]> code.delx.au - gnu-emacs/blob - doc/lispref/tips.texi
Convert consecutive FSF copyright years to ranges.
[gnu-emacs] / doc / lispref / tips.texi
1 @c -*-texinfo-*-
2 @c This is part of the GNU Emacs Lisp Reference Manual.
3 @c Copyright (C) 1990-1993, 1995, 1998-1999, 2001-2011
4 @c Free Software Foundation, Inc.
5 @c See the file elisp.texi for copying conditions.
6 @setfilename ../../info/tips
7 @node Tips, GNU Emacs Internals, GPL, Top
8 @appendix Tips and Conventions
9 @cindex tips for writing Lisp
10 @cindex standards of coding style
11 @cindex coding standards
12
13 This chapter describes no additional features of Emacs Lisp. Instead
14 it gives advice on making effective use of the features described in the
15 previous chapters, and describes conventions Emacs Lisp programmers
16 should follow.
17
18 You can automatically check some of the conventions described below by
19 running the command @kbd{M-x checkdoc RET} when visiting a Lisp file.
20 It cannot check all of the conventions, and not all the warnings it
21 gives necessarily correspond to problems, but it is worth examining them
22 all.
23
24 @menu
25 * Coding Conventions:: Conventions for clean and robust programs.
26 * Key Binding Conventions:: Which keys should be bound by which programs.
27 * Programming Tips:: Making Emacs code fit smoothly in Emacs.
28 * Compilation Tips:: Making compiled code run fast.
29 * Warning Tips:: Turning off compiler warnings.
30 * Documentation Tips:: Writing readable documentation strings.
31 * Comment Tips:: Conventions for writing comments.
32 * Library Headers:: Standard headers for library packages.
33 @end menu
34
35 @node Coding Conventions
36 @section Emacs Lisp Coding Conventions
37
38 @cindex coding conventions in Emacs Lisp
39 Here are conventions that you should follow when writing Emacs Lisp
40 code intended for widespread use:
41
42 @itemize @bullet
43 @item
44 Simply loading a package should not change Emacs's editing behavior.
45 Include a command or commands to enable and disable the feature,
46 or to invoke it.
47
48 This convention is mandatory for any file that includes custom
49 definitions. If fixing such a file to follow this convention requires
50 an incompatible change, go ahead and make the incompatible change;
51 don't postpone it.
52
53 @item
54 You should choose a short word to distinguish your program from other
55 Lisp programs. The names of all global variables, constants, and
56 functions in your program should begin with that chosen prefix.
57 Separate the prefix from the rest of the name with a hyphen, @samp{-}.
58 This practice helps avoid name conflicts, since all global variables
59 in Emacs Lisp share the same name space, and all functions share
60 another name space@footnote{The benefits of a Common Lisp-style
61 package system are considered not to outweigh the costs.}
62
63 Occasionally, for a command name intended for users to use, it is more
64 convenient if some words come before the package's name prefix. And
65 constructs that define functions, variables, etc., work better if they
66 start with @samp{defun} or @samp{defvar}, so put the name prefix later
67 on in the name.
68
69 This recommendation applies even to names for traditional Lisp
70 primitives that are not primitives in Emacs Lisp---such as
71 @code{copy-list}. Believe it or not, there is more than one plausible
72 way to define @code{copy-list}. Play it safe; append your name prefix
73 to produce a name like @code{foo-copy-list} or @code{mylib-copy-list}
74 instead.
75
76 If you write a function that you think ought to be added to Emacs under
77 a certain name, such as @code{twiddle-files}, don't call it by that name
78 in your program. Call it @code{mylib-twiddle-files} in your program,
79 and send mail to @samp{bug-gnu-emacs@@gnu.org} suggesting we add
80 it to Emacs. If and when we do, we can change the name easily enough.
81
82 If one prefix is insufficient, your package can use two or three
83 alternative common prefixes, so long as they make sense.
84
85 @item
86 Put a call to @code{provide} at the end of each separate Lisp file.
87 @xref{Named Features}.
88
89 @item
90 If a file requires certain other Lisp programs to be loaded
91 beforehand, then the comments at the beginning of the file should say
92 so. Also, use @code{require} to make sure they are loaded.
93 @xref{Named Features}.
94
95 @item
96 If a file @var{foo} uses a macro defined in another file @var{bar},
97 but does not use any functions or variables defined in @var{bar}, then
98 @var{foo} should contain the following expression:
99
100 @example
101 (eval-when-compile (require '@var{bar}))
102 @end example
103
104 @noindent
105 This tells Emacs to load @var{bar} just before byte-compiling
106 @var{foo}, so that the macro definition is available during
107 compilation. Using @code{eval-when-compile} avoids loading @var{bar}
108 when the compiled version of @var{foo} is @emph{used}. It should be
109 called before the first use of the macro in the file. @xref{Compiling
110 Macros}.
111
112 @item
113 Please don't require the @code{cl} package of Common Lisp extensions at
114 run time. Use of this package is optional, and it is not part of the
115 standard Emacs namespace. If your package loads @code{cl} at run time,
116 that could cause name clashes for users who don't use that package.
117
118 However, there is no problem with using the @code{cl} package at
119 compile time, with @code{(eval-when-compile (require 'cl))}. That's
120 sufficient for using the macros in the @code{cl} package, because the
121 compiler expands them before generating the byte-code.
