1 \input texinfo @c -*-texinfo-*-
3 @setfilename ../info/standards.info
4 @settitle GNU Coding Standards
5 @c This date is automagically updated when you save this file:
6 @set lastupdate June 24, 1999
12 * Standards: (standards). GNU coding standards.
17 @c @setchapternewpage odd
18 @setchapternewpage off
20 @c This is used by a cross ref in make-stds.texi
31 Copyright (C) 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999 Free Software Foundation, Inc.
33 Permission is granted to make and distribute verbatim copies of
34 this manual provided the copyright notice and this permission notice
35 are preserved on all copies.
38 Permission is granted to process this file through TeX and print the
39 results, provided the printed document carries copying permission
40 notice identical to this one except for the removal of this paragraph
41 (this paragraph not being relevant to the printed manual).
44 Permission is granted to copy and distribute modified versions of this
45 manual under the conditions for verbatim copying, provided that the entire
46 resulting derived work is distributed under the terms of a permission
47 notice identical to this one.
49 Permission is granted to copy and distribute translations of this manual
50 into another language, under the above conditions for modified versions,
51 except that this permission notice may be stated in a translation approved
52 by the Free Software Foundation.
56 @title GNU Coding Standards
57 @author Richard Stallman
58 @author last updated @value{lastupdate}
61 @vskip 0pt plus 1filll
62 Copyright @copyright{} 1992, 1993, 1994, 1995, 1996, 1997, 1998 Free Software Foundation, Inc.
64 Permission is granted to make and distribute verbatim copies of
65 this manual provided the copyright notice and this permission notice
66 are preserved on all copies.
68 Permission is granted to copy and distribute modified versions of this
69 manual under the conditions for verbatim copying, provided that the entire
70 resulting derived work is distributed under the terms of a permission
71 notice identical to this one.
73 Permission is granted to copy and distribute translations of this manual
74 into another language, under the above conditions for modified versions,
75 except that this permission notice may be stated in a translation approved
76 by the Free Software Foundation.
80 @node Top, Preface, (dir), (dir)
83 Last updated @value{lastupdate}.
87 * Preface:: About the GNU Coding Standards
88 * Legal Issues:: Keeping Free Software Free
89 * Design Advice:: General Program Design
90 * Program Behavior:: Program Behavior for All Programs
91 * Writing C:: Making The Best Use of C
92 * Documentation:: Documenting Programs
93 * Managing Releases:: The Release Process
94 * References:: References to Non-Free Software or Documentation
98 @chapter About the GNU Coding Standards
100 The GNU Coding Standards were written by Richard Stallman and other GNU
101 Project volunteers. Their purpose is to make the GNU system clean,
102 consistent, and easy to install. This document can also be read as a
103 guide to writing portable, robust and reliable programs. It focuses on
104 programs written in C, but many of the rules and principles are useful
105 even if you write in another programming language. The rules often
106 state reasons for writing in a certain way.
108 Corrections or suggestions for this document should be sent to
109 @email{gnu@@gnu.org}. If you make a suggestion, please include a
110 suggested new wording for it; our time is limited. We prefer a context
111 diff to the @file{standards.texi} or @file{make-stds.texi} files, but if
112 you don't have those files, please mail your suggestion anyway.
114 This release of the GNU Coding Standards was last updated
118 @chapter Keeping Free Software Free
120 This @value{CHAPTER} discusses how you can make sure that GNU software
121 remains unencumbered.
124 * Reading Non-Free Code:: Referring to Proprietary Programs
125 * Contributions:: Accepting Contributions
128 @node Reading Non-Free Code
129 @section Referring to Proprietary Programs
131 Don't in any circumstances refer to Unix source code for or during
132 your work on GNU! (Or to any other proprietary programs.)
134 If you have a vague recollection of the internals of a Unix program,
135 this does not absolutely mean you can't write an imitation of it, but
136 do try to organize the imitation internally along different lines,
137 because this is likely to make the details of the Unix version
138 irrelevant and dissimilar to your results.
140 For example, Unix utilities were generally optimized to minimize
141 memory use; if you go for speed instead, your program will be very
142 different. You could keep the entire input file in core and scan it
143 there instead of using stdio. Use a smarter algorithm discovered more
144 recently than the Unix program. Eliminate use of temporary files. Do
145 it in one pass instead of two (we did this in the assembler).
147 Or, on the contrary, emphasize simplicity instead of speed. For some
148 applications, the speed of today's computers makes simpler algorithms
151 Or go for generality. For example, Unix programs often have static
152 tables or fixed-size strings, which make for arbitrary limits; use
153 dynamic allocation instead. Make sure your program handles NULs and
154 other funny characters in the input files. Add a programming language
155 for extensibility and write part of the program in that language.
157 Or turn some parts of the program into independently usable libraries.
158 Or use a simple garbage collector instead of tracking precisely when
159 to free memory, or use a new GNU facility such as obstacks.
163 @section Accepting Contributions
165 If the program you are working on is copyrighted by the Free Software
166 Foundation, then when someone else sends you a piece of code to add to
167 the program, we need legal papers to use it---just as we asked you to
168 sign papers initially. @emph{Each} person who makes a nontrivial
169 contribution to a program must sign some sort of legal papers in order
170 for us to have clear title to the program; the main author alone is not
173 So, before adding in any contributions from other people, please tell
174 us, so we can arrange to get the papers. Then wait until we tell you
175 that we have received the signed papers, before you actually use the
178 This applies both before you release the program and afterward. If
179 you receive diffs to fix a bug, and they make significant changes, we
180 need legal papers for that change.
182 This also applies to comments and documentation files. For copyright
183 law, comments and code are just text. Copyright applies to all kinds of
184 text, so we need legal papers for all kinds.
186 We know it is frustrating to ask for legal papers; it's frustrating for
187 us as well. But if you don't wait, you are going out on a limb---for
188 example, what if the contributor's employer won't sign a disclaimer?
189 You might have to take that code out again!
191 You don't need papers for changes of a few lines here or there, since
192 they are not significant for copyright purposes. Also, you don't need
193 papers if all you get from the suggestion is some ideas, not actual code
194 which you use. For example, if someone send you one implementation, but
195 you write a different implementation of the same idea, you don't need to
198 The very worst thing is if you forget to tell us about the other
199 contributor. We could be very embarrassed in court some day as a
202 We have more detailed advice for maintainers of programs; if you have
203 reached the stage of actually maintaining a program for GNU (whether
204 released or not), please ask us for a copy.
207 @chapter General Program Design
209 This @value{CHAPTER} discusses some of the issues you should take into
210 account when designing your program.
213 * Compatibility:: Compatibility with other implementations
214 * Using Extensions:: Using non-standard features
215 * ANSI C:: Using ANSI C features
216 * Source Language:: Using languages other than C
220 @section Compatibility with Other Implementations
222 With occasional exceptions, utility programs and libraries for GNU
223 should be upward compatible with those in Berkeley Unix, and upward
224 compatible with @sc{ansi} C if @sc{ansi} C specifies their behavior, and
225 upward compatible with @sc{posix} if @sc{posix} specifies their
228 When these standards conflict, it is useful to offer compatibility
229 modes for each of them.
231 @sc{ansi} C and @sc{posix} prohibit many kinds of extensions. Feel free
232 to make the extensions anyway, and include a @samp{--ansi},
233 @samp{--posix}, or @samp{--compatible} option to turn them off.
234 However, if the extension has a significant chance of breaking any real
235 programs or scripts, then it is not really upward compatible. Try to
236 redesign its interface.
238 Many GNU programs suppress extensions that conflict with @sc{posix} if the
239 environment variable @code{POSIXLY_CORRECT} is defined (even if it is
240 defined with a null value). Please make your program recognize this
241 variable if appropriate.
243 When a feature is used only by users (not by programs or command
244 files), and it is done poorly in Unix, feel free to replace it
245 completely with something totally different and better. (For example,
246 @code{vi} is replaced with Emacs.) But it is nice to offer a compatible
247 feature as well. (There is a free @code{vi} clone, so we offer it.)
249 Additional useful features not in Berkeley Unix are welcome.
251 @node Using Extensions
252 @section Using Non-standard Features
254 Many GNU facilities that already exist support a number of convenient
255 extensions over the comparable Unix facilities. Whether to use these
256 extensions in implementing your program is a difficult question.
258 On the one hand, using the extensions can make a cleaner program.
259 On the other hand, people will not be able to build the program
260 unless the other GNU tools are available. This might cause the
261 program to work on fewer kinds of machines.
263 With some extensions, it might be easy to provide both alternatives.
264 For example, you can define functions with a ``keyword'' @code{INLINE}
265 and define that as a macro to expand into either @code{inline} or
266 nothing, depending on the compiler.
268 In general, perhaps it is best not to use the extensions if you can
269 straightforwardly do without them, but to use the extensions if they
270 are a big improvement.
272 An exception to this rule are the large, established programs (such as
273 Emacs) which run on a great variety of systems. Such programs would
274 be broken by use of GNU extensions.
276 Another exception is for programs that are used as part of
277 compilation: anything that must be compiled with other compilers in
278 order to bootstrap the GNU compilation facilities. If these require
279 the GNU compiler, then no one can compile them without having them
280 installed already. That would be no good.
283 @section @sc{ansi} C and pre-@sc{ansi} C
285 Do not ever use the ``trigraph'' feature of @sc{ansi} C.
287 @sc{ansi} C is widespread enough now that it is ok to write new programs
288 that use @sc{ansi} C features (and therefore will not work in
289 non-@sc{ansi} compilers). And if a program is already written in
290 @sc{ansi} C, there's no need to convert it to support non-@sc{ansi}
293 If you don't know non-@sc{ansi} C, there's no need to learn it; just
294 write in @sc{ansi} C.
296 However, it is easy to support non-@sc{ansi} compilers in most programs,
297 so you might still consider doing so when you write a program. And if a
298 program you are maintaining has such support, you should try to keep it
301 To support pre-@sc{ansi} C, instead of writing function definitions in
302 @sc{ansi} prototype form,
311 write the definition in pre-@sc{ansi} style like this,
321 and use a separate declaration to specify the argument prototype:
327 You need such a declaration anyway, in a header file, to get the benefit
328 of @sc{ansi} C prototypes in all the files where the function is called.
329 And once you have the declaration, you normally lose nothing by writing
330 the function definition in the pre-@sc{ansi} style.
332 This technique does not work for integer types narrower than @code{int}.
333 If you think of an argument as being of a type narrower than @code{int},
334 declare it as @code{int} instead.
336 There are a few special cases where this technique is hard to use. For
337 example, if a function argument needs to hold the system type
338 @code{dev_t}, you run into trouble, because @code{dev_t} is shorter than
339 @code{int} on some machines; but you cannot use @code{int} instead,
340 because @code{dev_t} is wider than @code{int} on some machines. There
341 is no type you can safely use on all machines in a non-@sc{ansi}
342 definition. The only way to support non-@sc{ansi} C and pass such an
343 argument is to check the width of @code{dev_t} using Autoconf and choose
344 the argument type accordingly. This may not be worth the trouble.
