-name)} is @file{"*info*"}, and the value returned by
2712 evaluating @code{(buffer-file-name)} is @file{nil}.
2714 On the other hand, while I am writing this Introduction, the value
2715 returned by evaluating @code{(buffer-name)} is
2716 @file{"introduction.texinfo"}, and the value returned by evaluating
2717 @code{(buffer-file-name)} is
2718 @file{"/gnu/work/intro/introduction.texinfo"}.
2720 @cindex @code{nil}, history of word
2721 The former is the name of the buffer and the latter is the name of the
2722 file. In Info, the buffer name is @file{"*info*"}. Info does not
2723 point to any file, so the result of evaluating
2724 @code{(buffer-file-name)} is @file{nil}. The symbol @code{nil} is
2725 from the Latin word for `nothing'; in this case, it means that the
2726 buffer is not associated with any file. (In Lisp, @code{nil} is also
2727 used to mean `false' and is a synonym for the empty list, @code{()}.)
2729 When I am writing, the name of my buffer is
2730 @file{"introduction.texinfo"}. The name of the file to which it
2731 points is @file{"/gnu/work/intro/introduction.texinfo"}.
2733 (In the expressions, the parentheses tell the Lisp interpreter to
2734 treat @w{@code{buffer-name}} and @w{@code{buffer-file-name}} as
2735 functions; without the parentheses, the interpreter would attempt to
2736 evaluate the symbols as variables. @xref{Variables}.)
2738 In spite of the distinction between files and buffers, you will often
2739 find that people refer to a file when they mean a buffer and vice-verse.
2740 Indeed, most people say, ``I am editing a file,'' rather than saying,
2741 ``I am editing a buffer which I will soon save to a file.'' It is
2742 almost always clear from context what people mean. When dealing with
2743 computer programs, however, it is important to keep the distinction in mind,
2744 since the computer is not as smart as a person.
2746 @cindex Buffer, history of word
2747 The word `buffer', by the way, comes from the meaning of the word as a
2748 cushion that deadens the force of a collision. In early computers, a
2749 buffer cushioned the interaction between files and the computer's
2750 central processing unit. The drums or tapes that held a file and the
2751 central processing unit were pieces of equipment that were very
2752 different from each other, working at their own speeds, in spurts. The
2753 buffer made it possible for them to work together effectively.
2754 Eventually, the buffer grew from being an intermediary, a temporary
2755 holding place, to being the place where work is done. This
2756 transformation is rather like that of a small seaport that grew into a
2757 great city: once it was merely the place where cargo was warehoused
2758 temporarily before being loaded onto ships; then it became a business
2759 and cultural center in its own right.
2761 Not all buffers are associated with files. For example, a
2762 @file{*scratch*} buffer does not visit any file. Similarly, a
2763 @file{*Help*} buffer is not associated with any file.
2765 In the old days, when you lacked a @file{~/.emacs} file and started an
2766 Emacs session by typing the command @code{emacs} alone, without naming
2767 any files, Emacs started with the @file{*scratch*} buffer visible.
2768 Nowadays, you will see a splash screen. You can follow one of the
2769 commands suggested on the splash screen, visit a file, or press the
2770 spacebar to reach the @file{*scratch*} buffer.
2772 If you switch to the @file{*scratch*} buffer, type
2773 @code{(buffer-name)}, position the cursor after it, and then type
2774 @kbd{C-x C-e} to evaluate the expression. The name @code{"*scratch*"}
2775 will be returned and will appear in the echo area. @code{"*scratch*"}
2776 is the name of the buffer. When you type @code{(buffer-file-name)} in
2777 the @file{*scratch*} buffer and evaluate that, @code{nil} will appear
2778 in the echo area, just as it does when you evaluate
2779 @code{(buffer-file-name)} in Info.
2781 Incidentally, if you are in the @file{*scratch*} buffer and want the
2782 value returned by an expression to appear in the @file{*scratch*}
2783 buffer itself rather than in the echo area, type @kbd{C-u C-x C-e}
2784 instead of @kbd{C-x C-e}. This causes the value returned to appear
2785 after the expression. The buffer will look like this:
2788 (buffer-name)"*scratch*"
2792 You cannot do this in Info since Info is read-only and it will not allow
2793 you to change the contents of the buffer. But you can do this in any
2794 buffer you can edit; and when you write code or documentation (such as
2795 this book), this feature is very useful.
2797 @node Getting Buffers, Switching Buffers, Buffer Names, Practicing Evaluation
2798 @comment node-name, next, previous, up
2799 @section Getting Buffers
2800 @findex current-buffer
2801 @findex other-buffer
2802 @cindex Getting a buffer
2804 The @code{buffer-name} function returns the @emph{name} of the buffer;
2805 to get the buffer @emph{itself}, a different function is needed: the
2806 @code{current-buffer} function. If you use this function in code, what
2807 you get is the buffer itself.
2809 A name and the object or entity to which the name refers are different
2810 from each other. You are not your name. You are a person to whom
2811 others refer by name. If you ask to speak to George and someone hands you
2812 a card with the letters @samp{G}, @samp{e}, @samp{o}, @samp{r},
2813 @samp{g}, and @samp{e} written on it, you might be amused, but you would
2814 not be satisfied. You do not want to speak to the name, but to the
2815 person to whom the name refers. A buffer is similar: the name of the
2816 scratch buffer is @file{*scratch*}, but the name is not the buffer. To
2817 get a buffer itself, you need to use a function such as
2818 @code{current-buffer}.
2820 However, there is a slight complication: if you evaluate
2821 @code{current-buffer} in an expression on its own, as we will do here,
2822 what you see is a printed representation of the name of the buffer
2823 without the contents of the buffer. Emacs works this way for two
2824 reasons: the buffer may be thousands of lines long---too long to be
2825 conveniently displayed; and, another buffer may have the same contents
2826 but a different name, and it is important to distinguish between them.
2829 Here is an expression containing the function:
2836 If you evaluate this expression in Info in Emacs in the usual way,
2837 @file{#<buffer *info*>} will appear in the echo area. The special
2838 format indicates that the buffer itself is being returned, rather than
2841 Incidentally, while you can type a number or symbol into a program, you
2842 cannot do that with the printed representation of a buffer: the only way
2843 to get a buffer itself is with a function such as @code{current-buffer}.
2845 A related function is @code{other-buffer}. This returns the most
2846 recently selected buffer other than the one you are in currently, not
2847 a printed representation of its name. If you have recently switched
2848 back and forth from the @file{*scratch*} buffer, @code{other-buffer}
2849 will return that buffer.
2852 You can see this by evaluating the expression:
2859 You should see @file{#<buffer *scratch*>} appear in the echo area, or
2860 the name of whatever other buffer you switched back from most
2861 recently@footnote{Actually, by default, if the buffer from which you
2862 just switched is visible to you in another window, @code{other-buffer}
2863 will choose the most recent buffer that you cannot see; this is a
2864 subtlety that I often forget.}.
2866 @node Switching Buffers, Buffer Size & Locations, Getting Buffers, Practicing Evaluation
2867 @comment node-name, next, previous, up
2868 @section Switching Buffers
2869 @findex switch-to-buffer
2871 @cindex Switching to a buffer
2873 The @code{other-buffer} function actually provides a buffer when it is
2874 used as an argument to a function that requires one. We can see this
2875 by using @code{other-buffer} and @code{switch-to-buffer} to switch to a
2878 But first, a brief introduction to the @code{switch-to-buffer}
2879 function. When you switched back and forth from Info to the
2880 @file{*scratch*} buffer to evaluate @code{(buffer-name)}, you most
2881 likely typed @kbd{C-x b} and then typed @file{*scratch*}@footnote{Or
2882 rather, to save typing, you probably only typed @kbd{RET} if the
2883 default buffer was @file{*scratch*}, or if it was different, then you
2884 typed just part of the name, such as @code{*sc}, pressed your
2885 @kbd{TAB} key to cause it to expand to the full name, and then typed
2886 your @kbd{RET} key.} when prompted in the minibuffer for the name of
2887 the buffer to which you wanted to switch. The keystrokes, @kbd{C-x
2888 b}, cause the Lisp interpreter to evaluate the interactive function
2889 @code{switch-to-buffer}. As we said before, this is how Emacs works:
2890 different keystrokes call or run different functions. For example,
2891 @kbd{C-f} calls @code{forward-char}, @kbd{M-e} calls
2892 @code{forward-sentence}, and so on.
2894 By writing @code{switch-to-buffer} in an expression, and giving it a
2895 buffer to switch to, we can switch buffers just the way @kbd{C-x b}
2899 Here is the Lisp expression:
2902 (switch-to-buffer (other-buffer))
2906 The symbol @code{switch-to-buffer} is the first element of the list,
2907 so the Lisp interpreter will treat it as a function and carry out the
2908 instructions that are attached to it. But before doing that, the
2909 interpreter will note that @code{other-buffer} is inside parentheses
2910 and work on that symbol first. @code{other-buffer} is the first (and
2911 in this case, the only) element of this list, so the Lisp interpreter
2912 calls or runs the function. It returns another buffer. Next, the
2913 interpreter runs @code{switch-to-buffer}, passing to it, as an
2914 argument, the other buffer, which is what Emacs will switch to. If
2915 you are reading this in Info, try this now. Evaluate the expression.
2916 (To get back, type @kbd{C-x b @key{RET}}.)@footnote{Remember, this
2917 expression will move you to your most recent other buffer that you
2918 cannot see. If you really want to go to your most recently selected
2919 buffer, even if you can still see it, you need to evaluate the
2920 following more complex expression:
2923 (switch-to-buffer (other-buffer (current-buffer) t))
2927 In this case, the first argument to @code{other-buffer} tells it which
2928 buffer to skip---the current one---and the second argument tells
2929 @code{other-buffer} it is OK to switch to a visible buffer.
2930 In regular use, @code{switch-to-buffer} takes you to an invisible
2931 window since you would most likely use @kbd{C-x o} (@code{other-window})
2932 to go to another visible buffer.}
2934 In the programming examples in later sections of this document, you will
2935 see the function @code{set-buffer} more often than
2936 @code{switch-to-buffer}. This is because of a difference between
2937 computer programs and humans: humans have eyes and expect to see the
2938 buffer on which they are working on their computer terminals. This is
2939 so obvious, it almost goes without saying. However, programs do not
2940 have eyes. When a computer program works on a buffer, that buffer does
2941 not need to be visible on the screen.
2943 @code{switch-to-buffer} is designed for humans and does two different
2944 things: it switches the buffer to which Emacs' attention is directed; and
2945 it switches the buffer displayed in the window to the new buffer.
2946 @code{set-buffer}, on the other hand, does only one thing: it switches
2947 the attention of the computer program to a different buffer. The buffer
2948 on the screen remains unchanged (of course, normally nothing happens
2949 there until the command finishes running).
2951 @cindex @samp{call} defined
2952 Also, we have just introduced another jargon term, the word @dfn{call}.
2953 When you evaluate a list in which the first symbol is a function, you
2954 are calling that function. The use of the term comes from the notion of
2955 the function as an entity that can do something for you if you `call'
2956 it---just as a plumber is an entity who can fix a leak if you call him
2959 @node Buffer Size & Locations, Evaluation Exercise, Switching Buffers, Practicing Evaluation
2960 @comment node-name, next, previous, up
2961 @section Buffer Size and the Location of Point
2962 @cindex Size of buffer
2964 @cindex Point location
2965 @cindex Location of point
2967 Finally, let's look at several rather simple functions,
2968 @code{buffer-size}, @code{point}, @code{point-min}, and
2969 @code{point-max}. These give information about the size of a buffer and
2970 the location of point within it.
2972 The function @code{buffer-size} tells you the size of the current
2973 buffer; that is, the function returns a count of the number of
2974 characters in the buffer.
2981 You can evaluate this in the usual way, by positioning the
2982 cursor after the expression and typing @kbd{C-x C-e}.
2984 @cindex @samp{point} defined
2985 In Emacs, the current position of the cursor is called @dfn{point}.
2986 The expression @code{(point)} returns a number that tells you where the
2987 cursor is located as a count of the number of characters from the
2988 beginning of the buffer up to point.
2991 You can see the character count for point in this buffer by evaluating
2992 the following expression in the usual way:
2999 As I write this, the value of @code{point} is 65724. The @code{point}
3000 function is frequently used in some of the examples later in this
3004 The value of point depends, of course, on its location within the
3005 buffer. If you evaluate point in this spot, the number will be larger:
3012 For me, the value of point in this location is 66043, which means that
3013 there are 319 characters (including spaces) between the two
3014 expressions. (Doubtless, you will see different numbers, since I will
3015 have edited this since I first evaluated point.)
3017 @cindex @samp{narrowing} defined
3018 The function @code{point-min} is somewhat similar to @code{point}, but
3019 it returns the value of the minimum permissible value of point in the
3020 current buffer. This is the number 1 unless @dfn{narrowing} is in
3021 effect. (Narrowing is a mechanism whereby you can restrict yourself,
3022 or a program, to operations on just a part of a buffer.
3023 @xref{Narrowing & Widening, , Narrowing and Widening}.) Likewise, the
3024 function @code{point-max} returns the value of the maximum permissible
3025 value of point in the current buffer.
3027 @node Evaluation Exercise, , Buffer Size & Locations, Practicing Evaluation
3030 Find a file with which you are working and move towards its middle.
3031 Find its buffer name, file name, length, and your position in the file.
3033 @node Writing Defuns, Buffer Walk Through, Practicing Evaluation, Top
3034 @comment node-name, next, previous, up
3035 @chapter How To Write Function Definitions
3036 @cindex Definition writing
3037 @cindex Function definition writing
3038 @cindex Writing a function definition
3040 When the Lisp interpreter evaluates a list, it looks to see whether the
3041 first symbol on the list has a function definition attached to it; or,
3042 put another way, whether the symbol points to a function definition. If
3043 it does, the computer carries out the instructions in the definition. A
3044 symbol that has a function definition is called, simply, a function
3045 (although, properly speaking, the definition is the function and the
3046 symbol refers to it.)
3049 * Primitive Functions::
3053 * Interactive Options::
3054 * Permanent Installation::
3058 * Truth & Falsehood::
3064 @node Primitive Functions, defun, Writing Defuns, Writing Defuns
3066 @unnumberedsec An Aside about Primitive Functions
3068 @cindex Primitive functions
3069 @cindex Functions, primitive
3071 @cindex C language primitives
3072 @cindex Primitives written in C
3073 All functions are defined in terms of other functions, except for a few
3074 @dfn{primitive} functions that are written in the C programming
3075 language. When you write functions' definitions, you will write them in
3076 Emacs Lisp and use other functions as your building blocks. Some of the
3077 functions you will use will themselves be written in Emacs Lisp (perhaps
3078 by you) and some will be primitives written in C. The primitive
3079 functions are used exactly like those written in Emacs Lisp and behave
3080 like them. They are written in C so we can easily run GNU Emacs on any
3081 computer that has sufficient power and can run C.
3083 Let me re-emphasize this: when you write code in Emacs Lisp, you do not
3084 distinguish between the use of functions written in C and the use of
3085 functions written in Emacs Lisp. The difference is irrelevant. I
3086 mention the distinction only because it is interesting to know. Indeed,
3087 unless you investigate, you won't know whether an already-written
3088 function is written in Emacs Lisp or C.
3090 @node defun, Install, Primitive Functions, Writing Defuns
3091 @comment node-name, next, previous, up
3092 @section The @code{defun} Special Form
3094 @cindex Special form of @code{defun}
3096 @cindex @samp{function definition} defined
3097 In Lisp, a symbol such as @code{mark-whole-buffer} has code attached to
3098 it that tells the computer what to do when the function is called.
3099 This code is called the @dfn{function definition} and is created by
3100 evaluating a Lisp expression that starts with the symbol @code{defun}
3101 (which is an abbreviation for @emph{define function}). Because
3102 @code{defun} does not evaluate its arguments in the usual way, it is
3103 called a @dfn{special form}.
3105 In subsequent sections, we will look at function definitions from the
3106 Emacs source code, such as @code{mark-whole-buffer}. In this section,
3107 we will describe a simple function definition so you can see how it
3108 looks. This function definition uses arithmetic because it makes for a
3109 simple example. Some people dislike examples using arithmetic; however,
3110 if you are such a person, do not despair. Hardly any of the code we
3111 will study in the remainder of this introduction involves arithmetic or
3112 mathematics. The examples mostly involve text in one way or another.
3114 A function definition has up to five parts following the word
3119 The name of the symbol to which the function definition should be
3123 A list of the arguments that will be passed to the function. If no
3124 arguments will be passed to the function, this is an empty list,
3128 Documentation describing the function. (Technically optional, but
3129 strongly recommended.)
3132 Optionally, an expression to make the function interactive so you can
3133 use it by typing @kbd{M-x} and then the name of the function; or by
3134 typing an appropriate key or keychord.
3136 @cindex @samp{body} defined
3138 The code that instructs the computer what to do: the @dfn{body} of the
3139 function definition.
3142 It is helpful to think of the five parts of a function definition as
3143 being organized in a template, with slots for each part:
3147 (defun @var{function-name} (@var{arguments}@dots{})
3148 "@var{optional-documentation}@dots{}"
3149 (interactive @var{argument-passing-info}) ; @r{optional}
3154 As an example, here is the code for a function that multiplies its
3155 argument by 7. (This example is not interactive. @xref{Interactive,
3156 , Making a Function Interactive}, for that information.)
3160 (defun multiply-by-seven (number)
3161 "Multiply NUMBER by seven."
3166 This definition begins with a parenthesis and the symbol @code{defun},
3167 followed by the name of the function.
3169 @cindex @samp{argument list} defined
3170 The name of the function is followed by a list that contains the
3171 arguments that will be passed to the function. This list is called
3172 the @dfn{argument list}. In this example, the list has only one
3173 element, the symbol, @code{number}. When the function is used, the
3174 symbol will be bound to the value that is used as the argument to the
3177 Instead of choosing the word @code{number} for the name of the argument,
3178 I could have picked any other name. For example, I could have chosen
3179 the word @code{multiplicand}. I picked the word `number' because it
3180 tells what kind of value is intended for this slot; but I could just as
3181 well have chosen the word `multiplicand' to indicate the role that the
3182 value placed in this slot will play in the workings of the function. I
3183 could have called it @code{foogle}, but that would have been a bad
3184 choice because it would not tell humans what it means. The choice of
3185 name is up to the programmer and should be chosen to make the meaning of
3188 Indeed, you can choose any name you wish for a symbol in an argument
3189 list, even the name of a symbol used in some other function: the name
3190 you use in an argument list is private to that particular definition.
3191 In that definition, the name refers to a different entity than any use
3192 of the same name outside the function definition. Suppose you have a
3193 nick-name `Shorty' in your family; when your family members refer to
3194 `Shorty', they mean you. But outside your family, in a movie, for
3195 example, the name `Shorty' refers to someone else. Because a name in an
3196 argument list is private to the function definition, you can change the
3197 value of such a symbol inside the body of a function without changing
3198 its value outside the function. The effect is similar to that produced
3199 by a @code{let} expression. (@xref{let, , @code{let}}.)
3202 Note also that we discuss the word `number' in two different ways: as a
3203 symbol that appears in the code, and as the name of something that will
3204 be replaced by a something else during the evaluation of the function.
3205 In the first case, @code{number} is a symbol, not a number; it happens
3206 that within the function, it is a variable who value is the number in
3207 question, but our primary interest in it is as a symbol. On the other
3208 hand, when we are talking about the function, our interest is that we
3209 will substitute a number for the word @var{number}. To keep this
3210 distinction clear, we use different typography for the two
3211 circumstances. When we talk about this function, or about how it works,
3212 we refer to this number by writing @var{number}. In the function
3213 itself, we refer to it by writing @code{number}.
3216 The argument list is followed by the documentation string that
3217 describes the function. This is what you see when you type
3218 @w{@kbd{C-h f}} and the name of a function. Incidentally, when you
3219 write a documentation string like this, you should make the first line
3220 a complete sentence since some commands, such as @code{apropos}, print
3221 only the first line of a multi-line documentation string. Also, you
3222 should not indent the second line of a documentation string, if you
3223 have one, because that looks odd when you use @kbd{C-h f}
3224 (@code{describe-function}). The documentation string is optional, but
3225 it is so useful, it should be included in almost every function you
3228 @findex * @r{(multiplication)}
3229 The third line of the example consists of the body of the function
3230 definition. (Most functions' definitions, of course, are longer than
3231 this.) In this function, the body is the list, @code{(* 7 number)}, which
3232 says to multiply the value of @var{number} by 7. (In Emacs Lisp,
3233 @code{*} is the function for multiplication, just as @code{+} is the
3234 function for addition.)
3236 When you use the @code{multiply-by-seven} function, the argument
3237 @code{number} evaluates to the actual number you want used. Here is an
3238 example that shows how @code{multiply-by-seven} is used; but don't try
3239 to evaluate this yet!
3242 (multiply-by-seven 3)
3246 The symbol @code{number}, specified in the function definition in the
3247 next section, is given or ``bound to'' the value 3 in the actual use of
3248 the function. Note that although @code{number} was inside parentheses
3249 in the function definition, the argument passed to the
3250 @code{multiply-by-seven} function is not in parentheses. The
3251 parentheses are written in the function definition so the computer can
3252 figure out where the argument list ends and the rest of the function
3255 If you evaluate this example, you are likely to get an error message.
3256 (Go ahead, try it!) This is because we have written the function
3257 definition, but not yet told the computer about the definition---we have
3258 not yet installed (or `loaded') the function definition in Emacs.
3259 Installing a function is the process that tells the Lisp interpreter the
3260 definition of the function. Installation is described in the next
3263 @node Install, Interactive, defun, Writing Defuns
3264 @comment node-name, next, previous, up
3265 @section Install a Function Definition
3266 @cindex Install a Function Definition
3267 @cindex Definition installation
3268 @cindex Function definition installation
3270 If you are reading this inside of Info in Emacs, you can try out the
3271 @code{multiply-by-seven} function by first evaluating the function
3272 definition and then evaluating @code{(multiply-by-seven 3)}. A copy of
3273 the function definition follows. Place the cursor after the last
3274 parenthesis of the function definition and type @kbd{C-x C-e}. When you
3275 do this, @code{multiply-by-seven} will appear in the echo area. (What
3276 this means is that when a function definition is evaluated, the value it
3277 returns is the name of the defined function.) At the same time, this
3278 action installs the function definition.
3282 (defun multiply-by-seven (number)
3283 "Multiply NUMBER by seven."
3289 By evaluating this @code{defun}, you have just installed
3290 @code{multiply-by-seven} in Emacs. The function is now just as much a
3291 part of Emacs as @code{forward-word} or any other editing function you
3292 use. (@code{multiply-by-seven} will stay installed until you quit
3293 Emacs. To reload code automatically whenever you start Emacs, see
3294 @ref{Permanent Installation, , Installing Code Permanently}.)
3297 * Effect of installation::
3301 @node Effect of installation, Change a defun, Install, Install
3303 @unnumberedsubsec The effect of installation
3306 You can see the effect of installing @code{multiply-by-seven} by
3307 evaluating the following sample. Place the cursor after the following
3308 expression and type @kbd{C-x C-e}. The number 21 will appear in the
3312 (multiply-by-seven 3)
3315 If you wish, you can read the documentation for the function by typing
3316 @kbd{C-h f} (@code{describe-function}) and then the name of the
3317 function, @code{multiply-by-seven}. When you do this, a
3318 @file{*Help*} window will appear on your screen that says:
3322 multiply-by-seven is a Lisp function.
3323 (multiply-by-seven NUMBER)
3325 Multiply NUMBER by seven.
3330 (To return to a single window on your screen, type @kbd{C-x 1}.)
3332 @node Change a defun, , Effect of installation, Install
3333 @comment node-name, next, previous, up
3334 @subsection Change a Function Definition
3335 @cindex Changing a function definition
3336 @cindex Function definition, how to change
3337 @cindex Definition, how to change
3339 If you want to change the code in @code{multiply-by-seven}, just rewrite
3340 it. To install the new version in place of the old one, evaluate the
3341 function definition again. This is how you modify code in Emacs. It is
3344 As an example, you can change the @code{multiply-by-seven} function to
3345 add the number to itself seven times instead of multiplying the number
3346 by seven. It produces the same answer, but by a different path. At
3347 the same time, we will add a comment to the code; a comment is text
3348 that the Lisp interpreter ignores, but that a human reader may find
3349 useful or enlightening. The comment is that this is the ``second
3354 (defun multiply-by-seven (number) ; @r{Second version.}
3355 "Multiply NUMBER by seven."
3356 (+ number number number number number number number))
3360 @cindex Comments in Lisp code
3361 The comment follows a semicolon, @samp{;}. In Lisp, everything on a
3362 line that follows a semicolon is a comment. The end of the line is the
3363 end of the comment. To stretch a comment over two or more lines, begin
3364 each line with a semicolon.
3366 @xref{Beginning a .emacs File, , Beginning a @file{.emacs}
3367 File}, and @ref{Comments, , Comments, elisp, The GNU Emacs Lisp
3368 Reference Manual}, for more about comments.
3370 You can install this version of the @code{multiply-by-seven} function by
3371 evaluating it in the same way you evaluated the first function: place
3372 the cursor after the last parenthesis and type @kbd{C-x C-e}.
3374 In summary, this is how you write code in Emacs Lisp: you write a
3375 function; install it; test it; and then make fixes or enhancements and
3378 @node Interactive, Interactive Options, Install, Writing Defuns
3379 @comment node-name, next, previous, up
3380 @section Make a Function Interactive
3381 @cindex Interactive functions
3384 You make a function interactive by placing a list that begins with
3385 the special form @code{interactive} immediately after the
3386 documentation. A user can invoke an interactive function by typing
3387 @kbd{M-x} and then the name of the function; or by typing the keys to
3388 which it is bound, for example, by typing @kbd{C-n} for
3389 @code{next-line} or @kbd{C-x h} for @code{mark-whole-buffer}.
3391 Interestingly, when you call an interactive function interactively,
3392 the value returned is not automatically displayed in the echo area.
3393 This is because you often call an interactive function for its side
3394 effects, such as moving forward by a word or line, and not for the
3395 value returned. If the returned value were displayed in the echo area
3396 each time you typed a key, it would be very distracting.
3399 * Interactive multiply-by-seven::
3400 * multiply-by-seven in detail::
3403 @node Interactive multiply-by-seven, multiply-by-seven in detail, Interactive, Interactive
3405 @unnumberedsubsec An Interactive @code{multiply-by-seven}, An Overview
3408 Both the use of the special form @code{interactive} and one way to
3409 display a value in the echo area can be illustrated by creating an
3410 interactive version of @code{multiply-by-seven}.
3417 (defun multiply-by-seven (number) ; @r{Interactive version.}
3418 "Multiply NUMBER by seven."
3420 (message "The result is %d" (* 7 number)))
3425 You can install this code by placing your cursor after it and typing
3426 @kbd{C-x C-e}. The name of the function will appear in your echo area.
3427 Then, you can use this code by typing @kbd{C-u} and a number and then
3428 typing @kbd{M-x multiply-by-seven} and pressing @key{RET}. The phrase
3429 @samp{The result is @dots{}} followed by the product will appear in the
3432 Speaking more generally, you invoke a function like this in either of two
3437 By typing a prefix argument that contains the number to be passed, and
3438 then typing @kbd{M-x} and the name of the function, as with
3439 @kbd{C-u 3 M-x forward-sentence}; or,
3442 By typing whatever key or keychord the function is bound to, as with
3447 Both the examples just mentioned work identically to move point forward
3448 three sentences. (Since @code{multiply-by-seven} is not bound to a key,
3449 it could not be used as an example of key binding.)
3451 (@xref{Keybindings, , Some Keybindings}, to learn how to bind a command
3454 A prefix argument is passed to an interactive function by typing the
3455 @key{META} key followed by a number, for example, @kbd{M-3 M-e}, or by
3456 typing @kbd{C-u} and then a number, for example, @kbd{C-u 3 M-e} (if you
3457 type @kbd{C-u} without a number, it defaults to 4).
3459 @node multiply-by-seven in detail, , Interactive multiply-by-seven, Interactive
3460 @comment node-name, next, previous, up
3461 @subsection An Interactive @code{multiply-by-seven}
3463 Let's look at the use of the special form @code{interactive} and then at
3464 the function @code{message} in the interactive version of
3465 @code{multiply-by-seven}. You will recall that the function definition
3470 (defun multiply-by-seven (number) ; @r{Interactive version.}
3471 "Multiply NUMBER by seven."
3473 (message "The result is %d" (* 7 number)))
3477 In this function, the expression, @code{(interactive "p")}, is a list of
3478 two elements. The @code{"p"} tells Emacs to pass the prefix argument to
3479 the function and use its value for the argument of the function.
3482 The argument will be a number. This means that the symbol
3483 @code{number} will be bound to a number in the line:
3486 (message "The result is %d" (* 7 number))
3491 For example, if your prefix argument is 5, the Lisp interpreter will
3492 evaluate the line as if it were:
3495 (message "The result is %d" (* 7 5))
3499 (If you are reading this in GNU Emacs, you can evaluate this expression
3500 yourself.) First, the interpreter will evaluate the inner list, which
3501 is @code{(* 7 5)}. This returns a value of 35. Next, it
3502 will evaluate the outer list, passing the values of the second and
3503 subsequent elements of the list to the function @code{message}.
3505 As we have seen, @code{message} is an Emacs Lisp function especially
3506 designed for sending a one line message to a user. (@xref{message, ,
3507 The @code{message} function}.) In summary, the @code{message}
3508 function prints its first argument in the echo area as is, except for
3509 occurrences of @samp{%d} or @samp{%s} (and various other %-sequences
3510 which we have not mentioned). When it sees a control sequence, the
3511 function looks to the second or subsequent arguments and prints the
3512 value of the argument in the location in the string where the control
3513 sequence is located.
3515 In the interactive @code{multiply-by-seven} function, the control string
3516 is @samp{%d}, which requires a number, and the value returned by
3517 evaluating @code{(* 7 5)} is the number 35. Consequently, the number 35
3518 is printed in place of the @samp{%d} and the message is @samp{The result
3521 (Note that when you call the function @code{multiply-by-seven}, the
3522 message is printed without quotes, but when you call @code{message}, the
3523 text is printed in double quotes. This is because the value returned by
3524 @code{message} is what appears in the echo area when you evaluate an
3525 expression whose first element is @code{message}; but when embedded in a
3526 function, @code{message} prints the text as a side effect without
3529 @node Interactive Options, Permanent Installation, Interactive, Writing Defuns
3530 @comment node-name, next, previous, up
3531 @section Different Options for @code{interactive}
3532 @cindex Options for @code{interactive}
3533 @cindex Interactive options
3535 In the example, @code{multiply-by-seven} used @code{"p"} as the
3536 argument to @code{interactive}. This argument told Emacs to interpret
3537 your typing either @kbd{C-u} followed by a number or @key{META}
3538 followed by a number as a command to pass that number to the function
3539 as its argument. Emacs has more than twenty characters predefined for
3540 use with @code{interactive}. In almost every case, one of these
3541 options will enable you to pass the right information interactively to
3542 a function. (@xref{Interactive Codes, , Code Characters for
3543 @code{interactive}, elisp, The GNU Emacs Lisp Reference Manual}.)
3546 Consider the function @code{zap-to-char}. Its interactive expression
3550 (interactive "p\ncZap to char: ")
3553 The first part of the argument to @code{interactive} is @samp{p}, with
3554 which you are already familiar. This argument tells Emacs to
3555 interpret a `prefix', as a number to be passed to the function. You
3556 can specify a prefix either by typing @kbd{C-u} followed by a number
3557 or by typing @key{META} followed by a number. The prefix is the
3558 number of specified characters. Thus, if your prefix is three and the
3559 specified character is @samp{x}, then you will delete all the text up
3560 to and including the third next @samp{x}. If you do not set a prefix,
3561 then you delete all the text up to and including the specified
3562 character, but no more.
3564 The @samp{c} tells the function the name of the character to which to delete.
3566 More formally, a function with two or more arguments can have
3567 information passed to each argument by adding parts to the string that
3568 follows @code{interactive}. When you do this, the information is
3569 passed to each argument in the same order it is specified in the
3570 @code{interactive} list. In the string, each part is separated from
3571 the next part by a @samp{\n}, which is a newline. For example, you
3572 can follow @samp{p} with a @samp{\n} and an @samp{cZap to char:@: }.
3573 This causes Emacs to pass the value of the prefix argument (if there
3574 is one) and the character.
3576 In this case, the function definition looks like the following, where
3577 @code{arg} and @code{char} are the symbols to which @code{interactive}
3578 binds the prefix argument and the specified character:
3582 (defun @var{name-of-function} (arg char)
3583 "@var{documentation}@dots{}"
3584 (interactive "p\ncZap to char: ")
3585 @var{body-of-function}@dots{})
3590 (The space after the colon in the prompt makes it look better when you
3591 are prompted. @xref{copy-to-buffer, , The Definition of
3592 @code{copy-to-buffer}}, for an example.)
3594 When a function does not take arguments, @code{interactive} does not
3595 require any. Such a function contains the simple expression
3596 @code{(interactive)}. The @code{mark-whole-buffer} function is like
3599 Alternatively, if the special letter-codes are not right for your
3600 application, you can pass your own arguments to @code{interactive} as
3603 @xref{append-to-buffer, , The Definition of @code{append-to-buffer}},
3604 for an example. @xref{Using Interactive, , Using @code{Interactive},
3605 elisp, The GNU Emacs Lisp Reference Manual}, for a more complete
3606 explanation about this technique.
3608 @node Permanent Installation, let, Interactive Options, Writing Defuns
3609 @comment node-name, next, previous, up
3610 @section Install Code Permanently
3611 @cindex Install code permanently
3612 @cindex Permanent code installation
3613 @cindex Code installation
3615 When you install a function definition by evaluating it, it will stay
3616 installed until you quit Emacs. The next time you start a new session
3617 of Emacs, the function will not be installed unless you evaluate the
3618 function definition again.
3620 At some point, you may want to have code installed automatically
3621 whenever you start a new session of Emacs. There are several ways of
3626 If you have code that is just for yourself, you can put the code for the
3627 function definition in your @file{.emacs} initialization file. When you
3628 start Emacs, your @file{.emacs} file is automatically evaluated and all
3629 the function definitions within it are installed.
3630 @xref{Emacs Initialization, , Your @file{.emacs} File}.
3633 Alternatively, you can put the function definitions that you want
3634 installed in one or more files of their own and use the @code{load}
3635 function to cause Emacs to evaluate and thereby install each of the
3636 functions in the files.
3637 @xref{Loading Files, , Loading Files}.
3640 Thirdly, if you have code that your whole site will use, it is usual
3641 to put it in a file called @file{site-init.el} that is loaded when
3642 Emacs is built. This makes the code available to everyone who uses
3643 your machine. (See the @file{INSTALL} file that is part of the Emacs
3647 Finally, if you have code that everyone who uses Emacs may want, you
3648 can post it on a computer network or send a copy to the Free Software
3649 Foundation. (When you do this, please license the code and its
3650 documentation under a license that permits other people to run, copy,
3651 study, modify, and redistribute the code and which protects you from
3652 having your work taken from you.) If you send a copy of your code to
3653 the Free Software Foundation, and properly protect yourself and
3654 others, it may be included in the next release of Emacs. In large
3655 part, this is how Emacs has grown over the past years, by donations.
3657 @node let, if, Permanent Installation, Writing Defuns
3658 @comment node-name, next, previous, up
3662 The @code{let} expression is a special form in Lisp that you will need
3663 to use in most function definitions.
3665 @code{let} is used to attach or bind a symbol to a value in such a way
3666 that the Lisp interpreter will not confuse the variable with a
3667 variable of the same name that is not part of the function.
3669 To understand why the @code{let} special form is necessary, consider
3670 the situation in which you own a home that you generally refer to as
3671 `the house', as in the sentence, ``The house needs painting.'' If you
3672 are visiting a friend and your host refers to `the house', he is
3673 likely to be referring to @emph{his} house, not yours, that is, to a
3676 If your friend is referring to his house and you think he is referring
3677 to your house, you may be in for some confusion. The same thing could
3678 happen in Lisp if a variable that is used inside of one function has
3679 the same name as a variable that is used inside of another function,
3680 and the two are not intended to refer to the same value. The
3681 @code{let} special form prevents this kind of confusion.
3684 * Prevent confusion::
3685 * Parts of let Expression::
3686 * Sample let Expression::
3687 * Uninitialized let Variables::
3690 @node Prevent confusion, Parts of let Expression, let, let
3692 @unnumberedsubsec @code{let} Prevents Confusion
3695 @cindex @samp{local variable} defined
3696 @cindex @samp{variable, local}, defined
3697 The @code{let} special form prevents confusion. @code{let} creates a
3698 name for a @dfn{local variable} that overshadows any use of the same
3699 name outside the @code{let} expression. This is like understanding
3700 that whenever your host refers to `the house', he means his house, not
3701 yours. (Symbols used in argument lists work the same way.
3702 @xref{defun, , The @code{defun} Special Form}.)
3704 Local variables created by a @code{let} expression retain their value
3705 @emph{only} within the @code{let} expression itself (and within
3706 expressions called within the @code{let} expression); the local
3707 variables have no effect outside the @code{let} expression.
3709 Another way to think about @code{let} is that it is like a @code{setq}
3710 that is temporary and local. The values set by @code{let} are
3711 automatically undone when the @code{let} is finished. The setting
3712 only affects expressions that are inside the bounds of the @code{let}
3713 expression. In computer science jargon, we would say ``the binding of
3714 a symbol is visible only in functions called in the @code{let} form;
3715 in Emacs Lisp, scoping is dynamic, not lexical.''
3717 @code{let} can create more than one variable at once. Also,
3718 @code{let} gives each variable it creates an initial value, either a
3719 value specified by you, or @code{nil}. (In the jargon, this is called
3720 `binding the variable to the value'.) After @code{let} has created
3721 and bound the variables, it executes the code in the body of the
3722 @code{let}, and returns the value of the last expression in the body,
3723 as the value of the whole @code{let} expression. (`Execute' is a jargon
3724 term that means to evaluate a list; it comes from the use of the word
3725 meaning `to give practical effect to' (@cite{Oxford English
3726 Dictionary}). Since you evaluate an expression to perform an action,
3727 `execute' has evolved as a synonym to `evaluate'.)
3729 @node Parts of let Expression, Sample let Expression, Prevent confusion, let
3730 @comment node-name, next, previous, up
3731 @subsection The Parts of a @code{let} Expression
3732 @cindex @code{let} expression, parts of
3733 @cindex Parts of @code{let} expression
3735 @cindex @samp{varlist} defined
3736 A @code{let} expression is a list of three parts. The first part is
3737 the symbol @code{let}. The second part is a list, called a
3738 @dfn{varlist}, each element of which is either a symbol by itself or a
3739 two-element list, the first element of which is a symbol. The third
3740 part of the @code{let} expression is the body of the @code{let}. The
3741 body usually consists of one or more lists.
3744 A template for a @code{let} expression looks like this:
3747 (let @var{varlist} @var{body}@dots{})
3751 The symbols in the varlist are the variables that are given initial
3752 values by the @code{let} special form. Symbols by themselves are given
3753 the initial value of @code{nil}; and each symbol that is the first
3754 element of a two-element list is bound to the value that is returned
3755 when the Lisp interpreter evaluates the second element.
3757 Thus, a varlist might look like this: @code{(thread (needles 3))}. In
3758 this case, in a @code{let} expression, Emacs binds the symbol
3759 @code{thread} to an initial value of @code{nil}, and binds the symbol
3760 @code{needles} to an initial value of 3.
3762 When you write a @code{let} expression, what you do is put the
3763 appropriate expressions in the slots of the @code{let} expression
3766 If the varlist is composed of two-element lists, as is often the case,
3767 the template for the @code{let} expression looks like this:
3771 (let ((@var{variable} @var{value})
3772 (@var{variable} @var{value})
3778 @node Sample let Expression, Uninitialized let Variables, Parts of let Expression, let
3779 @comment node-name, next, previous, up
3780 @subsection Sample @code{let} Expression
3781 @cindex Sample @code{let} expression
3782 @cindex @code{let} expression sample
3784 The following expression creates and gives initial values
3785 to the two variables @code{zebra} and @code{tiger}. The body of the
3786 @code{let} expression is a list which calls the @code{message} function.
3790 (let ((zebra 'stripes)
3792 (message "One kind of animal has %s and another is %s."
3797 Here, the varlist is @code{((zebra 'stripes) (tiger 'fierce))}.
3799 The two variables are @code{zebra} and @code{tiger}. Each variable is
3800 the first element of a two-element list and each value is the second
3801 element of its two-element list. In the varlist, Emacs binds the
3802 variable @code{zebra} to the value @code{stripes}@footnote{According
3803 to Jared Diamond in @cite{Guns, Germs, and Steel}, ``@dots{} zebras
3804 become impossibly dangerous as they grow older'' but the claim here is
3805 that they do not become fierce like a tiger. (1997, W. W. Norton and
3806 Co., ISBN 0-393-03894-2, page 171)}, and binds the
3807 variable @code{tiger} to the value @code{fierce}. In this example,
3808 both values are symbols preceded by a quote. The values could just as
3809 well have been another list or a string. The body of the @code{let}
3810 follows after the list holding the variables. In this example, the
3811 body is a list that uses the @code{message} function to print a string
3815 You may evaluate the example in the usual fashion, by placing the
3816 cursor after the last parenthesis and typing @kbd{C-x C-e}. When you do
3817 this, the following will appear in the echo area:
3820 "One kind of animal has stripes and another is fierce."
3823 As we have seen before, the @code{message} function prints its first
3824 argument, except for @samp{%s}. In this example, the value of the variable
3825 @code{zebra} is printed at the location of the first @samp{%s} and the
3826 value of the variable @code{tiger} is printed at the location of the
3829 @node Uninitialized let Variables, , Sample let Expression, let
3830 @comment node-name, next, previous, up
3831 @subsection Uninitialized Variables in a @code{let} Statement
3832 @cindex Uninitialized @code{let} variables
3833 @cindex @code{let} variables uninitialized
3835 If you do not bind the variables in a @code{let} statement to specific
3836 initial values, they will automatically be bound to an initial value of
3837 @code{nil}, as in the following expression:
3846 "Here are %d variables with %s, %s, and %s value."
3847 birch pine fir oak))
3852 Here, the varlist is @code{((birch 3) pine fir (oak 'some))}.
3855 If you evaluate this expression in the usual way, the following will
3856 appear in your echo area:
3859 "Here are 3 variables with nil, nil, and some value."
3863 In this example, Emacs binds the symbol @code{birch} to the number 3,
3864 binds the symbols @code{pine} and @code{fir} to @code{nil}, and binds
3865 the symbol @code{oak} to the value @code{some}.
3867 Note that in the first part of the @code{let}, the variables @code{pine}
3868 and @code{fir} stand alone as atoms that are not surrounded by
3869 parentheses; this is because they are being bound to @code{nil}, the
3870 empty list. But @code{oak} is bound to @code{some} and so is a part of
3871 the list @code{(oak 'some)}. Similarly, @code{birch} is bound to the
3872 number 3 and so is in a list with that number. (Since a number
3873 evaluates to itself, the number does not need to be quoted. Also, the
3874 number is printed in the message using a @samp{%d} rather than a
3875 @samp{%s}.) The four variables as a group are put into a list to
3876 delimit them from the body of the @code{let}.
3878 @node if, else, let, Writing Defuns
3879 @comment node-name, next, previous, up
3880 @section The @code{if} Special Form
3882 @cindex Conditional with @code{if}
3884 A third special form, in addition to @code{defun} and @code{let}, is the
3885 conditional @code{if}. This form is used to instruct the computer to
3886 make decisions. You can write function definitions without using
3887 @code{if}, but it is used often enough, and is important enough, to be
3888 included here. It is used, for example, in the code for the
3889 function @code{beginning-of-buffer}.
3891 The basic idea behind an @code{if}, is that ``@emph{if} a test is true,
3892 @emph{then} an expression is evaluated.'' If the test is not true, the
3893 expression is not evaluated. For example, you might make a decision
3894 such as, ``if it is warm and sunny, then go to the beach!''
3897 * if in more detail::
3898 * type-of-animal in detail::
3901 @node if in more detail, type-of-animal in detail, if, if
3903 @unnumberedsubsec @code{if} in more detail
3906 @cindex @samp{if-part} defined
3907 @cindex @samp{then-part} defined
3908 An @code{if} expression written in Lisp does not use the word `then';
3909 the test and the action are the second and third elements of the list
3910 whose first element is @code{if}. Nonetheless, the test part of an
3911 @code{if} expression is often called the @dfn{if-part} and the second
3912 argument is often called the @dfn{then-part}.
3914 Also, when an @code{if} expression is written, the true-or-false-test
3915 is usually written on the same line as the symbol @code{if}, but the
3916 action to carry out if the test is true, the ``then-part'', is written
3917 on the second and subsequent lines. This makes the @code{if}
3918 expression easier to read.
3922 (if @var{true-or-false-test}
3923 @var{action-to-carry-out-if-test-is-true})
3928 The true-or-false-test will be an expression that
3929 is evaluated by the Lisp interpreter.
3931 Here is an example that you can evaluate in the usual manner. The test
3932 is whether the number 5 is greater than the number 4. Since it is, the
3933 message @samp{5 is greater than 4!} will be printed.
3937 (if (> 5 4) ; @r{if-part}
3938 (message "5 is greater than 4!")) ; @r{then-part}
3943 (The function @code{>} tests whether its first argument is greater than
3944 its second argument and returns true if it is.)
3945 @findex > (greater than)
3947 Of course, in actual use, the test in an @code{if} expression will not
3948 be fixed for all time as it is by the expression @code{(> 5 4)}.
3949 Instead, at least one of the variables used in the test will be bound to
3950 a value that is not known ahead of time. (If the value were known ahead
3951 of time, we would not need to run the test!)
3953 For example, the value may be bound to an argument of a function
3954 definition. In the following function definition, the character of the
3955 animal is a value that is passed to the function. If the value bound to
3956 @code{characteristic} is @code{fierce}, then the message, @samp{It's a
3957 tiger!} will be printed; otherwise, @code{nil} will be returned.
3961 (defun type-of-animal (characteristic)
3962 "Print message in echo area depending on CHARACTERISTIC.
3963 If the CHARACTERISTIC is the symbol `fierce',
3964 then warn of a tiger."
3965 (if (equal characteristic 'fierce)
3966 (message "It's a tiger!")))
3972 If you are reading this inside of GNU Emacs, you can evaluate the
3973 function definition in the usual way to install it in Emacs, and then you
3974 can evaluate the following two expressions to see the results:
3978 (type-of-animal 'fierce)
3980 (type-of-animal 'zebra)
3985 @c Following sentences rewritten to prevent overfull hbox.
3987 When you evaluate @code{(type-of-animal 'fierce)}, you will see the
3988 following message printed in the echo area: @code{"It's a tiger!"}; and
3989 when you evaluate @code{(type-of-animal 'zebra)} you will see @code{nil}
3990 printed in the echo area.
3992 @node type-of-animal in detail, , if in more detail, if
3993 @comment node-name, next, previous, up
3994 @subsection The @code{type-of-animal} Function in Detail
3996 Let's look at the @code{type-of-animal} function in detail.
3998 The function definition for @code{type-of-animal} was written by filling
3999 the slots of two templates, one for a function definition as a whole, and
4000 a second for an @code{if} expression.
4003 The template for every function that is not interactive is:
4007 (defun @var{name-of-function} (@var{argument-list})
4008 "@var{documentation}@dots{}"
4014 The parts of the function that match this template look like this:
4018 (defun type-of-animal (characteristic)
4019 "Print message in echo area depending on CHARACTERISTIC.
4020 If the CHARACTERISTIC is the symbol `fierce',
4021 then warn of a tiger."
4022 @var{body: the} @code{if} @var{expression})
4026 The name of function is @code{type-of-animal}; it is passed the value
4027 of one argument. The argument list is followed by a multi-line
4028 documentation string. The documentation string is included in the
4029 example because it is a good habit to write documentation string for
4030 every function definition. The body of the function definition
4031 consists of the @code{if} expression.
4034 The template for an @code{if} expression looks like this:
4038 (if @var{true-or-false-test}
4039 @var{action-to-carry-out-if-the-test-returns-true})
4044 In the @code{type-of-animal} function, the code for the @code{if}
4049 (if (equal characteristic 'fierce)
4050 (message "It's a tiger!")))
4055 Here, the true-or-false-test is the expression:
4058 (equal characteristic 'fierce)
4062 In Lisp, @code{equal} is a function that determines whether its first
4063 argument is equal to its second argument. The second argument is the
4064 quoted symbol @code{'fierce} and the first argument is the value of the
4065 symbol @code{characteristic}---in other words, the argument passed to
4068 In the first exercise of @code{type-of-animal}, the argument
4069 @code{fierce} is passed to @code{type-of-animal}. Since @code{fierce}
4070 is equal to @code{fierce}, the expression, @code{(equal characteristic
4071 'fierce)}, returns a value of true. When this happens, the @code{if}
4072 evaluates the second argument or then-part of the @code{if}:
4073 @code{(message "It's tiger!")}.
4075 On the other hand, in the second exercise of @code{type-of-animal}, the
4076 argument @code{zebra} is passed to @code{type-of-animal}. @code{zebra}
4077 is not equal to @code{fierce}, so the then-part is not evaluated and
4078 @code{nil} is returned by the @code{if} expression.
4080 @node else, Truth & Falsehood, if, Writing Defuns
4081 @comment node-name, next, previous, up
4082 @section If--then--else Expressions
4085 An @code{if} expression may have an optional third argument, called
4086 the @dfn{else-part}, for the case when the true-or-false-test returns
4087 false. When this happens, the second argument or then-part of the
4088 overall @code{if} expression is @emph{not} evaluated, but the third or
4089 else-part @emph{is} evaluated. You might think of this as the cloudy
4090 day alternative for the decision ``if it is warm and sunny, then go to
4091 the beach, else read a book!''.
4093 The word ``else'' is not written in the Lisp code; the else-part of an
4094 @code{if} expression comes after the then-part. In the written Lisp, the
4095 else-part is usually written to start on a line of its own and is
4096 indented less than the then-part:
4100 (if @var{true-or-false-test}
4101 @var{action-to-carry-out-if-the-test-returns-true}
4102 @var{action-to-carry-out-if-the-test-returns-false})
4106 For example, the following @code{if} expression prints the message @samp{4
4107 is not greater than 5!} when you evaluate it in the usual way:
4111 (if (> 4 5) ; @r{if-part}
4112 (message "5 is greater than 4!") ; @r{then-part}
4113 (message "4 is not greater than 5!")) ; @r{else-part}
4118 Note that the different levels of indentation make it easy to
4119 distinguish the then-part from the else-part. (GNU Emacs has several
4120 commands that automatically indent @code{if} expressions correctly.
4121 @xref{Typing Lists, , GNU Emacs Helps You Type Lists}.)
4123 We can extend the @code{type-of-animal} function to include an
4124 else-part by simply incorporating an additional part to the @code{if}
4128 You can see the consequences of doing this if you evaluate the following
4129 version of the @code{type-of-animal} function definition to install it
4130 and then evaluate the two subsequent expressions to pass different
4131 arguments to the function.
4135 (defun type-of-animal (characteristic) ; @r{Second version.}
4136 "Print message in echo area depending on CHARACTERISTIC.
4137 If the CHARACTERISTIC is the symbol `fierce',
4138 then warn of a tiger;
4139 else say it's not fierce."
4140 (if (equal characteristic 'fierce)
4141 (message "It's a tiger!")
4142 (message "It's not fierce!")))
4149 (type-of-animal 'fierce)
4151 (type-of-animal 'zebra)
4156 @c Following sentence rewritten to prevent overfull hbox.
4158 When you evaluate @code{(type-of-animal 'fierce)}, you will see the
4159 following message printed in the echo area: @code{"It's a tiger!"}; but
4160 when you evaluate @code{(type-of-animal 'zebra)}, you will see
4161 @code{"It's not fierce!"}.
4163 (Of course, if the @var{characteristic} were @code{ferocious}, the
4164 message @code{"It's not fierce!"} would be printed; and it would be
4165 misleading! When you write code, you need to take into account the
4166 possibility that some such argument will be tested by the @code{if}
4167 and write your program accordingly.)
4169 @node Truth & Falsehood, save-excursion, else, Writing Defuns
4170 @comment node-name, next, previous, up
4171 @section Truth and Falsehood in Emacs Lisp
4172 @cindex Truth and falsehood in Emacs Lisp
4173 @cindex Falsehood and truth in Emacs Lisp
4176 There is an important aspect to the truth test in an @code{if}
4177 expression. So far, we have spoken of `true' and `false' as values of
4178 predicates as if they were new kinds of Emacs Lisp objects. In fact,
4179 `false' is just our old friend @code{nil}. Anything else---anything
4182 The expression that tests for truth is interpreted as @dfn{true}
4183 if the result of evaluating it is a value that is not @code{nil}. In
4184 other words, the result of the test is considered true if the value
4185 returned is a number such as 47, a string such as @code{"hello"}, or a
4186 symbol (other than @code{nil}) such as @code{flowers}, or a list (so
4187 long as it is not empty), or even a buffer!
4193 @node nil explained, , Truth & Falsehood, Truth & Falsehood
4195 @unnumberedsubsec An explanation of @code{nil}
4198 Before illustrating a test for truth, we need an explanation of @code{nil}.
4200 In Emacs Lisp, the symbol @code{nil} has two meanings. First, it means the
4201 empty list. Second, it means false and is the value returned when a
4202 true-or-false-test tests false. @code{nil} can be written as an empty
4203 list, @code{()}, or as @code{nil}. As far as the Lisp interpreter is
4204 concerned, @code{()} and @code{nil} are the same. Humans, however, tend
4205 to use @code{nil} for false and @code{()} for the empty list.
4207 In Emacs Lisp, any value that is not @code{nil}---is not the empty
4208 list---is considered true. This means that if an evaluation returns
4209 something that is not an empty list, an @code{if} expression will test
4210 true. For example, if a number is put in the slot for the test, it
4211 will be evaluated and will return itself, since that is what numbers
4212 do when evaluated. In this conditional, the @code{if} expression will
4213 test true. The expression tests false only when @code{nil}, an empty
4214 list, is returned by evaluating the expression.
4216 You can see this by evaluating the two expressions in the following examples.
4218 In the first example, the number 4 is evaluated as the test in the
4219 @code{if} expression and returns itself; consequently, the then-part
4220 of the expression is evaluated and returned: @samp{true} appears in
4221 the echo area. In the second example, the @code{nil} indicates false;
4222 consequently, the else-part of the expression is evaluated and
4223 returned: @samp{false} appears in the echo area.
4240 Incidentally, if some other useful value is not available for a test that
4241 returns true, then the Lisp interpreter will return the symbol @code{t}
4242 for true. For example, the expression @code{(> 5 4)} returns @code{t}
4243 when evaluated, as you can see by evaluating it in the usual way:
4251 On the other hand, this function returns @code{nil} if the test is false.
4257 @node save-excursion, Review, Truth & Falsehood, Writing Defuns
4258 @comment node-name, next, previous, up
4259 @section @code{save-excursion}
4260 @findex save-excursion
4261 @cindex Region, what it is
4262 @cindex Preserving point, mark, and buffer
4263 @cindex Point, mark, buffer preservation
4267 The @code{save-excursion} function is the fourth and final special form
4268 that we will discuss in this chapter.
4270 In Emacs Lisp programs used for editing, the @code{save-excursion}
4271 function is very common. It saves the location of point and mark,
4272 executes the body of the function, and then restores point and mark to
4273 their previous positions if their locations were changed. Its primary
4274 purpose is to keep the user from being surprised and disturbed by
4275 unexpected movement of point or mark.
4279 * Template for save-excursion::
4282 @node Point and mark, Template for save-excursion, save-excursion, save-excursion
4284 @unnumberedsubsec Point and Mark
4287 Before discussing @code{save-excursion}, however, it may be useful
4288 first to review what point and mark are in GNU Emacs. @dfn{Point} is
4289 the current location of the cursor. Wherever the cursor
4290 is, that is point. More precisely, on terminals where the cursor
4291 appears to be on top of a character, point is immediately before the
4292 character. In Emacs Lisp, point is an integer. The first character in
4293 a buffer is number one, the second is number two, and so on. The
4294 function @code{point} returns the current position of the cursor as a
4295 number. Each buffer has its own value for point.
4297 The @dfn{mark} is another position in the buffer; its value can be set
4298 with a command such as @kbd{C-@key{SPC}} (@code{set-mark-command}). If
4299 a mark has been set, you can use the command @kbd{C-x C-x}
4300 (@code{exchange-point-and-mark}) to cause the cursor to jump to the mark
4301 and set the mark to be the previous position of point. In addition, if
4302 you set another mark, the position of the previous mark is saved in the
4303 mark ring. Many mark positions can be saved this way. You can jump the
4304 cursor to a saved mark by typing @kbd{C-u C-@key{SPC}} one or more
4307 The part of the buffer between point and mark is called @dfn{the
4308 region}. Numerous commands work on the region, including
4309 @code{center-region}, @code{count-lines-region}, @code{kill-region}, and
4310 @code{print-region}.
4312 The @code{save-excursion} special form saves the locations of point and
4313 mark and restores those positions after the code within the body of the
4314 special form is evaluated by the Lisp interpreter. Thus, if point were
4315 in the beginning of a piece of text and some code moved point to the end
4316 of the buffer, the @code{save-excursion} would put point back to where
4317 it was before, after the expressions in the body of the function were
4320 In Emacs, a function frequently moves point as part of its internal
4321 workings even though a user would not expect this. For example,
4322 @code{count-lines-region} moves point. To prevent the user from being
4323 bothered by jumps that are both unexpected and (from the user's point of
4324 view) unnecessary, @code{save-excursion} is often used to keep point and
4325 mark in the location expected by the user. The use of
4326 @code{save-excursion} is good housekeeping.
4328 To make sure the house stays clean, @code{save-excursion} restores the
4329 values of point and mark even if something goes wrong in the code inside
4330 of it (or, to be more precise and to use the proper jargon, ``in case of
4331 abnormal exit''). This feature is very helpful.
4333 In addition to recording the values of point and mark,
4334 @code{save-excursion} keeps track of the current buffer, and restores
4335 it, too. This means you can write code that will change the buffer and
4336 have @code{save-excursion} switch you back to the original buffer.
4337 This is how @code{save-excursion} is used in @code{append-to-buffer}.
4338 (@xref{append-to-buffer, , The Definition of @code{append-to-buffer}}.)
4340 @node Template for save-excursion, , Point and mark, save-excursion
4341 @comment node-name, next, previous, up
4342 @subsection Template for a @code{save-excursion} Expression
4345 The template for code using @code{save-excursion} is simple:
4355 The body of the function is one or more expressions that will be
4356 evaluated in sequence by the Lisp interpreter. If there is more than
4357 one expression in the body, the value of the last one will be returned
4358 as the value of the @code{save-excursion} function. The other
4359 expressions in the body are evaluated only for their side effects; and
4360 @code{save-excursion} itself is used only for its side effect (which
4361 is restoring the positions of point and mark).
4364 In more detail, the template for a @code{save-excursion} expression
4370 @var{first-expression-in-body}
4371 @var{second-expression-in-body}
4372 @var{third-expression-in-body}
4374 @var{last-expression-in-body})
4379 An expression, of course, may be a symbol on its own or a list.
4381 In Emacs Lisp code, a @code{save-excursion} expression often occurs
4382 within the body of a @code{let} expression. It looks like this:
4392 @node Review, defun Exercises, save-excursion, Writing Defuns
4393 @comment node-name, next, previous, up
4396 In the last few chapters we have introduced a fair number of functions
4397 and special forms. Here they are described in brief, along with a few
4398 similar functions that have not been mentioned yet.
4401 @item eval-last-sexp
4402 Evaluate the last symbolic expression before the current location of
4403 point. The value is printed in the echo area unless the function is
4404 invoked with an argument; in that case, the output is printed in the
4405 current buffer. This command is normally bound to @kbd{C-x C-e}.
4408 Define function. This special form has up to five parts: the name,
4409 a template for the arguments that will be passed to the function,
4410 documentation, an optional interactive declaration, and the body of the
4414 For example, in an early version of Emacs, the function definition was
4415 as follows. (It is slightly more complex now that it seeks the first
4416 non-whitespace character rather than the first visible character.)
4420 (defun back-to-indentation ()
4421 "Move point to first visible character on line."
4423 (beginning-of-line 1)
4424 (skip-chars-forward " \t"))
4431 (defun backward-to-indentation (&optional arg)
4432 "Move backward ARG lines and position at first nonblank character."
4434 (forward-line (- (or arg 1)))
4435 (skip-chars-forward " \t"))
4437 (defun back-to-indentation ()
4438 "Move point to the first non-whitespace character on this line."
4440 (beginning-of-line 1)
4441 (skip-syntax-forward " " (line-end-position))
4442 ;; Move back over chars that have whitespace syntax but have the p flag.
4443 (backward-prefix-chars))
4447 Declare to the interpreter that the function can be used
4448 interactively. This special form may be followed by a string with one
4449 or more parts that pass the information to the arguments of the
4450 function, in sequence. These parts may also tell the interpreter to
4451 prompt for information. Parts of the string are separated by
4452 newlines, @samp{\n}.
4455 Common code characters are:
4459 The name of an existing buffer.
4462 The name of an existing file.
4465 The numeric prefix argument. (Note that this `p' is lower case.)
4468 Point and the mark, as two numeric arguments, smallest first. This
4469 is the only code letter that specifies two successive arguments
4473 @xref{Interactive Codes, , Code Characters for @samp{interactive},
4474 elisp, The GNU Emacs Lisp Reference Manual}, for a complete list of
4478 Declare that a list of variables is for use within the body of the
4479 @code{let} and give them an initial value, either @code{nil} or a
4480 specified value; then evaluate the rest of the expressions in the body
4481 of the @code{let} and return the value of the last one. Inside the
4482 body of the @code{let}, the Lisp interpreter does not see the values of
4483 the variables of the same names that are bound outside of the
4491 (let ((foo (buffer-name))
4492 (bar (buffer-size)))
4494 "This buffer is %s and has %d characters."
4499 @item save-excursion
4500 Record the values of point and mark and the current buffer before
4501 evaluating the body of this special form. Restore the values of point
4502 and mark and buffer afterward.
4509 (message "We are %d characters into this buffer."
4512 (goto-char (point-min)) (point))))
4517 Evaluate the first argument to the function; if it is true, evaluate
4518 the second argument; else evaluate the third argument, if there is one.
4520 The @code{if} special form is called a @dfn{conditional}. There are
4521 other conditionals in Emacs Lisp, but @code{if} is perhaps the most
4529 (if (= 22 emacs-major-version)
4530 (message "This is version 22 Emacs")
4531 (message "This is not version 22 Emacs"))
4540 The @code{<} function tests whether its first argument is smaller than
4541 its second argument. A corresponding function, @code{>}, tests whether
4542 the first argument is greater than the second. Likewise, @code{<=}
4543 tests whether the first argument is less than or equal to the second and
4544 @code{>=} tests whether the first argument is greater than or equal to
4545 the second. In all cases, both arguments must be numbers or markers
4546 (markers indicate positions in buffers).
4550 The @code{=} function tests whether two arguments, both numbers or
4556 Test whether two objects are the same. @code{equal} uses one meaning
4557 of the word `same' and @code{eq} uses another: @code{equal} returns
4558 true if the two objects have a similar structure and contents, such as
4559 two copies of the same book. On the other hand, @code{eq}, returns
4560 true if both arguments are actually the same object.
4569 The @code{string-lessp} function tests whether its first argument is
4570 smaller than the second argument. A shorter, alternative name for the
4571 same function (a @code{defalias}) is @code{string<}.
4573 The arguments to @code{string-lessp} must be strings or symbols; the
4574 ordering is lexicographic, so case is significant. The print names of
4575 symbols are used instead of the symbols themselves.
4577 @cindex @samp{empty string} defined
4578 An empty string, @samp{""}, a string with no characters in it, is
4579 smaller than any string of characters.
4581 @code{string-equal} provides the corresponding test for equality. Its
4582 shorter, alternative name is @code{string=}. There are no string test
4583 functions that correspond to @var{>}, @code{>=}, or @code{<=}.
4586 Print a message in the echo area. The first argument is a string that
4587 can contain @samp{%s}, @samp{%d}, or @samp{%c} to print the value of
4588 arguments that follow the string. The argument used by @samp{%s} must
4589 be a string or a symbol; the argument used by @samp{%d} must be a
4590 number. The argument used by @samp{%c} must be an @sc{ascii} code
4591 number; it will be printed as the character with that @sc{ascii} code.
4592 (Various other %-sequences have not been mentioned.)
4596 The @code{setq} function sets the value of its first argument to the
4597 value of the second argument. The first argument is automatically
4598 quoted by @code{setq}. It does the same for succeeding pairs of
4599 arguments. Another function, @code{set}, takes only two arguments and
4600 evaluates both of them before setting the value returned by its first
4601 argument to the value returned by its second argument.
4604 Without an argument, return the name of the buffer, as a string.
4606 @itemx buffer-file-name
4607 Without an argument, return the name of the file the buffer is
4610 @item current-buffer
4611 Return the buffer in which Emacs is active; it may not be
4612 the buffer that is visible on the screen.
4615 Return the most recently selected buffer (other than the buffer passed
4616 to @code{other-buffer} as an argument and other than the current
4619 @item switch-to-buffer
4620 Select a buffer for Emacs to be active in and display it in the current
4621 window so users can look at it. Usually bound to @kbd{C-x b}.
4624 Switch Emacs' attention to a buffer on which programs will run. Don't
4625 alter what the window is showing.
4628 Return the number of characters in the current buffer.
4631 Return the value of the current position of the cursor, as an
4632 integer counting the number of characters from the beginning of the
4636 Return the minimum permissible value of point in
4637 the current buffer. This is 1, unless narrowing is in effect.
4640 Return the value of the maximum permissible value of point in the
4641 current buffer. This is the end of the buffer, unless narrowing is in
4646 @node defun Exercises, , Review, Writing Defuns
4651 Write a non-interactive function that doubles the value of its
4652 argument, a number. Make that function interactive.
4655 Write a function that tests whether the current value of
4656 @code{fill-column} is greater than the argument passed to the function,
4657 and if so, prints an appropriate message.
4660 @node Buffer Walk Through, More Complex, Writing Defuns, Top
4661 @comment node-name, next, previous, up
4662 @chapter A Few Buffer--Related Functions
4664 In this chapter we study in detail several of the functions used in GNU
4665 Emacs. This is called a ``walk-through''. These functions are used as
4666 examples of Lisp code, but are not imaginary examples; with the
4667 exception of the first, simplified function definition, these functions
4668 show the actual code used in GNU Emacs. You can learn a great deal from
4669 these definitions. The functions described here are all related to
4670 buffers. Later, we will study other functions.
4674 * simplified-beginning-of-buffer::
4675 * mark-whole-buffer::
4676 * append-to-buffer::
4677 * Buffer Related Review::
4678 * Buffer Exercises::
4681 @node Finding More, simplified-beginning-of-buffer, Buffer Walk Through, Buffer Walk Through
4682 @section Finding More Information
4684 @findex describe-function, @r{introduced}
4685 @cindex Find function documentation
4686 In this walk-through, I will describe each new function as we come to
4687 it, sometimes in detail and sometimes briefly. If you are interested,
4688 you can get the full documentation of any Emacs Lisp function at any
4689 time by typing @kbd{C-h f} and then the name of the function (and then
4690 @key{RET}). Similarly, you can get the full documentation for a
4691 variable by typing @kbd{C-h v} and then the name of the variable (and
4694 @cindex Find source of function
4695 @c In version 22, tells location both of C and of Emacs Lisp
4696 Also, @code{describe-function} will tell you the location of the
4697 function definition.
4699 Put point into the name of the file that contains the function and
4700 press the @key{RET} key. In this case, @key{RET} means
4701 @code{push-button} rather than `return' or `enter'. Emacs will take
4702 you directly to the function definition.
4707 If you move point over the file name and press
4708 the @key{RET} key, which in this case means @code{help-follow} rather
4709 than `return' or `enter', Emacs will take you directly to the function
4713 More generally, if you want to see a function in its original source
4714 file, you can use the @code{find-tags} function to jump to it.
4715 @code{find-tags} works with a wide variety of languages, not just
4716 Lisp, and C, and it works with non-programming text as well. For
4717 example, @code{find-tags} will jump to the various nodes in the
4718 Texinfo source file of this document.
4719 The @code{find-tags} function depends on `tags tables' that record
4720 the locations of the functions, variables, and other items to which
4721 @code{find-tags} jumps.
4723 To use the @code{find-tags} command, type @kbd{M-.} (i.e., press the
4724 period key while holding down the @key{META} key, or else type the
4725 @key{ESC} key and then type the period key), and then, at the prompt,
4726 type in the name of the function whose source code you want to see,
4727 such as @code{mark-whole-buffer}, and then type @key{RET}. Emacs will
4728 switch buffers and display the source code for the function on your
4729 screen. To switch back to your current buffer, type @kbd{C-x b
4730 @key{RET}}. (On some keyboards, the @key{META} key is labelled
4733 @c !!! 22.1.1 tags table location in this paragraph
4734 @cindex TAGS table, specifying
4736 Depending on how the initial default values of your copy of Emacs are
4737 set, you may also need to specify the location of your `tags table',
4738 which is a file called @file{TAGS}. For example, if you are
4739 interested in Emacs sources, the tags table you will most likely want,
4740 if it has already been created for you, will be in a subdirectory of
4741 the @file{/usr/local/share/emacs/} directory; thus you would use the
4742 @code{M-x visit-tags-table} command and specify a pathname such as
4743 @file{/usr/local/share/emacs/22.1.1/lisp/TAGS}. If the tags table
4744 has not already been created, you will have to create it yourself. It
4745 will in a file such as @file{/usr/local/src/emacs/src/TAGS}.
4748 To create a @file{TAGS} file in a specific directory, switch to that
4749 directory in Emacs using @kbd{M-x cd} command, or list the directory
4750 with @kbd{C-x d} (@code{dired}). Then run the compile command, with
4751 @w{@code{etags *.el}} as the command to execute:
4754 M-x compile RET etags *.el RET
4757 For more information, see @ref{etags, , Create Your Own @file{TAGS} File}.
4759 After you become more familiar with Emacs Lisp, you will find that you will
4760 frequently use @code{find-tags} to navigate your way around source code;
4761 and you will create your own @file{TAGS} tables.
4763 @cindex Library, as term for `file'
4764 Incidentally, the files that contain Lisp code are conventionally
4765 called @dfn{libraries}. The metaphor is derived from that of a
4766 specialized library, such as a law library or an engineering library,
4767 rather than a general library. Each library, or file, contains
4768 functions that relate to a particular topic or activity, such as
4769 @file{abbrev.el} for handling abbreviations and other typing
4770 shortcuts, and @file{help.el} for on-line help. (Sometimes several
4771 libraries provide code for a single activity, as the various
4772 @file{rmail@dots{}} files provide code for reading electronic mail.)
4773 In @cite{The GNU Emacs Manual}, you will see sentences such as ``The
4774 @kbd{C-h p} command lets you search the standard Emacs Lisp libraries
4775 by topic keywords.''
4777 @node simplified-beginning-of-buffer, mark-whole-buffer, Finding More, Buffer Walk Through
4778 @comment node-name, next, previous, up
4779 @section A Simplified @code{beginning-of-buffer} Definition
4780 @findex simplified-beginning-of-buffer
4782 The @code{beginning-of-buffer} command is a good function to start with
4783 since you are likely to be familiar with it and it is easy to
4784 understand. Used as an interactive command, @code{beginning-of-buffer}
4785 moves the cursor to the beginning of the buffer, leaving the mark at the
4786 previous position. It is generally bound to @kbd{M-<}.
4788 In this section, we will discuss a shortened version of the function
4789 that shows how it is most frequently used. This shortened function
4790 works as written, but it does not contain the code for a complex option.
4791 In another section, we will describe the entire function.
4792 (@xref{beginning-of-buffer, , Complete Definition of
4793 @code{beginning-of-buffer}}.)
4795 Before looking at the code, let's consider what the function
4796 definition has to contain: it must include an expression that makes
4797 the function interactive so it can be called by typing @kbd{M-x
4798 beginning-of-buffer} or by typing a keychord such as @kbd{M-<}; it
4799 must include code to leave a mark at the original position in the
4800 buffer; and it must include code to move the cursor to the beginning
4804 Here is the complete text of the shortened version of the function:
4808 (defun simplified-beginning-of-buffer ()
4809 "Move point to the beginning of the buffer;
4810 leave mark at previous position."
4813 (goto-char (point-min)))
4817 Like all function definitions, this definition has five parts following
4818 the special form @code{defun}:
4822 The name: in this example, @code{simplified-beginning-of-buffer}.
4825 A list of the arguments: in this example, an empty list, @code{()},
4828 The documentation string.
4831 The interactive expression.
4838 In this function definition, the argument list is empty; this means that
4839 this function does not require any arguments. (When we look at the
4840 definition for the complete function, we will see that it may be passed
4841 an optional argument.)
4843 The interactive expression tells Emacs that the function is intended to
4844 be used interactively. In this example, @code{interactive} does not have
4845 an argument because @code{simplified-beginning-of-buffer} does not
4849 The body of the function consists of the two lines:
4854 (goto-char (point-min))
4858 The first of these lines is the expression, @code{(push-mark)}. When
4859 this expression is evaluated by the Lisp interpreter, it sets a mark at
4860 the current position of the cursor, wherever that may be. The position
4861 of this mark is saved in the mark ring.
4863 The next line is @code{(goto-char (point-min))}. This expression
4864 jumps the cursor to the minimum point in the buffer, that is, to the
4865 beginning of the buffer (or to the beginning of the accessible portion
4866 of the buffer if it is narrowed. @xref{Narrowing & Widening, ,
4867 Narrowing and Widening}.)
4869 The @code{push-mark} command sets a mark at the place where the cursor
4870 was located before it was moved to the beginning of the buffer by the
4871 @code{(goto-char (point-min))} expression. Consequently, you can, if
4872 you wish, go back to where you were originally by typing @kbd{C-x C-x}.
4874 That is all there is to the function definition!
4876 @findex describe-function
4877 When you are reading code such as this and come upon an unfamiliar
4878 function, such as @code{goto-char}, you can find out what it does by
4879 using the @code{describe-function} command. To use this command, type
4880 @kbd{C-h f} and then type in the name of the function and press
4881 @key{RET}. The @code{describe-function} command will print the
4882 function's documentation string in a @file{*Help*} window. For
4883 example, the documentation for @code{goto-char} is:
4887 Set point to POSITION, a number or marker.
4888 Beginning of buffer is position (point-min), end is (point-max).
4893 The function's one argument is the desired position.
4896 (The prompt for @code{describe-function} will offer you the symbol
4897 under or preceding the cursor, so you can save typing by positioning
4898 the cursor right over or after the function and then typing @kbd{C-h f
4901 The @code{end-of-buffer} function definition is written in the same way as
4902 the @code{beginning-of-buffer} definition except that the body of the
4903 function contains the expression @code{(goto-char (point-max))} in place
4904 of @code{(goto-char (point-min))}.
4906 @node mark-whole-buffer, append-to-buffer, simplified-beginning-of-buffer, Buffer Walk Through
4907 @comment node-name, next, previous, up
4908 @section The Definition of @code{mark-whole-buffer}
4909 @findex mark-whole-buffer
4911 The @code{mark-whole-buffer} function is no harder to understand than the
4912 @code{simplified-beginning-of-buffer} function. In this case, however,
4913 we will look at the complete function, not a shortened version.