122
123 @item
124 When defining a major mode, please follow the major mode
125 conventions. @xref{Major Mode Conventions}.
126
127 @item
128 When defining a minor mode, please follow the minor mode
129 conventions. @xref{Minor Mode Conventions}.
130
131 @item
132 If the purpose of a function is to tell you whether a certain
133 condition is true or false, give the function a name that ends in
134 @samp{p} (which stands for ``predicate''). If the name is one word,
135 add just @samp{p}; if the name is multiple words, add @samp{-p}.
136 Examples are @code{framep} and @code{frame-live-p}.
137
138 @item
139 If the purpose of a variable is to store a single function, give it a
140 name that ends in @samp{-function}. If the purpose of a variable is
141 to store a list of functions (i.e., the variable is a hook), please
142 follow the naming conventions for hooks. @xref{Hooks}.
143
144 @item
145 @cindex unloading packages, preparing for
146 If loading the file adds functions to hooks, define a function
147 @code{@var{feature}-unload-hook}, where @var{feature} is the name of
148 the feature the package provides, and make it undo any such changes.
149 Using @code{unload-feature} to unload the file will run this function.
150 @xref{Unloading}.
151
152 @item
153 It is a bad idea to define aliases for the Emacs primitives. Normally
154 you should use the standard names instead. The case where an alias
155 may be useful is where it facilitates backwards compatibility or
156 portability.
157
158 @item
159 If a package needs to define an alias or a new function for
160 compatibility with some other version of Emacs, name it with the package
161 prefix, not with the raw name with which it occurs in the other version.
162 Here is an example from Gnus, which provides many examples of such
163 compatibility issues.
164
165 @example
166 (defalias 'gnus-point-at-bol
167 (if (fboundp 'point-at-bol)
168 'point-at-bol
169 'line-beginning-position))
170 @end example
171
172 @item
173 Redefining or advising an Emacs primitive is a bad idea. It may do
174 the right thing for a particular program, but there is no telling what
175 other programs might break as a result.
176
177 @item
178 It is likewise a bad idea for one Lisp package to advise a function in
179 another Lisp package (@pxref{Advising Functions}).
180
181 @item
182 Avoid using @code{eval-after-load} in libraries and packages
183 (@pxref{Hooks for Loading}). This feature is meant for personal
184 customizations; using it in a Lisp program is unclean, because it
185 modifies the behavior of another Lisp file in a way that's not visible
186 in that file. This is an obstacle for debugging, much like advising a
187 function in the other package.
188
189 @item
190 If a file does replace any of the standard functions or library
191 programs of Emacs, prominent comments at the beginning of the file
192 should say which functions are replaced, and how the behavior of the
193 replacements differs from that of the originals.
194
195 @item
196 Constructs that define a function or variable should be macros,
197 not functions, and their names should start with @samp{def}.
198
199 @item
200 A macro that defines a function or variable should have a name that
201 starts with @samp{define-}. The macro should receive the name to be
202 defined as the first argument. That will help various tools find the
203 definition automatically. Avoid constructing the names in the macro
204 itself, since that would confuse these tools.
205
206 @item
207 Please keep the names of your Emacs Lisp source files to 13 characters
208 or less. This way, if the files are compiled, the compiled files' names
209 will be 14 characters or less, which is short enough to fit on all kinds
210 of Unix systems.
211
212 @item
213 In some other systems there is a convention of choosing variable names
214 that begin and end with @samp{*}. We don't use that convention in Emacs
215 Lisp, so please don't use it in your programs. (Emacs uses such names
216 only for special-purpose buffers.) The users will find Emacs more
217 coherent if all libraries use the same conventions.
218
219 @item
220 If your program contains non-ASCII characters in string or character
221 constants, you should make sure Emacs always decodes these characters
222 the same way, regardless of the user's settings. The easiest way to
223 do this is to use the coding system @code{utf-8-emacs} (@pxref{Coding
224 System Basics}), and specify that coding in the @samp{-*-} line or the
225 local variables list. @xref{File variables, , Local Variables in
226 Files, emacs, The GNU Emacs Manual}.
227
228 @example
229 ;; XXX.el -*- coding: utf-8-emacs; -*-
230 @end example
231
232 @item
233 Indent each function with @kbd{C-M-q} (@code{indent-sexp}) using the
234 default indentation parameters.
235
236 @item
237 Don't make a habit of putting close-parentheses on lines by
238 themselves; Lisp programmers find this disconcerting.
239
240 @item
241 Please put a copyright notice and copying permission notice on the
242 file if you distribute copies. Use a notice like this one:
243
244 @smallexample
245 ;; Copyright (C) @var{year} @var{name}
246
247 ;; This program is free software: you can redistribute it and/or
248 ;; modify it under the terms of the GNU General Public License as
249 ;; published by the Free Software Foundation, either version 3 of
250 ;; the License, or (at your option) any later version.
251
252 ;; This program is distributed in the hope that it will be useful,
253 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
254 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
255 ;; GNU General Public License for more details.
256
257 ;; You should have received a copy of the GNU General Public License
258 ;; along with this program. If not, see <http://www.gnu.org/licenses/>.
259 @end smallexample
260
261 If you have signed papers to assign the copyright to the Foundation,
262 then use @samp{Free Software Foundation, Inc.} as @var{name}.