346 @node Source Language
347 @section Using Languages Other Than C
349 Using a language other than C is like using a non-standard feature: it
350 will cause trouble for users. Even if GCC supports the other language,
351 users may find it inconvenient to have to install the compiler for that
352 other language in order to build your program. For example, if you
353 write your program in C++, people will have to install the C++ compiler
354 in order to compile your program. Thus, it is better if you write in C.
356 But there are three situations when there is no disadvantage in using
361 It is okay to use another language if your program contains an
362 interpreter for that language.
364 For example, if your program links with GUILE, it is ok to write part of
365 the program in Scheme or another language supported by GUILE.
368 It is okay to use another language in a tool specifically intended for
369 use with that language.
371 This is okay because the only people who want to build the tool will be
372 those who have installed the other language anyway.
375 If an application is of interest to a narrow community, then perhaps
376 it's not important if the application is inconvenient to install.
379 C has one other advantage over C++ and other compiled languages: more
380 people know C, so more people will find it easy to read and modify the
381 program if it is written in C.
383 @node Program Behavior
384 @chapter Program Behavior for All Programs
386 This @value{CHAPTER} describes how to write robust software. It also
387 describes general standards for error messages, the command line interface,
388 and how libraries should behave.
391 * Semantics:: Writing robust programs
392 * Libraries:: Library behavior
393 * Errors:: Formatting error messages
394 * User Interfaces:: Standards for command line interfaces
395 * Option Table:: Table of long options.
396 * Memory Usage:: When and how to care about memory needs
400 @section Writing Robust Programs
402 Avoid arbitrary limits on the length or number of @emph{any} data
403 structure, including file names, lines, files, and symbols, by allocating
404 all data structures dynamically. In most Unix utilities, ``long lines
405 are silently truncated''. This is not acceptable in a GNU utility.
407 Utilities reading files should not drop NUL characters, or any other
408 nonprinting characters @emph{including those with codes above 0177}.
409 The only sensible exceptions would be utilities specifically intended
410 for interface to certain types of terminals or printers
411 that can't handle those characters.
412 Whenever possible, try to make programs work properly with
413 sequences of bytes that represent multibyte characters, using encodings
414 such as UTF-8 and others.
416 Check every system call for an error return, unless you know you wish to
417 ignore errors. Include the system error text (from @code{perror} or
418 equivalent) in @emph{every} error message resulting from a failing
419 system call, as well as the name of the file if any and the name of the
420 utility. Just ``cannot open foo.c'' or ``stat failed'' is not
423 Check every call to @code{malloc} or @code{realloc} to see if it
424 returned zero. Check @code{realloc} even if you are making the block
425 smaller; in a system that rounds block sizes to a power of 2,
426 @code{realloc} may get a different block if you ask for less space.
428 In Unix, @code{realloc} can destroy the storage block if it returns
429 zero. GNU @code{realloc} does not have this bug: if it fails, the
430 original block is unchanged. Feel free to assume the bug is fixed. If
431 you wish to run your program on Unix, and wish to avoid lossage in this
432 case, you can use the GNU @code{malloc}.
434 You must expect @code{free} to alter the contents of the block that was
435 freed. Anything you want to fetch from the block, you must fetch before
438 If @code{malloc} fails in a noninteractive program, make that a fatal
439 error. In an interactive program (one that reads commands from the
440 user), it is better to abort the command and return to the command
441 reader loop. This allows the user to kill other processes to free up
442 virtual memory, and then try the command again.
444 Use @code{getopt_long} to decode arguments, unless the argument syntax
445 makes this unreasonable.
447 When static storage is to be written in during program execution, use
448 explicit C code to initialize it. Reserve C initialized declarations
449 for data that will not be changed.
452 Try to avoid low-level interfaces to obscure Unix data structures (such
453 as file directories, utmp, or the layout of kernel memory), since these
454 are less likely to work compatibly. If you need to find all the files
455 in a directory, use @code{readdir} or some other high-level interface.
456 These are supported compatibly by GNU.
458 The preferred signal handling facilities are the BSD variant of
459 @code{signal}, and the @sc{posix} @code{sigaction} function; the
460 alternative USG @code{signal} interface is an inferior design.
462 Nowadays, using the @sc{posix} signal functions may be the easiest way
463 to make a program portable. If you use @code{signal}, then on GNU/Linux
464 systems running GNU libc version 1, you should include
465 @file{bsd/signal.h} instead of @file{signal.h}, so as to get BSD
466 behavior. It is up to you whether to support systems where
467 @code{signal} has only the USG behavior, or give up on them.
469 In error checks that detect ``impossible'' conditions, just abort.
470 There is usually no point in printing any message. These checks
471 indicate the existence of bugs. Whoever wants to fix the bugs will have
472 to read the source code and run a debugger. So explain the problem with
473 comments in the source. The relevant data will be in variables, which
474 are easy to examine with the debugger, so there is no point moving them
477 Do not use a count of errors as the exit status for a program.
478 @emph{That does not work}, because exit status values are limited to 8
479 bits (0 through 255). A single run of the program might have 256
480 errors; if you try to return 256 as the exit status, the parent process
481 will see 0 as the status, and it will appear that the program succeeded.
483 If you make temporary files, check the @code{TMPDIR} environment
484 variable; if that variable is defined, use the specified directory
485 instead of @file{/tmp}.
488 @section Library Behavior
490 Try to make library functions reentrant. If they need to do dynamic
491 storage allocation, at least try to avoid any nonreentrancy aside from
492 that of @code{malloc} itself.
494 Here are certain name conventions for libraries, to avoid name
497 Choose a name prefix for the library, more than two characters long.
498 All external function and variable names should start with this
499 prefix. In addition, there should only be one of these in any given
500 library member. This usually means putting each one in a separate
503 An exception can be made when two external symbols are always used
504 together, so that no reasonable program could use one without the
505 other; then they can both go in the same file.
507 External symbols that are not documented entry points for the user
508 should have names beginning with @samp{_}. They should also contain
509 the chosen name prefix for the library, to prevent collisions with
510 other libraries. These can go in the same files with user entry
513 Static functions and variables can be used as you like and need not
514 fit any naming convention.
517 @section Formatting Error Messages
519 Error messages from compilers should look like this:
522 @var{source-file-name}:@var{lineno}: @var{message}
526 If you want to mention the column number, use this format:
529 @var{source-file-name}:@var{lineno}:@var{column}: @var{message}
533 Line numbers should start from 1 at the beginning of the file, and
534 column numbers should start from 1 at the beginning of the line. (Both
535 of these conventions are chosen for compatibility.) Calculate column
536 numbers assuming that space and all ASCII printing characters have
537 equal width, and assuming tab stops every 8 columns.
539 Error messages from other noninteractive programs should look like this:
542 @var{program}:@var{source-file-name}:@var{lineno}: @var{message}
546 when there is an appropriate source file, or like this:
549 @var{program}: @var{message}
553 when there is no relevant source file.
555 If you want to mention the column number, use this format:
558 @var{program}:@var{source-file-name}:@var{lineno}:@var{column}: @var{message}
561 In an interactive program (one that is reading commands from a
562 terminal), it is better not to include the program name in an error
563 message. The place to indicate which program is running is in the
564 prompt or with the screen layout. (When the same program runs with
565 input from a source other than a terminal, it is not interactive and
566 would do best to print error messages using the noninteractive style.)
568 The string @var{message} should not begin with a capital letter when
569 it follows a program name and/or file name. Also, it should not end
572 Error messages from interactive programs, and other messages such as
573 usage messages, should start with a capital letter. But they should not
576 @node User Interfaces
577 @section Standards for Command Line Interfaces
579 Please don't make the behavior of a utility depend on the name used
580 to invoke it. It is useful sometimes to make a link to a utility
581 with a different name, and that should not change what it does.
583 Instead, use a run time option or a compilation switch or both
584 to select among the alternate behaviors.
586 Likewise, please don't make the behavior of the program depend on the
587 type of output device it is used with. Device independence is an
588 important principle of the system's design; do not compromise it merely
589 to save someone from typing an option now and then. (Variation in error
590 message syntax when using a terminal is ok, because that is a side issue
591 that people do not depend on.)
593 If you think one behavior is most useful when the output is to a
594 terminal, and another is most useful when the output is a file or a
595 pipe, then it is usually best to make the default behavior the one that
596 is useful with output to a terminal, and have an option for the other
599 Compatibility requires certain programs to depend on the type of output
600 device. It would be disastrous if @code{ls} or @code{sh} did not do so
601 in the way all users expect. In some of these cases, we supplement the
602 program with a preferred alternate version that does not depend on the
603 output device type. For example, we provide a @code{dir} program much
604 like @code{ls} except that its default output format is always
607 It is a good idea to follow the @sc{posix} guidelines for the
608 command-line options of a program. The easiest way to do this is to use
609 @code{getopt} to parse them. Note that the GNU version of @code{getopt}
610 will normally permit options anywhere among the arguments unless the
611 special argument @samp{--} is used. This is not what @sc{posix}
612 specifies; it is a GNU extension.
614 Please define long-named options that are equivalent to the
615 single-letter Unix-style options. We hope to make GNU more user
616 friendly this way. This is easy to do with the GNU function
619 One of the advantages of long-named options is that they can be
620 consistent from program to program. For example, users should be able
621 to expect the ``verbose'' option of any GNU program which has one, to be
622 spelled precisely @samp{--verbose}. To achieve this uniformity, look at
623 the table of common long-option names when you choose the option names
624 for your program (@pxref{Option Table}).
626 It is usually a good idea for file names given as ordinary arguments to
627 be input files only; any output files would be specified using options
628 (preferably @samp{-o} or @samp{--output}). Even if you allow an output
629 file name as an ordinary argument for compatibility, try to provide an
630 option as another way to specify it. This will lead to more consistency
631 among GNU utilities, and fewer idiosyncracies for users to remember.
633 All programs should support two standard options: @samp{--version}
638 This option should direct the program to print information about its name,
639 version, origin and legal status, all on standard output, and then exit
640 successfully. Other options and arguments should be ignored once this
641 is seen, and the program should not perform its normal function.
643 The first line is meant to be easy for a program to parse; the version
644 number proper starts after the last space. In addition, it contains
645 the canonical name for this program, in this format:
652 The program's name should be a constant string; @emph{don't} compute it
653 from @code{argv[0]}. The idea is to state the standard or canonical
654 name for the program, not its file name. There are other ways to find
655 out the precise file name where a command is found in @code{PATH}.
657 If the program is a subsidiary part of a larger package, mention the
658 package name in parentheses, like this:
661 emacsserver (GNU Emacs) 19.30
665 If the package has a version number which is different from this
666 program's version number, you can mention the package version number
667 just before the close-parenthesis.
669 If you @strong{need} to mention the version numbers of libraries which
670 are distributed separately from the package which contains this program,
671 you can do so by printing an additional line of version info for each
672 library you want to mention. Use the same format for these lines as for
675 Please do not mention all of the libraries that the program uses ``just
676 for completeness''---that would produce a lot of unhelpful clutter.
677 Please mention library version numbers only if you find in practice that
678 they are very important to you in debugging.
680 The following line, after the version number line or lines, should be a
681 copyright notice. If more than one copyright notice is called for, put
682 each on a separate line.