4915 The @code{mark-whole-buffer} function is not as commonly used as the
4916 @code{beginning-of-buffer} function, but is useful nonetheless: it
4917 marks a whole buffer as a region by putting point at the beginning and
4918 a mark at the end of the buffer. It is generally bound to @kbd{C-x
4922 * mark-whole-buffer overview::
4923 * Body of mark-whole-buffer::
4926 @node mark-whole-buffer overview, Body of mark-whole-buffer, mark-whole-buffer, mark-whole-buffer
4928 @unnumberedsubsec An overview of @code{mark-whole-buffer}
4932 In GNU Emacs 22, the code for the complete function looks like this:
4936 (defun mark-whole-buffer ()
4937 "Put point at beginning and mark at end of buffer.
4938 You probably should not use this function in Lisp programs;
4939 it is usually a mistake for a Lisp function to use any subroutine
4940 that uses or sets the mark."
4943 (push-mark (point-max) nil t)
4944 (goto-char (point-min)))
4949 Like all other functions, the @code{mark-whole-buffer} function fits
4950 into the template for a function definition. The template looks like
4955 (defun @var{name-of-function} (@var{argument-list})
4956 "@var{documentation}@dots{}"
4957 (@var{interactive-expression}@dots{})
4962 Here is how the function works: the name of the function is
4963 @code{mark-whole-buffer}; it is followed by an empty argument list,
4964 @samp{()}, which means that the function does not require arguments.
4965 The documentation comes next.
4967 The next line is an @code{(interactive)} expression that tells Emacs
4968 that the function will be used interactively. These details are similar
4969 to the @code{simplified-beginning-of-buffer} function described in the
4973 @node Body of mark-whole-buffer, , mark-whole-buffer overview, mark-whole-buffer
4974 @comment node-name, next, previous, up
4975 @subsection Body of @code{mark-whole-buffer}
4977 The body of the @code{mark-whole-buffer} function consists of three
4984 (push-mark (point-max) nil t)
4985 (goto-char (point-min))
4989 The first of these lines is the expression, @code{(push-mark (point))}.
4991 This line does exactly the same job as the first line of the body of
4992 the @code{simplified-beginning-of-buffer} function, which is written
4993 @code{(push-mark)}. In both cases, the Lisp interpreter sets a mark
4994 at the current position of the cursor.
4996 I don't know why the expression in @code{mark-whole-buffer} is written
4997 @code{(push-mark (point))} and the expression in
4998 @code{beginning-of-buffer} is written @code{(push-mark)}. Perhaps
4999 whoever wrote the code did not know that the arguments for
5000 @code{push-mark} are optional and that if @code{push-mark} is not
5001 passed an argument, the function automatically sets mark at the
5002 location of point by default. Or perhaps the expression was written
5003 so as to parallel the structure of the next line. In any case, the
5004 line causes Emacs to determine the position of point and set a mark
5007 In earlier versions of GNU Emacs, the next line of
5008 @code{mark-whole-buffer} was @code{(push-mark (point-max))}. This
5009 expression sets a mark at the point in the buffer that has the highest
5010 number. This will be the end of the buffer (or, if the buffer is
5011 narrowed, the end of the accessible portion of the buffer.
5012 @xref{Narrowing & Widening, , Narrowing and Widening}, for more about
5013 narrowing.) After this mark has been set, the previous mark, the one
5014 set at point, is no longer set, but Emacs remembers its position, just
5015 as all other recent marks are always remembered. This means that you
5016 can, if you wish, go back to that position by typing @kbd{C-u
5020 In GNU Emacs 22, the @code{(point-max)} is slightly more complicated.
5024 (push-mark (point-max) nil t)
5028 The expression works nearly the same as before. It sets a mark at the
5029 highest numbered place in the buffer that it can. However, in this
5030 version, @code{push-mark} has two additional arguments. The second
5031 argument to @code{push-mark} is @code{nil}. This tells the function
5032 it @emph{should} display a message that says `Mark set' when it pushes
5033 the mark. The third argument is @code{t}. This tells
5034 @code{push-mark} to activate the mark when Transient Mark mode is
5035 turned on. Transient Mark mode highlights the currently active
5036 region. It is often turned off.
5038 Finally, the last line of the function is @code{(goto-char
5039 (point-min)))}. This is written exactly the same way as it is written
5040 in @code{beginning-of-buffer}. The expression moves the cursor to
5041 the minimum point in the buffer, that is, to the beginning of the buffer
5042 (or to the beginning of the accessible portion of the buffer). As a
5043 result of this, point is placed at the beginning of the buffer and mark
5044 is set at the end of the buffer. The whole buffer is, therefore, the
5047 @node append-to-buffer, Buffer Related Review, mark-whole-buffer, Buffer Walk Through
5048 @comment node-name, next, previous, up
5049 @section The Definition of @code{append-to-buffer}
5050 @findex append-to-buffer
5052 The @code{append-to-buffer} command is more complex than the
5053 @code{mark-whole-buffer} command. What it does is copy the region
5054 (that is, the part of the buffer between point and mark) from the
5055 current buffer to a specified buffer.
5058 * append-to-buffer overview::
5059 * append interactive::
5060 * append-to-buffer body::
5061 * append save-excursion::
5064 @node append-to-buffer overview, append interactive, append-to-buffer, append-to-buffer
5066 @unnumberedsubsec An Overview of @code{append-to-buffer}
5069 @findex insert-buffer-substring
5070 The @code{append-to-buffer} command uses the
5071 @code{insert-buffer-substring} function to copy the region.
5072 @code{insert-buffer-substring} is described by its name: it takes a
5073 string of characters from part of a buffer, a ``substring'', and
5074 inserts them into another buffer.
5076 Most of @code{append-to-buffer} is
5077 concerned with setting up the conditions for
5078 @code{insert-buffer-substring} to work: the code must specify both the
5079 buffer to which the text will go, the window it comes from and goes
5080 to, and the region that will be copied.
5083 Here is the complete text of the function:
5087 (defun append-to-buffer (buffer start end)
5088 "Append to specified buffer the text of the region.
5089 It is inserted into that buffer before its point.
5093 When calling from a program, give three arguments:
5094 BUFFER (or buffer name), START and END.
5095 START and END specify the portion of the current buffer to be copied."
5097 (list (read-buffer "Append to buffer: " (other-buffer
5098 (current-buffer) t))
5099 (region-beginning) (region-end)))
5102 (let ((oldbuf (current-buffer)))
5104 (let* ((append-to (get-buffer-create buffer))
5105 (windows (get-buffer-window-list append-to t t))
5107 (set-buffer append-to)
5108 (setq point (point))
5109 (barf-if-buffer-read-only)
5110 (insert-buffer-substring oldbuf start end)
5111 (dolist (window windows)
5112 (when (= (window-point window) point)
5113 (set-window-point window (point))))))))
5117 The function can be understood by looking at it as a series of
5118 filled-in templates.
5120 The outermost template is for the function definition. In this
5121 function, it looks like this (with several slots filled in):
5125 (defun append-to-buffer (buffer start end)
5126 "@var{documentation}@dots{}"
5127 (interactive @dots{})
5132 The first line of the function includes its name and three arguments.
5133 The arguments are the @code{buffer} to which the text will be copied, and
5134 the @code{start} and @code{end} of the region in the current buffer that
5137 The next part of the function is the documentation, which is clear and
5138 complete. As is conventional, the three arguments are written in
5139 upper case so you will notice them easily. Even better, they are
5140 described in the same order as in the argument list.
5142 Note that the documentation distinguishes between a buffer and its
5143 name. (The function can handle either.)
5145 @node append interactive, append-to-buffer body, append-to-buffer overview, append-to-buffer
5146 @comment node-name, next, previous, up
5147 @subsection The @code{append-to-buffer} Interactive Expression
5149 Since the @code{append-to-buffer} function will be used interactively,
5150 the function must have an @code{interactive} expression. (For a
5151 review of @code{interactive}, see @ref{Interactive, , Making a
5152 Function Interactive}.) The expression reads as follows:
5158 "Append to buffer: "
5159 (other-buffer (current-buffer) t))
5166 This expression is not one with letters standing for parts, as
5167 described earlier. Instead, it starts a list with thee parts.
5169 The first part of the list is an expression to read the name of a
5170 buffer and return it as a string. That is @code{read-buffer}. The
5171 function requires a prompt as its first argument, @samp{"Append to
5172 buffer: "}. Its second argument tells the command what value to
5173 provide if you don't specify anything.
5175 In this case that second argument is an expression containing the
5176 function @code{other-buffer}, an exception, and a @samp{t}, standing
5179 The first argument to @code{other-buffer}, the exception, is yet
5180 another function, @code{current-buffer}. That is not going to be
5181 returned. The second argument is the symbol for true, @code{t}. that
5182 tells @code{other-buffer} that it may show visible buffers (except in
5183 this case, it will not show the current buffer, which makes sense).
5186 The expression looks like this:
5189 (other-buffer (current-buffer) t)
5192 The second and third arguments to the @code{list} expression are
5193 @code{(region-beginning)} and @code{(region-end)}. These two
5194 functions specify the beginning and end of the text to be appended.
5197 Originally, the command used the letters @samp{B} and @samp{r}.
5198 The whole @code{interactive} expression looked like this:
5201 (interactive "BAppend to buffer:@: \nr")
5205 But when that was done, the default value of the buffer switched to
5206 was invisible. That was not wanted.
5208 (The prompt was separated from the second argument with a newline,
5209 @samp{\n}. It was followed by an @samp{r} that told Emacs to bind the
5210 two arguments that follow the symbol @code{buffer} in the function's
5211 argument list (that is, @code{start} and @code{end}) to the values of
5212 point and mark. That argument worked fine.)
5214 @node append-to-buffer body, append save-excursion, append interactive, append-to-buffer
5215 @comment node-name, next, previous, up
5216 @subsection The Body of @code{append-to-buffer}
5219 in GNU Emacs 22 in /usr/local/src/emacs/lisp/simple.el
5221 (defun append-to-buffer (buffer start end)
5222 "Append to specified buffer the text of the region.
5223 It is inserted into that buffer before its point.
5225 When calling from a program, give three arguments:
5226 BUFFER (or buffer name), START and END.
5227 START and END specify the portion of the current buffer to be copied."
5229 (list (read-buffer "Append to buffer: " (other-buffer (current-buffer) t))
5230 (region-beginning) (region-end)))
5231 (let ((oldbuf (current-buffer)))
5233 (let* ((append-to (get-buffer-create buffer))
5234 (windows (get-buffer-window-list append-to t t))
5236 (set-buffer append-to)
5237 (setq point (point))
5238 (barf-if-buffer-read-only)
5239 (insert-buffer-substring oldbuf start end)
5240 (dolist (window windows)
5241 (when (= (window-point window) point)
5242 (set-window-point window (point))))))))
5245 The body of the @code{append-to-buffer} function begins with @code{let}.
5247 As we have seen before (@pxref{let, , @code{let}}), the purpose of a
5248 @code{let} expression is to create and give initial values to one or
5249 more variables that will only be used within the body of the
5250 @code{let}. This means that such a variable will not be confused with
5251 any variable of the same name outside the @code{let} expression.
5253 We can see how the @code{let} expression fits into the function as a
5254 whole by showing a template for @code{append-to-buffer} with the
5255 @code{let} expression in outline:
5259 (defun append-to-buffer (buffer start end)
5260 "@var{documentation}@dots{}"
5261 (interactive @dots{})
5262 (let ((@var{variable} @var{value}))
5267 The @code{let} expression has three elements:
5271 The symbol @code{let};
5274 A varlist containing, in this case, a single two-element list,
5275 @code{(@var{variable} @var{value})};
5278 The body of the @code{let} expression.
5282 In the @code{append-to-buffer} function, the varlist looks like this:
5285 (oldbuf (current-buffer))
5289 In this part of the @code{let} expression, the one variable,
5290 @code{oldbuf}, is bound to the value returned by the
5291 @code{(current-buffer)} expression. The variable, @code{oldbuf}, is
5292 used to keep track of the buffer in which you are working and from
5293 which you will copy.
5295 The element or elements of a varlist are surrounded by a set of
5296 parentheses so the Lisp interpreter can distinguish the varlist from
5297 the body of the @code{let}. As a consequence, the two-element list
5298 within the varlist is surrounded by a circumscribing set of parentheses.
5299 The line looks like this:
5303 (let ((oldbuf (current-buffer)))
5309 The two parentheses before @code{oldbuf} might surprise you if you did
5310 not realize that the first parenthesis before @code{oldbuf} marks the
5311 boundary of the varlist and the second parenthesis marks the beginning
5312 of the two-element list, @code{(oldbuf (current-buffer))}.
5314 @node append save-excursion, , append-to-buffer body, append-to-buffer
5315 @comment node-name, next, previous, up
5316 @subsection @code{save-excursion} in @code{append-to-buffer}
5318 The body of the @code{let} expression in @code{append-to-buffer}
5319 consists of a @code{save-excursion} expression.
5321 The @code{save-excursion} function saves the locations of point and
5322 mark, and restores them to those positions after the expressions in the
5323 body of the @code{save-excursion} complete execution. In addition,
5324 @code{save-excursion} keeps track of the original buffer, and
5325 restores it. This is how @code{save-excursion} is used in
5326 @code{append-to-buffer}.
5329 @cindex Indentation for formatting
5330 @cindex Formatting convention
5331 Incidentally, it is worth noting here that a Lisp function is normally
5332 formatted so that everything that is enclosed in a multi-line spread is
5333 indented more to the right than the first symbol. In this function
5334 definition, the @code{let} is indented more than the @code{defun}, and
5335 the @code{save-excursion} is indented more than the @code{let}, like
5351 This formatting convention makes it easy to see that the lines in
5352 the body of the @code{save-excursion} are enclosed by the parentheses
5353 associated with @code{save-excursion}, just as the
5354 @code{save-excursion} itself is enclosed by the parentheses associated
5355 with the @code{let}:
5359 (let ((oldbuf (current-buffer)))
5362 (set-buffer @dots{})
5363 (insert-buffer-substring oldbuf start end)
5369 The use of the @code{save-excursion} function can be viewed as a process
5370 of filling in the slots of a template:
5375 @var{first-expression-in-body}
5376 @var{second-expression-in-body}
5378 @var{last-expression-in-body})
5384 In this function, the body of the @code{save-excursion} contains only
5385 one expression, the @code{let*} expression. You know about a
5386 @code{let} function. The @code{let*} function is different. It has a
5387 @samp{*} in its name. It enables Emacs to set each variable in its
5388 varlist in sequence, one after another.
5390 Its critical feature is that variables later in the varlist can make
5391 use of the values to which Emacs set variables earlier in the varlist.
5392 @xref{fwd-para let, , The @code{let*} expression}.
5394 We will skip functions like @code{let*} and focus on two: the
5395 @code{set-buffer} function and the @code{insert-buffer-substring}
5399 In the old days, the @code{set-buffer} expression was simply
5402 (set-buffer (get-buffer-create buffer))
5410 (set-buffer append-to)
5414 @code{append-to} is bound to @code{(get-buffer-create buffer)} earlier
5415 on in the @code{let*} expression. That extra binding would not be
5416 necessary except for that @code{append-to} is used later in the
5417 varlist as an argument to @code{get-buffer-window-list}.
5422 (let ((oldbuf (current-buffer)))
5424 (let* ((append-to (get-buffer-create buffer))
5425 (windows (get-buffer-window-list append-to t t))
5427 (set-buffer append-to)
5428 (setq point (point))
5429 (barf-if-buffer-read-only)
5430 (insert-buffer-substring oldbuf start end)
5431 (dolist (window windows)
5432 (when (= (window-point window) point)
5433 (set-window-point window (point))))))))
5436 The @code{append-to-buffer} function definition inserts text from the
5437 buffer in which you are currently to a named buffer. It happens that
5438 @code{insert-buffer-substring} copies text from another buffer to the
5439 current buffer, just the reverse---that is why the
5440 @code{append-to-buffer} definition starts out with a @code{let} that
5441 binds the local symbol @code{oldbuf} to the value returned by
5442 @code{current-buffer}.
5445 The @code{insert-buffer-substring} expression looks like this:
5448 (insert-buffer-substring oldbuf start end)
5452 The @code{insert-buffer-substring} function copies a string
5453 @emph{from} the buffer specified as its first argument and inserts the
5454 string into the present buffer. In this case, the argument to
5455 @code{insert-buffer-substring} is the value of the variable created
5456 and bound by the @code{let}, namely the value of @code{oldbuf}, which
5457 was the current buffer when you gave the @code{append-to-buffer}
5460 After @code{insert-buffer-substring} has done its work,
5461 @code{save-excursion} will restore the action to the original buffer
5462 and @code{append-to-buffer} will have done its job.
5465 Written in skeletal form, the workings of the body look like this:
5469 (let (@var{bind-}@code{oldbuf}@var{-to-value-of-}@code{current-buffer})
5470 (save-excursion ; @r{Keep track of buffer.}
5472 @var{insert-substring-from-}@code{oldbuf}@var{-into-buffer})
5474 @var{change-back-to-original-buffer-when-finished}
5475 @var{let-the-local-meaning-of-}@code{oldbuf}@var{-disappear-when-finished}
5479 In summary, @code{append-to-buffer} works as follows: it saves the
5480 value of the current buffer in the variable called @code{oldbuf}. It
5481 gets the new buffer (creating one if need be) and switches Emacs'
5482 attention to it. Using the value of @code{oldbuf}, it inserts the
5483 region of text from the old buffer into the new buffer; and then using
5484 @code{save-excursion}, it brings you back to your original buffer.
5486 In looking at @code{append-to-buffer}, you have explored a fairly
5487 complex function. It shows how to use @code{let} and
5488 @code{save-excursion}, and how to change to and come back from another
5489 buffer. Many function definitions use @code{let},
5490 @code{save-excursion}, and @code{set-buffer} this way.
5492 @node Buffer Related Review, Buffer Exercises, append-to-buffer, Buffer Walk Through
5493 @comment node-name, next, previous, up
5496 Here is a brief summary of the various functions discussed in this chapter.
5499 @item describe-function
5500 @itemx describe-variable
5501 Print the documentation for a function or variable.
5502 Conventionally bound to @kbd{C-h f} and @kbd{C-h v}.
5505 Find the file containing the source for a function or variable and
5506 switch buffers to it, positioning point at the beginning of the item.
5507 Conventionally bound to @kbd{M-.} (that's a period following the
5510 @item save-excursion
5511 Save the location of point and mark and restore their values after the
5512 arguments to @code{save-excursion} have been evaluated. Also, remember
5513 the current buffer and return to it.
5516 Set mark at a location and record the value of the previous mark on the
5517 mark ring. The mark is a location in the buffer that will keep its
5518 relative position even if text is added to or removed from the buffer.
5521 Set point to the location specified by the value of the argument, which
5522 can be a number, a marker, or an expression that returns the number of
5523 a position, such as @code{(point-min)}.
5525 @item insert-buffer-substring
5526 Copy a region of text from a buffer that is passed to the function as
5527 an argument and insert the region into the current buffer.
5529 @item mark-whole-buffer
5530 Mark the whole buffer as a region. Normally bound to @kbd{C-x h}.
5533 Switch the attention of Emacs to another buffer, but do not change the
5534 window being displayed. Used when the program rather than a human is
5535 to work on a different buffer.
5537 @item get-buffer-create
5539 Find a named buffer or create one if a buffer of that name does not
5540 exist. The @code{get-buffer} function returns @code{nil} if the named
5541 buffer does not exist.
5545 @node Buffer Exercises, , Buffer Related Review, Buffer Walk Through
5550 Write your own @code{simplified-end-of-buffer} function definition;
5551 then test it to see whether it works.
5554 Use @code{if} and @code{get-buffer} to write a function that prints a
5555 message telling you whether a buffer exists.
5558 Using @code{find-tag}, find the source for the @code{copy-to-buffer}
5562 @node More Complex, Narrowing & Widening, Buffer Walk Through, Top
5563 @comment node-name, next, previous, up
5564 @chapter A Few More Complex Functions
5566 In this chapter, we build on what we have learned in previous chapters
5567 by looking at more complex functions. The @code{copy-to-buffer}
5568 function illustrates use of two @code{save-excursion} expressions in
5569 one definition, while the @code{insert-buffer} function illustrates
5570 use of an asterisk in an @code{interactive} expression, use of
5571 @code{or}, and the important distinction between a name and the object
5572 to which the name refers.
5577 * beginning-of-buffer::
5578 * Second Buffer Related Review::
5579 * optional Exercise::
5582 @node copy-to-buffer, insert-buffer, More Complex, More Complex
5583 @comment node-name, next, previous, up
5584 @section The Definition of @code{copy-to-buffer}
5585 @findex copy-to-buffer
5587 After understanding how @code{append-to-buffer} works, it is easy to
5588 understand @code{copy-to-buffer}. This function copies text into a
5589 buffer, but instead of adding to the second buffer, it replaces all the
5590 previous text in the second buffer.
5593 The body of @code{copy-to-buffer} looks like this,
5598 (interactive "BCopy to buffer: \nr")
5599 (let ((oldbuf (current-buffer)))
5600 (with-current-buffer (get-buffer-create buffer)
5601 (barf-if-buffer-read-only)
5604 (insert-buffer-substring oldbuf start end)))))
5608 The @code{copy-to-buffer} function has a simpler @code{interactive}
5609 expression than @code{append-to-buffer}.
5612 The definition then says
5615 (with-current-buffer (get-buffer-create buffer) @dots{}
5618 First, look at the earliest inner expression; that is evaluated first.
5619 That expression starts with @code{get-buffer-create buffer}. The
5620 function tells the computer to use the buffer with the name specified
5621 as the one to which you are copying, or if such a buffer does not
5622 exist, to create it. Then, the @code{with-current-buffer} function
5623 evaluates its body with that buffer temporarily current.
5625 (This demonstrates another way to shift the computer's attention but
5626 not the user's. The @code{append-to-buffer} function showed how to do
5627 the same with @code{save-excursion} and @code{set-buffer}.
5628 @code{with-current-buffer} is a newer, and arguably easier,
5631 The @code{barf-if-buffer-read-only} function sends you an error
5632 message saying the buffer is read-only if you cannot modify it.
5634 The next line has the @code{erase-buffer} function as its sole
5635 contents. That function erases the buffer.
5637 Finally, the last two lines contain the @code{save-excursion}
5638 expression with @code{insert-buffer-substring} as its body.
5639 The @code{insert-buffer-substring} expression copies the text from
5640 the buffer you are in (and you have not seen the computer shift its
5641 attention, so you don't know that that buffer is now called
5644 Incidentally, this is what is meant by `replacement'. To replace text,
5645 Emacs erases the previous text and then inserts new text.
5648 In outline, the body of @code{copy-to-buffer} looks like this:
5652 (let (@var{bind-}@code{oldbuf}@var{-to-value-of-}@code{current-buffer})
5653 (@var{with-the-buffer-you-are-copying-to}
5654 (@var{but-do-not-erase-or-copy-to-a-read-only-buffer})
5657 @var{insert-substring-from-}@code{oldbuf}@var{-into-buffer})))
5661 @node insert-buffer, beginning-of-buffer, copy-to-buffer, More Complex
5662 @comment node-name, next, previous, up
5663 @section The Definition of @code{insert-buffer}
5664 @findex insert-buffer
5666 @code{insert-buffer} is yet another buffer-related function. This
5667 command copies another buffer @emph{into} the current buffer. It is the
5668 reverse of @code{append-to-buffer} or @code{copy-to-buffer}, since they
5669 copy a region of text @emph{from} the current buffer to another buffer.
5671 Here is a discussion based on the original code. The code was
5672 simplified in 2003 and is harder to understand.
5674 (@xref{New insert-buffer, , New Body for @code{insert-buffer}}, to see
5675 a discussion of the new body.)
5677 In addition, this code illustrates the use of @code{interactive} with a
5678 buffer that might be @dfn{read-only} and the important distinction
5679 between the name of an object and the object actually referred to.
5682 * insert-buffer code::
5683 * insert-buffer interactive::
5684 * insert-buffer body::
5688 * New insert-buffer::
5691 @node insert-buffer code, insert-buffer interactive, insert-buffer, insert-buffer
5693 @unnumberedsubsec The Code for @code{insert-buffer}
5697 Here is the earlier code:
5701 (defun insert-buffer (buffer)
5702 "Insert after point the contents of BUFFER.
5703 Puts mark after the inserted text.
5704 BUFFER may be a buffer or a buffer name."
5705 (interactive "*bInsert buffer:@: ")
5708 (or (bufferp buffer)
5709 (setq buffer (get-buffer buffer)))
5710 (let (start end newmark)
5714 (setq start (point-min) end (point-max)))
5717 (insert-buffer-substring buffer start end)
5718 (setq newmark (point)))
5719 (push-mark newmark)))
5724 As with other function definitions, you can use a template to see an
5725 outline of the function:
5729 (defun insert-buffer (buffer)
5730 "@var{documentation}@dots{}"
5731 (interactive "*bInsert buffer:@: ")
5736 @node insert-buffer interactive, insert-buffer body, insert-buffer code, insert-buffer
5737 @comment node-name, next, previous, up
5738 @subsection The Interactive Expression in @code{insert-buffer}
5739 @findex interactive, @r{example use of}
5741 In @code{insert-buffer}, the argument to the @code{interactive}
5742 declaration has two parts, an asterisk, @samp{*}, and @samp{bInsert
5746 * Read-only buffer::
5747 * b for interactive::
5750 @node Read-only buffer, b for interactive, insert-buffer interactive, insert-buffer interactive
5751 @comment node-name, next, previous, up
5752 @unnumberedsubsubsec A Read-only Buffer
5753 @cindex Read-only buffer
5754 @cindex Asterisk for read-only buffer
5755 @findex * @r{for read-only buffer}
5757 The asterisk is for the situation when the current buffer is a
5758 read-only buffer---a buffer that cannot be modified. If
5759 @code{insert-buffer} is called when the current buffer is read-only, a
5760 message to this effect is printed in the echo area and the terminal
5761 may beep or blink at you; you will not be permitted to insert anything
5762 into current buffer. The asterisk does not need to be followed by a
5763 newline to separate it from the next argument.
5765 @node b for interactive, , Read-only buffer, insert-buffer interactive
5766 @comment node-name, next, previous, up
5767 @unnumberedsubsubsec @samp{b} in an Interactive Expression
5769 The next argument in the interactive expression starts with a lower
5770 case @samp{b}. (This is different from the code for
5771 @code{append-to-buffer}, which uses an upper-case @samp{B}.
5772 @xref{append-to-buffer, , The Definition of @code{append-to-buffer}}.)
5773 The lower-case @samp{b} tells the Lisp interpreter that the argument
5774 for @code{insert-buffer} should be an existing buffer or else its
5775 name. (The upper-case @samp{B} option provides for the possibility
5776 that the buffer does not exist.) Emacs will prompt you for the name
5777 of the buffer, offering you a default buffer, with name completion
5778 enabled. If the buffer does not exist, you receive a message that
5779 says ``No match''; your terminal may beep at you as well.
5781 The new and simplified code generates a list for @code{interactive}.
5782 It uses the @code{barf-if-buffer-read-only} and @code{read-buffer}
5783 functions with which we are already familiar and the @code{progn}
5784 special form with which we are not. (It will be described later.)
5786 @node insert-buffer body, if & or, insert-buffer interactive, insert-buffer
5787 @comment node-name, next, previous, up
5788 @subsection The Body of the @code{insert-buffer} Function
5790 The body of the @code{insert-buffer} function has two major parts: an
5791 @code{or} expression and a @code{let} expression. The purpose of the
5792 @code{or} expression is to ensure that the argument @code{buffer} is
5793 bound to a buffer and not just the name of a buffer. The body of the
5794 @code{let} expression contains the code which copies the other buffer
5795 into the current buffer.
5798 In outline, the two expressions fit into the @code{insert-buffer}
5803 (defun insert-buffer (buffer)
5804 "@var{documentation}@dots{}"
5805 (interactive "*bInsert buffer:@: ")
5810 (let (@var{varlist})
5811 @var{body-of-}@code{let}@dots{} )
5815 To understand how the @code{or} expression ensures that the argument
5816 @code{buffer} is bound to a buffer and not to the name of a buffer, it
5817 is first necessary to understand the @code{or} function.
5819 Before doing this, let me rewrite this part of the function using
5820 @code{if} so that you can see what is done in a manner that will be familiar.
5822 @node if & or, Insert or, insert-buffer body, insert-buffer
5823 @comment node-name, next, previous, up
5824 @subsection @code{insert-buffer} With an @code{if} Instead of an @code{or}
5826 The job to be done is to make sure the value of @code{buffer} is a
5827 buffer itself and not the name of a buffer. If the value is the name,
5828 then the buffer itself must be got.
5830 You can imagine yourself at a conference where an usher is wandering
5831 around holding a list with your name on it and looking for you: the
5832 usher is ``bound'' to your name, not to you; but when the usher finds
5833 you and takes your arm, the usher becomes ``bound'' to you.
5836 In Lisp, you might describe this situation like this:
5840 (if (not (holding-on-to-guest))
5841 (find-and-take-arm-of-guest))
5845 We want to do the same thing with a buffer---if we do not have the
5846 buffer itself, we want to get it.
5849 Using a predicate called @code{bufferp} that tells us whether we have a
5850 buffer (rather than its name), we can write the code like this:
5854 (if (not (bufferp buffer)) ; @r{if-part}
5855 (setq buffer (get-buffer buffer))) ; @r{then-part}
5860 Here, the true-or-false-test of the @code{if} expression is
5861 @w{@code{(not (bufferp buffer))}}; and the then-part is the expression
5862 @w{@code{(setq buffer (get-buffer buffer))}}.
5864 In the test, the function @code{bufferp} returns true if its argument is
5865 a buffer---but false if its argument is the name of the buffer. (The
5866 last character of the function name @code{bufferp} is the character
5867 @samp{p}; as we saw earlier, such use of @samp{p} is a convention that
5868 indicates that the function is a predicate, which is a term that means
5869 that the function will determine whether some property is true or false.
5870 @xref{Wrong Type of Argument, , Using the Wrong Type Object as an
5874 The function @code{not} precedes the expression @code{(bufferp buffer)},
5875 so the true-or-false-test looks like this:
5878 (not (bufferp buffer))
5882 @code{not} is a function that returns true if its argument is false
5883 and false if its argument is true. So if @code{(bufferp buffer)}
5884 returns true, the @code{not} expression returns false and vice-verse:
5885 what is ``not true'' is false and what is ``not false'' is true.
5887 Using this test, the @code{if} expression works as follows: when the
5888 value of the variable @code{buffer} is actually a buffer rather than
5889 its name, the true-or-false-test returns false and the @code{if}
5890 expression does not evaluate the then-part. This is fine, since we do
5891 not need to do anything to the variable @code{buffer} if it really is
5894 On the other hand, when the value of @code{buffer} is not a buffer
5895 itself, but the name of a buffer, the true-or-false-test returns true
5896 and the then-part of the expression is evaluated. In this case, the
5897 then-part is @code{(setq buffer (get-buffer buffer))}. This
5898 expression uses the @code{get-buffer} function to return an actual
5899 buffer itself, given its name. The @code{setq} then sets the variable
5900 @code{buffer} to the value of the buffer itself, replacing its previous
5901 value (which was the name of the buffer).
5903 @node Insert or, Insert let, if & or, insert-buffer
5904 @comment node-name, next, previous, up
5905 @subsection The @code{or} in the Body
5907 The purpose of the @code{or} expression in the @code{insert-buffer}
5908 function is to ensure that the argument @code{buffer} is bound to a
5909 buffer and not just to the name of a buffer. The previous section shows
5910 how the job could have been done using an @code{if} expression.
5911 However, the @code{insert-buffer} function actually uses @code{or}.
5912 To understand this, it is necessary to understand how @code{or} works.
5915 An @code{or} function can have any number of arguments. It evaluates
5916 each argument in turn and returns the value of the first of its
5917 arguments that is not @code{nil}. Also, and this is a crucial feature
5918 of @code{or}, it does not evaluate any subsequent arguments after
5919 returning the first non-@code{nil} value.
5922 The @code{or} expression looks like this:
5926 (or (bufferp buffer)
5927 (setq buffer (get-buffer buffer)))
5932 The first argument to @code{or} is the expression @code{(bufferp buffer)}.
5933 This expression returns true (a non-@code{nil} value) if the buffer is
5934 actually a buffer, and not just the name of a buffer. In the @code{or}
5935 expression, if this is the case, the @code{or} expression returns this
5936 true value and does not evaluate the next expression---and this is fine
5937 with us, since we do not want to do anything to the value of
5938 @code{buffer} if it really is a buffer.
5940 On the other hand, if the value of @code{(bufferp buffer)} is @code{nil},
5941 which it will be if the value of @code{buffer} is the name of a buffer,
5942 the Lisp interpreter evaluates the next element of the @code{or}
5943 expression. This is the expression @code{(setq buffer (get-buffer
5944 buffer))}. This expression returns a non-@code{nil} value, which
5945 is the value to which it sets the variable @code{buffer}---and this
5946 value is a buffer itself, not the name of a buffer.
5948 The result of all this is that the symbol @code{buffer} is always
5949 bound to a buffer itself rather than to the name of a buffer. All
5950 this is necessary because the @code{set-buffer} function in a
5951 following line only works with a buffer itself, not with the name to a
5955 Incidentally, using @code{or}, the situation with the usher would be
5959 (or (holding-on-to-guest) (find-and-take-arm-of-guest))
5962 @node Insert let, New insert-buffer, Insert or, insert-buffer
5963 @comment node-name, next, previous, up
5964 @subsection The @code{let} Expression in @code{insert-buffer}
5966 After ensuring that the variable @code{buffer} refers to a buffer itself
5967 and not just to the name of a buffer, the @code{insert-buffer function}
5968 continues with a @code{let} expression. This specifies three local
5969 variables, @code{start}, @code{end}, and @code{newmark} and binds them
5970 to the initial value @code{nil}. These variables are used inside the
5971 remainder of the @code{let} and temporarily hide any other occurrence of
5972 variables of the same name in Emacs until the end of the @code{let}.
5975 The body of the @code{let} contains two @code{save-excursion}
5976 expressions. First, we will look at the inner @code{save-excursion}
5977 expression in detail. The expression looks like this:
5983 (setq start (point-min) end (point-max)))
5988 The expression @code{(set-buffer buffer)} changes Emacs' attention
5989 from the current buffer to the one from which the text will copied.
5990 In that buffer, the variables @code{start} and @code{end} are set to
5991 the beginning and end of the buffer, using the commands
5992 @code{point-min} and @code{point-max}. Note that we have here an
5993 illustration of how @code{setq} is able to set two variables in the
5994 same expression. The first argument of @code{setq} is set to the
5995 value of its second, and its third argument is set to the value of its
5998 After the body of the inner @code{save-excursion} is evaluated, the
5999 @code{save-excursion} restores the original buffer, but @code{start} and
6000 @code{end} remain set to the values of the beginning and end of the
6001 buffer from which the text will be copied.
6004 The outer @code{save-excursion} expression looks like this:
6009 (@var{inner-}@code{save-excursion}@var{-expression}
6010 (@var{go-to-new-buffer-and-set-}@code{start}@var{-and-}@code{end})
6011 (insert-buffer-substring buffer start end)
6012 (setq newmark (point)))
6017 The @code{insert-buffer-substring} function copies the text
6018 @emph{into} the current buffer @emph{from} the region indicated by
6019 @code{start} and @code{end} in @code{buffer}. Since the whole of the
6020 second buffer lies between @code{start} and @code{end}, the whole of
6021 the second buffer is copied into the buffer you are editing. Next,
6022 the value of point, which will be at the end of the inserted text, is
6023 recorded in the variable @code{newmark}.
6025 After the body of the outer @code{save-excursion} is evaluated, point
6026 and mark are relocated to their original places.
6028 However, it is convenient to locate a mark at the end of the newly
6029 inserted text and locate point at its beginning. The @code{newmark}
6030 variable records the end of the inserted text. In the last line of
6031 the @code{let} expression, the @code{(push-mark newmark)} expression
6032 function sets a mark to this location. (The previous location of the
6033 mark is still accessible; it is recorded on the mark ring and you can
6034 go back to it with @kbd{C-u C-@key{SPC}}.) Meanwhile, point is
6035 located at the beginning of the inserted text, which is where it was
6036 before you called the insert function, the position of which was saved
6037 by the first @code{save-excursion}.
6040 The whole @code{let} expression looks like this:
6044 (let (start end newmark)
6048 (setq start (point-min) end (point-max)))
6049 (insert-buffer-substring buffer start end)
6050 (setq newmark (point)))
6051 (push-mark newmark))
6055 Like the @code{append-to-buffer} function, the @code{insert-buffer}
6056 function uses @code{let}, @code{save-excursion}, and
6057 @code{set-buffer}. In addition, the function illustrates one way to
6058 use @code{or}. All these functions are building blocks that we will
6059 find and use again and again.
6061 @node New insert-buffer, , Insert let, insert-buffer
6062 @comment node-name, next, previous, up
6063 @subsection New Body for @code{insert-buffer}
6064 @findex insert-buffer, new version body
6065 @findex new version body for insert-buffer
6067 The body in the GNU Emacs 22 version is more confusing than the original.
6070 It consists of two expressions,
6076 (insert-buffer-substring (get-buffer buffer))
6084 except, and this is what confuses novices, very important work is done
6085 inside the @code{push-mark} expression.
6087 The @code{get-buffer} function returns a buffer with the name
6088 provided. You will note that the function is @emph{not} called
6089 @code{get-buffer-create}; it does not create a buffer if one does not
6090 already exist. The buffer returned by @code{get-buffer}, an existing
6091 buffer, is passed to @code{insert-buffer-substring}, which inserts the
6092 whole of the buffer (since you did not specify anything else).
6094 The location into which the buffer is inserted is recorded by
6095 @code{push-mark}. Then the function returns @code{nil}, the value of
6096 its last command. Put another way, the @code{insert-buffer} function
6097 exists only to produce a side effect, inserting another buffer, not to
6100 @node beginning-of-buffer, Second Buffer Related Review, insert-buffer, More Complex
6101 @comment node-name, next, previous, up
6102 @section Complete Definition of @code{beginning-of-buffer}
6103 @findex beginning-of-buffer
6105 The basic structure of the @code{beginning-of-buffer} function has
6106 already been discussed. (@xref{simplified-beginning-of-buffer, , A
6107 Simplified @code{beginning-of-buffer} Definition}.)
6108 This section describes the complex part of the definition.
6110 As previously described, when invoked without an argument,
6111 @code{beginning-of-buffer} moves the cursor to the beginning of the
6112 buffer (in truth, the beginning of the accessible portion of the
6113 buffer), leaving the mark at the previous position. However, when the
6114 command is invoked with a number between one and ten, the function
6115 considers that number to be a fraction of the length of the buffer,
6116 measured in tenths, and Emacs moves the cursor that fraction of the
6117 way from the beginning of the buffer. Thus, you can either call this
6118 function with the key command @kbd{M-<}, which will move the cursor to
6119 the beginning of the buffer, or with a key command such as @kbd{C-u 7
6120 M-<} which will move the cursor to a point 70% of the way through the
6121 buffer. If a number bigger than ten is used for the argument, it
6122 moves to the end of the buffer.
6124 The @code{beginning-of-buffer} function can be called with or without an
6125 argument. The use of the argument is optional.
6128 * Optional Arguments::
6129 * beginning-of-buffer opt arg::
6130 * beginning-of-buffer complete::
6133 @node Optional Arguments, beginning-of-buffer opt arg, beginning-of-buffer, beginning-of-buffer
6134 @subsection Optional Arguments
6136 Unless told otherwise, Lisp expects that a function with an argument in
6137 its function definition will be called with a value for that argument.
6138 If that does not happen, you get an error and a message that says
6139 @samp{Wrong number of arguments}.
6141 @cindex Optional arguments
6144 However, optional arguments are a feature of Lisp: a particular
6145 @dfn{keyword} is used to tell the Lisp interpreter that an argument is
6146 optional. The keyword is @code{&optional}. (The @samp{&} in front of
6147 @samp{optional} is part of the keyword.) In a function definition, if
6148 an argument follows the keyword @code{&optional}, no value need be
6149 passed to that argument when the function is called.
6152 The first line of the function definition of @code{beginning-of-buffer}
6153 therefore looks like this:
6156 (defun beginning-of-buffer (&optional arg)
6160 In outline, the whole function looks like this:
6164 (defun beginning-of-buffer (&optional arg)
6165 "@var{documentation}@dots{}"
6167 (or (@var{is-the-argument-a-cons-cell} arg)
6168 (and @var{are-both-transient-mark-mode-and-mark-active-true})
6170 (let (@var{determine-size-and-set-it})
6172 (@var{if-there-is-an-argument}
6173 @var{figure-out-where-to-go}
6180 The function is similar to the @code{simplified-beginning-of-buffer}
6181 function except that the @code{interactive} expression has @code{"P"}
6182 as an argument and the @code{goto-char} function is followed by an
6183 if-then-else expression that figures out where to put the cursor if
6184 there is an argument that is not a cons cell.
6186 (Since I do not explain a cons cell for many more chapters, please
6187 consider ignoring the function @code{consp}. @xref{List
6188 Implementation, , How Lists are Implemented}, and @ref{Cons Cell Type,
6189 , Cons Cell and List Types, elisp, The GNU Emacs Lisp Reference
6192 The @code{"P"} in the @code{interactive} expression tells Emacs to
6193 pass a prefix argument, if there is one, to the function in raw form.
6194 A prefix argument is made by typing the @key{META} key followed by a
6195 number, or by typing @kbd{C-u} and then a number. (If you don't type
6196 a number, @kbd{C-u} defaults to a cons cell with a 4. A lowercase
6197 @code{"p"} in the @code{interactive} expression causes the function to
6198 convert a prefix arg to a number.)
6200 The true-or-false-test of the @code{if} expression looks complex, but
6201 it is not: it checks whether @code{arg} has a value that is not
6202 @code{nil} and whether it is a cons cell. (That is what @code{consp}
6203 does; it checks whether its argument is a cons cell.) If @code{arg}
6204 has a value that is not @code{nil} (and is not a cons cell), which
6205 will be the case if @code{beginning-of-buffer} is called with a
6206 numeric argument, then this true-or-false-test will return true and
6207 the then-part of the @code{if} expression will be evaluated. On the
6208 other hand, if @code{beginning-of-buffer} is not called with an
6209 argument, the value of @code{arg} will be @code{nil} and the else-part
6210 of the @code{if} expression will be evaluated. The else-part is
6211 simply @code{point-min}, and when this is the outcome, the whole
6212 @code{goto-char} expression is @code{(goto-char (point-min))}, which
6213 is how we saw the @code{beginning-of-buffer} function in its
6216 @node beginning-of-buffer opt arg, beginning-of-buffer complete, Optional Arguments, beginning-of-buffer
6217 @subsection @code{beginning-of-buffer} with an Argument
6219 When @code{beginning-of-buffer} is called with an argument, an
6220 expression is evaluated which calculates what value to pass to
6221 @code{goto-char}. This expression is rather complicated at first sight.
6222 It includes an inner @code{if} expression and much arithmetic. It looks
6227 (if (> (buffer-size) 10000)
6228 ;; @r{Avoid overflow for large buffer sizes!}
6229 (* (prefix-numeric-value arg)
6234 size (prefix-numeric-value arg))) 10)))
6239 * Disentangle beginning-of-buffer::
6240 * Large buffer case::
6241 * Small buffer case::
6244 @node Disentangle beginning-of-buffer, Large buffer case, beginning-of-buffer opt arg, beginning-of-buffer opt arg
6246 @unnumberedsubsubsec Disentangle @code{beginning-of-buffer}
6249 Like other complex-looking expressions, the conditional expression
6250 within @code{beginning-of-buffer} can be disentangled by looking at it
6251 as parts of a template, in this case, the template for an if-then-else
6252 expression. In skeletal form, the expression looks like this:
6256 (if (@var{buffer-is-large}
6257 @var{divide-buffer-size-by-10-and-multiply-by-arg}
6258 @var{else-use-alternate-calculation}
6262 The true-or-false-test of this inner @code{if} expression checks the
6263 size of the buffer. The reason for this is that the old version 18
6264 Emacs used numbers that are no bigger than eight million or so and in
6265 the computation that followed, the programmer feared that Emacs might
6266 try to use over-large numbers if the buffer were large. The term
6267 `overflow', mentioned in the comment, means numbers that are over
6268 large. More recent versions of Emacs use larger numbers, but this
6269 code has not been touched, if only because people now look at buffers
6270 that are far, far larger than ever before.
6272 There are two cases: if the buffer is large and if it is not.
6274 @node Large buffer case, Small buffer case, Disentangle beginning-of-buffer, beginning-of-buffer opt arg
6275 @comment node-name, next, previous, up
6276 @unnumberedsubsubsec What happens in a large buffer
6278 In @code{beginning-of-buffer}, the inner @code{if} expression tests
6279 whether the size of the buffer is greater than 10,000 characters. To do
6280 this, it uses the @code{>} function and the computation of @code{size}
6281 that comes from the let expression.
6283 In the old days, the function @code{buffer-size} was used. Not only
6284 was that function called several times, it gave the size of the whole
6285 buffer, not the accessible part. The computation makes much more
6286 sense when it handles just the accessible part. (@xref{Narrowing &
6287 Widening, , Narrowing and Widening}, for more information on focusing
6288 attention to an `accessible' part.)
6291 The line looks like this:
6299 When the buffer is large, the then-part of the @code{if} expression is
6300 evaluated. It reads like this (after formatting for easy reading):
6305 (prefix-numeric-value arg)
6311 This expression is a multiplication, with two arguments to the function
6314 The first argument is @code{(prefix-numeric-value arg)}. When
6315 @code{"P"} is used as the argument for @code{interactive}, the value
6316 passed to the function as its argument is passed a ``raw prefix
6317 argument'', and not a number. (It is a number in a list.) To perform
6318 the arithmetic, a conversion is necessary, and
6319 @code{prefix-numeric-value} does the job.
6321 @findex / @r{(division)}
6323 The second argument is @code{(/ size 10)}. This expression divides
6324 the numeric value by ten --- the numeric value of the size of the
6325 accessible portion of the buffer. This produces a number that tells
6326 how many characters make up one tenth of the buffer size. (In Lisp,
6327 @code{/} is used for division, just as @code{*} is used for
6331 In the multiplication expression as a whole, this amount is multiplied
6332 by the value of the prefix argument---the multiplication looks like this:
6336 (* @var{numeric-value-of-prefix-arg}
6337 @var{number-of-characters-in-one-tenth-of-the-accessible-buffer})
6342 If, for example, the prefix argument is @samp{7}, the one-tenth value
6343 will be multiplied by 7 to give a position 70% of the way through.
6346 The result of all this is that if the accessible portion of the buffer
6347 is large, the @code{goto-char} expression reads like this:
6351 (goto-char (* (prefix-numeric-value arg)
6356 This puts the cursor where we want it.
6358 @node Small buffer case, , Large buffer case, beginning-of-buffer opt arg
6359 @comment node-name, next, previous, up
6360 @unnumberedsubsubsec What happens in a small buffer
6362 If the buffer contains fewer than 10,000 characters, a slightly
6363 different computation is performed. You might think this is not
6364 necessary, since the first computation could do the job. However, in
6365 a small buffer, the first method may not put the cursor on exactly the
6366 desired line; the second method does a better job.
6369 The code looks like this:
6371 @c Keep this on one line.
6373 (/ (+ 10 (* size (prefix-numeric-value arg))) 10))
6378 This is code in which you figure out what happens by discovering how the
6379 functions are embedded in parentheses. It is easier to read if you
6380 reformat it with each expression indented more deeply than its
6381 enclosing expression:
6389 (prefix-numeric-value arg)))
6396 Looking at parentheses, we see that the innermost operation is
6397 @code{(prefix-numeric-value arg)}, which converts the raw argument to
6398 a number. In the following expression, this number is multiplied by
6399 the size of the accessible portion of the buffer:
6402 (* size (prefix-numeric-value arg))
6406 This multiplication creates a number that may be larger than the size of
6407 the buffer---seven times larger if the argument is 7, for example. Ten
6408 is then added to this number and finally the large number is divided by
6409 ten to provide a value that is one character larger than the percentage
6410 position in the buffer.
6412 The number that results from all this is passed to @code{goto-char} and
6413 the cursor is moved to that point.
6416 @node beginning-of-buffer complete, , beginning-of-buffer opt arg, beginning-of-buffer
6417 @comment node-name, next, previous, up
6418 @subsection The Complete @code{beginning-of-buffer}
6421 Here is the complete text of the @code{beginning-of-buffer} function:
6427 (defun beginning-of-buffer (&optional arg)
6428 "Move point to the beginning of the buffer;
6429 leave mark at previous position.
6430 With \\[universal-argument] prefix,
6431 do not set mark at previous position.
6433 put point N/10 of the way from the beginning.
6435 If the buffer is narrowed,
6436 this command uses the beginning and size
6437 of the accessible part of the buffer.
6441 Don't use this command in Lisp programs!
6442 \(goto-char (point-min)) is faster
6443 and avoids clobbering the mark."
6446 (and transient-mark-mode mark-active)
6450 (let ((size (- (point-max) (point-min))))
6451 (goto-char (if (and arg (not (consp arg)))
6454 ;; Avoid overflow for large buffer sizes!
6455 (* (prefix-numeric-value arg)
6457 (/ (+ 10 (* size (prefix-numeric-value arg))) 10)))
6459 (if arg (forward-line 1)))
6464 From before GNU Emacs 22
6467 (defun beginning-of-buffer (&optional arg)
6468 "Move point to the beginning of the buffer;
6469 leave mark at previous position.
6470 With arg N, put point N/10 of the way
6471 from the true beginning.
6474 Don't use this in Lisp programs!
6475 \(goto-char (point-min)) is faster
6476 and does not set the mark."
6483 (if (> (buffer-size) 10000)
6484 ;; @r{Avoid overflow for large buffer sizes!}
6485 (* (prefix-numeric-value arg)
6486 (/ (buffer-size) 10))
6489 (/ (+ 10 (* (buffer-size)
6490 (prefix-numeric-value arg)))
6493 (if arg (forward-line 1)))
6499 Except for two small points, the previous discussion shows how this
6500 function works. The first point deals with a detail in the
6501 documentation string, and the second point concerns the last line of
6505 In the documentation string, there is reference to an expression:
6508 \\[universal-argument]
6512 A @samp{\\} is used before the first square bracket of this
6513 expression. This @samp{\\} tells the Lisp interpreter to substitute
6514 whatever key is currently bound to the @samp{[@dots{}]}. In the case
6515 of @code{universal-argument}, that is usually @kbd{C-u}, but it might
6516 be different. (@xref{Documentation Tips, , Tips for Documentation
6517 Strings, elisp, The GNU Emacs Lisp Reference Manual}, for more
6521 Finally, the last line of the @code{beginning-of-buffer} command says
6522 to move point to the beginning of the next line if the command is
6523 invoked with an argument:
6526 (if arg (forward-line 1)))
6530 This puts the cursor at the beginning of the first line after the
6531 appropriate tenths position in the buffer. This is a flourish that
6532 means that the cursor is always located @emph{at least} the requested
6533 tenths of the way through the buffer, which is a nicety that is,
6534 perhaps, not necessary, but which, if it did not occur, would be sure
6537 On the other hand, it also means that if you specify the command with
6538 a @kbd{C-u}, but without a number, that is to say, if the `raw prefix
6539 argument' is simply a cons cell, then the command puts you at the
6540 beginning of the second line @dots{} I don't know whether this is
6541 intended or whether no one has dealt with the code to avoid this
6544 @node Second Buffer Related Review, optional Exercise, beginning-of-buffer, More Complex
6545 @comment node-name, next, previous, up
6548 Here is a brief summary of some of the topics covered in this chapter.
6552 Evaluate each argument in sequence, and return the value of the first
6553 argument that is not @code{nil}; if none return a value that is not
6554 @code{nil}, return @code{nil}. In brief, return the first true value
6555 of the arguments; return a true value if one @emph{or} any of the
6559 Evaluate each argument in sequence, and if any are @code{nil}, return
6560 @code{nil}; if none are @code{nil}, return the value of the last
6561 argument. In brief, return a true value only if all the arguments are
6562 true; return a true value if one @emph{and} each of the others is
6566 A keyword used to indicate that an argument to a function definition
6567 is optional; this means that the function can be evaluated without the
6568 argument, if desired.
6570 @item prefix-numeric-value
6571 Convert the `raw prefix argument' produced by @code{(interactive
6572 "P")} to a numeric value.
6575 Move point forward to the beginning of the next line, or if the argument
6576 is greater than one, forward that many lines. If it can't move as far
6577 forward as it is supposed to, @code{forward-line} goes forward as far as
6578 it can and then returns a count of the number of additional lines it was
6579 supposed to move but couldn't.
6582 Delete the entire contents of the current buffer.
6585 Return @code{t} if its argument is a buffer; otherwise return @code{nil}.
6588 @node optional Exercise, , Second Buffer Related Review, More Complex
6589 @section @code{optional} Argument Exercise
6591 Write an interactive function with an optional argument that tests
6592 whether its argument, a number, is greater than or equal to, or else,
6593 less than the value of @code{fill-column}, and tells you which, in a
6594 message. However, if you do not pass an argument to the function, use
6595 56 as a default value.
6597 @node Narrowing & Widening, car cdr & cons, More Complex, Top
6598 @comment node-name, next, previous, up
6599 @chapter Narrowing and Widening
6600 @cindex Focusing attention (narrowing)
6604 Narrowing is a feature of Emacs that makes it possible for you to focus
6605 on a specific part of a buffer, and work without accidentally changing
6606 other parts. Narrowing is normally disabled since it can confuse
6610 * Narrowing advantages::
6611 * save-restriction::
6616 @node Narrowing advantages, save-restriction, Narrowing & Widening, Narrowing & Widening
6618 @unnumberedsec The Advantages of Narrowing
6621 With narrowing, the rest of a buffer is made invisible, as if it weren't
6622 there. This is an advantage if, for example, you want to replace a word
6623 in one part of a buffer but not in another: you narrow to the part you want
6624 and the replacement is carried out only in that section, not in the rest
6625 of the buffer. Searches will only work within a narrowed region, not
6626 outside of one, so if you are fixing a part of a document, you can keep
6627 yourself from accidentally finding parts you do not need to fix by
6628 narrowing just to the region you want.
6629 (The key binding for @code{narrow-to-region} is @kbd{C-x n n}.)
6631 However, narrowing does make the rest of the buffer invisible, which
6632 can scare people who inadvertently invoke narrowing and think they
6633 have deleted a part of their file. Moreover, the @code{undo} command
6634 (which is usually bound to @kbd{C-x u}) does not turn off narrowing
6635 (nor should it), so people can become quite desperate if they do not
6636 know that they can return the rest of a buffer to visibility with the
6637 @code{widen} command.
6638 (The key binding for @code{widen} is @kbd{C-x n w}.)
6640 Narrowing is just as useful to the Lisp interpreter as to a human.
6641 Often, an Emacs Lisp function is designed to work on just part of a
6642 buffer; or conversely, an Emacs Lisp function needs to work on all of a
6643 buffer that has been narrowed. The @code{what-line} function, for
6644 example, removes the narrowing from a buffer, if it has any narrowing
6645 and when it has finished its job, restores the narrowing to what it was.
6646 On the other hand, the @code{count-lines} function, which is called by
6647 @code{what-line}, uses narrowing to restrict itself to just that portion
6648 of the buffer in which it is interested and then restores the previous
6651 @node save-restriction, what-line, Narrowing advantages, Narrowing & Widening
6652 @comment node-name, next, previous, up
6653 @section The @code{save-restriction} Special Form
6654 @findex save-restriction
6656 In Emacs Lisp, you can use the @code{save-restriction} special form to
6657 keep track of whatever narrowing is in effect, if any. When the Lisp
6658 interpreter meets with @code{save-restriction}, it executes the code
6659 in the body of the @code{save-restriction} expression, and then undoes
6660 any changes to narrowing that the code caused. If, for example, the
6661 buffer is narrowed and the code that follows @code{save-restriction}
6662 gets rid of the narrowing, @code{save-restriction} returns the buffer
6663 to its narrowed region afterwards. In the @code{what-line} command,
6664 any narrowing the buffer may have is undone by the @code{widen}
6665 command that immediately follows the @code{save-restriction} command.
6666 Any original narrowing is restored just before the completion of the
6670 The template for a @code{save-restriction} expression is simple:
6680 The body of the @code{save-restriction} is one or more expressions that
6681 will be evaluated in sequence by the Lisp interpreter.
6683 Finally, a point to note: when you use both @code{save-excursion} and
6684 @code{save-restriction}, one right after the other, you should use
6685 @code{save-excursion} outermost. If you write them in reverse order,
6686 you may fail to record narrowing in the buffer to which Emacs switches
6687 after calling @code{save-excursion}. Thus, when written together,
6688 @code{save-excursion} and @code{save-restriction} should be written
6699 In other circumstances, when not written together, the
6700 @code{save-excursion} and @code{save-restriction} special forms must
6701 be written in the order appropriate to the function.
6717 /usr/local/src/emacs/lisp/simple.el
6720 "Print the current buffer line number and narrowed line number of point."
6722 (let ((start (point-min))
6723 (n (line-number-at-pos)))
6725 (message "Line %d" n)
6729 (message "line %d (narrowed line %d)"
6730 (+ n (line-number-at-pos start) -1) n))))))
6732 (defun line-number-at-pos (&optional pos)
6733 "Return (narrowed) buffer line number at position POS.
6734 If POS is nil, use current buffer location.
6735 Counting starts at (point-min), so the value refers
6736 to the contents of the accessible portion of the buffer."
6737 (let ((opoint (or pos (point))) start)
6739 (goto-char (point-min))
6740 (setq start (point))
6743 (1+ (count-lines start (point))))))
6745 (defun count-lines (start end)
6746 "Return number of lines between START and END.
6747 This is usually the number of newlines between them,
6748 but can be one more if START is not equal to END
6749 and the greater of them is not at the start of a line."
6752 (narrow-to-region start end)
6753 (goto-char (point-min))
6754 (if (eq selective-display t)
6757 (while (re-search-forward "[\n\C-m]" nil t 40)
6758 (setq done (+ 40 done)))
6759 (while (re-search-forward "[\n\C-m]" nil t 1)
6760 (setq done (+ 1 done)))
6761 (goto-char (point-max))
6762 (if (and (/= start end)
6766 (- (buffer-size) (forward-line (buffer-size)))))))
6769 @node what-line, narrow Exercise, save-restriction, Narrowing & Widening
6770 @comment node-name, next, previous, up
6771 @section @code{what-line}
6773 @cindex Widening, example of
6775 The @code{what-line} command tells you the number of the line in which
6776 the cursor is located. The function illustrates the use of the
6777 @code{save-restriction} and @code{save-excursion} commands. Here is the
6778 original text of the function:
6783 "Print the current line number (in the buffer) of point."
6790 (1+ (count-lines 1 (point)))))))
6794 (In recent versions of GNU Emacs, the @code{what-line} function has
6795 been expanded to tell you your line number in a narrowed buffer as
6796 well as your line number in a widened buffer. The recent version is
6797 more complex than the version shown here. If you feel adventurous,
6798 you might want to look at it after figuring out how this version
6799 works. You will probably need to use @kbd{C-h f}
6800 (@code{describe-function}). The newer version uses a conditional to
6801 determine whether the buffer has been narrowed.
6803 (Also, it uses @code{line-number-at-pos}, which among other simple
6804 expressions, such as @code{(goto-char (point-min))}, moves point to
6805 the beginning of the current line with @code{(forward-line 0)} rather
6806 than @code{beginning-of-line}.)
6808 The @code{what-line} function as shown here has a documentation line
6809 and is interactive, as you would expect. The next two lines use the
6810 functions @code{save-restriction} and @code{widen}.
6812 The @code{save-restriction} special form notes whatever narrowing is in
6813 effect, if any, in the current buffer and restores that narrowing after
6814 the code in the body of the @code{save-restriction} has been evaluated.
6816 The @code{save-restriction} special form is followed by @code{widen}.
6817 This function undoes any narrowing the current buffer may have had
6818 when @code{what-line} was called. (The narrowing that was there is
6819 the narrowing that @code{save-restriction} remembers.) This widening
6820 makes it possible for the line counting commands to count from the
6821 beginning of the buffer. Otherwise, they would have been limited to
6822 counting within the accessible region. Any original narrowing is
6823 restored just before the completion of the function by the
6824 @code{save-restriction} special form.
6826 The call to @code{widen} is followed by @code{save-excursion}, which
6827 saves the location of the cursor (i.e., of point) and of the mark, and
6828 restores them after the code in the body of the @code{save-excursion}
6829 uses the @code{beginning-of-line} function to move point.
6831 (Note that the @code{(widen)} expression comes between the
6832 @code{save-restriction} and @code{save-excursion} special forms. When
6833 you write the two @code{save- @dots{}} expressions in sequence, write
6834 @code{save-excursion} outermost.)
6837 The last two lines of the @code{what-line} function are functions to
6838 count the number of lines in the buffer and then print the number in the
6844 (1+ (count-lines 1 (point)))))))
6848 The @code{message} function prints a one-line message at the bottom of
6849 the Emacs screen. The first argument is inside of quotation marks and
6850 is printed as a string of characters. However, it may contain a
6851 @samp{%d} expression to print a following argument. @samp{%d} prints
6852 the argument as a decimal, so the message will say something such as
6856 The number that is printed in place of the @samp{%d} is computed by the
6857 last line of the function:
6860 (1+ (count-lines 1 (point)))
6866 (defun count-lines (start end)
6867 "Return number of lines between START and END.
6868 This is usually the number of newlines between them,
6869 but can be one more if START is not equal to END
6870 and the greater of them is not at the start of a line."
6873 (narrow-to-region start end)
6874 (goto-char (point-min))
6875 (if (eq selective-display t)
6878 (while (re-search-forward "[\n\C-m]" nil t 40)
6879 (setq done (+ 40 done)))
6880 (while (re-search-forward "[\n\C-m]" nil t 1)
6881 (setq done (+ 1 done)))
6882 (goto-char (point-max))
6883 (if (and (/= start end)
6887 (- (buffer-size) (forward-line (buffer-size)))))))
6891 What this does is count the lines from the first position of the
6892 buffer, indicated by the @code{1}, up to @code{(point)}, and then add
6893 one to that number. (The @code{1+} function adds one to its
6894 argument.) We add one to it because line 2 has only one line before
6895 it, and @code{count-lines} counts only the lines @emph{before} the
6898 After @code{count-lines} has done its job, and the message has been
6899 printed in the echo area, the @code{save-excursion} restores point and
6900 mark to their original positions; and @code{save-restriction} restores
6901 the original narrowing, if any.
6903 @node narrow Exercise, , what-line, Narrowing & Widening
6904 @section Exercise with Narrowing
6906 Write a function that will display the first 60 characters of the
6907 current buffer, even if you have narrowed the buffer to its latter
6908 half so that the first line is inaccessible. Restore point, mark, and
6909 narrowing. For this exercise, you need to use a whole potpourri of
6910 functions, including @code{save-restriction}, @code{widen},
6911 @code{goto-char}, @code{point-min}, @code{message}, and
6912 @code{buffer-substring}.
6914 @cindex Properties, mention of @code{buffer-substring-no-properties}
6915 (@code{buffer-substring} is a previously unmentioned function you will
6916 have to investigate yourself; or perhaps you will have to use
6917 @code{buffer-substring-no-properties} or
6918 @code{filter-buffer-substring} @dots{}, yet other functions. Text
6919 properties are a feature otherwise not discussed here. @xref{Text
6920 Properties, , Text Properties, elisp, The GNU Emacs Lisp Reference
6923 Additionally, do you really need @code{goto-char} or @code{point-min}?
6924 Or can you write the function without them?
6926 @node car cdr & cons, Cutting & Storing Text, Narrowing & Widening, Top
6927 @comment node-name, next, previous, up
6928 @chapter @code{car}, @code{cdr}, @code{cons}: Fundamental Functions
6929 @findex car, @r{introduced}
6930 @findex cdr, @r{introduced}
6932 In Lisp, @code{car}, @code{cdr}, and @code{cons} are fundamental
6933 functions. The @code{cons} function is used to construct lists, and
6934 the @code{car} and @code{cdr} functions are used to take them apart.
6936 In the walk through of the @code{copy-region-as-kill} function, we
6937 will see @code{cons} as well as two variants on @code{cdr},
6938 namely, @code{setcdr} and @code{nthcdr}. (@xref{copy-region-as-kill}.)
6951 @node Strange Names, car & cdr, car cdr & cons, car cdr & cons
6953 @unnumberedsec Strange Names
6956 The name of the @code{cons} function is not unreasonable: it is an
6957 abbreviation of the word `construct'. The origins of the names for
6958 @code{car} and @code{cdr}, on the other hand, are esoteric: @code{car}
6959 is an acronym from the phrase `Contents of the Address part of the
6960 Register'; and @code{cdr} (pronounced `could-er') is an acronym from
6961 the phrase `Contents of the Decrement part of the Register'. These
6962 phrases refer to specific pieces of hardware on the very early
6963 computer on which the original Lisp was developed. Besides being
6964 obsolete, the phrases have been completely irrelevant for more than 25
6965 years to anyone thinking about Lisp. Nonetheless, although a few
6966 brave scholars have begun to use more reasonable names for these
6967 functions, the old terms are still in use. In particular, since the
6968 terms are used in the Emacs Lisp source code, we will use them in this
6971 @node car & cdr, cons, Strange Names, car cdr & cons
6972 @comment node-name, next, previous, up
6973 @section @code{car} and @code{cdr}
6975 The @sc{car} of a list is, quite simply, the first item in the list.
6976 Thus the @sc{car} of the list @code{(rose violet daisy buttercup)} is
6980 If you are reading this in Info in GNU Emacs, you can see this by
6981 evaluating the following:
6984 (car '(rose violet daisy buttercup))
6988 After evaluating the expression, @code{rose} will appear in the echo
6991 Clearly, a more reasonable name for the @code{car} function would be
6992 @code{first} and this is often suggested.
6994 @code{car} does not remove the first item from the list; it only reports
6995 what it is. After @code{car} has been applied to a list, the list is
6996 still the same as it was. In the jargon, @code{car} is
6997 `non-destructive'. This feature turns out to be important.
6999 The @sc{cdr} of a list is the rest of the list, that is, the
7000 @code{cdr} function returns the part of the list that follows the
7001 first item. Thus, while the @sc{car} of the list @code{'(rose violet
7002 daisy buttercup)} is @code{rose}, the rest of the list, the value
7003 returned by the @code{cdr} function, is @code{(violet daisy
7007 You can see this by evaluating the following in the usual way:
7010 (cdr '(rose violet daisy buttercup))
7014 When you evaluate this, @code{(violet daisy buttercup)} will appear in
7017 Like @code{car}, @code{cdr} does not remove any elements from the
7018 list---it just returns a report of what the second and subsequent
7021 Incidentally, in the example, the list of flowers is quoted. If it were
7022 not, the Lisp interpreter would try to evaluate the list by calling
7023 @code{rose} as a function. In this example, we do not want to do that.
7025 Clearly, a more reasonable name for @code{cdr} would be @code{rest}.
7027 (There is a lesson here: when you name new functions, consider very
7028 carefully what you are doing, since you may be stuck with the names
7029 for far longer than you expect. The reason this document perpetuates
7030 these names is that the Emacs Lisp source code uses them, and if I did
7031 not use them, you would have a hard time reading the code; but do,
7032 please, try to avoid using these terms yourself. The people who come
7033 after you will be grateful to you.)
7035 When @code{car} and @code{cdr} are applied to a list made up of symbols,
7036 such as the list @code{(pine fir oak maple)}, the element of the list
7037 returned by the function @code{car} is the symbol @code{pine} without
7038 any parentheses around it. @code{pine} is the first element in the
7039 list. However, the @sc{cdr} of the list is a list itself, @code{(fir
7040 oak maple)}, as you can see by evaluating the following expressions in
7045 (car '(pine fir oak maple))
7047 (cdr '(pine fir oak maple))
7051 On the other hand, in a list of lists, the first element is itself a
7052 list. @code{car} returns this first element as a list. For example,
7053 the following list contains three sub-lists, a list of carnivores, a
7054 list of herbivores and a list of sea mammals:
7058 (car '((lion tiger cheetah)
7059 (gazelle antelope zebra)
7060 (whale dolphin seal)))
7065 In this example, the first element or @sc{car} of the list is the list of
7066 carnivores, @code{(lion tiger cheetah)}, and the rest of the list is
7067 @code{((gazelle antelope zebra) (whale dolphin seal))}.
7071 (cdr '((lion tiger cheetah)
7072 (gazelle antelope zebra)
7073 (whale dolphin seal)))
7077 It is worth saying again that @code{car} and @code{cdr} are
7078 non-destructive---that is, they do not modify or change lists to which
7079 they are applied. This is very important for how they are used.
7081 Also, in the first chapter, in the discussion about atoms, I said that
7082 in Lisp, ``certain kinds of atom, such as an array, can be separated
7083 into parts; but the mechanism for doing this is different from the
7084 mechanism for splitting a list. As far as Lisp is concerned, the
7085 atoms of a list are unsplittable.'' (@xref{Lisp Atoms}.) The
7086 @code{car} and @code{cdr} functions are used for splitting lists and
7087 are considered fundamental to Lisp. Since they cannot split or gain
7088 access to the parts of an array, an array is considered an atom.
7089 Conversely, the other fundamental function, @code{cons}, can put
7090 together or construct a list, but not an array. (Arrays are handled
7091 by array-specific functions. @xref{Arrays, , Arrays, elisp, The GNU
7092 Emacs Lisp Reference Manual}.)
7094 @node cons, nthcdr, car & cdr, car cdr & cons
7095 @comment node-name, next, previous, up
7096 @section @code{cons}
7097 @findex cons, @r{introduced}
7099 The @code{cons} function constructs lists; it is the inverse of
7100 @code{car} and @code{cdr}. For example, @code{cons} can be used to make
7101 a four element list from the three element list, @code{(fir oak maple)}:
7104 (cons 'pine '(fir oak maple))
7109 After evaluating this list, you will see
7112 (pine fir oak maple)
7116 appear in the echo area. @code{cons} causes the creation of a new
7117 list in which the element is followed by the elements of the original
7120 We often say that `@code{cons} puts a new element at the beginning of
7121 a list; it attaches or pushes elements onto the list', but this
7122 phrasing can be misleading, since @code{cons} does not change an
7123 existing list, but creates a new one.
7125 Like @code{car} and @code{cdr}, @code{cons} is non-destructive.
7132 @node Build a list, length, cons, cons
7134 @unnumberedsubsec Build a list
7137 @code{cons} must have a list to attach to.@footnote{Actually, you can
7138 @code{cons} an element to an atom to produce a dotted pair. Dotted
7139 pairs are not discussed here; see @ref{Dotted Pair Notation, , Dotted
7140 Pair Notation, elisp, The GNU Emacs Lisp Reference Manual}.} You
7141 cannot start from absolutely nothing. If you are building a list, you
7142 need to provide at least an empty list at the beginning. Here is a
7143 series of @code{cons} expressions that build up a list of flowers. If
7144 you are reading this in Info in GNU Emacs, you can evaluate each of
7145 the expressions in the usual way; the value is printed in this text
7146 after @samp{@result{}}, which you may read as `evaluates to'.
7150 (cons 'buttercup ())
7151 @result{} (buttercup)
7155 (cons 'daisy '(buttercup))
7156 @result{} (daisy buttercup)
7160 (cons 'violet '(daisy buttercup))
7161 @result{} (violet daisy buttercup)
7165 (cons 'rose '(violet daisy buttercup))
7166 @result{} (rose violet daisy buttercup)
7171 In the first example, the empty list is shown as @code{()} and a list
7172 made up of @code{buttercup} followed by the empty list is constructed.
7173 As you can see, the empty list is not shown in the list that was
7174 constructed. All that you see is @code{(buttercup)}. The empty list is
7175 not counted as an element of a list because there is nothing in an empty
7176 list. Generally speaking, an empty list is invisible.
7178 The second example, @code{(cons 'daisy '(buttercup))} constructs a new,
7179 two element list by putting @code{daisy} in front of @code{buttercup};
7180 and the third example constructs a three element list by putting
7181 @code{violet} in front of @code{daisy} and @code{buttercup}.
7183 @node length, , Build a list, cons
7184 @comment node-name, next, previous, up
7185 @subsection Find the Length of a List: @code{length}
7188 You can find out how many elements there are in a list by using the Lisp
7189 function @code{length}, as in the following examples:
7193 (length '(buttercup))
7198 (length '(daisy buttercup))
7203 (length (cons 'violet '(daisy buttercup)))
7209 In the third example, the @code{cons} function is used to construct a
7210 three element list which is then passed to the @code{length} function as
7214 We can also use @code{length} to count the number of elements in an
7225 As you would expect, the number of elements in an empty list is zero.
7227 An interesting experiment is to find out what happens if you try to find
7228 the length of no list at all; that is, if you try to call @code{length}
7229 without giving it an argument, not even an empty list:
7237 What you see, if you evaluate this, is the error message
7240 Lisp error: (wrong-number-of-arguments length 0)
7244 This means that the function receives the wrong number of
7245 arguments, zero, when it expects some other number of arguments. In
7246 this case, one argument is expected, the argument being a list whose
7247 length the function is measuring. (Note that @emph{one} list is
7248 @emph{one} argument, even if the list has many elements inside it.)
7250 The part of the error message that says @samp{length} is the name of
7254 @code{length} is still a subroutine, but you need C-h f to discover that.
7256 In an earlier version:
7257 This is written with a special notation, @samp{#<subr},
7258 that indicates that the function @code{length} is one of the primitive
7259 functions written in C rather than in Emacs Lisp. (@samp{subr} is an
7260 abbreviation for `subroutine'.) @xref{What Is a Function, , What Is a
7261 Function?, elisp , The GNU Emacs Lisp Reference Manual}, for more
7265 @node nthcdr, nth, cons, car cdr & cons
7266 @comment node-name, next, previous, up
7267 @section @code{nthcdr}
7270 The @code{nthcdr} function is associated with the @code{cdr} function.
7271 What it does is take the @sc{cdr} of a list repeatedly.
7273 If you take the @sc{cdr} of the list @code{(pine fir
7274 oak maple)}, you will be returned the list @code{(fir oak maple)}. If you
7275 repeat this on what was returned, you will be returned the list
7276 @code{(oak maple)}. (Of course, repeated @sc{cdr}ing on the original
7277 list will just give you the original @sc{cdr} since the function does
7278 not change the list. You need to evaluate the @sc{cdr} of the
7279 @sc{cdr} and so on.) If you continue this, eventually you will be
7280 returned an empty list, which in this case, instead of being shown as
7281 @code{()} is shown as @code{nil}.
7284 For review, here is a series of repeated @sc{cdr}s, the text following
7285 the @samp{@result{}} shows what is returned.