263 Otherwise, use your name. @xref{Library Headers}.
264 @end itemize
265
266 @node Key Binding Conventions
267 @section Key Binding Conventions
268 @cindex key binding, conventions for
269
270 @itemize @bullet
271 @item
272 @cindex mouse-2
273 @cindex references, following
274 Many special major modes, like Dired, Info, Compilation, and Occur,
275 are designed to handle read-only text that contains @dfn{hyper-links}.
276 Such a major mode should redefine @kbd{mouse-2} and @key{RET} to
277 follow the links. It should also set up a @code{follow-link}
278 condition, so that the link obeys @code{mouse-1-click-follows-link}.
279 @xref{Clickable Text}. @xref{Buttons}, for an easy method of
280 implementing such clickable links.
281
282 @item
283 @cindex reserved keys
284 @cindex keys, reserved
285 Don't define @kbd{C-c @var{letter}} as a key in Lisp programs.
286 Sequences consisting of @kbd{C-c} and a letter (either upper or lower
287 case) are reserved for users; they are the @strong{only} sequences
288 reserved for users, so do not block them.
289
290 Changing all the Emacs major modes to respect this convention was a
291 lot of work; abandoning this convention would make that work go to
292 waste, and inconvenience users. Please comply with it.
293
294 @item
295 Function keys @key{F5} through @key{F9} without modifier keys are
296 also reserved for users to define.
297
298 @item
299 Sequences consisting of @kbd{C-c} followed by a control character or a
300 digit are reserved for major modes.
301
302 @item
303 Sequences consisting of @kbd{C-c} followed by @kbd{@{}, @kbd{@}},
304 @kbd{<}, @kbd{>}, @kbd{:} or @kbd{;} are also reserved for major modes.
305
306 @item
307 Sequences consisting of @kbd{C-c} followed by any other punctuation
308 character are allocated for minor modes. Using them in a major mode is
309 not absolutely prohibited, but if you do that, the major mode binding
310 may be shadowed from time to time by minor modes.
311
312 @item
313 Don't bind @kbd{C-h} following any prefix character (including
314 @kbd{C-c}). If you don't bind @kbd{C-h}, it is automatically
315 available as a help character for listing the subcommands of the
316 prefix character.
317
318 @item
319 Don't bind a key sequence ending in @key{ESC} except following another
320 @key{ESC}. (That is, it is OK to bind a sequence ending in
321 @kbd{@key{ESC} @key{ESC}}.)
322
323 The reason for this rule is that a non-prefix binding for @key{ESC} in
324 any context prevents recognition of escape sequences as function keys in
325 that context.
326
327 @item
328 Anything which acts like a temporary mode or state which the user can
329 enter and leave should define @kbd{@key{ESC} @key{ESC}} or
330 @kbd{@key{ESC} @key{ESC} @key{ESC}} as a way to escape.
331
332 For a state which accepts ordinary Emacs commands, or more generally any
333 kind of state in which @key{ESC} followed by a function key or arrow key
334 is potentially meaningful, then you must not define @kbd{@key{ESC}
335 @key{ESC}}, since that would preclude recognizing an escape sequence
336 after @key{ESC}. In these states, you should define @kbd{@key{ESC}
337 @key{ESC} @key{ESC}} as the way to escape. Otherwise, define
338 @kbd{@key{ESC} @key{ESC}} instead.
339 @end itemize
340
341 @node Programming Tips
342 @section Emacs Programming Tips
343 @cindex programming conventions
344
345 Following these conventions will make your program fit better
346 into Emacs when it runs.
347
348 @itemize @bullet
349 @item
350 Don't use @code{next-line} or @code{previous-line} in programs; nearly
351 always, @code{forward-line} is more convenient as well as more
352 predictable and robust. @xref{Text Lines}.
353
354 @item
355 Don't call functions that set the mark, unless setting the mark is one
356 of the intended features of your program. The mark is a user-level
357 feature, so it is incorrect to change the mark except to supply a value
358 for the user's benefit. @xref{The Mark}.
359
360 In particular, don't use any of these functions:
361
362 @itemize @bullet
363 @item
364 @code{beginning-of-buffer}, @code{end-of-buffer}
365 @item
366 @code{replace-string}, @code{replace-regexp}
367 @item
368 @code{insert-file}, @code{insert-buffer}
369 @end itemize
370
371 If you just want to move point, or replace a certain string, or insert
372 a file or buffer's contents, without any of the other features
373 intended for interactive users, you can replace these functions with
374 one or two lines of simple Lisp code.
375
376 @item
377 Use lists rather than vectors, except when there is a particular reason
378 to use a vector. Lisp has more facilities for manipulating lists than
379 for vectors, and working with lists is usually more convenient.
380
381 Vectors are advantageous for tables that are substantial in size and are
382 accessed in random order (not searched front to back), provided there is
383 no need to insert or delete elements (only lists allow that).
384
385 @item
386 The recommended way to show a message in the echo area is with
387 the @code{message} function, not @code{princ}. @xref{The Echo Area}.
388
389 @item
390 When you encounter an error condition, call the function @code{error}
391 (or @code{signal}). The function @code{error} does not return.
392 @xref{Signaling Errors}.
393
394 Don't use @code{message}, @code{throw}, @code{sleep-for}, or
395 @code{beep} to report errors.