684 Next should follow a brief statement that the program is free software,
685 and that users are free to copy and change it on certain conditions. If
686 the program is covered by the GNU GPL, say so here. Also mention that
687 there is no warranty, to the extent permitted by law.
689 It is ok to finish the output with a list of the major authors of the
690 program, as a way of giving credit.
692 Here's an example of output that follows these rules:
696 Copyright (C) 1996 Free Software Foundation, Inc.
697 GNU Emacs comes with NO WARRANTY,
698 to the extent permitted by law.
699 You may redistribute copies of GNU Emacs
700 under the terms of the GNU General Public License.
701 For more information about these matters,
702 see the files named COPYING.
705 You should adapt this to your program, of course, filling in the proper
706 year, copyright holder, name of program, and the references to
707 distribution terms, and changing the rest of the wording as necessary.
709 This copyright notice only needs to mention the most recent year in
710 which changes were made---there's no need to list the years for previous
711 versions' changes. You don't have to mention the name of the program in
712 these notices, if that is inconvenient, since it appeared in the first
716 This option should output brief documentation for how to invoke the
717 program, on standard output, then exit successfully. Other options and
718 arguments should be ignored once this is seen, and the program should
719 not perform its normal function.
721 Near the end of the @samp{--help} option's output there should be a line
722 that says where to mail bug reports. It should have this format:
725 Report bugs to @var{mailing-address}.
730 @section Table of Long Options
732 Here is a table of long options used by GNU programs. It is surely
733 incomplete, but we aim to list all the options that a new program might
734 want to be compatible with. If you use names not already in the table,
735 please send @email{gnu@@gnu.org} a list of them, with their
736 meanings, so we can update the table.
738 @c Please leave newlines between items in this table; it's much easier
739 @c to update when it isn't completely squashed together and unreadable.
740 @c When there is more than one short option for a long option name, put
741 @c a semicolon between the lists of the programs that use them, not a
742 @c period. --friedman
746 @samp{-N} in @code{tar}.
749 @samp{-a} in @code{du}, @code{ls}, @code{nm}, @code{stty}, @code{uname},
753 @samp{-a} in @code{diff}.
756 @samp{-A} in @code{ls}.
759 @samp{-a} in @code{etags}, @code{tee}, @code{time};
760 @samp{-r} in @code{tar}.
763 @samp{-a} in @code{cp}.
766 @samp{-n} in @code{shar}.
769 @samp{-l} in @code{m4}.
772 @samp{-a} in @code{diff}.
775 @samp{-v} in @code{gawk}.
784 @samp{-a} in @code{recode}.
787 @samp{-a} in @code{wdiff}.
790 @samp{-A} in @code{ptx}.
793 @samp{-n} in @code{wdiff}.
796 For server programs, run in the background.
798 @item backward-search
799 @samp{-B} in @code{ctags}.
802 @samp{-f} in @code{shar}.
811 @samp{-b} in @code{tac}.
814 @samp{-b} in @code{cpio} and @code{diff}.
817 @samp{-b} in @code{shar}.
820 Used in @code{cpio} and @code{tar}.
823 @samp{-b} in @code{head} and @code{tail}.
826 @samp{-b} in @code{ptx}.
829 Used in various programs to make output shorter.
832 @samp{-c} in @code{head}, @code{split}, and @code{tail}.
835 @samp{-C} in @code{etags}.
838 @samp{-A} in @code{tar}.
841 Used in various programs to specify the directory to use.
844 @samp{-c} in @code{chgrp} and @code{chown}.
847 @samp{-F} in @code{ls}.
850 @samp{-c} in @code{recode}.
853 @samp{-c} in @code{su};
857 @samp{-d} in @code{tar}.
863 @samp{-Z} in @code{tar} and @code{shar}.
866 @samp{-A} in @code{tar}.
869 @samp{-w} in @code{tar}.
875 @samp{-W copyleft} in @code{gawk}.
878 @samp{-C} in @code{ptx}, @code{recode}, and @code{wdiff};
879 @samp{-W copyright} in @code{gawk}.
885 @samp{-q} in @code{who}.
888 @samp{-l} in @code{du}.
891 Used in @code{tar} and @code{cpio}.
894 @samp{-c} in @code{shar}.
897 @samp{-x} in @code{ctags}.
900 @samp{-d} in @code{touch}.
903 @samp{-d} in Make and @code{m4};
907 @samp{-D} in @code{m4}.
910 @samp{-d} in Bison and @code{ctags}.
913 @samp{-D} in @code{tar}.
916 @samp{-L} in @code{chgrp}, @code{chown}, @code{cpio}, @code{du},
917 @code{ls}, and @code{tar}.
919 @item dereference-args
920 @samp{-D} in @code{du}.
923 Specify an I/O device (special file name).
926 @samp{-d} in @code{recode}.
928 @item dictionary-order
929 @samp{-d} in @code{look}.
932 @samp{-d} in @code{tar}.
935 @samp{-n} in @code{csplit}.
938 Specify the directory to use, in various programs. In @code{ls}, it
939 means to show directories themselves rather than their contents. In
940 @code{rm} and @code{ln}, it means to not treat links to directories
944 @samp{-x} in @code{strip}.
947 @samp{-X} in @code{strip}.
953 @samp{-e} in @code{diff}.
955 @item elide-empty-files
956 @samp{-z} in @code{csplit}.
959 @samp{-x} in @code{wdiff}.
962 @samp{-z} in @code{wdiff}.
964 @item entire-new-file
965 @samp{-N} in @code{diff}.
967 @item environment-overrides
971 @samp{-e} in @code{xargs}.
977 Used in @code{makeinfo}.
980 @samp{-o} in @code{m4}.
983 @samp{-b} in @code{ls}.
986 @samp{-X} in @code{tar}.
992 @samp{-x} in @code{xargs}.
995 @samp{-e} in @code{unshar}.
998 @samp{-t} in @code{diff}.
1001 @samp{-e} in @code{sed}.
1004 @samp{-g} in @code{nm}.
1007 @samp{-i} in @code{cpio};
1008 @samp{-x} in @code{tar}.
1011 @samp{-f} in @code{finger}.
1014 @samp{-f} in @code{su}.
1016 @item fatal-warnings
1017 @samp{-E} in @code{m4}.
1020 @samp{-f} in @code{info}, @code{gawk}, Make, @code{mt}, and @code{tar};
1021 @samp{-n} in @code{sed};
1022 @samp{-r} in @code{touch}.
1024 @item field-separator
1025 @samp{-F} in @code{gawk}.
1031 @samp{-F} in @code{ls}.
1034 @samp{-T} in @code{tar}.
1037 Used in @code{makeinfo}.
1039 @item flag-truncation
1040 @samp{-F} in @code{ptx}.
1042 @item fixed-output-files
1046 @samp{-f} in @code{tail}.
1048 @item footnote-style
1049 Used in @code{makeinfo}.
1052 @samp{-f} in @code{cp}, @code{ln}, @code{mv}, and @code{rm}.
1055 @samp{-F} in @code{shar}.
1058 For server programs, run in the foreground;
1059 in other words, don't do anything special to run the server
1063 Used in @code{ls}, @code{time}, and @code{ptx}.
1066 @samp{-F} in @code{m4}.
1072 @samp{-g} in @code{ptx}.
1075 @samp{-x} in @code{tar}.
1078 @samp{-i} in @code{ul}.
1081 @samp{-g} in @code{recode}.
1084 @samp{-g} in @code{install}.
1087 @samp{-z} in @code{tar} and @code{shar}.
1090 @samp{-H} in @code{m4}.
1093 @samp{-h} in @code{objdump} and @code{recode}
1096 @samp{-H} in @code{who}.
1099 Used to ask for brief usage information.
1101 @item here-delimiter
1102 @samp{-d} in @code{shar}.
1104 @item hide-control-chars
1105 @samp{-q} in @code{ls}.
1108 @samp{-u} in @code{who}.
1111 @samp{-D} in @code{diff}.
1114 @samp{-I} in @code{ls};
1115 @samp{-x} in @code{recode}.
1117 @item ignore-all-space
1118 @samp{-w} in @code{diff}.
1120 @item ignore-backups
1121 @samp{-B} in @code{ls}.
1123 @item ignore-blank-lines
1124 @samp{-B} in @code{diff}.
1127 @samp{-f} in @code{look} and @code{ptx};
1128 @samp{-i} in @code{diff} and @code{wdiff}.
1134 @samp{-i} in @code{ptx}.
1136 @item ignore-indentation
1137 @samp{-I} in @code{etags}.
1139 @item ignore-init-file
1142 @item ignore-interrupts
1143 @samp{-i} in @code{tee}.
1145 @item ignore-matching-lines
1146 @samp{-I} in @code{diff}.
1148 @item ignore-space-change
1149 @samp{-b} in @code{diff}.
1152 @samp{-i} in @code{tar}.
1155 @samp{-i} in @code{etags};
1156 @samp{-I} in @code{m4}.
1162 @samp{-G} in @code{tar}.
1165 @samp{-i}, @samp{-l}, and @samp{-m} in Finger.
1168 @samp{-i} in @code{expand}.
1171 @samp{-T} in @code{diff}.
1174 @samp{-i} in @code{ls}.
1177 @samp{-i} in @code{cp}, @code{ln}, @code{mv}, @code{rm};
1178 @samp{-e} in @code{m4};
1179 @samp{-p} in @code{xargs};
1180 @samp{-w} in @code{tar}.
1183 @samp{-p} in @code{shar}.
1195 @samp{-k} in @code{csplit}.
1198 @samp{-k} in @code{du} and @code{ls}.
1201 @samp{-l} in @code{etags}.
1204 @samp{-l} in @code{wdiff}.
1206 @item level-for-gzip
1207 @samp{-g} in @code{shar}.
1210 @samp{-C} in @code{split}.
1213 Used in @code{split}, @code{head}, and @code{tail}.
1216 @samp{-l} in @code{cpio}.
1220 Used in @code{gawk}.
1223 @samp{-t} in @code{cpio};
1224 @samp{-l} in @code{recode}.
1227 @samp{-t} in @code{tar}.
1230 @samp{-N} in @code{ls}.
1239 No listing of which programs already use this;
1240 someone should check to
1241 see if any actually do, and tell @email{gnu@@gnu.org}.
1244 @samp{-M} in @code{ptx}.
1247 @samp{-m} in @code{hello} and @code{uname}.
1249 @item make-directories
1250 @samp{-d} in @code{cpio}.
1259 @samp{-n} in @code{xargs}.
1262 @samp{-n} in @code{xargs}.
1265 @samp{-l} in @code{xargs}.
1271 @samp{-P} in @code{xargs}.
1274 @samp{-T} in @code{who}.
1277 @samp{-T} in @code{who}.
1280 @samp{-d} in @code{diff}.
1282 @item mixed-uuencode
1283 @samp{-M} in @code{shar}.
1286 @samp{-m} in @code{install}, @code{mkdir}, and @code{mkfifo}.
1288 @item modification-time
1289 @samp{-m} in @code{tar}.
1292 @samp{-M} in @code{tar}.