7289 (cdr '(pine fir oak maple))
7290 @result{}(fir oak maple)
7294 (cdr '(fir oak maple))
7295 @result{} (oak maple)
7320 You can also do several @sc{cdr}s without printing the values in
7325 (cdr (cdr '(pine fir oak maple)))
7326 @result{} (oak maple)
7331 In this example, the Lisp interpreter evaluates the innermost list first.
7332 The innermost list is quoted, so it just passes the list as it is to the
7333 innermost @code{cdr}. This @code{cdr} passes a list made up of the
7334 second and subsequent elements of the list to the outermost @code{cdr},
7335 which produces a list composed of the third and subsequent elements of
7336 the original list. In this example, the @code{cdr} function is repeated
7337 and returns a list that consists of the original list without its
7340 The @code{nthcdr} function does the same as repeating the call to
7341 @code{cdr}. In the following example, the argument 2 is passed to the
7342 function @code{nthcdr}, along with the list, and the value returned is
7343 the list without its first two items, which is exactly the same
7344 as repeating @code{cdr} twice on the list:
7348 (nthcdr 2 '(pine fir oak maple))
7349 @result{} (oak maple)
7354 Using the original four element list, we can see what happens when
7355 various numeric arguments are passed to @code{nthcdr}, including 0, 1,
7360 ;; @r{Leave the list as it was.}
7361 (nthcdr 0 '(pine fir oak maple))
7362 @result{} (pine fir oak maple)
7366 ;; @r{Return a copy without the first element.}
7367 (nthcdr 1 '(pine fir oak maple))
7368 @result{} (fir oak maple)
7372 ;; @r{Return a copy of the list without three elements.}
7373 (nthcdr 3 '(pine fir oak maple))
7378 ;; @r{Return a copy lacking all four elements.}
7379 (nthcdr 4 '(pine fir oak maple))
7384 ;; @r{Return a copy lacking all elements.}
7385 (nthcdr 5 '(pine fir oak maple))
7390 @node nth, setcar, nthcdr, car cdr & cons
7391 @comment node-name, next, previous, up
7395 The @code{nthcdr} function takes the @sc{cdr} of a list repeatedly.
7396 The @code{nth} function takes the @sc{car} of the result returned by
7397 @code{nthcdr}. It returns the Nth element of the list.
7400 Thus, if it were not defined in C for speed, the definition of
7401 @code{nth} would be:
7406 "Returns the Nth element of LIST.
7407 N counts from zero. If LIST is not that long, nil is returned."
7408 (car (nthcdr n list)))
7413 (Originally, @code{nth} was defined in Emacs Lisp in @file{subr.el},
7414 but its definition was redone in C in the 1980s.)
7416 The @code{nth} function returns a single element of a list.
7417 This can be very convenient.
7419 Note that the elements are numbered from zero, not one. That is to
7420 say, the first element of a list, its @sc{car} is the zeroth element.
7421 This is called `zero-based' counting and often bothers people who
7422 are accustomed to the first element in a list being number one, which
7430 (nth 0 '("one" "two" "three"))
7433 (nth 1 '("one" "two" "three"))
7438 It is worth mentioning that @code{nth}, like @code{nthcdr} and
7439 @code{cdr}, does not change the original list---the function is
7440 non-destructive. This is in sharp contrast to the @code{setcar} and
7441 @code{setcdr} functions.
7443 @node setcar, setcdr, nth, car cdr & cons
7444 @comment node-name, next, previous, up
7445 @section @code{setcar}
7448 As you might guess from their names, the @code{setcar} and @code{setcdr}
7449 functions set the @sc{car} or the @sc{cdr} of a list to a new value.
7450 They actually change the original list, unlike @code{car} and @code{cdr}
7451 which leave the original list as it was. One way to find out how this
7452 works is to experiment. We will start with the @code{setcar} function.
7455 First, we can make a list and then set the value of a variable to the
7456 list, using the @code{setq} function. Here is a list of animals:
7459 (setq animals '(antelope giraffe lion tiger))
7463 If you are reading this in Info inside of GNU Emacs, you can evaluate
7464 this expression in the usual fashion, by positioning the cursor after
7465 the expression and typing @kbd{C-x C-e}. (I'm doing this right here
7466 as I write this. This is one of the advantages of having the
7467 interpreter built into the computing environment. Incidentally, when
7468 there is nothing on the line after the final parentheses, such as a
7469 comment, point can be on the next line. Thus, if your cursor is in
7470 the first column of the next line, you do not need to move it.
7471 Indeed, Emacs permits any amount of white space after the final
7475 When we evaluate the variable @code{animals}, we see that it is bound to
7476 the list @code{(antelope giraffe lion tiger)}:
7481 @result{} (antelope giraffe lion tiger)
7486 Put another way, the variable @code{animals} points to the list
7487 @code{(antelope giraffe lion tiger)}.
7489 Next, evaluate the function @code{setcar} while passing it two
7490 arguments, the variable @code{animals} and the quoted symbol
7491 @code{hippopotamus}; this is done by writing the three element list
7492 @code{(setcar animals 'hippopotamus)} and then evaluating it in the
7496 (setcar animals 'hippopotamus)
7501 After evaluating this expression, evaluate the variable @code{animals}
7502 again. You will see that the list of animals has changed:
7507 @result{} (hippopotamus giraffe lion tiger)
7512 The first element on the list, @code{antelope} is replaced by
7513 @code{hippopotamus}.
7515 So we can see that @code{setcar} did not add a new element to the list
7516 as @code{cons} would have; it replaced @code{antelope} with
7517 @code{hippopotamus}; it @emph{changed} the list.
7519 @node setcdr, cons Exercise, setcar, car cdr & cons
7520 @comment node-name, next, previous, up
7521 @section @code{setcdr}
7524 The @code{setcdr} function is similar to the @code{setcar} function,
7525 except that the function replaces the second and subsequent elements of
7526 a list rather than the first element.
7528 (To see how to change the last element of a list, look ahead to
7529 @ref{kill-new function, , The @code{kill-new} function}, which uses
7530 the @code{nthcdr} and @code{setcdr} functions.)
7533 To see how this works, set the value of the variable to a list of
7534 domesticated animals by evaluating the following expression:
7537 (setq domesticated-animals '(horse cow sheep goat))
7542 If you now evaluate the list, you will be returned the list
7543 @code{(horse cow sheep goat)}:
7547 domesticated-animals
7548 @result{} (horse cow sheep goat)
7553 Next, evaluate @code{setcdr} with two arguments, the name of the
7554 variable which has a list as its value, and the list to which the
7555 @sc{cdr} of the first list will be set;
7558 (setcdr domesticated-animals '(cat dog))
7562 If you evaluate this expression, the list @code{(cat dog)} will appear
7563 in the echo area. This is the value returned by the function. The
7564 result we are interested in is the ``side effect'', which we can see by
7565 evaluating the variable @code{domesticated-animals}:
7569 domesticated-animals
7570 @result{} (horse cat dog)
7575 Indeed, the list is changed from @code{(horse cow sheep goat)} to
7576 @code{(horse cat dog)}. The @sc{cdr} of the list is changed from
7577 @code{(cow sheep goat)} to @code{(cat dog)}.
7579 @node cons Exercise, , setcdr, car cdr & cons
7582 Construct a list of four birds by evaluating several expressions with
7583 @code{cons}. Find out what happens when you @code{cons} a list onto
7584 itself. Replace the first element of the list of four birds with a
7585 fish. Replace the rest of that list with a list of other fish.
7587 @node Cutting & Storing Text, List Implementation, car cdr & cons, Top
7588 @comment node-name, next, previous, up
7589 @chapter Cutting and Storing Text
7590 @cindex Cutting and storing text
7591 @cindex Storing and cutting text
7592 @cindex Killing text
7593 @cindex Clipping text
7594 @cindex Erasing text
7595 @cindex Deleting text
7597 Whenever you cut or clip text out of a buffer with a `kill' command in
7598 GNU Emacs, it is stored in a list and you can bring it back with a
7601 (The use of the word `kill' in Emacs for processes which specifically
7602 @emph{do not} destroy the values of the entities is an unfortunate
7603 historical accident. A much more appropriate word would be `clip' since
7604 that is what the kill commands do; they clip text out of a buffer and
7605 put it into storage from which it can be brought back. I have often
7606 been tempted to replace globally all occurrences of `kill' in the Emacs
7607 sources with `clip' and all occurrences of `killed' with `clipped'.)
7613 * copy-region-as-kill::
7614 * Digression into C::
7616 * cons & search-fwd Review::
7617 * search Exercises::
7620 @node Storing Text, zap-to-char, Cutting & Storing Text, Cutting & Storing Text
7622 @unnumberedsec Storing Text in a List
7625 When text is cut out of a buffer, it is stored on a list. Successive
7626 pieces of text are stored on the list successively, so the list might
7630 ("a piece of text" "previous piece")
7635 The function @code{cons} can be used to create a new list from a piece
7636 of text (an `atom', to use the jargon) and an existing list, like
7641 (cons "another piece"
7642 '("a piece of text" "previous piece"))
7648 If you evaluate this expression, a list of three elements will appear in
7652 ("another piece" "a piece of text" "previous piece")
7655 With the @code{car} and @code{nthcdr} functions, you can retrieve
7656 whichever piece of text you want. For example, in the following code,
7657 @code{nthcdr 1 @dots{}} returns the list with the first item removed;
7658 and the @code{car} returns the first element of that remainder---the
7659 second element of the original list:
7663 (car (nthcdr 1 '("another piece"
7666 @result{} "a piece of text"
7670 The actual functions in Emacs are more complex than this, of course.
7671 The code for cutting and retrieving text has to be written so that
7672 Emacs can figure out which element in the list you want---the first,
7673 second, third, or whatever. In addition, when you get to the end of
7674 the list, Emacs should give you the first element of the list, rather
7675 than nothing at all.
7677 The list that holds the pieces of text is called the @dfn{kill ring}.
7678 This chapter leads up to a description of the kill ring and how it is
7679 used by first tracing how the @code{zap-to-char} function works. This
7680 function uses (or `calls') a function that invokes a function that
7681 manipulates the kill ring. Thus, before reaching the mountains, we
7682 climb the foothills.
7684 A subsequent chapter describes how text that is cut from the buffer is
7685 retrieved. @xref{Yanking, , Yanking Text Back}.
7687 @node zap-to-char, kill-region, Storing Text, Cutting & Storing Text
7688 @comment node-name, next, previous, up
7689 @section @code{zap-to-char}
7692 The @code{zap-to-char} function changed little between GNU Emacs
7693 version 19 and GNU Emacs version 22. However, @code{zap-to-char}
7694 calls another function, @code{kill-region}, which enjoyed a major
7697 The @code{kill-region} function in Emacs 19 is complex, but does not
7698 use code that is important at this time. We will skip it.
7700 The @code{kill-region} function in Emacs 22 is easier to read than the
7701 same function in Emacs 19 and introduces a very important concept,
7702 that of error handling. We will walk through the function.
7704 But first, let us look at the interactive @code{zap-to-char} function.
7707 * Complete zap-to-char::
7708 * zap-to-char interactive::
7709 * zap-to-char body::
7712 * Summing up zap-to-char::
7715 @node Complete zap-to-char, zap-to-char interactive, zap-to-char, zap-to-char
7717 @unnumberedsubsec The Complete @code{zap-to-char} Implementation
7720 The @code{zap-to-char} function removes the text in the region between
7721 the location of the cursor (i.e., of point) up to and including the
7722 next occurrence of a specified character. The text that
7723 @code{zap-to-char} removes is put in the kill ring; and it can be
7724 retrieved from the kill ring by typing @kbd{C-y} (@code{yank}). If
7725 the command is given an argument, it removes text through that number
7726 of occurrences. Thus, if the cursor were at the beginning of this
7727 sentence and the character were @samp{s}, @samp{Thus} would be
7728 removed. If the argument were two, @samp{Thus, if the curs} would be
7729 removed, up to and including the @samp{s} in @samp{cursor}.
7731 If the specified character is not found, @code{zap-to-char} will say
7732 ``Search failed'', tell you the character you typed, and not remove
7735 In order to determine how much text to remove, @code{zap-to-char} uses
7736 a search function. Searches are used extensively in code that
7737 manipulates text, and we will focus attention on them as well as on the
7741 @c GNU Emacs version 19
7742 (defun zap-to-char (arg char) ; version 19 implementation
7743 "Kill up to and including ARG'th occurrence of CHAR.
7744 Goes backward if ARG is negative; error if CHAR not found."
7745 (interactive "*p\ncZap to char: ")
7746 (kill-region (point)
7749 (char-to-string char) nil nil arg)
7754 Here is the complete text of the version 22 implementation of the function:
7759 (defun zap-to-char (arg char)
7760 "Kill up to and including ARG'th occurrence of CHAR.
7761 Case is ignored if `case-fold-search' is non-nil in the current buffer.
7762 Goes backward if ARG is negative; error if CHAR not found."
7763 (interactive "p\ncZap to char: ")
7764 (if (char-table-p translation-table-for-input)
7765 (setq char (or (aref translation-table-for-input char) char)))
7766 (kill-region (point) (progn
7767 (search-forward (char-to-string char) nil nil arg)
7772 The documentation is thorough. You do need to know the jargon meaning
7775 @node zap-to-char interactive, zap-to-char body, Complete zap-to-char, zap-to-char
7776 @comment node-name, next, previous, up
7777 @subsection The @code{interactive} Expression
7780 The interactive expression in the @code{zap-to-char} command looks like
7784 (interactive "p\ncZap to char: ")
7787 The part within quotation marks, @code{"p\ncZap to char:@: "}, specifies
7788 two different things. First, and most simply, is the @samp{p}.
7789 This part is separated from the next part by a newline, @samp{\n}.
7790 The @samp{p} means that the first argument to the function will be
7791 passed the value of a `processed prefix'. The prefix argument is
7792 passed by typing @kbd{C-u} and a number, or @kbd{M-} and a number. If
7793 the function is called interactively without a prefix, 1 is passed to
7796 The second part of @code{"p\ncZap to char:@: "} is
7797 @samp{cZap to char:@: }. In this part, the lower case @samp{c}
7798 indicates that @code{interactive} expects a prompt and that the
7799 argument will be a character. The prompt follows the @samp{c} and is
7800 the string @samp{Zap to char:@: } (with a space after the colon to
7803 What all this does is prepare the arguments to @code{zap-to-char} so they
7804 are of the right type, and give the user a prompt.
7806 In a read-only buffer, the @code{zap-to-char} function copies the text
7807 to the kill ring, but does not remove it. The echo area displays a
7808 message saying that the buffer is read-only. Also, the terminal may
7809 beep or blink at you.
7811 @node zap-to-char body, search-forward, zap-to-char interactive, zap-to-char
7812 @comment node-name, next, previous, up
7813 @subsection The Body of @code{zap-to-char}
7815 The body of the @code{zap-to-char} function contains the code that
7816 kills (that is, removes) the text in the region from the current
7817 position of the cursor up to and including the specified character.
7819 The first part of the code looks like this:
7822 (if (char-table-p translation-table-for-input)
7823 (setq char (or (aref translation-table-for-input char) char)))
7824 (kill-region (point) (progn
7825 (search-forward (char-to-string char) nil nil arg)
7830 @code{char-table-p} is an hitherto unseen function. It determines
7831 whether its argument is a character table. When it is, it sets the
7832 character passed to @code{zap-to-char} to one of them, if that
7833 character exists, or to the character itself. (This becomes important
7834 for certain characters in non-European languages. The @code{aref}
7835 function extracts an element from an array. It is an array-specific
7836 function that is not described in this document. @xref{Arrays, ,
7837 Arrays, elisp, The GNU Emacs Lisp Reference Manual}.)
7840 @code{(point)} is the current position of the cursor.
7842 The next part of the code is an expression using @code{progn}. The body
7843 of the @code{progn} consists of calls to @code{search-forward} and
7846 It is easier to understand how @code{progn} works after learning about
7847 @code{search-forward}, so we will look at @code{search-forward} and
7848 then at @code{progn}.
7850 @node search-forward, progn, zap-to-char body, zap-to-char
7851 @comment node-name, next, previous, up
7852 @subsection The @code{search-forward} Function
7853 @findex search-forward
7855 The @code{search-forward} function is used to locate the
7856 zapped-for-character in @code{zap-to-char}. If the search is
7857 successful, @code{search-forward} leaves point immediately after the
7858 last character in the target string. (In @code{zap-to-char}, the
7859 target string is just one character long. @code{zap-to-char} uses the
7860 function @code{char-to-string} to ensure that the computer treats that
7861 character as a string.) If the search is backwards,
7862 @code{search-forward} leaves point just before the first character in
7863 the target. Also, @code{search-forward} returns @code{t} for true.
7864 (Moving point is therefore a `side effect'.)
7867 In @code{zap-to-char}, the @code{search-forward} function looks like this:
7870 (search-forward (char-to-string char) nil nil arg)
7873 The @code{search-forward} function takes four arguments:
7877 The first argument is the target, what is searched for. This must be a
7878 string, such as @samp{"z"}.
7880 As it happens, the argument passed to @code{zap-to-char} is a single
7881 character. Because of the way computers are built, the Lisp
7882 interpreter may treat a single character as being different from a
7883 string of characters. Inside the computer, a single character has a
7884 different electronic format than a string of one character. (A single
7885 character can often be recorded in the computer using exactly one
7886 byte; but a string may be longer, and the computer needs to be ready
7887 for this.) Since the @code{search-forward} function searches for a
7888 string, the character that the @code{zap-to-char} function receives as
7889 its argument must be converted inside the computer from one format to
7890 the other; otherwise the @code{search-forward} function will fail.
7891 The @code{char-to-string} function is used to make this conversion.
7894 The second argument bounds the search; it is specified as a position in
7895 the buffer. In this case, the search can go to the end of the buffer,
7896 so no bound is set and the second argument is @code{nil}.
7899 The third argument tells the function what it should do if the search
7900 fails---it can signal an error (and print a message) or it can return
7901 @code{nil}. A @code{nil} as the third argument causes the function to
7902 signal an error when the search fails.
7905 The fourth argument to @code{search-forward} is the repeat count---how
7906 many occurrences of the string to look for. This argument is optional
7907 and if the function is called without a repeat count, this argument is
7908 passed the value 1. If this argument is negative, the search goes
7913 In template form, a @code{search-forward} expression looks like this:
7917 (search-forward "@var{target-string}"
7918 @var{limit-of-search}
7919 @var{what-to-do-if-search-fails}
7924 We will look at @code{progn} next.
7926 @node progn, Summing up zap-to-char, search-forward, zap-to-char
7927 @comment node-name, next, previous, up
7928 @subsection The @code{progn} Special Form
7931 @code{progn} is a special form that causes each of its arguments to be
7932 evaluated in sequence and then returns the value of the last one. The
7933 preceding expressions are evaluated only for the side effects they
7934 perform. The values produced by them are discarded.
7937 The template for a @code{progn} expression is very simple:
7946 In @code{zap-to-char}, the @code{progn} expression has to do two things:
7947 put point in exactly the right position; and return the location of
7948 point so that @code{kill-region} will know how far to kill to.
7950 The first argument to the @code{progn} is @code{search-forward}. When
7951 @code{search-forward} finds the string, the function leaves point
7952 immediately after the last character in the target string. (In this
7953 case the target string is just one character long.) If the search is
7954 backwards, @code{search-forward} leaves point just before the first
7955 character in the target. The movement of point is a side effect.
7957 The second and last argument to @code{progn} is the expression
7958 @code{(point)}. This expression returns the value of point, which in
7959 this case will be the location to which it has been moved by
7960 @code{search-forward}. (In the source, a line that tells the function
7961 to go to the previous character, if it is going forward, was commented
7962 out in 1999; I don't remember whether that feature or mis-feature was
7963 ever a part of the distributed source.) The value of @code{point} is
7964 returned by the @code{progn} expression and is passed to
7965 @code{kill-region} as @code{kill-region}'s second argument.
7967 @node Summing up zap-to-char, , progn, zap-to-char
7968 @comment node-name, next, previous, up
7969 @subsection Summing up @code{zap-to-char}
7971 Now that we have seen how @code{search-forward} and @code{progn} work,
7972 we can see how the @code{zap-to-char} function works as a whole.
7974 The first argument to @code{kill-region} is the position of the cursor
7975 when the @code{zap-to-char} command is given---the value of point at
7976 that time. Within the @code{progn}, the search function then moves
7977 point to just after the zapped-to-character and @code{point} returns the
7978 value of this location. The @code{kill-region} function puts together
7979 these two values of point, the first one as the beginning of the region
7980 and the second one as the end of the region, and removes the region.
7982 The @code{progn} special form is necessary because the
7983 @code{kill-region} command takes two arguments; and it would fail if
7984 @code{search-forward} and @code{point} expressions were written in
7985 sequence as two additional arguments. The @code{progn} expression is
7986 a single argument to @code{kill-region} and returns the one value that
7987 @code{kill-region} needs for its second argument.
7989 @node kill-region, copy-region-as-kill, zap-to-char, Cutting & Storing Text
7990 @comment node-name, next, previous, up
7991 @section @code{kill-region}
7994 The @code{zap-to-char} function uses the @code{kill-region} function.
7995 This function clips text from a region and copies that text to
7996 the kill ring, from which it may be retrieved.
8001 (defun kill-region (beg end &optional yank-handler)
8002 "Kill (\"cut\") text between point and mark.
8003 This deletes the text from the buffer and saves it in the kill ring.
8004 The command \\[yank] can retrieve it from there.
8005 \(If you want to kill and then yank immediately, use \\[kill-ring-save].)
8007 If you want to append the killed region to the last killed text,
8008 use \\[append-next-kill] before \\[kill-region].
8010 If the buffer is read-only, Emacs will beep and refrain from deleting
8011 the text, but put the text in the kill ring anyway. This means that
8012 you can use the killing commands to copy text from a read-only buffer.
8014 This is the primitive for programs to kill text (as opposed to deleting it).
8015 Supply two arguments, character positions indicating the stretch of text
8017 Any command that calls this function is a \"kill command\".
8018 If the previous command was also a kill command,
8019 the text killed this time appends to the text killed last time
8020 to make one entry in the kill ring.
8022 In Lisp code, optional third arg YANK-HANDLER, if non-nil,
8023 specifies the yank-handler text property to be set on the killed
8024 text. See `insert-for-yank'."
8025 ;; Pass point first, then mark, because the order matters
8026 ;; when calling kill-append.
8027 (interactive (list (point) (mark)))
8028 (unless (and beg end)
8029 (error "The mark is not set now, so there is no region"))
8031 (let ((string (filter-buffer-substring beg end t)))
8032 (when string ;STRING is nil if BEG = END
8033 ;; Add that string to the kill ring, one way or another.
8034 (if (eq last-command 'kill-region)
8035 (kill-append string (< end beg) yank-handler)
8036 (kill-new string nil yank-handler)))
8037 (when (or string (eq last-command 'kill-region))
8038 (setq this-command 'kill-region))
8040 ((buffer-read-only text-read-only)
8041 ;; The code above failed because the buffer, or some of the characters
8042 ;; in the region, are read-only.
8043 ;; We should beep, in case the user just isn't aware of this.
8044 ;; However, there's no harm in putting
8045 ;; the region's text in the kill ring, anyway.
8046 (copy-region-as-kill beg end)
8047 ;; Set this-command now, so it will be set even if we get an error.
8048 (setq this-command 'kill-region)
8049 ;; This should barf, if appropriate, and give us the correct error.
8050 (if kill-read-only-ok
8051 (progn (message "Read only text copied to kill ring") nil)
8052 ;; Signal an error if the buffer is read-only.
8053 (barf-if-buffer-read-only)
8054 ;; If the buffer isn't read-only, the text is.
8055 (signal 'text-read-only (list (current-buffer)))))))
8058 The Emacs 22 version of that function uses @code{condition-case} and
8059 @code{copy-region-as-kill}, both of which we will explain.
8060 @code{condition-case} is an important special form.
8062 In essence, the @code{kill-region} function calls
8063 @code{condition-case}, which takes three arguments. In this function,
8064 the first argument does nothing. The second argument contains the
8065 code that does the work when all goes well. The third argument
8066 contains the code that is called in the event of an error.
8069 * Complete kill-region::
8074 @node Complete kill-region, condition-case, kill-region, kill-region
8076 @unnumberedsubsec The Complete @code{kill-region} Definition
8080 We will go through the @code{condition-case} code in a moment. First,
8081 let us look at the definition of @code{kill-region}, with comments
8087 (defun kill-region (beg end)
8088 "Kill (\"cut\") text between point and mark.
8089 This deletes the text from the buffer and saves it in the kill ring.
8090 The command \\[yank] can retrieve it from there. @dots{} "
8094 ;; @bullet{} Since order matters, pass point first.
8095 (interactive (list (point) (mark)))
8096 ;; @bullet{} And tell us if we cannot cut the text.
8097 ;; `unless' is an `if' without a then-part.
8098 (unless (and beg end)
8099 (error "The mark is not set now, so there is no region"))
8103 ;; @bullet{} `condition-case' takes three arguments.
8104 ;; If the first argument is nil, as it is here,
8105 ;; information about the error signal is not
8106 ;; stored for use by another function.
8111 ;; @bullet{} The second argument to `condition-case' tells the
8112 ;; Lisp interpreter what to do when all goes well.
8116 ;; It starts with a `let' function that extracts the string
8117 ;; and tests whether it exists. If so (that is what the
8118 ;; `when' checks), it calls an `if' function that determines
8119 ;; whether the previous command was another call to
8120 ;; `kill-region'; if it was, then the new text is appended to
8121 ;; the previous text; if not, then a different function,
8122 ;; `kill-new', is called.
8126 ;; The `kill-append' function concatenates the new string and
8127 ;; the old. The `kill-new' function inserts text into a new
8128 ;; item in the kill ring.
8132 ;; `when' is an `if' without an else-part. The second `when'
8133 ;; again checks whether the current string exists; in
8134 ;; addition, it checks whether the previous command was
8135 ;; another call to `kill-region'. If one or the other
8136 ;; condition is true, then it sets the current command to
8137 ;; be `kill-region'.
8140 (let ((string (filter-buffer-substring beg end t)))
8141 (when string ;STRING is nil if BEG = END
8142 ;; Add that string to the kill ring, one way or another.
8143 (if (eq last-command 'kill-region)
8146 ;; @minus{} `yank-handler' is an optional argument to
8147 ;; `kill-region' that tells the `kill-append' and
8148 ;; `kill-new' functions how deal with properties
8149 ;; added to the text, such as `bold' or `italics'.
8150 (kill-append string (< end beg) yank-handler)
8151 (kill-new string nil yank-handler)))
8152 (when (or string (eq last-command 'kill-region))
8153 (setq this-command 'kill-region))
8158 ;; @bullet{} The third argument to `condition-case' tells the interpreter
8159 ;; what to do with an error.
8162 ;; The third argument has a conditions part and a body part.
8163 ;; If the conditions are met (in this case,
8164 ;; if text or buffer are read-only)
8165 ;; then the body is executed.
8168 ;; The first part of the third argument is the following:
8169 ((buffer-read-only text-read-only) ;; the if-part
8170 ;; @dots{} the then-part
8171 (copy-region-as-kill beg end)
8174 ;; Next, also as part of the then-part, set this-command, so
8175 ;; it will be set in an error
8176 (setq this-command 'kill-region)
8177 ;; Finally, in the then-part, send a message if you may copy
8178 ;; the text to the kill ring without signally an error, but
8179 ;; don't if you may not.
8182 (if kill-read-only-ok
8183 (progn (message "Read only text copied to kill ring") nil)
8184 (barf-if-buffer-read-only)
8185 ;; If the buffer isn't read-only, the text is.
8186 (signal 'text-read-only (list (current-buffer)))))
8194 (defun kill-region (beg end)
8195 "Kill between point and mark.
8196 The text is deleted but saved in the kill ring."
8201 ;; 1. `condition-case' takes three arguments.
8202 ;; If the first argument is nil, as it is here,
8203 ;; information about the error signal is not
8204 ;; stored for use by another function.
8209 ;; 2. The second argument to `condition-case'
8210 ;; tells the Lisp interpreter what to do when all goes well.
8214 ;; The `delete-and-extract-region' function usually does the
8215 ;; work. If the beginning and ending of the region are both
8216 ;; the same, then the variable `string' will be empty, or nil
8217 (let ((string (delete-and-extract-region beg end)))
8221 ;; `when' is an `if' clause that cannot take an `else-part'.
8222 ;; Emacs normally sets the value of `last-command' to the
8223 ;; previous command.
8226 ;; `kill-append' concatenates the new string and the old.
8227 ;; `kill-new' inserts text into a new item in the kill ring.
8229 (if (eq last-command 'kill-region)
8230 ;; if true, prepend string
8231 (kill-append string (< end beg))
8233 (setq this-command 'kill-region))
8237 ;; 3. The third argument to `condition-case' tells the interpreter
8238 ;; what to do with an error.
8241 ;; The third argument has a conditions part and a body part.
8242 ;; If the conditions are met (in this case,
8243 ;; if text or buffer are read-only)
8244 ;; then the body is executed.
8247 ((buffer-read-only text-read-only) ;; this is the if-part
8249 (copy-region-as-kill beg end)
8252 (if kill-read-only-ok ;; usually this variable is nil
8253 (message "Read only text copied to kill ring")
8254 ;; or else, signal an error if the buffer is read-only;
8255 (barf-if-buffer-read-only)
8256 ;; and, in any case, signal that the text is read-only.
8257 (signal 'text-read-only (list (current-buffer)))))))
8262 @node condition-case, Lisp macro, Complete kill-region, kill-region
8263 @comment node-name, next, previous, up
8264 @subsection @code{condition-case}
8265 @findex condition-case
8267 As we have seen earlier (@pxref{Making Errors, , Generate an Error
8268 Message}), when the Emacs Lisp interpreter has trouble evaluating an
8269 expression, it provides you with help; in the jargon, this is called
8270 ``signaling an error''. Usually, the computer stops the program and
8271 shows you a message.
8273 However, some programs undertake complicated actions. They should not
8274 simply stop on an error. In the @code{kill-region} function, the most
8275 likely error is that you will try to kill text that is read-only and
8276 cannot be removed. So the @code{kill-region} function contains code
8277 to handle this circumstance. This code, which makes up the body of
8278 the @code{kill-region} function, is inside of a @code{condition-case}
8282 The template for @code{condition-case} looks like this:
8289 @var{error-handler}@dots{})
8293 The second argument, @var{bodyform}, is straightforward. The
8294 @code{condition-case} special form causes the Lisp interpreter to
8295 evaluate the code in @var{bodyform}. If no error occurs, the special
8296 form returns the code's value and produces the side-effects, if any.
8298 In short, the @var{bodyform} part of a @code{condition-case}
8299 expression determines what should happen when everything works
8302 However, if an error occurs, among its other actions, the function
8303 generating the error signal will define one or more error condition
8306 An error handler is the third argument to @code{condition case}.
8307 An error handler has two parts, a @var{condition-name} and a
8308 @var{body}. If the @var{condition-name} part of an error handler
8309 matches a condition name generated by an error, then the @var{body}
8310 part of the error handler is run.
8312 As you will expect, the @var{condition-name} part of an error handler
8313 may be either a single condition name or a list of condition names.
8315 Also, a complete @code{condition-case} expression may contain more
8316 than one error handler. When an error occurs, the first applicable
8319 Lastly, the first argument to the @code{condition-case} expression,
8320 the @var{var} argument, is sometimes bound to a variable that
8321 contains information about the error. However, if that argument is
8322 nil, as is the case in @code{kill-region}, that information is
8326 In brief, in the @code{kill-region} function, the code
8327 @code{condition-case} works like this:
8331 @var{If no errors}, @var{run only this code}
8332 @var{but}, @var{if errors}, @var{run this other code}.
8339 copy-region-as-kill is short, 12 lines, and uses
8340 filter-buffer-substring, which is longer, 39 lines
8341 and has delete-and-extract-region in it.
8342 delete-and-extract-region is written in C.
8344 see Initializing a Variable with @code{defvar}
8346 Initializing a Variable with @code{defvar} includes line 8350
8349 @node Lisp macro, , condition-case, kill-region
8350 @comment node-name, next, previous, up
8351 @subsection Lisp macro
8355 The part of the @code{condition-case} expression that is evaluated in
8356 the expectation that all goes well has a @code{when}. The code uses
8357 @code{when} to determine whether the @code{string} variable points to
8360 A @code{when} expression is simply a programmers' convenience. It is
8361 an @code{if} without the possibility of an else clause. In your mind,
8362 you can replace @code{when} with @code{if} and understand what goes
8363 on. That is what the Lisp interpreter does.
8365 Technically speaking, @code{when} is a Lisp macro. A Lisp @dfn{macro}
8366 enables you to define new control constructs and other language
8367 features. It tells the interpreter how to compute another Lisp
8368 expression which will in turn compute the value. In this case, the
8369 `other expression' is an @code{if} expression.
8371 The @code{kill-region} function definition also has an @code{unless}
8372 macro; it is the converse of @code{when}. The @code{unless} macro is
8373 an @code{if} without a then clause
8375 For more about Lisp macros, see @ref{Macros, , Macros, elisp, The GNU
8376 Emacs Lisp Reference Manual}. The C programming language also
8377 provides macros. These are different, but also useful.
8380 We will briefly look at C macros in
8381 @ref{Digression into C}.
8385 Regarding the @code{when} macro, in the @code{condition-case}
8386 expression, when the string has content, then another conditional
8387 expression is executed. This is an @code{if} with both a then-part
8392 (if (eq last-command 'kill-region)
8393 (kill-append string (< end beg) yank-handler)
8394 (kill-new string nil yank-handler))
8398 The then-part is evaluated if the previous command was another call to
8399 @code{kill-region}; if not, the else-part is evaluated.
8401 @code{yank-handler} is an optional argument to @code{kill-region} that
8402 tells the @code{kill-append} and @code{kill-new} functions how deal
8403 with properties added to the text, such as `bold' or `italics'.
8405 @code{last-command} is a variable that comes with Emacs that we have
8406 not seen before. Normally, whenever a function is executed, Emacs
8407 sets the value of @code{last-command} to the previous command.
8410 In this segment of the definition, the @code{if} expression checks
8411 whether the previous command was @code{kill-region}. If it was,
8414 (kill-append string (< end beg) yank-handler)
8418 concatenates a copy of the newly clipped text to the just previously
8419 clipped text in the kill ring.
8421 @node copy-region-as-kill, Digression into C, kill-region, Cutting & Storing Text
8422 @comment node-name, next, previous, up
8423 @section @code{copy-region-as-kill}
8424 @findex copy-region-as-kill
8427 The @code{copy-region-as-kill} function copies a region of text from a
8428 buffer and (via either @code{kill-append} or @code{kill-new}) saves it
8429 in the @code{kill-ring}.
8431 If you call @code{copy-region-as-kill} immediately after a
8432 @code{kill-region} command, Emacs appends the newly copied text to the
8433 previously copied text. This means that if you yank back the text, you
8434 get it all, from both this and the previous operation. On the other
8435 hand, if some other command precedes the @code{copy-region-as-kill},
8436 the function copies the text into a separate entry in the kill ring.
8439 * Complete copy-region-as-kill::
8440 * copy-region-as-kill body::
8443 @node Complete copy-region-as-kill, copy-region-as-kill body, copy-region-as-kill, copy-region-as-kill
8445 @unnumberedsubsec The complete @code{copy-region-as-kill} function definition
8449 Here is the complete text of the version 22 @code{copy-region-as-kill}
8454 (defun copy-region-as-kill (beg end)
8455 "Save the region as if killed, but don't kill it.
8456 In Transient Mark mode, deactivate the mark.
8457 If `interprogram-cut-function' is non-nil, also save the text for a window
8458 system cut and paste."
8462 (if (eq last-command 'kill-region)
8463 (kill-append (filter-buffer-substring beg end) (< end beg))
8464 (kill-new (filter-buffer-substring beg end)))
8467 (if transient-mark-mode
8468 (setq deactivate-mark t))
8474 As usual, this function can be divided into its component parts:
8478 (defun copy-region-as-kill (@var{argument-list})
8479 "@var{documentation}@dots{}"
8485 The arguments are @code{beg} and @code{end} and the function is
8486 interactive with @code{"r"}, so the two arguments must refer to the
8487 beginning and end of the region. If you have been reading though this
8488 document from the beginning, understanding these parts of a function is
8489 almost becoming routine.
8491 The documentation is somewhat confusing unless you remember that the
8492 word `kill' has a meaning different from usual. The `Transient Mark'
8493 and @code{interprogram-cut-function} comments explain certain
8496 After you once set a mark, a buffer always contains a region. If you
8497 wish, you can use Transient Mark mode to highlight the region
8498 temporarily. (No one wants to highlight the region all the time, so
8499 Transient Mark mode highlights it only at appropriate times. Many
8500 people turn off Transient Mark mode, so the region is never
8503 Also, a windowing system allows you to copy, cut, and paste among
8504 different programs. In the X windowing system, for example, the
8505 @code{interprogram-cut-function} function is @code{x-select-text},
8506 which works with the windowing system's equivalent of the Emacs kill
8509 The body of the @code{copy-region-as-kill} function starts with an
8510 @code{if} clause. What this clause does is distinguish between two
8511 different situations: whether or not this command is executed
8512 immediately after a previous @code{kill-region} command. In the first
8513 case, the new region is appended to the previously copied text.
8514 Otherwise, it is inserted into the beginning of the kill ring as a
8515 separate piece of text from the previous piece.
8517 The last two lines of the function prevent the region from lighting up
8518 if Transient Mark mode is turned on.
8520 The body of @code{copy-region-as-kill} merits discussion in detail.
8522 @node copy-region-as-kill body, , Complete copy-region-as-kill, copy-region-as-kill
8523 @comment node-name, next, previous, up
8524 @subsection The Body of @code{copy-region-as-kill}
8526 The @code{copy-region-as-kill} function works in much the same way as
8527 the @code{kill-region} function. Both are written so that two or more
8528 kills in a row combine their text into a single entry. If you yank
8529 back the text from the kill ring, you get it all in one piece.
8530 Moreover, kills that kill forward from the current position of the
8531 cursor are added to the end of the previously copied text and commands
8532 that copy text backwards add it to the beginning of the previously
8533 copied text. This way, the words in the text stay in the proper
8536 Like @code{kill-region}, the @code{copy-region-as-kill} function makes
8537 use of the @code{last-command} variable that keeps track of the
8538 previous Emacs command.
8541 * last-command & this-command::
8542 * kill-append function::
8543 * kill-new function::
8546 @node last-command & this-command, kill-append function, copy-region-as-kill body, copy-region-as-kill body
8548 @unnumberedsubsubsec @code{last-command} and @code{this-command}
8551 Normally, whenever a function is executed, Emacs sets the value of
8552 @code{this-command} to the function being executed (which in this case
8553 would be @code{copy-region-as-kill}). At the same time, Emacs sets
8554 the value of @code{last-command} to the previous value of
8555 @code{this-command}.
8557 In the first part of the body of the @code{copy-region-as-kill}
8558 function, an @code{if} expression determines whether the value of
8559 @code{last-command} is @code{kill-region}. If so, the then-part of
8560 the @code{if} expression is evaluated; it uses the @code{kill-append}
8561 function to concatenate the text copied at this call to the function
8562 with the text already in the first element (the @sc{car}) of the kill
8563 ring. On the other hand, if the value of @code{last-command} is not
8564 @code{kill-region}, then the @code{copy-region-as-kill} function
8565 attaches a new element to the kill ring using the @code{kill-new}
8569 The @code{if} expression reads as follows; it uses @code{eq}:
8573 (if (eq last-command 'kill-region)
8575 (kill-append (filter-buffer-substring beg end) (< end beg))
8577 (kill-new (filter-buffer-substring beg end)))
8581 @findex filter-buffer-substring
8582 (The @code{filter-buffer-substring} function returns a filtered
8583 substring of the buffer, if any. Optionally---the arguments are not
8584 here, so neither is done---the function may delete the initial text or
8585 return the text without its properties; this function is a replacement
8586 for the older @code{buffer-substring} function, which came before text
8587 properties were implemented.)
8589 @findex eq @r{(example of use)}
8591 The @code{eq} function tests whether its first argument is the same Lisp
8592 object as its second argument. The @code{eq} function is similar to the
8593 @code{equal} function in that it is used to test for equality, but
8594 differs in that it determines whether two representations are actually
8595 the same object inside the computer, but with different names.
8596 @code{equal} determines whether the structure and contents of two
8597 expressions are the same.
8599 If the previous command was @code{kill-region}, then the Emacs Lisp
8600 interpreter calls the @code{kill-append} function
8602 @node kill-append function, kill-new function, last-command & this-command, copy-region-as-kill body
8603 @unnumberedsubsubsec The @code{kill-append} function
8607 The @code{kill-append} function looks like this:
8612 (defun kill-append (string before-p &optional yank-handler)
8613 "Append STRING to the end of the latest kill in the kill ring.
8614 If BEFORE-P is non-nil, prepend STRING to the kill.
8616 (let* ((cur (car kill-ring)))
8617 (kill-new (if before-p (concat string cur) (concat cur string))
8618 (or (= (length cur) 0)
8620 (get-text-property 0 'yank-handler cur)))
8627 (defun kill-append (string before-p)
8628 "Append STRING to the end of the latest kill in the kill ring.
8629 If BEFORE-P is non-nil, prepend STRING to the kill.
8630 If `interprogram-cut-function' is set, pass the resulting kill to
8632 (kill-new (if before-p
8633 (concat string (car kill-ring))
8634 (concat (car kill-ring) string))
8639 The @code{kill-append} function is fairly straightforward. It uses
8640 the @code{kill-new} function, which we will discuss in more detail in
8643 (Also, the function provides an optional argument called
8644 @code{yank-handler}; when invoked, this argument tells the function
8645 how to deal with properties added to the text, such as `bold' or
8648 @c !!! bug in GNU Emacs 22 version of kill-append ?
8649 It has a @code{let*} function to set the value of the first element of
8650 the kill ring to @code{cur}. (I do not know why the function does not
8651 use @code{let} instead; only one value is set in the expression.
8652 Perhaps this is a bug that produces no problems?)
8654 Consider the conditional that is one of the two arguments to
8655 @code{kill-new}. It uses @code{concat} to concatenate the new text to
8656 the @sc{car} of the kill ring. Whether it prepends or appends the
8657 text depends on the results of an @code{if} expression:
8661 (if before-p ; @r{if-part}
8662 (concat string cur) ; @r{then-part}
8663 (concat cur string)) ; @r{else-part}
8668 If the region being killed is before the region that was killed in the
8669 last command, then it should be prepended before the material that was
8670 saved in the previous kill; and conversely, if the killed text follows
8671 what was just killed, it should be appended after the previous text.
8672 The @code{if} expression depends on the predicate @code{before-p} to
8673 decide whether the newly saved text should be put before or after the
8674 previously saved text.
8676 The symbol @code{before-p} is the name of one of the arguments to
8677 @code{kill-append}. When the @code{kill-append} function is
8678 evaluated, it is bound to the value returned by evaluating the actual
8679 argument. In this case, this is the expression @code{(< end beg)}.
8680 This expression does not directly determine whether the killed text in
8681 this command is located before or after the kill text of the last
8682 command; what it does is determine whether the value of the variable
8683 @code{end} is less than the value of the variable @code{beg}. If it
8684 is, it means that the user is most likely heading towards the
8685 beginning of the buffer. Also, the result of evaluating the predicate
8686 expression, @code{(< end beg)}, will be true and the text will be
8687 prepended before the previous text. On the other hand, if the value of
8688 the variable @code{end} is greater than the value of the variable
8689 @code{beg}, the text will be appended after the previous text.
8692 When the newly saved text will be prepended, then the string with the new
8693 text will be concatenated before the old text:
8701 But if the text will be appended, it will be concatenated
8705 (concat cur string))
8708 To understand how this works, we first need to review the
8709 @code{concat} function. The @code{concat} function links together or
8710 unites two strings of text. The result is a string. For example:
8714 (concat "abc" "def")
8720 (car '("first element" "second element")))
8721 @result{} "new first element"
8724 '("first element" "second element")) " modified")
8725 @result{} "first element modified"
8729 We can now make sense of @code{kill-append}: it modifies the contents
8730 of the kill ring. The kill ring is a list, each element of which is
8731 saved text. The @code{kill-append} function uses the @code{kill-new}
8732 function which in turn uses the @code{setcar} function.
8734 @node kill-new function, , kill-append function, copy-region-as-kill body
8735 @unnumberedsubsubsec The @code{kill-new} function
8738 @c in GNU Emacs 22, additional documentation to kill-new:
8740 Optional third arguments YANK-HANDLER controls how the STRING is later
8741 inserted into a buffer; see `insert-for-yank' for details.
8742 When a yank handler is specified, STRING must be non-empty (the yank
8743 handler, if non-nil, is stored as a `yank-handler' text property on STRING).
8745 When the yank handler has a non-nil PARAM element, the original STRING
8746 argument is not used by `insert-for-yank'. However, since Lisp code
8747 may access and use elements from the kill ring directly, the STRING
8748 argument should still be a \"useful\" string for such uses."
8751 The @code{kill-new} function looks like this:
8755 (defun kill-new (string &optional replace yank-handler)
8756 "Make STRING the latest kill in the kill ring.
8757 Set `kill-ring-yank-pointer' to point to it.
8759 If `interprogram-cut-function' is non-nil, apply it to STRING.
8760 Optional second argument REPLACE non-nil means that STRING will replace
8761 the front of the kill ring, rather than being added to the list.
8765 (if (> (length string) 0)
8767 (put-text-property 0 (length string)
8768 'yank-handler yank-handler string))
8770 (signal 'args-out-of-range
8771 (list string "yank-handler specified for empty string"))))
8774 (if (fboundp 'menu-bar-update-yank-menu)
8775 (menu-bar-update-yank-menu string (and replace (car kill-ring))))
8778 (if (and replace kill-ring)
8779 (setcar kill-ring string)
8780 (push string kill-ring)
8781 (if (> (length kill-ring) kill-ring-max)
8782 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)))
8785 (setq kill-ring-yank-pointer kill-ring)
8786 (if interprogram-cut-function
8787 (funcall interprogram-cut-function string (not replace))))
8792 (defun kill-new (string &optional replace)
8793 "Make STRING the latest kill in the kill ring.
8794 Set the kill-ring-yank pointer to point to it.
8795 If `interprogram-cut-function' is non-nil, apply it to STRING.
8796 Optional second argument REPLACE non-nil means that STRING will replace
8797 the front of the kill ring, rather than being added to the list."
8798 (and (fboundp 'menu-bar-update-yank-menu)
8799 (menu-bar-update-yank-menu string (and replace (car kill-ring))))
8800 (if (and replace kill-ring)
8801 (setcar kill-ring string)
8802 (setq kill-ring (cons string kill-ring))
8803 (if (> (length kill-ring) kill-ring-max)
8804 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)))
8805 (setq kill-ring-yank-pointer kill-ring)
8806 (if interprogram-cut-function
8807 (funcall interprogram-cut-function string (not replace))))
8810 (Notice that the function is not interactive.)
8812 As usual, we can look at this function in parts.
8814 The function definition has an optional @code{yank-handler} argument,
8815 which when invoked tells the function how to deal with properties
8816 added to the text, such as `bold' or `italics'. We will skip that.
8819 The first line of the documentation makes sense:
8822 Make STRING the latest kill in the kill ring.
8826 Let's skip over the rest of the documentation for the moment.
8829 Also, let's skip over the initial @code{if} expression and those lines
8830 of code involving @code{menu-bar-update-yank-menu}. We will explain
8834 The critical lines are these:
8838 (if (and replace kill-ring)
8840 (setcar kill-ring string)
8844 (push string kill-ring)
8847 (setq kill-ring (cons string kill-ring))
8848 (if (> (length kill-ring) kill-ring-max)
8849 ;; @r{avoid overly long kill ring}
8850 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)))
8853 (setq kill-ring-yank-pointer kill-ring)
8854 (if interprogram-cut-function
8855 (funcall interprogram-cut-function string (not replace))))
8859 The conditional test is @w{@code{(and replace kill-ring)}}.
8860 This will be true when two conditions are met: the kill ring has
8861 something in it, and the @code{replace} variable is true.
8864 When the @code{kill-append} function sets @code{replace} to be true
8865 and when the kill ring has at least one item in it, the @code{setcar}
8866 expression is executed:
8869 (setcar kill-ring string)
8872 The @code{setcar} function actually changes the first element of the
8873 @code{kill-ring} list to the value of @code{string}. It replaces the
8877 On the other hand, if the kill ring is empty, or replace is false, the
8878 else-part of the condition is executed:
8881 (push string kill-ring)
8886 @code{push} puts its first argument onto the second. It is similar to
8890 (setq kill-ring (cons string kill-ring))
8898 (add-to-list kill-ring string)
8902 When it is false, the expression first constructs a new version of the
8903 kill ring by prepending @code{string} to the existing kill ring as a
8904 new element (that is what the @code{push} does). Then it executes a
8905 second @code{if} clause. This second @code{if} clause keeps the kill
8906 ring from growing too long.
8908 Let's look at these two expressions in order.
8910 The @code{push} line of the else-part sets the new value of the kill
8911 ring to what results from adding the string being killed to the old
8914 We can see how this works with an example.
8920 (setq example-list '("here is a clause" "another clause"))
8925 After evaluating this expression with @kbd{C-x C-e}, you can evaluate
8926 @code{example-list} and see what it returns:
8931 @result{} ("here is a clause" "another clause")
8937 Now, we can add a new element on to this list by evaluating the
8938 following expression:
8939 @findex push, @r{example}
8942 (push "a third clause" example-list)
8947 When we evaluate @code{example-list}, we find its value is:
8952 @result{} ("a third clause" "here is a clause" "another clause")
8957 Thus, the third clause is added to the list by @code{push}.
8960 Now for the second part of the @code{if} clause. This expression
8961 keeps the kill ring from growing too long. It looks like this:
8965 (if (> (length kill-ring) kill-ring-max)
8966 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil))
8970 The code checks whether the length of the kill ring is greater than
8971 the maximum permitted length. This is the value of
8972 @code{kill-ring-max} (which is 60, by default). If the length of the
8973 kill ring is too long, then this code sets the last element of the
8974 kill ring to @code{nil}. It does this by using two functions,
8975 @code{nthcdr} and @code{setcdr}.
8977 We looked at @code{setcdr} earlier (@pxref{setcdr, , @code{setcdr}}).
8978 It sets the @sc{cdr} of a list, just as @code{setcar} sets the
8979 @sc{car} of a list. In this case, however, @code{setcdr} will not be
8980 setting the @sc{cdr} of the whole kill ring; the @code{nthcdr}
8981 function is used to cause it to set the @sc{cdr} of the next to last
8982 element of the kill ring---this means that since the @sc{cdr} of the
8983 next to last element is the last element of the kill ring, it will set
8984 the last element of the kill ring.
8986 @findex nthcdr, @r{example}
8987 The @code{nthcdr} function works by repeatedly taking the @sc{cdr} of a
8988 list---it takes the @sc{cdr} of the @sc{cdr} of the @sc{cdr}
8989 @dots{} It does this @var{N} times and returns the results.
8990 (@xref{nthcdr, , @code{nthcdr}}.)
8992 @findex setcdr, @r{example}
8993 Thus, if we had a four element list that was supposed to be three
8994 elements long, we could set the @sc{cdr} of the next to last element
8995 to @code{nil}, and thereby shorten the list. (If you set the last
8996 element to some other value than @code{nil}, which you could do, then
8997 you would not have shortened the list. @xref{setcdr, ,
9000 You can see shortening by evaluating the following three expressions
9001 in turn. First set the value of @code{trees} to @code{(maple oak pine
9002 birch)}, then set the @sc{cdr} of its second @sc{cdr} to @code{nil}
9003 and then find the value of @code{trees}:
9007 (setq trees '(maple oak pine birch))
9008 @result{} (maple oak pine birch)
9012 (setcdr (nthcdr 2 trees) nil)
9016 @result{} (maple oak pine)
9021 (The value returned by the @code{setcdr} expression is @code{nil} since
9022 that is what the @sc{cdr} is set to.)
9024 To repeat, in @code{kill-new}, the @code{nthcdr} function takes the
9025 @sc{cdr} a number of times that is one less than the maximum permitted
9026 size of the kill ring and @code{setcdr} sets the @sc{cdr} of that
9027 element (which will be the rest of the elements in the kill ring) to
9028 @code{nil}. This prevents the kill ring from growing too long.
9031 The next to last expression in the @code{kill-new} function is
9034 (setq kill-ring-yank-pointer kill-ring)
9037 The @code{kill-ring-yank-pointer} is a global variable that is set to be
9038 the @code{kill-ring}.
9040 Even though the @code{kill-ring-yank-pointer} is called a
9041 @samp{pointer}, it is a variable just like the kill ring. However, the
9042 name has been chosen to help humans understand how the variable is used.
9045 Now, to return to an early expression in the body of the function:
9049 (if (fboundp 'menu-bar-update-yank-menu)
9050 (menu-bar-update-yank-menu string (and replace (car kill-ring))))
9055 It starts with an @code{if} expression
9057 In this case, the expression tests first to see whether
9058 @code{menu-bar-update-yank-menu} exists as a function, and if so,
9059 calls it. The @code{fboundp} function returns true if the symbol it
9060 is testing has a function definition that `is not void'. If the
9061 symbol's function definition were void, we would receive an error
9062 message, as we did when we created errors intentionally (@pxref{Making
9063 Errors, , Generate an Error Message}).
9066 The then-part contains an expression whose first element is the
9067 function @code{and}.
9070 The @code{and} special form evaluates each of its arguments until one
9071 of the arguments returns a value of @code{nil}, in which case the
9072 @code{and} expression returns @code{nil}; however, if none of the
9073 arguments returns a value of @code{nil}, the value resulting from
9074 evaluating the last argument is returned. (Since such a value is not
9075 @code{nil}, it is considered true in Emacs Lisp.) In other words, an
9076 @code{and} expression returns a true value only if all its arguments
9077 are true. (@xref{Second Buffer Related Review}.)
9079 The expression determines whether the second argument to
9080 @code{menu-bar-update-yank-menu} is true or not.
9082 ;; If we're supposed to be extending an existing string, and that
9083 ;; string really is at the front of the menu, then update it in place.
9086 @code{menu-bar-update-yank-menu} is one of the functions that make it
9087 possible to use the `Select and Paste' menu in the Edit item of a menu
9088 bar; using a mouse, you can look at the various pieces of text you
9089 have saved and select one piece to paste.
9091 The last expression in the @code{kill-new} function adds the newly
9092 copied string to whatever facility exists for copying and pasting
9093 among different programs running in a windowing system. In the X
9094 Windowing system, for example, the @code{x-select-text} function takes
9095 the string and stores it in memory operated by X. You can paste the
9096 string in another program, such as an Xterm.
9099 The expression looks like this:
9103 (if interprogram-cut-function
9104 (funcall interprogram-cut-function string (not replace))))
9108 If an @code{interprogram-cut-function} exists, then Emacs executes
9109 @code{funcall}, which in turn calls its first argument as a function
9110 and passes the remaining arguments to it. (Incidentally, as far as I
9111 can see, this @code{if} expression could be replaced by an @code{and}
9112 expression similar to the one in the first part of the function.)
9114 We are not going to discuss windowing systems and other programs
9115 further, but merely note that this is a mechanism that enables GNU
9116 Emacs to work easily and well with other programs.
9118 This code for placing text in the kill ring, either concatenated with
9119 an existing element or as a new element, leads us to the code for
9120 bringing back text that has been cut out of the buffer---the yank
9121 commands. However, before discussing the yank commands, it is better
9122 to learn how lists are implemented in a computer. This will make
9123 clear such mysteries as the use of the term `pointer'. But before
9124 that, we will digress into C.
9127 @c is this true in Emacs 22? Does not seems to be
9129 (If the @w{@code{(< end beg))}}
9130 expression is true, @code{kill-append} prepends the string to the just
9131 previously clipped text. For a detailed discussion, see
9132 @ref{kill-append function, , The @code{kill-append} function}.)
9134 If you then yank back the text, i.e., `paste' it, you get both
9135 pieces of text at once. That way, if you delete two words in a row,
9136 and then yank them back, you get both words, in their proper order,
9137 with one yank. (The @w{@code{(< end beg))}} expression makes sure the
9140 On the other hand, if the previous command is not @code{kill-region},
9141 then the @code{kill-new} function is called, which adds the text to
9142 the kill ring as the latest item, and sets the
9143 @code{kill-ring-yank-pointer} variable to point to it.
9147 @c Evidently, changed for Emacs 22. The zap-to-char command does not
9148 @c use the delete-and-extract-region function
9150 2006 Oct 26, the Digression into C is now OK but should come after
9151 copy-region-as-kill and filter-buffer-substring
9155 copy-region-as-kill is short, 12 lines, and uses
9156 filter-buffer-substring, which is longer, 39 lines
9157 and has delete-and-extract-region in it.
9158 delete-and-extract-region is written in C.
9160 see Initializing a Variable with @code{defvar}
9163 @node Digression into C, defvar, copy-region-as-kill, Cutting & Storing Text
9164 @comment node-name, next, previous, up
9165 @section Digression into C
9166 @findex delete-and-extract-region
9167 @cindex C, a digression into
9168 @cindex Digression into C
9170 The @code{copy-region-as-kill} function (@pxref{copy-region-as-kill, ,
9171 @code{copy-region-as-kill}}) uses the @code{filter-buffer-substring}
9172 function, which in turn uses the @code{delete-and-extract-region}
9173 function. It removes the contents of a region and you cannot get them
9176 Unlike the other code discussed here, the
9177 @code{delete-and-extract-region} function is not written in Emacs
9178 Lisp; it is written in C and is one of the primitives of the GNU Emacs
9179 system. Since it is very simple, I will digress briefly from Lisp and
9182 @c GNU Emacs 22 in /usr/local/src/emacs/src/editfns.c
9183 @c the DEFUN for buffer-substring-no-properties
9186 Like many of the other Emacs primitives,
9187 @code{delete-and-extract-region} is written as an instance of a C
9188 macro, a macro being a template for code. The complete macro looks
9193 DEFUN ("buffer-substring-no-properties", Fbuffer_substring_no_properties,
9194 Sbuffer_substring_no_properties, 2, 2, 0,
9195 doc: /* Return the characters of part of the buffer,
9196 without the text properties.
9197 The two arguments START and END are character positions;
9198 they can be in either order. */)
9200 Lisp_Object start, end;
9204 validate_region (&start, &end);
9208 return make_buffer_string (b, e, 0);
9213 Without going into the details of the macro writing process, let me
9214 point out that this macro starts with the word @code{DEFUN}. The word
9215 @code{DEFUN} was chosen since the code serves the same purpose as
9216 @code{defun} does in Lisp. (The @code{DEFUN} C macro is defined in
9217 @file{emacs/src/lisp.h}.)
9219 The word @code{DEFUN} is followed by seven parts inside of
9224 The first part is the name given to the function in Lisp,
9225 @code{delete-and-extract-region}.
9228 The second part is the name of the function in C,
9229 @code{Fdelete_and_extract_region}. By convention, it starts with
9230 @samp{F}. Since C does not use hyphens in names, underscores are used
9234 The third part is the name for the C constant structure that records
9235 information on this function for internal use. It is the name of the
9236 function in C but begins with an @samp{S} instead of an @samp{F}.
9239 The fourth and fifth parts specify the minimum and maximum number of
9240 arguments the function can have. This function demands exactly 2
9244 The sixth part is nearly like the argument that follows the
9245 @code{interactive} declaration in a function written in Lisp: a letter
9246 followed, perhaps, by a prompt. The only difference from the Lisp is
9247 when the macro is called with no arguments. Then you write a @code{0}
9248 (which is a `null string'), as in this macro.
9250 If you were to specify arguments, you would place them between
9251 quotation marks. The C macro for @code{goto-char} includes
9252 @code{"NGoto char: "} in this position to indicate that the function
9253 expects a raw prefix, in this case, a numerical location in a buffer,
9254 and provides a prompt.
9257 The seventh part is a documentation string, just like the one for a
9258 function written in Emacs Lisp, except that every newline must be
9259 written explicitly as @samp{\n} followed by a backslash and carriage
9263 Thus, the first two lines of documentation for @code{goto-char} are
9268 "Set point to POSITION, a number or marker.\n\
9269 Beginning of buffer is position (point-min), end is (point-max)."
9275 In a C macro, the formal parameters come next, with a statement of
9276 what kind of object they are, followed by what might be called the `body'
9277 of the macro. For @code{delete-and-extract-region} the `body'
9278 consists of the following four lines:
9282 validate_region (&start, &end);
9283 if (XINT (start) == XINT (end))
9284 return build_string ("");
9285 return del_range_1 (XINT (start), XINT (end), 1, 1);
9289 The @code{validate_region} function checks whether the values
9290 passed as the beginning and end of the region are the proper type and
9291 are within range. If the beginning and end positions are the same,
9292 then return and empty string.
9294 The @code{del_range_1} function actually deletes the text. It is a
9295 complex function we will not look into. It updates the buffer and
9296 does other things. However, it is worth looking at the two arguments
9297 passed to @code{del_range}. These are @w{@code{XINT (start)}} and
9298 @w{@code{XINT (end)}}.
9300 As far as the C language is concerned, @code{start} and @code{end} are
9301 two integers that mark the beginning and end of the region to be
9302 deleted@footnote{More precisely, and requiring more expert knowledge
9303 to understand, the two integers are of type `Lisp_Object', which can
9304 also be a C union instead of an integer type.}.
9306 In early versions of Emacs, these two numbers were thirty-two bits
9307 long, but the code is slowly being generalized to handle other
9308 lengths. Three of the available bits are used to specify the type of
9309 information; the remaining bits are used as `content'.
9311 @samp{XINT} is a C macro that extracts the relevant number from the
9312 longer collection of bits; the three other bits are discarded.
9315 The command in @code{delete-and-extract-region} looks like this:
9318 del_range_1 (XINT (start), XINT (end), 1, 1);
9322 It deletes the region between the beginning position, @code{start},
9323 and the ending position, @code{end}.
9325 From the point of view of the person writing Lisp, Emacs is all very
9326 simple; but hidden underneath is a great deal of complexity to make it
9329 @node defvar, cons & search-fwd Review, Digression into C, Cutting & Storing Text
9330 @comment node-name, next, previous, up
9331 @section Initializing a Variable with @code{defvar}
9333 @cindex Initializing a variable
9334 @cindex Variable initialization
9339 copy-region-as-kill is short, 12 lines, and uses
9340 filter-buffer-substring, which is longer, 39 lines
9341 and has delete-and-extract-region in it.
9342 delete-and-extract-region is written in C.
9344 see Initializing a Variable with @code{defvar}
9348 The @code{copy-region-as-kill} function is written in Emacs Lisp. Two
9349 functions within it, @code{kill-append} and @code{kill-new}, copy a
9350 region in a buffer and save it in a variable called the
9351 @code{kill-ring}. This section describes how the @code{kill-ring}
9352 variable is created and initialized using the @code{defvar} special
9355 (Again we note that the term @code{kill-ring} is a misnomer. The text
9356 that is clipped out of the buffer can be brought back; it is not a ring
9357 of corpses, but a ring of resurrectable text.)
9359 In Emacs Lisp, a variable such as the @code{kill-ring} is created and
9360 given an initial value by using the @code{defvar} special form. The
9361 name comes from ``define variable''.
9363 The @code{defvar} special form is similar to @code{setq} in that it sets
9364 the value of a variable. It is unlike @code{setq} in two ways: first,
9365 it only sets the value of the variable if the variable does not already
9366 have a value. If the variable already has a value, @code{defvar} does
9367 not override the existing value. Second, @code{defvar} has a
9368 documentation string.
9370 (Another special form, @code{defcustom}, is designed for variables
9371 that people customize. It has more features than @code{defvar}.
9372 (@xref{defcustom, , Setting Variables with @code{defcustom}}.)
9375 * See variable current value::
9376 * defvar and asterisk::
9379 @node See variable current value, defvar and asterisk, defvar, defvar
9381 @unnumberedsubsec Seeing the Current Value of a Variable
9384 You can see the current value of a variable, any variable, by using
9385 the @code{describe-variable} function, which is usually invoked by
9386 typing @kbd{C-h v}. If you type @kbd{C-h v} and then @code{kill-ring}
9387 (followed by @key{RET}) when prompted, you will see what is in your
9388 current kill ring---this may be quite a lot! Conversely, if you have
9389 been doing nothing this Emacs session except read this document, you
9390 may have nothing in it. Also, you will see the documentation for
9396 List of killed text sequences.
9397 Since the kill ring is supposed to interact nicely with cut-and-paste
9398 facilities offered by window systems, use of this variable should
9401 interact nicely with `interprogram-cut-function' and
9402 `interprogram-paste-function'. The functions `kill-new',
9403 `kill-append', and `current-kill' are supposed to implement this
9404 interaction; you may want to use them instead of manipulating the kill
9410 The kill ring is defined by a @code{defvar} in the following way:
9414 (defvar kill-ring nil
9415 "List of killed text sequences.
9421 In this variable definition, the variable is given an initial value of
9422 @code{nil}, which makes sense, since if you have saved nothing, you want
9423 nothing back if you give a @code{yank} command. The documentation
9424 string is written just like the documentation string of a @code{defun}.
9425 As with the documentation string of the @code{defun}, the first line of
9426 the documentation should be a complete sentence, since some commands,
9427 like @code{apropos}, print only the first line of documentation.
9428 Succeeding lines should not be indented; otherwise they look odd when
9429 you use @kbd{C-h v} (@code{describe-variable}).
9431 @node defvar and asterisk, , See variable current value, defvar
9432 @subsection @code{defvar} and an asterisk
9433 @findex defvar @r{for a user customizable variable}
9434 @findex defvar @r{with an asterisk}
9436 In the past, Emacs used the @code{defvar} special form both for
9437 internal variables that you would not expect a user to change and for
9438 variables that you do expect a user to change. Although you can still
9439 use @code{defvar} for user customizable variables, please use
9440 @code{defcustom} instead, since that special form provides a path into
9441 the Customization commands. (@xref{defcustom, , Specifying Variables
9442 using @code{defcustom}}.)
9444 When you specified a variable using the @code{defvar} special form,
9445 you could distinguish a readily settable variable from others by
9446 typing an asterisk, @samp{*}, in the first column of its documentation
9447 string. For example:
9451 (defvar shell-command-default-error-buffer nil
9452 "*Buffer name for `shell-command' @dots{} error output.
9457 @findex set-variable
9459 You could (and still can) use the @code{set-variable} command to
9460 change the value of @code{shell-command-default-error-buffer}
9461 temporarily. However, options set using @code{set-variable} are set
9462 only for the duration of your editing session. The new values are not
9463 saved between sessions. Each time Emacs starts, it reads the original
9464 value, unless you change the value within your @file{.emacs} file,
9465 either by setting it manually or by using @code{customize}.
9466 @xref{Emacs Initialization, , Your @file{.emacs} File}.
9468 For me, the major use of the @code{set-variable} command is to suggest
9469 variables that I might want to set in my @file{.emacs} file. There
9470 are now more than 700 such variables --- far too many to remember
9471 readily. Fortunately, you can press @key{TAB} after calling the
9472 @code{M-x set-variable} command to see the list of variables.
9473 (@xref{Examining, , Examining and Setting Variables, emacs,
9474 The GNU Emacs Manual}.)
9477 @node cons & search-fwd Review, search Exercises, defvar, Cutting & Storing Text
9478 @comment node-name, next, previous, up
9481 Here is a brief summary of some recently introduced functions.
9486 @code{car} returns the first element of a list; @code{cdr} returns the
9487 second and subsequent elements of a list.
9494 (car '(1 2 3 4 5 6 7))
9496 (cdr '(1 2 3 4 5 6 7))
9497 @result{} (2 3 4 5 6 7)
9502 @code{cons} constructs a list by prepending its first argument to its
9516 @code{funcall} evaluates its first argument as a function. It passes
9517 its remaining arguments to its first argument.
9520 Return the result of taking @sc{cdr} `n' times on a list.
9528 The `rest of the rest', as it were.
9535 (nthcdr 3 '(1 2 3 4 5 6 7))
9542 @code{setcar} changes the first element of a list; @code{setcdr}
9543 changes the second and subsequent elements of a list.
9550 (setq triple '(1 2 3))
9557 (setcdr triple '("foo" "bar"))
9560 @result{} (37 "foo" "bar")
9565 Evaluate each argument in sequence and then return the value of the
9578 @item save-restriction
9579 Record whatever narrowing is in effect in the current buffer, if any,
9580 and restore that narrowing after evaluating the arguments.
9582 @item search-forward
9583 Search for a string, and if the string is found, move point. With a
9584 regular expression, use the similar @code{re-search-forward}.
9585 (@xref{Regexp Search, , Regular Expression Searches}, for an
9586 explanation of regular expression patterns and searches.)
9590 @code{search-forward} and @code{re-search-forward} take four
9595 The string or regular expression to search for.
9598 Optionally, the limit of the search.
9601 Optionally, what to do if the search fails, return @code{nil} or an
9605 Optionally, how many times to repeat the search; if negative, the
9606 search goes backwards.
9610 @itemx delete-and-extract-region
9611 @itemx copy-region-as-kill
9613 @code{kill-region} cuts the text between point and mark from the
9614 buffer and stores that text in the kill ring, so you can get it back
9617 @code{copy-region-as-kill} copies the text between point and mark into
9618 the kill ring, from which you can get it by yanking. The function
9619 does not cut or remove the text from the buffer.
9622 @code{delete-and-extract-region} removes the text between point and
9623 mark from the buffer and throws it away. You cannot get it back.
9624 (This is not an interactive command.)
9627 @node search Exercises, , cons & search-fwd Review, Cutting & Storing Text
9628 @section Searching Exercises
9632 Write an interactive function that searches for a string. If the
9633 search finds the string, leave point after it and display a message
9634 that says ``Found!''. (Do not use @code{search-forward} for the name
9635 of this function; if you do, you will overwrite the existing version of
9636 @code{search-forward} that comes with Emacs. Use a name such as
9637 @code{test-search} instead.)
9640 Write a function that prints the third element of the kill ring in the
9641 echo area, if any; if the kill ring does not contain a third element,
9642 print an appropriate message.
9645 @node List Implementation, Yanking, Cutting & Storing Text, Top
9646 @comment node-name, next, previous, up
9647 @chapter How Lists are Implemented
9648 @cindex Lists in a computer
9650 In Lisp, atoms are recorded in a straightforward fashion; if the
9651 implementation is not straightforward in practice, it is, nonetheless,
9652 straightforward in theory. The atom @samp{rose}, for example, is
9653 recorded as the four contiguous letters @samp{r}, @samp{o}, @samp{s},
9654 @samp{e}. A list, on the other hand, is kept differently. The mechanism
9655 is equally simple, but it takes a moment to get used to the idea. A
9656 list is kept using a series of pairs of pointers. In the series, the
9657 first pointer in each pair points to an atom or to another list, and the
9658 second pointer in each pair points to the next pair, or to the symbol
9659 @code{nil}, which marks the end of the list.
9661 A pointer itself is quite simply the electronic address of what is
9662 pointed to. Hence, a list is kept as a series of electronic addresses.
9665 * Lists diagrammed::
9666 * Symbols as Chest::
9670 @node Lists diagrammed, Symbols as Chest, List Implementation, List Implementation
9672 @unnumberedsec Lists diagrammed
9675 For example, the list @code{(rose violet buttercup)} has three elements,
9676 @samp{rose}, @samp{violet}, and @samp{buttercup}. In the computer, the
9677 electronic address of @samp{rose} is recorded in a segment of computer
9678 memory along with the address that gives the electronic address of where
9679 the atom @samp{violet} is located; and that address (the one that tells
9680 where @samp{violet} is located) is kept along with an address that tells
9681 where the address for the atom @samp{buttercup} is located.
9684 This sounds more complicated than it is and is easier seen in a diagram:
9686 @c clear print-postscript-figures
9687 @c !!! cons-cell-diagram #1
9691 ___ ___ ___ ___ ___ ___
9692 |___|___|--> |___|___|--> |___|___|--> nil
9695 --> rose --> violet --> buttercup
9699 @ifset print-postscript-figures
9702 @center @image{cons-1}
9703 %%%% old method of including an image
9704 % \input /usr/local/lib/tex/inputs/psfig.tex
9705 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-1.eps}}
9710 @ifclear print-postscript-figures
9714 ___ ___ ___ ___ ___ ___
9715 |___|___|--> |___|___|--> |___|___|--> nil
9718 --> rose --> violet --> buttercup
9725 In the diagram, each box represents a word of computer memory that
9726 holds a Lisp object, usually in the form of a memory address. The boxes,
9727 i.e.@: the addresses, are in pairs. Each arrow points to what the address
9728 is the address of, either an atom or another pair of addresses. The
9729 first box is the electronic address of @samp{rose} and the arrow points
9730 to @samp{rose}; the second box is the address of the next pair of boxes,
9731 the first part of which is the address of @samp{violet} and the second
9732 part of which is the address of the next pair. The very last box
9733 points to the symbol @code{nil}, which marks the end of the list.
9736 When a variable is set to a list with a function such as @code{setq},
9737 it stores the address of the first box in the variable. Thus,
9738 evaluation of the expression
9741 (setq bouquet '(rose violet buttercup))
9746 creates a situation like this:
9748 @c cons-cell-diagram #2
9754 | ___ ___ ___ ___ ___ ___
9755 --> |___|___|--> |___|___|--> |___|___|--> nil
9758 --> rose --> violet --> buttercup
9762 @ifset print-postscript-figures
9765 @center @image{cons-2}
9766 %%%% old method of including an image
9767 % \input /usr/local/lib/tex/inputs/psfig.tex
9768 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-2.eps}}
9773 @ifclear print-postscript-figures
9779 | ___ ___ ___ ___ ___ ___
9780 --> |___|___|--> |___|___|--> |___|___|--> nil
9783 --> rose --> violet --> buttercup
9790 In this example, the symbol @code{bouquet} holds the address of the first
9794 This same list can be illustrated in a different sort of box notation
9797 @c cons-cell-diagram #2a
9803 | -------------- --------------- ----------------
9804 | | car | cdr | | car | cdr | | car | cdr |
9805 -->| rose | o------->| violet | o------->| butter- | nil |
9806 | | | | | | | cup | |
9807 -------------- --------------- ----------------
9811 @ifset print-postscript-figures
9814 @center @image{cons-2a}
9815 %%%% old method of including an image
9816 % \input /usr/local/lib/tex/inputs/psfig.tex
9817 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-2a.eps}}
9822 @ifclear print-postscript-figures
9828 | -------------- --------------- ----------------
9829 | | car | cdr | | car | cdr | | car | cdr |
9830 -->| rose | o------->| violet | o------->| butter- | nil |
9831 | | | | | | | cup | |
9832 -------------- --------------- ----------------
9838 (Symbols consist of more than pairs of addresses, but the structure of
9839 a symbol is made up of addresses. Indeed, the symbol @code{bouquet}
9840 consists of a group of address-boxes, one of which is the address of
9841 the printed word @samp{bouquet}, a second of which is the address of a
9842 function definition attached to the symbol, if any, a third of which
9843 is the address of the first pair of address-boxes for the list
9844 @code{(rose violet buttercup)}, and so on. Here we are showing that
9845 the symbol's third address-box points to the first pair of
9846 address-boxes for the list.)
9848 If a symbol is set to the @sc{cdr} of a list, the list itself is not
9849 changed; the symbol simply has an address further down the list. (In
9850 the jargon, @sc{car} and @sc{cdr} are `non-destructive'.) Thus,
9851 evaluation of the following expression
9854 (setq flowers (cdr bouquet))
9861 @c cons-cell-diagram #3
9868 | ___ ___ | ___ ___ ___ ___
9869 --> | | | --> | | | | | |
9870 |___|___|----> |___|___|--> |___|___|--> nil
9873 --> rose --> violet --> buttercup
9878 @ifset print-postscript-figures
9881 @center @image{cons-3}
9882 %%%% old method of including an image
9883 % \input /usr/local/lib/tex/inputs/psfig.tex
9884 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-3.eps}}
9889 @ifclear print-postscript-figures
9896 | ___ ___ | ___ ___ ___ ___
9897 --> | | | --> | | | | | |
9898 |___|___|----> |___|___|--> |___|___|--> nil
9901 --> rose --> violet --> buttercup
9909 The value of @code{flowers} is @code{(violet buttercup)}, which is
9910 to say, the symbol @code{flowers} holds the address of the pair of
9911 address-boxes, the first of which holds the address of @code{violet},
9912 and the second of which holds the address of @code{buttercup}.