396
397 @item
398 An error message should start with a capital letter but should not end
399 with a period.
400
401 @item
402 A question asked in the minibuffer with @code{y-or-n-p} or
403 @code{yes-or-no-p} should start with a capital letter and end with
404 @samp{? }.
405
406 @item
407 When you mention a default value in a minibuffer prompt,
408 put it and the word @samp{default} inside parentheses.
409 It should look like this:
410
411 @example
412 Enter the answer (default 42):
413 @end example
414
415 @item
416 In @code{interactive}, if you use a Lisp expression to produce a list
417 of arguments, don't try to provide the ``correct'' default values for
418 region or position arguments. Instead, provide @code{nil} for those
419 arguments if they were not specified, and have the function body
420 compute the default value when the argument is @code{nil}. For
421 instance, write this:
422
423 @example
424 (defun foo (pos)
425 (interactive
426 (list (if @var{specified} @var{specified-pos})))
427 (unless pos (setq pos @var{default-pos}))
428 ...)
429 @end example
430
431 @noindent
432 rather than this:
433
434 @example
435 (defun foo (pos)
436 (interactive
437 (list (if @var{specified} @var{specified-pos}
438 @var{default-pos})))
439 ...)
440 @end example
441
442 @noindent
443 This is so that repetition of the command will recompute
444 these defaults based on the current circumstances.
445
446 You do not need to take such precautions when you use interactive
447 specs @samp{d}, @samp{m} and @samp{r}, because they make special
448 arrangements to recompute the argument values on repetition of the
449 command.
450
451 @item
452 Many commands that take a long time to execute display a message that
453 says something like @samp{Operating...} when they start, and change it
454 to @samp{Operating...done} when they finish. Please keep the style of
455 these messages uniform: @emph{no} space around the ellipsis, and
456 @emph{no} period after @samp{done}. @xref{Progress}, for an easy way
457 to generate such messages.
458
459 @item
460 Try to avoid using recursive edits. Instead, do what the Rmail @kbd{e}
461 command does: use a new local keymap that contains one command defined
462 to switch back to the old local keymap. Or do what the
463 @code{edit-options} command does: switch to another buffer and let the
464 user switch back at will. @xref{Recursive Editing}.
465 @end itemize
466
467 @node Compilation Tips
468 @section Tips for Making Compiled Code Fast
469 @cindex execution speed
470 @cindex speedups
471
472 Here are ways of improving the execution speed of byte-compiled
473 Lisp programs.
474
475 @itemize @bullet
476 @item
477 @cindex profiling
478 @cindex timing programs
479 @cindex @file{elp.el}
480 Profile your program with the @file{elp} library. See the file
481 @file{elp.el} for instructions.
482
483 @item
484 @cindex @file{benchmark.el}
485 @cindex benchmarking
486 Check the speed of individual Emacs Lisp forms using the
487 @file{benchmark} library. See the functions @code{benchmark-run} and
488 @code{benchmark-run-compiled} in @file{benchmark.el}.
489
490 @item
491 Use iteration rather than recursion whenever possible.
492 Function calls are slow in Emacs Lisp even when a compiled function
493 is calling another compiled function.
494
495 @item
496 Using the primitive list-searching functions @code{memq}, @code{member},
497 @code{assq}, or @code{assoc} is even faster than explicit iteration. It
498 can be worth rearranging a data structure so that one of these primitive
499 search functions can be used.
500
501 @item
502 Certain built-in functions are handled specially in byte-compiled code,
503 avoiding the need for an ordinary function call. It is a good idea to
504 use these functions rather than alternatives. To see whether a function
505 is handled specially by the compiler, examine its @code{byte-compile}
506 property. If the property is non-@code{nil}, then the function is
507 handled specially.
508
509 For example, the following input will show you that @code{aref} is
510 compiled specially (@pxref{Array Functions}):
511
512 @example
513 @group
514 (get 'aref 'byte-compile)
515 @result{} byte-compile-two-args
516 @end group
517 @end example
518
519 @item
520 If calling a small function accounts for a substantial part of your
521 program's running time, make the function inline. This eliminates
522 the function call overhead. Since making a function inline reduces
523 the flexibility of changing the program, don't do it unless it gives
524 a noticeable speedup in something slow enough that users care about
525 the speed. @xref{Inline Functions}.
526 @end itemize
527
528 @node Warning Tips
529 @section Tips for Avoiding Compiler Warnings
530 @cindex byte compiler warnings, how to avoid
531
532 @itemize @bullet
533 @item
534 Try to avoid compiler warnings about undefined free variables, by adding
535 dummy @code{defvar} definitions for these variables, like this:
536
537 @example
538 (defvar foo)
539 @end example
540
541 Such a definition has no effect except to tell the compiler
542 not to warn about uses of the variable @code{foo} in this file.
543
544 @item
545 If you use many functions and variables from a certain file, you can
546 add a @code{require} for that package to avoid compilation warnings
547 for them. For instance,
548
549 @example
550 (eval-when-compile
551 (require 'foo))
552 @end example
553
554 @item
555 If you bind a variable in one function, and use it or set it in
556 another function, the compiler warns about the latter function unless
557 the variable has a definition. But adding a definition would be
558 unclean if the variable has a short name, since Lisp packages should
559 not define short variable names. The right thing to do is to rename
560 this variable to start with the name prefix used for the other
561 functions and variables in your package.