1298 @samp{-L} in @code{m4}.
1301 @samp{-a} in @code{shar}.
1306 @item no-builtin-rules
1309 @item no-character-count
1310 @samp{-w} in @code{shar}.
1312 @item no-check-existing
1313 @samp{-x} in @code{shar}.
1316 @samp{-3} in @code{wdiff}.
1319 @samp{-c} in @code{touch}.
1322 @samp{-D} in @code{etags}.
1325 @samp{-1} in @code{wdiff}.
1327 @item no-dereference
1328 @samp{-d} in @code{cp}.
1331 @samp{-2} in @code{wdiff}.
1340 @samp{-P} in @code{shar}.
1343 @samp{-e} in @code{gprof}.
1346 @samp{-R} in @code{etags}.
1349 @samp{-p} in @code{nm}.
1352 Used in @code{makeinfo}.
1355 @samp{-a} in @code{gprof}.
1358 @samp{-E} in @code{gprof}.
1361 @samp{-m} in @code{shar}.
1364 Used in @code{makeinfo}.
1367 Used in @code{emacsclient}.
1370 Used in various programs to inhibit warnings.
1373 @samp{-n} in @code{info}.
1376 @samp{-n} in @code{uname}.
1379 @samp{-f} in @code{cpio}.
1382 @samp{-n} in @code{objdump}.
1385 @samp{-0} in @code{xargs}.
1388 @samp{-n} in @code{cat}.
1390 @item number-nonblank
1391 @samp{-b} in @code{cat}.
1394 @samp{-n} in @code{nm}.
1396 @item numeric-uid-gid
1397 @samp{-n} in @code{cpio} and @code{ls}.
1403 @samp{-o} in @code{tar}.
1408 @item one-file-system
1409 @samp{-l} in @code{tar}, @code{cp}, and @code{du}.
1412 @samp{-o} in @code{ptx}.
1415 @samp{-f} in @code{gprof}.
1418 @samp{-F} in @code{gprof}.
1421 @samp{-o} in @code{getopt}, @code{fdlist}, @code{fdmount},
1422 @code{fdmountd}, and @code{fdumount}.
1425 In various programs, specify the output file name.
1428 @samp{-o} in @code{shar}.
1431 @samp{-o} in @code{rm}.
1434 @samp{-c} in @code{unshar}.
1437 @samp{-o} in @code{install}.
1440 @samp{-l} in @code{diff}.
1442 @item paragraph-indent
1443 Used in @code{makeinfo}.
1446 @samp{-p} in @code{mkdir} and @code{rmdir}.
1449 @samp{-p} in @code{ul}.
1452 @samp{-p} in @code{cpio}.
1455 @samp{-P} in @code{finger}.
1458 @samp{-c} in @code{cpio} and @code{tar}.
1461 Used in @code{gawk}.
1463 @item prefix-builtins
1464 @samp{-P} in @code{m4}.
1467 @samp{-f} in @code{csplit}.
1470 Used in @code{tar} and @code{cp}.
1472 @item preserve-environment
1473 @samp{-p} in @code{su}.
1475 @item preserve-modification-time
1476 @samp{-m} in @code{cpio}.
1478 @item preserve-order
1479 @samp{-s} in @code{tar}.
1481 @item preserve-permissions
1482 @samp{-p} in @code{tar}.
1485 @samp{-l} in @code{diff}.
1488 @samp{-L} in @code{cmp}.
1490 @item print-data-base
1493 @item print-directory
1496 @item print-file-name
1497 @samp{-o} in @code{nm}.
1500 @samp{-s} in @code{nm}.
1503 @samp{-p} in @code{wdiff}.
1506 @samp{-p} in @code{ed}.
1509 Specify an HTTP proxy.
1512 @samp{-X} in @code{shar}.
1518 Used in many programs to inhibit the usual output. @strong{Note:} every
1519 program accepting @samp{--quiet} should accept @samp{--silent} as a
1523 @samp{-Q} in @code{shar}
1526 @samp{-Q} in @code{ls}.
1529 @samp{-n} in @code{diff}.
1532 Used in @code{gawk}.
1534 @item read-full-blocks
1535 @samp{-B} in @code{tar}.
1544 @samp{-R} in @code{tar}.
1547 Used in @code{chgrp}, @code{chown}, @code{cp}, @code{ls}, @code{diff},
1550 @item reference-limit
1551 Used in @code{makeinfo}.
1554 @samp{-r} in @code{ptx}.
1557 @samp{-r} in @code{tac} and @code{etags}.
1560 @samp{-r} in @code{uname}.
1563 @samp{-R} in @code{m4}.
1566 @samp{-r} in @code{objdump}.
1569 @samp{-r} in @code{cpio}.
1572 @samp{-i} in @code{xargs}.
1574 @item report-identical-files
1575 @samp{-s} in @code{diff}.
1577 @item reset-access-time
1578 @samp{-a} in @code{cpio}.
1581 @samp{-r} in @code{ls} and @code{nm}.
1584 @samp{-f} in @code{diff}.
1586 @item right-side-defs
1587 @samp{-R} in @code{ptx}.
1590 @samp{-s} in @code{tar}.
1592 @item same-permissions
1593 @samp{-p} in @code{tar}.
1596 @samp{-g} in @code{stty}.
1601 @item sentence-regexp
1602 @samp{-S} in @code{ptx}.
1605 @samp{-S} in @code{du}.
1608 @samp{-s} in @code{tac}.
1611 Used by @code{recode} to chose files or pipes for sequencing passes.
1614 @samp{-s} in @code{su}.
1617 @samp{-A} in @code{cat}.
1619 @item show-c-function
1620 @samp{-p} in @code{diff}.
1623 @samp{-E} in @code{cat}.
1625 @item show-function-line
1626 @samp{-F} in @code{diff}.
1629 @samp{-T} in @code{cat}.
1632 Used in many programs to inhibit the usual output.
1633 @strong{Note:} every program accepting
1634 @samp{--silent} should accept @samp{--quiet} as a synonym.
1637 @samp{-s} in @code{ls}.
1640 Specify a file descriptor for a network server to use for its socket,
1641 instead of opening and binding a new socket. This provides a way to
1642 run, in a nonpriveledged process, a server that normally needs a
1643 reserved port number.
1649 @samp{-W source} in @code{gawk}.
1652 @samp{-S} in @code{tar}.
1654 @item speed-large-files
1655 @samp{-H} in @code{diff}.
1658 @samp{-E} in @code{unshar}.
1660 @item split-size-limit
1661 @samp{-L} in @code{shar}.
1664 @samp{-s} in @code{cat}.
1667 @samp{-w} in @code{wdiff}.
1670 @samp{-y} in @code{wdiff}.
1673 Used in @code{tar} and @code{diff} to specify which file within
1674 a directory to start processing with.
1677 @samp{-s} in @code{wdiff}.
1679 @item stdin-file-list
1680 @samp{-S} in @code{shar}.
1686 @samp{-s} in @code{recode}.
1689 @samp{-s} in @code{install}.
1692 @samp{-s} in @code{strip}.
1695 @samp{-S} in @code{strip}.
1698 @samp{-s} in @code{shar}.
1701 @samp{-S} in @code{cp}, @code{ln}, @code{mv}.
1704 @samp{-b} in @code{csplit}.
1707 @samp{-s} in @code{gprof}.
1710 @samp{-s} in @code{du}.
1713 @samp{-s} in @code{ln}.
1716 Used in GDB and @code{objdump}.
1719 @samp{-s} in @code{m4}.
1722 @samp{-s} in @code{uname}.
1725 @samp{-t} in @code{expand} and @code{unexpand}.
1728 @samp{-T} in @code{ls}.
1731 @samp{-T} in @code{tput} and @code{ul}.
1732 @samp{-t} in @code{wdiff}.
1735 @samp{-a} in @code{diff}.
1738 @samp{-T} in @code{shar}.
1741 Used in @code{ls} and @code{touch}.
1744 Specify how long to wait before giving up on some operation.
1747 @samp{-O} in @code{tar}.
1750 @samp{-c} in @code{du}.
1753 @samp{-t} in Make, @code{ranlib}, and @code{recode}.
1756 @samp{-t} in @code{m4}.
1759 @samp{-t} in @code{hello};
1760 @samp{-W traditional} in @code{gawk};
1761 @samp{-G} in @code{ed}, @code{m4}, and @code{ptx}.
1767 @samp{-t} in @code{ctags}.
1769 @item typedefs-and-c++
1770 @samp{-T} in @code{ctags}.
1773 @samp{-t} in @code{ptx}.
1776 @samp{-z} in @code{tar}.
1779 @samp{-u} in @code{cpio}.
1782 @samp{-U} in @code{m4}.
1784 @item undefined-only
1785 @samp{-u} in @code{nm}.
1788 @samp{-u} in @code{cp}, @code{ctags}, @code{mv}, @code{tar}.
1791 Used in @code{gawk}; same as @samp{--help}.
1794 @samp{-B} in @code{shar}.
1796 @item vanilla-operation
1797 @samp{-V} in @code{shar}.
1800 Print more information about progress. Many programs support this.
1803 @samp{-W} in @code{tar}.
1806 Print the version number.
1808 @item version-control
1809 @samp{-V} in @code{cp}, @code{ln}, @code{mv}.
1812 @samp{-v} in @code{ctags}.
1815 @samp{-V} in @code{tar}.
1820 @item whole-size-limit
1821 @samp{-l} in @code{shar}.
1824 @samp{-w} in @code{ls} and @code{ptx}.
1827 @samp{-W} in @code{ptx}.
1830 @samp{-T} in @code{who}.
1833 @samp{-z} in @code{gprof}.
1837 @section Memory Usage
1839 If it typically uses just a few meg of memory, don't bother making any
1840 effort to reduce memory usage. For example, if it is impractical for
1841 other reasons to operate on files more than a few meg long, it is
1842 reasonable to read entire input files into core to operate on them.
1844 However, for programs such as @code{cat} or @code{tail}, that can
1845 usefully operate on very large files, it is important to avoid using a
1846 technique that would artificially limit the size of files it can handle.
1847 If a program works by lines and could be applied to arbitrary
1848 user-supplied input files, it should keep only a line in memory, because
1849 this is not very hard and users will want to be able to operate on input
1850 files that are bigger than will fit in core all at once.
1852 If your program creates complicated data structures, just make them in
1853 core and give a fatal error if @code{malloc} returns zero.
1856 @chapter Making The Best Use of C
1858 This @value{CHAPTER} provides advice on how best to use the C language
1859 when writing GNU software.
1862 * Formatting:: Formatting Your Source Code
1863 * Comments:: Commenting Your Work
1864 * Syntactic Conventions:: Clean Use of C Constructs
1865 * Names:: Naming Variables and Functions
1866 * System Portability:: Portability between different operating systems
1867 * CPU Portability:: Supporting the range of CPU types
1868 * System Functions:: Portability and ``standard'' library functions
1869 * Internationalization:: Techniques for internationalization
1870 * Mmap:: How you can safely use @code{mmap}.