9914 A pair of address-boxes is called a @dfn{cons cell} or @dfn{dotted
9915 pair}. @xref{Cons Cell Type, , Cons Cell and List Types, elisp, The GNU Emacs Lisp
9916 Reference Manual}, and @ref{Dotted Pair Notation, , Dotted Pair
9917 Notation, elisp, The GNU Emacs Lisp Reference Manual}, for more
9918 information about cons cells and dotted pairs.
9921 The function @code{cons} adds a new pair of addresses to the front of
9922 a series of addresses like that shown above. For example, evaluating
9926 (setq bouquet (cons 'lily bouquet))
9933 @c cons-cell-diagram #4
9940 | ___ ___ ___ ___ | ___ ___ ___ ___
9941 --> | | | | | | --> | | | | | |
9942 |___|___|----> |___|___|----> |___|___|---->|___|___|--> nil
9945 --> lily --> rose --> violet --> buttercup
9950 @ifset print-postscript-figures
9953 @center @image{cons-4}
9954 %%%% old method of including an image
9955 % \input /usr/local/lib/tex/inputs/psfig.tex
9956 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-4.eps}}
9961 @ifclear print-postscript-figures
9968 | ___ ___ ___ ___ | ___ ___ ___ ___
9969 --> | | | | | | --> | | | | | |
9970 |___|___|----> |___|___|----> |___|___|---->|___|___|--> nil
9973 --> lily --> rose --> violet --> buttercup
9982 However, this does not change the value of the symbol
9983 @code{flowers}, as you can see by evaluating the following,
9986 (eq (cdr (cdr bouquet)) flowers)
9990 which returns @code{t} for true.
9992 Until it is reset, @code{flowers} still has the value
9993 @code{(violet buttercup)}; that is, it has the address of the cons
9994 cell whose first address is of @code{violet}. Also, this does not
9995 alter any of the pre-existing cons cells; they are all still there.
9997 Thus, in Lisp, to get the @sc{cdr} of a list, you just get the address
9998 of the next cons cell in the series; to get the @sc{car} of a list,
9999 you get the address of the first element of the list; to @code{cons} a
10000 new element on a list, you add a new cons cell to the front of the list.
10001 That is all there is to it! The underlying structure of Lisp is
10002 brilliantly simple!
10004 And what does the last address in a series of cons cells refer to? It
10005 is the address of the empty list, of @code{nil}.
10007 In summary, when a Lisp variable is set to a value, it is provided with
10008 the address of the list to which the variable refers.
10010 @node Symbols as Chest, List Exercise, Lists diagrammed, List Implementation
10011 @section Symbols as a Chest of Drawers
10012 @cindex Symbols as a Chest of Drawers
10013 @cindex Chest of Drawers, metaphor for a symbol
10014 @cindex Drawers, Chest of, metaphor for a symbol
10016 In an earlier section, I suggested that you might imagine a symbol as
10017 being a chest of drawers. The function definition is put in one
10018 drawer, the value in another, and so on. What is put in the drawer
10019 holding the value can be changed without affecting the contents of the
10020 drawer holding the function definition, and vice-verse.
10022 Actually, what is put in each drawer is the address of the value or
10023 function definition. It is as if you found an old chest in the attic,
10024 and in one of its drawers you found a map giving you directions to
10025 where the buried treasure lies.
10027 (In addition to its name, symbol definition, and variable value, a
10028 symbol has a `drawer' for a @dfn{property list} which can be used to
10029 record other information. Property lists are not discussed here; see
10030 @ref{Property Lists, , Property Lists, elisp, The GNU Emacs Lisp
10031 Reference Manual}.)
10034 Here is a fanciful representation:
10036 @c chest-of-drawers diagram
10041 Chest of Drawers Contents of Drawers
10045 ---------------------
10046 | directions to | [map to]
10047 | symbol name | bouquet
10049 +---------------------+
10051 | symbol definition | [none]
10053 +---------------------+
10054 | directions to | [map to]
10055 | variable value | (rose violet buttercup)
10057 +---------------------+
10059 | property list | [not described here]
10061 +---------------------+
10067 @ifset print-postscript-figures
10070 @center @image{drawers}
10071 %%%% old method of including an image
10072 % \input /usr/local/lib/tex/inputs/psfig.tex
10073 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/drawers.eps}}
10078 @ifclear print-postscript-figures
10083 Chest of Drawers Contents of Drawers
10087 ---------------------
10088 | directions to | [map to]
10089 | symbol name | bouquet
10091 +---------------------+
10093 | symbol definition | [none]
10095 +---------------------+
10096 | directions to | [map to]
10097 | variable value | (rose violet buttercup)
10099 +---------------------+
10101 | property list | [not described here]
10103 +---------------------+
10111 @node List Exercise, , Symbols as Chest, List Implementation
10114 Set @code{flowers} to @code{violet} and @code{buttercup}. Cons two
10115 more flowers on to this list and set this new list to
10116 @code{more-flowers}. Set the @sc{car} of @code{flowers} to a fish.
10117 What does the @code{more-flowers} list now contain?
10119 @node Yanking, Loops & Recursion, List Implementation, Top
10120 @comment node-name, next, previous, up
10121 @chapter Yanking Text Back
10123 @cindex Text retrieval
10124 @cindex Retrieving text
10125 @cindex Pasting text
10127 Whenever you cut text out of a buffer with a `kill' command in GNU Emacs,
10128 you can bring it back with a `yank' command. The text that is cut out of
10129 the buffer is put in the kill ring and the yank commands insert the
10130 appropriate contents of the kill ring back into a buffer (not necessarily
10131 the original buffer).
10133 A simple @kbd{C-y} (@code{yank}) command inserts the first item from
10134 the kill ring into the current buffer. If the @kbd{C-y} command is
10135 followed immediately by @kbd{M-y}, the first element is replaced by
10136 the second element. Successive @kbd{M-y} commands replace the second
10137 element with the third, fourth, or fifth element, and so on. When the
10138 last element in the kill ring is reached, it is replaced by the first
10139 element and the cycle is repeated. (Thus the kill ring is called a
10140 `ring' rather than just a `list'. However, the actual data structure
10141 that holds the text is a list.
10142 @xref{Kill Ring, , Handling the Kill Ring}, for the details of how the
10143 list is handled as a ring.)
10146 * Kill Ring Overview::
10147 * kill-ring-yank-pointer::
10148 * yank nthcdr Exercises::
10151 @node Kill Ring Overview, kill-ring-yank-pointer, Yanking, Yanking
10152 @comment node-name, next, previous, up
10153 @section Kill Ring Overview
10154 @cindex Kill ring overview
10156 The kill ring is a list of textual strings. This is what it looks like:
10159 ("some text" "a different piece of text" "yet more text")
10162 If this were the contents of my kill ring and I pressed @kbd{C-y}, the
10163 string of characters saying @samp{some text} would be inserted in this
10164 buffer where my cursor is located.
10166 The @code{yank} command is also used for duplicating text by copying it.
10167 The copied text is not cut from the buffer, but a copy of it is put on the
10168 kill ring and is inserted by yanking it back.
10170 Three functions are used for bringing text back from the kill ring:
10171 @code{yank}, which is usually bound to @kbd{C-y}; @code{yank-pop},
10172 which is usually bound to @kbd{M-y}; and @code{rotate-yank-pointer},
10173 which is used by the two other functions.
10175 These functions refer to the kill ring through a variable called the
10176 @code{kill-ring-yank-pointer}. Indeed, the insertion code for both the
10177 @code{yank} and @code{yank-pop} functions is:
10180 (insert (car kill-ring-yank-pointer))
10184 (Well, no more. In GNU Emacs 22, the function has been replaced by
10185 @code{insert-for-yank} which calls @code{insert-for-yank-1}
10186 repetitively for each @code{yank-handler} segment. In turn,
10187 @code{insert-for-yank-1} strips text properties from the inserted text
10188 according to @code{yank-excluded-properties}. Otherwise, it is just
10189 like @code{insert}. We will stick with plain @code{insert} since it
10190 is easier to understand.)
10192 To begin to understand how @code{yank} and @code{yank-pop} work, it is
10193 first necessary to look at the @code{kill-ring-yank-pointer} variable.
10195 @node kill-ring-yank-pointer, yank nthcdr Exercises, Kill Ring Overview, Yanking
10196 @comment node-name, next, previous, up
10197 @section The @code{kill-ring-yank-pointer} Variable
10199 @code{kill-ring-yank-pointer} is a variable, just as @code{kill-ring} is
10200 a variable. It points to something by being bound to the value of what
10201 it points to, like any other Lisp variable.
10204 Thus, if the value of the kill ring is:
10207 ("some text" "a different piece of text" "yet more text")
10212 and the @code{kill-ring-yank-pointer} points to the second clause, the
10213 value of @code{kill-ring-yank-pointer} is:
10216 ("a different piece of text" "yet more text")
10219 As explained in the previous chapter (@pxref{List Implementation}), the
10220 computer does not keep two different copies of the text being pointed to
10221 by both the @code{kill-ring} and the @code{kill-ring-yank-pointer}. The
10222 words ``a different piece of text'' and ``yet more text'' are not
10223 duplicated. Instead, the two Lisp variables point to the same pieces of
10224 text. Here is a diagram:
10226 @c cons-cell-diagram #5
10230 kill-ring kill-ring-yank-pointer
10232 | ___ ___ | ___ ___ ___ ___
10233 ---> | | | --> | | | | | |
10234 |___|___|----> |___|___|--> |___|___|--> nil
10237 | | --> "yet more text"
10239 | --> "a different piece of text"
10246 @ifset print-postscript-figures
10249 @center @image{cons-5}
10250 %%%% old method of including an image
10251 % \input /usr/local/lib/tex/inputs/psfig.tex
10252 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-5.eps}}
10257 @ifclear print-postscript-figures
10261 kill-ring kill-ring-yank-pointer
10263 | ___ ___ | ___ ___ ___ ___
10264 ---> | | | --> | | | | | |
10265 |___|___|----> |___|___|--> |___|___|--> nil
10268 | | --> "yet more text"
10270 | --> "a different piece of text
10279 Both the variable @code{kill-ring} and the variable
10280 @code{kill-ring-yank-pointer} are pointers. But the kill ring itself is
10281 usually described as if it were actually what it is composed of. The
10282 @code{kill-ring} is spoken of as if it were the list rather than that it
10283 points to the list. Conversely, the @code{kill-ring-yank-pointer} is
10284 spoken of as pointing to a list.
10286 These two ways of talking about the same thing sound confusing at first but
10287 make sense on reflection. The kill ring is generally thought of as the
10288 complete structure of data that holds the information of what has recently
10289 been cut out of the Emacs buffers. The @code{kill-ring-yank-pointer}
10290 on the other hand, serves to indicate---that is, to `point to'---that part
10291 of the kill ring of which the first element (the @sc{car}) will be
10295 In GNU Emacs 22, the @code{kill-new} function calls
10297 @code{(setq kill-ring-yank-pointer kill-ring)}
10299 (defun rotate-yank-pointer (arg)
10300 "Rotate the yanking point in the kill ring.
10301 With argument, rotate that many kills forward (or backward, if negative)."
10303 (current-kill arg))
10305 (defun current-kill (n &optional do-not-move)
10306 "Rotate the yanking point by N places, and then return that kill.
10307 If N is zero, `interprogram-paste-function' is set, and calling it
10308 returns a string, then that string is added to the front of the
10309 kill ring and returned as the latest kill.
10310 If optional arg DO-NOT-MOVE is non-nil, then don't actually move the
10311 yanking point; just return the Nth kill forward."
10312 (let ((interprogram-paste (and (= n 0)
10313 interprogram-paste-function
10314 (funcall interprogram-paste-function))))
10315 (if interprogram-paste
10317 ;; Disable the interprogram cut function when we add the new
10318 ;; text to the kill ring, so Emacs doesn't try to own the
10319 ;; selection, with identical text.
10320 (let ((interprogram-cut-function nil))
10321 (kill-new interprogram-paste))
10322 interprogram-paste)
10323 (or kill-ring (error "Kill ring is empty"))
10324 (let ((ARGth-kill-element
10325 (nthcdr (mod (- n (length kill-ring-yank-pointer))
10326 (length kill-ring))
10329 (setq kill-ring-yank-pointer ARGth-kill-element))
10330 (car ARGth-kill-element)))))
10335 @node yank nthcdr Exercises, , kill-ring-yank-pointer, Yanking
10336 @section Exercises with @code{yank} and @code{nthcdr}
10340 Using @kbd{C-h v} (@code{describe-variable}), look at the value of
10341 your kill ring. Add several items to your kill ring; look at its
10342 value again. Using @kbd{M-y} (@code{yank-pop)}, move all the way
10343 around the kill ring. How many items were in your kill ring? Find
10344 the value of @code{kill-ring-max}. Was your kill ring full, or could
10345 you have kept more blocks of text within it?
10348 Using @code{nthcdr} and @code{car}, construct a series of expressions
10349 to return the first, second, third, and fourth elements of a list.
10352 @node Loops & Recursion, Regexp Search, Yanking, Top
10353 @comment node-name, next, previous, up
10354 @chapter Loops and Recursion
10355 @cindex Loops and recursion
10356 @cindex Recursion and loops
10357 @cindex Repetition (loops)
10359 Emacs Lisp has two primary ways to cause an expression, or a series of
10360 expressions, to be evaluated repeatedly: one uses a @code{while}
10361 loop, and the other uses @dfn{recursion}.
10363 Repetition can be very valuable. For example, to move forward four
10364 sentences, you need only write a program that will move forward one
10365 sentence and then repeat the process four times. Since a computer does
10366 not get bored or tired, such repetitive action does not have the
10367 deleterious effects that excessive or the wrong kinds of repetition can
10370 People mostly write Emacs Lisp functions using @code{while} loops and
10371 their kin; but you can use recursion, which provides a very powerful
10372 way to think about and then to solve problems@footnote{You can write
10373 recursive functions to be frugal or wasteful of mental or computer
10374 resources; as it happens, methods that people find easy---that are
10375 frugal of `mental resources'---sometimes use considerable computer
10376 resources. Emacs was designed to run on machines that we now consider
10377 limited and its default settings are conservative. You may want to
10378 increase the values of @code{max-specpdl-size} and
10379 @code{max-lisp-eval-depth}. In my @file{.emacs} file, I set them to
10380 15 and 30 times their default value.}.
10386 * Looping exercise::
10389 @node while, dolist dotimes, Loops & Recursion, Loops & Recursion
10390 @comment node-name, next, previous, up
10391 @section @code{while}
10395 The @code{while} special form tests whether the value returned by
10396 evaluating its first argument is true or false. This is similar to what
10397 the Lisp interpreter does with an @code{if}; what the interpreter does
10398 next, however, is different.
10400 In a @code{while} expression, if the value returned by evaluating the
10401 first argument is false, the Lisp interpreter skips the rest of the
10402 expression (the @dfn{body} of the expression) and does not evaluate it.
10403 However, if the value is true, the Lisp interpreter evaluates the body
10404 of the expression and then again tests whether the first argument to
10405 @code{while} is true or false. If the value returned by evaluating the
10406 first argument is again true, the Lisp interpreter again evaluates the
10407 body of the expression.
10410 The template for a @code{while} expression looks like this:
10414 (while @var{true-or-false-test}
10420 * Looping with while::
10422 * print-elements-of-list::
10423 * Incrementing Loop::
10424 * Incrementing Loop Details::
10425 * Decrementing Loop::
10428 @node Looping with while, Loop Example, while, while
10430 @unnumberedsubsec Looping with @code{while}
10433 So long as the true-or-false-test of the @code{while} expression
10434 returns a true value when it is evaluated, the body is repeatedly
10435 evaluated. This process is called a loop since the Lisp interpreter
10436 repeats the same thing again and again, like an airplane doing a loop.
10437 When the result of evaluating the true-or-false-test is false, the
10438 Lisp interpreter does not evaluate the rest of the @code{while}
10439 expression and `exits the loop'.
10441 Clearly, if the value returned by evaluating the first argument to
10442 @code{while} is always true, the body following will be evaluated
10443 again and again @dots{} and again @dots{} forever. Conversely, if the
10444 value returned is never true, the expressions in the body will never
10445 be evaluated. The craft of writing a @code{while} loop consists of
10446 choosing a mechanism such that the true-or-false-test returns true
10447 just the number of times that you want the subsequent expressions to
10448 be evaluated, and then have the test return false.
10450 The value returned by evaluating a @code{while} is the value of the
10451 true-or-false-test. An interesting consequence of this is that a
10452 @code{while} loop that evaluates without error will return @code{nil}
10453 or false regardless of whether it has looped 1 or 100 times or none at
10454 all. A @code{while} expression that evaluates successfully never
10455 returns a true value! What this means is that @code{while} is always
10456 evaluated for its side effects, which is to say, the consequences of
10457 evaluating the expressions within the body of the @code{while} loop.
10458 This makes sense. It is not the mere act of looping that is desired,
10459 but the consequences of what happens when the expressions in the loop
10460 are repeatedly evaluated.
10462 @node Loop Example, print-elements-of-list, Looping with while, while
10463 @comment node-name, next, previous, up
10464 @subsection A @code{while} Loop and a List
10466 A common way to control a @code{while} loop is to test whether a list
10467 has any elements. If it does, the loop is repeated; but if it does not,
10468 the repetition is ended. Since this is an important technique, we will
10469 create a short example to illustrate it.
10471 A simple way to test whether a list has elements is to evaluate the
10472 list: if it has no elements, it is an empty list and will return the
10473 empty list, @code{()}, which is a synonym for @code{nil} or false. On
10474 the other hand, a list with elements will return those elements when it
10475 is evaluated. Since Emacs Lisp considers as true any value that is not
10476 @code{nil}, a list that returns elements will test true in a
10480 For example, you can set the variable @code{empty-list} to @code{nil} by
10481 evaluating the following @code{setq} expression:
10484 (setq empty-list ())
10488 After evaluating the @code{setq} expression, you can evaluate the
10489 variable @code{empty-list} in the usual way, by placing the cursor after
10490 the symbol and typing @kbd{C-x C-e}; @code{nil} will appear in your
10497 On the other hand, if you set a variable to be a list with elements, the
10498 list will appear when you evaluate the variable, as you can see by
10499 evaluating the following two expressions:
10503 (setq animals '(gazelle giraffe lion tiger))
10509 Thus, to create a @code{while} loop that tests whether there are any
10510 items in the list @code{animals}, the first part of the loop will be
10521 When the @code{while} tests its first argument, the variable
10522 @code{animals} is evaluated. It returns a list. So long as the list
10523 has elements, the @code{while} considers the results of the test to be
10524 true; but when the list is empty, it considers the results of the test
10527 To prevent the @code{while} loop from running forever, some mechanism
10528 needs to be provided to empty the list eventually. An oft-used
10529 technique is to have one of the subsequent forms in the @code{while}
10530 expression set the value of the list to be the @sc{cdr} of the list.
10531 Each time the @code{cdr} function is evaluated, the list will be made
10532 shorter, until eventually only the empty list will be left. At this
10533 point, the test of the @code{while} loop will return false, and the
10534 arguments to the @code{while} will no longer be evaluated.
10536 For example, the list of animals bound to the variable @code{animals}
10537 can be set to be the @sc{cdr} of the original list with the
10538 following expression:
10541 (setq animals (cdr animals))
10545 If you have evaluated the previous expressions and then evaluate this
10546 expression, you will see @code{(giraffe lion tiger)} appear in the echo
10547 area. If you evaluate the expression again, @code{(lion tiger)} will
10548 appear in the echo area. If you evaluate it again and yet again,
10549 @code{(tiger)} appears and then the empty list, shown by @code{nil}.
10551 A template for a @code{while} loop that uses the @code{cdr} function
10552 repeatedly to cause the true-or-false-test eventually to test false
10557 (while @var{test-whether-list-is-empty}
10559 @var{set-list-to-cdr-of-list})
10563 This test and use of @code{cdr} can be put together in a function that
10564 goes through a list and prints each element of the list on a line of its
10567 @node print-elements-of-list, Incrementing Loop, Loop Example, while
10568 @subsection An Example: @code{print-elements-of-list}
10569 @findex print-elements-of-list
10571 The @code{print-elements-of-list} function illustrates a @code{while}
10574 @cindex @file{*scratch*} buffer
10575 The function requires several lines for its output. If you are
10576 reading this in a recent instance of GNU Emacs,
10577 @c GNU Emacs 21, GNU Emacs 22, or a later version,
10578 you can evaluate the following expression inside of Info, as usual.
10580 If you are using an earlier version of Emacs, you need to copy the
10581 necessary expressions to your @file{*scratch*} buffer and evaluate
10582 them there. This is because the echo area had only one line in the
10585 You can copy the expressions by marking the beginning of the region
10586 with @kbd{C-@key{SPC}} (@code{set-mark-command}), moving the cursor to
10587 the end of the region and then copying the region using @kbd{M-w}
10588 (@code{kill-ring-save}, which calls @code{copy-region-as-kill} and
10589 then provides visual feedback). In the @file{*scratch*}
10590 buffer, you can yank the expressions back by typing @kbd{C-y}
10593 After you have copied the expressions to the @file{*scratch*} buffer,
10594 evaluate each expression in turn. Be sure to evaluate the last
10595 expression, @code{(print-elements-of-list animals)}, by typing
10596 @kbd{C-u C-x C-e}, that is, by giving an argument to
10597 @code{eval-last-sexp}. This will cause the result of the evaluation
10598 to be printed in the @file{*scratch*} buffer instead of being printed
10599 in the echo area. (Otherwise you will see something like this in your
10600 echo area: @code{^Jgazelle^J^Jgiraffe^J^Jlion^J^Jtiger^Jnil}, in which
10601 each @samp{^J} stands for a `newline'.)
10604 In a recent instance of GNU Emacs, you can evaluate these expressions
10605 directly in the Info buffer, and the echo area will grow to show the
10610 (setq animals '(gazelle giraffe lion tiger))
10612 (defun print-elements-of-list (list)
10613 "Print each element of LIST on a line of its own."
10616 (setq list (cdr list))))
10618 (print-elements-of-list animals)
10624 When you evaluate the three expressions in sequence, you will see
10640 Each element of the list is printed on a line of its own (that is what
10641 the function @code{print} does) and then the value returned by the
10642 function is printed. Since the last expression in the function is the
10643 @code{while} loop, and since @code{while} loops always return
10644 @code{nil}, a @code{nil} is printed after the last element of the list.
10646 @node Incrementing Loop, Incrementing Loop Details, print-elements-of-list, while
10647 @comment node-name, next, previous, up
10648 @subsection A Loop with an Incrementing Counter
10650 A loop is not useful unless it stops when it ought. Besides
10651 controlling a loop with a list, a common way of stopping a loop is to
10652 write the first argument as a test that returns false when the correct
10653 number of repetitions are complete. This means that the loop must
10654 have a counter---an expression that counts how many times the loop
10657 @node Incrementing Loop Details, Decrementing Loop, Incrementing Loop, while
10659 @unnumberedsubsec Details of an Incrementing Loop
10662 The test for a loop with an incrementing counter can be an expression
10663 such as @code{(< count desired-number)} which returns @code{t} for
10664 true if the value of @code{count} is less than the
10665 @code{desired-number} of repetitions and @code{nil} for false if the
10666 value of @code{count} is equal to or is greater than the
10667 @code{desired-number}. The expression that increments the count can
10668 be a simple @code{setq} such as @code{(setq count (1+ count))}, where
10669 @code{1+} is a built-in function in Emacs Lisp that adds 1 to its
10670 argument. (The expression @w{@code{(1+ count)}} has the same result
10671 as @w{@code{(+ count 1)}}, but is easier for a human to read.)
10674 The template for a @code{while} loop controlled by an incrementing
10675 counter looks like this:
10679 @var{set-count-to-initial-value}
10680 (while (< count desired-number) ; @r{true-or-false-test}
10682 (setq count (1+ count))) ; @r{incrementer}
10687 Note that you need to set the initial value of @code{count}; usually it
10691 * Incrementing Example::
10692 * Inc Example parts::
10693 * Inc Example altogether::
10696 @node Incrementing Example, Inc Example parts, Incrementing Loop Details, Incrementing Loop Details
10697 @unnumberedsubsubsec Example with incrementing counter
10699 Suppose you are playing on the beach and decide to make a triangle of
10700 pebbles, putting one pebble in the first row, two in the second row,
10701 three in the third row and so on, like this:
10719 @bullet{} @bullet{}
10720 @bullet{} @bullet{} @bullet{}
10721 @bullet{} @bullet{} @bullet{} @bullet{}
10728 (About 2500 years ago, Pythagoras and others developed the beginnings of
10729 number theory by considering questions such as this.)
10731 Suppose you want to know how many pebbles you will need to make a
10732 triangle with 7 rows?
10734 Clearly, what you need to do is add up the numbers from 1 to 7. There
10735 are two ways to do this; start with the smallest number, one, and add up
10736 the list in sequence, 1, 2, 3, 4 and so on; or start with the largest
10737 number and add the list going down: 7, 6, 5, 4 and so on. Because both
10738 mechanisms illustrate common ways of writing @code{while} loops, we will
10739 create two examples, one counting up and the other counting down. In
10740 this first example, we will start with 1 and add 2, 3, 4 and so on.
10742 If you are just adding up a short list of numbers, the easiest way to do
10743 it is to add up all the numbers at once. However, if you do not know
10744 ahead of time how many numbers your list will have, or if you want to be
10745 prepared for a very long list, then you need to design your addition so
10746 that what you do is repeat a simple process many times instead of doing
10747 a more complex process once.
10749 For example, instead of adding up all the pebbles all at once, what you
10750 can do is add the number of pebbles in the first row, 1, to the number
10751 in the second row, 2, and then add the total of those two rows to the
10752 third row, 3. Then you can add the number in the fourth row, 4, to the
10753 total of the first three rows; and so on.
10755 The critical characteristic of the process is that each repetitive
10756 action is simple. In this case, at each step we add only two numbers,
10757 the number of pebbles in the row and the total already found. This
10758 process of adding two numbers is repeated again and again until the last
10759 row has been added to the total of all the preceding rows. In a more
10760 complex loop the repetitive action might not be so simple, but it will
10761 be simpler than doing everything all at once.
10763 @node Inc Example parts, Inc Example altogether, Incrementing Example, Incrementing Loop Details
10764 @unnumberedsubsubsec The parts of the function definition
10766 The preceding analysis gives us the bones of our function definition:
10767 first, we will need a variable that we can call @code{total} that will
10768 be the total number of pebbles. This will be the value returned by
10771 Second, we know that the function will require an argument: this
10772 argument will be the total number of rows in the triangle. It can be
10773 called @code{number-of-rows}.
10775 Finally, we need a variable to use as a counter. We could call this
10776 variable @code{counter}, but a better name is @code{row-number}. That
10777 is because what the counter does in this function is count rows, and a
10778 program should be written to be as understandable as possible.
10780 When the Lisp interpreter first starts evaluating the expressions in the
10781 function, the value of @code{total} should be set to zero, since we have
10782 not added anything to it. Then the function should add the number of
10783 pebbles in the first row to the total, and then add the number of
10784 pebbles in the second to the total, and then add the number of
10785 pebbles in the third row to the total, and so on, until there are no
10786 more rows left to add.
10788 Both @code{total} and @code{row-number} are used only inside the
10789 function, so they can be declared as local variables with @code{let}
10790 and given initial values. Clearly, the initial value for @code{total}
10791 should be 0. The initial value of @code{row-number} should be 1,
10792 since we start with the first row. This means that the @code{let}
10793 statement will look like this:
10803 After the internal variables are declared and bound to their initial
10804 values, we can begin the @code{while} loop. The expression that serves
10805 as the test should return a value of @code{t} for true so long as the
10806 @code{row-number} is less than or equal to the @code{number-of-rows}.
10807 (If the expression tests true only so long as the row number is less
10808 than the number of rows in the triangle, the last row will never be
10809 added to the total; hence the row number has to be either less than or
10810 equal to the number of rows.)
10813 @findex <= @r{(less than or equal)}
10814 Lisp provides the @code{<=} function that returns true if the value of
10815 its first argument is less than or equal to the value of its second
10816 argument and false otherwise. So the expression that the @code{while}
10817 will evaluate as its test should look like this:
10820 (<= row-number number-of-rows)
10823 The total number of pebbles can be found by repeatedly adding the number
10824 of pebbles in a row to the total already found. Since the number of
10825 pebbles in the row is equal to the row number, the total can be found by
10826 adding the row number to the total. (Clearly, in a more complex
10827 situation, the number of pebbles in the row might be related to the row
10828 number in a more complicated way; if this were the case, the row number
10829 would be replaced by the appropriate expression.)
10832 (setq total (+ total row-number))
10836 What this does is set the new value of @code{total} to be equal to the
10837 sum of adding the number of pebbles in the row to the previous total.
10839 After setting the value of @code{total}, the conditions need to be
10840 established for the next repetition of the loop, if there is one. This
10841 is done by incrementing the value of the @code{row-number} variable,
10842 which serves as a counter. After the @code{row-number} variable has
10843 been incremented, the true-or-false-test at the beginning of the
10844 @code{while} loop tests whether its value is still less than or equal to
10845 the value of the @code{number-of-rows} and if it is, adds the new value
10846 of the @code{row-number} variable to the @code{total} of the previous
10847 repetition of the loop.
10850 The built-in Emacs Lisp function @code{1+} adds 1 to a number, so the
10851 @code{row-number} variable can be incremented with this expression:
10854 (setq row-number (1+ row-number))
10857 @node Inc Example altogether, , Inc Example parts, Incrementing Loop Details
10858 @unnumberedsubsubsec Putting the function definition together
10860 We have created the parts for the function definition; now we need to
10864 First, the contents of the @code{while} expression:
10868 (while (<= row-number number-of-rows) ; @r{true-or-false-test}
10869 (setq total (+ total row-number))
10870 (setq row-number (1+ row-number))) ; @r{incrementer}
10874 Along with the @code{let} expression varlist, this very nearly
10875 completes the body of the function definition. However, it requires
10876 one final element, the need for which is somewhat subtle.
10878 The final touch is to place the variable @code{total} on a line by
10879 itself after the @code{while} expression. Otherwise, the value returned
10880 by the whole function is the value of the last expression that is
10881 evaluated in the body of the @code{let}, and this is the value
10882 returned by the @code{while}, which is always @code{nil}.
10884 This may not be evident at first sight. It almost looks as if the
10885 incrementing expression is the last expression of the whole function.
10886 But that expression is part of the body of the @code{while}; it is the
10887 last element of the list that starts with the symbol @code{while}.
10888 Moreover, the whole of the @code{while} loop is a list within the body
10892 In outline, the function will look like this:
10896 (defun @var{name-of-function} (@var{argument-list})
10897 "@var{documentation}@dots{}"
10898 (let (@var{varlist})
10899 (while (@var{true-or-false-test})
10900 @var{body-of-while}@dots{} )
10901 @dots{} )) ; @r{Need final expression here.}
10905 The result of evaluating the @code{let} is what is going to be returned
10906 by the @code{defun} since the @code{let} is not embedded within any
10907 containing list, except for the @code{defun} as a whole. However, if
10908 the @code{while} is the last element of the @code{let} expression, the
10909 function will always return @code{nil}. This is not what we want!
10910 Instead, what we want is the value of the variable @code{total}. This
10911 is returned by simply placing the symbol as the last element of the list
10912 starting with @code{let}. It gets evaluated after the preceding
10913 elements of the list are evaluated, which means it gets evaluated after
10914 it has been assigned the correct value for the total.
10916 It may be easier to see this by printing the list starting with
10917 @code{let} all on one line. This format makes it evident that the
10918 @var{varlist} and @code{while} expressions are the second and third
10919 elements of the list starting with @code{let}, and the @code{total} is
10924 (let (@var{varlist}) (while (@var{true-or-false-test}) @var{body-of-while}@dots{} ) total)
10929 Putting everything together, the @code{triangle} function definition
10934 (defun triangle (number-of-rows) ; @r{Version with}
10935 ; @r{ incrementing counter.}
10936 "Add up the number of pebbles in a triangle.
10937 The first row has one pebble, the second row two pebbles,
10938 the third row three pebbles, and so on.
10939 The argument is NUMBER-OF-ROWS."
10944 (while (<= row-number number-of-rows)
10945 (setq total (+ total row-number))
10946 (setq row-number (1+ row-number)))
10952 After you have installed @code{triangle} by evaluating the function, you
10953 can try it out. Here are two examples:
10964 The sum of the first four numbers is 10 and the sum of the first seven
10967 @node Decrementing Loop, , Incrementing Loop Details, while
10968 @comment node-name, next, previous, up
10969 @subsection Loop with a Decrementing Counter
10971 Another common way to write a @code{while} loop is to write the test
10972 so that it determines whether a counter is greater than zero. So long
10973 as the counter is greater than zero, the loop is repeated. But when
10974 the counter is equal to or less than zero, the loop is stopped. For
10975 this to work, the counter has to start out greater than zero and then
10976 be made smaller and smaller by a form that is evaluated
10979 The test will be an expression such as @code{(> counter 0)} which
10980 returns @code{t} for true if the value of @code{counter} is greater
10981 than zero, and @code{nil} for false if the value of @code{counter} is
10982 equal to or less than zero. The expression that makes the number
10983 smaller and smaller can be a simple @code{setq} such as @code{(setq
10984 counter (1- counter))}, where @code{1-} is a built-in function in
10985 Emacs Lisp that subtracts 1 from its argument.
10988 The template for a decrementing @code{while} loop looks like this:
10992 (while (> counter 0) ; @r{true-or-false-test}
10994 (setq counter (1- counter))) ; @r{decrementer}
10999 * Decrementing Example::
11000 * Dec Example parts::
11001 * Dec Example altogether::
11004 @node Decrementing Example, Dec Example parts, Decrementing Loop, Decrementing Loop
11005 @unnumberedsubsubsec Example with decrementing counter
11007 To illustrate a loop with a decrementing counter, we will rewrite the
11008 @code{triangle} function so the counter decreases to zero.
11010 This is the reverse of the earlier version of the function. In this
11011 case, to find out how many pebbles are needed to make a triangle with
11012 3 rows, add the number of pebbles in the third row, 3, to the number
11013 in the preceding row, 2, and then add the total of those two rows to
11014 the row that precedes them, which is 1.
11016 Likewise, to find the number of pebbles in a triangle with 7 rows, add
11017 the number of pebbles in the seventh row, 7, to the number in the
11018 preceding row, which is 6, and then add the total of those two rows to
11019 the row that precedes them, which is 5, and so on. As in the previous
11020 example, each addition only involves adding two numbers, the total of
11021 the rows already added up and the number of pebbles in the row that is
11022 being added to the total. This process of adding two numbers is
11023 repeated again and again until there are no more pebbles to add.
11025 We know how many pebbles to start with: the number of pebbles in the
11026 last row is equal to the number of rows. If the triangle has seven
11027 rows, the number of pebbles in the last row is 7. Likewise, we know how
11028 many pebbles are in the preceding row: it is one less than the number in
11031 @node Dec Example parts, Dec Example altogether, Decrementing Example, Decrementing Loop
11032 @unnumberedsubsubsec The parts of the function definition
11034 We start with three variables: the total number of rows in the
11035 triangle; the number of pebbles in a row; and the total number of
11036 pebbles, which is what we want to calculate. These variables can be
11037 named @code{number-of-rows}, @code{number-of-pebbles-in-row}, and
11038 @code{total}, respectively.
11040 Both @code{total} and @code{number-of-pebbles-in-row} are used only
11041 inside the function and are declared with @code{let}. The initial
11042 value of @code{total} should, of course, be zero. However, the
11043 initial value of @code{number-of-pebbles-in-row} should be equal to
11044 the number of rows in the triangle, since the addition will start with
11048 This means that the beginning of the @code{let} expression will look
11054 (number-of-pebbles-in-row number-of-rows))
11059 The total number of pebbles can be found by repeatedly adding the number
11060 of pebbles in a row to the total already found, that is, by repeatedly
11061 evaluating the following expression:
11064 (setq total (+ total number-of-pebbles-in-row))
11068 After the @code{number-of-pebbles-in-row} is added to the @code{total},
11069 the @code{number-of-pebbles-in-row} should be decremented by one, since
11070 the next time the loop repeats, the preceding row will be
11071 added to the total.
11073 The number of pebbles in a preceding row is one less than the number of
11074 pebbles in a row, so the built-in Emacs Lisp function @code{1-} can be
11075 used to compute the number of pebbles in the preceding row. This can be
11076 done with the following expression:
11080 (setq number-of-pebbles-in-row
11081 (1- number-of-pebbles-in-row))
11085 Finally, we know that the @code{while} loop should stop making repeated
11086 additions when there are no pebbles in a row. So the test for
11087 the @code{while} loop is simply:
11090 (while (> number-of-pebbles-in-row 0)
11093 @node Dec Example altogether, , Dec Example parts, Decrementing Loop
11094 @unnumberedsubsubsec Putting the function definition together
11096 We can put these expressions together to create a function definition
11097 that works. However, on examination, we find that one of the local
11098 variables is unneeded!
11101 The function definition looks like this:
11105 ;;; @r{First subtractive version.}
11106 (defun triangle (number-of-rows)
11107 "Add up the number of pebbles in a triangle."
11109 (number-of-pebbles-in-row number-of-rows))
11110 (while (> number-of-pebbles-in-row 0)
11111 (setq total (+ total number-of-pebbles-in-row))
11112 (setq number-of-pebbles-in-row
11113 (1- number-of-pebbles-in-row)))
11118 As written, this function works.
11120 However, we do not need @code{number-of-pebbles-in-row}.
11122 @cindex Argument as local variable
11123 When the @code{triangle} function is evaluated, the symbol
11124 @code{number-of-rows} will be bound to a number, giving it an initial
11125 value. That number can be changed in the body of the function as if
11126 it were a local variable, without any fear that such a change will
11127 effect the value of the variable outside of the function. This is a
11128 very useful characteristic of Lisp; it means that the variable
11129 @code{number-of-rows} can be used anywhere in the function where
11130 @code{number-of-pebbles-in-row} is used.
11133 Here is a second version of the function written a bit more cleanly:
11137 (defun triangle (number) ; @r{Second version.}
11138 "Return sum of numbers 1 through NUMBER inclusive."
11140 (while (> number 0)
11141 (setq total (+ total number))
11142 (setq number (1- number)))
11147 In brief, a properly written @code{while} loop will consist of three parts:
11151 A test that will return false after the loop has repeated itself the
11152 correct number of times.
11155 An expression the evaluation of which will return the value desired
11156 after being repeatedly evaluated.
11159 An expression to change the value passed to the true-or-false-test so
11160 that the test returns false after the loop has repeated itself the right
11164 @node dolist dotimes, Recursion, while, Loops & Recursion
11165 @comment node-name, next, previous, up
11166 @section Save your time: @code{dolist} and @code{dotimes}
11168 In addition to @code{while}, both @code{dolist} and @code{dotimes}
11169 provide for looping. Sometimes these are quicker to write than the
11170 equivalent @code{while} loop. Both are Lisp macros. (@xref{Macros, ,
11171 Macros, elisp, The GNU Emacs Lisp Reference Manual}. )
11173 @code{dolist} works like a @code{while} loop that `@sc{cdr}s down a
11174 list': @code{dolist} automatically shortens the list each time it
11175 loops---takes the @sc{cdr} of the list---and binds the @sc{car} of
11176 each shorter version of the list to the first of its arguments.
11178 @code{dotimes} loops a specific number of times: you specify the number.
11185 @node dolist, dotimes, dolist dotimes, dolist dotimes
11186 @unnumberedsubsubsec The @code{dolist} Macro
11189 Suppose, for example, you want to reverse a list, so that
11190 ``first'' ``second'' ``third'' becomes ``third'' ``second'' ``first''.
11193 In practice, you would use the @code{reverse} function, like this:
11197 (setq animals '(gazelle giraffe lion tiger))
11205 Here is how you could reverse the list using a @code{while} loop:
11209 (setq animals '(gazelle giraffe lion tiger))
11211 (defun reverse-list-with-while (list)
11212 "Using while, reverse the order of LIST."
11213 (let (value) ; make sure list starts empty
11215 (setq value (cons (car list) value))
11216 (setq list (cdr list)))
11219 (reverse-list-with-while animals)
11225 And here is how you could use the @code{dolist} macro:
11229 (setq animals '(gazelle giraffe lion tiger))
11231 (defun reverse-list-with-dolist (list)
11232 "Using dolist, reverse the order of LIST."
11233 (let (value) ; make sure list starts empty
11234 (dolist (element list value)
11235 (setq value (cons element value)))))
11237 (reverse-list-with-dolist animals)
11243 In Info, you can place your cursor after the closing parenthesis of
11244 each expression and type @kbd{C-x C-e}; in each case, you should see
11247 (tiger lion giraffe gazelle)
11253 For this example, the existing @code{reverse} function is obviously best.
11254 The @code{while} loop is just like our first example (@pxref{Loop
11255 Example, , A @code{while} Loop and a List}). The @code{while} first
11256 checks whether the list has elements; if so, it constructs a new list
11257 by adding the first element of the list to the existing list (which in
11258 the first iteration of the loop is @code{nil}). Since the second
11259 element is prepended in front of the first element, and the third
11260 element is prepended in front of the second element, the list is reversed.
11262 In the expression using a @code{while} loop,
11263 the @w{@code{(setq list (cdr list))}}
11264 expression shortens the list, so the @code{while} loop eventually
11265 stops. In addition, it provides the @code{cons} expression with a new
11266 first element by creating a new and shorter list at each repetition of
11269 The @code{dolist} expression does very much the same as the
11270 @code{while} expression, except that the @code{dolist} macro does some
11271 of the work you have to do when writing a @code{while} expression.
11273 Like a @code{while} loop, a @code{dolist} loops. What is different is
11274 that it automatically shortens the list each time it loops --- it
11275 `@sc{cdr}s down the list' on its own --- and it automatically binds
11276 the @sc{car} of each shorter version of the list to the first of its
11279 In the example, the @sc{car} of each shorter version of the list is
11280 referred to using the symbol @samp{element}, the list itself is called
11281 @samp{list}, and the value returned is called @samp{value}. The
11282 remainder of the @code{dolist} expression is the body.
11284 The @code{dolist} expression binds the @sc{car} of each shorter
11285 version of the list to @code{element} and then evaluates the body of
11286 the expression; and repeats the loop. The result is returned in
11289 @node dotimes, , dolist, dolist dotimes
11290 @unnumberedsubsubsec The @code{dotimes} Macro
11293 The @code{dotimes} macro is similar to @code{dolist}, except that it
11294 loops a specific number of times.
11296 The first argument to @code{dotimes} is assigned the numbers 0, 1, 2
11297 and so forth each time around the loop, and the value of the third
11298 argument is returned. You need to provide the value of the second
11299 argument, which is how many times the macro loops.
11302 For example, the following binds the numbers from 0 up to, but not
11303 including, the number 3 to the first argument, @var{number}, and then
11304 constructs a list of the three numbers. (The first number is 0, the
11305 second number is 1, and the third number is 2; this makes a total of
11306 three numbers in all, starting with zero as the first number.)
11310 (let (value) ; otherwise a value is a void variable
11311 (dotimes (number 3 value)
11312 (setq value (cons number value))))
11319 @code{dotimes} returns @code{value}, so the way to use
11320 @code{dotimes} is to operate on some expression @var{number} number of
11321 times and then return the result, either as a list or an atom.
11324 Here is an example of a @code{defun} that uses @code{dotimes} to add
11325 up the number of pebbles in a triangle.
11329 (defun triangle-using-dotimes (number-of-rows)
11330 "Using dotimes, add up the number of pebbles in a triangle."
11331 (let ((total 0)) ; otherwise a total is a void variable
11332 (dotimes (number number-of-rows total)
11333 (setq total (+ total (1+ number))))))
11335 (triangle-using-dotimes 4)
11339 @node Recursion, Looping exercise, dolist dotimes, Loops & Recursion
11340 @comment node-name, next, previous, up
11344 A recursive function contains code that tells the Lisp interpreter to
11345 call a program that runs exactly like itself, but with slightly
11346 different arguments. The code runs exactly the same because it has
11347 the same name. However, even though the program has the same name, it
11348 is not the same entity. It is different. In the jargon, it is a
11349 different `instance'.
11351 Eventually, if the program is written correctly, the `slightly
11352 different arguments' will become sufficiently different from the first
11353 arguments that the final instance will stop.
11356 * Building Robots::
11357 * Recursive Definition Parts::
11358 * Recursion with list::
11359 * Recursive triangle function::
11360 * Recursion with cond::
11361 * Recursive Patterns::
11363 * No deferment solution::
11366 @node Building Robots, Recursive Definition Parts, Recursion, Recursion
11367 @comment node-name, next, previous, up
11368 @subsection Building Robots: Extending the Metaphor
11369 @cindex Building robots
11370 @cindex Robots, building
11372 It is sometimes helpful to think of a running program as a robot that
11373 does a job. In doing its job, a recursive function calls on a second
11374 robot to help it. The second robot is identical to the first in every
11375 way, except that the second robot helps the first and has been
11376 passed different arguments than the first.
11378 In a recursive function, the second robot may call a third; and the
11379 third may call a fourth, and so on. Each of these is a different
11380 entity; but all are clones.
11382 Since each robot has slightly different instructions---the arguments
11383 will differ from one robot to the next---the last robot should know
11386 Let's expand on the metaphor in which a computer program is a robot.
11388 A function definition provides the blueprints for a robot. When you
11389 install a function definition, that is, when you evaluate a
11390 @code{defun} special form, you install the necessary equipment to
11391 build robots. It is as if you were in a factory, setting up an
11392 assembly line. Robots with the same name are built according to the
11393 same blueprints. So they have, as it were, the same `model number',
11394 but a different `serial number'.
11396 We often say that a recursive function `calls itself'. What we mean
11397 is that the instructions in a recursive function cause the Lisp
11398 interpreter to run a different function that has the same name and
11399 does the same job as the first, but with different arguments.
11401 It is important that the arguments differ from one instance to the
11402 next; otherwise, the process will never stop.
11404 @node Recursive Definition Parts, Recursion with list, Building Robots, Recursion
11405 @comment node-name, next, previous, up
11406 @subsection The Parts of a Recursive Definition
11407 @cindex Parts of a Recursive Definition
11408 @cindex Recursive Definition Parts
11410 A recursive function typically contains a conditional expression which
11415 A true-or-false-test that determines whether the function is called
11416 again, here called the @dfn{do-again-test}.
11419 The name of the function. When this name is called, a new instance of
11420 the function---a new robot, as it were---is created and told what to do.
11423 An expression that returns a different value each time the function is
11424 called, here called the @dfn{next-step-expression}. Consequently, the
11425 argument (or arguments) passed to the new instance of the function
11426 will be different from that passed to the previous instance. This
11427 causes the conditional expression, the @dfn{do-again-test}, to test
11428 false after the correct number of repetitions.
11431 Recursive functions can be much simpler than any other kind of
11432 function. Indeed, when people first start to use them, they often look
11433 so mysteriously simple as to be incomprehensible. Like riding a
11434 bicycle, reading a recursive function definition takes a certain knack
11435 which is hard at first but then seems simple.
11438 There are several different common recursive patterns. A very simple
11439 pattern looks like this:
11443 (defun @var{name-of-recursive-function} (@var{argument-list})
11444 "@var{documentation}@dots{}"
11445 (if @var{do-again-test}
11447 (@var{name-of-recursive-function}
11448 @var{next-step-expression})))
11452 Each time a recursive function is evaluated, a new instance of it is
11453 created and told what to do. The arguments tell the instance what to do.
11455 An argument is bound to the value of the next-step-expression. Each
11456 instance runs with a different value of the next-step-expression.
11458 The value in the next-step-expression is used in the do-again-test.
11460 The value returned by the next-step-expression is passed to the new
11461 instance of the function, which evaluates it (or some
11462 transmogrification of it) to determine whether to continue or stop.
11463 The next-step-expression is designed so that the do-again-test returns
11464 false when the function should no longer be repeated.
11466 The do-again-test is sometimes called the @dfn{stop condition},
11467 since it stops the repetitions when it tests false.
11469 @node Recursion with list, Recursive triangle function, Recursive Definition Parts, Recursion
11470 @comment node-name, next, previous, up
11471 @subsection Recursion with a List
11473 The example of a @code{while} loop that printed the elements of a list
11474 of numbers can be written recursively. Here is the code, including
11475 an expression to set the value of the variable @code{animals} to a list.
11477 If you are using GNU Emacs 20 or before, this example must be copied
11478 to the @file{*scratch*} buffer and each expression must be evaluated
11479 there. Use @kbd{C-u C-x C-e} to evaluate the
11480 @code{(print-elements-recursively animals)} expression so that the
11481 results are printed in the buffer; otherwise the Lisp interpreter will
11482 try to squeeze the results into the one line of the echo area.
11484 Also, place your cursor immediately after the last closing parenthesis
11485 of the @code{print-elements-recursively} function, before the comment.
11486 Otherwise, the Lisp interpreter will try to evaluate the comment.
11488 If you are using a more recent version of Emacs, you can evaluate this
11489 expression directly in Info.
11491 @findex print-elements-recursively
11494 (setq animals '(gazelle giraffe lion tiger))
11496 (defun print-elements-recursively (list)
11497 "Print each element of LIST on a line of its own.
11499 (when list ; @r{do-again-test}
11500 (print (car list)) ; @r{body}
11501 (print-elements-recursively ; @r{recursive call}
11502 (cdr list)))) ; @r{next-step-expression}
11504 (print-elements-recursively animals)
11508 The @code{print-elements-recursively} function first tests whether
11509 there is any content in the list; if there is, the function prints the
11510 first element of the list, the @sc{car} of the list. Then the
11511 function `invokes itself', but gives itself as its argument, not the
11512 whole list, but the second and subsequent elements of the list, the
11513 @sc{cdr} of the list.
11515 Put another way, if the list is not empty, the function invokes
11516 another instance of code that is similar to the initial code, but is a
11517 different thread of execution, with different arguments than the first
11520 Put in yet another way, if the list is not empty, the first robot
11521 assemblies a second robot and tells it what to do; the second robot is
11522 a different individual from the first, but is the same model.
11524 When the second evaluation occurs, the @code{when} expression is
11525 evaluated and if true, prints the first element of the list it
11526 receives as its argument (which is the second element of the original
11527 list). Then the function `calls itself' with the @sc{cdr} of the list
11528 it is invoked with, which (the second time around) is the @sc{cdr} of
11529 the @sc{cdr} of the original list.
11531 Note that although we say that the function `calls itself', what we
11532 mean is that the Lisp interpreter assembles and instructs a new
11533 instance of the program. The new instance is a clone of the first,
11534 but is a separate individual.
11536 Each time the function `invokes itself', it invokes itself on a
11537 shorter version of the original list. It creates a new instance that
11538 works on a shorter list.
11540 Eventually, the function invokes itself on an empty list. It creates
11541 a new instance whose argument is @code{nil}. The conditional expression
11542 tests the value of @code{list}. Since the value of @code{list} is
11543 @code{nil}, the @code{when} expression tests false so the then-part is
11544 not evaluated. The function as a whole then returns @code{nil}.
11547 When you evaluate @code{(print-elements-recursively animals)} in the
11548 @file{*scratch*} buffer, you see this result:
11564 @node Recursive triangle function, Recursion with cond, Recursion with list, Recursion
11565 @comment node-name, next, previous, up
11566 @subsection Recursion in Place of a Counter
11567 @findex triangle-recursively
11570 The @code{triangle} function described in a previous section can also
11571 be written recursively. It looks like this:
11575 (defun triangle-recursively (number)
11576 "Return the sum of the numbers 1 through NUMBER inclusive.
11578 (if (= number 1) ; @r{do-again-test}
11580 (+ number ; @r{else-part}
11581 (triangle-recursively ; @r{recursive call}
11582 (1- number))))) ; @r{next-step-expression}
11584 (triangle-recursively 7)
11589 You can install this function by evaluating it and then try it by
11590 evaluating @code{(triangle-recursively 7)}. (Remember to put your
11591 cursor immediately after the last parenthesis of the function
11592 definition, before the comment.) The function evaluates to 28.
11594 To understand how this function works, let's consider what happens in the
11595 various cases when the function is passed 1, 2, 3, or 4 as the value of
11599 * Recursive Example arg of 1 or 2::
11600 * Recursive Example arg of 3 or 4::
11603 @node Recursive Example arg of 1 or 2, Recursive Example arg of 3 or 4, Recursive triangle function, Recursive triangle function
11605 @unnumberedsubsubsec An argument of 1 or 2
11608 First, what happens if the value of the argument is 1?
11610 The function has an @code{if} expression after the documentation
11611 string. It tests whether the value of @code{number} is equal to 1; if
11612 so, Emacs evaluates the then-part of the @code{if} expression, which
11613 returns the number 1 as the value of the function. (A triangle with
11614 one row has one pebble in it.)
11616 Suppose, however, that the value of the argument is 2. In this case,
11617 Emacs evaluates the else-part of the @code{if} expression.
11620 The else-part consists of an addition, the recursive call to
11621 @code{triangle-recursively} and a decrementing action; and it looks like
11625 (+ number (triangle-recursively (1- number)))
11628 When Emacs evaluates this expression, the innermost expression is
11629 evaluated first; then the other parts in sequence. Here are the steps
11633 @item Step 1 @w{ } Evaluate the innermost expression.
11635 The innermost expression is @code{(1- number)} so Emacs decrements the
11636 value of @code{number} from 2 to 1.
11638 @item Step 2 @w{ } Evaluate the @code{triangle-recursively} function.
11640 The Lisp interpreter creates an individual instance of
11641 @code{triangle-recursively}. It does not matter that this function is
11642 contained within itself. Emacs passes the result Step 1 as the
11643 argument used by this instance of the @code{triangle-recursively}
11646 In this case, Emacs evaluates @code{triangle-recursively} with an
11647 argument of 1. This means that this evaluation of
11648 @code{triangle-recursively} returns 1.
11650 @item Step 3 @w{ } Evaluate the value of @code{number}.
11652 The variable @code{number} is the second element of the list that
11653 starts with @code{+}; its value is 2.
11655 @item Step 4 @w{ } Evaluate the @code{+} expression.
11657 The @code{+} expression receives two arguments, the first
11658 from the evaluation of @code{number} (Step 3) and the second from the
11659 evaluation of @code{triangle-recursively} (Step 2).
11661 The result of the addition is the sum of 2 plus 1, and the number 3 is
11662 returned, which is correct. A triangle with two rows has three
11666 @node Recursive Example arg of 3 or 4, , Recursive Example arg of 1 or 2, Recursive triangle function
11667 @unnumberedsubsubsec An argument of 3 or 4
11669 Suppose that @code{triangle-recursively} is called with an argument of
11673 @item Step 1 @w{ } Evaluate the do-again-test.
11675 The @code{if} expression is evaluated first. This is the do-again
11676 test and returns false, so the else-part of the @code{if} expression
11677 is evaluated. (Note that in this example, the do-again-test causes
11678 the function to call itself when it tests false, not when it tests
11681 @item Step 2 @w{ } Evaluate the innermost expression of the else-part.
11683 The innermost expression of the else-part is evaluated, which decrements
11684 3 to 2. This is the next-step-expression.
11686 @item Step 3 @w{ } Evaluate the @code{triangle-recursively} function.
11688 The number 2 is passed to the @code{triangle-recursively} function.
11690 We know what happens when Emacs evaluates @code{triangle-recursively} with
11691 an argument of 2. After going through the sequence of actions described
11692 earlier, it returns a value of 3. So that is what will happen here.
11694 @item Step 4 @w{ } Evaluate the addition.
11696 3 will be passed as an argument to the addition and will be added to the
11697 number with which the function was called, which is 3.
11701 The value returned by the function as a whole will be 6.
11703 Now that we know what will happen when @code{triangle-recursively} is
11704 called with an argument of 3, it is evident what will happen if it is
11705 called with an argument of 4:
11709 In the recursive call, the evaluation of
11712 (triangle-recursively (1- 4))
11717 will return the value of evaluating
11720 (triangle-recursively 3)
11724 which is 6 and this value will be added to 4 by the addition in the
11729 The value returned by the function as a whole will be 10.
11731 Each time @code{triangle-recursively} is evaluated, it evaluates a
11732 version of itself---a different instance of itself---with a smaller
11733 argument, until the argument is small enough so that it does not
11736 Note that this particular design for a recursive function
11737 requires that operations be deferred.
11739 Before @code{(triangle-recursively 7)} can calculate its answer, it
11740 must call @code{(triangle-recursively 6)}; and before
11741 @code{(triangle-recursively 6)} can calculate its answer, it must call
11742 @code{(triangle-recursively 5)}; and so on. That is to say, the
11743 calculation that @code{(triangle-recursively 7)} makes must be
11744 deferred until @code{(triangle-recursively 6)} makes its calculation;
11745 and @code{(triangle-recursively 6)} must defer until
11746 @code{(triangle-recursively 5)} completes; and so on.
11748 If each of these instances of @code{triangle-recursively} are thought
11749 of as different robots, the first robot must wait for the second to
11750 complete its job, which must wait until the third completes, and so
11753 There is a way around this kind of waiting, which we will discuss in
11754 @ref{No Deferment, , Recursion without Deferments}.
11756 @node Recursion with cond, Recursive Patterns, Recursive triangle function, Recursion
11757 @comment node-name, next, previous, up
11758 @subsection Recursion Example Using @code{cond}
11761 The version of @code{triangle-recursively} described earlier is written
11762 with the @code{if} special form. It can also be written using another
11763 special form called @code{cond}. The name of the special form
11764 @code{cond} is an abbreviation of the word @samp{conditional}.
11766 Although the @code{cond} special form is not used as often in the
11767 Emacs Lisp sources as @code{if}, it is used often enough to justify
11771 The template for a @code{cond} expression looks like this:
11781 where the @var{body} is a series of lists.
11784 Written out more fully, the template looks like this:
11789 (@var{first-true-or-false-test} @var{first-consequent})
11790 (@var{second-true-or-false-test} @var{second-consequent})
11791 (@var{third-true-or-false-test} @var{third-consequent})
11796 When the Lisp interpreter evaluates the @code{cond} expression, it
11797 evaluates the first element (the @sc{car} or true-or-false-test) of
11798 the first expression in a series of expressions within the body of the
11801 If the true-or-false-test returns @code{nil} the rest of that
11802 expression, the consequent, is skipped and the true-or-false-test of the
11803 next expression is evaluated. When an expression is found whose
11804 true-or-false-test returns a value that is not @code{nil}, the
11805 consequent of that expression is evaluated. The consequent can be one
11806 or more expressions. If the consequent consists of more than one
11807 expression, the expressions are evaluated in sequence and the value of
11808 the last one is returned. If the expression does not have a consequent,
11809 the value of the true-or-false-test is returned.
11811 If none of the true-or-false-tests test true, the @code{cond} expression
11812 returns @code{nil}.
11815 Written using @code{cond}, the @code{triangle} function looks like this:
11819 (defun triangle-using-cond (number)
11820 (cond ((<= number 0) 0)
11823 (+ number (triangle-using-cond (1- number))))))
11828 In this example, the @code{cond} returns 0 if the number is less than or
11829 equal to 0, it returns 1 if the number is 1 and it evaluates @code{(+
11830 number (triangle-using-cond (1- number)))} if the number is greater than
11833 @node Recursive Patterns, No Deferment, Recursion with cond, Recursion
11834 @comment node-name, next, previous, up
11835 @subsection Recursive Patterns
11836 @cindex Recursive Patterns
11838 Here are three common recursive patterns. Each involves a list.
11839 Recursion does not need to involve lists, but Lisp is designed for lists
11840 and this provides a sense of its primal capabilities.
11848 @node Every, Accumulate, Recursive Patterns, Recursive Patterns
11849 @comment node-name, next, previous, up
11850 @unnumberedsubsubsec Recursive Pattern: @emph{every}
11851 @cindex Every, type of recursive pattern
11852 @cindex Recursive pattern: every
11854 In the @code{every} recursive pattern, an action is performed on every
11858 The basic pattern is:
11862 If a list be empty, return @code{nil}.
11864 Else, act on the beginning of the list (the @sc{car} of the list)
11867 through a recursive call by the function on the rest (the
11868 @sc{cdr}) of the list,
11870 and, optionally, combine the acted-on element, using @code{cons},
11871 with the results of acting on the rest.
11880 (defun square-each (numbers-list)
11881 "Square each of a NUMBERS LIST, recursively."
11882 (if (not numbers-list) ; do-again-test
11885 (* (car numbers-list) (car numbers-list))
11886 (square-each (cdr numbers-list))))) ; next-step-expression
11890 (square-each '(1 2 3))
11897 If @code{numbers-list} is empty, do nothing. But if it has content,
11898 construct a list combining the square of the first number in the list
11899 with the result of the recursive call.
11901 (The example follows the pattern exactly: @code{nil} is returned if
11902 the numbers' list is empty. In practice, you would write the
11903 conditional so it carries out the action when the numbers' list is not
11906 The @code{print-elements-recursively} function (@pxref{Recursion with
11907 list, , Recursion with a List}) is another example of an @code{every}
11908 pattern, except in this case, rather than bring the results together
11909 using @code{cons}, we print each element of output.
11912 The @code{print-elements-recursively} function looks like this:
11916 (setq animals '(gazelle giraffe lion tiger))
11920 (defun print-elements-recursively (list)
11921 "Print each element of LIST on a line of its own.
11923 (when list ; @r{do-again-test}
11924 (print (car list)) ; @r{body}
11925 (print-elements-recursively ; @r{recursive call}
11926 (cdr list)))) ; @r{next-step-expression}
11928 (print-elements-recursively animals)
11933 The pattern for @code{print-elements-recursively} is:
11937 When the list is empty, do nothing.
11939 But when the list has at least one element,
11942 act on the beginning of the list (the @sc{car} of the list),
11944 and make a recursive call on the rest (the @sc{cdr}) of the list.
11948 @node Accumulate, Keep, Every, Recursive Patterns
11949 @comment node-name, next, previous, up
11950 @unnumberedsubsubsec Recursive Pattern: @emph{accumulate}
11951 @cindex Accumulate, type of recursive pattern
11952 @cindex Recursive pattern: accumulate
11954 Another recursive pattern is called the @code{accumulate} pattern. In
11955 the @code{accumulate} recursive pattern, an action is performed on
11956 every element of a list and the result of that action is accumulated
11957 with the results of performing the action on the other elements.
11959 This is very like the `every' pattern using @code{cons}, except that
11960 @code{cons} is not used, but some other combiner.
11967 If a list be empty, return zero or some other constant.
11969 Else, act on the beginning of the list (the @sc{car} of the list),
11972 and combine that acted-on element, using @code{+} or
11973 some other combining function, with
11975 a recursive call by the function on the rest (the @sc{cdr}) of the list.
11980 Here is an example:
11984 (defun add-elements (numbers-list)
11985 "Add the elements of NUMBERS-LIST together."
11986 (if (not numbers-list)
11988 (+ (car numbers-list) (add-elements (cdr numbers-list)))))
11992 (add-elements '(1 2 3 4))
11997 @xref{Files List, , Making a List of Files}, for an example of the
11998 accumulate pattern.
12000 @node Keep, , Accumulate, Recursive Patterns
12001 @comment node-name, next, previous, up
12002 @unnumberedsubsubsec Recursive Pattern: @emph{keep}
12003 @cindex Keep, type of recursive pattern
12004 @cindex Recursive pattern: keep
12006 A third recursive pattern is called the @code{keep} pattern.
12007 In the @code{keep} recursive pattern, each element of a list is tested;
12008 the element is acted on and the results are kept only if the element
12011 Again, this is very like the `every' pattern, except the element is
12012 skipped unless it meets a criterion.
12015 The pattern has three parts:
12019 If a list be empty, return @code{nil}.
12021 Else, if the beginning of the list (the @sc{car} of the list) passes
12025 act on that element and combine it, using @code{cons} with
12027 a recursive call by the function on the rest (the @sc{cdr}) of the list.
12030 Otherwise, if the beginning of the list (the @sc{car} of the list) fails
12034 skip on that element,
12036 and, recursively call the function on the rest (the @sc{cdr}) of the list.
12041 Here is an example that uses @code{cond}:
12045 (defun keep-three-letter-words (word-list)
12046 "Keep three letter words in WORD-LIST."
12048 ;; First do-again-test: stop-condition
12049 ((not word-list) nil)
12051 ;; Second do-again-test: when to act
12052 ((eq 3 (length (symbol-name (car word-list))))
12053 ;; combine acted-on element with recursive call on shorter list
12054 (cons (car word-list) (keep-three-letter-words (cdr word-list))))
12056 ;; Third do-again-test: when to skip element;
12057 ;; recursively call shorter list with next-step expression
12058 (t (keep-three-letter-words (cdr word-list)))))
12062 (keep-three-letter-words '(one two three four five six))
12063 @result{} (one two six)
12067 It goes without saying that you need not use @code{nil} as the test for
12068 when to stop; and you can, of course, combine these patterns.
12070 @node No Deferment, No deferment solution, Recursive Patterns, Recursion
12071 @subsection Recursion without Deferments
12072 @cindex Deferment in recursion
12073 @cindex Recursion without Deferments
12075 Let's consider again what happens with the @code{triangle-recursively}
12076 function. We will find that the intermediate calculations are
12077 deferred until all can be done.
12080 Here is the function definition:
12084 (defun triangle-recursively (number)
12085 "Return the sum of the numbers 1 through NUMBER inclusive.
12087 (if (= number 1) ; @r{do-again-test}
12089 (+ number ; @r{else-part}
12090 (triangle-recursively ; @r{recursive call}
12091 (1- number))))) ; @r{next-step-expression}
12095 What happens when we call this function with a argument of 7?
12097 The first instance of the @code{triangle-recursively} function adds
12098 the number 7 to the value returned by a second instance of
12099 @code{triangle-recursively}, an instance that has been passed an
12100 argument of 6. That is to say, the first calculation is:
12103 (+ 7 (triangle-recursively 6))
12107 The first instance of @code{triangle-recursively}---you may want to
12108 think of it as a little robot---cannot complete its job. It must hand
12109 off the calculation for @code{(triangle-recursively 6)} to a second
12110 instance of the program, to a second robot. This second individual is
12111 completely different from the first one; it is, in the jargon, a
12112 `different instantiation'. Or, put another way, it is a different
12113 robot. It is the same model as the first; it calculates triangle
12114 numbers recursively; but it has a different serial number.
12116 And what does @code{(triangle-recursively 6)} return? It returns the
12117 number 6 added to the value returned by evaluating
12118 @code{triangle-recursively} with an argument of 5. Using the robot
12119 metaphor, it asks yet another robot to help it.
12125 (+ 7 6 (triangle-recursively 5))
12129 And what happens next?
12132 (+ 7 6 5 (triangle-recursively 4))
12135 Each time @code{triangle-recursively} is called, except for the last
12136 time, it creates another instance of the program---another robot---and
12137 asks it to make a calculation.
12140 Eventually, the full addition is set up and performed:
12146 This design for the function defers the calculation of the first step
12147 until the second can be done, and defers that until the third can be
12148 done, and so on. Each deferment means the computer must remember what
12149 is being waited on. This is not a problem when there are only a few
12150 steps, as in this example. But it can be a problem when there are
12153 @node No deferment solution, , No Deferment, Recursion
12154 @subsection No Deferment Solution
12155 @cindex No deferment solution
12156 @cindex Defermentless solution
12157 @cindex Solution without deferment
12159 The solution to the problem of deferred operations is to write in a
12160 manner that does not defer operations@footnote{The phrase @dfn{tail
12161 recursive} is used to describe such a process, one that uses
12162 `constant space'.}. This requires
12163 writing to a different pattern, often one that involves writing two
12164 function definitions, an `initialization' function and a `helper'
12167 The `initialization' function sets up the job; the `helper' function
12171 Here are the two function definitions for adding up numbers. They are
12172 so simple, I find them hard to understand.
12176 (defun triangle-initialization (number)
12177 "Return the sum of the numbers 1 through NUMBER inclusive.
12178 This is the `initialization' component of a two function
12179 duo that uses recursion."
12180 (triangle-recursive-helper 0 0 number))
12186 (defun triangle-recursive-helper (sum counter number)
12187 "Return SUM, using COUNTER, through NUMBER inclusive.
12188 This is the `helper' component of a two function duo
12189 that uses recursion."
12190 (if (> counter number)
12192 (triangle-recursive-helper (+ sum counter) ; @r{sum}
12193 (1+ counter) ; @r{counter}
12194 number))) ; @r{number}
12199 Install both function definitions by evaluating them, then call
12200 @code{triangle-initialization} with 2 rows:
12204 (triangle-initialization 2)
12209 The `initialization' function calls the first instance of the `helper'
12210 function with three arguments: zero, zero, and a number which is the
12211 number of rows in the triangle.
12213 The first two arguments passed to the `helper' function are
12214 initialization values. These values are changed when
12215 @code{triangle-recursive-helper} invokes new instances.@footnote{The
12216 jargon is mildly confusing: @code{triangle-recursive-helper} uses a
12217 process that is iterative in a procedure that is recursive. The
12218 process is called iterative because the computer need only record the
12219 three values, @code{sum}, @code{counter}, and @code{number}; the
12220 procedure is recursive because the function `calls itself'. On the
12221 other hand, both the process and the procedure used by
12222 @code{triangle-recursively} are called recursive. The word
12223 `recursive' has different meanings in the two contexts.}
12225 Let's see what happens when we have a triangle that has one row. (This
12226 triangle will have one pebble in it!)
12229 @code{triangle-initialization} will call its helper with
12230 the arguments @w{@code{0 0 1}}. That function will run the conditional
12231 test whether @code{(> counter number)}:
12239 and find that the result is false, so it will invoke
12240 the else-part of the @code{if} clause:
12244 (triangle-recursive-helper
12245 (+ sum counter) ; @r{sum plus counter} @result{} @r{sum}
12246 (1+ counter) ; @r{increment counter} @result{} @r{counter}
12247 number) ; @r{number stays the same}
12253 which will first compute:
12257 (triangle-recursive-helper (+ 0 0) ; @r{sum}
12258 (1+ 0) ; @r{counter}
12262 (triangle-recursive-helper 0 1 1)
12266 Again, @code{(> counter number)} will be false, so again, the Lisp
12267 interpreter will evaluate @code{triangle-recursive-helper}, creating a
12268 new instance with new arguments.
12271 This new instance will be;
12275 (triangle-recursive-helper
12276 (+ sum counter) ; @r{sum plus counter} @result{} @r{sum}
12277 (1+ counter) ; @r{increment counter} @result{} @r{counter}
12278 number) ; @r{number stays the same}
12282 (triangle-recursive-helper 1 2 1)
12286 In this case, the @code{(> counter number)} test will be true! So the
12287 instance will return the value of the sum, which will be 1, as
12290 Now, let's pass @code{triangle-initialization} an argument
12291 of 2, to find out how many pebbles there are in a triangle with two rows.
12293 That function calls @code{(triangle-recursive-helper 0 0 2)}.
12296 In stages, the instances called will be:
12300 @r{sum counter number}
12301 (triangle-recursive-helper 0 1 2)
12303 (triangle-recursive-helper 1 2 2)
12305 (triangle-recursive-helper 3 3 2)
12309 When the last instance is called, the @code{(> counter number)} test
12310 will be true, so the instance will return the value of @code{sum},
12313 This kind of pattern helps when you are writing functions that can use
12314 many resources in a computer.
12317 @node Looping exercise, , Recursion, Loops & Recursion
12318 @section Looping Exercise
12322 Write a function similar to @code{triangle} in which each row has a
12323 value which is the square of the row number. Use a @code{while} loop.
12326 Write a function similar to @code{triangle} that multiplies instead of
12330 Rewrite these two functions recursively. Rewrite these functions
12333 @c comma in printed title causes problem in Info cross reference
12335 Write a function for Texinfo mode that creates an index entry at the
12336 beginning of a paragraph for every @samp{@@dfn} within the paragraph.
12337 (In a Texinfo file, @samp{@@dfn} marks a definition. This book is
12338 written in Texinfo.)
12340 Many of the functions you will need are described in two of the
12341 previous chapters, @ref{Cutting & Storing Text, , Cutting and Storing
12342 Text}, and @ref{Yanking, , Yanking Text Back}. If you use
12343 @code{forward-paragraph} to put the index entry at the beginning of
12344 the paragraph, you will have to use @w{@kbd{C-h f}}
12345 (@code{describe-function}) to find out how to make the command go
12348 For more information, see
12350 @ref{Indicating, , Indicating Definitions, texinfo}.
12353 @ref{Indicating, , Indicating, texinfo, Texinfo Manual}, which goes to
12354 a Texinfo manual in the current directory. Or, if you are on the
12356 @uref{http://www.gnu.org/software/texinfo/manual/texinfo/}
12359 ``Indicating Definitions, Commands, etc.'' in @cite{Texinfo, The GNU
12360 Documentation Format}.
12364 @node Regexp Search, Counting Words, Loops & Recursion, Top
12365 @comment node-name, next, previous, up
12366 @chapter Regular Expression Searches
12367 @cindex Searches, illustrating
12368 @cindex Regular expression searches
12369 @cindex Patterns, searching for
12370 @cindex Motion by sentence and paragraph
12371 @cindex Sentences, movement by
12372 @cindex Paragraphs, movement by
12374 Regular expression searches are used extensively in GNU Emacs. The
12375 two functions, @code{forward-sentence} and @code{forward-paragraph},
12376 illustrate these searches well. They use regular expressions to find
12377 where to move point. The phrase `regular expression' is often written
12380 Regular expression searches are described in @ref{Regexp Search, ,
12381 Regular Expression Search, emacs, The GNU Emacs Manual}, as well as in
12382 @ref{Regular Expressions, , , elisp, The GNU Emacs Lisp Reference
12383 Manual}. In writing this chapter, I am presuming that you have at
12384 least a mild acquaintance with them. The major point to remember is
12385 that regular expressions permit you to search for patterns as well as
12386 for literal strings of characters. For example, the code in
12387 @code{forward-sentence} searches for the pattern of possible
12388 characters that could mark the end of a sentence, and moves point to
12391 Before looking at the code for the @code{forward-sentence} function, it
12392 is worth considering what the pattern that marks the end of a sentence
12393 must be. The pattern is discussed in the next section; following that
12394 is a description of the regular expression search function,
12395 @code{re-search-forward}. The @code{forward-sentence} function
12396 is described in the section following. Finally, the
12397 @code{forward-paragraph} function is described in the last section of
12398 this chapter. @code{forward-paragraph} is a complex function that
12399 introduces several new features.
12403 * re-search-forward::
12404 * forward-sentence::
12405 * forward-paragraph::
12408 * re-search Exercises::
12411 @node sentence-end, re-search-forward, Regexp Search, Regexp Search
12412 @comment node-name, next, previous, up
12413 @section The Regular Expression for @code{sentence-end}
12414 @findex sentence-end
12416 The symbol @code{sentence-end} is bound to the pattern that marks the
12417 end of a sentence. What should this regular expression be?
12419 Clearly, a sentence may be ended by a period, a question mark, or an
12420 exclamation mark. Indeed, in English, only clauses that end with one
12421 of those three characters should be considered the end of a sentence.
12422 This means that the pattern should include the character set:
12428 However, we do not want @code{forward-sentence} merely to jump to a
12429 period, a question mark, or an exclamation mark, because such a character
12430 might be used in the middle of a sentence. A period, for example, is
12431 used after abbreviations. So other information is needed.
12433 According to convention, you type two spaces after every sentence, but
12434 only one space after a period, a question mark, or an exclamation mark in
12435 the body of a sentence. So a period, a question mark, or an exclamation
12436 mark followed by two spaces is a good indicator of an end of sentence.
12437 However, in a file, the two spaces may instead be a tab or the end of a
12438 line. This means that the regular expression should include these three
12439 items as alternatives.
12442 This group of alternatives will look like this:
12453 Here, @samp{$} indicates the end of the line, and I have pointed out
12454 where the tab and two spaces are inserted in the expression. Both are
12455 inserted by putting the actual characters into the expression.