562
563 @item
564 The last resort for avoiding a warning, when you want to do something
565 that usually is a mistake but it's not a mistake in this one case,
566 is to put a call to @code{with-no-warnings} around it.
567 @end itemize
568
569 @node Documentation Tips
570 @section Tips for Documentation Strings
571 @cindex documentation strings, conventions and tips
572
573 @findex checkdoc-minor-mode
574 Here are some tips and conventions for the writing of documentation
575 strings. You can check many of these conventions by running the command
576 @kbd{M-x checkdoc-minor-mode}.
577
578 @itemize @bullet
579 @item
580 Every command, function, or variable intended for users to know about
581 should have a documentation string.
582
583 @item
584 An internal variable or subroutine of a Lisp program might as well have
585 a documentation string. In earlier Emacs versions, you could save space
586 by using a comment instead of a documentation string, but that is no
587 longer the case---documentation strings now take up very little space in
588 a running Emacs.
589
590 @item
591 Format the documentation string so that it fits in an Emacs window on an
592 80-column screen. It is a good idea for most lines to be no wider than
593 60 characters. The first line should not be wider than 67 characters
594 or it will look bad in the output of @code{apropos}.
595
596 You can fill the text if that looks good. However, rather than blindly
597 filling the entire documentation string, you can often make it much more
598 readable by choosing certain line breaks with care. Use blank lines
599 between topics if the documentation string is long.
600
601 @item
602 The first line of the documentation string should consist of one or two
603 complete sentences that stand on their own as a summary. @kbd{M-x
604 apropos} displays just the first line, and if that line's contents don't
605 stand on their own, the result looks bad. In particular, start the
606 first line with a capital letter and end with a period.
607
608 For a function, the first line should briefly answer the question,
609 ``What does this function do?'' For a variable, the first line should
610 briefly answer the question, ``What does this value mean?''
611
612 Don't limit the documentation string to one line; use as many lines as
613 you need to explain the details of how to use the function or
614 variable. Please use complete sentences for the rest of the text too.
615
616 @item
617 When the user tries to use a disabled command, Emacs displays just the
618 first paragraph of its documentation string---everything through the
619 first blank line. If you wish, you can choose which information to
620 include before the first blank line so as to make this display useful.
621
622 @item
623 The first line should mention all the important arguments of the
624 function, and should mention them in the order that they are written
625 in a function call. If the function has many arguments, then it is
626 not feasible to mention them all in the first line; in that case, the
627 first line should mention the first few arguments, including the most
628 important arguments.
629
630 @item
631 When a function's documentation string mentions the value of an argument
632 of the function, use the argument name in capital letters as if it were
633 a name for that value. Thus, the documentation string of the function
634 @code{eval} refers to its second argument as @samp{FORM}, because the
635 actual argument name is @code{form}:
636
637 @example
638 Evaluate FORM and return its value.
639 @end example
640
641 Also write metasyntactic variables in capital letters, such as when you
642 show the decomposition of a list or vector into subunits, some of which
643 may vary. @samp{KEY} and @samp{VALUE} in the following example
644 illustrate this practice:
645
646 @example
647 The argument TABLE should be an alist whose elements
648 have the form (KEY . VALUE). Here, KEY is ...
649 @end example
650
651 @item
652 Never change the case of a Lisp symbol when you mention it in a doc
653 string. If the symbol's name is @code{foo}, write ``foo,'' not
654 ``Foo'' (which is a different symbol).
655
656 This might appear to contradict the policy of writing function
657 argument values, but there is no real contradiction; the argument
658 @emph{value} is not the same thing as the @emph{symbol} which the
659 function uses to hold the value.
660
661 If this puts a lower-case letter at the beginning of a sentence
662 and that annoys you, rewrite the sentence so that the symbol
663 is not at the start of it.
664
665 @item
666 Do not start or end a documentation string with whitespace.
667
668 @item
669 @strong{Do not} indent subsequent lines of a documentation string so
670 that the text is lined up in the source code with the text of the first
671 line. This looks nice in the source code, but looks bizarre when users
672 view the documentation. Remember that the indentation before the
673 starting double-quote is not part of the string!
674
675 @anchor{Docstring hyperlinks}
676 @item
677 @iftex
678 When a documentation string refers to a Lisp symbol, write it as it
679 would be printed (which usually means in lower case), with single-quotes
680 around it. For example: @samp{`lambda'}. There are two exceptions:
681 write @code{t} and @code{nil} without single-quotes.
682 @end iftex
683 @ifnottex
684 When a documentation string refers to a Lisp symbol, write it as it
685 would be printed (which usually means in lower case), with single-quotes
686 around it. For example: @samp{lambda}. There are two exceptions: write
687 t and nil without single-quotes. (In this manual, we use a different
688 convention, with single-quotes for all symbols.)
689 @end ifnottex
690
691 @cindex hyperlinks in documentation strings
692 Help mode automatically creates a hyperlink when a documentation string
693 uses a symbol name inside single quotes, if the symbol has either a
694 function or a variable definition. You do not need to do anything
695 special to make use of this feature. However, when a symbol has both a
696 function definition and a variable definition, and you want to refer to
697 just one of them, you can specify which one by writing one of the words
698 @samp{variable}, @samp{option}, @samp{function}, or @samp{command},
699 immediately before the symbol name. (Case makes no difference in
700 recognizing these indicator words.) For example, if you write
701
702 @example
703 This function sets the variable `buffer-file-name'.
704 @end example
705
706 @noindent
707 then the hyperlink will refer only to the variable documentation of
708 @code{buffer-file-name}, and not to its function documentation.