1874 @section Formatting Your Source Code
1876 It is important to put the open-brace that starts the body of a C
1877 function in column zero, and avoid putting any other open-brace or
1878 open-parenthesis or open-bracket in column zero. Several tools look
1879 for open-braces in column zero to find the beginnings of C functions.
1880 These tools will not work on code not formatted that way.
1882 It is also important for function definitions to start the name of the
1883 function in column zero. This helps people to search for function
1884 definitions, and may also help certain tools recognize them. Thus,
1885 the proper format is this:
1889 concat (s1, s2) /* Name starts in column zero here */
1891 @{ /* Open brace in column zero here */
1897 or, if you want to use @sc{ansi} C, format the definition like this:
1901 concat (char *s1, char *s2)
1907 In @sc{ansi} C, if the arguments don't fit nicely on one line,
1912 lots_of_args (int an_integer, long a_long, short a_short,
1913 double a_double, float a_float)
1917 For the body of the function, we prefer code formatted like this:
1929 return ++x + bar ();
1933 We find it easier to read a program when it has spaces before the
1934 open-parentheses and after the commas. Especially after the commas.
1936 When you split an expression into multiple lines, split it
1937 before an operator, not after one. Here is the right way:
1940 if (foo_this_is_long && bar > win (x, y, z)
1941 && remaining_condition)
1944 Try to avoid having two operators of different precedence at the same
1945 level of indentation. For example, don't write this:
1948 mode = (inmode[j] == VOIDmode
1949 || GET_MODE_SIZE (outmode[j]) > GET_MODE_SIZE (inmode[j])
1950 ? outmode[j] : inmode[j]);
1953 Instead, use extra parentheses so that the indentation shows the nesting:
1956 mode = ((inmode[j] == VOIDmode
1957 || (GET_MODE_SIZE (outmode[j]) > GET_MODE_SIZE (inmode[j])))
1958 ? outmode[j] : inmode[j]);
1961 Insert extra parentheses so that Emacs will indent the code properly.
1962 For example, the following indentation looks nice if you do it by hand,
1963 but Emacs would mess it up:
1966 v = rup->ru_utime.tv_sec*1000 + rup->ru_utime.tv_usec/1000
1967 + rup->ru_stime.tv_sec*1000 + rup->ru_stime.tv_usec/1000;
1970 But adding a set of parentheses solves the problem:
1973 v = (rup->ru_utime.tv_sec*1000 + rup->ru_utime.tv_usec/1000
1974 + rup->ru_stime.tv_sec*1000 + rup->ru_stime.tv_usec/1000);
1977 Format do-while statements like this:
1987 Please use formfeed characters (control-L) to divide the program into
1988 pages at logical places (but not within a function). It does not matter
1989 just how long the pages are, since they do not have to fit on a printed
1990 page. The formfeeds should appear alone on lines by themselves.
1994 @section Commenting Your Work
1996 Every program should start with a comment saying briefly what it is for.
1997 Example: @samp{fmt - filter for simple filling of text}.
1999 Please write the comments in a GNU program in English, because English
2000 is the one language that nearly all programmers in all countries can
2001 read. If you do not write English well, please write comments in
2002 English as well as you can, then ask other people to help rewrite them.
2003 If you can't write comments in English, please find someone to work with
2004 you and translate your comments into English.
2006 Please put a comment on each function saying what the function does,
2007 what sorts of arguments it gets, and what the possible values of
2008 arguments mean and are used for. It is not necessary to duplicate in
2009 words the meaning of the C argument declarations, if a C type is being
2010 used in its customary fashion. If there is anything nonstandard about
2011 its use (such as an argument of type @code{char *} which is really the
2012 address of the second character of a string, not the first), or any
2013 possible values that would not work the way one would expect (such as,
2014 that strings containing newlines are not guaranteed to work), be sure
2017 Also explain the significance of the return value, if there is one.
2019 Please put two spaces after the end of a sentence in your comments, so
2020 that the Emacs sentence commands will work. Also, please write
2021 complete sentences and capitalize the first word. If a lower-case
2022 identifier comes at the beginning of a sentence, don't capitalize it!
2023 Changing the spelling makes it a different identifier. If you don't
2024 like starting a sentence with a lower case letter, write the sentence
2025 differently (e.g., ``The identifier lower-case is @dots{}'').
2027 The comment on a function is much clearer if you use the argument
2028 names to speak about the argument values. The variable name itself
2029 should be lower case, but write it in upper case when you are speaking
2030 about the value rather than the variable itself. Thus, ``the inode
2031 number NODE_NUM'' rather than ``an inode''.
2033 There is usually no purpose in restating the name of the function in
2034 the comment before it, because the reader can see that for himself.
2035 There might be an exception when the comment is so long that the function
2036 itself would be off the bottom of the screen.
2038 There should be a comment on each static variable as well, like this:
2041 /* Nonzero means truncate lines in the display;
2042 zero means continue them. */
2046 Every @samp{#endif} should have a comment, except in the case of short
2047 conditionals (just a few lines) that are not nested. The comment should
2048 state the condition of the conditional that is ending, @emph{including
2049 its sense}. @samp{#else} should have a comment describing the condition
2050 @emph{and sense} of the code that follows. For example:
2058 #endif /* not foo */
2068 but, by contrast, write the comments this way for a @samp{#ifndef}:
2081 #endif /* not foo */
2085 @node Syntactic Conventions
2086 @section Clean Use of C Constructs
2088 Please explicitly declare all arguments to functions.
2089 Don't omit them just because they are @code{int}s.
2091 Declarations of external functions and functions to appear later in the
2092 source file should all go in one place near the beginning of the file
2093 (somewhere before the first function definition in the file), or else
2094 should go in a header file. Don't put @code{extern} declarations inside
2097 It used to be common practice to use the same local variables (with
2098 names like @code{tem}) over and over for different values within one
2099 function. Instead of doing this, it is better declare a separate local
2100 variable for each distinct purpose, and give it a name which is
2101 meaningful. This not only makes programs easier to understand, it also
2102 facilitates optimization by good compilers. You can also move the
2103 declaration of each local variable into the smallest scope that includes
2104 all its uses. This makes the program even cleaner.
2106 Don't use local variables or parameters that shadow global identifiers.
2108 Don't declare multiple variables in one declaration that spans lines.
2109 Start a new declaration on each line, instead. For example, instead
2135 (If they are global variables, each should have a comment preceding it
2138 When you have an @code{if}-@code{else} statement nested in another
2139 @code{if} statement, always put braces around the @code{if}-@code{else}.
2140 Thus, never write like this:
2163 If you have an @code{if} statement nested inside of an @code{else}
2164 statement, either write @code{else if} on one line, like this,
2174 with its @code{then}-part indented like the preceding @code{then}-part,
2175 or write the nested @code{if} within braces like this:
2187 Don't declare both a structure tag and variables or typedefs in the
2188 same declaration. Instead, declare the structure tag separately
2189 and then use it to declare the variables or typedefs.
2191 Try to avoid assignments inside @code{if}-conditions. For example,
2195 if ((foo = (char *) malloc (sizeof *foo)) == 0)
2196 fatal ("virtual memory exhausted");
2200 instead, write this:
2203 foo = (char *) malloc (sizeof *foo);
2205 fatal ("virtual memory exhausted");
2208 Don't make the program ugly to placate @code{lint}. Please don't insert any
2209 casts to @code{void}. Zero without a cast is perfectly fine as a null
2210 pointer constant, except when calling a varargs function.
2213 @section Naming Variables and Functions
2215 The names of global variables and functions in a program serve as
2216 comments of a sort. So don't choose terse names---instead, look for
2217 names that give useful information about the meaning of the variable or
2218 function. In a GNU program, names should be English, like other
2221 Local variable names can be shorter, because they are used only within
2222 one context, where (presumably) comments explain their purpose.
2224 Try to limit your use of abbreviations in symbol names. It is ok to
2225 make a few abbreviations, explain what they mean, and then use them
2226 frequently, but don't use lots of obscure abbreviations.
2228 Please use underscores to separate words in a name, so that the Emacs
2229 word commands can be useful within them. Stick to lower case; reserve
2230 upper case for macros and @code{enum} constants, and for name-prefixes
2231 that follow a uniform convention.
2233 For example, you should use names like @code{ignore_space_change_flag};
2234 don't use names like @code{iCantReadThis}.
2236 Variables that indicate whether command-line options have been
2237 specified should be named after the meaning of the option, not after
2238 the option-letter. A comment should state both the exact meaning of
2239 the option and its letter. For example,
2243 /* Ignore changes in horizontal whitespace (-b). */
2244 int ignore_space_change_flag;
2248 When you want to define names with constant integer values, use
2249 @code{enum} rather than @samp{#define}. GDB knows about enumeration
2252 Use file names of 14 characters or less, to avoid creating gratuitous
2253 problems on older System V systems. You can use the program
2254 @code{doschk} to test for this. @code{doschk} also tests for potential
2255 name conflicts if the files were loaded onto an MS-DOS file
2256 system---something you may or may not care about.
2258 @node System Portability
2259 @section Portability between System Types
2261 In the Unix world, ``portability'' refers to porting to different Unix
2262 versions. For a GNU program, this kind of portability is desirable, but
2265 The primary purpose of GNU software is to run on top of the GNU kernel,
2266 compiled with the GNU C compiler, on various types of @sc{cpu}. The
2267 amount and kinds of variation among GNU systems on different @sc{cpu}s
2268 will be comparable to the variation among Linux-based GNU systems or
2269 among BSD systems today. So the kinds of portability that are absolutely
2270 necessary are quite limited.
2272 But many users do run GNU software on non-GNU Unix or Unix-like systems.
2273 So supporting a variety of Unix-like systems is desirable, although not
2276 The easiest way to achieve portability to most Unix-like systems is to
2277 use Autoconf. It's unlikely that your program needs to know more
2278 information about the host platform than Autoconf can provide, simply
2279 because most of the programs that need such knowledge have already been
2282 Avoid using the format of semi-internal data bases (e.g., directories)
2283 when there is a higher-level alternative (@code{readdir}).
2285 As for systems that are not like Unix, such as MSDOS, Windows, the
2286 Macintosh, VMS, and MVS, supporting them is often a lot of work. When
2287 that is the case, it is better to spend your time adding features that
2288 will be useful on GNU and GNU/Linux, rather than on supporting other
2289 incompatible systems.
2291 @node CPU Portability
2292 @section Portability between @sc{cpu}s
2294 Even GNU systems will differ because of differences among @sc{cpu}
2295 types---for example, difference in byte ordering and alignment
2296 requirements. It is absolutely essential to handle these differences.
2297 However, don't make any effort to cater to the possibility that an
2298 @code{int} will be less than 32 bits. We don't support 16-bit machines
2301 Don't assume that the address of an @code{int} object is also the
2302 address of its least-significant byte. This is false on big-endian
2303 machines. Thus, don't make the following mistake:
2308 while ((c = getchar()) != EOF)
2309 write(file_descriptor, &c, 1);
2312 When calling functions, you need not worry about the difference between
2313 pointers of various types, or between pointers and integers. On most
2314 machines, there's no difference anyway. As for the few machines where
2315 there is a difference, all of them support @sc{ansi} C, so you can use
2316 prototypes (conditionalized to be active only in @sc{ansi} C) to make
2317 the code work on those systems.