12457 Two backslashes, @samp{\\}, are required before the parentheses and
12458 vertical bars: the first backslash quotes the following backslash in
12459 Emacs; and the second indicates that the following character, the
12460 parenthesis or the vertical bar, is special.
12463 Also, a sentence may be followed by one or more carriage returns, like
12474 Like tabs and spaces, a carriage return is inserted into a regular
12475 expression by inserting it literally. The asterisk indicates that the
12476 @key{RET} is repeated zero or more times.
12478 But a sentence end does not consist only of a period, a question mark or
12479 an exclamation mark followed by appropriate space: a closing quotation
12480 mark or a closing brace of some kind may precede the space. Indeed more
12481 than one such mark or brace may precede the space. These require a
12482 expression that looks like this:
12488 In this expression, the first @samp{]} is the first character in the
12489 expression; the second character is @samp{"}, which is preceded by a
12490 @samp{\} to tell Emacs the @samp{"} is @emph{not} special. The last
12491 three characters are @samp{'}, @samp{)}, and @samp{@}}.
12493 All this suggests what the regular expression pattern for matching the
12494 end of a sentence should be; and, indeed, if we evaluate
12495 @code{sentence-end} we find that it returns the following value:
12500 @result{} "[.?!][]\"')@}]*\\($\\| \\| \\)[
12506 (Well, not in GNU Emacs 22; that is because of an effort to make the
12507 process simpler and to handle more glyphs and languages. When the
12508 value of @code{sentence-end} is @code{nil}, then use the value defined
12509 by the function @code{sentence-end}. (Here is a use of the difference
12510 between a value and a function in Emacs Lisp.) The function returns a
12511 value constructed from the variables @code{sentence-end-base},
12512 @code{sentence-end-double-space}, @code{sentence-end-without-period},
12513 and @code{sentence-end-without-space}. The critical variable is
12514 @code{sentence-end-base}; its global value is similar to the one
12515 described above but it also contains two additional quotation marks.
12516 These have differing degrees of curliness. The
12517 @code{sentence-end-without-period} variable, when true, tells Emacs
12518 that a sentence may end without a period, such as text in Thai.)
12522 (Note that here the @key{TAB}, two spaces, and @key{RET} are shown
12523 literally in the pattern.)
12525 This regular expression can be deciphered as follows:
12529 The first part of the pattern is the three characters, a period, a question
12530 mark and an exclamation mark, within square brackets. The pattern must
12531 begin with one or other of these characters.
12534 The second part of the pattern is the group of closing braces and
12535 quotation marks, which can appear zero or more times. These may follow
12536 the period, question mark or exclamation mark. In a regular expression,
12537 the backslash, @samp{\}, followed by the double quotation mark,
12538 @samp{"}, indicates the class of string-quote characters. Usually, the
12539 double quotation mark is the only character in this class. The
12540 asterisk, @samp{*}, indicates that the items in the previous group (the
12541 group surrounded by square brackets, @samp{[]}) may be repeated zero or
12544 @item \\($\\| \\| \\)
12545 The third part of the pattern is one or other of: either the end of a
12546 line, or two blank spaces, or a tab. The double back-slashes are used
12547 to prevent Emacs from reading the parentheses and vertical bars as part
12548 of the search pattern; the parentheses are used to mark the group and
12549 the vertical bars are used to indicated that the patterns to either side
12550 of them are alternatives. The dollar sign is used to indicate the end
12551 of a line and both the two spaces and the tab are each inserted as is to
12552 indicate what they are.
12555 Finally, the last part of the pattern indicates that the end of the line
12556 or the whitespace following the period, question mark or exclamation
12557 mark may, but need not, be followed by one or more carriage returns. In
12558 the pattern, the carriage return is inserted as an actual carriage
12559 return between square brackets but here it is shown as @key{RET}.
12563 @node re-search-forward, forward-sentence, sentence-end, Regexp Search
12564 @comment node-name, next, previous, up
12565 @section The @code{re-search-forward} Function
12566 @findex re-search-forward
12568 The @code{re-search-forward} function is very like the
12569 @code{search-forward} function. (@xref{search-forward, , The
12570 @code{search-forward} Function}.)
12572 @code{re-search-forward} searches for a regular expression. If the
12573 search is successful, it leaves point immediately after the last
12574 character in the target. If the search is backwards, it leaves point
12575 just before the first character in the target. You may tell
12576 @code{re-search-forward} to return @code{t} for true. (Moving point
12577 is therefore a `side effect'.)
12579 Like @code{search-forward}, the @code{re-search-forward} function takes
12584 The first argument is the regular expression that the function searches
12585 for. The regular expression will be a string between quotations marks.
12588 The optional second argument limits how far the function will search; it is a
12589 bound, which is specified as a position in the buffer.
12592 The optional third argument specifies how the function responds to
12593 failure: @code{nil} as the third argument causes the function to
12594 signal an error (and print a message) when the search fails; any other
12595 value causes it to return @code{nil} if the search fails and @code{t}
12596 if the search succeeds.
12599 The optional fourth argument is the repeat count. A negative repeat
12600 count causes @code{re-search-forward} to search backwards.
12604 The template for @code{re-search-forward} looks like this:
12608 (re-search-forward "@var{regular-expression}"
12609 @var{limit-of-search}
12610 @var{what-to-do-if-search-fails}
12611 @var{repeat-count})
12615 The second, third, and fourth arguments are optional. However, if you
12616 want to pass a value to either or both of the last two arguments, you
12617 must also pass a value to all the preceding arguments. Otherwise, the
12618 Lisp interpreter will mistake which argument you are passing the value
12622 In the @code{forward-sentence} function, the regular expression will be
12623 the value of the variable @code{sentence-end}. In simple form, that is:
12627 "[.?!][]\"')@}]*\\($\\| \\| \\)[
12633 The limit of the search will be the end of the paragraph (since a
12634 sentence cannot go beyond a paragraph). If the search fails, the
12635 function will return @code{nil}; and the repeat count will be provided
12636 by the argument to the @code{forward-sentence} function.
12638 @node forward-sentence, forward-paragraph, re-search-forward, Regexp Search
12639 @comment node-name, next, previous, up
12640 @section @code{forward-sentence}
12641 @findex forward-sentence
12643 The command to move the cursor forward a sentence is a straightforward
12644 illustration of how to use regular expression searches in Emacs Lisp.
12645 Indeed, the function looks longer and more complicated than it is; this
12646 is because the function is designed to go backwards as well as forwards;
12647 and, optionally, over more than one sentence. The function is usually
12648 bound to the key command @kbd{M-e}.
12651 * Complete forward-sentence::
12652 * fwd-sentence while loops::
12653 * fwd-sentence re-search::
12656 @node Complete forward-sentence, fwd-sentence while loops, forward-sentence, forward-sentence
12658 @unnumberedsubsec Complete @code{forward-sentence} function definition
12662 Here is the code for @code{forward-sentence}:
12667 (defun forward-sentence (&optional arg)
12668 "Move forward to next `sentence-end'. With argument, repeat.
12669 With negative argument, move backward repeatedly to `sentence-beginning'.
12671 The variable `sentence-end' is a regular expression that matches ends of
12672 sentences. Also, every paragraph boundary terminates sentences as well."
12676 (or arg (setq arg 1))
12677 (let ((opoint (point))
12678 (sentence-end (sentence-end)))
12680 (let ((pos (point))
12681 (par-beg (save-excursion (start-of-paragraph-text) (point))))
12682 (if (and (re-search-backward sentence-end par-beg t)
12683 (or (< (match-end 0) pos)
12684 (re-search-backward sentence-end par-beg t)))
12685 (goto-char (match-end 0))
12686 (goto-char par-beg)))
12687 (setq arg (1+ arg)))
12691 (let ((par-end (save-excursion (end-of-paragraph-text) (point))))
12692 (if (re-search-forward sentence-end par-end t)
12693 (skip-chars-backward " \t\n")
12694 (goto-char par-end)))
12695 (setq arg (1- arg)))
12696 (constrain-to-field nil opoint t)))
12704 (defun forward-sentence (&optional arg)
12705 "Move forward to next sentence-end. With argument, repeat.
12706 With negative argument, move backward repeatedly to sentence-beginning.
12707 Sentence ends are identified by the value of sentence-end
12708 treated as a regular expression. Also, every paragraph boundary
12709 terminates sentences as well."
12713 (or arg (setq arg 1))
12716 (save-excursion (start-of-paragraph-text) (point))))
12717 (if (re-search-backward
12718 (concat sentence-end "[^ \t\n]") par-beg t)
12719 (goto-char (1- (match-end 0)))
12720 (goto-char par-beg)))
12721 (setq arg (1+ arg)))
12724 (save-excursion (end-of-paragraph-text) (point))))
12725 (if (re-search-forward sentence-end par-end t)
12726 (skip-chars-backward " \t\n")
12727 (goto-char par-end)))
12728 (setq arg (1- arg))))
12733 The function looks long at first sight and it is best to look at its
12734 skeleton first, and then its muscle. The way to see the skeleton is to
12735 look at the expressions that start in the left-most columns:
12739 (defun forward-sentence (&optional arg)
12740 "@var{documentation}@dots{}"
12742 (or arg (setq arg 1))
12743 (let ((opoint (point)) (sentence-end (sentence-end)))
12745 (let ((pos (point))
12746 (par-beg (save-excursion (start-of-paragraph-text) (point))))
12747 @var{rest-of-body-of-while-loop-when-going-backwards}
12749 (let ((par-end (save-excursion (end-of-paragraph-text) (point))))
12750 @var{rest-of-body-of-while-loop-when-going-forwards}
12751 @var{handle-forms-and-equivalent}
12755 This looks much simpler! The function definition consists of
12756 documentation, an @code{interactive} expression, an @code{or}
12757 expression, a @code{let} expression, and @code{while} loops.
12759 Let's look at each of these parts in turn.
12761 We note that the documentation is thorough and understandable.
12763 The function has an @code{interactive "p"} declaration. This means
12764 that the processed prefix argument, if any, is passed to the
12765 function as its argument. (This will be a number.) If the function
12766 is not passed an argument (it is optional) then the argument
12767 @code{arg} will be bound to 1.
12769 When @code{forward-sentence} is called non-interactively without an
12770 argument, @code{arg} is bound to @code{nil}. The @code{or} expression
12771 handles this. What it does is either leave the value of @code{arg} as
12772 it is, but only if @code{arg} is bound to a value; or it sets the
12773 value of @code{arg} to 1, in the case when @code{arg} is bound to
12776 Next is a @code{let}. That specifies the values of two local
12777 variables, @code{point} and @code{sentence-end}. The local value of
12778 point, from before the search, is used in the
12779 @code{constrain-to-field} function which handles forms and
12780 equivalents. The @code{sentence-end} variable is set by the
12781 @code{sentence-end} function.
12783 @node fwd-sentence while loops, fwd-sentence re-search, Complete forward-sentence, forward-sentence
12784 @unnumberedsubsec The @code{while} loops
12786 Two @code{while} loops follow. The first @code{while} has a
12787 true-or-false-test that tests true if the prefix argument for
12788 @code{forward-sentence} is a negative number. This is for going
12789 backwards. The body of this loop is similar to the body of the second
12790 @code{while} clause, but it is not exactly the same. We will skip
12791 this @code{while} loop and concentrate on the second @code{while}
12795 The second @code{while} loop is for moving point forward. Its skeleton
12800 (while (> arg 0) ; @r{true-or-false-test}
12802 (if (@var{true-or-false-test})
12805 (setq arg (1- arg)))) ; @code{while} @r{loop decrementer}
12809 The @code{while} loop is of the decrementing kind.
12810 (@xref{Decrementing Loop, , A Loop with a Decrementing Counter}.) It
12811 has a true-or-false-test that tests true so long as the counter (in
12812 this case, the variable @code{arg}) is greater than zero; and it has a
12813 decrementer that subtracts 1 from the value of the counter every time
12816 If no prefix argument is given to @code{forward-sentence}, which is
12817 the most common way the command is used, this @code{while} loop will
12818 run once, since the value of @code{arg} will be 1.
12820 The body of the @code{while} loop consists of a @code{let} expression,
12821 which creates and binds a local variable, and has, as its body, an
12822 @code{if} expression.
12825 The body of the @code{while} loop looks like this:
12830 (save-excursion (end-of-paragraph-text) (point))))
12831 (if (re-search-forward sentence-end par-end t)
12832 (skip-chars-backward " \t\n")
12833 (goto-char par-end)))
12837 The @code{let} expression creates and binds the local variable
12838 @code{par-end}. As we shall see, this local variable is designed to
12839 provide a bound or limit to the regular expression search. If the
12840 search fails to find a proper sentence ending in the paragraph, it will
12841 stop on reaching the end of the paragraph.
12843 But first, let us examine how @code{par-end} is bound to the value of
12844 the end of the paragraph. What happens is that the @code{let} sets the
12845 value of @code{par-end} to the value returned when the Lisp interpreter
12846 evaluates the expression
12850 (save-excursion (end-of-paragraph-text) (point))
12855 In this expression, @code{(end-of-paragraph-text)} moves point to the
12856 end of the paragraph, @code{(point)} returns the value of point, and then
12857 @code{save-excursion} restores point to its original position. Thus,
12858 the @code{let} binds @code{par-end} to the value returned by the
12859 @code{save-excursion} expression, which is the position of the end of
12860 the paragraph. (The @code{end-of-paragraph-text} function uses
12861 @code{forward-paragraph}, which we will discuss shortly.)
12864 Emacs next evaluates the body of the @code{let}, which is an @code{if}
12865 expression that looks like this:
12869 (if (re-search-forward sentence-end par-end t) ; @r{if-part}
12870 (skip-chars-backward " \t\n") ; @r{then-part}
12871 (goto-char par-end))) ; @r{else-part}
12875 The @code{if} tests whether its first argument is true and if so,
12876 evaluates its then-part; otherwise, the Emacs Lisp interpreter
12877 evaluates the else-part. The true-or-false-test of the @code{if}
12878 expression is the regular expression search.
12880 It may seem odd to have what looks like the `real work' of
12881 the @code{forward-sentence} function buried here, but this is a common
12882 way this kind of operation is carried out in Lisp.
12884 @node fwd-sentence re-search, , fwd-sentence while loops, forward-sentence
12885 @unnumberedsubsec The regular expression search
12887 The @code{re-search-forward} function searches for the end of the
12888 sentence, that is, for the pattern defined by the @code{sentence-end}
12889 regular expression. If the pattern is found---if the end of the sentence is
12890 found---then the @code{re-search-forward} function does two things:
12894 The @code{re-search-forward} function carries out a side effect, which
12895 is to move point to the end of the occurrence found.
12898 The @code{re-search-forward} function returns a value of true. This is
12899 the value received by the @code{if}, and means that the search was
12904 The side effect, the movement of point, is completed before the
12905 @code{if} function is handed the value returned by the successful
12906 conclusion of the search.
12908 When the @code{if} function receives the value of true from a successful
12909 call to @code{re-search-forward}, the @code{if} evaluates the then-part,
12910 which is the expression @code{(skip-chars-backward " \t\n")}. This
12911 expression moves backwards over any blank spaces, tabs or carriage
12912 returns until a printed character is found and then leaves point after
12913 the character. Since point has already been moved to the end of the
12914 pattern that marks the end of the sentence, this action leaves point
12915 right after the closing printed character of the sentence, which is
12918 On the other hand, if the @code{re-search-forward} function fails to
12919 find a pattern marking the end of the sentence, the function returns
12920 false. The false then causes the @code{if} to evaluate its third
12921 argument, which is @code{(goto-char par-end)}: it moves point to the
12922 end of the paragraph.
12924 (And if the text is in a form or equivalent, and point may not move
12925 fully, then the @code{constrain-to-field} function comes into play.)
12927 Regular expression searches are exceptionally useful and the pattern
12928 illustrated by @code{re-search-forward}, in which the search is the
12929 test of an @code{if} expression, is handy. You will see or write code
12930 incorporating this pattern often.
12932 @node forward-paragraph, etags, forward-sentence, Regexp Search
12933 @comment node-name, next, previous, up
12934 @section @code{forward-paragraph}: a Goldmine of Functions
12935 @findex forward-paragraph
12939 (defun forward-paragraph (&optional arg)
12940 "Move forward to end of paragraph.
12941 With argument ARG, do it ARG times;
12942 a negative argument ARG = -N means move backward N paragraphs.
12944 A line which `paragraph-start' matches either separates paragraphs
12945 \(if `paragraph-separate' matches it also) or is the first line of a paragraph.
12946 A paragraph end is the beginning of a line which is not part of the paragraph
12947 to which the end of the previous line belongs, or the end of the buffer.
12948 Returns the count of paragraphs left to move."
12950 (or arg (setq arg 1))
12951 (let* ((opoint (point))
12952 (fill-prefix-regexp
12953 (and fill-prefix (not (equal fill-prefix ""))
12954 (not paragraph-ignore-fill-prefix)
12955 (regexp-quote fill-prefix)))
12956 ;; Remove ^ from paragraph-start and paragraph-sep if they are there.
12957 ;; These regexps shouldn't be anchored, because we look for them
12958 ;; starting at the left-margin. This allows paragraph commands to
12959 ;; work normally with indented text.
12960 ;; This hack will not find problem cases like "whatever\\|^something".
12961 (parstart (if (and (not (equal "" paragraph-start))
12962 (equal ?^ (aref paragraph-start 0)))
12963 (substring paragraph-start 1)
12965 (parsep (if (and (not (equal "" paragraph-separate))
12966 (equal ?^ (aref paragraph-separate 0)))
12967 (substring paragraph-separate 1)
12968 paragraph-separate))
12970 (if fill-prefix-regexp
12971 (concat parsep "\\|"
12972 fill-prefix-regexp "[ \t]*$")
12974 ;; This is used for searching.
12975 (sp-parstart (concat "^[ \t]*\\(?:" parstart "\\|" parsep "\\)"))
12977 (while (and (< arg 0) (not (bobp)))
12978 (if (and (not (looking-at parsep))
12979 (re-search-backward "^\n" (max (1- (point)) (point-min)) t)
12980 (looking-at parsep))
12981 (setq arg (1+ arg))
12982 (setq start (point))
12983 ;; Move back over paragraph-separating lines.
12984 (forward-char -1) (beginning-of-line)
12985 (while (and (not (bobp))
12986 (progn (move-to-left-margin)
12987 (looking-at parsep)))
12991 (setq arg (1+ arg))
12992 ;; Go to end of the previous (non-separating) line.
12994 ;; Search back for line that starts or separates paragraphs.
12995 (if (if fill-prefix-regexp
12996 ;; There is a fill prefix; it overrides parstart.
12997 (let (multiple-lines)
12998 (while (and (progn (beginning-of-line) (not (bobp)))
12999 (progn (move-to-left-margin)
13000 (not (looking-at parsep)))
13001 (looking-at fill-prefix-regexp))
13002 (unless (= (point) start)
13003 (setq multiple-lines t))
13005 (move-to-left-margin)
13006 ;; This deleted code caused a long hanging-indent line
13007 ;; not to be filled together with the following lines.
13008 ;; ;; Don't move back over a line before the paragraph
13009 ;; ;; which doesn't start with fill-prefix
13010 ;; ;; unless that is the only line we've moved over.
13011 ;; (and (not (looking-at fill-prefix-regexp))
13013 ;; (forward-line 1))
13015 (while (and (re-search-backward sp-parstart nil 1)
13016 (setq found-start t)
13017 ;; Found a candidate, but need to check if it is a
13019 (progn (setq start (point))
13020 (move-to-left-margin)
13021 (not (looking-at parsep)))
13022 (not (and (looking-at parstart)
13023 (or (not use-hard-newlines)
13026 (1- start) 'hard)))))
13027 (setq found-start nil)
13032 ;; Move forward over paragraph separators.
13033 ;; We know this cannot reach the place we started
13034 ;; because we know we moved back over a non-separator.
13035 (while (and (not (eobp))
13036 (progn (move-to-left-margin)
13037 (looking-at parsep)))
13039 ;; If line before paragraph is just margin, back up to there.
13041 (if (> (current-column) (current-left-margin))
13043 (skip-chars-backward " \t")
13045 (forward-line 1))))
13046 ;; No starter or separator line => use buffer beg.
13047 (goto-char (point-min))))))
13049 (while (and (> arg 0) (not (eobp)))
13050 ;; Move forward over separator lines...
13051 (while (and (not (eobp))
13052 (progn (move-to-left-margin) (not (eobp)))
13053 (looking-at parsep))
13055 (unless (eobp) (setq arg (1- arg)))
13056 ;; ... and one more line.
13058 (if fill-prefix-regexp
13059 ;; There is a fill prefix; it overrides parstart.
13060 (while (and (not (eobp))
13061 (progn (move-to-left-margin) (not (eobp)))
13062 (not (looking-at parsep))
13063 (looking-at fill-prefix-regexp))
13065 (while (and (re-search-forward sp-parstart nil 1)
13066 (progn (setq start (match-beginning 0))
13069 (progn (move-to-left-margin)
13070 (not (looking-at parsep)))
13071 (or (not (looking-at parstart))
13072 (and use-hard-newlines
13073 (not (get-text-property (1- start) 'hard)))))
13075 (if (< (point) (point-max))
13076 (goto-char start))))
13077 (constrain-to-field nil opoint t)
13078 ;; Return the number of steps that could not be done.
13082 The @code{forward-paragraph} function moves point forward to the end
13083 of the paragraph. It is usually bound to @kbd{M-@}} and makes use of a
13084 number of functions that are important in themselves, including
13085 @code{let*}, @code{match-beginning}, and @code{looking-at}.
13087 The function definition for @code{forward-paragraph} is considerably
13088 longer than the function definition for @code{forward-sentence}
13089 because it works with a paragraph, each line of which may begin with a
13092 A fill prefix consists of a string of characters that are repeated at
13093 the beginning of each line. For example, in Lisp code, it is a
13094 convention to start each line of a paragraph-long comment with
13095 @samp{;;; }. In Text mode, four blank spaces make up another common
13096 fill prefix, creating an indented paragraph. (@xref{Fill Prefix, , ,
13097 emacs, The GNU Emacs Manual}, for more information about fill
13100 The existence of a fill prefix means that in addition to being able to
13101 find the end of a paragraph whose lines begin on the left-most
13102 column, the @code{forward-paragraph} function must be able to find the
13103 end of a paragraph when all or many of the lines in the buffer begin
13104 with the fill prefix.
13106 Moreover, it is sometimes practical to ignore a fill prefix that
13107 exists, especially when blank lines separate paragraphs.
13108 This is an added complication.
13111 * forward-paragraph in brief::
13116 @node forward-paragraph in brief, fwd-para let, forward-paragraph, forward-paragraph
13118 @unnumberedsubsec Shortened @code{forward-paragraph} function definition
13121 Rather than print all of the @code{forward-paragraph} function, we
13122 will only print parts of it. Read without preparation, the function
13126 In outline, the function looks like this:
13130 (defun forward-paragraph (&optional arg)
13131 "@var{documentation}@dots{}"
13133 (or arg (setq arg 1))
13136 (while (and (< arg 0) (not (bobp))) ; @r{backward-moving-code}
13138 (while (and (> arg 0) (not (eobp))) ; @r{forward-moving-code}
13143 The first parts of the function are routine: the function's argument
13144 list consists of one optional argument. Documentation follows.
13146 The lower case @samp{p} in the @code{interactive} declaration means
13147 that the processed prefix argument, if any, is passed to the function.
13148 This will be a number, and is the repeat count of how many paragraphs
13149 point will move. The @code{or} expression in the next line handles
13150 the common case when no argument is passed to the function, which occurs
13151 if the function is called from other code rather than interactively.
13152 This case was described earlier. (@xref{forward-sentence, The
13153 @code{forward-sentence} function}.) Now we reach the end of the
13154 familiar part of this function.
13156 @node fwd-para let, fwd-para while, forward-paragraph in brief, forward-paragraph
13157 @unnumberedsubsec The @code{let*} expression
13159 The next line of the @code{forward-paragraph} function begins a
13160 @code{let*} expression. This is a different than @code{let}. The
13161 symbol is @code{let*} not @code{let}.
13163 The @code{let*} special form is like @code{let} except that Emacs sets
13164 each variable in sequence, one after another, and variables in the
13165 latter part of the varlist can make use of the values to which Emacs
13166 set variables in the earlier part of the varlist.
13169 ( refappend save-excursion, , code save-excursion in code append-to-buffer .)
13172 (@ref{append save-excursion, , @code{save-excursion} in @code{append-to-buffer}}.)
13174 In the @code{let*} expression in this function, Emacs binds a total of
13175 seven variables: @code{opoint}, @code{fill-prefix-regexp},
13176 @code{parstart}, @code{parsep}, @code{sp-parstart}, @code{start}, and
13177 @code{found-start}.
13179 The variable @code{parsep} appears twice, first, to remove instances
13180 of @samp{^}, and second, to handle fill prefixes.
13182 The variable @code{opoint} is just the value of @code{point}. As you
13183 can guess, it is used in a @code{constrain-to-field} expression, just
13184 as in @code{forward-sentence}.
13186 The variable @code{fill-prefix-regexp} is set to the value returned by
13187 evaluating the following list:
13192 (not (equal fill-prefix ""))
13193 (not paragraph-ignore-fill-prefix)
13194 (regexp-quote fill-prefix))
13199 This is an expression whose first element is the @code{and} special form.
13201 As we learned earlier (@pxref{kill-new function, , The @code{kill-new}
13202 function}), the @code{and} special form evaluates each of its
13203 arguments until one of the arguments returns a value of @code{nil}, in
13204 which case the @code{and} expression returns @code{nil}; however, if
13205 none of the arguments returns a value of @code{nil}, the value
13206 resulting from evaluating the last argument is returned. (Since such
13207 a value is not @code{nil}, it is considered true in Lisp.) In other
13208 words, an @code{and} expression returns a true value only if all its
13209 arguments are true.
13212 In this case, the variable @code{fill-prefix-regexp} is bound to a
13213 non-@code{nil} value only if the following four expressions produce a
13214 true (i.e., a non-@code{nil}) value when they are evaluated; otherwise,
13215 @code{fill-prefix-regexp} is bound to @code{nil}.
13219 When this variable is evaluated, the value of the fill prefix, if any,
13220 is returned. If there is no fill prefix, this variable returns
13223 @item (not (equal fill-prefix "")
13224 This expression checks whether an existing fill prefix is an empty
13225 string, that is, a string with no characters in it. An empty string is
13226 not a useful fill prefix.
13228 @item (not paragraph-ignore-fill-prefix)
13229 This expression returns @code{nil} if the variable
13230 @code{paragraph-ignore-fill-prefix} has been turned on by being set to a
13231 true value such as @code{t}.
13233 @item (regexp-quote fill-prefix)
13234 This is the last argument to the @code{and} special form. If all the
13235 arguments to the @code{and} are true, the value resulting from
13236 evaluating this expression will be returned by the @code{and} expression
13237 and bound to the variable @code{fill-prefix-regexp},
13240 @findex regexp-quote
13242 The result of evaluating this @code{and} expression successfully is that
13243 @code{fill-prefix-regexp} will be bound to the value of
13244 @code{fill-prefix} as modified by the @code{regexp-quote} function.
13245 What @code{regexp-quote} does is read a string and return a regular
13246 expression that will exactly match the string and match nothing else.
13247 This means that @code{fill-prefix-regexp} will be set to a value that
13248 will exactly match the fill prefix if the fill prefix exists.
13249 Otherwise, the variable will be set to @code{nil}.
13251 The next two local variables in the @code{let*} expression are
13252 designed to remove instances of @samp{^} from @code{parstart} and
13253 @code{parsep}, the local variables which indicate the paragraph start
13254 and the paragraph separator. The next expression sets @code{parsep}
13255 again. That is to handle fill prefixes.
13257 This is the setting that requires the definition call @code{let*}
13258 rather than @code{let}. The true-or-false-test for the @code{if}
13259 depends on whether the variable @code{fill-prefix-regexp} evaluates to
13260 @code{nil} or some other value.
13262 If @code{fill-prefix-regexp} does not have a value, Emacs evaluates
13263 the else-part of the @code{if} expression and binds @code{parsep} to
13264 its local value. (@code{parsep} is a regular expression that matches
13265 what separates paragraphs.)
13267 But if @code{fill-prefix-regexp} does have a value, Emacs evaluates
13268 the then-part of the @code{if} expression and binds @code{parsep} to a
13269 regular expression that includes the @code{fill-prefix-regexp} as part
13272 Specifically, @code{parsep} is set to the original value of the
13273 paragraph separate regular expression concatenated with an alternative
13274 expression that consists of the @code{fill-prefix-regexp} followed by
13275 optional whitespace to the end of the line. The whitespace is defined
13276 by @w{@code{"[ \t]*$"}}.) The @samp{\\|} defines this portion of the
13277 regexp as an alternative to @code{parsep}.
13279 According to a comment in the code, the next local variable,
13280 @code{sp-parstart}, is used for searching, and then the final two,
13281 @code{start} and @code{found-start}, are set to @code{nil}.
13283 Now we get into the body of the @code{let*}. The first part of the body
13284 of the @code{let*} deals with the case when the function is given a
13285 negative argument and is therefore moving backwards. We will skip this
13288 @node fwd-para while, , fwd-para let, forward-paragraph
13289 @unnumberedsubsec The forward motion @code{while} loop
13291 The second part of the body of the @code{let*} deals with forward
13292 motion. It is a @code{while} loop that repeats itself so long as the
13293 value of @code{arg} is greater than zero. In the most common use of
13294 the function, the value of the argument is 1, so the body of the
13295 @code{while} loop is evaluated exactly once, and the cursor moves
13296 forward one paragraph.
13299 (while (and (> arg 0) (not (eobp)))
13301 ;; Move forward over separator lines...
13302 (while (and (not (eobp))
13303 (progn (move-to-left-margin) (not (eobp)))
13304 (looking-at parsep))
13306 (unless (eobp) (setq arg (1- arg)))
13307 ;; ... and one more line.
13310 (if fill-prefix-regexp
13311 ;; There is a fill prefix; it overrides parstart.
13312 (while (and (not (eobp))
13313 (progn (move-to-left-margin) (not (eobp)))
13314 (not (looking-at parsep))
13315 (looking-at fill-prefix-regexp))
13318 (while (and (re-search-forward sp-parstart nil 1)
13319 (progn (setq start (match-beginning 0))
13322 (progn (move-to-left-margin)
13323 (not (looking-at parsep)))
13324 (or (not (looking-at parstart))
13325 (and use-hard-newlines
13326 (not (get-text-property (1- start) 'hard)))))
13329 (if (< (point) (point-max))
13330 (goto-char start))))
13333 This part handles three situations: when point is between paragraphs,
13334 when there is a fill prefix and when there is no fill prefix.
13337 The @code{while} loop looks like this:
13341 ;; @r{going forwards and not at the end of the buffer}
13342 (while (and (> arg 0) (not (eobp)))
13344 ;; @r{between paragraphs}
13345 ;; Move forward over separator lines...
13346 (while (and (not (eobp))
13347 (progn (move-to-left-margin) (not (eobp)))
13348 (looking-at parsep))
13350 ;; @r{This decrements the loop}
13351 (unless (eobp) (setq arg (1- arg)))
13352 ;; ... and one more line.
13357 (if fill-prefix-regexp
13358 ;; There is a fill prefix; it overrides parstart;
13359 ;; we go forward line by line
13360 (while (and (not (eobp))
13361 (progn (move-to-left-margin) (not (eobp)))
13362 (not (looking-at parsep))
13363 (looking-at fill-prefix-regexp))
13368 ;; There is no fill prefix;
13369 ;; we go forward character by character
13370 (while (and (re-search-forward sp-parstart nil 1)
13371 (progn (setq start (match-beginning 0))
13374 (progn (move-to-left-margin)
13375 (not (looking-at parsep)))
13376 (or (not (looking-at parstart))
13377 (and use-hard-newlines
13378 (not (get-text-property (1- start) 'hard)))))
13383 ;; and if there is no fill prefix and if we are not at the end,
13384 ;; go to whatever was found in the regular expression search
13386 (if (< (point) (point-max))
13387 (goto-char start))))
13392 We can see that this is a decrementing counter @code{while} loop,
13393 using the expression @code{(setq arg (1- arg))} as the decrementer.
13394 That expression is not far from the @code{while}, but is hidden in
13395 another Lisp macro, an @code{unless} macro. Unless we are at the end
13396 of the buffer --- that is what the @code{eobp} function determines; it
13397 is an abbreviation of @samp{End Of Buffer P} --- we decrease the value
13398 of @code{arg} by one.
13400 (If we are at the end of the buffer, we cannot go forward any more and
13401 the next loop of the @code{while} expression will test false since the
13402 test is an @code{and} with @code{(not (eobp))}. The @code{not}
13403 function means exactly as you expect; it is another name for
13404 @code{null}, a function that returns true when its argument is false.)
13406 Interestingly, the loop count is not decremented until we leave the
13407 space between paragraphs, unless we come to the end of buffer or stop
13408 seeing the local value of the paragraph separator.
13410 That second @code{while} also has a @code{(move-to-left-margin)}
13411 expression. The function is self-explanatory. It is inside a
13412 @code{progn} expression and not the last element of its body, so it is
13413 only invoked for its side effect, which is to move point to the left
13414 margin of the current line.
13417 The @code{looking-at} function is also self-explanatory; it returns
13418 true if the text after point matches the regular expression given as
13421 The rest of the body of the loop looks difficult at first, but makes
13422 sense as you come to understand it.
13425 First consider what happens if there is a fill prefix:
13429 (if fill-prefix-regexp
13430 ;; There is a fill prefix; it overrides parstart;
13431 ;; we go forward line by line
13432 (while (and (not (eobp))
13433 (progn (move-to-left-margin) (not (eobp)))
13434 (not (looking-at parsep))
13435 (looking-at fill-prefix-regexp))
13441 This expression moves point forward line by line so long
13442 as four conditions are true:
13446 Point is not at the end of the buffer.
13449 We can move to the left margin of the text and are
13450 not at the end of the buffer.
13453 The text following point does not separate paragraphs.
13456 The pattern following point is the fill prefix regular expression.
13459 The last condition may be puzzling, until you remember that point was
13460 moved to the beginning of the line early in the @code{forward-paragraph}
13461 function. This means that if the text has a fill prefix, the
13462 @code{looking-at} function will see it.
13465 Consider what happens when there is no fill prefix.
13469 (while (and (re-search-forward sp-parstart nil 1)
13470 (progn (setq start (match-beginning 0))
13473 (progn (move-to-left-margin)
13474 (not (looking-at parsep)))
13475 (or (not (looking-at parstart))
13476 (and use-hard-newlines
13477 (not (get-text-property (1- start) 'hard)))))
13483 This @code{while} loop has us searching forward for
13484 @code{sp-parstart}, which is the combination of possible whitespace
13485 with a the local value of the start of a paragraph or of a paragraph
13486 separator. (The latter two are within an expression starting
13487 @code{\(?:} so that they are not referenced by the
13488 @code{match-beginning} function.)
13491 The two expressions,
13495 (setq start (match-beginning 0))
13501 mean go to the start of the text matched by the regular expression
13504 The @code{(match-beginning 0)} expression is new. It returns a number
13505 specifying the location of the start of the text that was matched by
13508 The @code{match-beginning} function is used here because of a
13509 characteristic of a forward search: a successful forward search,
13510 regardless of whether it is a plain search or a regular expression
13511 search, moves point to the end of the text that is found. In this
13512 case, a successful search moves point to the end of the pattern for
13513 @code{sp-parstart}.
13515 However, we want to put point at the end of the current paragraph, not
13516 somewhere else. Indeed, since the search possibly includes the
13517 paragraph separator, point may end up at the beginning of the next one
13518 unless we use an expression that includes @code{match-beginning}.
13520 @findex match-beginning
13521 When given an argument of 0, @code{match-beginning} returns the
13522 position that is the start of the text matched by the most recent
13523 search. In this case, the most recent search looks for
13524 @code{sp-parstart}. The @code{(match-beginning 0)} expression returns
13525 the beginning position of that pattern, rather than the end position
13528 (Incidentally, when passed a positive number as an argument, the
13529 @code{match-beginning} function returns the location of point at that
13530 parenthesized expression in the last search unless that parenthesized
13531 expression begins with @code{\(?:}. I don't know why @code{\(?:}
13532 appears here since the argument is 0.)
13535 The last expression when there is no fill prefix is
13539 (if (< (point) (point-max))
13540 (goto-char start))))
13545 This says that if there is no fill prefix and if we are not at the
13546 end, point should move to the beginning of whatever was found by the
13547 regular expression search for @code{sp-parstart}.
13549 The full definition for the @code{forward-paragraph} function not only
13550 includes code for going forwards, but also code for going backwards.
13552 If you are reading this inside of GNU Emacs and you want to see the
13553 whole function, you can type @kbd{C-h f} (@code{describe-function})
13554 and the name of the function. This gives you the function
13555 documentation and the name of the library containing the function's
13556 source. Place point over the name of the library and press the RET
13557 key; you will be taken directly to the source. (Be sure to install
13558 your sources! Without them, you are like a person who tries to drive
13559 a car with his eyes shut!)
13561 @node etags, Regexp Review, forward-paragraph, Regexp Search
13562 @section Create Your Own @file{TAGS} File
13564 @cindex @file{TAGS} file, create own
13566 Besides @kbd{C-h f} (@code{describe-function}), another way to see the
13567 source of a function is to type @kbd{M-.} (@code{find-tag}) and the
13568 name of the function when prompted for it. This is a good habit to
13569 get into. The @kbd{M-.} (@code{find-tag}) command takes you directly
13570 to the source for a function, variable, or node. The function depends
13571 on tags tables to tell it where to go.
13573 If the @code{find-tag} function first asks you for the name of a
13574 @file{TAGS} table, give it the name of a @file{TAGS} file such as
13575 @file{/usr/local/src/emacs/src/TAGS}. (The exact path to your
13576 @file{TAGS} file depends on how your copy of Emacs was installed. I
13577 just told you the location that provides both my C and my Emacs Lisp
13580 You can also create your own @file{TAGS} file for directories that
13583 You often need to build and install tags tables yourself. They are
13584 not built automatically. A tags table is called a @file{TAGS} file;
13585 the name is in upper case letters.
13587 You can create a @file{TAGS} file by calling the @code{etags} program
13588 that comes as a part of the Emacs distribution. Usually, @code{etags}
13589 is compiled and installed when Emacs is built. (@code{etags} is not
13590 an Emacs Lisp function or a part of Emacs; it is a C program.)
13593 To create a @file{TAGS} file, first switch to the directory in which
13594 you want to create the file. In Emacs you can do this with the
13595 @kbd{M-x cd} command, or by visiting a file in the directory, or by
13596 listing the directory with @kbd{C-x d} (@code{dired}). Then run the
13597 compile command, with @w{@code{etags *.el}} as the command to execute
13600 M-x compile RET etags *.el RET
13604 to create a @file{TAGS} file for Emacs Lisp.
13606 For example, if you have a large number of files in your
13607 @file{~/emacs} directory, as I do---I have 137 @file{.el} files in it,
13608 of which I load 12---you can create a @file{TAGS} file for the Emacs
13609 Lisp files in that directory.
13612 The @code{etags} program takes all the usual shell `wildcards'. For
13613 example, if you have two directories for which you want a single
13614 @file{TAGS} file, type @w{@code{etags *.el ../elisp/*.el}}, where
13615 @file{../elisp/} is the second directory:
13618 M-x compile RET etags *.el ../elisp/*.el RET
13625 M-x compile RET etags --help RET
13629 to see a list of the options accepted by @code{etags} as well as a
13630 list of supported languages.
13632 The @code{etags} program handles more than 20 languages, including
13633 Emacs Lisp, Common Lisp, Scheme, C, C++, Ada, Fortran, HTML, Java,
13634 LaTeX, Pascal, Perl, Postscript, Python, TeX, Texinfo, makefiles, and
13635 most assemblers. The program has no switches for specifying the
13636 language; it recognizes the language in an input file according to its
13637 file name and contents.
13639 @file{etags} is very helpful when you are writing code yourself and
13640 want to refer back to functions you have already written. Just run
13641 @code{etags} again at intervals as you write new functions, so they
13642 become part of the @file{TAGS} file.
13644 If you think an appropriate @file{TAGS} file already exists for what
13645 you want, but do not know where it is, you can use the @code{locate}
13646 program to attempt to find it.
13648 Type @w{@kbd{M-x locate @key{RET} TAGS @key{RET}}} and Emacs will list
13649 for you the full path names of all your @file{TAGS} files. On my
13650 system, this command lists 34 @file{TAGS} files. On the other hand, a
13651 `plain vanilla' system I recently installed did not contain any
13654 If the tags table you want has been created, you can use the @code{M-x
13655 visit-tags-table} command to specify it. Otherwise, you will need to
13656 create the tag table yourself and then use @code{M-x
13659 @subsubheading Building Tags in the Emacs sources
13660 @cindex Building Tags in the Emacs sources
13661 @cindex Tags in the Emacs sources
13664 The GNU Emacs sources come with a @file{Makefile} that contains a
13665 sophisticated @code{etags} command that creates, collects, and merges
13666 tags tables from all over the Emacs sources and puts the information
13667 into one @file{TAGS} file in the @file{src/} directory. (The
13668 @file{src/} directory is below the top level of your Emacs directory.)
13671 To build this @file{TAGS} file, go to the top level of your Emacs
13672 source directory and run the compile command @code{make tags}:
13675 M-x compile RET make tags RET
13679 (The @code{make tags} command works well with the GNU Emacs sources,
13680 as well as with some other source packages.)
13682 For more information, see @ref{Tags, , Tag Tables, emacs, The GNU Emacs
13685 @node Regexp Review, re-search Exercises, etags, Regexp Search
13686 @comment node-name, next, previous, up
13689 Here is a brief summary of some recently introduced functions.
13693 Repeatedly evaluate the body of the expression so long as the first
13694 element of the body tests true. Then return @code{nil}. (The
13695 expression is evaluated only for its side effects.)
13704 (insert (format "foo is %d.\n" foo))
13705 (setq foo (1- foo))))
13707 @result{} foo is 2.
13714 (The @code{insert} function inserts its arguments at point; the
13715 @code{format} function returns a string formatted from its arguments
13716 the way @code{message} formats its arguments; @code{\n} produces a new
13719 @item re-search-forward
13720 Search for a pattern, and if the pattern is found, move point to rest
13724 Takes four arguments, like @code{search-forward}:
13728 A regular expression that specifies the pattern to search for.
13729 (Remember to put quotation marks around this argument!)
13732 Optionally, the limit of the search.
13735 Optionally, what to do if the search fails, return @code{nil} or an
13739 Optionally, how many times to repeat the search; if negative, the
13740 search goes backwards.
13744 Bind some variables locally to particular values,
13745 and then evaluate the remaining arguments, returning the value of the
13746 last one. While binding the local variables, use the local values of
13747 variables bound earlier, if any.
13756 (message "`bar' is %d." bar))
13757 @result{} `bar' is 21.
13761 @item match-beginning
13762 Return the position of the start of the text found by the last regular
13766 Return @code{t} for true if the text after point matches the argument,
13767 which should be a regular expression.
13770 Return @code{t} for true if point is at the end of the accessible part
13771 of a buffer. The end of the accessible part is the end of the buffer
13772 if the buffer is not narrowed; it is the end of the narrowed part if
13773 the buffer is narrowed.
13777 @node re-search Exercises, , Regexp Review, Regexp Search
13778 @section Exercises with @code{re-search-forward}
13782 Write a function to search for a regular expression that matches two
13783 or more blank lines in sequence.
13786 Write a function to search for duplicated words, such as `the the'.
13787 @xref{Regexps, , Syntax of Regular Expressions, emacs, The GNU Emacs
13788 Manual}, for information on how to write a regexp (a regular
13789 expression) to match a string that is composed of two identical
13790 halves. You can devise several regexps; some are better than others.
13791 The function I use is described in an appendix, along with several
13792 regexps. @xref{the-the, , @code{the-the} Duplicated Words Function}.
13795 @node Counting Words, Words in a defun, Regexp Search, Top
13796 @chapter Counting: Repetition and Regexps
13797 @cindex Repetition for word counting
13798 @cindex Regular expressions for word counting
13800 Repetition and regular expression searches are powerful tools that you
13801 often use when you write code in Emacs Lisp. This chapter illustrates
13802 the use of regular expression searches through the construction of
13803 word count commands using @code{while} loops and recursion.
13806 * Why Count Words::
13807 * count-words-region::
13808 * recursive-count-words::
13809 * Counting Exercise::
13812 @node Why Count Words, count-words-region, Counting Words, Counting Words
13814 @unnumberedsec Counting words
13817 The standard Emacs distribution contains a function for counting the
13818 number of lines within a region. However, there is no corresponding
13819 function for counting words.
13821 Certain types of writing ask you to count words. Thus, if you write
13822 an essay, you may be limited to 800 words; if you write a novel, you
13823 may discipline yourself to write 1000 words a day. It seems odd to me
13824 that Emacs lacks a word count command. Perhaps people use Emacs
13825 mostly for code or types of documentation that do not require word
13826 counts; or perhaps they restrict themselves to the operating system
13827 word count command, @code{wc}. Alternatively, people may follow
13828 the publishers' convention and compute a word count by dividing the
13829 number of characters in a document by five. In any event, here are
13830 commands to count words.
13832 @node count-words-region, recursive-count-words, Why Count Words, Counting Words
13833 @comment node-name, next, previous, up
13834 @section The @code{count-words-region} Function
13835 @findex count-words-region
13837 A word count command could count words in a line, paragraph, region,
13838 or buffer. What should the command cover? You could design the
13839 command to count the number of words in a complete buffer. However,
13840 the Emacs tradition encourages flexibility---you may want to count
13841 words in just a section, rather than all of a buffer. So it makes
13842 more sense to design the command to count the number of words in a
13843 region. Once you have a @code{count-words-region} command, you can,
13844 if you wish, count words in a whole buffer by marking it with
13845 @w{@kbd{C-x h}} (@code{mark-whole-buffer}).
13847 Clearly, counting words is a repetitive act: starting from the
13848 beginning of the region, you count the first word, then the second
13849 word, then the third word, and so on, until you reach the end of the
13850 region. This means that word counting is ideally suited to recursion
13851 or to a @code{while} loop.
13854 * Design count-words-region::
13858 @node Design count-words-region, Whitespace Bug, count-words-region, count-words-region
13860 @unnumberedsubsec Designing @code{count-words-region}
13863 First, we will implement the word count command with a @code{while}
13864 loop, then with recursion. The command will, of course, be
13868 The template for an interactive function definition is, as always:
13872 (defun @var{name-of-function} (@var{argument-list})
13873 "@var{documentation}@dots{}"
13874 (@var{interactive-expression}@dots{})
13879 What we need to do is fill in the slots.
13881 The name of the function should be self-explanatory and similar to the
13882 existing @code{count-lines-region} name. This makes the name easier
13883 to remember. @code{count-words-region} is a good choice.
13885 The function counts words within a region. This means that the
13886 argument list must contain symbols that are bound to the two
13887 positions, the beginning and end of the region. These two positions
13888 can be called @samp{beginning} and @samp{end} respectively. The first
13889 line of the documentation should be a single sentence, since that is
13890 all that is printed as documentation by a command such as
13891 @code{apropos}. The interactive expression will be of the form
13892 @samp{(interactive "r")}, since that will cause Emacs to pass the
13893 beginning and end of the region to the function's argument list. All
13896 The body of the function needs to be written to do three tasks:
13897 first, to set up conditions under which the @code{while} loop can
13898 count words, second, to run the @code{while} loop, and third, to send
13899 a message to the user.
13901 When a user calls @code{count-words-region}, point may be at the
13902 beginning or the end of the region. However, the counting process
13903 must start at the beginning of the region. This means we will want
13904 to put point there if it is not already there. Executing
13905 @code{(goto-char beginning)} ensures this. Of course, we will want to
13906 return point to its expected position when the function finishes its
13907 work. For this reason, the body must be enclosed in a
13908 @code{save-excursion} expression.
13910 The central part of the body of the function consists of a
13911 @code{while} loop in which one expression jumps point forward word by
13912 word, and another expression counts those jumps. The true-or-false-test
13913 of the @code{while} loop should test true so long as point should jump
13914 forward, and false when point is at the end of the region.
13916 We could use @code{(forward-word 1)} as the expression for moving point
13917 forward word by word, but it is easier to see what Emacs identifies as a
13918 `word' if we use a regular expression search.
13920 A regular expression search that finds the pattern for which it is
13921 searching leaves point after the last character matched. This means
13922 that a succession of successful word searches will move point forward
13925 As a practical matter, we want the regular expression search to jump
13926 over whitespace and punctuation between words as well as over the
13927 words themselves. A regexp that refuses to jump over interword
13928 whitespace would never jump more than one word! This means that
13929 the regexp should include the whitespace and punctuation that follows
13930 a word, if any, as well as the word itself. (A word may end a buffer
13931 and not have any following whitespace or punctuation, so that part of
13932 the regexp must be optional.)
13934 Thus, what we want for the regexp is a pattern defining one or more
13935 word constituent characters followed, optionally, by one or more
13936 characters that are not word constituents. The regular expression for
13944 The buffer's syntax table determines which characters are and are not
13945 word constituents. (@xref{Syntax, , What Constitutes a Word or
13946 Symbol?}, for more about syntax. Also, see @ref{Syntax, Syntax, The
13947 Syntax Table, emacs, The GNU Emacs Manual}, and @ref{Syntax Tables, ,
13948 Syntax Tables, elisp, The GNU Emacs Lisp Reference Manual}.)
13951 The search expression looks like this:
13954 (re-search-forward "\\w+\\W*")
13958 (Note that paired backslashes precede the @samp{w} and @samp{W}. A
13959 single backslash has special meaning to the Emacs Lisp interpreter.
13960 It indicates that the following character is interpreted differently
13961 than usual. For example, the two characters, @samp{\n}, stand for
13962 @samp{newline}, rather than for a backslash followed by @samp{n}. Two
13963 backslashes in a row stand for an ordinary, `unspecial' backslash, so
13964 Emacs Lisp interpreter ends of seeing a single backslash followed by a
13965 letter. So it discovers the letter is special.)
13967 We need a counter to count how many words there are; this variable
13968 must first be set to 0 and then incremented each time Emacs goes
13969 around the @code{while} loop. The incrementing expression is simply:
13972 (setq count (1+ count))
13975 Finally, we want to tell the user how many words there are in the
13976 region. The @code{message} function is intended for presenting this
13977 kind of information to the user. The message has to be phrased so
13978 that it reads properly regardless of how many words there are in the
13979 region: we don't want to say that ``there are 1 words in the region''.
13980 The conflict between singular and plural is ungrammatical. We can
13981 solve this problem by using a conditional expression that evaluates
13982 different messages depending on the number of words in the region.
13983 There are three possibilities: no words in the region, one word in the
13984 region, and more than one word. This means that the @code{cond}
13985 special form is appropriate.
13988 All this leads to the following function definition:
13992 ;;; @r{First version; has bugs!}
13993 (defun count-words-region (beginning end)
13994 "Print number of words in the region.
13995 Words are defined as at least one word-constituent
13996 character followed by at least one character that
13997 is not a word-constituent. The buffer's syntax
13998 table determines which characters these are."
14000 (message "Counting words in region ... ")
14004 ;;; @r{1. Set up appropriate conditions.}
14006 (goto-char beginning)
14011 ;;; @r{2. Run the} while @r{loop.}
14012 (while (< (point) end)
14013 (re-search-forward "\\w+\\W*")
14014 (setq count (1+ count)))
14018 ;;; @r{3. Send a message to the user.}
14019 (cond ((zerop count)
14021 "The region does NOT have any words."))
14024 "The region has 1 word."))
14027 "The region has %d words." count))))))
14032 As written, the function works, but not in all circumstances.
14034 @node Whitespace Bug, , Design count-words-region, count-words-region
14035 @comment node-name, next, previous, up
14036 @subsection The Whitespace Bug in @code{count-words-region}
14038 The @code{count-words-region} command described in the preceding
14039 section has two bugs, or rather, one bug with two manifestations.
14040 First, if you mark a region containing only whitespace in the middle
14041 of some text, the @code{count-words-region} command tells you that the
14042 region contains one word! Second, if you mark a region containing
14043 only whitespace at the end of the buffer or the accessible portion of
14044 a narrowed buffer, the command displays an error message that looks
14048 Search failed: "\\w+\\W*"
14051 If you are reading this in Info in GNU Emacs, you can test for these
14054 First, evaluate the function in the usual manner to install it.
14056 Here is a copy of the definition. Place your cursor after the closing
14057 parenthesis and type @kbd{C-x C-e} to install it.
14061 ;; @r{First version; has bugs!}
14062 (defun count-words-region (beginning end)
14063 "Print number of words in the region.
14064 Words are defined as at least one word-constituent character followed
14065 by at least one character that is not a word-constituent. The buffer's
14066 syntax table determines which characters these are."
14070 (message "Counting words in region ... ")
14074 ;;; @r{1. Set up appropriate conditions.}
14076 (goto-char beginning)
14081 ;;; @r{2. Run the} while @r{loop.}
14082 (while (< (point) end)
14083 (re-search-forward "\\w+\\W*")
14084 (setq count (1+ count)))
14088 ;;; @r{3. Send a message to the user.}
14089 (cond ((zerop count)
14090 (message "The region does NOT have any words."))
14091 ((= 1 count) (message "The region has 1 word."))
14092 (t (message "The region has %d words." count))))))
14098 If you wish, you can also install this keybinding by evaluating it:
14101 (global-set-key "\C-c=" 'count-words-region)
14104 To conduct the first test, set mark and point to the beginning and end
14105 of the following line and then type @kbd{C-c =} (or @kbd{M-x
14106 count-words-region} if you have not bound @kbd{C-c =}):
14113 Emacs will tell you, correctly, that the region has three words.
14115 Repeat the test, but place mark at the beginning of the line and place
14116 point just @emph{before} the word @samp{one}. Again type the command
14117 @kbd{C-c =} (or @kbd{M-x count-words-region}). Emacs should tell you
14118 that the region has no words, since it is composed only of the
14119 whitespace at the beginning of the line. But instead Emacs tells you
14120 that the region has one word!
14122 For the third test, copy the sample line to the end of the
14123 @file{*scratch*} buffer and then type several spaces at the end of the
14124 line. Place mark right after the word @samp{three} and point at the
14125 end of line. (The end of the line will be the end of the buffer.)
14126 Type @kbd{C-c =} (or @kbd{M-x count-words-region}) as you did before.
14127 Again, Emacs should tell you that the region has no words, since it is
14128 composed only of the whitespace at the end of the line. Instead,
14129 Emacs displays an error message saying @samp{Search failed}.
14131 The two bugs stem from the same problem.
14133 Consider the first manifestation of the bug, in which the command
14134 tells you that the whitespace at the beginning of the line contains
14135 one word. What happens is this: The @code{M-x count-words-region}
14136 command moves point to the beginning of the region. The @code{while}
14137 tests whether the value of point is smaller than the value of
14138 @code{end}, which it is. Consequently, the regular expression search
14139 looks for and finds the first word. It leaves point after the word.
14140 @code{count} is set to one. The @code{while} loop repeats; but this
14141 time the value of point is larger than the value of @code{end}, the
14142 loop is exited; and the function displays a message saying the number
14143 of words in the region is one. In brief, the regular expression
14144 search looks for and finds the word even though it is outside
14147 In the second manifestation of the bug, the region is whitespace at
14148 the end of the buffer. Emacs says @samp{Search failed}. What happens
14149 is that the true-or-false-test in the @code{while} loop tests true, so
14150 the search expression is executed. But since there are no more words
14151 in the buffer, the search fails.
14153 In both manifestations of the bug, the search extends or attempts to
14154 extend outside of the region.
14156 The solution is to limit the search to the region---this is a fairly
14157 simple action, but as you may have come to expect, it is not quite as
14158 simple as you might think.
14160 As we have seen, the @code{re-search-forward} function takes a search
14161 pattern as its first argument. But in addition to this first,
14162 mandatory argument, it accepts three optional arguments. The optional
14163 second argument bounds the search. The optional third argument, if
14164 @code{t}, causes the function to return @code{nil} rather than signal
14165 an error if the search fails. The optional fourth argument is a
14166 repeat count. (In Emacs, you can see a function's documentation by
14167 typing @kbd{C-h f}, the name of the function, and then @key{RET}.)
14169 In the @code{count-words-region} definition, the value of the end of
14170 the region is held by the variable @code{end} which is passed as an
14171 argument to the function. Thus, we can add @code{end} as an argument
14172 to the regular expression search expression:
14175 (re-search-forward "\\w+\\W*" end)
14178 However, if you make only this change to the @code{count-words-region}
14179 definition and then test the new version of the definition on a
14180 stretch of whitespace, you will receive an error message saying
14181 @samp{Search failed}.
14183 What happens is this: the search is limited to the region, and fails
14184 as you expect because there are no word-constituent characters in the
14185 region. Since it fails, we receive an error message. But we do not
14186 want to receive an error message in this case; we want to receive the
14187 message that "The region does NOT have any words."
14189 The solution to this problem is to provide @code{re-search-forward}
14190 with a third argument of @code{t}, which causes the function to return
14191 @code{nil} rather than signal an error if the search fails.
14193 However, if you make this change and try it, you will see the message
14194 ``Counting words in region ... '' and @dots{} you will keep on seeing
14195 that message @dots{}, until you type @kbd{C-g} (@code{keyboard-quit}).
14197 Here is what happens: the search is limited to the region, as before,
14198 and it fails because there are no word-constituent characters in the
14199 region, as expected. Consequently, the @code{re-search-forward}
14200 expression returns @code{nil}. It does nothing else. In particular,
14201 it does not move point, which it does as a side effect if it finds the
14202 search target. After the @code{re-search-forward} expression returns
14203 @code{nil}, the next expression in the @code{while} loop is evaluated.
14204 This expression increments the count. Then the loop repeats. The
14205 true-or-false-test tests true because the value of point is still less
14206 than the value of end, since the @code{re-search-forward} expression
14207 did not move point. @dots{} and the cycle repeats @dots{}
14209 The @code{count-words-region} definition requires yet another
14210 modification, to cause the true-or-false-test of the @code{while} loop
14211 to test false if the search fails. Put another way, there are two
14212 conditions that must be satisfied in the true-or-false-test before the
14213 word count variable is incremented: point must still be within the
14214 region and the search expression must have found a word to count.
14216 Since both the first condition and the second condition must be true
14217 together, the two expressions, the region test and the search
14218 expression, can be joined with an @code{and} special form and embedded in
14219 the @code{while} loop as the true-or-false-test, like this:
14222 (and (< (point) end) (re-search-forward "\\w+\\W*" end t))
14225 @c colon in printed section title causes problem in Info cross reference
14226 @c also trouble with an overfull hbox
14229 (For information about @code{and}, see
14230 @ref{kill-new function, , The @code{kill-new} function}.)
14234 (@xref{kill-new function, , The @code{kill-new} function}, for
14235 information about @code{and}.)
14238 The @code{re-search-forward} expression returns @code{t} if the search
14239 succeeds and as a side effect moves point. Consequently, as words are
14240 found, point is moved through the region. When the search expression
14241 fails to find another word, or when point reaches the end of the
14242 region, the true-or-false-test tests false, the @code{while} loop
14243 exits, and the @code{count-words-region} function displays one or
14244 other of its messages.
14246 After incorporating these final changes, the @code{count-words-region}
14247 works without bugs (or at least, without bugs that I have found!).
14248 Here is what it looks like:
14252 ;;; @r{Final version:} @code{while}
14253 (defun count-words-region (beginning end)
14254 "Print number of words in the region."
14256 (message "Counting words in region ... ")
14260 ;;; @r{1. Set up appropriate conditions.}
14263 (goto-char beginning)
14267 ;;; @r{2. Run the} while @r{loop.}
14268 (while (and (< (point) end)
14269 (re-search-forward "\\w+\\W*" end t))
14270 (setq count (1+ count)))
14274 ;;; @r{3. Send a message to the user.}
14275 (cond ((zerop count)
14277 "The region does NOT have any words."))
14280 "The region has 1 word."))
14283 "The region has %d words." count))))))
14287 @node recursive-count-words, Counting Exercise, count-words-region, Counting Words
14288 @comment node-name, next, previous, up
14289 @section Count Words Recursively
14290 @cindex Count words recursively
14291 @cindex Recursively counting words
14292 @cindex Words, counted recursively
14294 You can write the function for counting words recursively as well as
14295 with a @code{while} loop. Let's see how this is done.
14297 First, we need to recognize that the @code{count-words-region}
14298 function has three jobs: it sets up the appropriate conditions for
14299 counting to occur; it counts the words in the region; and it sends a
14300 message to the user telling how many words there are.
14302 If we write a single recursive function to do everything, we will
14303 receive a message for every recursive call. If the region contains 13
14304 words, we will receive thirteen messages, one right after the other.
14305 We don't want this! Instead, we must write two functions to do the
14306 job, one of which (the recursive function) will be used inside of the
14307 other. One function will set up the conditions and display the
14308 message; the other will return the word count.
14310 Let us start with the function that causes the message to be displayed.
14311 We can continue to call this @code{count-words-region}.
14313 This is the function that the user will call. It will be interactive.
14314 Indeed, it will be similar to our previous versions of this
14315 function, except that it will call @code{recursive-count-words} to
14316 determine how many words are in the region.
14319 We can readily construct a template for this function, based on our
14324 ;; @r{Recursive version; uses regular expression search}
14325 (defun count-words-region (beginning end)
14326 "@var{documentation}@dots{}"
14327 (@var{interactive-expression}@dots{})
14331 ;;; @r{1. Set up appropriate conditions.}
14332 (@var{explanatory message})
14333 (@var{set-up functions}@dots{}
14337 ;;; @r{2. Count the words.}
14338 @var{recursive call}
14342 ;;; @r{3. Send a message to the user.}
14343 @var{message providing word count}))
14347 The definition looks straightforward, except that somehow the count
14348 returned by the recursive call must be passed to the message
14349 displaying the word count. A little thought suggests that this can be
14350 done by making use of a @code{let} expression: we can bind a variable
14351 in the varlist of a @code{let} expression to the number of words in
14352 the region, as returned by the recursive call; and then the
14353 @code{cond} expression, using binding, can display the value to the
14356 Often, one thinks of the binding within a @code{let} expression as
14357 somehow secondary to the `primary' work of a function. But in this
14358 case, what you might consider the `primary' job of the function,
14359 counting words, is done within the @code{let} expression.
14362 Using @code{let}, the function definition looks like this:
14366 (defun count-words-region (beginning end)
14367 "Print number of words in the region."
14372 ;;; @r{1. Set up appropriate conditions.}
14373 (message "Counting words in region ... ")
14375 (goto-char beginning)
14379 ;;; @r{2. Count the words.}
14380 (let ((count (recursive-count-words end)))
14384 ;;; @r{3. Send a message to the user.}
14385 (cond ((zerop count)
14387 "The region does NOT have any words."))
14390 "The region has 1 word."))
14393 "The region has %d words." count))))))
14397 Next, we need to write the recursive counting function.
14399 A recursive function has at least three parts: the `do-again-test', the
14400 `next-step-expression', and the recursive call.
14402 The do-again-test determines whether the function will or will not be
14403 called again. Since we are counting words in a region and can use a
14404 function that moves point forward for every word, the do-again-test
14405 can check whether point is still within the region. The do-again-test
14406 should find the value of point and determine whether point is before,
14407 at, or after the value of the end of the region. We can use the
14408 @code{point} function to locate point. Clearly, we must pass the
14409 value of the end of the region to the recursive counting function as an
14412 In addition, the do-again-test should also test whether the search finds a
14413 word. If it does not, the function should not call itself again.
14415 The next-step-expression changes a value so that when the recursive
14416 function is supposed to stop calling itself, it stops. More
14417 precisely, the next-step-expression changes a value so that at the
14418 right time, the do-again-test stops the recursive function from
14419 calling itself again. In this case, the next-step-expression can be
14420 the expression that moves point forward, word by word.
14422 The third part of a recursive function is the recursive call.
14424 Somewhere, also, we also need a part that does the `work' of the
14425 function, a part that does the counting. A vital part!
14428 But already, we have an outline of the recursive counting function:
14432 (defun recursive-count-words (region-end)
14433 "@var{documentation}@dots{}"
14434 @var{do-again-test}
14435 @var{next-step-expression}
14436 @var{recursive call})
14440 Now we need to fill in the slots. Let's start with the simplest cases
14441 first: if point is at or beyond the end of the region, there cannot
14442 be any words in the region, so the function should return zero.
14443 Likewise, if the search fails, there are no words to count, so the
14444 function should return zero.
14446 On the other hand, if point is within the region and the search
14447 succeeds, the function should call itself again.
14450 Thus, the do-again-test should look like this:
14454 (and (< (point) region-end)
14455 (re-search-forward "\\w+\\W*" region-end t))
14459 Note that the search expression is part of the do-again-test---the
14460 function returns @code{t} if its search succeeds and @code{nil} if it
14461 fails. (@xref{Whitespace Bug, , The Whitespace Bug in
14462 @code{count-words-region}}, for an explanation of how
14463 @code{re-search-forward} works.)
14465 The do-again-test is the true-or-false test of an @code{if} clause.
14466 Clearly, if the do-again-test succeeds, the then-part of the @code{if}
14467 clause should call the function again; but if it fails, the else-part
14468 should return zero since either point is outside the region or the
14469 search failed because there were no words to find.
14471 But before considering the recursive call, we need to consider the
14472 next-step-expression. What is it? Interestingly, it is the search
14473 part of the do-again-test.
14475 In addition to returning @code{t} or @code{nil} for the
14476 do-again-test, @code{re-search-forward} moves point forward as a side
14477 effect of a successful search. This is the action that changes the
14478 value of point so that the recursive function stops calling itself
14479 when point completes its movement through the region. Consequently,
14480 the @code{re-search-forward} expression is the next-step-expression.
14483 In outline, then, the body of the @code{recursive-count-words}
14484 function looks like this:
14488 (if @var{do-again-test-and-next-step-combined}
14490 @var{recursive-call-returning-count}
14496 How to incorporate the mechanism that counts?
14498 If you are not used to writing recursive functions, a question like
14499 this can be troublesome. But it can and should be approached
14502 We know that the counting mechanism should be associated in some way
14503 with the recursive call. Indeed, since the next-step-expression moves
14504 point forward by one word, and since a recursive call is made for
14505 each word, the counting mechanism must be an expression that adds one
14506 to the value returned by a call to @code{recursive-count-words}.
14509 Consider several cases:
14513 If there are two words in the region, the function should return
14514 a value resulting from adding one to the value returned when it counts
14515 the first word, plus the number returned when it counts the remaining
14516 words in the region, which in this case is one.
14519 If there is one word in the region, the function should return
14520 a value resulting from adding one to the value returned when it counts
14521 that word, plus the number returned when it counts the remaining
14522 words in the region, which in this case is zero.
14525 If there are no words in the region, the function should return zero.
14528 From the sketch we can see that the else-part of the @code{if} returns
14529 zero for the case of no words. This means that the then-part of the
14530 @code{if} must return a value resulting from adding one to the value
14531 returned from a count of the remaining words.
14534 The expression will look like this, where @code{1+} is a function that
14535 adds one to its argument.
14538 (1+ (recursive-count-words region-end))
14542 The whole @code{recursive-count-words} function will then look like
14547 (defun recursive-count-words (region-end)
14548 "@var{documentation}@dots{}"
14550 ;;; @r{1. do-again-test}
14551 (if (and (< (point) region-end)
14552 (re-search-forward "\\w+\\W*" region-end t))
14556 ;;; @r{2. then-part: the recursive call}
14557 (1+ (recursive-count-words region-end))
14559 ;;; @r{3. else-part}
14565 Let's examine how this works:
14567 If there are no words in the region, the else part of the @code{if}
14568 expression is evaluated and consequently the function returns zero.
14570 If there is one word in the region, the value of point is less than
14571 the value of @code{region-end} and the search succeeds. In this case,
14572 the true-or-false-test of the @code{if} expression tests true, and the
14573 then-part of the @code{if} expression is evaluated. The counting
14574 expression is evaluated. This expression returns a value (which will
14575 be the value returned by the whole function) that is the sum of one
14576 added to the value returned by a recursive call.
14578 Meanwhile, the next-step-expression has caused point to jump over the
14579 first (and in this case only) word in the region. This means that
14580 when @code{(recursive-count-words region-end)} is evaluated a second
14581 time, as a result of the recursive call, the value of point will be
14582 equal to or greater than the value of region end. So this time,
14583 @code{recursive-count-words} will return zero. The zero will be added
14584 to one, and the original evaluation of @code{recursive-count-words}
14585 will return one plus zero, which is one, which is the correct amount.
14587 Clearly, if there are two words in the region, the first call to
14588 @code{recursive-count-words} returns one added to the value returned
14589 by calling @code{recursive-count-words} on a region containing the
14590 remaining word---that is, it adds one to one, producing two, which is
14591 the correct amount.
14593 Similarly, if there are three words in the region, the first call to
14594 @code{recursive-count-words} returns one added to the value returned
14595 by calling @code{recursive-count-words} on a region containing the
14596 remaining two words---and so on and so on.
14600 With full documentation the two functions look like this:
14604 The recursive function:
14606 @findex recursive-count-words
14609 (defun recursive-count-words (region-end)
14610 "Number of words between point and REGION-END."
14614 ;;; @r{1. do-again-test}
14615 (if (and (< (point) region-end)
14616 (re-search-forward "\\w+\\W*" region-end t))
14620 ;;; @r{2. then-part: the recursive call}
14621 (1+ (recursive-count-words region-end))
14623 ;;; @r{3. else-part}
14634 ;;; @r{Recursive version}
14635 (defun count-words-region (beginning end)
14636 "Print number of words in the region.
14640 Words are defined as at least one word-constituent
14641 character followed by at least one character that is
14642 not a word-constituent. The buffer's syntax table
14643 determines which characters these are."
14647 (message "Counting words in region ... ")
14649 (goto-char beginning)
14650 (let ((count (recursive-count-words end)))
14653 (cond ((zerop count)
14655 "The region does NOT have any words."))
14659 (message "The region has 1 word."))
14662 "The region has %d words." count))))))
14666 @node Counting Exercise, , recursive-count-words, Counting Words
14667 @section Exercise: Counting Punctuation
14669 Using a @code{while} loop, write a function to count the number of
14670 punctuation marks in a region---period, comma, semicolon, colon,
14671 exclamation mark, and question mark. Do the same using recursion.
14673 @node Words in a defun, Readying a Graph, Counting Words, Top
14674 @chapter Counting Words in a @code{defun}
14675 @cindex Counting words in a @code{defun}
14676 @cindex Word counting in a @code{defun}
14678 Our next project is to count the number of words in a function
14679 definition. Clearly, this can be done using some variant of
14680 @code{count-word-region}. @xref{Counting Words, , Counting Words:
14681 Repetition and Regexps}. If we are just going to count the words in
14682 one definition, it is easy enough to mark the definition with the
14683 @kbd{C-M-h} (@code{mark-defun}) command, and then call
14684 @code{count-word-region}.
14686 However, I am more ambitious: I want to count the words and symbols in
14687 every definition in the Emacs sources and then print a graph that
14688 shows how many functions there are of each length: how many contain 40
14689 to 49 words or symbols, how many contain 50 to 59 words or symbols,
14690 and so on. I have often been curious how long a typical function is,
14691 and this will tell.
14694 * Divide and Conquer::
14695 * Words and Symbols::
14697 * count-words-in-defun::
14700 * lengths-list-file::
14702 * Several files recursively::
14703 * Prepare the data::
14706 @node Divide and Conquer, Words and Symbols, Words in a defun, Words in a defun
14708 @unnumberedsec Divide and Conquer
14711 Described in one phrase, the histogram project is daunting; but
14712 divided into numerous small steps, each of which we can take one at a
14713 time, the project becomes less fearsome. Let us consider what the
14718 First, write a function to count the words in one definition. This
14719 includes the problem of handling symbols as well as words.
14722 Second, write a function to list the numbers of words in each function
14723 in a file. This function can use the @code{count-words-in-defun}
14727 Third, write a function to list the numbers of words in each function
14728 in each of several files. This entails automatically finding the
14729 various files, switching to them, and counting the words in the
14730 definitions within them.
14733 Fourth, write a function to convert the list of numbers that we
14734 created in step three to a form that will be suitable for printing as
14738 Fifth, write a function to print the results as a graph.
14741 This is quite a project! But if we take each step slowly, it will not
14744 @node Words and Symbols, Syntax, Divide and Conquer, Words in a defun
14745 @section What to Count?
14746 @cindex Words and symbols in defun
14748 When we first start thinking about how to count the words in a
14749 function definition, the first question is (or ought to be) what are
14750 we going to count? When we speak of `words' with respect to a Lisp
14751 function definition, we are actually speaking, in large part, of
14752 `symbols'. For example, the following @code{multiply-by-seven}
14753 function contains the five symbols @code{defun},
14754 @code{multiply-by-seven}, @code{number}, @code{*}, and @code{7}. In
14755 addition, in the documentation string, it contains the four words
14756 @samp{Multiply}, @samp{NUMBER}, @samp{by}, and @samp{seven}. The
14757 symbol @samp{number} is repeated, so the definition contains a total
14758 of ten words and symbols.
14762 (defun multiply-by-seven (number)
14763 "Multiply NUMBER by seven."
14769 However, if we mark the @code{multiply-by-seven} definition with
14770 @kbd{C-M-h} (@code{mark-defun}), and then call
14771 @code{count-words-region} on it, we will find that
14772 @code{count-words-region} claims the definition has eleven words, not
14773 ten! Something is wrong!
14775 The problem is twofold: @code{count-words-region} does not count the
14776 @samp{*} as a word, and it counts the single symbol,
14777 @code{multiply-by-seven}, as containing three words. The hyphens are
14778 treated as if they were interword spaces rather than intraword
14779 connectors: @samp{multiply-by-seven} is counted as if it were written
14780 @samp{multiply by seven}.
14782 The cause of this confusion is the regular expression search within
14783 the @code{count-words-region} definition that moves point forward word
14784 by word. In the canonical version of @code{count-words-region}, the
14792 This regular expression is a pattern defining one or more word
14793 constituent characters possibly followed by one or more characters
14794 that are not word constituents. What is meant by `word constituent
14795 characters' brings us to the issue of syntax, which is worth a section
14798 @node Syntax, count-words-in-defun, Words and Symbols, Words in a defun
14799 @section What Constitutes a Word or Symbol?
14800 @cindex Syntax categories and tables
14802 Emacs treats different characters as belonging to different
14803 @dfn{syntax categories}. For example, the regular expression,
14804 @samp{\\w+}, is a pattern specifying one or more @emph{word
14805 constituent} characters. Word constituent characters are members of
14806 one syntax category. Other syntax categories include the class of
14807 punctuation characters, such as the period and the comma, and the
14808 class of whitespace characters, such as the blank space and the tab
14809 character. (For more information, see @ref{Syntax, Syntax, The Syntax
14810 Table, emacs, The GNU Emacs Manual}, and @ref{Syntax Tables, , Syntax
14811 Tables, elisp, The GNU Emacs Lisp Reference Manual}.)
14813 Syntax tables specify which characters belong to which categories.
14814 Usually, a hyphen is not specified as a `word constituent character'.
14815 Instead, it is specified as being in the `class of characters that are
14816 part of symbol names but not words.' This means that the
14817 @code{count-words-region} function treats it in the same way it treats
14818 an interword white space, which is why @code{count-words-region}
14819 counts @samp{multiply-by-seven} as three words.
14821 There are two ways to cause Emacs to count @samp{multiply-by-seven} as
14822 one symbol: modify the syntax table or modify the regular expression.