709
710 If a symbol has a function definition and/or a variable definition, but
711 those are irrelevant to the use of the symbol that you are documenting,
712 you can write the words @samp{symbol} or @samp{program} before the
713 symbol name to prevent making any hyperlink. For example,
714
715 @example
716 If the argument KIND-OF-RESULT is the symbol `list',
717 this function returns a list of all the objects
718 that satisfy the criterion.
719 @end example
720
721 @noindent
722 does not make a hyperlink to the documentation, irrelevant here, of the
723 function @code{list}.
724
725 Normally, no hyperlink is made for a variable without variable
726 documentation. You can force a hyperlink for such variables by
727 preceding them with one of the words @samp{variable} or
728 @samp{option}.
729
730 Hyperlinks for faces are only made if the face name is preceded or
731 followed by the word @samp{face}. In that case, only the face
732 documentation will be shown, even if the symbol is also defined as a
733 variable or as a function.
734
735 To make a hyperlink to Info documentation, write the name of the Info
736 node (or anchor) in single quotes, preceded by @samp{info node},
737 @samp{Info node}, @samp{info anchor} or @samp{Info anchor}. The Info
738 file name defaults to @samp{emacs}. For example,
739
740 @smallexample
741 See Info node `Font Lock' and Info node `(elisp)Font Lock Basics'.
742 @end smallexample
743
744 Finally, to create a hyperlink to URLs, write the URL in single
745 quotes, preceded by @samp{URL}. For example,
746
747 @smallexample
748 The home page for the GNU project has more information (see URL
749 `http://www.gnu.org/').
750 @end smallexample
751
752 @item
753 Don't write key sequences directly in documentation strings. Instead,
754 use the @samp{\\[@dots{}]} construct to stand for them. For example,
755 instead of writing @samp{C-f}, write the construct
756 @samp{\\[forward-char]}. When Emacs displays the documentation string,
757 it substitutes whatever key is currently bound to @code{forward-char}.
758 (This is normally @samp{C-f}, but it may be some other character if the
759 user has moved key bindings.) @xref{Keys in Documentation}.
760
761 @item
762 In documentation strings for a major mode, you will want to refer to the
763 key bindings of that mode's local map, rather than global ones.
764 Therefore, use the construct @samp{\\<@dots{}>} once in the
765 documentation string to specify which key map to use. Do this before
766 the first use of @samp{\\[@dots{}]}. The text inside the
767 @samp{\\<@dots{}>} should be the name of the variable containing the
768 local keymap for the major mode.
769
770 It is not practical to use @samp{\\[@dots{}]} very many times, because
771 display of the documentation string will become slow. So use this to
772 describe the most important commands in your major mode, and then use
773 @samp{\\@{@dots{}@}} to display the rest of the mode's keymap.
774
775 @item
776 For consistency, phrase the verb in the first sentence of a function's
777 documentation string as an imperative---for instance, use ``Return the
778 cons of A and B.'' in preference to ``Returns the cons of A and B@.''
779 Usually it looks good to do likewise for the rest of the first
780 paragraph. Subsequent paragraphs usually look better if each sentence
781 is indicative and has a proper subject.
782
783 @item
784 The documentation string for a function that is a yes-or-no predicate
785 should start with words such as ``Return t if,'' to indicate
786 explicitly what constitutes ``truth.'' The word ``return'' avoids
787 starting the sentence with lower-case ``t,'' which could be somewhat
788 distracting.
789
790 @item
791 If a line in a documentation string begins with an open-parenthesis,
792 write a backslash before the open-parenthesis, like this:
793
794 @example
795 The argument FOO can be either a number
796 \(a buffer position) or a string (a file name).
797 @end example
798
799 This prevents the open-parenthesis from being treated as the start of a
800 defun (@pxref{Defuns,, Defuns, emacs, The GNU Emacs Manual}).
801
802 @item
803 Write documentation strings in the active voice, not the passive, and in
804 the present tense, not the future. For instance, use ``Return a list
805 containing A and B.'' instead of ``A list containing A and B will be
806 returned.''
807
808 @item
809 Avoid using the word ``cause'' (or its equivalents) unnecessarily.
810 Instead of, ``Cause Emacs to display text in boldface,'' write just
811 ``Display text in boldface.''
812
813 @item
814 Avoid using ``iff'' (a mathematics term meaning ``if and only if''),
815 since many people are unfamiliar with it and mistake it for a typo. In
816 most cases, the meaning is clear with just ``if''. Otherwise, try to
817 find an alternate phrasing that conveys the meaning.
818
819 @item
820 When a command is meaningful only in a certain mode or situation,
821 do mention that in the documentation string. For example,
822 the documentation of @code{dired-find-file} is:
823
824 @example
825 In Dired, visit the file or directory named on this line.
826 @end example
827
828 @item
829 When you define a variable that users ought to set interactively, you
830 normally should use @code{defcustom}. However, if for some reason you
831 use @code{defvar} instead, start the doc string with a @samp{*}.