2319 In certain cases, it is ok to pass integer and pointer arguments
2320 indiscriminately to the same function, and use no prototype on any
2321 system. For example, many GNU programs have error-reporting functions
2322 that pass their arguments along to @code{printf} and friends:
2325 error (s, a1, a2, a3)
2329 fprintf (stderr, "error: ");
2330 fprintf (stderr, s, a1, a2, a3);
2335 In practice, this works on all machines, since a pointer is generally
2336 the widest possible kind of argument, and it is much simpler than any
2337 ``correct'' alternative. Be sure @emph{not} to use a prototype for such
2340 However, avoid casting pointers to integers unless you really need to.
2341 Outside of special situations, such casts greatly reduce portability,
2342 and in most programs they are easy to avoid. In the cases where casting
2343 pointers to integers is essential---such as, a Lisp interpreter which
2344 stores type information as well as an address in one word---it is ok to
2345 do it, but you'll have to make explicit provisions to handle different
2348 @node System Functions
2349 @section Calling System Functions
2351 C implementations differ substantially. @sc{ansi} C reduces but does not
2352 eliminate the incompatibilities; meanwhile, many users wish to compile
2353 GNU software with pre-@sc{ansi} compilers. This chapter gives
2354 recommendations for how to use the more or less standard C library
2355 functions to avoid unnecessary loss of portability.
2359 Don't use the value of @code{sprintf}. It returns the number of
2360 characters written on some systems, but not on all systems.
2363 @code{main} should be declared to return type @code{int}. It should
2364 terminate either by calling @code{exit} or by returning the integer
2365 status code; make sure it cannot ever return an undefined value.
2368 Don't declare system functions explicitly.
2370 Almost any declaration for a system function is wrong on some system.
2371 To minimize conflicts, leave it to the system header files to declare
2372 system functions. If the headers don't declare a function, let it
2375 While it may seem unclean to use a function without declaring it, in
2376 practice this works fine for most system library functions on the
2377 systems where this really happens; thus, the disadvantage is only
2378 theoretical. By contrast, actual declarations have frequently caused
2382 If you must declare a system function, don't specify the argument types.
2383 Use an old-style declaration, not an @sc{ansi} prototype. The more you
2384 specify about the function, the more likely a conflict.
2387 In particular, don't unconditionally declare @code{malloc} or
2390 Most GNU programs use those functions just once, in functions
2391 conventionally named @code{xmalloc} and @code{xrealloc}. These
2392 functions call @code{malloc} and @code{realloc}, respectively, and
2395 Because @code{xmalloc} and @code{xrealloc} are defined in your program,
2396 you can declare them in other files without any risk of type conflict.
2398 On most systems, @code{int} is the same length as a pointer; thus, the
2399 calls to @code{malloc} and @code{realloc} work fine. For the few
2400 exceptional systems (mostly 64-bit machines), you can use
2401 @strong{conditionalized} declarations of @code{malloc} and
2402 @code{realloc}---or put these declarations in configuration files
2403 specific to those systems.
2406 The string functions require special treatment. Some Unix systems have
2407 a header file @file{string.h}; others have @file{strings.h}. Neither
2408 file name is portable. There are two things you can do: use Autoconf to
2409 figure out which file to include, or don't include either file.
2412 If you don't include either strings file, you can't get declarations for
2413 the string functions from the header file in the usual way.
2415 That causes less of a problem than you might think. The newer @sc{ansi}
2416 string functions should be avoided anyway because many systems still
2417 don't support them. The string functions you can use are these:
2420 strcpy strncpy strcat strncat
2421 strlen strcmp strncmp
2425 The copy and concatenate functions work fine without a declaration as
2426 long as you don't use their values. Using their values without a
2427 declaration fails on systems where the width of a pointer differs from
2428 the width of @code{int}, and perhaps in other cases. It is trivial to
2429 avoid using their values, so do that.
2431 The compare functions and @code{strlen} work fine without a declaration
2432 on most systems, possibly all the ones that GNU software runs on.
2433 You may find it necessary to declare them @strong{conditionally} on a
2436 The search functions must be declared to return @code{char *}. Luckily,
2437 there is no variation in the data type they return. But there is
2438 variation in their names. Some systems give these functions the names
2439 @code{index} and @code{rindex}; other systems use the names
2440 @code{strchr} and @code{strrchr}. Some systems support both pairs of
2441 names, but neither pair works on all systems.
2443 You should pick a single pair of names and use it throughout your
2444 program. (Nowadays, it is better to choose @code{strchr} and
2445 @code{strrchr} for new programs, since those are the standard @sc{ansi}
2446 names.) Declare both of those names as functions returning @code{char
2447 *}. On systems which don't support those names, define them as macros
2448 in terms of the other pair. For example, here is what to put at the
2449 beginning of your file (or in a header) if you want to use the names
2450 @code{strchr} and @code{strrchr} throughout:
2454 #define strchr index
2456 #ifndef HAVE_STRRCHR
2457 #define strrchr rindex
2465 Here we assume that @code{HAVE_STRCHR} and @code{HAVE_STRRCHR} are
2466 macros defined in systems where the corresponding functions exist.
2467 One way to get them properly defined is to use Autoconf.
2469 @node Internationalization
2470 @section Internationalization
2472 GNU has a library called GNU gettext that makes it easy to translate the
2473 messages in a program into various languages. You should use this
2474 library in every program. Use English for the messages as they appear
2475 in the program, and let gettext provide the way to translate them into
2478 Using GNU gettext involves putting a call to the @code{gettext} macro
2479 around each string that might need translation---like this:
2482 printf (gettext ("Processing file `%s'..."));
2486 This permits GNU gettext to replace the string @code{"Processing file
2487 `%s'..."} with a translated version.
2489 Once a program uses gettext, please make a point of writing calls to
2490 @code{gettext} when you add new strings that call for translation.
2492 Using GNU gettext in a package involves specifying a @dfn{text domain
2493 name} for the package. The text domain name is used to separate the
2494 translations for this package from the translations for other packages.
2495 Normally, the text domain name should be the same as the name of the
2496 package---for example, @samp{fileutils} for the GNU file utilities.
2498 To enable gettext to work well, avoid writing code that makes
2499 assumptions about the structure of words or sentences. When you want
2500 the precise text of a sentence to vary depending on the data, use two or
2501 more alternative string constants each containing a complete sentences,
2502 rather than inserting conditionalized words or phrases into a single
2505 Here is an example of what not to do:
2508 printf ("%d file%s processed", nfiles,
2509 nfiles != 1 ? "s" : "");
2513 The problem with that example is that it assumes that plurals are made
2514 by adding `s'. If you apply gettext to the format string, like this,
2517 printf (gettext ("%d file%s processed"), nfiles,
2518 nfiles != 1 ? "s" : "");
2522 the message can use different words, but it will still be forced to use
2523 `s' for the plural. Here is a better way:
2526 printf ((nfiles != 1 ? "%d files processed"
2527 : "%d file processed"),
2532 This way, you can apply gettext to each of the two strings
2536 printf ((nfiles != 1 ? gettext ("%d files processed")
2537 : gettext ("%d file processed")),
2542 This can be any method of forming the plural of the word for ``file'', and
2543 also handles languages that require agreement in the word for
2546 A similar problem appears at the level of sentence structure with this
2550 printf ("# Implicit rule search has%s been done.\n",
2551 f->tried_implicit ? "" : " not");
2555 Adding @code{gettext} calls to this code cannot give correct results for
2556 all languages, because negation in some languages requires adding words
2557 at more than one place in the sentence. By contrast, adding
2558 @code{gettext} calls does the job straightfowardly if the code starts
2562 printf (f->tried_implicit
2563 ? "# Implicit rule search has been done.\n",
2564 : "# Implicit rule search has not been done.\n");
2570 Don't assume that @code{mmap} either works on all files or fails
2571 for all files. It may work on some files and fail on others.
2573 The proper way to use @code{mmap} is to try it on the specific file for
2574 which you want to use it---and if @code{mmap} doesn't work, fall back on
2575 doing the job in another way using @code{read} and @code{write}.
2577 The reason this precaution is needed is that the GNU kernel (the HURD)
2578 provides a user-extensible file system, in which there can be many
2579 different kinds of ``ordinary files.'' Many of them support
2580 @code{mmap}, but some do not. It is important to make programs handle
2581 all these kinds of files.
2584 @chapter Documenting Programs
2587 * GNU Manuals:: Writing proper manuals.
2588 * Manual Structure Details:: Specific structure conventions.
2589 * License for Manuals:: Writing the distribution terms for a manual.
2590 * NEWS File:: NEWS files supplement manuals.
2591 * Change Logs:: Recording Changes
2592 * Man Pages:: Man pages are secondary.
2593 * Reading other Manuals:: How far you can go in learning
2598 @section GNU Manuals
2600 The preferred way to document part of the GNU system is to write a
2601 manual in the Texinfo formatting language. This makes it possible to
2602 produce a good quality formatted book, using @TeX{}, and to generate an
2603 Info file. It is also possible to generate HTML output from Texinfo
2604 source. See the Texinfo manual, either the hardcopy, or the on-line
2605 version available through @code{info} or the Emacs Info subsystem
2608 Programmers often find it most natural to structure the documentation
2609 following the structure of the implementation, which they know. But
2610 this structure is not necessarily good for explaining how to use the
2611 program; it may be irrelevant and confusing for a user.
2613 At every level, from the sentences in a paragraph to the grouping of
2614 topics into separate manuals, the right way to structure documentation
2615 is according to the concepts and questions that a user will have in mind
2616 when reading it. Sometimes this structure of ideas matches the
2617 structure of the implementation of the software being documented---but
2618 often they are different. Often the most important part of learning to
2619 write good documentation is learning to notice when you are structuring
2620 the documentation like the implementation, and think about better
2623 For example, each program in the GNU system probably ought to be
2624 documented in one manual; but this does not mean each program should
2625 have its own manual. That would be following the structure of the
2626 implementation, rather than the structure that helps the user
2629 Instead, each manual should cover a coherent @emph{topic}. For example,
2630 instead of a manual for @code{diff} and a manual for @code{diff3}, we
2631 have one manual for ``comparison of files'' which covers both of those
2632 programs, as well as @code{cmp}. By documenting these programs
2633 together, we can make the whole subject clearer.
2635 The manual which discusses a program should document all of the
2636 program's command-line options and all of its commands. It should give
2637 examples of their use. But don't organize the manual as a list of
2638 features. Instead, organize it logically, by subtopics. Address the
2639 questions that a user will ask when thinking about the job that the
2642 In general, a GNU manual should serve both as tutorial and reference.
2643 It should be set up for convenient access to each topic through Info,
2644 and for reading straight through (appendixes aside). A GNU manual
2645 should give a good introduction to a beginner reading through from the
2646 start, and should also provide all the details that hackers want.
2647 The Bison manual is a good example of this---please take a look at it
2648 to see what we mean.