14824 We could redefine a hyphen as a word constituent character by
14825 modifying the syntax table that Emacs keeps for each mode. This
14826 action would serve our purpose, except that a hyphen is merely the
14827 most common character within symbols that is not typically a word
14828 constituent character; there are others, too.
14830 Alternatively, we can redefine the regular expression used in the
14831 @code{count-words} definition so as to include symbols. This
14832 procedure has the merit of clarity, but the task is a little tricky.
14835 The first part is simple enough: the pattern must match ``at least one
14836 character that is a word or symbol constituent''. Thus:
14839 "\\(\\w\\|\\s_\\)+"
14843 The @samp{\\(} is the first part of the grouping construct that
14844 includes the @samp{\\w} and the @samp{\\s_} as alternatives, separated
14845 by the @samp{\\|}. The @samp{\\w} matches any word-constituent
14846 character and the @samp{\\s_} matches any character that is part of a
14847 symbol name but not a word-constituent character. The @samp{+}
14848 following the group indicates that the word or symbol constituent
14849 characters must be matched at least once.
14851 However, the second part of the regexp is more difficult to design.
14852 What we want is to follow the first part with ``optionally one or more
14853 characters that are not constituents of a word or symbol''. At first,
14854 I thought I could define this with the following:
14857 "\\(\\W\\|\\S_\\)*"
14861 The upper case @samp{W} and @samp{S} match characters that are
14862 @emph{not} word or symbol constituents. Unfortunately, this
14863 expression matches any character that is either not a word constituent
14864 or not a symbol constituent. This matches any character!
14866 I then noticed that every word or symbol in my test region was
14867 followed by white space (blank space, tab, or newline). So I tried
14868 placing a pattern to match one or more blank spaces after the pattern
14869 for one or more word or symbol constituents. This failed, too. Words
14870 and symbols are often separated by whitespace, but in actual code
14871 parentheses may follow symbols and punctuation may follow words. So
14872 finally, I designed a pattern in which the word or symbol constituents
14873 are followed optionally by characters that are not white space and
14874 then followed optionally by white space.
14877 Here is the full regular expression:
14880 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*"
14883 @node count-words-in-defun, Several defuns, Syntax, Words in a defun
14884 @section The @code{count-words-in-defun} Function
14885 @cindex Counting words in a @code{defun}
14887 We have seen that there are several ways to write a
14888 @code{count-word-region} function. To write a
14889 @code{count-words-in-defun}, we need merely adapt one of these
14892 The version that uses a @code{while} loop is easy to understand, so I
14893 am going to adapt that. Because @code{count-words-in-defun} will be
14894 part of a more complex program, it need not be interactive and it need
14895 not display a message but just return the count. These considerations
14896 simplify the definition a little.
14898 On the other hand, @code{count-words-in-defun} will be used within a
14899 buffer that contains function definitions. Consequently, it is
14900 reasonable to ask that the function determine whether it is called
14901 when point is within a function definition, and if it is, to return
14902 the count for that definition. This adds complexity to the
14903 definition, but saves us from needing to pass arguments to the
14907 These considerations lead us to prepare the following template:
14911 (defun count-words-in-defun ()
14912 "@var{documentation}@dots{}"
14913 (@var{set up}@dots{}
14914 (@var{while loop}@dots{})
14915 @var{return count})
14920 As usual, our job is to fill in the slots.
14924 We are presuming that this function will be called within a buffer
14925 containing function definitions. Point will either be within a
14926 function definition or not. For @code{count-words-in-defun} to work,
14927 point must move to the beginning of the definition, a counter must
14928 start at zero, and the counting loop must stop when point reaches the
14929 end of the definition.
14931 The @code{beginning-of-defun} function searches backwards for an
14932 opening delimiter such as a @samp{(} at the beginning of a line, and
14933 moves point to that position, or else to the limit of the search. In
14934 practice, this means that @code{beginning-of-defun} moves point to the
14935 beginning of an enclosing or preceding function definition, or else to
14936 the beginning of the buffer. We can use @code{beginning-of-defun} to
14937 place point where we wish to start.
14939 The @code{while} loop requires a counter to keep track of the words or
14940 symbols being counted. A @code{let} expression can be used to create
14941 a local variable for this purpose, and bind it to an initial value of zero.
14943 The @code{end-of-defun} function works like @code{beginning-of-defun}
14944 except that it moves point to the end of the definition.
14945 @code{end-of-defun} can be used as part of an expression that
14946 determines the position of the end of the definition.
14948 The set up for @code{count-words-in-defun} takes shape rapidly: first
14949 we move point to the beginning of the definition, then we create a
14950 local variable to hold the count, and finally, we record the position
14951 of the end of the definition so the @code{while} loop will know when to stop
14955 The code looks like this:
14959 (beginning-of-defun)
14961 (end (save-excursion (end-of-defun) (point))))
14966 The code is simple. The only slight complication is likely to concern
14967 @code{end}: it is bound to the position of the end of the definition
14968 by a @code{save-excursion} expression that returns the value of point
14969 after @code{end-of-defun} temporarily moves it to the end of the
14972 The second part of the @code{count-words-in-defun}, after the set up,
14973 is the @code{while} loop.
14975 The loop must contain an expression that jumps point forward word by
14976 word and symbol by symbol, and another expression that counts the
14977 jumps. The true-or-false-test for the @code{while} loop should test
14978 true so long as point should jump forward, and false when point is at
14979 the end of the definition. We have already redefined the regular
14980 expression for this (@pxref{Syntax}), so the loop is straightforward:
14984 (while (and (< (point) end)
14986 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*" end t)
14987 (setq count (1+ count)))
14991 The third part of the function definition returns the count of words
14992 and symbols. This part is the last expression within the body of the
14993 @code{let} expression, and can be, very simply, the local variable
14994 @code{count}, which when evaluated returns the count.
14997 Put together, the @code{count-words-in-defun} definition looks like this:
14999 @findex count-words-in-defun
15002 (defun count-words-in-defun ()
15003 "Return the number of words and symbols in a defun."
15004 (beginning-of-defun)
15006 (end (save-excursion (end-of-defun) (point))))
15010 (and (< (point) end)
15012 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*"
15014 (setq count (1+ count)))
15019 How to test this? The function is not interactive, but it is easy to
15020 put a wrapper around the function to make it interactive; we can use
15021 almost the same code as for the recursive version of
15022 @code{count-words-region}:
15026 ;;; @r{Interactive version.}
15027 (defun count-words-defun ()
15028 "Number of words and symbols in a function definition."
15031 "Counting words and symbols in function definition ... ")
15034 (let ((count (count-words-in-defun)))
15038 "The definition does NOT have any words or symbols."))
15043 "The definition has 1 word or symbol."))
15046 "The definition has %d words or symbols." count)))))
15052 Let's re-use @kbd{C-c =} as a convenient keybinding:
15055 (global-set-key "\C-c=" 'count-words-defun)
15058 Now we can try out @code{count-words-defun}: install both
15059 @code{count-words-in-defun} and @code{count-words-defun}, and set the
15060 keybinding, and then place the cursor within the following definition:
15064 (defun multiply-by-seven (number)
15065 "Multiply NUMBER by seven."
15072 Success! The definition has 10 words and symbols.
15074 The next problem is to count the numbers of words and symbols in
15075 several definitions within a single file.
15077 @node Several defuns, Find a File, count-words-in-defun, Words in a defun
15078 @section Count Several @code{defuns} Within a File
15080 A file such as @file{simple.el} may have a hundred or more function
15081 definitions within it. Our long term goal is to collect statistics on
15082 many files, but as a first step, our immediate goal is to collect
15083 statistics on one file.
15085 The information will be a series of numbers, each number being the
15086 length of a function definition. We can store the numbers in a list.
15088 We know that we will want to incorporate the information regarding one
15089 file with information about many other files; this means that the
15090 function for counting definition lengths within one file need only
15091 return the list of lengths. It need not and should not display any
15094 The word count commands contain one expression to jump point forward
15095 word by word and another expression to count the jumps. The function
15096 to return the lengths of definitions can be designed to work the same
15097 way, with one expression to jump point forward definition by
15098 definition and another expression to construct the lengths' list.
15100 This statement of the problem makes it elementary to write the
15101 function definition. Clearly, we will start the count at the
15102 beginning of the file, so the first command will be @code{(goto-char
15103 (point-min))}. Next, we start the @code{while} loop; and the
15104 true-or-false test of the loop can be a regular expression search for
15105 the next function definition---so long as the search succeeds, point
15106 is moved forward and then the body of the loop is evaluated. The body
15107 needs an expression that constructs the lengths' list. @code{cons},
15108 the list construction command, can be used to create the list. That
15109 is almost all there is to it.
15112 Here is what this fragment of code looks like:
15116 (goto-char (point-min))
15117 (while (re-search-forward "^(defun" nil t)
15119 (cons (count-words-in-defun) lengths-list)))
15123 What we have left out is the mechanism for finding the file that
15124 contains the function definitions.
15126 In previous examples, we either used this, the Info file, or we
15127 switched back and forth to some other buffer, such as the
15128 @file{*scratch*} buffer.
15130 Finding a file is a new process that we have not yet discussed.
15132 @node Find a File, lengths-list-file, Several defuns, Words in a defun
15133 @comment node-name, next, previous, up
15134 @section Find a File
15135 @cindex Find a File
15137 To find a file in Emacs, you use the @kbd{C-x C-f} (@code{find-file})
15138 command. This command is almost, but not quite right for the lengths
15142 Let's look at the source for @code{find-file}:
15146 (defun find-file (filename)
15147 "Edit file FILENAME.
15148 Switch to a buffer visiting file FILENAME,
15149 creating one if none already exists."
15150 (interactive "FFind file: ")
15151 (switch-to-buffer (find-file-noselect filename)))
15156 (The most recent version of the @code{find-file} function definition
15157 permits you to specify optional wildcards to visit multiple files; that
15158 makes the definition more complex and we will not discuss it here,
15159 since it is not relevant. You can see its source using either
15160 @kbd{M-.} (@code{find-tag}) or @kbd{C-h f} (@code{describe-function}).)
15164 (defun find-file (filename &optional wildcards)
15165 "Edit file FILENAME.
15166 Switch to a buffer visiting file FILENAME,
15167 creating one if none already exists.
15168 Interactively, the default if you just type RET is the current directory,
15169 but the visited file name is available through the minibuffer history:
15170 type M-n to pull it into the minibuffer.
15172 Interactively, or if WILDCARDS is non-nil in a call from Lisp,
15173 expand wildcards (if any) and visit multiple files. You can
15174 suppress wildcard expansion by setting `find-file-wildcards' to nil.
15176 To visit a file without any kind of conversion and without
15177 automatically choosing a major mode, use \\[find-file-literally]."
15178 (interactive (find-file-read-args "Find file: " nil))
15179 (let ((value (find-file-noselect filename nil nil wildcards)))
15181 (mapcar 'switch-to-buffer (nreverse value))
15182 (switch-to-buffer value))))
15185 The definition I am showing possesses short but complete documentation
15186 and an interactive specification that prompts you for a file name when
15187 you use the command interactively. The body of the definition
15188 contains two functions, @code{find-file-noselect} and
15189 @code{switch-to-buffer}.
15191 According to its documentation as shown by @kbd{C-h f} (the
15192 @code{describe-function} command), the @code{find-file-noselect}
15193 function reads the named file into a buffer and returns the buffer.
15194 (Its most recent version includes an optional wildcards argument,
15195 too, as well as another to read a file literally and an other you
15196 suppress warning messages. These optional arguments are irrelevant.)
15198 However, the @code{find-file-noselect} function does not select the
15199 buffer in which it puts the file. Emacs does not switch its attention
15200 (or yours if you are using @code{find-file-noselect}) to the selected
15201 buffer. That is what @code{switch-to-buffer} does: it switches the
15202 buffer to which Emacs attention is directed; and it switches the
15203 buffer displayed in the window to the new buffer. We have discussed
15204 buffer switching elsewhere. (@xref{Switching Buffers}.)
15206 In this histogram project, we do not need to display each file on the
15207 screen as the program determines the length of each definition within
15208 it. Instead of employing @code{switch-to-buffer}, we can work with
15209 @code{set-buffer}, which redirects the attention of the computer
15210 program to a different buffer but does not redisplay it on the screen.
15211 So instead of calling on @code{find-file} to do the job, we must write
15212 our own expression.
15214 The task is easy: use @code{find-file-noselect} and @code{set-buffer}.
15216 @node lengths-list-file, Several files, Find a File, Words in a defun
15217 @section @code{lengths-list-file} in Detail
15219 The core of the @code{lengths-list-file} function is a @code{while}
15220 loop containing a function to move point forward `defun by defun' and
15221 a function to count the number of words and symbols in each defun.
15222 This core must be surrounded by functions that do various other tasks,
15223 including finding the file, and ensuring that point starts out at the
15224 beginning of the file. The function definition looks like this:
15225 @findex lengths-list-file
15229 (defun lengths-list-file (filename)
15230 "Return list of definitions' lengths within FILE.
15231 The returned list is a list of numbers.
15232 Each number is the number of words or
15233 symbols in one function definition."
15236 (message "Working on `%s' ... " filename)
15238 (let ((buffer (find-file-noselect filename))
15240 (set-buffer buffer)
15241 (setq buffer-read-only t)
15243 (goto-char (point-min))
15244 (while (re-search-forward "^(defun" nil t)
15246 (cons (count-words-in-defun) lengths-list)))
15247 (kill-buffer buffer)
15253 The function is passed one argument, the name of the file on which it
15254 will work. It has four lines of documentation, but no interactive
15255 specification. Since people worry that a computer is broken if they
15256 don't see anything going on, the first line of the body is a
15259 The next line contains a @code{save-excursion} that returns Emacs'
15260 attention to the current buffer when the function completes. This is
15261 useful in case you embed this function in another function that
15262 presumes point is restored to the original buffer.
15264 In the varlist of the @code{let} expression, Emacs finds the file and
15265 binds the local variable @code{buffer} to the buffer containing the
15266 file. At the same time, Emacs creates @code{lengths-list} as a local
15269 Next, Emacs switches its attention to the buffer.
15271 In the following line, Emacs makes the buffer read-only. Ideally,
15272 this line is not necessary. None of the functions for counting words
15273 and symbols in a function definition should change the buffer.
15274 Besides, the buffer is not going to be saved, even if it were changed.
15275 This line is entirely the consequence of great, perhaps excessive,
15276 caution. The reason for the caution is that this function and those
15277 it calls work on the sources for Emacs and it is inconvenient if they
15278 are inadvertently modified. It goes without saying that I did not
15279 realize a need for this line until an experiment went awry and started
15280 to modify my Emacs source files @dots{}
15282 Next comes a call to widen the buffer if it is narrowed. This
15283 function is usually not needed---Emacs creates a fresh buffer if none
15284 already exists; but if a buffer visiting the file already exists Emacs
15285 returns that one. In this case, the buffer may be narrowed and must
15286 be widened. If we wanted to be fully `user-friendly', we would
15287 arrange to save the restriction and the location of point, but we
15290 The @code{(goto-char (point-min))} expression moves point to the
15291 beginning of the buffer.
15293 Then comes a @code{while} loop in which the `work' of the function is
15294 carried out. In the loop, Emacs determines the length of each
15295 definition and constructs a lengths' list containing the information.
15297 Emacs kills the buffer after working through it. This is to save
15298 space inside of Emacs. My version of GNU Emacs 19 contained over 300
15299 source files of interest; GNU Emacs 22 contains over a thousand source
15300 files. Another function will apply @code{lengths-list-file} to each
15303 Finally, the last expression within the @code{let} expression is the
15304 @code{lengths-list} variable; its value is returned as the value of
15305 the whole function.
15307 You can try this function by installing it in the usual fashion. Then
15308 place your cursor after the following expression and type @kbd{C-x
15309 C-e} (@code{eval-last-sexp}).
15311 @c !!! 22.1.1 lisp sources location here
15314 "/usr/local/share/emacs/22.1.1/lisp/emacs-lisp/debug.el")
15318 (You may need to change the pathname of the file; the one here is for
15319 GNU Emacs version 22.1.1. To change the expression, copy it to
15320 the @file{*scratch*} buffer and edit it.
15324 (Also, to see the full length of the list, rather than a truncated
15325 version, you may have to evaluate the following:
15328 (custom-set-variables '(eval-expression-print-length nil))
15332 (@xref{defcustom, , Specifying Variables using @code{defcustom}}.
15333 Then evaluate the @code{lengths-list-file} expression.)
15336 The lengths' list for @file{debug.el} takes less than a second to
15337 produce and looks like this in GNU Emacs 22:
15340 (83 113 105 144 289 22 30 97 48 89 25 52 52 88 28 29 77 49 43 290 232 587)
15344 (Using my old machine, the version 19 lengths' list for @file{debug.el}
15345 took seven seconds to produce and looked like this:
15348 (75 41 80 62 20 45 44 68 45 12 34 235)
15351 (The newer version of @file{debug.el} contains more defuns than the
15352 earlier one; and my new machine is much faster than the old one.)
15354 Note that the length of the last definition in the file is first in
15357 @node Several files, Several files recursively, lengths-list-file, Words in a defun
15358 @section Count Words in @code{defuns} in Different Files
15360 In the previous section, we created a function that returns a list of
15361 the lengths of each definition in a file. Now, we want to define a
15362 function to return a master list of the lengths of the definitions in
15365 Working on each of a list of files is a repetitious act, so we can use
15366 either a @code{while} loop or recursion.
15369 * lengths-list-many-files::
15373 @node lengths-list-many-files, append, Several files, Several files
15375 @unnumberedsubsec Determine the lengths of @code{defuns}
15378 The design using a @code{while} loop is routine. The argument passed
15379 the function is a list of files. As we saw earlier (@pxref{Loop
15380 Example}), you can write a @code{while} loop so that the body of the
15381 loop is evaluated if such a list contains elements, but to exit the
15382 loop if the list is empty. For this design to work, the body of the
15383 loop must contain an expression that shortens the list each time the
15384 body is evaluated, so that eventually the list is empty. The usual
15385 technique is to set the value of the list to the value of the @sc{cdr}
15386 of the list each time the body is evaluated.
15389 The template looks like this:
15393 (while @var{test-whether-list-is-empty}
15395 @var{set-list-to-cdr-of-list})
15399 Also, we remember that a @code{while} loop returns @code{nil} (the
15400 result of evaluating the true-or-false-test), not the result of any
15401 evaluation within its body. (The evaluations within the body of the
15402 loop are done for their side effects.) However, the expression that
15403 sets the lengths' list is part of the body---and that is the value
15404 that we want returned by the function as a whole. To do this, we
15405 enclose the @code{while} loop within a @code{let} expression, and
15406 arrange that the last element of the @code{let} expression contains
15407 the value of the lengths' list. (@xref{Incrementing Example, , Loop
15408 Example with an Incrementing Counter}.)
15410 @findex lengths-list-many-files
15412 These considerations lead us directly to the function itself:
15416 ;;; @r{Use @code{while} loop.}
15417 (defun lengths-list-many-files (list-of-files)
15418 "Return list of lengths of defuns in LIST-OF-FILES."
15421 (let (lengths-list)
15423 ;;; @r{true-or-false-test}
15424 (while list-of-files
15429 ;;; @r{Generate a lengths' list.}
15431 (expand-file-name (car list-of-files)))))
15435 ;;; @r{Make files' list shorter.}
15436 (setq list-of-files (cdr list-of-files)))
15438 ;;; @r{Return final value of lengths' list.}
15443 @code{expand-file-name} is a built-in function that converts a file
15444 name to the absolute, long, path name form. The function employs the
15445 name of the directory in which the function is called.
15447 @c !!! 22.1.1 lisp sources location here
15449 Thus, if @code{expand-file-name} is called on @code{debug.el} when
15450 Emacs is visiting the
15451 @file{/usr/local/share/emacs/22.1.1/lisp/emacs-lisp/} directory,
15461 @c !!! 22.1.1 lisp sources location here
15463 /usr/local/share/emacs/22.1.1/lisp/emacs-lisp/debug.el
15466 The only other new element of this function definition is the as yet
15467 unstudied function @code{append}, which merits a short section for
15470 @node append, , lengths-list-many-files, Several files
15471 @subsection The @code{append} Function
15474 The @code{append} function attaches one list to another. Thus,
15477 (append '(1 2 3 4) '(5 6 7 8))
15488 This is exactly how we want to attach two lengths' lists produced by
15489 @code{lengths-list-file} to each other. The results contrast with
15493 (cons '(1 2 3 4) '(5 6 7 8))
15498 which constructs a new list in which the first argument to @code{cons}
15499 becomes the first element of the new list:
15502 ((1 2 3 4) 5 6 7 8)
15505 @node Several files recursively, Prepare the data, Several files, Words in a defun
15506 @section Recursively Count Words in Different Files
15508 Besides a @code{while} loop, you can work on each of a list of files
15509 with recursion. A recursive version of @code{lengths-list-many-files}
15510 is short and simple.
15512 The recursive function has the usual parts: the `do-again-test', the
15513 `next-step-expression', and the recursive call. The `do-again-test'
15514 determines whether the function should call itself again, which it
15515 will do if the @code{list-of-files} contains any remaining elements;
15516 the `next-step-expression' resets the @code{list-of-files} to the
15517 @sc{cdr} of itself, so eventually the list will be empty; and the
15518 recursive call calls itself on the shorter list. The complete
15519 function is shorter than this description!
15520 @findex recursive-lengths-list-many-files
15524 (defun recursive-lengths-list-many-files (list-of-files)
15525 "Return list of lengths of each defun in LIST-OF-FILES."
15526 (if list-of-files ; @r{do-again-test}
15529 (expand-file-name (car list-of-files)))
15530 (recursive-lengths-list-many-files
15531 (cdr list-of-files)))))
15536 In a sentence, the function returns the lengths' list for the first of
15537 the @code{list-of-files} appended to the result of calling itself on
15538 the rest of the @code{list-of-files}.
15540 Here is a test of @code{recursive-lengths-list-many-files}, along with
15541 the results of running @code{lengths-list-file} on each of the files
15544 Install @code{recursive-lengths-list-many-files} and
15545 @code{lengths-list-file}, if necessary, and then evaluate the
15546 following expressions. You may need to change the files' pathnames;
15547 those here work when this Info file and the Emacs sources are located
15548 in their customary places. To change the expressions, copy them to
15549 the @file{*scratch*} buffer, edit them, and then evaluate them.
15551 The results are shown after the @samp{@result{}}. (These results are
15552 for files from Emacs version 22.1.1; files from other versions of
15553 Emacs may produce different results.)
15555 @c !!! 22.1.1 lisp sources location here
15558 (cd "/usr/local/share/emacs/22.1.1/")
15560 (lengths-list-file "./lisp/macros.el")
15561 @result{} (283 263 480 90)
15565 (lengths-list-file "./lisp/mail/mailalias.el")
15566 @result{} (38 32 29 95 178 180 321 218 324)
15570 (lengths-list-file "./lisp/makesum.el")
15575 (recursive-lengths-list-many-files
15576 '("./lisp/macros.el"
15577 "./lisp/mail/mailalias.el"
15578 "./lisp/makesum.el"))
15579 @result{} (283 263 480 90 38 32 29 95 178 180 321 218 324 85 181)
15583 The @code{recursive-lengths-list-many-files} function produces the
15586 The next step is to prepare the data in the list for display in a graph.
15588 @node Prepare the data, , Several files recursively, Words in a defun
15589 @section Prepare the Data for Display in a Graph
15591 The @code{recursive-lengths-list-many-files} function returns a list
15592 of numbers. Each number records the length of a function definition.
15593 What we need to do now is transform this data into a list of numbers
15594 suitable for generating a graph. The new list will tell how many
15595 functions definitions contain less than 10 words and
15596 symbols, how many contain between 10 and 19 words and symbols, how
15597 many contain between 20 and 29 words and symbols, and so on.
15599 In brief, we need to go through the lengths' list produced by the
15600 @code{recursive-lengths-list-many-files} function and count the number
15601 of defuns within each range of lengths, and produce a list of those
15605 * Data for Display in Detail::
15608 * Counting function definitions::
15611 @node Data for Display in Detail, Sorting, Prepare the data, Prepare the data
15613 @unnumberedsubsec The Data for Display in Detail
15616 Based on what we have done before, we can readily foresee that it
15617 should not be too hard to write a function that `@sc{cdr}s' down the
15618 lengths' list, looks at each element, determines which length range it
15619 is in, and increments a counter for that range.
15621 However, before beginning to write such a function, we should consider
15622 the advantages of sorting the lengths' list first, so the numbers are
15623 ordered from smallest to largest. First, sorting will make it easier
15624 to count the numbers in each range, since two adjacent numbers will
15625 either be in the same length range or in adjacent ranges. Second, by
15626 inspecting a sorted list, we can discover the highest and lowest
15627 number, and thereby determine the largest and smallest length range
15630 @node Sorting, Files List, Data for Display in Detail, Prepare the data
15631 @subsection Sorting Lists
15634 Emacs contains a function to sort lists, called (as you might guess)
15635 @code{sort}. The @code{sort} function takes two arguments, the list
15636 to be sorted, and a predicate that determines whether the first of
15637 two list elements is ``less'' than the second.
15639 As we saw earlier (@pxref{Wrong Type of Argument, , Using the Wrong
15640 Type Object as an Argument}), a predicate is a function that
15641 determines whether some property is true or false. The @code{sort}
15642 function will reorder a list according to whatever property the
15643 predicate uses; this means that @code{sort} can be used to sort
15644 non-numeric lists by non-numeric criteria---it can, for example,
15645 alphabetize a list.
15648 The @code{<} function is used when sorting a numeric list. For example,
15651 (sort '(4 8 21 17 33 7 21 7) '<)
15659 (4 7 7 8 17 21 21 33)
15663 (Note that in this example, both the arguments are quoted so that the
15664 symbols are not evaluated before being passed to @code{sort} as
15667 Sorting the list returned by the
15668 @code{recursive-lengths-list-many-files} function is straightforward;
15669 it uses the @code{<} function:
15673 In GNU Emacs 22, eval
15675 (cd "/usr/local/share/emacs/22.0.50/")
15677 (recursive-lengths-list-many-files
15678 '("./lisp/macros.el"
15679 "./lisp/mail/mailalias.el"
15680 "./lisp/makesum.el"))
15688 (recursive-lengths-list-many-files
15689 '("./lisp/macros.el"
15690 "./lisp/mailalias.el"
15691 "./lisp/makesum.el"))
15701 (29 32 38 85 90 95 178 180 181 218 263 283 321 324 480)
15705 (Note that in this example, the first argument to @code{sort} is not
15706 quoted, since the expression must be evaluated so as to produce the
15707 list that is passed to @code{sort}.)
15709 @node Files List, Counting function definitions, Sorting, Prepare the data
15710 @subsection Making a List of Files
15712 The @code{recursive-lengths-list-many-files} function requires a list
15713 of files as its argument. For our test examples, we constructed such
15714 a list by hand; but the Emacs Lisp source directory is too large for
15715 us to do for that. Instead, we will write a function to do the job
15716 for us. In this function, we will use both a @code{while} loop and a
15719 @findex directory-files
15720 We did not have to write a function like this for older versions of
15721 GNU Emacs, since they placed all the @samp{.el} files in one
15722 directory. Instead, we were able to use the @code{directory-files}
15723 function, which lists the names of files that match a specified
15724 pattern within a single directory.
15726 However, recent versions of Emacs place Emacs Lisp files in
15727 sub-directories of the top level @file{lisp} directory. This
15728 re-arrangement eases navigation. For example, all the mail related
15729 files are in a @file{lisp} sub-directory called @file{mail}. But at
15730 the same time, this arrangement forces us to create a file listing
15731 function that descends into the sub-directories.
15733 @findex files-in-below-directory
15734 We can create this function, called @code{files-in-below-directory},
15735 using familiar functions such as @code{car}, @code{nthcdr}, and
15736 @code{substring} in conjunction with an existing function called
15737 @code{directory-files-and-attributes}. This latter function not only
15738 lists all the filenames in a directory, including the names
15739 of sub-directories, but also their attributes.
15741 To restate our goal: to create a function that will enable us
15742 to feed filenames to @code{recursive-lengths-list-many-files}
15743 as a list that looks like this (but with more elements):
15747 ("./lisp/macros.el"
15748 "./lisp/mail/rmail.el"
15749 "./lisp/makesum.el")
15753 The @code{directory-files-and-attributes} function returns a list of
15754 lists. Each of the lists within the main list consists of 13
15755 elements. The first element is a string that contains the name of the
15756 file -- which, in GNU/Linux, may be a `directory file', that is to
15757 say, a file with the special attributes of a directory. The second
15758 element of the list is @code{t} for a directory, a string
15759 for symbolic link (the string is the name linked to), or @code{nil}.
15761 For example, the first @samp{.el} file in the @file{lisp/} directory
15762 is @file{abbrev.el}. Its name is
15763 @file{/usr/local/share/emacs/22.1.1/lisp/abbrev.el} and it is not a
15764 directory or a symbolic link.
15767 This is how @code{directory-files-and-attributes} lists that file and
15793 On the other hand, @file{mail/} is a directory within the @file{lisp/}
15794 directory. The beginning of its listing looks like this:
15805 (To learn about the different attributes, look at the documentation of
15806 @code{file-attributes}. Bear in mind that the @code{file-attributes}
15807 function does not list the filename, so its first element is
15808 @code{directory-files-and-attributes}'s second element.)
15810 We will want our new function, @code{files-in-below-directory}, to
15811 list the @samp{.el} files in the directory it is told to check, and in
15812 any directories below that directory.
15814 This gives us a hint on how to construct
15815 @code{files-in-below-directory}: within a directory, the function
15816 should add @samp{.el} filenames to a list; and if, within a directory,
15817 the function comes upon a sub-directory, it should go into that
15818 sub-directory and repeat its actions.
15820 However, we should note that every directory contains a name that
15821 refers to itself, called @file{.}, (``dot'') and a name that refers to
15822 its parent directory, called @file{..} (``double dot''). (In
15823 @file{/}, the root directory, @file{..} refers to itself, since
15824 @file{/} has no parent.) Clearly, we do not want our
15825 @code{files-in-below-directory} function to enter those directories,
15826 since they always lead us, directly or indirectly, to the current
15829 Consequently, our @code{files-in-below-directory} function must do
15834 Check to see whether it is looking at a filename that ends in
15835 @samp{.el}; and if so, add its name to a list.
15838 Check to see whether it is looking at a filename that is the name of a
15839 directory; and if so,
15843 Check to see whether it is looking at @file{.} or @file{..}; and if
15847 Or else, go into that directory and repeat the process.
15851 Let's write a function definition to do these tasks. We will use a
15852 @code{while} loop to move from one filename to another within a
15853 directory, checking what needs to be done; and we will use a recursive
15854 call to repeat the actions on each sub-directory. The recursive
15855 pattern is `accumulate'
15856 (@pxref{Accumulate, , Recursive Pattern: @emph{accumulate}}),
15857 using @code{append} as the combiner.
15860 (directory-files "/usr/local/src/emacs/lisp/" t "\\.el$")
15861 (shell-command "find /usr/local/src/emacs/lisp/ -name '*.el'")
15863 (directory-files "/usr/local/share/emacs/22.1.1/lisp/" t "\\.el$")
15864 (shell-command "find /usr/local/share/emacs/22.1.1/lisp/ -name '*.el'")
15867 @c /usr/local/share/emacs/22.1.1/lisp/
15870 Here is the function:
15874 (defun files-in-below-directory (directory)
15875 "List the .el files in DIRECTORY and in its sub-directories."
15876 ;; Although the function will be used non-interactively,
15877 ;; it will be easier to test if we make it interactive.
15878 ;; The directory will have a name such as
15879 ;; "/usr/local/share/emacs/22.1.1/lisp/"
15880 (interactive "DDirectory name: ")
15883 (let (el-files-list
15884 (current-directory-list
15885 (directory-files-and-attributes directory t)))
15886 ;; while we are in the current directory
15887 (while current-directory-list
15891 ;; check to see whether filename ends in `.el'
15892 ;; and if so, append its name to a list.
15893 ((equal ".el" (substring (car (car current-directory-list)) -3))
15894 (setq el-files-list
15895 (cons (car (car current-directory-list)) el-files-list)))
15898 ;; check whether filename is that of a directory
15899 ((eq t (car (cdr (car current-directory-list))))
15900 ;; decide whether to skip or recurse
15903 (substring (car (car current-directory-list)) -1))
15904 ;; then do nothing since filename is that of
15905 ;; current directory or parent, "." or ".."
15909 ;; else descend into the directory and repeat the process
15910 (setq el-files-list
15912 (files-in-below-directory
15913 (car (car current-directory-list)))
15915 ;; move to the next filename in the list; this also
15916 ;; shortens the list so the while loop eventually comes to an end
15917 (setq current-directory-list (cdr current-directory-list)))
15918 ;; return the filenames
15923 @c (files-in-below-directory "/usr/local/src/emacs/lisp/")
15924 @c (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/")
15926 The @code{files-in-below-directory} @code{directory-files} function
15927 takes one argument, the name of a directory.
15930 Thus, on my system,
15932 @c (length (files-in-below-directory "/usr/local/src/emacs/lisp/"))
15934 @c !!! 22.1.1 lisp sources location here
15938 (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/"))
15943 tells me that in and below my Lisp sources directory are 1031
15946 @code{files-in-below-directory} returns a list in reverse alphabetical
15947 order. An expression to sort the list in alphabetical order looks
15953 (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/")
15960 "Test how long it takes to find lengths of all sorted elisp defuns."
15961 (insert "\n" (current-time-string) "\n")
15964 (recursive-lengths-list-many-files
15965 (files-in-below-directory "/usr/local/src/emacs/lisp/"))
15967 (insert (format "%s" (current-time-string))))
15970 @node Counting function definitions, , Files List, Prepare the data
15971 @subsection Counting function definitions
15973 Our immediate goal is to generate a list that tells us how many
15974 function definitions contain fewer than 10 words and symbols, how many
15975 contain between 10 and 19 words and symbols, how many contain between
15976 20 and 29 words and symbols, and so on.
15978 With a sorted list of numbers, this is easy: count how many elements
15979 of the list are smaller than 10, then, after moving past the numbers
15980 just counted, count how many are smaller than 20, then, after moving
15981 past the numbers just counted, count how many are smaller than 30, and
15982 so on. Each of the numbers, 10, 20, 30, 40, and the like, is one
15983 larger than the top of that range. We can call the list of such
15984 numbers the @code{top-of-ranges} list.
15987 If we wished, we could generate this list automatically, but it is
15988 simpler to write a list manually. Here it is:
15989 @vindex top-of-ranges
15993 (defvar top-of-ranges
15996 110 120 130 140 150
15997 160 170 180 190 200
15998 210 220 230 240 250
15999 260 270 280 290 300)
16000 "List specifying ranges for `defuns-per-range'.")
16004 To change the ranges, we edit this list.
16006 Next, we need to write the function that creates the list of the
16007 number of definitions within each range. Clearly, this function must
16008 take the @code{sorted-lengths} and the @code{top-of-ranges} lists
16011 The @code{defuns-per-range} function must do two things again and
16012 again: it must count the number of definitions within a range
16013 specified by the current top-of-range value; and it must shift to the
16014 next higher value in the @code{top-of-ranges} list after counting the
16015 number of definitions in the current range. Since each of these
16016 actions is repetitive, we can use @code{while} loops for the job.
16017 One loop counts the number of definitions in the range defined by the
16018 current top-of-range value, and the other loop selects each of the
16019 top-of-range values in turn.
16021 Several entries of the @code{sorted-lengths} list are counted for each
16022 range; this means that the loop for the @code{sorted-lengths} list
16023 will be inside the loop for the @code{top-of-ranges} list, like a
16024 small gear inside a big gear.
16026 The inner loop counts the number of definitions within the range. It
16027 is a simple counting loop of the type we have seen before.
16028 (@xref{Incrementing Loop, , A loop with an incrementing counter}.)
16029 The true-or-false test of the loop tests whether the value from the
16030 @code{sorted-lengths} list is smaller than the current value of the
16031 top of the range. If it is, the function increments the counter and
16032 tests the next value from the @code{sorted-lengths} list.
16035 The inner loop looks like this:
16039 (while @var{length-element-smaller-than-top-of-range}
16040 (setq number-within-range (1+ number-within-range))
16041 (setq sorted-lengths (cdr sorted-lengths)))
16045 The outer loop must start with the lowest value of the
16046 @code{top-of-ranges} list, and then be set to each of the succeeding
16047 higher values in turn. This can be done with a loop like this:
16051 (while top-of-ranges
16052 @var{body-of-loop}@dots{}
16053 (setq top-of-ranges (cdr top-of-ranges)))
16058 Put together, the two loops look like this:
16062 (while top-of-ranges
16064 ;; @r{Count the number of elements within the current range.}
16065 (while @var{length-element-smaller-than-top-of-range}
16066 (setq number-within-range (1+ number-within-range))
16067 (setq sorted-lengths (cdr sorted-lengths)))
16069 ;; @r{Move to next range.}
16070 (setq top-of-ranges (cdr top-of-ranges)))
16074 In addition, in each circuit of the outer loop, Emacs should record
16075 the number of definitions within that range (the value of
16076 @code{number-within-range}) in a list. We can use @code{cons} for
16077 this purpose. (@xref{cons, , @code{cons}}.)
16079 The @code{cons} function works fine, except that the list it
16080 constructs will contain the number of definitions for the highest
16081 range at its beginning and the number of definitions for the lowest
16082 range at its end. This is because @code{cons} attaches new elements
16083 of the list to the beginning of the list, and since the two loops are
16084 working their way through the lengths' list from the lower end first,
16085 the @code{defuns-per-range-list} will end up largest number first.
16086 But we will want to print our graph with smallest values first and the
16087 larger later. The solution is to reverse the order of the
16088 @code{defuns-per-range-list}. We can do this using the
16089 @code{nreverse} function, which reverses the order of a list.
16096 (nreverse '(1 2 3 4))
16107 Note that the @code{nreverse} function is ``destructive''---that is,
16108 it changes the list to which it is applied; this contrasts with the
16109 @code{car} and @code{cdr} functions, which are non-destructive. In
16110 this case, we do not want the original @code{defuns-per-range-list},
16111 so it does not matter that it is destroyed. (The @code{reverse}
16112 function provides a reversed copy of a list, leaving the original list
16117 Put all together, the @code{defuns-per-range} looks like this:
16121 (defun defuns-per-range (sorted-lengths top-of-ranges)
16122 "SORTED-LENGTHS defuns in each TOP-OF-RANGES range."
16123 (let ((top-of-range (car top-of-ranges))
16124 (number-within-range 0)
16125 defuns-per-range-list)
16130 (while top-of-ranges
16136 ;; @r{Need number for numeric test.}
16137 (car sorted-lengths)
16138 (< (car sorted-lengths) top-of-range))
16142 ;; @r{Count number of definitions within current range.}
16143 (setq number-within-range (1+ number-within-range))
16144 (setq sorted-lengths (cdr sorted-lengths)))
16146 ;; @r{Exit inner loop but remain within outer loop.}
16150 (setq defuns-per-range-list
16151 (cons number-within-range defuns-per-range-list))
16152 (setq number-within-range 0) ; @r{Reset count to zero.}
16156 ;; @r{Move to next range.}
16157 (setq top-of-ranges (cdr top-of-ranges))
16158 ;; @r{Specify next top of range value.}
16159 (setq top-of-range (car top-of-ranges)))
16163 ;; @r{Exit outer loop and count the number of defuns larger than}
16164 ;; @r{ the largest top-of-range value.}
16165 (setq defuns-per-range-list
16167 (length sorted-lengths)
16168 defuns-per-range-list))
16172 ;; @r{Return a list of the number of definitions within each range,}
16173 ;; @r{ smallest to largest.}
16174 (nreverse defuns-per-range-list)))
16180 The function is straightforward except for one subtle feature. The
16181 true-or-false test of the inner loop looks like this:
16185 (and (car sorted-lengths)
16186 (< (car sorted-lengths) top-of-range))
16192 instead of like this:
16195 (< (car sorted-lengths) top-of-range)
16198 The purpose of the test is to determine whether the first item in the
16199 @code{sorted-lengths} list is less than the value of the top of the
16202 The simple version of the test works fine unless the
16203 @code{sorted-lengths} list has a @code{nil} value. In that case, the
16204 @code{(car sorted-lengths)} expression function returns
16205 @code{nil}. The @code{<} function cannot compare a number to
16206 @code{nil}, which is an empty list, so Emacs signals an error and
16207 stops the function from attempting to continue to execute.
16209 The @code{sorted-lengths} list always becomes @code{nil} when the
16210 counter reaches the end of the list. This means that any attempt to
16211 use the @code{defuns-per-range} function with the simple version of
16212 the test will fail.
16214 We solve the problem by using the @code{(car sorted-lengths)}
16215 expression in conjunction with the @code{and} expression. The
16216 @code{(car sorted-lengths)} expression returns a non-@code{nil}
16217 value so long as the list has at least one number within it, but
16218 returns @code{nil} if the list is empty. The @code{and} expression
16219 first evaluates the @code{(car sorted-lengths)} expression, and
16220 if it is @code{nil}, returns false @emph{without} evaluating the
16221 @code{<} expression. But if the @code{(car sorted-lengths)}
16222 expression returns a non-@code{nil} value, the @code{and} expression
16223 evaluates the @code{<} expression, and returns that value as the value
16224 of the @code{and} expression.
16226 @c colon in printed section title causes problem in Info cross reference
16227 This way, we avoid an error.
16230 (For information about @code{and}, see
16231 @ref{kill-new function, , The @code{kill-new} function}.)
16235 (@xref{kill-new function, , The @code{kill-new} function}, for
16236 information about @code{and}.)
16239 Here is a short test of the @code{defuns-per-range} function. First,
16240 evaluate the expression that binds (a shortened)
16241 @code{top-of-ranges} list to the list of values, then evaluate the
16242 expression for binding the @code{sorted-lengths} list, and then
16243 evaluate the @code{defuns-per-range} function.
16247 ;; @r{(Shorter list than we will use later.)}
16248 (setq top-of-ranges
16249 '(110 120 130 140 150
16250 160 170 180 190 200))
16252 (setq sorted-lengths
16253 '(85 86 110 116 122 129 154 176 179 200 265 300 300))
16255 (defuns-per-range sorted-lengths top-of-ranges)
16261 The list returned looks like this:
16264 (2 2 2 0 0 1 0 2 0 0 4)
16268 Indeed, there are two elements of the @code{sorted-lengths} list
16269 smaller than 110, two elements between 110 and 119, two elements
16270 between 120 and 129, and so on. There are four elements with a value
16273 @c The next step is to turn this numbers' list into a graph.
16274 @node Readying a Graph, Emacs Initialization, Words in a defun, Top
16275 @chapter Readying a Graph
16276 @cindex Readying a graph
16277 @cindex Graph prototype
16278 @cindex Prototype graph
16279 @cindex Body of graph
16281 Our goal is to construct a graph showing the numbers of function
16282 definitions of various lengths in the Emacs lisp sources.
16284 As a practical matter, if you were creating a graph, you would
16285 probably use a program such as @code{gnuplot} to do the job.
16286 (@code{gnuplot} is nicely integrated into GNU Emacs.) In this case,
16287 however, we create one from scratch, and in the process we will
16288 re-acquaint ourselves with some of what we learned before and learn
16291 In this chapter, we will first write a simple graph printing function.
16292 This first definition will be a @dfn{prototype}, a rapidly written
16293 function that enables us to reconnoiter this unknown graph-making
16294 territory. We will discover dragons, or find that they are myth.
16295 After scouting the terrain, we will feel more confident and enhance
16296 the function to label the axes automatically.
16299 * Columns of a graph::
16300 * graph-body-print::
16301 * recursive-graph-body-print::
16303 * Line Graph Exercise::
16306 @node Columns of a graph, graph-body-print, Readying a Graph, Readying a Graph
16308 @unnumberedsec Printing the Columns of a Graph
16311 Since Emacs is designed to be flexible and work with all kinds of
16312 terminals, including character-only terminals, the graph will need to
16313 be made from one of the `typewriter' symbols. An asterisk will do; as
16314 we enhance the graph-printing function, we can make the choice of
16315 symbol a user option.
16317 We can call this function @code{graph-body-print}; it will take a
16318 @code{numbers-list} as its only argument. At this stage, we will not
16319 label the graph, but only print its body.
16321 The @code{graph-body-print} function inserts a vertical column of
16322 asterisks for each element in the @code{numbers-list}. The height of
16323 each line is determined by the value of that element of the
16324 @code{numbers-list}.
16326 Inserting columns is a repetitive act; that means that this function can
16327 be written either with a @code{while} loop or recursively.
16329 Our first challenge is to discover how to print a column of asterisks.
16330 Usually, in Emacs, we print characters onto a screen horizontally,
16331 line by line, by typing. We have two routes we can follow: write our
16332 own column-insertion function or discover whether one exists in Emacs.
16334 To see whether there is one in Emacs, we can use the @kbd{M-x apropos}
16335 command. This command is like the @kbd{C-h a} (@code{command-apropos})
16336 command, except that the latter finds only those functions that are
16337 commands. The @kbd{M-x apropos} command lists all symbols that match
16338 a regular expression, including functions that are not interactive.
16341 What we want to look for is some command that prints or inserts
16342 columns. Very likely, the name of the function will contain either
16343 the word `print' or the word `insert' or the word `column'.
16344 Therefore, we can simply type @kbd{M-x apropos RET
16345 print\|insert\|column RET} and look at the result. On my system, this
16346 command once too takes quite some time, and then produced a list of 79
16347 functions and variables. Now it does not take much time at all and
16348 produces a list of 211 functions and variables. Scanning down the
16349 list, the only function that looks as if it might do the job is
16350 @code{insert-rectangle}.
16353 Indeed, this is the function we want; its documentation says:
16358 Insert text of RECTANGLE with upper left corner at point.
16359 RECTANGLE's first line is inserted at point,
16360 its second line is inserted at a point vertically under point, etc.
16361 RECTANGLE should be a list of strings.
16362 After this command, the mark is at the upper left corner
16363 and point is at the lower right corner.
16367 We can run a quick test, to make sure it does what we expect of it.
16369 Here is the result of placing the cursor after the
16370 @code{insert-rectangle} expression and typing @kbd{C-u C-x C-e}
16371 (@code{eval-last-sexp}). The function inserts the strings
16372 @samp{"first"}, @samp{"second"}, and @samp{"third"} at and below
16373 point. Also the function returns @code{nil}.
16377 (insert-rectangle '("first" "second" "third"))first
16384 Of course, we won't be inserting the text of the
16385 @code{insert-rectangle} expression itself into the buffer in which we
16386 are making the graph, but will call the function from our program. We
16387 shall, however, have to make sure that point is in the buffer at the
16388 place where the @code{insert-rectangle} function will insert its
16391 If you are reading this in Info, you can see how this works by
16392 switching to another buffer, such as the @file{*scratch*} buffer,
16393 placing point somewhere in the buffer, typing @kbd{M-:}, typing the
16394 @code{insert-rectangle} expression into the minibuffer at the prompt,
16395 and then typing @key{RET}. This causes Emacs to evaluate the
16396 expression in the minibuffer, but to use as the value of point the
16397 position of point in the @file{*scratch*} buffer. (@kbd{M-:} is the
16398 keybinding for @code{eval-expression}. Also, @code{nil} does not
16399 appear in the @file{*scratch*} buffer since the expression is
16400 evaluated in the minibuffer.)
16402 We find when we do this that point ends up at the end of the last
16403 inserted line---that is to say, this function moves point as a
16404 side-effect. If we were to repeat the command, with point at this
16405 position, the next insertion would be below and to the right of the
16406 previous insertion. We don't want this! If we are going to make a
16407 bar graph, the columns need to be beside each other.
16409 So we discover that each cycle of the column-inserting @code{while}
16410 loop must reposition point to the place we want it, and that place
16411 will be at the top, not the bottom, of the column. Moreover, we
16412 remember that when we print a graph, we do not expect all the columns
16413 to be the same height. This means that the top of each column may be
16414 at a different height from the previous one. We cannot simply
16415 reposition point to the same line each time, but moved over to the
16416 right---or perhaps we can@dots{}
16418 We are planning to make the columns of the bar graph out of asterisks.
16419 The number of asterisks in the column is the number specified by the
16420 current element of the @code{numbers-list}. We need to construct a
16421 list of asterisks of the right length for each call to
16422 @code{insert-rectangle}. If this list consists solely of the requisite
16423 number of asterisks, then we will have position point the right number
16424 of lines above the base for the graph to print correctly. This could
16427 Alternatively, if we can figure out some way to pass
16428 @code{insert-rectangle} a list of the same length each time, then we
16429 can place point on the same line each time, but move it over one
16430 column to the right for each new column. If we do this, however, some
16431 of the entries in the list passed to @code{insert-rectangle} must be
16432 blanks rather than asterisks. For example, if the maximum height of
16433 the graph is 5, but the height of the column is 3, then
16434 @code{insert-rectangle} requires an argument that looks like this:
16437 (" " " " "*" "*" "*")
16440 This last proposal is not so difficult, so long as we can determine
16441 the column height. There are two ways for us to specify the column
16442 height: we can arbitrarily state what it will be, which would work
16443 fine for graphs of that height; or we can search through the list of
16444 numbers and use the maximum height of the list as the maximum height
16445 of the graph. If the latter operation were difficult, then the former
16446 procedure would be easiest, but there is a function built into Emacs
16447 that determines the maximum of its arguments. We can use that
16448 function. The function is called @code{max} and it returns the
16449 largest of all its arguments, which must be numbers. Thus, for
16457 returns 7. (A corresponding function called @code{min} returns the
16458 smallest of all its arguments.)
16462 However, we cannot simply call @code{max} on the @code{numbers-list};
16463 the @code{max} function expects numbers as its argument, not a list of
16464 numbers. Thus, the following expression,
16467 (max '(3 4 6 5 7 3))
16472 produces the following error message;
16475 Wrong type of argument: number-or-marker-p, (3 4 6 5 7 3)
16479 We need a function that passes a list of arguments to a function.
16480 This function is @code{apply}. This function `applies' its first
16481 argument (a function) to its remaining arguments, the last of which
16488 (apply 'max 3 4 7 3 '(4 8 5))
16494 (Incidentally, I don't know how you would learn of this function
16495 without a book such as this. It is possible to discover other
16496 functions, like @code{search-forward} or @code{insert-rectangle}, by
16497 guessing at a part of their names and then using @code{apropos}. Even
16498 though its base in metaphor is clear---`apply' its first argument to
16499 the rest---I doubt a novice would come up with that particular word
16500 when using @code{apropos} or other aid. Of course, I could be wrong;
16501 after all, the function was first named by someone who had to invent
16504 The second and subsequent arguments to @code{apply} are optional, so
16505 we can use @code{apply} to call a function and pass the elements of a
16506 list to it, like this, which also returns 8:
16509 (apply 'max '(4 8 5))
16512 This latter way is how we will use @code{apply}. The
16513 @code{recursive-lengths-list-many-files} function returns a numbers'
16514 list to which we can apply @code{max} (we could also apply @code{max} to
16515 the sorted numbers' list; it does not matter whether the list is
16519 Hence, the operation for finding the maximum height of the graph is this:
16522 (setq max-graph-height (apply 'max numbers-list))
16525 Now we can return to the question of how to create a list of strings
16526 for a column of the graph. Told the maximum height of the graph
16527 and the number of asterisks that should appear in the column, the
16528 function should return a list of strings for the
16529 @code{insert-rectangle} command to insert.
16531 Each column is made up of asterisks or blanks. Since the function is
16532 passed the value of the height of the column and the number of
16533 asterisks in the column, the number of blanks can be found by
16534 subtracting the number of asterisks from the height of the column.
16535 Given the number of blanks and the number of asterisks, two
16536 @code{while} loops can be used to construct the list:
16540 ;;; @r{First version.}
16541 (defun column-of-graph (max-graph-height actual-height)
16542 "Return list of strings that is one column of a graph."
16543 (let ((insert-list nil)
16544 (number-of-top-blanks
16545 (- max-graph-height actual-height)))
16549 ;; @r{Fill in asterisks.}
16550 (while (> actual-height 0)
16551 (setq insert-list (cons "*" insert-list))
16552 (setq actual-height (1- actual-height)))
16556 ;; @r{Fill in blanks.}
16557 (while (> number-of-top-blanks 0)
16558 (setq insert-list (cons " " insert-list))
16559 (setq number-of-top-blanks
16560 (1- number-of-top-blanks)))
16564 ;; @r{Return whole list.}
16569 If you install this function and then evaluate the following
16570 expression you will see that it returns the list as desired:
16573 (column-of-graph 5 3)
16581 (" " " " "*" "*" "*")
16584 As written, @code{column-of-graph} contains a major flaw: the symbols
16585 used for the blank and for the marked entries in the column are
16586 `hard-coded' as a space and asterisk. This is fine for a prototype,
16587 but you, or another user, may wish to use other symbols. For example,
16588 in testing the graph function, you many want to use a period in place
16589 of the space, to make sure the point is being repositioned properly
16590 each time the @code{insert-rectangle} function is called; or you might
16591 want to substitute a @samp{+} sign or other symbol for the asterisk.
16592 You might even want to make a graph-column that is more than one
16593 display column wide. The program should be more flexible. The way to
16594 do that is to replace the blank and the asterisk with two variables
16595 that we can call @code{graph-blank} and @code{graph-symbol} and define
16596 those variables separately.
16598 Also, the documentation is not well written. These considerations
16599 lead us to the second version of the function:
16603 (defvar graph-symbol "*"
16604 "String used as symbol in graph, usually an asterisk.")
16608 (defvar graph-blank " "
16609 "String used as blank in graph, usually a blank space.
16610 graph-blank must be the same number of columns wide
16616 (For an explanation of @code{defvar}, see
16617 @ref{defvar, , Initializing a Variable with @code{defvar}}.)
16621 ;;; @r{Second version.}
16622 (defun column-of-graph (max-graph-height actual-height)
16623 "Return MAX-GRAPH-HEIGHT strings; ACTUAL-HEIGHT are graph-symbols.
16627 The graph-symbols are contiguous entries at the end
16629 The list will be inserted as one column of a graph.
16630 The strings are either graph-blank or graph-symbol."
16634 (let ((insert-list nil)
16635 (number-of-top-blanks
16636 (- max-graph-height actual-height)))
16640 ;; @r{Fill in @code{graph-symbols}.}
16641 (while (> actual-height 0)
16642 (setq insert-list (cons graph-symbol insert-list))
16643 (setq actual-height (1- actual-height)))
16647 ;; @r{Fill in @code{graph-blanks}.}
16648 (while (> number-of-top-blanks 0)
16649 (setq insert-list (cons graph-blank insert-list))
16650 (setq number-of-top-blanks
16651 (1- number-of-top-blanks)))
16653 ;; @r{Return whole list.}
16658 If we wished, we could rewrite @code{column-of-graph} a third time to
16659 provide optionally for a line graph as well as for a bar graph. This
16660 would not be hard to do. One way to think of a line graph is that it
16661 is no more than a bar graph in which the part of each bar that is
16662 below the top is blank. To construct a column for a line graph, the
16663 function first constructs a list of blanks that is one shorter than
16664 the value, then it uses @code{cons} to attach a graph symbol to the
16665 list; then it uses @code{cons} again to attach the `top blanks' to
16668 It is easy to see how to write such a function, but since we don't
16669 need it, we will not do it. But the job could be done, and if it were
16670 done, it would be done with @code{column-of-graph}. Even more
16671 important, it is worth noting that few changes would have to be made
16672 anywhere else. The enhancement, if we ever wish to make it, is
16675 Now, finally, we come to our first actual graph printing function.
16676 This prints the body of a graph, not the labels for the vertical and
16677 horizontal axes, so we can call this @code{graph-body-print}.
16679 @node graph-body-print, recursive-graph-body-print, Columns of a graph, Readying a Graph
16680 @section The @code{graph-body-print} Function
16681 @findex graph-body-print
16683 After our preparation in the preceding section, the
16684 @code{graph-body-print} function is straightforward. The function
16685 will print column after column of asterisks and blanks, using the
16686 elements of a numbers' list to specify the number of asterisks in each
16687 column. This is a repetitive act, which means we can use a
16688 decrementing @code{while} loop or recursive function for the job. In
16689 this section, we will write the definition using a @code{while} loop.
16691 The @code{column-of-graph} function requires the height of the graph
16692 as an argument, so we should determine and record that as a local variable.
16694 This leads us to the following template for the @code{while} loop
16695 version of this function:
16699 (defun graph-body-print (numbers-list)
16700 "@var{documentation}@dots{}"
16701 (let ((height @dots{}
16706 (while numbers-list
16707 @var{insert-columns-and-reposition-point}
16708 (setq numbers-list (cdr numbers-list)))))
16713 We need to fill in the slots of the template.
16715 Clearly, we can use the @code{(apply 'max numbers-list)} expression to
16716 determine the height of the graph.
16718 The @code{while} loop will cycle through the @code{numbers-list} one
16719 element at a time. As it is shortened by the @code{(setq numbers-list
16720 (cdr numbers-list))} expression, the @sc{car} of each instance of the
16721 list is the value of the argument for @code{column-of-graph}.
16723 At each cycle of the @code{while} loop, the @code{insert-rectangle}
16724 function inserts the list returned by @code{column-of-graph}. Since
16725 the @code{insert-rectangle} function moves point to the lower right of
16726 the inserted rectangle, we need to save the location of point at the
16727 time the rectangle is inserted, move back to that position after the
16728 rectangle is inserted, and then move horizontally to the next place
16729 from which @code{insert-rectangle} is called.
16731 If the inserted columns are one character wide, as they will be if
16732 single blanks and asterisks are used, the repositioning command is
16733 simply @code{(forward-char 1)}; however, the width of a column may be
16734 greater than one. This means that the repositioning command should be
16735 written @code{(forward-char symbol-width)}. The @code{symbol-width}
16736 itself is the length of a @code{graph-blank} and can be found using
16737 the expression @code{(length graph-blank)}. The best place to bind
16738 the @code{symbol-width} variable to the value of the width of graph
16739 column is in the varlist of the @code{let} expression.
16742 These considerations lead to the following function definition:
16746 (defun graph-body-print (numbers-list)
16747 "Print a bar graph of the NUMBERS-LIST.
16748 The numbers-list consists of the Y-axis values."
16750 (let ((height (apply 'max numbers-list))
16751 (symbol-width (length graph-blank))
16756 (while numbers-list
16757 (setq from-position (point))
16759 (column-of-graph height (car numbers-list)))
16760 (goto-char from-position)
16761 (forward-char symbol-width)
16764 ;; @r{Draw graph column by column.}
16766 (setq numbers-list (cdr numbers-list)))
16769 ;; @r{Place point for X axis labels.}
16770 (forward-line height)
16777 The one unexpected expression in this function is the
16778 @w{@code{(sit-for 0)}} expression in the @code{while} loop. This
16779 expression makes the graph printing operation more interesting to
16780 watch than it would be otherwise. The expression causes Emacs to
16781 `sit' or do nothing for a zero length of time and then redraw the
16782 screen. Placed here, it causes Emacs to redraw the screen column by
16783 column. Without it, Emacs would not redraw the screen until the
16786 We can test @code{graph-body-print} with a short list of numbers.
16790 Install @code{graph-symbol}, @code{graph-blank},
16791 @code{column-of-graph}, which are in
16793 @ref{Readying a Graph, , Readying a Graph},
16796 @ref{Columns of a graph},
16798 and @code{graph-body-print}.
16802 Copy the following expression:
16805 (graph-body-print '(1 2 3 4 6 4 3 5 7 6 5 2 3))
16809 Switch to the @file{*scratch*} buffer and place the cursor where you
16810 want the graph to start.
16813 Type @kbd{M-:} (@code{eval-expression}).
16816 Yank the @code{graph-body-print} expression into the minibuffer
16817 with @kbd{C-y} (@code{yank)}.
16820 Press @key{RET} to evaluate the @code{graph-body-print} expression.
16824 Emacs will print a graph like this:
16838 @node recursive-graph-body-print, Printed Axes, graph-body-print, Readying a Graph
16839 @section The @code{recursive-graph-body-print} Function
16840 @findex recursive-graph-body-print
16842 The @code{graph-body-print} function may also be written recursively.
16843 The recursive solution is divided into two parts: an outside `wrapper'
16844 that uses a @code{let} expression to determine the values of several
16845 variables that need only be found once, such as the maximum height of
16846 the graph, and an inside function that is called recursively to print
16850 The `wrapper' is uncomplicated:
16854 (defun recursive-graph-body-print (numbers-list)
16855 "Print a bar graph of the NUMBERS-LIST.
16856 The numbers-list consists of the Y-axis values."
16857 (let ((height (apply 'max numbers-list))
16858 (symbol-width (length graph-blank))
16860 (recursive-graph-body-print-internal
16867 The recursive function is a little more difficult. It has four parts:
16868 the `do-again-test', the printing code, the recursive call, and the
16869 `next-step-expression'. The `do-again-test' is a @code{when}
16870 expression that determines whether the @code{numbers-list} contains
16871 any remaining elements; if it does, the function prints one column of
16872 the graph using the printing code and calls itself again. The
16873 function calls itself again according to the value produced by the
16874 `next-step-expression' which causes the call to act on a shorter
16875 version of the @code{numbers-list}.
16879 (defun recursive-graph-body-print-internal
16880 (numbers-list height symbol-width)
16881 "Print a bar graph.
16882 Used within recursive-graph-body-print function."
16887 (setq from-position (point))
16889 (column-of-graph height (car numbers-list)))
16892 (goto-char from-position)
16893 (forward-char symbol-width)
16894 (sit-for 0) ; @r{Draw graph column by column.}
16895 (recursive-graph-body-print-internal
16896 (cdr numbers-list) height symbol-width)))
16901 After installation, this expression can be tested; here is a sample:
16904 (recursive-graph-body-print '(3 2 5 6 7 5 3 4 6 4 3 2 1))
16908 Here is what @code{recursive-graph-body-print} produces:
16922 Either of these two functions, @code{graph-body-print} or
16923 @code{recursive-graph-body-print}, create the body of a graph.
16925 @node Printed Axes, Line Graph Exercise, recursive-graph-body-print, Readying a Graph
16926 @section Need for Printed Axes
16928 A graph needs printed axes, so you can orient yourself. For a do-once
16929 project, it may be reasonable to draw the axes by hand using Emacs'
16930 Picture mode; but a graph drawing function may be used more than once.
16932 For this reason, I have written enhancements to the basic
16933 @code{print-graph-body} function that automatically print labels for
16934 the horizontal and vertical axes. Since the label printing functions
16935 do not contain much new material, I have placed their description in
16936 an appendix. @xref{Full Graph, , A Graph with Labelled Axes}.
16938 @node Line Graph Exercise, , Printed Axes, Readying a Graph
16941 Write a line graph version of the graph printing functions.
16943 @node Emacs Initialization, Debugging, Readying a Graph, Top
16944 @chapter Your @file{.emacs} File
16945 @cindex @file{.emacs} file
16946 @cindex Customizing your @file{.emacs} file
16947 @cindex Initialization file
16949 ``You don't have to like Emacs to like it'' -- this seemingly
16950 paradoxical statement is the secret of GNU Emacs. The plain, `out of
16951 the box' Emacs is a generic tool. Most people who use it, customize
16952 it to suit themselves.
16954 GNU Emacs is mostly written in Emacs Lisp; this means that by writing
16955 expressions in Emacs Lisp you can change or extend Emacs.
16958 * Default Configuration::
16961 * Beginning a .emacs File::
16962 * Text and Auto-fill::
16964 * Indent Tabs Mode::
16969 * Simple Extension::
16975 @node Default Configuration, Site-wide Init, Emacs Initialization, Emacs Initialization
16977 @unnumberedsec Emacs' Default Configuration
16980 There are those who appreciate Emacs' default configuration. After
16981 all, Emacs starts you in C mode when you edit a C file, starts you in
16982 Fortran mode when you edit a Fortran file, and starts you in
16983 Fundamental mode when you edit an unadorned file. This all makes
16984 sense, if you do not know who is going to use Emacs. Who knows what a
16985 person hopes to do with an unadorned file? Fundamental mode is the
16986 right default for such a file, just as C mode is the right default for
16987 editing C code. (Enough programming languages have syntaxes
16988 that enable them to share or nearly share features, so C mode is
16989 now provided by by CC mode, the `C Collection'.)
16991 But when you do know who is going to use Emacs---you,
16992 yourself---then it makes sense to customize Emacs.
16994 For example, I seldom want Fundamental mode when I edit an
16995 otherwise undistinguished file; I want Text mode. This is why I
16996 customize Emacs: so it suits me.
16998 You can customize and extend Emacs by writing or adapting a
16999 @file{~/.emacs} file. This is your personal initialization file; its
17000 contents, written in Emacs Lisp, tell Emacs what to do.@footnote{You
17001 may also add @file{.el} to @file{~/.emacs} and call it a
17002 @file{~/.emacs.el} file. In the past, you were forbidden to type the
17003 extra keystrokes that the name @file{~/.emacs.el} requires, but now
17004 you may. The new format is consistent with the Emacs Lisp file
17005 naming conventions; the old format saves typing.}
17007 A @file{~/.emacs} file contains Emacs Lisp code. You can write this
17008 code yourself; or you can use Emacs' @code{customize} feature to write
17009 the code for you. You can combine your own expressions and
17010 auto-written Customize expressions in your @file{.emacs} file.
17012 (I myself prefer to write my own expressions, except for those,
17013 particularly fonts, that I find easier to manipulate using the
17014 @code{customize} command. I combine the two methods.)
17016 Most of this chapter is about writing expressions yourself. It
17017 describes a simple @file{.emacs} file; for more information, see
17018 @ref{Init File, , The Init File, emacs, The GNU Emacs Manual}, and
17019 @ref{Init File, , The Init File, elisp, The GNU Emacs Lisp Reference
17022 @node Site-wide Init, defcustom, Default Configuration, Emacs Initialization
17023 @section Site-wide Initialization Files
17025 @cindex @file{default.el} init file
17026 @cindex @file{site-init.el} init file
17027 @cindex @file{site-load.el} init file
17028 In addition to your personal initialization file, Emacs automatically
17029 loads various site-wide initialization files, if they exist. These
17030 have the same form as your @file{.emacs} file, but are loaded by
17033 Two site-wide initialization files, @file{site-load.el} and
17034 @file{site-init.el}, are loaded into Emacs and then `dumped' if a
17035 `dumped' version of Emacs is created, as is most common. (Dumped
17036 copies of Emacs load more quickly. However, once a file is loaded and
17037 dumped, a change to it does not lead to a change in Emacs unless you
17038 load it yourself or re-dump Emacs. @xref{Building Emacs, , Building
17039 Emacs, elisp, The GNU Emacs Lisp Reference Manual}, and the
17040 @file{INSTALL} file.)
17042 Three other site-wide initialization files are loaded automatically
17043 each time you start Emacs, if they exist. These are
17044 @file{site-start.el}, which is loaded @emph{before} your @file{.emacs}
17045 file, and @file{default.el}, and the terminal type file, which are both
17046 loaded @emph{after} your @file{.emacs} file.
17048 Settings and definitions in your @file{.emacs} file will overwrite
17049 conflicting settings and definitions in a @file{site-start.el} file,
17050 if it exists; but the settings and definitions in a @file{default.el}
17051 or terminal type file will overwrite those in your @file{.emacs} file.
17052 (You can prevent interference from a terminal type file by setting
17053 @code{term-file-prefix} to @code{nil}. @xref{Simple Extension, , A
17054 Simple Extension}.)
17056 @c Rewritten to avoid overfull hbox.
17057 The @file{INSTALL} file that comes in the distribution contains
17058 descriptions of the @file{site-init.el} and @file{site-load.el} files.
17060 The @file{loadup.el}, @file{startup.el}, and @file{loaddefs.el} files
17061 control loading. These files are in the @file{lisp} directory of the
17062 Emacs distribution and are worth perusing.
17064 The @file{loaddefs.el} file contains a good many suggestions as to
17065 what to put into your own @file{.emacs} file, or into a site-wide
17066 initialization file.
17068 @node defcustom, Beginning a .emacs File, Site-wide Init, Emacs Initialization
17069 @section Specifying Variables using @code{defcustom}
17072 You can specify variables using @code{defcustom} so that you and
17073 others can then use Emacs' @code{customize} feature to set their
17074 values. (You cannot use @code{customize} to write function
17075 definitions; but you can write @code{defuns} in your @file{.emacs}
17076 file. Indeed, you can write any Lisp expression in your @file{.emacs}
17079 The @code{customize} feature depends on the @code{defcustom} special
17080 form. Although you can use @code{defvar} or @code{setq} for variables
17081 that users set, the @code{defcustom} special form is designed for the
17084 You can use your knowledge of @code{defvar} for writing the
17085 first three arguments for @code{defcustom}. The first argument to
17086 @code{defcustom} is the name of the variable. The second argument is
17087 the variable's initial value, if any; and this value is set only if
17088 the value has not already been set. The third argument is the
17091 The fourth and subsequent arguments to @code{defcustom} specify types
17092 and options; these are not featured in @code{defvar}. (These
17093 arguments are optional.)
17095 Each of these arguments consists of a keyword followed by a value.
17096 Each keyword starts with the colon character @samp{:}.
17099 For example, the customizable user option variable
17100 @code{text-mode-hook} looks like this:
17104 (defcustom text-mode-hook nil
17105 "Normal hook run when entering Text mode and many related modes."
17107 :options '(turn-on-auto-fill flyspell-mode)
17113 The name of the variable is @code{text-mode-hook}; it has no default
17114 value; and its documentation string tells you what it does.
17116 The @code{:type} keyword tells Emacs the kind of data to which
17117 @code{text-mode-hook} should be set and how to display the value in a
17118 Customization buffer.
17120 The @code{:options} keyword specifies a suggested list of values for
17121 the variable. Usually, @code{:options} applies to a hook.
17122 The list is only a suggestion; it is not exclusive; a person who sets
17123 the variable may set it to other values; the list shown following the
17124 @code{:options} keyword is intended to offer convenient choices to a
17127 Finally, the @code{:group} keyword tells the Emacs Customization
17128 command in which group the variable is located. This tells where to
17131 The @code{defcustom} function recognizes more than a dozen keywords.
17132 For more information, see @ref{Customization, , Writing Customization
17133 Definitions, elisp, The GNU Emacs Lisp Reference Manual}.
17135 Consider @code{text-mode-hook} as an example.
17137 There are two ways to customize this variable. You can use the
17138 customization command or write the appropriate expressions yourself.
17141 Using the customization command, you can type:
17148 and find that the group for editing files of data is called `data'.
17149 Enter that group. Text Mode Hook is the first member. You can click
17150 on its various options, such as @code{turn-on-auto-fill}, to set the
17151 values. After you click on the button to
17154 Save for Future Sessions
17158 Emacs will write an expression into your @file{.emacs} file.
17159 It will look like this:
17163 (custom-set-variables
17164 ;; custom-set-variables was added by Custom.
17165 ;; If you edit it by hand, you could mess it up, so be careful.
17166 ;; Your init file should contain only one such instance.
17167 ;; If there is more than one, they won't work right.
17168 '(text-mode-hook (quote (turn-on-auto-fill text-mode-hook-identify))))
17173 (The @code{text-mode-hook-identify} function tells
17174 @code{toggle-text-mode-auto-fill} which buffers are in Text mode.
17175 It comes on automatically.)
17177 The @code{custom-set-variables} function works somewhat differently
17178 than a @code{setq}. While I have never learned the differences, I
17179 modify the @code{custom-set-variables} expressions in my @file{.emacs}
17180 file by hand: I make the changes in what appears to me to be a
17181 reasonable manner and have not had any problems. Others prefer to use
17182 the Customization command and let Emacs do the work for them.
17184 Another @code{custom-set-@dots{}} function is @code{custom-set-faces}.
17185 This function sets the various font faces. Over time, I have set a
17186 considerable number of faces. Some of the time, I re-set them using
17187 @code{customize}; other times, I simply edit the
17188 @code{custom-set-faces} expression in my @file{.emacs} file itself.
17190 The second way to customize your @code{text-mode-hook} is to set it
17191 yourself in your @file{.emacs} file using code that has nothing to do
17192 with the @code{custom-set-@dots{}} functions.
17195 When you do this, and later use @code{customize}, you will see a
17199 CHANGED outside Customize; operating on it here may be unreliable.
17203 This message is only a warning. If you click on the button to
17206 Save for Future Sessions
17210 Emacs will write a @code{custom-set-@dots{}} expression near the end
17211 of your @file{.emacs} file that will be evaluated after your
17212 hand-written expression. It will, therefore, overrule your
17213 hand-written expression. No harm will be done. When you do this,
17214 however, be careful to remember which expression is active; if you
17215 forget, you may confuse yourself.
17217 So long as you remember where the values are set, you will have no
17218 trouble. In any event, the values are always set in your
17219 initialization file, which is usually called @file{.emacs}.
17221 I myself use @code{customize} for hardly anything. Mostly, I write
17222 expressions myself.
17226 Incidentally, to be more complete concerning defines: @code{defsubst}
17227 defines an inline function. The syntax is just like that of
17228 @code{defun}. @code{defconst} defines a symbol as a constant. The
17229 intent is that neither programs nor users should ever change a value
17230 set by @code{defconst}. (You can change it; the value set is a
17231 variable; but please do not.)
17233 @node Beginning a .emacs File, Text and Auto-fill, defcustom, Emacs Initialization
17234 @section Beginning a @file{.emacs} File
17235 @cindex @file{.emacs} file, beginning of
17237 When you start Emacs, it loads your @file{.emacs} file unless you tell
17238 it not to by specifying @samp{-q} on the command line. (The
17239 @code{emacs -q} command gives you a plain, out-of-the-box Emacs.)
17241 A @file{.emacs} file contains Lisp expressions. Often, these are no
17242 more than expressions to set values; sometimes they are function
17245 @xref{Init File, , The Init File @file{~/.emacs}, emacs, The GNU Emacs
17246 Manual}, for a short description of initialization files.
17248 This chapter goes over some of the same ground, but is a walk among
17249 extracts from a complete, long-used @file{.emacs} file---my own.
17251 The first part of the file consists of comments: reminders to myself.
17252 By now, of course, I remember these things, but when I started, I did
17258 ;;;; Bob's .emacs file
17259 ; Robert J. Chassell
17260 ; 26 September 1985
17265 Look at that date! I started this file a long time ago. I have been
17266 adding to it ever since.
17270 ; Each section in this file is introduced by a
17271 ; line beginning with four semicolons; and each
17272 ; entry is introduced by a line beginning with
17273 ; three semicolons.
17278 This describes the usual conventions for comments in Emacs Lisp.
17279 Everything on a line that follows a semicolon is a comment. Two,
17280 three, and four semicolons are used as subsection and section markers.
17281 (@xref{Comments, ,, elisp, The GNU Emacs Lisp Reference Manual}, for
17282 more about comments.)
17287 ; Control-h is the help key;
17288 ; after typing control-h, type a letter to
17289 ; indicate the subject about which you want help.
17290 ; For an explanation of the help facility,
17291 ; type control-h two times in a row.
17296 Just remember: type @kbd{C-h} two times for help.
17300 ; To find out about any mode, type control-h m
17301 ; while in that mode. For example, to find out
17302 ; about mail mode, enter mail mode and then type
17308 `Mode help', as I call this, is very helpful. Usually, it tells you
17309 all you need to know.
17311 Of course, you don't need to include comments like these in your
17312 @file{.emacs} file. I included them in mine because I kept forgetting
17313 about Mode help or the conventions for comments---but I was able to
17314 remember to look here to remind myself.
17316 @node Text and Auto-fill, Mail Aliases, Beginning a .emacs File, Emacs Initialization
17317 @section Text and Auto Fill Mode
17319 Now we come to the part that `turns on' Text mode and
17324 ;;; Text mode and Auto Fill mode
17325 ; The next two lines put Emacs into Text mode
17326 ; and Auto Fill mode, and are for writers who
17327 ; want to start writing prose rather than code.