832 @xref{Defining Variables}.
833
834 @item
835 The documentation string for a variable that is a yes-or-no flag should
836 start with words such as ``Non-nil means,'' to make it clear that
837 all non-@code{nil} values are equivalent and indicate explicitly what
838 @code{nil} and non-@code{nil} mean.
839 @end itemize
840
841 @node Comment Tips
842 @section Tips on Writing Comments
843 @cindex comments, Lisp convention for
844
845 We recommend these conventions for where to put comments and how to
846 indent them:
847
848 @table @samp
849 @item ;
850 Comments that start with a single semicolon, @samp{;}, should all be
851 aligned to the same column on the right of the source code. Such
852 comments usually explain how the code on the same line does its job. In
853 Lisp mode and related modes, the @kbd{M-;} (@code{indent-for-comment})
854 command automatically inserts such a @samp{;} in the right place, or
855 aligns such a comment if it is already present.
856
857 This and following examples are taken from the Emacs sources.
858
859 @smallexample
860 @group
861 (setq base-version-list ; there was a base
862 (assoc (substring fn 0 start-vn) ; version to which
863 file-version-assoc-list)) ; this looks like
864 ; a subversion
865 @end group
866 @end smallexample
867
868 @item ;;
869 Comments that start with two semicolons, @samp{;;}, should be aligned to
870 the same level of indentation as the code. Such comments usually
871 describe the purpose of the following lines or the state of the program
872 at that point. For example:
873
874 @smallexample
875 @group
876 (prog1 (setq auto-fill-function
877 @dots{}
878 @dots{}
879 ;; update mode line
880 (force-mode-line-update)))
881 @end group
882 @end smallexample
883
884 We also normally use two semicolons for comments outside functions.
885
886 @smallexample
887 @group
888 ;; This Lisp code is run in Emacs
889 ;; when it is to operate as a server
890 ;; for other processes.
891 @end group
892 @end smallexample
893
894 Every function that has no documentation string (presumably one that is
895 used only internally within the package it belongs to), should instead
896 have a two-semicolon comment right before the function, explaining what
897 the function does and how to call it properly. Explain precisely what
898 each argument means and how the function interprets its possible values.
899
900 @item ;;;
901 Comments that start with three semicolons, @samp{;;;}, should start at
902 the left margin. These are used, occasionally, for comments within
903 functions that should start at the margin. We also use them sometimes
904 for comments that are between functions---whether to use two or three
905 semicolons depends on whether the comment should be considered a
906 ``heading'' by Outline minor mode. By default, comments starting with
907 at least three semicolons (followed by a single space and a
908 non-whitespace character) are considered headings, comments starting
909 with two or less are not.
910
911 Another use for triple-semicolon comments is for commenting out lines
912 within a function. We use three semicolons for this precisely so that
913 they remain at the left margin. By default, Outline minor mode does
914 not consider a comment to be a heading (even if it starts with at
915 least three semicolons) if the semicolons are followed by at least two
916 spaces. Thus, if you add an introductory comment to the commented out
917 code, make sure to indent it by at least two spaces after the three
918 semicolons.
919
920 @smallexample
921 (defun foo (a)
922 ;;; This is no longer necessary.
923 ;;; (force-mode-line-update)
924 (message "Finished with %s" a))
925 @end smallexample
926
927 When commenting out entire functions, use two semicolons.
928
929 @item ;;;;
930 Comments that start with four semicolons, @samp{;;;;}, should be aligned
931 to the left margin and are used for headings of major sections of a
932 program. For example:
933
934 @smallexample
935 ;;;; The kill ring
936 @end smallexample
937 @end table
938
939 @noindent
940 The indentation commands of the Lisp modes in Emacs, such as @kbd{M-;}
941 (@code{indent-for-comment}) and @key{TAB} (@code{lisp-indent-line}),
942 automatically indent comments according to these conventions,
943 depending on the number of semicolons. @xref{Comments,,
944 Manipulating Comments, emacs, The GNU Emacs Manual}.
945
946 @node Library Headers
947 @section Conventional Headers for Emacs Libraries
948 @cindex header comments
949 @cindex library header comments
950
951 Emacs has conventions for using special comments in Lisp libraries
952 to divide them into sections and give information such as who wrote
953 them. This section explains these conventions.
954
955 We'll start with an example, a package that is included in the Emacs
956 distribution.
957
958 Parts of this example reflect its status as part of Emacs; for
959 example, the copyright notice lists the Free Software Foundation as the
960 copyright holder, and the copying permission says the file is part of
961 Emacs. When you write a package and post it, the copyright holder would
962 be you (unless your employer claims to own it instead), and you should
963 get the suggested copying permission from the end of the GNU General
964 Public License itself. Don't say your file is part of Emacs
965 if we haven't installed it in Emacs yet!
966
967 With that warning out of the way, on to the example:
968
969 @smallexample
970 @group
971 ;;; lisp-mnt.el --- minor mode for Emacs Lisp maintainers
972
973 ;; Copyright (C) 1992 Free Software Foundation, Inc.
974 @end group
975
976 ;; Author: Eric S. Raymond <esr@@snark.thyrsus.com>
977 ;; Maintainer: Eric S. Raymond <esr@@snark.thyrsus.com>
978 ;; Created: 14 Jul 1992
979 ;; Version: 1.2
980 @group
981 ;; Keywords: docs
982
983 ;; This file is part of GNU Emacs.