2650 That is not as hard as it first sounds. Arrange each chapter as a
2651 logical breakdown of its topic, but order the sections, and write their
2652 text, so that reading the chapter straight through makes sense. Do
2653 likewise when structuring the book into chapters, and when structuring a
2654 section into paragraphs. The watchword is, @emph{at each point, address
2655 the most fundamental and important issue raised by the preceding text.}
2657 If necessary, add extra chapters at the beginning of the manual which
2658 are purely tutorial and cover the basics of the subject. These provide
2659 the framework for a beginner to understand the rest of the manual. The
2660 Bison manual provides a good example of how to do this.
2662 Don't use Unix man pages as a model for how to write GNU documentation;
2663 most of them are terse, badly structured, and give inadequate
2664 explanation of the underlying concepts. (There are, of course
2665 exceptions.) Also Unix man pages use a particular format which is
2666 different from what we use in GNU manuals.
2668 Please include an email address in the manual for where to report
2669 bugs @emph{in the manual}.
2671 Please do not use the term ``pathname'' that is used in Unix
2672 documentation; use ``file name'' (two words) instead. We use the term
2673 ``path'' only for search paths, which are lists of directory names.
2675 Please do not use the term ``illegal'' to refer to erroneous input to a
2676 computer program. Please use ``invalid'' for this, and reserve the term
2677 ``illegal'' for violations of law.
2679 @node Manual Structure Details
2680 @section Manual Structure Details
2682 The title page of the manual should state the version of the programs or
2683 packages documented in the manual. The Top node of the manual should
2684 also contain this information. If the manual is changing more
2685 frequently than or independent of the program, also state a version
2686 number for the manual in both of these places.
2688 Each program documented in the manual should have a node named
2689 @samp{@var{program} Invocation} or @samp{Invoking @var{program}}. This
2690 node (together with its subnodes, if any) should describe the program's
2691 command line arguments and how to run it (the sort of information people
2692 would look in a man page for). Start with an @samp{@@example}
2693 containing a template for all the options and arguments that the program
2696 Alternatively, put a menu item in some menu whose item name fits one of
2697 the above patterns. This identifies the node which that item points to
2698 as the node for this purpose, regardless of the node's actual name.
2700 There will be automatic features for specifying a program name and
2701 quickly reading just this part of its manual.
2703 If one manual describes several programs, it should have such a node for
2704 each program described.
2706 @node License for Manuals
2707 @section License for Manuals
2709 If the manual contains a copy of the GNU GPL or GNU LGPL, or if it
2710 contains chapters that make political or personal statements, please
2711 copy the distribution terms of the GNU Emacs Manual, and adapt it by
2712 modifying appropriately the list of special chapters that may not be
2713 modified or deleted.
2715 If the manual does not contain any such chapters, then imitate the
2716 simpler distribution terms of the Texinfo manual.
2719 @section The NEWS File
2721 In addition to its manual, the package should have a file named
2722 @file{NEWS} which contains a list of user-visible changes worth
2723 mentioning. In each new release, add items to the front of the file and
2724 identify the version they pertain to. Don't discard old items; leave
2725 them in the file after the newer items. This way, a user upgrading from
2726 any previous version can see what is new.
2728 If the @file{NEWS} file gets very long, move some of the older items
2729 into a file named @file{ONEWS} and put a note at the end referring the
2733 @section Change Logs
2735 Keep a change log to describe all the changes made to program source
2736 files. The purpose of this is so that people investigating bugs in the
2737 future will know about the changes that might have introduced the bug.
2738 Often a new bug can be found by looking at what was recently changed.
2739 More importantly, change logs can help you eliminate conceptual
2740 inconsistencies between different parts of a program, by giving you a
2741 history of how the conflicting concepts arose and who they came from.
2744 * Change Log Concepts::
2745 * Style of Change Logs::
2747 * Conditional Changes::
2750 @node Change Log Concepts
2751 @subsection Change Log Concepts
2753 You can think of the change log as a conceptual ``undo list'' which
2754 explains how earlier versions were different from the current version.
2755 People can see the current version; they don't need the change log
2756 to tell them what is in it. What they want from a change log is a
2757 clear explanation of how the earlier version differed.
2759 The change log file is normally called @file{ChangeLog} and covers an
2760 entire directory. Each directory can have its own change log, or a
2761 directory can use the change log of its parent directory--it's up to
2764 Another alternative is to record change log information with a version
2765 control system such as RCS or CVS. This can be converted automatically
2766 to a @file{ChangeLog} file using @code{rcs2log}; in Emacs, the command
2767 @kbd{C-x v a} (@code{vc-update-change-log}) does the job.
2769 There's no need to describe the full purpose of the changes or how they
2770 work together. If you think that a change calls for explanation, you're
2771 probably right. Please do explain it---but please put the explanation
2772 in comments in the code, where people will see it whenever they see the
2773 code. For example, ``New function'' is enough for the change log when
2774 you add a function, because there should be a comment before the
2775 function definition to explain what it does.
2777 However, sometimes it is useful to write one line to describe the
2778 overall purpose of a batch of changes.
2780 The easiest way to add an entry to @file{ChangeLog} is with the Emacs
2781 command @kbd{M-x add-change-log-entry}. An entry should have an
2782 asterisk, the name of the changed file, and then in parentheses the name
2783 of the changed functions, variables or whatever, followed by a colon.
2784 Then describe the changes you made to that function or variable.
2786 @node Style of Change Logs
2787 @subsection Style of Change Logs
2789 Here are some examples of change log entries:
2792 * register.el (insert-register): Return nil.
2793 (jump-to-register): Likewise.
2795 * sort.el (sort-subr): Return nil.
2797 * tex-mode.el (tex-bibtex-file, tex-file, tex-region):
2798 Restart the tex shell if process is gone or stopped.
2799 (tex-shell-running): New function.
2801 * expr.c (store_one_arg): Round size up for move_block_to_reg.
2802 (expand_call): Round up when emitting USE insns.
2803 * stmt.c (assign_parms): Round size up for move_block_from_reg.
2806 It's important to name the changed function or variable in full. Don't
2807 abbreviate function or variable names, and don't combine them.
2808 Subsequent maintainers will often search for a function name to find all
2809 the change log entries that pertain to it; if you abbreviate the name,
2810 they won't find it when they search.
2812 For example, some people are tempted to abbreviate groups of function
2813 names by writing @samp{* register.el (@{insert,jump-to@}-register)};
2814 this is not a good idea, since searching for @code{jump-to-register} or
2815 @code{insert-register} would not find that entry.
2817 Separate unrelated change log entries with blank lines. When two
2818 entries represent parts of the same change, so that they work together,
2819 then don't put blank lines between them. Then you can omit the file
2820 name and the asterisk when successive entries are in the same file.
2822 @node Simple Changes
2823 @subsection Simple Changes
2825 Certain simple kinds of changes don't need much detail in the change
2828 When you change the calling sequence of a function in a simple fashion,
2829 and you change all the callers of the function, there is no need to make
2830 individual entries for all the callers that you changed. Just write in
2831 the entry for the function being called, ``All callers changed.''
2834 * keyboard.c (Fcommand_execute): New arg SPECIAL.
2835 All callers changed.
2838 When you change just comments or doc strings, it is enough to write an
2839 entry for the file, without mentioning the functions. Just ``Doc
2840 fixes'' is enough for the change log.
2842 There's no need to make change log entries for documentation files.
2843 This is because documentation is not susceptible to bugs that are hard
2844 to fix. Documentation does not consist of parts that must interact in a
2845 precisely engineered fashion. To correct an error, you need not know
2846 the history of the erroneous passage; it is enough to compare what the
2847 documentation says with the way the program actually works.
2849 @node Conditional Changes
2850 @subsection Conditional Changes
2852 C programs often contain compile-time @code{#if} conditionals. Many
2853 changes are conditional; sometimes you add a new definition which is
2854 entirely contained in a conditional. It is very useful to indicate in
2855 the change log the conditions for which the change applies.
2857 Our convention for indicating conditional changes is to use square
2858 brackets around the name of the condition.
2860 Here is a simple example, describing a change which is conditional but
2861 does not have a function or entity name associated with it:
2864 * xterm.c [SOLARIS2]: Include string.h.
2867 Here is an entry describing a new definition which is entirely
2868 conditional. This new definition for the macro @code{FRAME_WINDOW_P} is
2869 used only when @code{HAVE_X_WINDOWS} is defined:
2872 * frame.h [HAVE_X_WINDOWS] (FRAME_WINDOW_P): Macro defined.
2875 Here is an entry for a change within the function @code{init_display},
2876 whose definition as a whole is unconditional, but the changes themselves
2877 are contained in a @samp{#ifdef HAVE_LIBNCURSES} conditional:
2880 * dispnew.c (init_display) [HAVE_LIBNCURSES]: If X, call tgetent.
2883 Here is an entry for a change that takes affect only when
2884 a certain macro is @emph{not} defined:
2887 (gethostname) [!HAVE_SOCKETS]: Replace with winsock version.
2893 In the GNU project, man pages are secondary. It is not necessary or
2894 expected for every GNU program to have a man page, but some of them do.
2895 It's your choice whether to include a man page in your program.
2897 When you make this decision, consider that supporting a man page
2898 requires continual effort each time the program is changed. The time
2899 you spend on the man page is time taken away from more useful work.
2901 For a simple program which changes little, updating the man page may be
2902 a small job. Then there is little reason not to include a man page, if
2905 For a large program that changes a great deal, updating a man page may
2906 be a substantial burden. If a user offers to donate a man page, you may
2907 find this gift costly to accept. It may be better to refuse the man
2908 page unless the same person agrees to take full responsibility for
2909 maintaining it---so that you can wash your hands of it entirely. If
2910 this volunteer later ceases to do the job, then don't feel obliged to
2911 pick it up yourself; it may be better to withdraw the man page from the
2912 distribution until someone else agrees to update it.
2914 When a program changes only a little, you may feel that the
2915 discrepancies are small enough that the man page remains useful without
2916 updating. If so, put a prominent note near the beginning of the man
2917 page explaining that you don't maintain it and that the Texinfo manual
2918 is more authoritative. The note should say how to access the Texinfo
2921 @node Reading other Manuals
2922 @section Reading other Manuals
2924 There may be non-free books or documentation files that describe the
2925 program you are documenting.
2927 It is ok to use these documents for reference, just as the author of a
2928 new algebra textbook can read other books on algebra. A large portion
2929 of any non-fiction book consists of facts, in this case facts about how
2930 a certain program works, and these facts are necessarily the same for
2931 everyone who writes about the subject. But be careful not to copy your
2932 outline structure, wording, tables or examples from preexisting non-free
2933 documentation. Copying from free documentation may be ok; please check
2934 with the FSF about the individual case.
2936 @node Managing Releases
2937 @chapter The Release Process
2939 Making a release is more than just bundling up your source files in a
2940 tar file and putting it up for FTP. You should set up your software so
2941 that it can be configured to run on a variety of systems. Your Makefile
2942 should conform to the GNU standards described below, and your directory
2943 layout should also conform to the standards discussed below. Doing so
2944 makes it easy to include your package into the larger framework of
2948 * Configuration:: How Configuration Should Work
2949 * Makefile Conventions:: Makefile Conventions
2950 * Releases:: Making Releases
2954 @section How Configuration Should Work
2956 Each GNU distribution should come with a shell script named
2957 @code{configure}. This script is given arguments which describe the
2958 kind of machine and system you want to compile the program for.