17328 (setq default-major-mode 'text-mode)
17329 (add-hook 'text-mode-hook 'turn-on-auto-fill)
17333 Here is the first part of this @file{.emacs} file that does something
17334 besides remind a forgetful human!
17336 The first of the two lines in parentheses tells Emacs to turn on Text
17337 mode when you find a file, @emph{unless} that file should go into some
17338 other mode, such as C mode.
17340 @cindex Per-buffer, local variables list
17341 @cindex Local variables list, per-buffer,
17342 @cindex Automatic mode selection
17343 @cindex Mode selection, automatic
17344 When Emacs reads a file, it looks at the extension to the file name,
17345 if any. (The extension is the part that comes after a @samp{.}.) If
17346 the file ends with a @samp{.c} or @samp{.h} extension then Emacs turns
17347 on C mode. Also, Emacs looks at first nonblank line of the file; if
17348 the line says @w{@samp{-*- C -*-}}, Emacs turns on C mode. Emacs
17349 possesses a list of extensions and specifications that it uses
17350 automatically. In addition, Emacs looks near the last page for a
17351 per-buffer, ``local variables list'', if any.
17354 @xref{Choosing Modes, , How Major Modes are Chosen, emacs, The GNU
17357 @xref{File Variables, , Local Variables in Files, emacs, The GNU Emacs
17361 See sections ``How Major Modes are Chosen'' and ``Local Variables in
17362 Files'' in @cite{The GNU Emacs Manual}.
17365 Now, back to the @file{.emacs} file.
17368 Here is the line again; how does it work?
17370 @cindex Text Mode turned on
17372 (setq default-major-mode 'text-mode)
17376 This line is a short, but complete Emacs Lisp expression.
17378 We are already familiar with @code{setq}. It sets the following variable,
17379 @code{default-major-mode}, to the subsequent value, which is
17380 @code{text-mode}. The single quote mark before @code{text-mode} tells
17381 Emacs to deal directly with the @code{text-mode} variable, not with
17382 whatever it might stand for. @xref{set & setq, , Setting the Value of
17383 a Variable}, for a reminder of how @code{setq} works. The main point
17384 is that there is no difference between the procedure you use to set
17385 a value in your @file{.emacs} file and the procedure you use anywhere
17389 Here is the next line:
17391 @cindex Auto Fill mode turned on
17394 (add-hook 'text-mode-hook 'turn-on-auto-fill)
17398 In this line, the @code{add-hook} command adds
17399 @code{turn-on-auto-fill} to the variable.
17401 @code{turn-on-auto-fill} is the name of a program, that, you guessed
17402 it!, turns on Auto Fill mode.
17404 Every time Emacs turns on Text mode, Emacs runs the commands `hooked'
17405 onto Text mode. So every time Emacs turns on Text mode, Emacs also
17406 turns on Auto Fill mode.
17408 In brief, the first line causes Emacs to enter Text mode when you edit a
17409 file, unless the file name extension, a first non-blank line, or local
17410 variables to tell Emacs otherwise.
17412 Text mode among other actions, sets the syntax table to work
17413 conveniently for writers. In Text mode, Emacs considers an apostrophe
17414 as part of a word like a letter; but Emacs does not consider a period
17415 or a space as part of a word. Thus, @kbd{M-f} moves you over
17416 @samp{it's}. On the other hand, in C mode, @kbd{M-f} stops just after
17417 the @samp{t} of @samp{it's}.
17419 The second line causes Emacs to turn on Auto Fill mode when it turns
17420 on Text mode. In Auto Fill mode, Emacs automatically breaks a line
17421 that is too wide and brings the excessively wide part of the line down
17422 to the next line. Emacs breaks lines between words, not within them.
17424 When Auto Fill mode is turned off, lines continue to the right as you
17425 type them. Depending on how you set the value of
17426 @code{truncate-lines}, the words you type either disappear off the
17427 right side of the screen, or else are shown, in a rather ugly and
17428 unreadable manner, as a continuation line on the screen.
17431 In addition, in this part of my @file{.emacs} file, I tell the Emacs
17432 fill commands to insert two spaces after a colon:
17435 (setq colon-double-space t)
17438 @node Mail Aliases, Indent Tabs Mode, Text and Auto-fill, Emacs Initialization
17439 @section Mail Aliases
17441 Here is a @code{setq} that `turns on' mail aliases, along with more
17447 ; To enter mail mode, type `C-x m'
17448 ; To enter RMAIL (for reading mail),
17450 (setq mail-aliases t)
17454 @cindex Mail aliases
17456 This @code{setq} command sets the value of the variable
17457 @code{mail-aliases} to @code{t}. Since @code{t} means true, the line
17458 says, in effect, ``Yes, use mail aliases.''
17460 Mail aliases are convenient short names for long email addresses or
17461 for lists of email addresses. The file where you keep your `aliases'
17462 is @file{~/.mailrc}. You write an alias like this:
17465 alias geo george@@foobar.wiz.edu
17469 When you write a message to George, address it to @samp{geo}; the
17470 mailer will automatically expand @samp{geo} to the full address.
17472 @node Indent Tabs Mode, Keybindings, Mail Aliases, Emacs Initialization
17473 @section Indent Tabs Mode
17474 @cindex Tabs, preventing
17475 @findex indent-tabs-mode
17477 By default, Emacs inserts tabs in place of multiple spaces when it
17478 formats a region. (For example, you might indent many lines of text
17479 all at once with the @code{indent-region} command.) Tabs look fine on
17480 a terminal or with ordinary printing, but they produce badly indented
17481 output when you use @TeX{} or Texinfo since @TeX{} ignores tabs.
17484 The following turns off Indent Tabs mode:
17488 ;;; Prevent Extraneous Tabs
17489 (setq-default indent-tabs-mode nil)
17493 Note that this line uses @code{setq-default} rather than the
17494 @code{setq} command that we have seen before. The @code{setq-default}
17495 command sets values only in buffers that do not have their own local
17496 values for the variable.
17499 @xref{Just Spaces, , Tabs vs. Spaces, emacs, The GNU Emacs Manual}.
17501 @xref{File Variables, , Local Variables in Files, emacs, The GNU Emacs
17505 See sections ``Tabs vs.@: Spaces'' and ``Local Variables in
17506 Files'' in @cite{The GNU Emacs Manual}.
17510 @node Keybindings, Keymaps, Indent Tabs Mode, Emacs Initialization
17511 @section Some Keybindings
17513 Now for some personal keybindings:
17517 ;;; Compare windows
17518 (global-set-key "\C-cw" 'compare-windows)
17522 @findex compare-windows
17523 @code{compare-windows} is a nifty command that compares the text in
17524 your current window with text in the next window. It makes the
17525 comparison by starting at point in each window, moving over text in
17526 each window as far as they match. I use this command all the time.
17528 This also shows how to set a key globally, for all modes.
17530 @cindex Setting a key globally
17531 @cindex Global set key
17532 @cindex Key setting globally
17533 @findex global-set-key
17534 The command is @code{global-set-key}. It is followed by the
17535 keybinding. In a @file{.emacs} file, the keybinding is written as
17536 shown: @code{\C-c} stands for `control-c', which means `press the
17537 control key and the @key{c} key at the same time'. The @code{w} means
17538 `press the @key{w} key'. The keybinding is surrounded by double
17539 quotation marks. In documentation, you would write this as
17540 @w{@kbd{C-c w}}. (If you were binding a @key{META} key, such as
17541 @kbd{M-c}, rather than a @key{CTRL} key, you would write
17542 @w{@code{\M-c}} in your @file{.emacs} file. @xref{Init Rebinding, ,
17543 Rebinding Keys in Your Init File, emacs, The GNU Emacs Manual}, for
17546 The command invoked by the keys is @code{compare-windows}. Note that
17547 @code{compare-windows} is preceded by a single quote; otherwise, Emacs
17548 would first try to evaluate the symbol to determine its value.
17550 These three things, the double quotation marks, the backslash before
17551 the @samp{C}, and the single quote mark are necessary parts of
17552 keybinding that I tend to forget. Fortunately, I have come to
17553 remember that I should look at my existing @file{.emacs} file, and
17554 adapt what is there.
17556 As for the keybinding itself: @kbd{C-c w}. This combines the prefix
17557 key, @kbd{C-c}, with a single character, in this case, @kbd{w}. This
17558 set of keys, @kbd{C-c} followed by a single character, is strictly
17559 reserved for individuals' own use. (I call these `own' keys, since
17560 these are for my own use.) You should always be able to create such a
17561 keybinding for your own use without stomping on someone else's
17562 keybinding. If you ever write an extension to Emacs, please avoid
17563 taking any of these keys for public use. Create a key like @kbd{C-c
17564 C-w} instead. Otherwise, we will run out of `own' keys.
17567 Here is another keybinding, with a comment:
17571 ;;; Keybinding for `occur'
17572 ; I use occur a lot, so let's bind it to a key:
17573 (global-set-key "\C-co" 'occur)
17578 The @code{occur} command shows all the lines in the current buffer
17579 that contain a match for a regular expression. Matching lines are
17580 shown in a buffer called @file{*Occur*}. That buffer serves as a menu
17581 to jump to occurrences.
17583 @findex global-unset-key
17584 @cindex Unbinding key
17585 @cindex Key unbinding
17587 Here is how to unbind a key, so it does not
17593 (global-unset-key "\C-xf")
17597 There is a reason for this unbinding: I found I inadvertently typed
17598 @w{@kbd{C-x f}} when I meant to type @kbd{C-x C-f}. Rather than find a
17599 file, as I intended, I accidentally set the width for filled text,
17600 almost always to a width I did not want. Since I hardly ever reset my
17601 default width, I simply unbound the key.
17603 @findex list-buffers, @r{rebound}
17604 @findex buffer-menu, @r{bound to key}
17606 The following rebinds an existing key:
17610 ;;; Rebind `C-x C-b' for `buffer-menu'
17611 (global-set-key "\C-x\C-b" 'buffer-menu)
17615 By default, @kbd{C-x C-b} runs the
17616 @code{list-buffers} command. This command lists
17617 your buffers in @emph{another} window. Since I
17618 almost always want to do something in that
17619 window, I prefer the @code{buffer-menu}
17620 command, which not only lists the buffers,
17621 but moves point into that window.
17623 @node Keymaps, Loading Files, Keybindings, Emacs Initialization
17626 @cindex Rebinding keys
17628 Emacs uses @dfn{keymaps} to record which keys call which commands.
17629 When you use @code{global-set-key} to set the keybinding for a single
17630 command in all parts of Emacs, you are specifying the keybinding in
17631 @code{current-global-map}.
17633 Specific modes, such as C mode or Text mode, have their own keymaps;
17634 the mode-specific keymaps override the global map that is shared by
17637 The @code{global-set-key} function binds, or rebinds, the global
17638 keymap. For example, the following binds the key @kbd{C-x C-b} to the
17639 function @code{buffer-menu}:
17642 (global-set-key "\C-x\C-b" 'buffer-menu)
17645 Mode-specific keymaps are bound using the @code{define-key} function,
17646 which takes a specific keymap as an argument, as well as the key and
17647 the command. For example, my @file{.emacs} file contains the
17648 following expression to bind the @code{texinfo-insert-@@group} command
17649 to @kbd{C-c C-c g}:
17653 (define-key texinfo-mode-map "\C-c\C-cg" 'texinfo-insert-@@group)
17658 The @code{texinfo-insert-@@group} function itself is a little extension
17659 to Texinfo mode that inserts @samp{@@group} into a Texinfo file. I
17660 use this command all the time and prefer to type the three strokes
17661 @kbd{C-c C-c g} rather than the six strokes @kbd{@@ g r o u p}.
17662 (@samp{@@group} and its matching @samp{@@end group} are commands that
17663 keep all enclosed text together on one page; many multi-line examples
17664 in this book are surrounded by @samp{@@group @dots{} @@end group}.)
17667 Here is the @code{texinfo-insert-@@group} function definition:
17671 (defun texinfo-insert-@@group ()
17672 "Insert the string @@group in a Texinfo buffer."
17674 (beginning-of-line)
17675 (insert "@@group\n"))
17679 (Of course, I could have used Abbrev mode to save typing, rather than
17680 write a function to insert a word; but I prefer key strokes consistent
17681 with other Texinfo mode key bindings.)
17683 You will see numerous @code{define-key} expressions in
17684 @file{loaddefs.el} as well as in the various mode libraries, such as
17685 @file{cc-mode.el} and @file{lisp-mode.el}.
17687 @xref{Key Bindings, , Customizing Key Bindings, emacs, The GNU Emacs
17688 Manual}, and @ref{Keymaps, , Keymaps, elisp, The GNU Emacs Lisp
17689 Reference Manual}, for more information about keymaps.
17691 @node Loading Files, Autoload, Keymaps, Emacs Initialization
17692 @section Loading Files
17693 @cindex Loading files
17696 Many people in the GNU Emacs community have written extensions to
17697 Emacs. As time goes by, these extensions are often included in new
17698 releases. For example, the Calendar and Diary packages are now part
17699 of the standard GNU Emacs, as is Calc.
17701 You can use a @code{load} command to evaluate a complete file and
17702 thereby install all the functions and variables in the file into Emacs.
17705 @c (auto-compression-mode t)
17708 (load "~/emacs/slowsplit")
17711 This evaluates, i.e.@: loads, the @file{slowsplit.el} file or if it
17712 exists, the faster, byte compiled @file{slowsplit.elc} file from the
17713 @file{emacs} sub-directory of your home directory. The file contains
17714 the function @code{split-window-quietly}, which John Robinson wrote in
17717 The @code{split-window-quietly} function splits a window with the
17718 minimum of redisplay. I installed it in 1989 because it worked well
17719 with the slow 1200 baud terminals I was then using. Nowadays, I only
17720 occasionally come across such a slow connection, but I continue to use
17721 the function because I like the way it leaves the bottom half of a
17722 buffer in the lower of the new windows and the top half in the upper
17726 To replace the key binding for the default
17727 @code{split-window-vertically}, you must also unset that key and bind
17728 the keys to @code{split-window-quietly}, like this:
17732 (global-unset-key "\C-x2")
17733 (global-set-key "\C-x2" 'split-window-quietly)
17738 If you load many extensions, as I do, then instead of specifying the
17739 exact location of the extension file, as shown above, you can specify
17740 that directory as part of Emacs' @code{load-path}. Then, when Emacs
17741 loads a file, it will search that directory as well as its default
17742 list of directories. (The default list is specified in @file{paths.h}
17743 when Emacs is built.)
17746 The following command adds your @file{~/emacs} directory to the
17747 existing load path:
17751 ;;; Emacs Load Path
17752 (setq load-path (cons "~/emacs" load-path))
17756 Incidentally, @code{load-library} is an interactive interface to the
17757 @code{load} function. The complete function looks like this:
17759 @findex load-library
17762 (defun load-library (library)
17763 "Load the library named LIBRARY.
17764 This is an interface to the function `load'."
17766 (list (completing-read "Load library: "
17767 'locate-file-completion
17768 (cons load-path (get-load-suffixes)))))
17773 The name of the function, @code{load-library}, comes from the use of
17774 `library' as a conventional synonym for `file'. The source for the
17775 @code{load-library} command is in the @file{files.el} library.
17777 Another interactive command that does a slightly different job is
17778 @code{load-file}. @xref{Lisp Libraries, , Libraries of Lisp Code for
17779 Emacs, emacs, The GNU Emacs Manual}, for information on the
17780 distinction between @code{load-library} and this command.
17782 @node Autoload, Simple Extension, Loading Files, Emacs Initialization
17783 @section Autoloading
17786 Instead of installing a function by loading the file that contains it,
17787 or by evaluating the function definition, you can make the function
17788 available but not actually install it until it is first called. This
17789 is called @dfn{autoloading}.
17791 When you execute an autoloaded function, Emacs automatically evaluates
17792 the file that contains the definition, and then calls the function.
17794 Emacs starts quicker with autoloaded functions, since their libraries
17795 are not loaded right away; but you need to wait a moment when you
17796 first use such a function, while its containing file is evaluated.
17798 Rarely used functions are frequently autoloaded. The
17799 @file{loaddefs.el} library contains hundreds of autoloaded functions,
17800 from @code{bookmark-set} to @code{wordstar-mode}. Of course, you may
17801 come to use a `rare' function frequently. When you do, you should
17802 load that function's file with a @code{load} expression in your
17803 @file{.emacs} file.
17805 In my @file{.emacs} file, I load 14 libraries that contain functions
17806 that would otherwise be autoloaded. (Actually, it would have been
17807 better to include these files in my `dumped' Emacs, but I forgot.
17808 @xref{Building Emacs, , Building Emacs, elisp, The GNU Emacs Lisp
17809 Reference Manual}, and the @file{INSTALL} file for more about
17812 You may also want to include autoloaded expressions in your @file{.emacs}
17813 file. @code{autoload} is a built-in function that takes up to five
17814 arguments, the final three of which are optional. The first argument
17815 is the name of the function to be autoloaded; the second is the name
17816 of the file to be loaded. The third argument is documentation for the
17817 function, and the fourth tells whether the function can be called
17818 interactively. The fifth argument tells what type of
17819 object---@code{autoload} can handle a keymap or macro as well as a
17820 function (the default is a function).
17823 Here is a typical example:
17827 (autoload 'html-helper-mode
17828 "html-helper-mode" "Edit HTML documents" t)
17833 (@code{html-helper-mode} is an older alternative to @code{html-mode},
17834 which is a standard part of the distribution.)
17837 This expression autoloads the @code{html-helper-mode} function. It
17838 takes it from the @file{html-helper-mode.el} file (or from the byte
17839 compiled file @file{html-helper-mode.elc}, if it exists.) The file
17840 must be located in a directory specified by @code{load-path}. The
17841 documentation says that this is a mode to help you edit documents
17842 written in the HyperText Markup Language. You can call this mode
17843 interactively by typing @kbd{M-x html-helper-mode}. (You need to
17844 duplicate the function's regular documentation in the autoload
17845 expression because the regular function is not yet loaded, so its
17846 documentation is not available.)
17848 @xref{Autoload, , Autoload, elisp, The GNU Emacs Lisp Reference
17849 Manual}, for more information.
17851 @node Simple Extension, X11 Colors, Autoload, Emacs Initialization
17852 @section A Simple Extension: @code{line-to-top-of-window}
17853 @findex line-to-top-of-window
17854 @cindex Simple extension in @file{.emacs} file
17856 Here is a simple extension to Emacs that moves the line point is on to
17857 the top of the window. I use this all the time, to make text easier
17860 You can put the following code into a separate file and then load it
17861 from your @file{.emacs} file, or you can include it within your
17862 @file{.emacs} file.
17865 Here is the definition:
17869 ;;; Line to top of window;
17870 ;;; replace three keystroke sequence C-u 0 C-l
17871 (defun line-to-top-of-window ()
17872 "Move the line point is on to top of window."
17879 Now for the keybinding.
17881 Nowadays, function keys as well as mouse button events and
17882 non-@sc{ascii} characters are written within square brackets, without
17883 quotation marks. (In Emacs version 18 and before, you had to write
17884 different function key bindings for each different make of terminal.)
17886 I bind @code{line-to-top-of-window} to my @key{F6} function key like
17890 (global-set-key [f6] 'line-to-top-of-window)
17893 For more information, see @ref{Init Rebinding, , Rebinding Keys in
17894 Your Init File, emacs, The GNU Emacs Manual}.
17896 @cindex Conditional 'twixt two versions of Emacs
17897 @cindex Version of Emacs, choosing
17898 @cindex Emacs version, choosing
17899 If you run two versions of GNU Emacs, such as versions 21 and 22, and
17900 use one @file{.emacs} file, you can select which code to evaluate with
17901 the following conditional:
17906 (= 21 emacs-major-version)
17907 ;; evaluate version 21 code
17909 (= 22 emacs-major-version)
17910 ;; evaluate version 22 code
17915 For example, in contrast to version 20, more recent versions blink
17916 their cursors by default. I hate such blinking, as well as other
17917 features, so I placed the following in my @file{.emacs}
17918 file@footnote{When I start instances of Emacs that do not load my
17919 @file{.emacs} file or any site file, I also turn off blinking:
17922 emacs -q --no-site-file -eval '(blink-cursor-mode nil)'
17924 @exdent Or nowadays, using an even more sophisticated set of options,
17932 (when (or (= 21 emacs-major-version)
17933 (= 22 emacs-major-version))
17934 (blink-cursor-mode 0)
17935 ;; Insert newline when you press `C-n' (next-line)
17936 ;; at the end of the buffer
17937 (setq next-line-add-newlines t)
17940 ;; Turn on image viewing
17941 (auto-image-file-mode t)
17944 ;; Turn on menu bar (this bar has text)
17945 ;; (Use numeric argument to turn on)
17949 ;; Turn off tool bar (this bar has icons)
17950 ;; (Use numeric argument to turn on)
17951 (tool-bar-mode nil)
17954 ;; Turn off tooltip mode for tool bar
17955 ;; (This mode causes icon explanations to pop up)
17956 ;; (Use numeric argument to turn on)
17958 ;; If tooltips turned on, make tips appear promptly
17959 (setq tooltip-delay 0.1) ; default is 0.7 second
17965 Alternatively, since @code{blink-cursor-mode} has existed since Emacs
17966 version 21 and is likely to continue, you could write
17970 (when (>= emacs-major-version 21)
17971 (blink-cursor-mode 0)
17976 and add other expressions, too.
17979 @node X11 Colors, Miscellaneous, Simple Extension, Emacs Initialization
17980 @section X11 Colors
17982 You can specify colors when you use Emacs with the MIT X Windowing
17985 I dislike the default colors and specify my own.
17988 Here are the expressions in my @file{.emacs}
17989 file that set values:
17993 ;; Set cursor color
17994 (set-cursor-color "white")
17997 (set-mouse-color "white")
17999 ;; Set foreground and background
18000 (set-foreground-color "white")
18001 (set-background-color "darkblue")
18005 ;;; Set highlighting colors for isearch and drag
18006 (set-face-foreground 'highlight "white")
18007 (set-face-background 'highlight "blue")
18011 (set-face-foreground 'region "cyan")
18012 (set-face-background 'region "blue")
18016 (set-face-foreground 'secondary-selection "skyblue")
18017 (set-face-background 'secondary-selection "darkblue")
18021 ;; Set calendar highlighting colors
18022 (setq calendar-load-hook
18024 (set-face-foreground 'diary-face "skyblue")
18025 (set-face-background 'holiday-face "slate blue")
18026 (set-face-foreground 'holiday-face "white")))
18030 The various shades of blue soothe my eye and prevent me from seeing
18031 the screen flicker.
18033 Alternatively, I could have set my specifications in various X
18034 initialization files. For example, I could set the foreground,
18035 background, cursor, and pointer (i.e., mouse) colors in my
18036 @file{~/.Xresources} file like this:
18040 Emacs*foreground: white
18041 Emacs*background: darkblue
18042 Emacs*cursorColor: white
18043 Emacs*pointerColor: white
18047 In any event, since it is not part of Emacs, I set the root color of
18048 my X window in my @file{~/.xinitrc} file, like this@footnote{I also
18049 run more modern window managers, such as Enlightenment, Gnome, or KDE;
18050 in those cases, I often specify an image rather than a plain color.}:
18053 xsetroot -solid Navy -fg white &
18057 @node Miscellaneous, Mode Line, X11 Colors, Emacs Initialization
18058 @section Miscellaneous Settings for a @file{.emacs} File
18061 Here are a few miscellaneous settings:
18066 Set the shape and color of the mouse cursor:
18070 ; Cursor shapes are defined in
18071 ; `/usr/include/X11/cursorfont.h';
18072 ; for example, the `target' cursor is number 128;
18073 ; the `top_left_arrow' cursor is number 132.
18077 (let ((mpointer (x-get-resource "*mpointer"
18078 "*emacs*mpointer")))
18079 ;; If you have not set your mouse pointer
18080 ;; then set it, otherwise leave as is:
18081 (if (eq mpointer nil)
18082 (setq mpointer "132")) ; top_left_arrow
18085 (setq x-pointer-shape (string-to-int mpointer))
18086 (set-mouse-color "white"))
18091 Or you can set the values of a variety of features in an alist, like
18097 default-frame-alist
18098 '((cursor-color . "white")
18099 (mouse-color . "white")
18100 (foreground-color . "white")
18101 (background-color . "DodgerBlue4")
18102 ;; (cursor-type . bar)
18103 (cursor-type . box)
18106 (tool-bar-lines . 0)
18107 (menu-bar-lines . 1)
18111 "-Misc-Fixed-Medium-R-Normal--20-200-75-75-C-100-ISO8859-1")
18117 Convert @kbd{@key{CTRL}-h} into @key{DEL} and @key{DEL}
18118 into @kbd{@key{CTRL}-h}.@*
18119 (Some older keyboards needed this, although I have not seen the
18124 ;; Translate `C-h' to <DEL>.
18125 ; (keyboard-translate ?\C-h ?\C-?)
18127 ;; Translate <DEL> to `C-h'.
18128 (keyboard-translate ?\C-? ?\C-h)
18132 @item Turn off a blinking cursor!
18136 (if (fboundp 'blink-cursor-mode)
18137 (blink-cursor-mode -1))
18142 or start GNU Emacs with the command @code{emacs -nbc}.
18145 @item When using `grep'@*
18146 @samp{-i}@w{ } Ignore case distinctions@*
18147 @samp{-n}@w{ } Prefix each line of output with line number@*
18148 @samp{-H}@w{ } Print the filename for each match.@*
18149 @samp{-e}@w{ } Protect patterns beginning with a hyphen character, @samp{-}
18152 (setq grep-command "grep -i -nH -e ")
18156 @c Evidently, no longer needed in GNU Emacs 22
18158 item Automatically uncompress compressed files when visiting them
18161 (load "uncompress")
18166 @item Find an existing buffer, even if it has a different name@*
18167 This avoids problems with symbolic links.
18170 (setq find-file-existing-other-name t)
18173 @item Set your language environment and default input method
18177 (set-language-environment "latin-1")
18178 ;; Remember you can enable or disable multilingual text input
18179 ;; with the @code{toggle-input-method'} (@kbd{C-\}) command
18180 (setq default-input-method "latin-1-prefix")
18184 If you want to write with Chinese `GB' characters, set this instead:
18188 (set-language-environment "Chinese-GB")
18189 (setq default-input-method "chinese-tonepy")
18194 @subsubheading Fixing Unpleasant Key Bindings
18195 @cindex Key bindings, fixing
18196 @cindex Bindings, key, fixing unpleasant
18198 Some systems bind keys unpleasantly. Sometimes, for example, the
18199 @key{CTRL} key appears in an awkward spot rather than at the far left
18202 Usually, when people fix these sorts of keybindings, they do not
18203 change their @file{~/.emacs} file. Instead, they bind the proper keys
18204 on their consoles with the @code{loadkeys} or @code{install-keymap}
18205 commands in their boot script and then include @code{xmodmap} commands
18206 in their @file{.xinitrc} or @file{.Xsession} file for X Windows.
18214 loadkeys /usr/share/keymaps/i386/qwerty/emacs2.kmap.gz
18216 install-keymap emacs2
18222 For a @file{.xinitrc} or @file{.Xsession} file when the @key{Caps
18223 Lock} key is at the far left of the home row:
18227 # Bind the key labeled `Caps Lock' to `Control'
18228 # (Such a broken user interface suggests that keyboard manufacturers
18229 # think that computers are typewriters from 1885.)
18231 xmodmap -e "clear Lock"
18232 xmodmap -e "add Control = Caps_Lock"
18238 In a @file{.xinitrc} or @file{.Xsession} file, to convert an @key{ALT}
18239 key to a @key{META} key:
18243 # Some ill designed keyboards have a key labeled ALT and no Meta
18244 xmodmap -e "keysym Alt_L = Meta_L Alt_L"
18249 @node Mode Line, , Miscellaneous, Emacs Initialization
18250 @section A Modified Mode Line
18251 @vindex default-mode-line-format
18252 @cindex Mode line format
18254 Finally, a feature I really like: a modified mode line.
18256 When I work over a network, I forget which machine I am using. Also,
18257 I tend to I lose track of where I am, and which line point is on.
18259 So I reset my mode line to look like this:
18262 -:-- foo.texi rattlesnake:/home/bob/ Line 1 (Texinfo Fill) Top
18265 I am visiting a file called @file{foo.texi}, on my machine
18266 @file{rattlesnake} in my @file{/home/bob} buffer. I am on line 1, in
18267 Texinfo mode, and am at the top of the buffer.
18270 My @file{.emacs} file has a section that looks like this:
18274 ;; Set a Mode Line that tells me which machine, which directory,
18275 ;; and which line I am on, plus the other customary information.
18276 (setq default-mode-line-format
18280 "mouse-1: select window, mouse-2: delete others ..."))
18281 mode-line-mule-info
18283 mode-line-frame-identification
18287 mode-line-buffer-identification
18290 (system-name) 0 (string-match "\\..+" (system-name))))
18295 "mouse-1: select window, mouse-2: delete others ..."))
18296 (line-number-mode " Line %l ")
18302 "mouse-1: select window, mouse-2: delete others ..."))
18303 (:eval (mode-line-mode-name))
18306 #("%n" 0 2 (help-echo "mouse-2: widen" local-map (keymap ...)))
18315 Here, I redefine the default mode line. Most of the parts are from
18316 the original; but I make a few changes. I set the @emph{default} mode
18317 line format so as to permit various modes, such as Info, to override
18320 Many elements in the list are self-explanatory:
18321 @code{mode-line-modified} is a variable that tells whether the buffer
18322 has been modified, @code{mode-name} tells the name of the mode, and so
18323 on. However, the format looks complicated because of two features we
18324 have not discussed.
18326 @cindex Properties, in mode line example
18327 The first string in the mode line is a dash or hyphen, @samp{-}. In
18328 the old days, it would have been specified simply as @code{"-"}. But
18329 nowadays, Emacs can add properties to a string, such as highlighting
18330 or, as in this case, a help feature. If you place your mouse cursor
18331 over the hyphen, some help information appears (By default, you must
18332 wait seven-tenths of a second before the information appears. You can
18333 change that timing by changing the value of @code{tooltip-delay}.)
18336 The new string format has a special syntax:
18339 #("-" 0 1 (help-echo "mouse-1: select window, ..."))
18343 The @code{#(} begins a list. The first element of the list is the
18344 string itself, just one @samp{-}. The second and third
18345 elements specify the range over which the fourth element applies. A
18346 range starts @emph{after} a character, so a zero means the range
18347 starts just before the first character; a 1 means that the range ends
18348 just after the first character. The third element is the property for
18349 the range. It consists of a property list, a
18350 property name, in this case, @samp{help-echo}, followed by a value, in this
18351 case, a string. The second, third, and fourth elements of this new
18352 string format can be repeated.
18354 @xref{Text Properties, , Text Properties, elisp, The GNU Emacs Lisp
18355 Reference Manual}, and see @ref{Mode Line Format, , Mode Line Format,
18356 elisp, The GNU Emacs Lisp Reference Manual}, for more information.
18358 @code{mode-line-buffer-identification}
18359 displays the current buffer name. It is a list
18360 beginning @code{(#("%12b" 0 4 @dots{}}.
18361 The @code{#(} begins the list.
18363 The @samp{"%12b"} displays the current buffer name, using the
18364 @code{buffer-name} function with which we are familiar; the `12'
18365 specifies the maximum number of characters that will be displayed.
18366 When a name has fewer characters, whitespace is added to fill out to
18367 this number. (Buffer names can and often should be longer than 12
18368 characters; this length works well in a typical 80 column wide
18371 @code{:eval} says to evaluate the following form and use the result as
18372 a string to display. In this case, the expression displays the first
18373 component of the full system name. The end of the first component is
18374 a @samp{.} (`period'), so I use the @code{string-match} function to
18375 tell me the length of the first component. The substring from the
18376 zeroth character to that length is the name of the machine.
18379 This is the expression:
18384 (system-name) 0 (string-match "\\..+" (system-name))))
18388 @samp{%[} and @samp{%]} cause a pair of square brackets
18389 to appear for each recursive editing level. @samp{%n} says `Narrow'
18390 when narrowing is in effect. @samp{%P} tells you the percentage of
18391 the buffer that is above the bottom of the window, or `Top', `Bottom',
18392 or `All'. (A lower case @samp{p} tell you the percentage above the
18393 @emph{top} of the window.) @samp{%-} inserts enough dashes to fill
18396 Remember, ``You don't have to like Emacs to like it'' --- your own
18397 Emacs can have different colors, different commands, and different
18398 keys than a default Emacs.
18400 On the other hand, if you want to bring up a plain `out of the box'
18401 Emacs, with no customization, type:
18408 This will start an Emacs that does @emph{not} load your
18409 @file{~/.emacs} initialization file. A plain, default Emacs. Nothing
18412 @node Debugging, Conclusion, Emacs Initialization, Top
18416 GNU Emacs has two debuggers, @code{debug} and @code{edebug}. The
18417 first is built into the internals of Emacs and is always with you;
18418 the second requires that you instrument a function before you can use it.
18420 Both debuggers are described extensively in @ref{Debugging, ,
18421 Debugging Lisp Programs, elisp, The GNU Emacs Lisp Reference Manual}.
18422 In this chapter, I will walk through a short example of each.
18429 * Debugging Exercises::
18432 @node debug, debug-on-entry, Debugging, Debugging
18433 @section @code{debug}
18436 Suppose you have written a function definition that is intended to
18437 return the sum of the numbers 1 through a given number. (This is the
18438 @code{triangle} function discussed earlier. @xref{Decrementing
18439 Example, , Example with Decrementing Counter}, for a discussion.)
18440 @c xref{Decrementing Loop,, Loop with a Decrementing Counter}, for a discussion.)
18442 However, your function definition has a bug. You have mistyped
18443 @samp{1=} for @samp{1-}. Here is the broken definition:
18445 @findex triangle-bugged
18448 (defun triangle-bugged (number)
18449 "Return sum of numbers 1 through NUMBER inclusive."
18451 (while (> number 0)
18452 (setq total (+ total number))
18453 (setq number (1= number))) ; @r{Error here.}
18458 If you are reading this in Info, you can evaluate this definition in
18459 the normal fashion. You will see @code{triangle-bugged} appear in the
18463 Now evaluate the @code{triangle-bugged} function with an
18467 (triangle-bugged 4)
18471 In a recent GNU Emacs, you will create and enter a @file{*Backtrace*}
18477 ---------- Buffer: *Backtrace* ----------
18478 Debugger entered--Lisp error: (void-function 1=)
18480 (setq number (1= number))
18481 (while (> number 0) (setq total (+ total number))
18482 (setq number (1= number)))
18483 (let ((total 0)) (while (> number 0) (setq total ...)
18484 (setq number ...)) total)
18488 eval((triangle-bugged 4))
18489 eval-last-sexp-1(nil)
18490 eval-last-sexp(nil)
18491 call-interactively(eval-last-sexp)
18492 ---------- Buffer: *Backtrace* ----------
18497 (I have reformatted this example slightly; the debugger does not fold
18498 long lines. As usual, you can quit the debugger by typing @kbd{q} in
18499 the @file{*Backtrace*} buffer.)
18501 In practice, for a bug as simple as this, the `Lisp error' line will
18502 tell you what you need to know to correct the definition. The
18503 function @code{1=} is `void'.
18507 In GNU Emacs 20 and before, you will see:
18510 Symbol's function definition is void:@: 1=
18514 which has the same meaning as the @file{*Backtrace*} buffer line in
18518 However, suppose you are not quite certain what is going on?
18519 You can read the complete backtrace.
18521 In this case, you need to run a recent GNU Emacs, which automatically
18522 starts the debugger that puts you in the @file{*Backtrace*} buffer; or
18523 else, you need to start the debugger manually as described below.
18525 Read the @file{*Backtrace*} buffer from the bottom up; it tells you
18526 what Emacs did that led to the error. Emacs made an interactive call
18527 to @kbd{C-x C-e} (@code{eval-last-sexp}), which led to the evaluation
18528 of the @code{triangle-bugged} expression. Each line above tells you
18529 what the Lisp interpreter evaluated next.
18532 The third line from the top of the buffer is
18535 (setq number (1= number))
18539 Emacs tried to evaluate this expression; in order to do so, it tried
18540 to evaluate the inner expression shown on the second line from the
18549 This is where the error occurred; as the top line says:
18552 Debugger entered--Lisp error: (void-function 1=)
18556 You can correct the mistake, re-evaluate the function definition, and
18557 then run your test again.
18559 @node debug-on-entry, debug-on-quit, debug, Debugging
18560 @section @code{debug-on-entry}
18561 @findex debug-on-entry
18563 A recent GNU Emacs starts the debugger automatically when your
18564 function has an error.
18567 GNU Emacs version 20 and before did not; it simply
18568 presented you with an error message. You had to start the debugger
18572 Incidentally, you can start the debugger manually for all versions of
18573 Emacs; the advantage is that the debugger runs even if you do not have
18574 a bug in your code. Sometimes your code will be free of bugs!
18576 You can enter the debugger when you call the function by calling
18577 @code{debug-on-entry}.
18584 M-x debug-on-entry RET triangle-bugged RET
18589 Now, evaluate the following:
18592 (triangle-bugged 5)
18596 All versions of Emacs will create a @file{*Backtrace*} buffer and tell
18597 you that it is beginning to evaluate the @code{triangle-bugged}
18602 ---------- Buffer: *Backtrace* ----------
18603 Debugger entered--entering a function:
18604 * triangle-bugged(5)
18605 eval((triangle-bugged 5))
18608 eval-last-sexp-1(nil)
18609 eval-last-sexp(nil)
18610 call-interactively(eval-last-sexp)
18611 ---------- Buffer: *Backtrace* ----------
18615 In the @file{*Backtrace*} buffer, type @kbd{d}. Emacs will evaluate
18616 the first expression in @code{triangle-bugged}; the buffer will look
18621 ---------- Buffer: *Backtrace* ----------
18622 Debugger entered--beginning evaluation of function call form:
18623 * (let ((total 0)) (while (> number 0) (setq total ...)
18624 (setq number ...)) total)
18625 * triangle-bugged(5)
18626 eval((triangle-bugged 5))
18629 eval-last-sexp-1(nil)
18630 eval-last-sexp(nil)
18631 call-interactively(eval-last-sexp)
18632 ---------- Buffer: *Backtrace* ----------
18637 Now, type @kbd{d} again, eight times, slowly. Each time you type
18638 @kbd{d}, Emacs will evaluate another expression in the function
18642 Eventually, the buffer will look like this:
18646 ---------- Buffer: *Backtrace* ----------
18647 Debugger entered--beginning evaluation of function call form:
18648 * (setq number (1= number))
18649 * (while (> number 0) (setq total (+ total number))
18650 (setq number (1= number)))
18653 * (let ((total 0)) (while (> number 0) (setq total ...)
18654 (setq number ...)) total)
18655 * triangle-bugged(5)
18656 eval((triangle-bugged 5))
18659 eval-last-sexp-1(nil)
18660 eval-last-sexp(nil)
18661 call-interactively(eval-last-sexp)
18662 ---------- Buffer: *Backtrace* ----------
18668 Finally, after you type @kbd{d} two more times, Emacs will reach the
18669 error, and the top two lines of the @file{*Backtrace*} buffer will look
18674 ---------- Buffer: *Backtrace* ----------
18675 Debugger entered--Lisp error: (void-function 1=)
18678 ---------- Buffer: *Backtrace* ----------
18682 By typing @kbd{d}, you were able to step through the function.
18684 You can quit a @file{*Backtrace*} buffer by typing @kbd{q} in it; this
18685 quits the trace, but does not cancel @code{debug-on-entry}.
18687 @findex cancel-debug-on-entry
18688 To cancel the effect of @code{debug-on-entry}, call
18689 @code{cancel-debug-on-entry} and the name of the function, like this:
18692 M-x cancel-debug-on-entry RET triangle-bugged RET
18696 (If you are reading this in Info, cancel @code{debug-on-entry} now.)
18698 @node debug-on-quit, edebug, debug-on-entry, Debugging
18699 @section @code{debug-on-quit} and @code{(debug)}
18701 In addition to setting @code{debug-on-error} or calling @code{debug-on-entry},
18702 there are two other ways to start @code{debug}.
18704 @findex debug-on-quit
18705 You can start @code{debug} whenever you type @kbd{C-g}
18706 (@code{keyboard-quit}) by setting the variable @code{debug-on-quit} to
18707 @code{t}. This is useful for debugging infinite loops.
18710 @cindex @code{(debug)} in code
18711 Or, you can insert a line that says @code{(debug)} into your code
18712 where you want the debugger to start, like this:
18716 (defun triangle-bugged (number)
18717 "Return sum of numbers 1 through NUMBER inclusive."
18719 (while (> number 0)
18720 (setq total (+ total number))
18721 (debug) ; @r{Start debugger.}
18722 (setq number (1= number))) ; @r{Error here.}
18727 The @code{debug} function is described in detail in @ref{Debugger, ,
18728 The Lisp Debugger, elisp, The GNU Emacs Lisp Reference Manual}.
18730 @node edebug, Debugging Exercises, debug-on-quit, Debugging
18731 @section The @code{edebug} Source Level Debugger
18732 @cindex Source level debugger
18735 Edebug is a source level debugger. Edebug normally displays the
18736 source of the code you are debugging, with an arrow at the left that
18737 shows which line you are currently executing.
18739 You can walk through the execution of a function, line by line, or run
18740 quickly until reaching a @dfn{breakpoint} where execution stops.
18742 Edebug is described in @ref{edebug, , Edebug, elisp, The GNU Emacs
18743 Lisp Reference Manual}.
18746 Here is a bugged function definition for @code{triangle-recursively}.
18747 @xref{Recursive triangle function, , Recursion in place of a counter},
18748 for a review of it.
18752 (defun triangle-recursively-bugged (number)
18753 "Return sum of numbers 1 through NUMBER inclusive.
18758 (triangle-recursively-bugged
18759 (1= number))))) ; @r{Error here.}
18764 Normally, you would install this definition by positioning your cursor
18765 after the function's closing parenthesis and typing @kbd{C-x C-e}
18766 (@code{eval-last-sexp}) or else by positioning your cursor within the
18767 definition and typing @kbd{C-M-x} (@code{eval-defun}). (By default,
18768 the @code{eval-defun} command works only in Emacs Lisp mode or in Lisp
18772 However, to prepare this function definition for Edebug, you must
18773 first @dfn{instrument} the code using a different command. You can do
18774 this by positioning your cursor within or just after the definition
18778 M-x edebug-defun RET
18782 This will cause Emacs to load Edebug automatically if it is not
18783 already loaded, and properly instrument the function.
18785 After instrumenting the function, place your cursor after the
18786 following expression and type @kbd{C-x C-e} (@code{eval-last-sexp}):
18789 (triangle-recursively-bugged 3)
18793 You will be jumped back to the source for
18794 @code{triangle-recursively-bugged} and the cursor positioned at the
18795 beginning of the @code{if} line of the function. Also, you will see
18796 an arrowhead at the left hand side of that line. The arrowhead marks
18797 the line where the function is executing. (In the following examples,
18798 we show the arrowhead with @samp{=>}; in a windowing system, you may
18799 see the arrowhead as a solid triangle in the window `fringe'.)
18802 =>@point{}(if (= number 1)
18807 In the example, the location of point is displayed with a star,
18808 @samp{@point{}} (in Info, it is displayed as @samp{-!-}).
18811 In the example, the location of point is displayed as @samp{@point{}}
18812 (in a printed book, it is displayed with a five pointed star).
18815 If you now press @key{SPC}, point will move to the next expression to
18816 be executed; the line will look like this:
18819 =>(if @point{}(= number 1)
18823 As you continue to press @key{SPC}, point will move from expression to
18824 expression. At the same time, whenever an expression returns a value,
18825 that value will be displayed in the echo area. For example, after you
18826 move point past @code{number}, you will see the following:
18829 Result: 3 (#o3, #x3, ?\C-c)
18833 This means the value of @code{number} is 3, which is octal three,
18834 hexadecimal three, and @sc{ascii} `control-c' (the third letter of the
18835 alphabet, in case you need to know this information).
18837 You can continue moving through the code until you reach the line with
18838 the error. Before evaluation, that line looks like this:
18841 => @point{}(1= number))))) ; @r{Error here.}
18846 When you press @key{SPC} once again, you will produce an error message
18850 Symbol's function definition is void:@: 1=
18856 Press @kbd{q} to quit Edebug.
18858 To remove instrumentation from a function definition, simply
18859 re-evaluate it with a command that does not instrument it.
18860 For example, you could place your cursor after the definition's
18861 closing parenthesis and type @kbd{C-x C-e}.
18863 Edebug does a great deal more than walk with you through a function.
18864 You can set it so it races through on its own, stopping only at an
18865 error or at specified stopping points; you can cause it to display the
18866 changing values of various expressions; you can find out how many
18867 times a function is called, and more.
18869 Edebug is described in @ref{edebug, , Edebug, elisp, The GNU Emacs
18870 Lisp Reference Manual}.
18873 @node Debugging Exercises, , edebug, Debugging
18874 @section Debugging Exercises
18878 Install the @code{count-words-region} function and then cause it to
18879 enter the built-in debugger when you call it. Run the command on a
18880 region containing two words. You will need to press @kbd{d} a
18881 remarkable number of times. On your system, is a `hook' called after
18882 the command finishes? (For information on hooks, see @ref{Command
18883 Overview, , Command Loop Overview, elisp, The GNU Emacs Lisp Reference
18887 Copy @code{count-words-region} into the @file{*scratch*} buffer,
18888 instrument the function for Edebug, and walk through its execution.
18889 The function does not need to have a bug, although you can introduce
18890 one if you wish. If the function lacks a bug, the walk-through
18891 completes without problems.
18894 While running Edebug, type @kbd{?} to see a list of all the Edebug commands.
18895 (The @code{global-edebug-prefix} is usually @kbd{C-x X}, i.e.@:
18896 @kbd{@key{CTRL}-x} followed by an upper case @kbd{X}; use this prefix
18897 for commands made outside of the Edebug debugging buffer.)
18900 In the Edebug debugging buffer, use the @kbd{p}
18901 (@code{edebug-bounce-point}) command to see where in the region the
18902 @code{count-words-region} is working.
18905 Move point to some spot further down the function and then type the
18906 @kbd{h} (@code{edebug-goto-here}) command to jump to that location.
18909 Use the @kbd{t} (@code{edebug-trace-mode}) command to cause Edebug to
18910 walk through the function on its own; use an upper case @kbd{T} for
18911 @code{edebug-Trace-fast-mode}.
18914 Set a breakpoint, then run Edebug in Trace mode until it reaches the
18918 @node Conclusion, the-the, Debugging, Top
18919 @chapter Conclusion
18921 We have now reached the end of this Introduction. You have now
18922 learned enough about programming in Emacs Lisp to set values, to write
18923 simple @file{.emacs} files for yourself and your friends, and write
18924 simple customizations and extensions to Emacs.
18926 This is a place to stop. Or, if you wish, you can now go onward, and
18929 You have learned some of the basic nuts and bolts of programming. But
18930 only some. There are a great many more brackets and hinges that are
18931 easy to use that we have not touched.
18933 A path you can follow right now lies among the sources to GNU Emacs
18936 @cite{The GNU Emacs Lisp Reference Manual}.
18939 @ref{Top, , The GNU Emacs Lisp Reference Manual, elisp, The GNU
18940 Emacs Lisp Reference Manual}.
18943 The Emacs Lisp sources are an adventure. When you read the sources and
18944 come across a function or expression that is unfamiliar, you need to
18945 figure out or find out what it does.
18947 Go to the Reference Manual. It is a thorough, complete, and fairly
18948 easy-to-read description of Emacs Lisp. It is written not only for
18949 experts, but for people who know what you know. (The @cite{Reference
18950 Manual} comes with the standard GNU Emacs distribution. Like this
18951 introduction, it comes as a Texinfo source file, so you can read it
18952 on-line and as a typeset, printed book.)
18954 Go to the other on-line help that is part of GNU Emacs: the on-line
18955 documentation for all functions and variables, and @code{find-tags},
18956 the program that takes you to sources.
18958 Here is an example of how I explore the sources. Because of its name,
18959 @file{simple.el} is the file I looked at first, a long time ago. As
18960 it happens some of the functions in @file{simple.el} are complicated,
18961 or at least look complicated at first sight. The @code{open-line}
18962 function, for example, looks complicated.
18964 You may want to walk through this function slowly, as we did with the
18965 @code{forward-sentence} function. (@xref{forward-sentence, The
18966 @code{forward-sentence} function}.) Or you may want to skip that
18967 function and look at another, such as @code{split-line}. You don't
18968 need to read all the functions. According to
18969 @code{count-words-in-defun}, the @code{split-line} function contains
18970 102 words and symbols.
18972 Even though it is short, @code{split-line} contains expressions
18973 we have not studied: @code{skip-chars-forward}, @code{indent-to},
18974 @code{current-column} and @code{insert-and-inherit}.
18976 Consider the @code{skip-chars-forward} function. (It is part of the
18977 function definition for @code{back-to-indentation}, which is shown in
18978 @ref{Review, , Review}.)
18980 In GNU Emacs, you can find out more about @code{skip-chars-forward} by
18981 typing @kbd{C-h f} (@code{describe-function}) and the name of the
18982 function. This gives you the function documentation.
18984 You may be able to guess what is done by a well named function such as
18985 @code{indent-to}; or you can look it up, too. Incidentally, the
18986 @code{describe-function} function itself is in @file{help.el}; it is
18987 one of those long, but decipherable functions. You can look up
18988 @code{describe-function} using the @kbd{C-h f} command!
18990 In this instance, since the code is Lisp, the @file{*Help*} buffer
18991 contains the name of the library containing the function's source.
18992 You can put point over the name of the library and press the RET key,
18993 which in this situation is bound to @code{help-follow}, and be taken
18994 directly to the source, in the same way as @kbd{M-.}
18997 The definition for @code{describe-function} illustrates how to
18998 customize the @code{interactive} expression without using the standard
18999 character codes; and it shows how to create a temporary buffer.
19001 (The @code{indent-to} function is written in C rather than Emacs Lisp;
19002 it is a `built-in' function. @code{help-follow} takes you to its
19003 source as does @code{find-tag}, when properly set up.)
19005 You can look at a function's source using @code{find-tag}, which is
19006 bound to @kbd{M-.} Finally, you can find out what the Reference
19007 Manual has to say by visiting the manual in Info, and typing @kbd{i}
19008 (@code{Info-index}) and the name of the function, or by looking up the
19009 function in the index to a printed copy of the manual.
19011 Similarly, you can find out what is meant by
19012 @code{insert-and-inherit}.
19014 Other interesting source files include @file{paragraphs.el},
19015 @file{loaddefs.el}, and @file{loadup.el}. The @file{paragraphs.el}
19016 file includes short, easily understood functions as well as longer
19017 ones. The @file{loaddefs.el} file contains the many standard
19018 autoloads and many keymaps. I have never looked at it all; only at
19019 parts. @file{loadup.el} is the file that loads the standard parts of
19020 Emacs; it tells you a great deal about how Emacs is built.
19021 (@xref{Building Emacs, , Building Emacs, elisp, The GNU Emacs Lisp
19022 Reference Manual}, for more about building.)
19024 As I said, you have learned some nuts and bolts; however, and very
19025 importantly, we have hardly touched major aspects of programming; I
19026 have said nothing about how to sort information, except to use the
19027 predefined @code{sort} function; I have said nothing about how to store
19028 information, except to use variables and lists; I have said nothing
19029 about how to write programs that write programs. These are topics for
19030 another, and different kind of book, a different kind of learning.
19032 What you have done is learn enough for much practical work with GNU
19033 Emacs. What you have done is get started. This is the end of a
19036 @c ================ Appendix ================
19038 @node the-the, Kill Ring, Conclusion, Top
19039 @appendix The @code{the-the} Function
19041 @cindex Duplicated words function
19042 @cindex Words, duplicated
19044 Sometimes when you you write text, you duplicate words---as with ``you
19045 you'' near the beginning of this sentence. I find that most
19046 frequently, I duplicate ``the''; hence, I call the function for
19047 detecting duplicated words, @code{the-the}.
19050 As a first step, you could use the following regular expression to
19051 search for duplicates:
19054 \\(\\w+[ \t\n]+\\)\\1
19058 This regexp matches one or more word-constituent characters followed
19059 by one or more spaces, tabs, or newlines. However, it does not detect
19060 duplicated words on different lines, since the ending of the first
19061 word, the end of the line, is different from the ending of the second
19062 word, a space. (For more information about regular expressions, see
19063 @ref{Regexp Search, , Regular Expression Searches}, as well as
19064 @ref{Regexps, , Syntax of Regular Expressions, emacs, The GNU Emacs
19065 Manual}, and @ref{Regular Expressions, , Regular Expressions, elisp,
19066 The GNU Emacs Lisp Reference Manual}.)
19068 You might try searching just for duplicated word-constituent
19069 characters but that does not work since the pattern detects doubles
19070 such as the two occurrences of `th' in `with the'.
19072 Another possible regexp searches for word-constituent characters
19073 followed by non-word-constituent characters, reduplicated. Here,
19074 @w{@samp{\\w+}} matches one or more word-constituent characters and
19075 @w{@samp{\\W*}} matches zero or more non-word-constituent characters.
19078 \\(\\(\\w+\\)\\W*\\)\\1
19084 Here is the pattern that I use. It is not perfect, but good enough.
19085 @w{@samp{\\b}} matches the empty string, provided it is at the beginning
19086 or end of a word; @w{@samp{[^@@ \n\t]+}} matches one or more occurrences of
19087 any characters that are @emph{not} an @@-sign, space, newline, or tab.
19090 \\b\\([^@@ \n\t]+\\)[ \n\t]+\\1\\b
19093 One can write more complicated expressions, but I found that this
19094 expression is good enough, so I use it.
19096 Here is the @code{the-the} function, as I include it in my
19097 @file{.emacs} file, along with a handy global key binding:
19102 "Search forward for for a duplicated word."
19104 (message "Searching for for duplicated words ...")
19108 ;; This regexp is not perfect
19109 ;; but is fairly good over all:
19110 (if (re-search-forward
19111 "\\b\\([^@@ \n\t]+\\)[ \n\t]+\\1\\b" nil 'move)
19112 (message "Found duplicated word.")
19113 (message "End of buffer")))
19117 ;; Bind `the-the' to C-c \
19118 (global-set-key "\C-c\\" 'the-the)
19127 one two two three four five
19132 You can substitute the other regular expressions shown above in the
19133 function definition and try each of them on this list.
19135 @node Kill Ring, Full Graph, the-the, Top
19136 @appendix Handling the Kill Ring
19137 @cindex Kill ring handling
19138 @cindex Handling the kill ring
19139 @cindex Ring, making a list like a
19141 The kill ring is a list that is transformed into a ring by the
19142 workings of the @code{current-kill} function. The @code{yank} and
19143 @code{yank-pop} commands use the @code{current-kill} function.
19145 This appendix describes the @code{current-kill} function as well as
19146 both the @code{yank} and the @code{yank-pop} commands, but first,
19147 consider the workings of the kill ring.
19150 * What the Kill Ring Does::
19157 @node What the Kill Ring Does, current-kill, Kill Ring, Kill Ring
19159 @unnumberedsec What the Kill Ring Does
19163 The kill ring has a default maximum length of sixty items; this number
19164 is too large for an explanation. Instead, set it to four. Please
19165 evaluate the following:
19169 (setq old-kill-ring-max kill-ring-max)
19170 (setq kill-ring-max 4)
19175 Then, please copy each line of the following indented example into the
19176 kill ring. You may kill each line with @kbd{C-k} or mark it and copy
19180 (In a read-only buffer, such as the @file{*info*} buffer, the kill
19181 command, @kbd{C-k} (@code{kill-line}), will not remove the text,
19182 merely copy it to the kill ring. However, your machine may beep at
19183 you. Alternatively, for silence, you may copy the region of each line
19184 with the @kbd{M-w} (@code{kill-ring-save}) command. You must mark
19185 each line for this command to succeed, but it does not matter at which
19186 end you put point or mark.)
19190 Please invoke the calls in order, so that five elements attempt to
19191 fill the kill ring:
19196 second piece of text
19198 fourth line of text
19205 Then find the value of @code{kill-ring} by evaluating
19217 ("fifth bit of text" "fourth line of text"
19218 "third line" "second piece of text")
19223 The first element, @samp{first some text}, was dropped.
19226 To return to the old value for the length of the kill ring, evaluate:
19229 (setq kill-ring-max old-kill-ring-max)
19232 @node current-kill, yank, What the Kill Ring Does, Kill Ring
19233 @comment node-name, next, previous, up
19234 @appendixsec The @code{current-kill} Function
19235 @findex current-kill
19237 The @code{current-kill} function changes the element in the kill ring
19238 to which @code{kill-ring-yank-pointer} points. (Also, the
19239 @code{kill-new} function sets @code{kill-ring-yank-pointer} to point
19240 to the latest element of the the kill ring. The @code{kill-new}
19241 function is used directly or indirectly by @code{kill-append},
19242 @code{copy-region-as-kill}, @code{kill-ring-save}, @code{kill-line},
19243 and @code{kill-region}.)
19246 * Code for current-kill::
19247 * Understanding current-kill::
19250 @node Code for current-kill, Understanding current-kill, current-kill, current-kill
19252 @unnumberedsubsec The code for @code{current-kill}
19257 The @code{current-kill} function is used by @code{yank} and by
19258 @code{yank-pop}. Here is the code for @code{current-kill}:
19262 (defun current-kill (n &optional do-not-move)
19263 "Rotate the yanking point by N places, and then return that kill.
19264 If N is zero, `interprogram-paste-function' is set, and calling it
19265 returns a string, then that string is added to the front of the
19266 kill ring and returned as the latest kill.
19269 If optional arg DO-NOT-MOVE is non-nil, then don't actually move the
19270 yanking point; just return the Nth kill forward."
19271 (let ((interprogram-paste (and (= n 0)
19272 interprogram-paste-function
19273 (funcall interprogram-paste-function))))
19276 (if interprogram-paste
19278 ;; Disable the interprogram cut function when we add the new
19279 ;; text to the kill ring, so Emacs doesn't try to own the
19280 ;; selection, with identical text.
19281 (let ((interprogram-cut-function nil))
19282 (kill-new interprogram-paste))
19283 interprogram-paste)
19286 (or kill-ring (error "Kill ring is empty"))
19287 (let ((ARGth-kill-element
19288 (nthcdr (mod (- n (length kill-ring-yank-pointer))
19289 (length kill-ring))
19292 (setq kill-ring-yank-pointer ARGth-kill-element))
19293 (car ARGth-kill-element)))))
19297 Remember also that the @code{kill-new} function sets
19298 @code{kill-ring-yank-pointer} to the latest element of the the kill
19299 ring, which means that all the functions that call it set the value
19300 indirectly: @code{kill-append}, @code{copy-region-as-kill},
19301 @code{kill-ring-save}, @code{kill-line}, and @code{kill-region}.
19304 Here is the line in @code{kill-new}, which is explained in
19305 @ref{kill-new function, , The @code{kill-new} function}.
19308 (setq kill-ring-yank-pointer kill-ring)
19311 @node Understanding current-kill, , Code for current-kill, current-kill
19313 @unnumberedsubsec @code{current-kill} in Outline
19316 The @code{current-kill} function looks complex, but as usual, it can
19317 be understood by taking it apart piece by piece. First look at it in
19322 (defun current-kill (n &optional do-not-move)
19323 "Rotate the yanking point by N places, and then return that kill."
19329 This function takes two arguments, one of which is optional. It has a
19330 documentation string. It is @emph{not} interactive.
19333 * Body of current-kill::
19334 * Digression concerning error::
19335 * Determining the Element::
19338 @node Body of current-kill, Digression concerning error, Understanding current-kill, Understanding current-kill
19340 @unnumberedsubsubsec The Body of @code{current-kill}
19343 The body of the function definition is a @code{let} expression, which
19344 itself has a body as well as a @var{varlist}.
19346 The @code{let} expression declares a variable that will be only usable
19347 within the bounds of this function. This variable is called
19348 @code{interprogram-paste} and is for copying to another program. It
19349 is not for copying within this instance of GNU Emacs. Most window
19350 systems provide a facility for interprogram pasting. Sadly, that
19351 facility usually provides only for the last element. Most windowing
19352 systems have not adopted a ring of many possibilities, even though
19353 Emacs has provided it for decades.
19355 The @code{if} expression has two parts, one if there exists
19356 @code{interprogram-paste} and one if not.
19359 Let us consider the `if not' or else-part of the @code{current-kill}
19360 function. (The then-part uses the the @code{kill-new} function, which
19361 we have already described. @xref{kill-new function, , The
19362 @code{kill-new} function}.)
19366 (or kill-ring (error "Kill ring is empty"))
19367 (let ((ARGth-kill-element
19368 (nthcdr (mod (- n (length kill-ring-yank-pointer))
19369 (length kill-ring))
19372 (setq kill-ring-yank-pointer ARGth-kill-element))
19373 (car ARGth-kill-element))
19378 The code first checks whether the kill ring has content; otherwise it
19382 Note that the @code{or} expression is very similar to testing length
19389 (if (zerop (length kill-ring)) ; @r{if-part}
19390 (error "Kill ring is empty")) ; @r{then-part}
19396 If there is not anything in the kill ring, its length must be zero and
19397 an error message sent to the user: @samp{Kill ring is empty}. The
19398 @code{current-kill} function uses an @code{or} expression which is
19399 simpler. But an @code{if} expression reminds us what goes on.
19401 This @code{if} expression uses the function @code{zerop} which returns
19402 true if the value it is testing is zero. When @code{zerop} tests
19403 true, the then-part of the @code{if} is evaluated. The then-part is a
19404 list starting with the function @code{error}, which is a function that
19405 is similar to the @code{message} function
19406 (@pxref{message, , The @code{message} Function}) in that
19407 it prints a one-line message in the echo area. However, in addition
19408 to printing a message, @code{error} also stops evaluation of the
19409 function within which it is embedded. This means that the rest of the
19410 function will not be evaluated if the length of the kill ring is zero.
19412 Then the @code{current-kill} function selects the element to return.
19413 The selection depends on the number of places that @code{current-kill}
19414 rotates and on where @code{kill-ring-yank-pointer} points.
19416 Next, either the optional @code{do-not-move} argument is true or the
19417 current value of @code{kill-ring-yank-pointer} is set to point to the
19418 list. Finally, another expression returns the first element of the
19419 list even if the @code{do-not-move} argument is true.
19421 @node Digression concerning error, Determining the Element, Body of current-kill, Understanding current-kill
19423 @unnumberedsubsubsec Digression about the word `error'
19426 In my opinion, it is slightly misleading, at least to humans, to use
19427 the term `error' as the name of the @code{error} function. A better
19428 term would be `cancel'. Strictly speaking, of course, you cannot
19429 point to, much less rotate a pointer to a list that has no length, so
19430 from the point of view of the computer, the word `error' is correct.
19431 But a human expects to attempt this sort of thing, if only to find out
19432 whether the kill ring is full or empty. This is an act of
19435 From the human point of view, the act of exploration and discovery is
19436 not necessarily an error, and therefore should not be labelled as one,
19437 even in the bowels of a computer. As it is, the code in Emacs implies
19438 that a human who is acting virtuously, by exploring his or her
19439 environment, is making an error. This is bad. Even though the computer
19440 takes the same steps as it does when there is an `error', a term such as
19441 `cancel' would have a clearer connotation.
19443 @node Determining the Element, , Digression concerning error, Understanding current-kill
19445 @unnumberedsubsubsec Determining the Element
19448 Among other actions, the else-part of the @code{if} expression sets
19449 the value of @code{kill-ring-yank-pointer} to
19450 @code{ARGth-kill-element} when the kill ring has something in it and
19451 the value of @code{do-not-move} is @code{nil}.
19454 The code looks like this:
19458 (nthcdr (mod (- n (length kill-ring-yank-pointer))
19459 (length kill-ring))
19464 This needs some examination. Unless it is not supposed to move the
19465 pointer, the @code{current-kill} function changes where
19466 @code{kill-ring-yank-pointer} points.
19468 @w{@code{(setq kill-ring-yank-pointer ARGth-kill-element))}}
19469 expression does. Also, clearly, @code{ARGth-kill-element} is being
19470 set to be equal to some @sc{cdr} of the kill ring, using the
19471 @code{nthcdr} function that is described in an earlier section.
19472 (@xref{copy-region-as-kill}.) How does it do this?
19474 As we have seen before (@pxref{nthcdr}), the @code{nthcdr} function
19475 works by repeatedly taking the @sc{cdr} of a list---it takes the
19476 @sc{cdr} of the @sc{cdr} of the @sc{cdr} @dots{}
19479 The two following expressions produce the same result:
19483 (setq kill-ring-yank-pointer (cdr kill-ring))
19485 (setq kill-ring-yank-pointer (nthcdr 1 kill-ring))
19489 However, the @code{nthcdr} expression is more complicated. It uses
19490 the @code{mod} function to determine which @sc{cdr} to select.
19492 (You will remember to look at inner functions first; indeed, we will
19493 have to go inside the @code{mod}.)
19495 The @code{mod} function returns the value of its first argument modulo
19496 the second; that is to say, it returns the remainder after dividing
19497 the first argument by the second. The value returned has the same
19498 sign as the second argument.
19506 @result{} 0 ;; @r{because there is no remainder}
19513 In this case, the first argument is often smaller than the second.
19525 We can guess what the @code{-} function does. It is like @code{+} but
19526 subtracts instead of adds; the @code{-} function subtracts its second
19527 argument from its first. Also, we already know what the @code{length}
19528 function does (@pxref{length}). It returns the length of a list.
19530 And @code{n} is the name of the required argument to the
19531 @code{current-kill} function.
19534 So when the first argument to @code{nthcdr} is zero, the @code{nthcdr}
19535 expression returns the whole list, as you can see by evaluating the
19540 ;; kill-ring-yank-pointer @r{and} kill-ring @r{have a length of four}
19541 ;; @r{and} (mod (- 0 4) 4) @result{} 0
19542 (nthcdr (mod (- 0 4) 4)
19543 '("fourth line of text"
19545 "second piece of text"
19546 "first some text"))
19551 When the first argument to the @code{current-kill} function is one,
19552 the @code{nthcdr} expression returns the list without its first
19557 (nthcdr (mod (- 1 4) 4)
19558 '("fourth line of text"
19560 "second piece of text"
19561 "first some text"))
19565 @cindex @samp{global variable} defined
19566 @cindex @samp{variable, global}, defined
19567 Incidentally, both @code{kill-ring} and @code{kill-ring-yank-pointer}
19568 are @dfn{global variables}. That means that any expression in Emacs
19569 Lisp can access them. They are not like the local variables set by
19570 @code{let} or like the symbols in an argument list.
19571 Local variables can only be accessed
19572 within the @code{let} that defines them or the function that specifies
19573 them in an argument list (and within expressions called by them).
19576 @c texi2dvi fails when the name of the section is within ifnottex ...
19577 (@xref{Prevent confusion, , @code{let} Prevents Confusion}, and
19578 @ref{defun, , The @code{defun} Special Form}.)
19581 @node yank, yank-pop, current-kill, Kill Ring
19582 @comment node-name, next, previous, up
19583 @appendixsec @code{yank}
19586 After learning about @code{current-kill}, the code for the
19587 @code{yank} function is almost easy.
19589 The @code{yank} function does not use the
19590 @code{kill-ring-yank-pointer} variable directly. It calls
19591 @code{insert-for-yank} which calls @code{current-kill} which sets the
19592 @code{kill-ring-yank-pointer} variable.
19595 The code looks like this:
19600 (defun yank (&optional arg)
19601 "Reinsert (\"paste\") the last stretch of killed text.
19602 More precisely, reinsert the stretch of killed text most recently
19603 killed OR yanked. Put point at end, and set mark at beginning.
19604 With just \\[universal-argument] as argument, same but put point at
19605 beginning (and mark at end). With argument N, reinsert the Nth most
19606 recently killed stretch of killed text.
19608 When this command inserts killed text into the buffer, it honors
19609 `yank-excluded-properties' and `yank-handler' as described in the
19610 doc string for `insert-for-yank-1', which see.
19612 See also the command \\[yank-pop]."
19616 (setq yank-window-start (window-start))
19617 ;; If we don't get all the way thru, make last-command indicate that
19618 ;; for the following command.
19619 (setq this-command t)
19620 (push-mark (point))
19623 (insert-for-yank (current-kill (cond
19628 ;; This is like exchange-point-and-mark,
19629 ;; but doesn't activate the mark.
19630 ;; It is cleaner to avoid activation, even though the command
19631 ;; loop would deactivate the mark because we inserted text.
19632 (goto-char (prog1 (mark t)
19633 (set-marker (mark-marker) (point) (current-buffer)))))
19636 ;; If we do get all the way thru, make this-command indicate that.
19637 (if (eq this-command t)
19638 (setq this-command 'yank))
19643 The key expression is @code{insert-for-yank}, which inserts the string
19644 returned by @code{current-kill}, but removes some text properties from
19647 However, before getting to that expression, the function sets the value
19648 of @code{yank-window-start} to the position returned by the
19649 @code{(window-start)} expression, the position at which the display
19650 currently starts. The @code{yank} function also sets
19651 @code{this-command} and pushes the mark.
19653 After it yanks the appropriate element, if the optional argument is a
19654 @sc{cons} rather than a number or nothing, it puts point at beginning
19655 of the yanked text and mark at its end.
19657 (The @code{prog1} function is like @code{progn} but returns the value
19658 of its first argument rather than the value of its last argument. Its
19659 first argument is forced to return the buffer's mark as an integer.
19660 You can see the documentation for these functions by placing point
19661 over them in this buffer and then typing @kbd{C-h f}
19662 (@code{describe-function}) followed by a @kbd{RET}; the default is the
19665 The last part of the function tells what to do when it succeeds.
19667 @node yank-pop, ring file, yank, Kill Ring
19668 @comment node-name, next, previous, up
19669 @appendixsec @code{yank-pop}
19672 After understanding @code{yank} and @code{current-kill}, you know how
19673 to approach the @code{yank-pop} function. Leaving out the
19674 documentation to save space, it looks like this:
19679 (defun yank-pop (&optional arg)
19682 (if (not (eq last-command 'yank))
19683 (error "Previous command was not a yank"))
19686 (setq this-command 'yank)
19687 (unless arg (setq arg 1))
19688 (let ((inhibit-read-only t)
19689 (before (< (point) (mark t))))
19693 (funcall (or yank-undo-function 'delete-region) (point) (mark t))
19694 (funcall (or yank-undo-function 'delete-region) (mark t) (point)))
19695 (setq yank-undo-function nil)
19698 (set-marker (mark-marker) (point) (current-buffer))
19699 (insert-for-yank (current-kill arg))
19700 ;; Set the window start back where it was in the yank command,
19702 (set-window-start (selected-window) yank-window-start t)
19706 ;; This is like exchange-point-and-mark,
19707 ;; but doesn't activate the mark.
19708 ;; It is cleaner to avoid activation, even though the command
19709 ;; loop would deactivate the mark because we inserted text.
19710 (goto-char (prog1 (mark t)
19711 (set-marker (mark-marker)
19713 (current-buffer))))))
19718 The function is interactive with a small @samp{p} so the prefix
19719 argument is processed and passed to the function. The command can
19720 only be used after a previous yank; otherwise an error message is
19721 sent. This check uses the variable @code{last-command} which is set
19722 by @code{yank} and is discussed elsewhere.
19723 (@xref{copy-region-as-kill}.)
19725 The @code{let} clause sets the variable @code{before} to true or false
19726 depending whether point is before or after mark and then the region
19727 between point and mark is deleted. This is the region that was just
19728 inserted by the previous yank and it is this text that will be
19731 @code{funcall} calls its first argument as a function, passing
19732 remaining arguments to it. The first argument is whatever the
19733 @code{or} expression returns. The two remaining arguments are the
19734 positions of point and mark set by the preceding @code{yank} command.
19736 There is more, but that is the hardest part.
19738 @node ring file, , yank-pop, Kill Ring
19739 @comment node-name, next, previous, up
19740 @appendixsec The @file{ring.el} File
19741 @cindex @file{ring.el} file
19743 Interestingly, GNU Emacs posses a file called @file{ring.el} that
19744 provides many of the features we just discussed. But functions such
19745 as @code{kill-ring-yank-pointer} do not use this library, possibly
19746 because they were written earlier.
19748 @node Full Graph, Free Software and Free Manuals, Kill Ring, Top
19749 @appendix A Graph with Labelled Axes
19751 Printed axes help you understand a graph. They convey scale. In an
19752 earlier chapter (@pxref{Readying a Graph, , Readying a Graph}), we
19753 wrote the code to print the body of a graph. Here we write the code
19754 for printing and labelling vertical and horizontal axes, along with the
19758 * Labelled Example::
19759 * print-graph Varlist::
19762 * Print Whole Graph::
19765 @node Labelled Example, print-graph Varlist, Full Graph, Full Graph
19767 @unnumberedsec Labelled Example Graph
19770 Since insertions fill a buffer to the right and below point, the new
19771 graph printing function should first print the Y or vertical axis,
19772 then the body of the graph, and finally the X or horizontal axis.
19773 This sequence lays out for us the contents of the function:
19783 Print body of graph.
19790 Here is an example of how a finished graph should look:
19803 1 - ****************
19810 In this graph, both the vertical and the horizontal axes are labelled
19811 with numbers. However, in some graphs, the horizontal axis is time
19812 and would be better labelled with months, like this:
19826 Indeed, with a little thought, we can easily come up with a variety of
19827 vertical and horizontal labelling schemes. Our task could become
19828 complicated. But complications breed confusion. Rather than permit
19829 this, it is better choose a simple labelling scheme for our first
19830 effort, and to modify or replace it later.
19833 These considerations suggest the following outline for the
19834 @code{print-graph} function:
19838 (defun print-graph (numbers-list)
19839 "@var{documentation}@dots{}"
19840 (let ((height @dots{}
19844 (print-Y-axis height @dots{} )
19845 (graph-body-print numbers-list)
19846 (print-X-axis @dots{} )))
19850 We can work on each part of the @code{print-graph} function definition
19853 @node print-graph Varlist, print-Y-axis, Labelled Example, Full Graph
19854 @comment node-name, next, previous, up
19855 @appendixsec The @code{print-graph} Varlist
19856 @cindex @code{print-graph} varlist
19858 In writing the @code{print-graph} function, the first task is to write
19859 the varlist in the @code{let} expression. (We will leave aside for the
19860 moment any thoughts about making the function interactive or about the
19861 contents of its documentation string.)
19863 The varlist should set several values. Clearly, the top of the label
19864 for the vertical axis must be at least the height of the graph, which
19865 means that we must obtain this information here. Note that the
19866 @code{print-graph-body} function also requires this information. There
19867 is no reason to calculate the height of the graph in two different
19868 places, so we should change @code{print-graph-body} from the way we
19869 defined it earlier to take advantage of the calculation.
19871 Similarly, both the function for printing the X axis labels and the
19872 @code{print-graph-body} function need to learn the value of the width of
19873 each symbol. We can perform the calculation here and change the
19874 definition for @code{print-graph-body} from the way we defined it in the
19877 The length of the label for the horizontal axis must be at least as long
19878 as the graph. However, this information is used only in the function
19879 that prints the horizontal axis, so it does not need to be calculated here.
19881 These thoughts lead us directly to the following form for the varlist
19882 in the @code{let} for @code{print-graph}:
19886 (let ((height (apply 'max numbers-list)) ; @r{First version.}
19887 (symbol-width (length graph-blank)))
19892 As we shall see, this expression is not quite right.
19895 @node print-Y-axis, print-X-axis, print-graph Varlist, Full Graph
19896 @comment node-name, next, previous, up
19897 @appendixsec The @code{print-Y-axis} Function
19898 @cindex Axis, print vertical
19899 @cindex Y axis printing
19900 @cindex Vertical axis printing
19901 @cindex Print vertical axis
19903 The job of the @code{print-Y-axis} function is to print a label for
19904 the vertical axis that looks like this:
19922 The function should be passed the height of the graph, and then should
19923 construct and insert the appropriate numbers and marks.