984 @dots{}
985 ;; along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>.
986 @end group
987 @end smallexample
988
989 The very first line should have this format:
990
991 @example
992 ;;; @var{filename} --- @var{description}
993 @end example
994
995 @noindent
996 The description should be complete in one line. If the file
997 needs a @samp{-*-} specification, put it after @var{description}.
998
999 After the copyright notice come several @dfn{header comment} lines,
1000 each beginning with @samp{;; @var{header-name}:}. Here is a table of
1001 the conventional possibilities for @var{header-name}:
1002
1003 @table @samp
1004 @item Author
1005 This line states the name and net address of at least the principal
1006 author of the library.
1007
1008 If there are multiple authors, you can list them on continuation lines
1009 led by @code{;;} and a tab character, like this:
1010
1011 @smallexample
1012 @group
1013 ;; Author: Ashwin Ram <Ram-Ashwin@@cs.yale.edu>
1014 ;; Dave Sill <de5@@ornl.gov>
1015 ;; Dave Brennan <brennan@@hal.com>
1016 ;; Eric Raymond <esr@@snark.thyrsus.com>
1017 @end group
1018 @end smallexample
1019
1020 @item Maintainer
1021 This line should contain a single name/address as in the Author line, or
1022 an address only, or the string @samp{FSF}. If there is no maintainer
1023 line, the person(s) in the Author field are presumed to be the
1024 maintainers. The example above is mildly bogus because the maintainer
1025 line is redundant.
1026
1027 The idea behind the @samp{Author} and @samp{Maintainer} lines is to make
1028 possible a Lisp function to ``send mail to the maintainer'' without
1029 having to mine the name out by hand.
1030
1031 Be sure to surround the network address with @samp{<@dots{}>} if
1032 you include the person's full name as well as the network address.
1033
1034 @item Created
1035 This optional line gives the original creation date of the
1036 file. For historical interest only.
1037
1038 @item Version
1039 If you wish to record version numbers for the individual Lisp program, put
1040 them in this line.
1041
1042 @item Adapted-By
1043 In this header line, place the name of the person who adapted the
1044 library for installation (to make it fit the style conventions, for
1045 example).
1046
1047 @item Keywords
1048 This line lists keywords for the @code{finder-by-keyword} help command.
1049 Please use that command to see a list of the meaningful keywords.
1050
1051 This field is important; it's how people will find your package when
1052 they're looking for things by topic area. To separate the keywords, you
1053 can use spaces, commas, or both.
1054
1055 @item Package-Version
1056 If @samp{Version} is not suitable for use by the package manager, then
1057 a package can define @samp{Package-Version}; it will be used instead.
1058 This is handy if @samp{Version} is an RCS id or something else that
1059 cannot be parsed by @code{version-to-list}. @xref{Packaging Basics}.
1060
1061 @item Package-Requires
1062 If this exists, it names packages on which the current package depends
1063 for proper operation. @xref{Packaging Basics}. This is used by the
1064 package manager both at download time (to ensure that a complete set
1065 of packages is downloaded) and at activation time (to ensure that a
1066 package is activated if and only if all its dependencies have been).
1067
1068 Its format is a list of lists. The @code{car} of each sub-list is the
1069 name of a package, as a symbol. The @code{cadr} of each sub-list is
1070 the minimum acceptable version number, as a string. For instance:
1071
1072 @smallexample
1073 ;; Package-Requires: ((gnus "1.0") (bubbles "2.7.2"))
1074 @end smallexample
1075
1076 The package code automatically defines a package named @samp{emacs}
1077 with the version number of the currently running Emacs. This can be
1078 used to require a minimal version of Emacs for a package.
1079 @end table
1080
1081 Just about every Lisp library ought to have the @samp{Author} and
1082 @samp{Keywords} header comment lines. Use the others if they are
1083 appropriate. You can also put in header lines with other header
1084 names---they have no standard meanings, so they can't do any harm.
1085
1086 We use additional stylized comments to subdivide the contents of the
1087 library file. These should be separated by blank lines from anything
1088 else. Here is a table of them:
1089
1090 @table @samp
1091 @item ;;; Commentary:
1092 This begins introductory comments that explain how the library works.
1093 It should come right after the copying permissions, terminated by a
1094 @samp{Change Log}, @samp{History} or @samp{Code} comment line. This
1095 text is used by the Finder package, so it should make sense in that
1096 context.
1097
1098 @item ;;; Documentation:
1099 This was used in some files in place of @samp{;;; Commentary:},
1100 but it is deprecated.
1101
1102 @item ;;; Change Log:
1103 This begins change log information stored in the library file (if you
1104 store the change history there). For Lisp files distributed with Emacs,
1105 the change history is kept in the file @file{ChangeLog} and not in the
1106 source file at all; these files generally do not have a @samp{;;; Change
1107 Log:} line. @samp{History} is an alternative to @samp{Change Log}.
1108
1109 @item ;;; Code:
1110 This begins the actual code of the program.
1111
1112 @item ;;; @var{filename} ends here
1113 This is the @dfn{footer line}; it appears at the very end of the file.
1114 Its purpose is to enable people to detect truncated versions of the file
1115 from the lack of a footer line.
1116 @end table