2960 The @code{configure} script must record the configuration options so
2961 that they affect compilation.
2963 One way to do this is to make a link from a standard name such as
2964 @file{config.h} to the proper configuration file for the chosen system.
2965 If you use this technique, the distribution should @emph{not} contain a
2966 file named @file{config.h}. This is so that people won't be able to
2967 build the program without configuring it first.
2969 Another thing that @code{configure} can do is to edit the Makefile. If
2970 you do this, the distribution should @emph{not} contain a file named
2971 @file{Makefile}. Instead, it should include a file @file{Makefile.in} which
2972 contains the input used for editing. Once again, this is so that people
2973 won't be able to build the program without configuring it first.
2975 If @code{configure} does write the @file{Makefile}, then @file{Makefile}
2976 should have a target named @file{Makefile} which causes @code{configure}
2977 to be rerun, setting up the same configuration that was set up last
2978 time. The files that @code{configure} reads should be listed as
2979 dependencies of @file{Makefile}.
2981 All the files which are output from the @code{configure} script should
2982 have comments at the beginning explaining that they were generated
2983 automatically using @code{configure}. This is so that users won't think
2984 of trying to edit them by hand.
2986 The @code{configure} script should write a file named @file{config.status}
2987 which describes which configuration options were specified when the
2988 program was last configured. This file should be a shell script which,
2989 if run, will recreate the same configuration.
2991 The @code{configure} script should accept an option of the form
2992 @samp{--srcdir=@var{dirname}} to specify the directory where sources are found
2993 (if it is not the current directory). This makes it possible to build
2994 the program in a separate directory, so that the actual source directory
2997 If the user does not specify @samp{--srcdir}, then @code{configure} should
2998 check both @file{.} and @file{..} to see if it can find the sources. If
2999 it finds the sources in one of these places, it should use them from
3000 there. Otherwise, it should report that it cannot find the sources, and
3001 should exit with nonzero status.
3003 Usually the easy way to support @samp{--srcdir} is by editing a
3004 definition of @code{VPATH} into the Makefile. Some rules may need to
3005 refer explicitly to the specified source directory. To make this
3006 possible, @code{configure} can add to the Makefile a variable named
3007 @code{srcdir} whose value is precisely the specified directory.
3009 The @code{configure} script should also take an argument which specifies the
3010 type of system to build the program for. This argument should look like
3014 @var{cpu}-@var{company}-@var{system}
3017 For example, a Sun 3 might be @samp{m68k-sun-sunos4.1}.
3019 The @code{configure} script needs to be able to decode all plausible
3020 alternatives for how to describe a machine. Thus, @samp{sun3-sunos4.1}
3021 would be a valid alias. For many programs, @samp{vax-dec-ultrix} would
3022 be an alias for @samp{vax-dec-bsd}, simply because the differences
3023 between Ultrix and BSD are rarely noticeable, but a few programs
3024 might need to distinguish them.
3025 @c Real 4.4BSD now runs on some Suns.
3027 There is a shell script called @file{config.sub} that you can use
3028 as a subroutine to validate system types and canonicalize aliases.
3030 Other options are permitted to specify in more detail the software
3031 or hardware present on the machine, and include or exclude optional
3032 parts of the package:
3035 @item --enable-@var{feature}@r{[}=@var{parameter}@r{]}
3036 Configure the package to build and install an optional user-level
3037 facility called @var{feature}. This allows users to choose which
3038 optional features to include. Giving an optional @var{parameter} of
3039 @samp{no} should omit @var{feature}, if it is built by default.
3041 No @samp{--enable} option should @strong{ever} cause one feature to
3042 replace another. No @samp{--enable} option should ever substitute one
3043 useful behavior for another useful behavior. The only proper use for
3044 @samp{--enable} is for questions of whether to build part of the program
3047 @item --with-@var{package}
3048 @c @r{[}=@var{parameter}@r{]}
3049 The package @var{package} will be installed, so configure this package
3050 to work with @var{package}.
3052 @c Giving an optional @var{parameter} of
3053 @c @samp{no} should omit @var{package}, if it is used by default.
3055 Possible values of @var{package} include
3056 @samp{gnu-as} (or @samp{gas}), @samp{gnu-ld}, @samp{gnu-libc},
3062 Do not use a @samp{--with} option to specify the file name to use to
3063 find certain files. That is outside the scope of what @samp{--with}
3067 The target machine has no floating point processor.
3070 The target machine assembler is GAS, the GNU assembler.
3071 This is obsolete; users should use @samp{--with-gnu-as} instead.
3074 The target machine has the X Window System installed.
3075 This is obsolete; users should use @samp{--with-x} instead.
3078 All @code{configure} scripts should accept all of these ``detail''
3079 options, whether or not they make any difference to the particular
3080 package at hand. In particular, they should accept any option that
3081 starts with @samp{--with-} or @samp{--enable-}. This is so users will
3082 be able to configure an entire GNU source tree at once with a single set
3085 You will note that the categories @samp{--with-} and @samp{--enable-}
3086 are narrow: they @strong{do not} provide a place for any sort of option
3087 you might think of. That is deliberate. We want to limit the possible
3088 configuration options in GNU software. We do not want GNU programs to
3089 have idiosyncratic configuration options.
3091 Packages that perform part of the compilation process may support cross-compilation.
3092 In such a case, the host and target machines for the program may be
3093 different. The @code{configure} script should normally treat the
3094 specified type of system as both the host and the target, thus producing
3095 a program which works for the same type of machine that it runs on.
3097 The way to build a cross-compiler, cross-assembler, or what have you, is
3098 to specify the option @samp{--host=@var{hosttype}} when running
3099 @code{configure}. This specifies the host system without changing the
3100 type of target system. The syntax for @var{hosttype} is the same as
3103 Bootstrapping a cross-compiler requires compiling it on a machine other
3104 than the host it will run on. Compilation packages accept a
3105 configuration option @samp{--build=@var{hosttype}} for specifying the
3106 configuration on which you will compile them, in case that is different
3109 Programs for which cross-operation is not meaningful need not accept the
3110 @samp{--host} option, because configuring an entire operating system for
3111 cross-operation is not a meaningful thing.
3113 Some programs have ways of configuring themselves automatically. If
3114 your program is set up to do this, your @code{configure} script can simply
3115 ignore most of its arguments.
3117 @comment The makefile standards are in a separate file that is also
3118 @comment included by make.texinfo. Done by roland@gnu.ai.mit.edu on 1/6/93.
3119 @comment For this document, turn chapters into sections, etc.
3121 @include make-stds.texi
3125 @section Making Releases
3127 Package the distribution of @code{Foo version 69.96} up in a gzipped tar
3128 file with the name @file{foo-69.96.tar.gz}. It should unpack into a
3129 subdirectory named @file{foo-69.96}.
3131 Building and installing the program should never modify any of the files
3132 contained in the distribution. This means that all the files that form
3133 part of the program in any way must be classified into @dfn{source
3134 files} and @dfn{non-source files}. Source files are written by humans
3135 and never changed automatically; non-source files are produced from
3136 source files by programs under the control of the Makefile.
3138 The distribution should contain a file named @file{README} which gives
3139 the name of the package, and a general description of what it does. It
3140 is also good to explain the purpose of each of the first-level
3141 subdirectories in the package, if there are any. The @file{README} file
3142 should either state the version number of the package, or refer to where
3143 in the package it can be found.
3145 The @file{README} file should refer to the file @file{INSTALL}, which
3146 should contain an explanation of the installation procedure.
3148 The @file{README} file should also refer to the file which contains the
3149 copying conditions. The GNU GPL, if used, should be in a file called
3150 @file{COPYING}. If the GNU LGPL is used, it should be in a file called
3153 Naturally, all the source files must be in the distribution. It is okay
3154 to include non-source files in the distribution, provided they are
3155 up-to-date and machine-independent, so that building the distribution
3156 normally will never modify them. We commonly include non-source files
3157 produced by Bison, @code{lex}, @TeX{}, and @code{makeinfo}; this helps avoid
3158 unnecessary dependencies between our distributions, so that users can
3159 install whichever packages they want to install.
3161 Non-source files that might actually be modified by building and
3162 installing the program should @strong{never} be included in the
3163 distribution. So if you do distribute non-source files, always make
3164 sure they are up to date when you make a new distribution.
3166 Make sure that the directory into which the distribution unpacks (as
3167 well as any subdirectories) are all world-writable (octal mode 777).
3168 This is so that old versions of @code{tar} which preserve the
3169 ownership and permissions of the files from the tar archive will be
3170 able to extract all the files even if the user is unprivileged.
3172 Make sure that all the files in the distribution are world-readable.
3174 Make sure that no file name in the distribution is more than 14
3175 characters long. Likewise, no file created by building the program
3176 should have a name longer than 14 characters. The reason for this is
3177 that some systems adhere to a foolish interpretation of the @sc{posix}
3178 standard, and refuse to open a longer name, rather than truncating as
3179 they did in the past.
3181 Don't include any symbolic links in the distribution itself. If the tar
3182 file contains symbolic links, then people cannot even unpack it on
3183 systems that don't support symbolic links. Also, don't use multiple
3184 names for one file in different directories, because certain file
3185 systems cannot handle this and that prevents unpacking the
3188 Try to make sure that all the file names will be unique on MS-DOS. A
3189 name on MS-DOS consists of up to 8 characters, optionally followed by a
3190 period and up to three characters. MS-DOS will truncate extra
3191 characters both before and after the period. Thus,
3192 @file{foobarhacker.c} and @file{foobarhacker.o} are not ambiguous; they
3193 are truncated to @file{foobarha.c} and @file{foobarha.o}, which are
3196 Include in your distribution a copy of the @file{texinfo.tex} you used
3197 to test print any @file{*.texinfo} or @file{*.texi} files.
3199 Likewise, if your program uses small GNU software packages like regex,
3200 getopt, obstack, or termcap, include them in the distribution file.
3201 Leaving them out would make the distribution file a little smaller at
3202 the expense of possible inconvenience to a user who doesn't know what
3206 @chapter References to Non-Free Software and Documentation
3208 A GNU program should not recommend use of any non-free program. We
3209 can't stop some people from writing proprietary programs, or stop other
3210 people from using them. But we can and should avoid helping to
3211 advertise them to new customers.
3213 Sometimes it is important to mention how to build your package on top of
3214 some non-free operating system or other non-free base package. In such
3215 cases, please mention the name of the non-free package or system in the
3216 briefest possible way. Don't include any references for where to find
3217 more information about the proprietary program. The goal should be that
3218 people already using the proprietary program will get the advice they
3219 need about how to use your free program, while people who don't already
3220 use the proprietary program will not see anything to encourage them to
3221 take an interest in it.
3223 Likewise, a GNU package should not refer the user to any non-free
3224 documentation for free software. The need for free documentation to go
3225 with free software is now a major focus of the GNU project; to show that
3226 we are serious about the need for free documentation, we must not
3227 undermine our position by recommending use of documentation that isn't