19926 * print-Y-axis in Detail::
19927 * Height of label::
19928 * Compute a Remainder::
19931 * print-Y-axis Penultimate::
19934 @node print-Y-axis in Detail, Height of label, print-Y-axis, print-Y-axis
19936 @unnumberedsubsec The @code{print-Y-axis} Function in Detail
19939 It is easy enough to see in the figure what the Y axis label should
19940 look like; but to say in words, and then to write a function
19941 definition to do the job is another matter. It is not quite true to
19942 say that we want a number and a tic every five lines: there are only
19943 three lines between the @samp{1} and the @samp{5} (lines 2, 3, and 4),
19944 but four lines between the @samp{5} and the @samp{10} (lines 6, 7, 8,
19945 and 9). It is better to say that we want a number and a tic mark on
19946 the base line (number 1) and then that we want a number and a tic on
19947 the fifth line from the bottom and on every line that is a multiple of
19950 @node Height of label, Compute a Remainder, print-Y-axis in Detail, print-Y-axis
19952 @unnumberedsubsec What height should the label be?
19955 The next issue is what height the label should be? Suppose the maximum
19956 height of tallest column of the graph is seven. Should the highest
19957 label on the Y axis be @samp{5 -}, and should the graph stick up above
19958 the label? Or should the highest label be @samp{7 -}, and mark the peak
19959 of the graph? Or should the highest label be @code{10 -}, which is a
19960 multiple of five, and be higher than the topmost value of the graph?
19962 The latter form is preferred. Most graphs are drawn within rectangles
19963 whose sides are an integral number of steps long---5, 10, 15, and so
19964 on for a step distance of five. But as soon as we decide to use a
19965 step height for the vertical axis, we discover that the simple
19966 expression in the varlist for computing the height is wrong. The
19967 expression is @code{(apply 'max numbers-list)}. This returns the
19968 precise height, not the maximum height plus whatever is necessary to
19969 round up to the nearest multiple of five. A more complex expression
19972 As usual in cases like this, a complex problem becomes simpler if it is
19973 divided into several smaller problems.
19975 First, consider the case when the highest value of the graph is an
19976 integral multiple of five---when it is 5, 10, 15, or some higher
19977 multiple of five. We can use this value as the Y axis height.
19979 A fairly simply way to determine whether a number is a multiple of
19980 five is to divide it by five and see if the division results in a
19981 remainder. If there is no remainder, the number is a multiple of
19982 five. Thus, seven divided by five has a remainder of two, and seven
19983 is not an integral multiple of five. Put in slightly different
19984 language, more reminiscent of the classroom, five goes into seven
19985 once, with a remainder of two. However, five goes into ten twice,
19986 with no remainder: ten is an integral multiple of five.
19988 @node Compute a Remainder, Y Axis Element, Height of label, print-Y-axis
19989 @appendixsubsec Side Trip: Compute a Remainder
19991 @findex % @r{(remainder function)}
19992 @cindex Remainder function, @code{%}
19993 In Lisp, the function for computing a remainder is @code{%}. The
19994 function returns the remainder of its first argument divided by its
19995 second argument. As it happens, @code{%} is a function in Emacs Lisp
19996 that you cannot discover using @code{apropos}: you find nothing if you
19997 type @kbd{M-x apropos @key{RET} remainder @key{RET}}. The only way to
19998 learn of the existence of @code{%} is to read about it in a book such
19999 as this or in the Emacs Lisp sources.
20001 You can try the @code{%} function by evaluating the following two
20013 The first expression returns 2 and the second expression returns 0.
20015 To test whether the returned value is zero or some other number, we
20016 can use the @code{zerop} function. This function returns @code{t} if
20017 its argument, which must be a number, is zero.
20029 Thus, the following expression will return @code{t} if the height
20030 of the graph is evenly divisible by five:
20033 (zerop (% height 5))
20037 (The value of @code{height}, of course, can be found from @code{(apply
20038 'max numbers-list)}.)
20040 On the other hand, if the value of @code{height} is not a multiple of
20041 five, we want to reset the value to the next higher multiple of five.
20042 This is straightforward arithmetic using functions with which we are
20043 already familiar. First, we divide the value of @code{height} by five
20044 to determine how many times five goes into the number. Thus, five
20045 goes into twelve twice. If we add one to this quotient and multiply by
20046 five, we will obtain the value of the next multiple of five that is
20047 larger than the height. Five goes into twelve twice. Add one to two,
20048 and multiply by five; the result is fifteen, which is the next multiple
20049 of five that is higher than twelve. The Lisp expression for this is:
20052 (* (1+ (/ height 5)) 5)
20056 For example, if you evaluate the following, the result is 15:
20059 (* (1+ (/ 12 5)) 5)
20062 All through this discussion, we have been using `five' as the value
20063 for spacing labels on the Y axis; but we may want to use some other
20064 value. For generality, we should replace `five' with a variable to
20065 which we can assign a value. The best name I can think of for this
20066 variable is @code{Y-axis-label-spacing}.
20069 Using this term, and an @code{if} expression, we produce the
20074 (if (zerop (% height Y-axis-label-spacing))
20077 (* (1+ (/ height Y-axis-label-spacing))
20078 Y-axis-label-spacing))
20083 This expression returns the value of @code{height} itself if the height
20084 is an even multiple of the value of the @code{Y-axis-label-spacing} or
20085 else it computes and returns a value of @code{height} that is equal to
20086 the next higher multiple of the value of the @code{Y-axis-label-spacing}.
20088 We can now include this expression in the @code{let} expression of the
20089 @code{print-graph} function (after first setting the value of
20090 @code{Y-axis-label-spacing}):
20091 @vindex Y-axis-label-spacing
20095 (defvar Y-axis-label-spacing 5
20096 "Number of lines from one Y axis label to next.")
20101 (let* ((height (apply 'max numbers-list))
20102 (height-of-top-line
20103 (if (zerop (% height Y-axis-label-spacing))
20108 (* (1+ (/ height Y-axis-label-spacing))
20109 Y-axis-label-spacing)))
20110 (symbol-width (length graph-blank))))
20116 (Note use of the @code{let*} function: the initial value of height is
20117 computed once by the @code{(apply 'max numbers-list)} expression and
20118 then the resulting value of @code{height} is used to compute its
20119 final value. @xref{fwd-para let, , The @code{let*} expression}, for
20120 more about @code{let*}.)
20122 @node Y Axis Element, Y-axis-column, Compute a Remainder, print-Y-axis
20123 @appendixsubsec Construct a Y Axis Element
20125 When we print the vertical axis, we want to insert strings such as
20126 @w{@samp{5 -}} and @w{@samp{10 - }} every five lines.
20127 Moreover, we want the numbers and dashes to line up, so shorter
20128 numbers must be padded with leading spaces. If some of the strings
20129 use two digit numbers, the strings with single digit numbers must
20130 include a leading blank space before the number.
20132 @findex number-to-string
20133 To figure out the length of the number, the @code{length} function is
20134 used. But the @code{length} function works only with a string, not with
20135 a number. So the number has to be converted from being a number to
20136 being a string. This is done with the @code{number-to-string} function.
20141 (length (number-to-string 35))
20144 (length (number-to-string 100))
20150 (@code{number-to-string} is also called @code{int-to-string}; you will
20151 see this alternative name in various sources.)
20153 In addition, in each label, each number is followed by a string such
20154 as @w{@samp{ - }}, which we will call the @code{Y-axis-tic} marker.
20155 This variable is defined with @code{defvar}:
20160 (defvar Y-axis-tic " - "
20161 "String that follows number in a Y axis label.")
20165 The length of the Y label is the sum of the length of the Y axis tic
20166 mark and the length of the number of the top of the graph.
20169 (length (concat (number-to-string height) Y-axis-tic)))
20172 This value will be calculated by the @code{print-graph} function in
20173 its varlist as @code{full-Y-label-width} and passed on. (Note that we
20174 did not think to include this in the varlist when we first proposed it.)
20176 To make a complete vertical axis label, a tic mark is concatenated
20177 with a number; and the two together may be preceded by one or more
20178 spaces depending on how long the number is. The label consists of
20179 three parts: the (optional) leading spaces, the number, and the tic
20180 mark. The function is passed the value of the number for the specific
20181 row, and the value of the width of the top line, which is calculated
20182 (just once) by @code{print-graph}.
20186 (defun Y-axis-element (number full-Y-label-width)
20187 "Construct a NUMBERed label element.
20188 A numbered element looks like this ` 5 - ',
20189 and is padded as needed so all line up with
20190 the element for the largest number."
20193 (let* ((leading-spaces
20194 (- full-Y-label-width
20196 (concat (number-to-string number)
20201 (make-string leading-spaces ? )
20202 (number-to-string number)
20207 The @code{Y-axis-element} function concatenates together the leading
20208 spaces, if any; the number, as a string; and the tic mark.
20210 To figure out how many leading spaces the label will need, the
20211 function subtracts the actual length of the label---the length of the
20212 number plus the length of the tic mark---from the desired label width.
20214 @findex make-string
20215 Blank spaces are inserted using the @code{make-string} function. This
20216 function takes two arguments: the first tells it how long the string
20217 will be and the second is a symbol for the character to insert, in a
20218 special format. The format is a question mark followed by a blank
20219 space, like this, @samp{? }. @xref{Character Type, , Character Type,
20220 elisp, The GNU Emacs Lisp Reference Manual}, for a description of the
20221 syntax for characters. (Of course, you might want to replace the
20222 blank space by some other character @dots{} You know what to do.)
20224 The @code{number-to-string} function is used in the concatenation
20225 expression, to convert the number to a string that is concatenated
20226 with the leading spaces and the tic mark.
20228 @node Y-axis-column, print-Y-axis Penultimate, Y Axis Element, print-Y-axis
20229 @appendixsubsec Create a Y Axis Column
20231 The preceding functions provide all the tools needed to construct a
20232 function that generates a list of numbered and blank strings to insert
20233 as the label for the vertical axis:
20235 @findex Y-axis-column
20238 (defun Y-axis-column (height width-of-label)
20239 "Construct list of Y axis labels and blank strings.
20240 For HEIGHT of line above base and WIDTH-OF-LABEL."
20244 (while (> height 1)
20245 (if (zerop (% height Y-axis-label-spacing))
20246 ;; @r{Insert label.}
20249 (Y-axis-element height width-of-label)
20253 ;; @r{Else, insert blanks.}
20256 (make-string width-of-label ? )
20258 (setq height (1- height)))
20259 ;; @r{Insert base line.}
20261 (cons (Y-axis-element 1 width-of-label) Y-axis))
20262 (nreverse Y-axis)))
20266 In this function, we start with the value of @code{height} and
20267 repetitively subtract one from its value. After each subtraction, we
20268 test to see whether the value is an integral multiple of the
20269 @code{Y-axis-label-spacing}. If it is, we construct a numbered label
20270 using the @code{Y-axis-element} function; if not, we construct a
20271 blank label using the @code{make-string} function. The base line
20272 consists of the number one followed by a tic mark.
20275 @node print-Y-axis Penultimate, , Y-axis-column, print-Y-axis
20276 @appendixsubsec The Not Quite Final Version of @code{print-Y-axis}
20278 The list constructed by the @code{Y-axis-column} function is passed to
20279 the @code{print-Y-axis} function, which inserts the list as a column.
20281 @findex print-Y-axis
20284 (defun print-Y-axis (height full-Y-label-width)
20285 "Insert Y axis using HEIGHT and FULL-Y-LABEL-WIDTH.
20286 Height must be the maximum height of the graph.
20287 Full width is the width of the highest label element."
20288 ;; Value of height and full-Y-label-width
20289 ;; are passed by `print-graph'.
20292 (let ((start (point)))
20294 (Y-axis-column height full-Y-label-width))
20295 ;; @r{Place point ready for inserting graph.}
20297 ;; @r{Move point forward by value of} full-Y-label-width
20298 (forward-char full-Y-label-width)))
20302 The @code{print-Y-axis} uses the @code{insert-rectangle} function to
20303 insert the Y axis labels created by the @code{Y-axis-column} function.
20304 In addition, it places point at the correct position for printing the body of
20307 You can test @code{print-Y-axis}:
20315 Y-axis-label-spacing
20324 Copy the following expression:
20327 (print-Y-axis 12 5)
20331 Switch to the @file{*scratch*} buffer and place the cursor where you
20332 want the axis labels to start.
20335 Type @kbd{M-:} (@code{eval-expression}).
20338 Yank the @code{graph-body-print} expression into the minibuffer
20339 with @kbd{C-y} (@code{yank)}.
20342 Press @key{RET} to evaluate the expression.
20345 Emacs will print labels vertically, the top one being @w{@samp{10 -@w{
20346 }}}. (The @code{print-graph} function will pass the value of
20347 @code{height-of-top-line}, which in this case will end up as 15,
20348 thereby getting rid of what might appear as a bug.)
20351 @node print-X-axis, Print Whole Graph, print-Y-axis, Full Graph
20352 @appendixsec The @code{print-X-axis} Function
20353 @cindex Axis, print horizontal
20354 @cindex X axis printing
20355 @cindex Print horizontal axis
20356 @cindex Horizontal axis printing
20358 X axis labels are much like Y axis labels, except that the ticks are on a
20359 line above the numbers. Labels should look like this:
20368 The first tic is under the first column of the graph and is preceded by
20369 several blank spaces. These spaces provide room in rows above for the Y
20370 axis labels. The second, third, fourth, and subsequent ticks are all
20371 spaced equally, according to the value of @code{X-axis-label-spacing}.
20373 The second row of the X axis consists of numbers, preceded by several
20374 blank spaces and also separated according to the value of the variable
20375 @code{X-axis-label-spacing}.
20377 The value of the variable @code{X-axis-label-spacing} should itself be
20378 measured in units of @code{symbol-width}, since you may want to change
20379 the width of the symbols that you are using to print the body of the
20380 graph without changing the ways the graph is labelled.
20383 * Similarities differences::
20384 * X Axis Tic Marks::
20387 @node Similarities differences, X Axis Tic Marks, print-X-axis, print-X-axis
20389 @unnumberedsubsec Similarities and differences
20392 The @code{print-X-axis} function is constructed in more or less the
20393 same fashion as the @code{print-Y-axis} function except that it has
20394 two lines: the line of tic marks and the numbers. We will write a
20395 separate function to print each line and then combine them within the
20396 @code{print-X-axis} function.
20398 This is a three step process:
20402 Write a function to print the X axis tic marks, @code{print-X-axis-tic-line}.
20405 Write a function to print the X numbers, @code{print-X-axis-numbered-line}.
20408 Write a function to print both lines, the @code{print-X-axis} function,
20409 using @code{print-X-axis-tic-line} and
20410 @code{print-X-axis-numbered-line}.
20413 @node X Axis Tic Marks, , Similarities differences, print-X-axis
20414 @appendixsubsec X Axis Tic Marks
20416 The first function should print the X axis tic marks. We must specify
20417 the tic marks themselves and their spacing:
20421 (defvar X-axis-label-spacing
20422 (if (boundp 'graph-blank)
20423 (* 5 (length graph-blank)) 5)
20424 "Number of units from one X axis label to next.")
20429 (Note that the value of @code{graph-blank} is set by another
20430 @code{defvar}. The @code{boundp} predicate checks whether it has
20431 already been set; @code{boundp} returns @code{nil} if it has not. If
20432 @code{graph-blank} were unbound and we did not use this conditional
20433 construction, in a recent GNU Emacs, we would enter the debugger and
20434 see an error message saying @samp{@w{Debugger entered--Lisp error:}
20435 @w{(void-variable graph-blank)}}.)
20438 Here is the @code{defvar} for @code{X-axis-tic-symbol}:
20442 (defvar X-axis-tic-symbol "|"
20443 "String to insert to point to a column in X axis.")
20448 The goal is to make a line that looks like this:
20454 The first tic is indented so that it is under the first column, which is
20455 indented to provide space for the Y axis labels.
20457 A tic element consists of the blank spaces that stretch from one tic to
20458 the next plus a tic symbol. The number of blanks is determined by the
20459 width of the tic symbol and the @code{X-axis-label-spacing}.
20462 The code looks like this:
20466 ;;; X-axis-tic-element
20470 ;; @r{Make a string of blanks.}
20471 (- (* symbol-width X-axis-label-spacing)
20472 (length X-axis-tic-symbol))
20474 ;; @r{Concatenate blanks with tic symbol.}
20480 Next, we determine how many blanks are needed to indent the first tic
20481 mark to the first column of the graph. This uses the value of
20482 @code{full-Y-label-width} passed it by the @code{print-graph} function.
20485 The code to make @code{X-axis-leading-spaces}
20490 ;; X-axis-leading-spaces
20492 (make-string full-Y-label-width ? )
20497 We also need to determine the length of the horizontal axis, which is
20498 the length of the numbers list, and the number of ticks in the horizontal
20505 (length numbers-list)
20511 (* symbol-width X-axis-label-spacing)
20515 ;; number-of-X-ticks
20516 (if (zerop (% (X-length tic-width)))
20517 (/ (X-length tic-width))
20518 (1+ (/ (X-length tic-width))))
20523 All this leads us directly to the function for printing the X axis tic line:
20525 @findex print-X-axis-tic-line
20528 (defun print-X-axis-tic-line
20529 (number-of-X-tics X-axis-leading-spaces X-axis-tic-element)
20530 "Print ticks for X axis."
20531 (insert X-axis-leading-spaces)
20532 (insert X-axis-tic-symbol) ; @r{Under first column.}
20535 ;; @r{Insert second tic in the right spot.}
20538 (- (* symbol-width X-axis-label-spacing)
20539 ;; @r{Insert white space up to second tic symbol.}
20540 (* 2 (length X-axis-tic-symbol)))
20542 X-axis-tic-symbol))
20545 ;; @r{Insert remaining ticks.}
20546 (while (> number-of-X-tics 1)
20547 (insert X-axis-tic-element)
20548 (setq number-of-X-tics (1- number-of-X-tics))))
20552 The line of numbers is equally straightforward:
20555 First, we create a numbered element with blank spaces before each number:
20557 @findex X-axis-element
20560 (defun X-axis-element (number)
20561 "Construct a numbered X axis element."
20562 (let ((leading-spaces
20563 (- (* symbol-width X-axis-label-spacing)
20564 (length (number-to-string number)))))
20565 (concat (make-string leading-spaces ? )
20566 (number-to-string number))))
20570 Next, we create the function to print the numbered line, starting with
20571 the number ``1'' under the first column:
20573 @findex print-X-axis-numbered-line
20576 (defun print-X-axis-numbered-line
20577 (number-of-X-tics X-axis-leading-spaces)
20578 "Print line of X-axis numbers"
20579 (let ((number X-axis-label-spacing))
20580 (insert X-axis-leading-spaces)
20586 ;; @r{Insert white space up to next number.}
20587 (- (* symbol-width X-axis-label-spacing) 2)
20589 (number-to-string number)))
20592 ;; @r{Insert remaining numbers.}
20593 (setq number (+ number X-axis-label-spacing))
20594 (while (> number-of-X-tics 1)
20595 (insert (X-axis-element number))
20596 (setq number (+ number X-axis-label-spacing))
20597 (setq number-of-X-tics (1- number-of-X-tics)))))
20601 Finally, we need to write the @code{print-X-axis} that uses
20602 @code{print-X-axis-tic-line} and
20603 @code{print-X-axis-numbered-line}.
20605 The function must determine the local values of the variables used by both
20606 @code{print-X-axis-tic-line} and @code{print-X-axis-numbered-line}, and
20607 then it must call them. Also, it must print the carriage return that
20608 separates the two lines.
20610 The function consists of a varlist that specifies five local variables,
20611 and calls to each of the two line printing functions:
20613 @findex print-X-axis
20616 (defun print-X-axis (numbers-list)
20617 "Print X axis labels to length of NUMBERS-LIST."
20618 (let* ((leading-spaces
20619 (make-string full-Y-label-width ? ))
20622 ;; symbol-width @r{is provided by} graph-body-print
20623 (tic-width (* symbol-width X-axis-label-spacing))
20624 (X-length (length numbers-list))
20632 ;; @r{Make a string of blanks.}
20633 (- (* symbol-width X-axis-label-spacing)
20634 (length X-axis-tic-symbol))
20638 ;; @r{Concatenate blanks with tic symbol.}
20639 X-axis-tic-symbol))
20643 (if (zerop (% X-length tic-width))
20644 (/ X-length tic-width)
20645 (1+ (/ X-length tic-width)))))
20648 (print-X-axis-tic-line tic-number leading-spaces X-tic)
20650 (print-X-axis-numbered-line tic-number leading-spaces)))
20655 You can test @code{print-X-axis}:
20659 Install @code{X-axis-tic-symbol}, @code{X-axis-label-spacing},
20660 @code{print-X-axis-tic-line}, as well as @code{X-axis-element},
20661 @code{print-X-axis-numbered-line}, and @code{print-X-axis}.
20664 Copy the following expression:
20669 (let ((full-Y-label-width 5)
20672 '(1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16))))
20677 Switch to the @file{*scratch*} buffer and place the cursor where you
20678 want the axis labels to start.
20681 Type @kbd{M-:} (@code{eval-expression}).
20684 Yank the test expression into the minibuffer
20685 with @kbd{C-y} (@code{yank)}.
20688 Press @key{RET} to evaluate the expression.
20692 Emacs will print the horizontal axis like this:
20702 @node Print Whole Graph, , print-X-axis, Full Graph
20703 @appendixsec Printing the Whole Graph
20704 @cindex Printing the whole graph
20705 @cindex Whole graph printing
20706 @cindex Graph, printing all
20708 Now we are nearly ready to print the whole graph.
20710 The function to print the graph with the proper labels follows the
20711 outline we created earlier (@pxref{Full Graph, , A Graph with Labelled
20712 Axes}), but with additions.
20715 Here is the outline:
20719 (defun print-graph (numbers-list)
20720 "@var{documentation}@dots{}"
20721 (let ((height @dots{}
20725 (print-Y-axis height @dots{} )
20726 (graph-body-print numbers-list)
20727 (print-X-axis @dots{} )))
20732 * The final version::
20733 * Test print-graph::
20734 * Graphing words in defuns::
20738 * Final printed graph::
20741 @node The final version, Test print-graph, Print Whole Graph, Print Whole Graph
20743 @unnumberedsubsec Changes for the Final Version
20746 The final version is different from what we planned in two ways:
20747 first, it contains additional values calculated once in the varlist;
20748 second, it carries an option to specify the labels' increment per row.
20749 This latter feature turns out to be essential; otherwise, a graph may
20750 have more rows than fit on a display or on a sheet of paper.
20753 This new feature requires a change to the @code{Y-axis-column}
20754 function, to add @code{vertical-step} to it. The function looks like
20757 @findex Y-axis-column @r{Final version.}
20760 ;;; @r{Final version.}
20761 (defun Y-axis-column
20762 (height width-of-label &optional vertical-step)
20763 "Construct list of labels for Y axis.
20764 HEIGHT is maximum height of graph.
20765 WIDTH-OF-LABEL is maximum width of label.
20766 VERTICAL-STEP, an option, is a positive integer
20767 that specifies how much a Y axis label increments
20768 for each line. For example, a step of 5 means
20769 that each line is five units of the graph."
20773 (number-per-line (or vertical-step 1)))
20774 (while (> height 1)
20775 (if (zerop (% height Y-axis-label-spacing))
20778 ;; @r{Insert label.}
20782 (* height number-per-line)
20787 ;; @r{Else, insert blanks.}
20790 (make-string width-of-label ? )
20792 (setq height (1- height)))
20795 ;; @r{Insert base line.}
20796 (setq Y-axis (cons (Y-axis-element
20797 (or vertical-step 1)
20800 (nreverse Y-axis)))
20804 The values for the maximum height of graph and the width of a symbol
20805 are computed by @code{print-graph} in its @code{let} expression; so
20806 @code{graph-body-print} must be changed to accept them.
20808 @findex graph-body-print @r{Final version.}
20811 ;;; @r{Final version.}
20812 (defun graph-body-print (numbers-list height symbol-width)
20813 "Print a bar graph of the NUMBERS-LIST.
20814 The numbers-list consists of the Y-axis values.
20815 HEIGHT is maximum height of graph.
20816 SYMBOL-WIDTH is number of each column."
20819 (let (from-position)
20820 (while numbers-list
20821 (setq from-position (point))
20823 (column-of-graph height (car numbers-list)))
20824 (goto-char from-position)
20825 (forward-char symbol-width)
20828 ;; @r{Draw graph column by column.}
20830 (setq numbers-list (cdr numbers-list)))
20831 ;; @r{Place point for X axis labels.}
20832 (forward-line height)
20838 Finally, the code for the @code{print-graph} function:
20840 @findex print-graph @r{Final version.}
20843 ;;; @r{Final version.}
20845 (numbers-list &optional vertical-step)
20846 "Print labelled bar graph of the NUMBERS-LIST.
20847 The numbers-list consists of the Y-axis values.
20851 Optionally, VERTICAL-STEP, a positive integer,
20852 specifies how much a Y axis label increments for
20853 each line. For example, a step of 5 means that
20854 each row is five units."
20857 (let* ((symbol-width (length graph-blank))
20858 ;; @code{height} @r{is both the largest number}
20859 ;; @r{and the number with the most digits.}
20860 (height (apply 'max numbers-list))
20863 (height-of-top-line
20864 (if (zerop (% height Y-axis-label-spacing))
20867 (* (1+ (/ height Y-axis-label-spacing))
20868 Y-axis-label-spacing)))
20871 (vertical-step (or vertical-step 1))
20872 (full-Y-label-width
20878 (* height-of-top-line vertical-step))
20884 height-of-top-line full-Y-label-width vertical-step)
20888 numbers-list height-of-top-line symbol-width)
20889 (print-X-axis numbers-list)))
20893 @node Test print-graph, Graphing words in defuns, The final version, Print Whole Graph
20894 @appendixsubsec Testing @code{print-graph}
20897 We can test the @code{print-graph} function with a short list of numbers:
20901 Install the final versions of @code{Y-axis-column},
20902 @code{graph-body-print}, and @code{print-graph} (in addition to the
20906 Copy the following expression:
20909 (print-graph '(3 2 5 6 7 5 3 4 6 4 3 2 1))
20913 Switch to the @file{*scratch*} buffer and place the cursor where you
20914 want the axis labels to start.
20917 Type @kbd{M-:} (@code{eval-expression}).
20920 Yank the test expression into the minibuffer
20921 with @kbd{C-y} (@code{yank)}.
20924 Press @key{RET} to evaluate the expression.
20928 Emacs will print a graph that looks like this:
20949 On the other hand, if you pass @code{print-graph} a
20950 @code{vertical-step} value of 2, by evaluating this expression:
20953 (print-graph '(3 2 5 6 7 5 3 4 6 4 3 2 1) 2)
20958 The graph looks like this:
20979 (A question: is the `2' on the bottom of the vertical axis a bug or a
20980 feature? If you think it is a bug, and should be a `1' instead, (or
20981 even a `0'), you can modify the sources.)
20983 @node Graphing words in defuns, lambda, Test print-graph, Print Whole Graph
20984 @appendixsubsec Graphing Numbers of Words and Symbols
20986 Now for the graph for which all this code was written: a graph that
20987 shows how many function definitions contain fewer than 10 words and
20988 symbols, how many contain between 10 and 19 words and symbols, how
20989 many contain between 20 and 29 words and symbols, and so on.
20991 This is a multi-step process. First make sure you have loaded all the
20995 It is a good idea to reset the value of @code{top-of-ranges} in case
20996 you have set it to some different value. You can evaluate the
21001 (setq top-of-ranges
21004 110 120 130 140 150
21005 160 170 180 190 200
21006 210 220 230 240 250
21007 260 270 280 290 300)
21012 Next create a list of the number of words and symbols in each range.
21016 Evaluate the following:
21020 (setq list-for-graph
21023 (recursive-lengths-list-many-files
21024 (directory-files "/usr/local/emacs/lisp"
21032 On my old machine, this took about an hour. It looked though 303 Lisp
21033 files in my copy of Emacs version 19.23. After all that computing,
21034 the @code{list-for-graph} had this value:
21038 (537 1027 955 785 594 483 349 292 224 199 166 120 116 99
21039 90 80 67 48 52 45 41 33 28 26 25 20 12 28 11 13 220)
21044 This means that my copy of Emacs had 537 function definitions with
21045 fewer than 10 words or symbols in them, 1,027 function definitions
21046 with 10 to 19 words or symbols in them, 955 function definitions with
21047 20 to 29 words or symbols in them, and so on.
21049 Clearly, just by looking at this list we can see that most function
21050 definitions contain ten to thirty words and symbols.
21052 Now for printing. We do @emph{not} want to print a graph that is
21053 1,030 lines high @dots{} Instead, we should print a graph that is
21054 fewer than twenty-five lines high. A graph that height can be
21055 displayed on almost any monitor, and easily printed on a sheet of paper.
21057 This means that each value in @code{list-for-graph} must be reduced to
21058 one-fiftieth its present value.
21060 Here is a short function to do just that, using two functions we have
21061 not yet seen, @code{mapcar} and @code{lambda}.
21065 (defun one-fiftieth (full-range)
21066 "Return list, each number one-fiftieth of previous."
21067 (mapcar '(lambda (arg) (/ arg 50)) full-range))
21071 @node lambda, mapcar, Graphing words in defuns, Print Whole Graph
21072 @appendixsubsec A @code{lambda} Expression: Useful Anonymity
21073 @cindex Anonymous function
21076 @code{lambda} is the symbol for an anonymous function, a function
21077 without a name. Every time you use an anonymous function, you need to
21078 include its whole body.
21085 (lambda (arg) (/ arg 50))
21089 is a function definition that says `return the value resulting from
21090 dividing whatever is passed to me as @code{arg} by 50'.
21093 Earlier, for example, we had a function @code{multiply-by-seven}; it
21094 multiplied its argument by 7. This function is similar, except it
21095 divides its argument by 50; and, it has no name. The anonymous
21096 equivalent of @code{multiply-by-seven} is:
21099 (lambda (number) (* 7 number))
21103 (@xref{defun, , The @code{defun} Special Form}.)
21107 If we want to multiply 3 by 7, we can write:
21109 @c !!! Clear print-postscript-figures if the computer formatting this
21110 @c document is too small and cannot handle all the diagrams and figures.
21111 @c clear print-postscript-figures
21112 @c set print-postscript-figures
21113 @c lambda example diagram #1
21117 (multiply-by-seven 3)
21118 \_______________/ ^
21124 @ifset print-postscript-figures
21127 @center @image{lambda-1}
21128 %%%% old method of including an image
21129 % \input /usr/local/lib/tex/inputs/psfig.tex
21130 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/lambda-1.eps}}
21135 @ifclear print-postscript-figures
21139 (multiply-by-seven 3)
21140 \_______________/ ^
21149 This expression returns 21.
21153 Similarly, we can write:
21155 @c lambda example diagram #2
21159 ((lambda (number) (* 7 number)) 3)
21160 \____________________________/ ^
21162 anonymous function argument
21166 @ifset print-postscript-figures
21169 @center @image{lambda-2}
21170 %%%% old method of including an image
21171 % \input /usr/local/lib/tex/inputs/psfig.tex
21172 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/lambda-2.eps}}
21177 @ifclear print-postscript-figures
21181 ((lambda (number) (* 7 number)) 3)
21182 \____________________________/ ^
21184 anonymous function argument
21192 If we want to divide 100 by 50, we can write:
21194 @c lambda example diagram #3
21198 ((lambda (arg) (/ arg 50)) 100)
21199 \______________________/ \_/
21201 anonymous function argument
21205 @ifset print-postscript-figures
21208 @center @image{lambda-3}
21209 %%%% old method of including an image
21210 % \input /usr/local/lib/tex/inputs/psfig.tex
21211 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/lambda-3.eps}}
21216 @ifclear print-postscript-figures
21220 ((lambda (arg) (/ arg 50)) 100)
21221 \______________________/ \_/
21223 anonymous function argument
21230 This expression returns 2. The 100 is passed to the function, which
21231 divides that number by 50.
21233 @xref{Lambda Expressions, , Lambda Expressions, elisp, The GNU Emacs
21234 Lisp Reference Manual}, for more about @code{lambda}. Lisp and lambda
21235 expressions derive from the Lambda Calculus.
21237 @node mapcar, Another Bug, lambda, Print Whole Graph
21238 @appendixsubsec The @code{mapcar} Function
21241 @code{mapcar} is a function that calls its first argument with each
21242 element of its second argument, in turn. The second argument must be
21245 The @samp{map} part of the name comes from the mathematical phrase,
21246 `mapping over a domain', meaning to apply a function to each of the
21247 elements in a domain. The mathematical phrase is based on the
21248 metaphor of a surveyor walking, one step at a time, over an area he is
21249 mapping. And @samp{car}, of course, comes from the Lisp notion of the
21258 (mapcar '1+ '(2 4 6))
21264 The function @code{1+} which adds one to its argument, is executed on
21265 @emph{each} element of the list, and a new list is returned.
21267 Contrast this with @code{apply}, which applies its first argument to
21269 (@xref{Readying a Graph, , Readying a Graph}, for a explanation of
21273 In the definition of @code{one-fiftieth}, the first argument is the
21274 anonymous function:
21277 (lambda (arg) (/ arg 50))
21281 and the second argument is @code{full-range}, which will be bound to
21282 @code{list-for-graph}.
21285 The whole expression looks like this:
21288 (mapcar '(lambda (arg) (/ arg 50)) full-range))
21291 @xref{Mapping Functions, , Mapping Functions, elisp, The GNU Emacs
21292 Lisp Reference Manual}, for more about @code{mapcar}.
21294 Using the @code{one-fiftieth} function, we can generate a list in
21295 which each element is one-fiftieth the size of the corresponding
21296 element in @code{list-for-graph}.
21300 (setq fiftieth-list-for-graph
21301 (one-fiftieth list-for-graph))
21306 The resulting list looks like this:
21310 (10 20 19 15 11 9 6 5 4 3 3 2 2
21311 1 1 1 1 0 1 0 0 0 0 0 0 0 0 0 0 0 4)
21316 This, we are almost ready to print! (We also notice the loss of
21317 information: many of the higher ranges are 0, meaning that fewer than
21318 50 defuns had that many words or symbols---but not necessarily meaning
21319 that none had that many words or symbols.)
21321 @node Another Bug, Final printed graph, mapcar, Print Whole Graph
21322 @appendixsubsec Another Bug @dots{} Most Insidious
21323 @cindex Bug, most insidious type
21324 @cindex Insidious type of bug
21326 I said `almost ready to print'! Of course, there is a bug in the
21327 @code{print-graph} function @dots{} It has a @code{vertical-step}
21328 option, but not a @code{horizontal-step} option. The
21329 @code{top-of-range} scale goes from 10 to 300 by tens. But the
21330 @code{print-graph} function will print only by ones.
21332 This is a classic example of what some consider the most insidious
21333 type of bug, the bug of omission. This is not the kind of bug you can
21334 find by studying the code, for it is not in the code; it is an omitted
21335 feature. Your best actions are to try your program early and often;
21336 and try to arrange, as much as you can, to write code that is easy to
21337 understand and easy to change. Try to be aware, whenever you can,
21338 that whatever you have written, @emph{will} be rewritten, if not soon,
21339 eventually. A hard maxim to follow.
21341 It is the @code{print-X-axis-numbered-line} function that needs the
21342 work; and then the @code{print-X-axis} and the @code{print-graph}
21343 functions need to be adapted. Not much needs to be done; there is one
21344 nicety: the numbers ought to line up under the tic marks. This takes
21348 Here is the corrected @code{print-X-axis-numbered-line}:
21352 (defun print-X-axis-numbered-line
21353 (number-of-X-tics X-axis-leading-spaces
21354 &optional horizontal-step)
21355 "Print line of X-axis numbers"
21356 (let ((number X-axis-label-spacing)
21357 (horizontal-step (or horizontal-step 1)))
21360 (insert X-axis-leading-spaces)
21361 ;; @r{Delete extra leading spaces.}
21364 (length (number-to-string horizontal-step)))))
21369 ;; @r{Insert white space.}
21371 X-axis-label-spacing)
21374 (number-to-string horizontal-step)))
21378 (* number horizontal-step))))
21381 ;; @r{Insert remaining numbers.}
21382 (setq number (+ number X-axis-label-spacing))
21383 (while (> number-of-X-tics 1)
21384 (insert (X-axis-element
21385 (* number horizontal-step)))
21386 (setq number (+ number X-axis-label-spacing))
21387 (setq number-of-X-tics (1- number-of-X-tics)))))
21392 If you are reading this in Info, you can see the new versions of
21393 @code{print-X-axis} @code{print-graph} and evaluate them. If you are
21394 reading this in a printed book, you can see the changed lines here
21395 (the full text is too much to print).
21400 (defun print-X-axis (numbers-list horizontal-step)
21402 (print-X-axis-numbered-line
21403 tic-number leading-spaces horizontal-step))
21411 &optional vertical-step horizontal-step)
21413 (print-X-axis numbers-list horizontal-step))
21421 (defun print-X-axis (numbers-list horizontal-step)
21422 "Print X axis labels to length of NUMBERS-LIST.
21423 Optionally, HORIZONTAL-STEP, a positive integer,
21424 specifies how much an X axis label increments for
21428 ;; Value of symbol-width and full-Y-label-width
21429 ;; are passed by `print-graph'.
21430 (let* ((leading-spaces
21431 (make-string full-Y-label-width ? ))
21432 ;; symbol-width @r{is provided by} graph-body-print
21433 (tic-width (* symbol-width X-axis-label-spacing))
21434 (X-length (length numbers-list))
21440 ;; @r{Make a string of blanks.}
21441 (- (* symbol-width X-axis-label-spacing)
21442 (length X-axis-tic-symbol))
21446 ;; @r{Concatenate blanks with tic symbol.}
21447 X-axis-tic-symbol))
21449 (if (zerop (% X-length tic-width))
21450 (/ X-length tic-width)
21451 (1+ (/ X-length tic-width)))))
21455 (print-X-axis-tic-line
21456 tic-number leading-spaces X-tic)
21458 (print-X-axis-numbered-line
21459 tic-number leading-spaces horizontal-step)))
21466 (numbers-list &optional vertical-step horizontal-step)
21467 "Print labelled bar graph of the NUMBERS-LIST.
21468 The numbers-list consists of the Y-axis values.
21472 Optionally, VERTICAL-STEP, a positive integer,
21473 specifies how much a Y axis label increments for
21474 each line. For example, a step of 5 means that
21475 each row is five units.
21479 Optionally, HORIZONTAL-STEP, a positive integer,
21480 specifies how much an X axis label increments for
21482 (let* ((symbol-width (length graph-blank))
21483 ;; @code{height} @r{is both the largest number}
21484 ;; @r{and the number with the most digits.}
21485 (height (apply 'max numbers-list))
21488 (height-of-top-line
21489 (if (zerop (% height Y-axis-label-spacing))
21492 (* (1+ (/ height Y-axis-label-spacing))
21493 Y-axis-label-spacing)))
21496 (vertical-step (or vertical-step 1))
21497 (full-Y-label-width
21501 (* height-of-top-line vertical-step))
21506 height-of-top-line full-Y-label-width vertical-step)
21508 numbers-list height-of-top-line symbol-width)
21509 (print-X-axis numbers-list horizontal-step)))
21516 Graphing Definitions Re-listed
21519 Here are all the graphing definitions in their final form:
21523 (defvar top-of-ranges
21526 110 120 130 140 150
21527 160 170 180 190 200
21528 210 220 230 240 250)
21529 "List specifying ranges for `defuns-per-range'.")
21533 (defvar graph-symbol "*"
21534 "String used as symbol in graph, usually an asterisk.")
21538 (defvar graph-blank " "
21539 "String used as blank in graph, usually a blank space.
21540 graph-blank must be the same number of columns wide
21545 (defvar Y-axis-tic " - "
21546 "String that follows number in a Y axis label.")
21550 (defvar Y-axis-label-spacing 5
21551 "Number of lines from one Y axis label to next.")
21555 (defvar X-axis-tic-symbol "|"
21556 "String to insert to point to a column in X axis.")
21560 (defvar X-axis-label-spacing
21561 (if (boundp 'graph-blank)
21562 (* 5 (length graph-blank)) 5)
21563 "Number of units from one X axis label to next.")
21569 (defun count-words-in-defun ()
21570 "Return the number of words and symbols in a defun."
21571 (beginning-of-defun)
21573 (end (save-excursion (end-of-defun) (point))))
21578 (and (< (point) end)
21580 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*"
21582 (setq count (1+ count)))
21589 (defun lengths-list-file (filename)
21590 "Return list of definitions' lengths within FILE.
21591 The returned list is a list of numbers.
21592 Each number is the number of words or
21593 symbols in one function definition."
21597 (message "Working on `%s' ... " filename)
21599 (let ((buffer (find-file-noselect filename))
21601 (set-buffer buffer)
21602 (setq buffer-read-only t)
21604 (goto-char (point-min))
21608 (while (re-search-forward "^(defun" nil t)
21610 (cons (count-words-in-defun) lengths-list)))
21611 (kill-buffer buffer)
21618 (defun lengths-list-many-files (list-of-files)
21619 "Return list of lengths of defuns in LIST-OF-FILES."
21620 (let (lengths-list)
21621 ;;; @r{true-or-false-test}
21622 (while list-of-files
21628 ;;; @r{Generate a lengths' list.}
21630 (expand-file-name (car list-of-files)))))
21631 ;;; @r{Make files' list shorter.}
21632 (setq list-of-files (cdr list-of-files)))
21633 ;;; @r{Return final value of lengths' list.}
21640 (defun defuns-per-range (sorted-lengths top-of-ranges)
21641 "SORTED-LENGTHS defuns in each TOP-OF-RANGES range."
21642 (let ((top-of-range (car top-of-ranges))
21643 (number-within-range 0)
21644 defuns-per-range-list)
21649 (while top-of-ranges
21653 ;; @r{Need number for numeric test.}
21654 (car sorted-lengths)
21655 (< (car sorted-lengths) top-of-range))
21657 ;; @r{Count number of definitions within current range.}
21658 (setq number-within-range (1+ number-within-range))
21659 (setq sorted-lengths (cdr sorted-lengths)))
21663 ;; @r{Exit inner loop but remain within outer loop.}
21665 (setq defuns-per-range-list
21666 (cons number-within-range defuns-per-range-list))
21667 (setq number-within-range 0) ; @r{Reset count to zero.}
21669 ;; @r{Move to next range.}
21670 (setq top-of-ranges (cdr top-of-ranges))
21671 ;; @r{Specify next top of range value.}
21672 (setq top-of-range (car top-of-ranges)))
21676 ;; @r{Exit outer loop and count the number of defuns larger than}
21677 ;; @r{ the largest top-of-range value.}
21678 (setq defuns-per-range-list
21680 (length sorted-lengths)
21681 defuns-per-range-list))
21683 ;; @r{Return a list of the number of definitions within each range,}
21684 ;; @r{ smallest to largest.}
21685 (nreverse defuns-per-range-list)))
21691 (defun column-of-graph (max-graph-height actual-height)
21692 "Return list of MAX-GRAPH-HEIGHT strings;
21693 ACTUAL-HEIGHT are graph-symbols.
21694 The graph-symbols are contiguous entries at the end
21696 The list will be inserted as one column of a graph.
21697 The strings are either graph-blank or graph-symbol."
21701 (let ((insert-list nil)
21702 (number-of-top-blanks
21703 (- max-graph-height actual-height)))
21705 ;; @r{Fill in @code{graph-symbols}.}
21706 (while (> actual-height 0)
21707 (setq insert-list (cons graph-symbol insert-list))
21708 (setq actual-height (1- actual-height)))
21712 ;; @r{Fill in @code{graph-blanks}.}
21713 (while (> number-of-top-blanks 0)
21714 (setq insert-list (cons graph-blank insert-list))
21715 (setq number-of-top-blanks
21716 (1- number-of-top-blanks)))
21718 ;; @r{Return whole list.}
21725 (defun Y-axis-element (number full-Y-label-width)
21726 "Construct a NUMBERed label element.
21727 A numbered element looks like this ` 5 - ',
21728 and is padded as needed so all line up with
21729 the element for the largest number."
21732 (let* ((leading-spaces
21733 (- full-Y-label-width
21735 (concat (number-to-string number)
21740 (make-string leading-spaces ? )
21741 (number-to-string number)
21748 (defun print-Y-axis
21749 (height full-Y-label-width &optional vertical-step)
21750 "Insert Y axis by HEIGHT and FULL-Y-LABEL-WIDTH.
21751 Height must be the maximum height of the graph.
21752 Full width is the width of the highest label element.
21753 Optionally, print according to VERTICAL-STEP."
21756 ;; Value of height and full-Y-label-width
21757 ;; are passed by `print-graph'.
21758 (let ((start (point)))
21760 (Y-axis-column height full-Y-label-width vertical-step))
21763 ;; @r{Place point ready for inserting graph.}
21765 ;; @r{Move point forward by value of} full-Y-label-width
21766 (forward-char full-Y-label-width)))
21772 (defun print-X-axis-tic-line
21773 (number-of-X-tics X-axis-leading-spaces X-axis-tic-element)
21774 "Print ticks for X axis."
21775 (insert X-axis-leading-spaces)
21776 (insert X-axis-tic-symbol) ; @r{Under first column.}
21779 ;; @r{Insert second tic in the right spot.}
21782 (- (* symbol-width X-axis-label-spacing)
21783 ;; @r{Insert white space up to second tic symbol.}
21784 (* 2 (length X-axis-tic-symbol)))
21786 X-axis-tic-symbol))
21789 ;; @r{Insert remaining ticks.}
21790 (while (> number-of-X-tics 1)
21791 (insert X-axis-tic-element)
21792 (setq number-of-X-tics (1- number-of-X-tics))))
21798 (defun X-axis-element (number)
21799 "Construct a numbered X axis element."
21800 (let ((leading-spaces
21801 (- (* symbol-width X-axis-label-spacing)
21802 (length (number-to-string number)))))
21803 (concat (make-string leading-spaces ? )
21804 (number-to-string number))))
21810 (defun graph-body-print (numbers-list height symbol-width)
21811 "Print a bar graph of the NUMBERS-LIST.
21812 The numbers-list consists of the Y-axis values.
21813 HEIGHT is maximum height of graph.
21814 SYMBOL-WIDTH is number of each column."
21817 (let (from-position)
21818 (while numbers-list
21819 (setq from-position (point))
21821 (column-of-graph height (car numbers-list)))
21822 (goto-char from-position)
21823 (forward-char symbol-width)
21826 ;; @r{Draw graph column by column.}
21828 (setq numbers-list (cdr numbers-list)))
21829 ;; @r{Place point for X axis labels.}
21830 (forward-line height)
21837 (defun Y-axis-column
21838 (height width-of-label &optional vertical-step)
21839 "Construct list of labels for Y axis.
21840 HEIGHT is maximum height of graph.
21841 WIDTH-OF-LABEL is maximum width of label.
21844 VERTICAL-STEP, an option, is a positive integer
21845 that specifies how much a Y axis label increments
21846 for each line. For example, a step of 5 means
21847 that each line is five units of the graph."
21849 (number-per-line (or vertical-step 1)))
21852 (while (> height 1)
21853 (if (zerop (% height Y-axis-label-spacing))
21854 ;; @r{Insert label.}
21858 (* height number-per-line)
21863 ;; @r{Else, insert blanks.}
21866 (make-string width-of-label ? )
21868 (setq height (1- height)))
21871 ;; @r{Insert base line.}
21872 (setq Y-axis (cons (Y-axis-element
21873 (or vertical-step 1)
21876 (nreverse Y-axis)))
21882 (defun print-X-axis-numbered-line
21883 (number-of-X-tics X-axis-leading-spaces
21884 &optional horizontal-step)
21885 "Print line of X-axis numbers"
21886 (let ((number X-axis-label-spacing)
21887 (horizontal-step (or horizontal-step 1)))
21890 (insert X-axis-leading-spaces)
21892 (delete-char (- (1- (length (number-to-string horizontal-step)))))
21895 ;; @r{Insert white space up to next number.}
21896 (- (* symbol-width X-axis-label-spacing)
21897 (1- (length (number-to-string horizontal-step)))
21900 (number-to-string (* number horizontal-step))))
21903 ;; @r{Insert remaining numbers.}
21904 (setq number (+ number X-axis-label-spacing))
21905 (while (> number-of-X-tics 1)
21906 (insert (X-axis-element (* number horizontal-step)))
21907 (setq number (+ number X-axis-label-spacing))
21908 (setq number-of-X-tics (1- number-of-X-tics)))))
21914 (defun print-X-axis (numbers-list horizontal-step)
21915 "Print X axis labels to length of NUMBERS-LIST.
21916 Optionally, HORIZONTAL-STEP, a positive integer,
21917 specifies how much an X axis label increments for
21921 ;; Value of symbol-width and full-Y-label-width
21922 ;; are passed by `print-graph'.
21923 (let* ((leading-spaces
21924 (make-string full-Y-label-width ? ))
21925 ;; symbol-width @r{is provided by} graph-body-print
21926 (tic-width (* symbol-width X-axis-label-spacing))
21927 (X-length (length numbers-list))
21933 ;; @r{Make a string of blanks.}
21934 (- (* symbol-width X-axis-label-spacing)
21935 (length X-axis-tic-symbol))
21939 ;; @r{Concatenate blanks with tic symbol.}
21940 X-axis-tic-symbol))
21942 (if (zerop (% X-length tic-width))
21943 (/ X-length tic-width)
21944 (1+ (/ X-length tic-width)))))
21948 (print-X-axis-tic-line
21949 tic-number leading-spaces X-tic)
21951 (print-X-axis-numbered-line
21952 tic-number leading-spaces horizontal-step)))
21958 (defun one-fiftieth (full-range)
21959 "Return list, each number of which is 1/50th previous."
21960 (mapcar '(lambda (arg) (/ arg 50)) full-range))
21967 (numbers-list &optional vertical-step horizontal-step)
21968 "Print labelled bar graph of the NUMBERS-LIST.
21969 The numbers-list consists of the Y-axis values.
21973 Optionally, VERTICAL-STEP, a positive integer,
21974 specifies how much a Y axis label increments for
21975 each line. For example, a step of 5 means that
21976 each row is five units.
21980 Optionally, HORIZONTAL-STEP, a positive integer,
21981 specifies how much an X axis label increments for
21983 (let* ((symbol-width (length graph-blank))
21984 ;; @code{height} @r{is both the largest number}
21985 ;; @r{and the number with the most digits.}
21986 (height (apply 'max numbers-list))
21989 (height-of-top-line
21990 (if (zerop (% height Y-axis-label-spacing))
21993 (* (1+ (/ height Y-axis-label-spacing))
21994 Y-axis-label-spacing)))
21997 (vertical-step (or vertical-step 1))
21998 (full-Y-label-width
22002 (* height-of-top-line vertical-step))
22008 height-of-top-line full-Y-label-width vertical-step)
22010 numbers-list height-of-top-line symbol-width)
22011 (print-X-axis numbers-list horizontal-step)))
22018 @node Final printed graph, , Another Bug, Print Whole Graph
22019 @appendixsubsec The Printed Graph
22021 When made and installed, you can call the @code{print-graph} command
22027 (print-graph fiftieth-list-for-graph 50 10)
22057 50 - ***************** * *
22059 10 50 100 150 200 250 300 350
22066 The largest group of functions contain 10 -- 19 words and symbols each.
22068 @node Free Software and Free Manuals, GNU Free Documentation License, Full Graph, Top
22069 @appendix Free Software and Free Manuals
22071 @strong{by Richard M. Stallman}
22074 The biggest deficiency in free operating systems is not in the
22075 software---it is the lack of good free manuals that we can include in
22076 these systems. Many of our most important programs do not come with
22077 full manuals. Documentation is an essential part of any software
22078 package; when an important free software package does not come with a
22079 free manual, that is a major gap. We have many such gaps today.
22081 Once upon a time, many years ago, I thought I would learn Perl. I got
22082 a copy of a free manual, but I found it hard to read. When I asked
22083 Perl users about alternatives, they told me that there were better
22084 introductory manuals---but those were not free.
22086 Why was this? The authors of the good manuals had written them for
22087 O'Reilly Associates, which published them with restrictive terms---no
22088 copying, no modification, source files not available---which exclude
22089 them from the free software community.
22091 That wasn't the first time this sort of thing has happened, and (to
22092 our community's great loss) it was far from the last. Proprietary
22093 manual publishers have enticed a great many authors to restrict their
22094 manuals since then. Many times I have heard a GNU user eagerly tell me
22095 about a manual that he is writing, with which he expects to help the
22096 GNU project---and then had my hopes dashed, as he proceeded to explain
22097 that he had signed a contract with a publisher that would restrict it
22098 so that we cannot use it.
22100 Given that writing good English is a rare skill among programmers, we
22101 can ill afford to lose manuals this way.
22104 (The Free Software Foundation
22105 @uref{http://www.gnu.org/doc/doc.html#DescriptionsOfGNUDocumentation, ,
22106 sells printed copies} of free @uref{http://www.gnu.org/doc/doc.html,
22107 GNU manuals}, too.)
22109 Free documentation, like free software, is a matter of freedom, not
22110 price. The problem with these manuals was not that O'Reilly Associates
22111 charged a price for printed copies---that in itself is fine. (The Free
22112 Software Foundation sells printed copies of free GNU manuals, too.)
22113 But GNU manuals are available in source code form, while these manuals
22114 are available only on paper. GNU manuals come with permission to copy
22115 and modify; the Perl manuals do not. These restrictions are the
22118 The criterion for a free manual is pretty much the same as for free
22119 software: it is a matter of giving all users certain
22120 freedoms. Redistribution (including commercial redistribution) must be
22121 permitted, so that the manual can accompany every copy of the program,
22122 on-line or on paper. Permission for modification is crucial too.
22124 As a general rule, I don't believe that it is essential for people to
22125 have permission to modify all sorts of articles and books. The issues
22126 for writings are not necessarily the same as those for software. For
22127 example, I don't think you or I are obliged to give permission to
22128 modify articles like this one, which describe our actions and our
22131 But there is a particular reason why the freedom to modify is crucial
22132 for documentation for free software. When people exercise their right
22133 to modify the software, and add or change its features, if they are
22134 conscientious they will change the manual too---so they can provide
22135 accurate and usable documentation with the modified program. A manual
22136 which forbids programmers to be conscientious and finish the job, or
22137 more precisely requires them to write a new manual from scratch if
22138 they change the program, does not fill our community's needs.
22140 While a blanket prohibition on modification is unacceptable, some
22141 kinds of limits on the method of modification pose no problem. For
22142 example, requirements to preserve the original author's copyright
22143 notice, the distribution terms, or the list of authors, are ok. It is
22144 also no problem to require modified versions to include notice that
22145 they were modified, even to have entire sections that may not be
22146 deleted or changed, as long as these sections deal with nontechnical
22147 topics. (Some GNU manuals have them.)
22149 These kinds of restrictions are not a problem because, as a practical
22150 matter, they don't stop the conscientious programmer from adapting the
22151 manual to fit the modified program. In other words, they don't block
22152 the free software community from making full use of the manual.
22154 However, it must be possible to modify all the technical content of
22155 the manual, and then distribute the result in all the usual media,
22156 through all the usual channels; otherwise, the restrictions do block
22157 the community, the manual is not free, and so we need another manual.
22159 Unfortunately, it is often hard to find someone to write another
22160 manual when a proprietary manual exists. The obstacle is that many
22161 users think that a proprietary manual is good enough---so they don't
22162 see the need to write a free manual. They do not see that the free
22163 operating system has a gap that needs filling.
22165 Why do users think that proprietary manuals are good enough? Some have
22166 not considered the issue. I hope this article will do something to
22169 Other users consider proprietary manuals acceptable for the same
22170 reason so many people consider proprietary software acceptable: they
22171 judge in purely practical terms, not using freedom as a
22172 criterion. These people are entitled to their opinions, but since
22173 those opinions spring from values which do not include freedom, they
22174 are no guide for those of us who do value freedom.
22176 Please spread the word about this issue. We continue to lose manuals
22177 to proprietary publishing. If we spread the word that proprietary
22178 manuals are not sufficient, perhaps the next person who wants to help
22179 GNU by writing documentation will realize, before it is too late, that
22180 he must above all make it free.
22182 We can also encourage commercial publishers to sell free, copylefted
22183 manuals instead of proprietary ones. One way you can help this is to
22184 check the distribution terms of a manual before you buy it, and prefer
22185 copylefted manuals to non-copylefted ones.
22189 Note: The Free Software Foundation maintains a page on its Web site
22190 that lists free books available from other publishers:@*
22191 @uref{http://www.gnu.org/doc/other-free-books.html}
22193 @node GNU Free Documentation License, Index, Free Software and Free Manuals, Top
22194 @appendix GNU Free Documentation License
22196 @cindex FDL, GNU Free Documentation License
22197 @center Version 1.2, November 2002
22200 Copyright @copyright{} 2000,2001,2002 Free Software Foundation, Inc.
22201 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA
22203 Everyone is permitted to copy and distribute verbatim copies
22204 of this license document, but changing it is not allowed.
22211 The purpose of this License is to make a manual, textbook, or other
22212 functional and useful document @dfn{free} in the sense of freedom: to
22213 assure everyone the effective freedom to copy and redistribute it,
22214 with or without modifying it, either commercially or noncommercially.
22215 Secondarily, this License preserves for the author and publisher a way
22216 to get credit for their work, while not being considered responsible
22217 for modifications made by others.
22219 This License is a kind of ``copyleft'', which means that derivative
22220 works of the document must themselves be free in the same sense. It
22221 complements the GNU General Public License, which is a copyleft
22222 license designed for free software.
22224 We have designed this License in order to use it for manuals for free
22225 software, because free software needs free documentation: a free
22226 program should come with manuals providing the same freedoms that the
22227 software does. But this License is not limited to software manuals;
22228 it can be used for any textual work, regardless of subject matter or
22229 whether it is published as a printed book. We recommend this License
22230 principally for works whose purpose is instruction or reference.
22233 APPLICABILITY AND DEFINITIONS
22235 This License applies to any manual or other work, in any medium, that
22236 contains a notice placed by the copyright holder saying it can be
22237 distributed under the terms of this License. Such a notice grants a
22238 world-wide, royalty-free license, unlimited in duration, to use that
22239 work under the conditions stated herein. The ``Document'', below,
22240 refers to any such manual or work. Any member of the public is a
22241 licensee, and is addressed as ``you''. You accept the license if you
22242 copy, modify or distribute the work in a way requiring permission
22243 under copyright law.
22245 A ``Modified Version'' of the Document means any work containing the
22246 Document or a portion of it, either copied verbatim, or with
22247 modifications and/or translated into another language.
22249 A ``Secondary Section'' is a named appendix or a front-matter section
22250 of the Document that deals exclusively with the relationship of the
22251 publishers or authors of the Document to the Document's overall
22252 subject (or to related matters) and contains nothing that could fall
22253 directly within that overall subject. (Thus, if the Document is in
22254 part a textbook of mathematics, a Secondary Section may not explain
22255 any mathematics.) The relationship could be a matter of historical
22256 connection with the subject or with related matters, or of legal,
22257 commercial, philosophical, ethical or political position regarding
22260 The ``Invariant Sections'' are certain Secondary Sections whose titles
22261 are designated, as being those of Invariant Sections, in the notice
22262 that says that the Document is released under this License. If a
22263 section does not fit the above definition of Secondary then it is not
22264 allowed to be designated as Invariant. The Document may contain zero
22265 Invariant Sections. If the Document does not identify any Invariant
22266 Sections then there are none.
22268 The ``Cover Texts'' are certain short passages of text that are listed,
22269 as Front-Cover Texts or Back-Cover Texts, in the notice that says that
22270 the Document is released under this License. A Front-Cover Text may
22271 be at most 5 words, and a Back-Cover Text may be at most 25 words.
22273 A ``Transparent'' copy of the Document means a machine-readable copy,
22274 represented in a format whose specification is available to the
22275 general public, that is suitable for revising the document
22276 straightforwardly with generic text editors or (for images composed of
22277 pixels) generic paint programs or (for drawings) some widely available
22278 drawing editor, and that is suitable for input to text formatters or
22279 for automatic translation to a variety of formats suitable for input
22280 to text formatters. A copy made in an otherwise Transparent file
22281 format whose markup, or absence of markup, has been arranged to thwart
22282 or discourage subsequent modification by readers is not Transparent.
22283 An image format is not Transparent if used for any substantial amount
22284 of text. A copy that is not ``Transparent'' is called ``Opaque''.
22286 Examples of suitable formats for Transparent copies include plain
22287 @sc{ascii} without markup, Texinfo input format, La@TeX{} input
22288 format, @acronym{SGML} or @acronym{XML} using a publicly available
22289 @acronym{DTD}, and standard-conforming simple @acronym{HTML},
22290 PostScript or @acronym{PDF} designed for human modification. Examples
22291 of transparent image formats include @acronym{PNG}, @acronym{XCF} and
22292 @acronym{JPG}. Opaque formats include proprietary formats that can be
22293 read and edited only by proprietary word processors, @acronym{SGML} or
22294 @acronym{XML} for which the @acronym{DTD} and/or processing tools are
22295 not generally available, and the machine-generated @acronym{HTML},
22296 PostScript or @acronym{PDF} produced by some word processors for
22297 output purposes only.
22299 The ``Title Page'' means, for a printed book, the title page itself,
22300 plus such following pages as are needed to hold, legibly, the material
22301 this License requires to appear in the title page. For works in
22302 formats which do not have any title page as such, ``Title Page'' means
22303 the text near the most prominent appearance of the work's title,
22304 preceding the beginning of the body of the text.
22306 A section ``Entitled XYZ'' means a named subunit of the Document whose
22307 title either is precisely XYZ or contains XYZ in parentheses following
22308 text that translates XYZ in another language. (Here XYZ stands for a
22309 specific section name mentioned below, such as ``Acknowledgements'',
22310 ``Dedications'', ``Endorsements'', or ``History''.) To ``Preserve the Title''
22311 of such a section when you modify the Document means that it remains a
22312 section ``Entitled XYZ'' according to this definition.
22314 The Document may include Warranty Disclaimers next to the notice which
22315 states that this License applies to the Document. These Warranty
22316 Disclaimers are considered to be included by reference in this
22317 License, but only as regards disclaiming warranties: any other
22318 implication that these Warranty Disclaimers may have is void and has
22319 no effect on the meaning of this License.
22324 You may copy and distribute the Document in any medium, either
22325 commercially or noncommercially, provided that this License, the
22326 copyright notices, and the license notice saying this License applies
22327 to the Document are reproduced in all copies, and that you add no other
22328 conditions whatsoever to those of this License. You may not use
22329 technical measures to obstruct or control the reading or further
22330 copying of the copies you make or distribute. However, you may accept
22331 compensation in exchange for copies. If you distribute a large enough
22332 number of copies you must also follow the conditions in section 3.
22334 You may also lend copies, under the same conditions stated above, and
22335 you may publicly display copies.
22338 COPYING IN QUANTITY
22340 If you publish printed copies (or copies in media that commonly have
22341 printed covers) of the Document, numbering more than 100, and the
22342 Document's license notice requires Cover Texts, you must enclose the
22343 copies in covers that carry, clearly and legibly, all these Cover
22344 Texts: Front-Cover Texts on the front cover, and Back-Cover Texts on
22345 the back cover. Both covers must also clearly and legibly identify
22346 you as the publisher of these copies. The front cover must present
22347 the full title with all words of the title equally prominent and
22348 visible. You may add other material on the covers in addition.
22349 Copying with changes limited to the covers, as long as they preserve
22350 the title of the Document and satisfy these conditions, can be treated
22351 as verbatim copying in other respects.
22353 If the required texts for either cover are too voluminous to fit
22354 legibly, you should put the first ones listed (as many as fit
22355 reasonably) on the actual cover, and continue the rest onto adjacent
22358 If you publish or distribute Opaque copies of the Document numbering
22359 more than 100, you must either include a machine-readable Transparent
22360 copy along with each Opaque copy, or state in or with each Opaque copy
22361 a computer-network location from which the general network-using
22362 public has access to download using public-standard network protocols
22363 a complete Transparent copy of the Document, free of added material.
22364 If you use the latter option, you must take reasonably prudent steps,
22365 when you begin distribution of Opaque copies in quantity, to ensure
22366 that this Transparent copy will remain thus accessible at the stated
22367 location until at least one year after the last time you distribute an
22368 Opaque copy (directly or through your agents or retailers) of that
22369 edition to the public.
22371 It is requested, but not required, that you contact the authors of the
22372 Document well before redistributing any large number of copies, to give
22373 them a chance to provide you with an updated version of the Document.
22378 You may copy and distribute a Modified Version of the Document under
22379 the conditions of sections 2 and 3 above, provided that you release
22380 the Modified Version under precisely this License, with the Modified
22381 Version filling the role of the Document, thus licensing distribution
22382 and modification of the Modified Version to whoever possesses a copy
22383 of it. In addition, you must do these things in the Modified Version:
22387 Use in the Title Page (and on the covers, if any) a title distinct
22388 from that of the Document, and from those of previous versions
22389 (which should, if there were any, be listed in the History section
22390 of the Document). You may use the same title as a previous version
22391 if the original publisher of that version gives permission.
22394 List on the Title Page, as authors, one or more persons or entities
22395 responsible for authorship of the modifications in the Modified
22396 Version, together with at least five of the principal authors of the
22397 Document (all of its principal authors, if it has fewer than five),
22398 unless they release you from this requirement.
22401 State on the Title page the name of the publisher of the
22402 Modified Version, as the publisher.
22405 Preserve all the copyright notices of the Document.
22408 Add an appropriate copyright notice for your modifications
22409 adjacent to the other copyright notices.
22412 Include, immediately after the copyright notices, a license notice
22413 giving the public permission to use the Modified Version under the
22414 terms of this License, in the form shown in the Addendum below.
22417 Preserve in that license notice the full lists of Invariant Sections
22418 and required Cover Texts given in the Document's license notice.
22421 Include an unaltered copy of this License.
22424 Preserve the section Entitled ``History'', Preserve its Title, and add
22425 to it an item stating at least the title, year, new authors, and
22426 publisher of the Modified Version as given on the Title Page. If
22427 there is no section Entitled ``History'' in the Document, create one
22428 stating the title, year, authors, and publisher of the Document as
22429 given on its Title Page, then add an item describing the Modified
22430 Version as stated in the previous sentence.
22433 Preserve the network location, if any, given in the Document for
22434 public access to a Transparent copy of the Document, and likewise
22435 the network locations given in the Document for previous versions
22436 it was based on. These may be placed in the ``History'' section.
22437 You may omit a network location for a work that was published at
22438 least four years before the Document itself, or if the original
22439 publisher of the version it refers to gives permission.
22442 For any section Entitled ``Acknowledgements'' or ``Dedications'', Preserve
22443 the Title of the section, and preserve in the section all the
22444 substance and tone of each of the contributor acknowledgements and/or
22445 dedications given therein.
22448 Preserve all the Invariant Sections of the Document,
22449 unaltered in their text and in their titles. Section numbers
22450 or the equivalent are not considered part of the section titles.
22453 Delete any section Entitled ``Endorsements''. Such a section
22454 may not be included in the Modified Version.
22457 Do not retitle any existing section to be Entitled ``Endorsements'' or
22458 to conflict in title with any Invariant Section.
22461 Preserve any Warranty Disclaimers.
22464 If the Modified Version includes new front-matter sections or
22465 appendices that qualify as Secondary Sections and contain no material
22466 copied from the Document, you may at your option designate some or all
22467 of these sections as invariant. To do this, add their titles to the
22468 list of Invariant Sections in the Modified Version's license notice.
22469 These titles must be distinct from any other section titles.
22471 You may add a section Entitled ``Endorsements'', provided it contains
22472 nothing but endorsements of your Modified Version by various
22473 parties---for example, statements of peer review or that the text has
22474 been approved by an organization as the authoritative definition of a
22477 You may add a passage of up to five words as a Front-Cover Text, and a
22478 passage of up to 25 words as a Back-Cover Text, to the end of the list
22479 of Cover Texts in the Modified Version. Only one passage of
22480 Front-Cover Text and one of Back-Cover Text may be added by (or
22481 through arrangements made by) any one entity. If the Document already
22482 includes a cover text for the same cover, previously added by you or
22483 by arrangement made by the same entity you are acting on behalf of,
22484 you may not add another; but you may replace the old one, on explicit
22485 permission from the previous publisher that added the old one.
22487 The author(s) and publisher(s) of the Document do not by this License
22488 give permission to use their names for publicity for or to assert or
22489 imply endorsement of any Modified Version.
22492 COMBINING DOCUMENTS
22494 You may combine the Document with other documents released under this
22495 License, under the terms defined in section 4 above for modified
22496 versions, provided that you include in the combination all of the
22497 Invariant Sections of all of the original documents, unmodified, and
22498 list them all as Invariant Sections of your combined work in its
22499 license notice, and that you preserve all their Warranty Disclaimers.
22501 The combined work need only contain one copy of this License, and
22502 multiple identical Invariant Sections may be replaced with a single
22503 copy. If there are multiple Invariant Sections with the same name but
22504 different contents, make the title of each such section unique by
22505 adding at the end of it, in parentheses, the name of the original
22506 author or publisher of that section if known, or else a unique number.
22507 Make the same adjustment to the section titles in the list of
22508 Invariant Sections in the license notice of the combined work.
22510 In the combination, you must combine any sections Entitled ``History''
22511 in the various original documents, forming one section Entitled
22512 ``History''; likewise combine any sections Entitled ``Acknowledgements'',
22513 and any sections Entitled ``Dedications''. You must delete all
22514 sections Entitled ``Endorsements.''
22517 COLLECTIONS OF DOCUMENTS
22519 You may make a collection consisting of the Document and other documents
22520 released under this License, and replace the individual copies of this
22521 License in the various documents with a single copy that is included in
22522 the collection, provided that you follow the rules of this License for
22523 verbatim copying of each of the documents in all other respects.
22525 You may extract a single document from such a collection, and distribute
22526 it individually under this License, provided you insert a copy of this
22527 License into the extracted document, and follow this License in all
22528 other respects regarding verbatim copying of that document.
22531 AGGREGATION WITH INDEPENDENT WORKS
22533 A compilation of the Document or its derivatives with other separate
22534 and independent documents or works, in or on a volume of a storage or
22535 distribution medium, is called an ``aggregate'' if the copyright
22536 resulting from the compilation is not used to limit the legal rights
22537 of the compilation's users beyond what the individual works permit.
22538 When the Document is included in an aggregate, this License does not
22539 apply to the other works in the aggregate which are not themselves
22540 derivative works of the Document.
22542 If the Cover Text requirement of section 3 is applicable to these
22543 copies of the Document, then if the Document is less than one half of
22544 the entire aggregate, the Document's Cover Texts may be placed on
22545 covers that bracket the Document within the aggregate, or the
22546 electronic equivalent of covers if the Document is in electronic form.
22547 Otherwise they must appear on printed covers that bracket the whole
22553 Translation is considered a kind of modification, so you may
22554 distribute translations of the Document under the terms of section 4.
22555 Replacing Invariant Sections with translations requires special
22556 permission from their copyright holders, but you may include
22557 translations of some or all Invariant Sections in addition to the
22558 original versions of these Invariant Sections. You may include a
22559 translation of this License, and all the license notices in the
22560 Document, and any Warranty Disclaimers, provided that you also include
22561 the original English version of this License and the original versions
22562 of those notices and disclaimers. In case of a disagreement between
22563 the translation and the original version of this License or a notice
22564 or disclaimer, the original version will prevail.
22566 If a section in the Document is Entitled ``Acknowledgements'',
22567 ``Dedications'', or ``History'', the requirement (section 4) to Preserve
22568 its Title (section 1) will typically require changing the actual
22574 You may not copy, modify, sublicense, or distribute the Document except
22575 as expressly provided for under this License. Any other attempt to
22576 copy, modify, sublicense or distribute the Document is void, and will
22577 automatically terminate your rights under this License. However,
22578 parties who have received copies, or rights, from you under this
22579 License will not have their licenses terminated so long as such
22580 parties remain in full compliance.
22583 FUTURE REVISIONS OF THIS LICENSE
22585 The Free Software Foundation may publish new, revised versions
22586 of the GNU Free Documentation License from time to time. Such new
22587 versions will be similar in spirit to the present version, but may
22588 differ in detail to address new problems or concerns. See
22589 @uref{http://www.gnu.org/copyleft/}.
22591 Each version of the License is given a distinguishing version number.
22592 If the Document specifies that a particular numbered version of this
22593 License ``or any later version'' applies to it, you have the option of
22594 following the terms and conditions either of that specified version or
22595 of any later version that has been published (not as a draft) by the
22596 Free Software Foundation. If the Document does not specify a version
22597 number of this License, you may choose any version ever published (not
22598 as a draft) by the Free Software Foundation.
22602 @appendixsubsec ADDENDUM: How to use this License for your documents
22604 To use this License in a document you have written, include a copy of
22605 the License in the document and put the following copyright and
22606 license notices just after the title page:
22610 Copyright (C) @var{year} @var{your name}.
22611 Permission is granted to copy, distribute and/or modify this document
22612 under the terms of the GNU Free Documentation License, Version 1.2
22613 or any later version published by the Free Software Foundation;
22614 with no Invariant Sections, no Front-Cover Texts, and no Back-Cover Texts.
22615 A copy of the license is included in the section entitled ``GNU
22616 Free Documentation License''.
22620 If you have Invariant Sections, Front-Cover Texts and Back-Cover Texts,
22621 replace the ``with...Texts.'' line with this:
22625 with the Invariant Sections being @var{list their titles}, with
22626 the Front-Cover Texts being @var{list}, and with the Back-Cover Texts
22631 If you have Invariant Sections without Cover Texts, or some other
22632 combination of the three, merge those two alternatives to suit the
22635 If your document contains nontrivial examples of program code, we
22636 recommend releasing these examples in parallel under your choice of
22637 free software license, such as the GNU General Public License,
22638 to permit their use in free software.
22640 @node Index, About the Author, GNU Free Documentation License, Top
22641 @comment node-name, next, previous, up
22645 MENU ENTRY: NODE NAME.
22651 @c Place biographical information on right-hand (verso) page
22655 \par\vfill\supereject
22656 \global\evenheadline={\hfil} \global\evenfootline={\hfil}
22657 \global\oddheadline={\hfil} \global\oddfootline={\hfil}
22660 \par\vfill\supereject
22661 \par\vfill\supereject
22662 \global\evenheadline={\hfil} \global\evenfootline={\hfil}
22663 \global\oddheadline={\hfil} \global\oddfootline={\hfil}
22672 @c ================ Biographical information ================
22676 @center About the Author
22681 @node About the Author, , Index, Top
22682 @unnumbered About the Author
22686 Robert J. Chassell has worked with GNU Emacs since 1985. He writes
22687 and edits, teaches Emacs and Emacs Lisp, and speaks throughout the
22688 world on software freedom. Chassell was a founding Director and
22689 Treasurer of the Free Software Foundation, Inc. He is co-author of
22690 the @cite{Texinfo} manual, and has edited more than a dozen other
22691 books. He graduated from Cambridge University, in England. He has an
22692 abiding interest in social and economic history and flies his own
22699 @c Prevent page number on blank verso, so eject it first.
22701 \par\vfill\supereject
22706 @evenheading @thispage @| @| @thistitle
22707 @oddheading @| @| @thispage
22713 arch-tag: da1a2154-531f-43a8-8e33-fc7faad10acf