1 \input texinfo @c -*-texinfo-*-
3 @setfilename standards.info
4 @settitle GNU Coding Standards
5 @c This date is automagically updated when you save this file:
6 @set lastupdate April 9, 2006
9 @dircategory GNU organization
11 * Standards: (standards). GNU coding standards.
14 @c @setchapternewpage odd
15 @setchapternewpage off
17 @c Put everything in one index (arbitrarily chosen to be the concept index).
23 @c This is used by a cross ref in make-stds.texi
33 The GNU coding standards, last updated @value{lastupdate}.
35 Copyright (C) 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000,
36 2001, 2002, 2003, 2004, 2005, 2006 Free Software Foundation, Inc.
38 Permission is granted to copy, distribute and/or modify this document
39 under the terms of the GNU Free Documentation License, Version 1.1
40 or any later version published by the Free Software Foundation;
41 with no Invariant Sections, with no
42 Front-Cover Texts, and with no Back-Cover Texts.
43 A copy of the license is included in the section entitled ``GNU
44 Free Documentation License''.
48 @title GNU Coding Standards
49 @author Richard Stallman, et al.
50 @author last updated @value{lastupdate}
52 @vskip 0pt plus 1filll
59 @node Top, Preface, (dir), (dir)
66 * Preface:: About the GNU Coding Standards
67 * Legal Issues:: Keeping Free Software Free
68 * Design Advice:: General Program Design
69 * Program Behavior:: Program Behavior for All Programs
70 * Writing C:: Making The Best Use of C
71 * Documentation:: Documenting Programs
72 * Managing Releases:: The Release Process
73 * References:: References to Non-Free Software or Documentation
74 * Copying This Manual:: How to Make Copies of This Manual
80 @chapter About the GNU Coding Standards
82 The GNU Coding Standards were written by Richard Stallman and other GNU
83 Project volunteers. Their purpose is to make the GNU system clean,
84 consistent, and easy to install. This document can also be read as a
85 guide to writing portable, robust and reliable programs. It focuses on
86 programs written in C, but many of the rules and principles are useful
87 even if you write in another programming language. The rules often
88 state reasons for writing in a certain way.
90 This release of the GNU Coding Standards was last updated
93 @cindex where to obtain @code{standards.texi}
94 @cindex downloading this manual
95 If you did not obtain this file directly from the GNU project and
96 recently, please check for a newer version. You can get the GNU
97 Coding Standards from the GNU web server in many
98 different formats, including the Texinfo source, PDF, HTML, DVI, plain
99 text, and more, at: @uref{http://www.gnu.org/prep/standards/}.
101 Corrections or suggestions for this document should be sent to
102 @email{bug-standards@@gnu.org}. If you make a suggestion, please include a
103 suggested new wording for it; our time is limited. We prefer a context
104 diff to the @file{standards.texi} or @file{make-stds.texi} files, but if
105 you don't have those files, please mail your suggestion anyway.
107 These standards cover the minimum of what is important when writing a
108 GNU package. Likely, the needs for additional standards will come up.
109 Sometimes, you might suggest that such standards be added to this
110 document. If you think your standards would be generally useful, please
113 You should also set standards for your package on many questions not
114 addressed or not firmly specified here. The most important point is to
115 be self-consistent---try to stick to the conventions you pick, and try
116 to document them as much as possible. That way, your program will be
117 more maintainable by others.
119 The GNU Hello program serves as an example of how to follow the GNU
120 coding standards for a trivial program.
121 @uref{http://www.gnu.org/software/hello/hello.html}.
124 @chapter Keeping Free Software Free
125 @cindex legal aspects
127 This chapter discusses how you can make sure that GNU software
128 avoids legal difficulties, and other related issues.
131 * Reading Non-Free Code:: Referring to Proprietary Programs
132 * Contributions:: Accepting Contributions
133 * Trademarks:: How We Deal with Trademark Issues
136 @node Reading Non-Free Code
137 @section Referring to Proprietary Programs
138 @cindex proprietary programs
139 @cindex avoiding proprietary code
141 Don't in any circumstances refer to Unix source code for or during
142 your work on GNU! (Or to any other proprietary programs.)
144 If you have a vague recollection of the internals of a Unix program,
145 this does not absolutely mean you can't write an imitation of it, but
146 do try to organize the imitation internally along different lines,
147 because this is likely to make the details of the Unix version
148 irrelevant and dissimilar to your results.
150 For example, Unix utilities were generally optimized to minimize
151 memory use; if you go for speed instead, your program will be very
152 different. You could keep the entire input file in core and scan it
153 there instead of using stdio. Use a smarter algorithm discovered more
154 recently than the Unix program. Eliminate use of temporary files. Do
155 it in one pass instead of two (we did this in the assembler).
157 Or, on the contrary, emphasize simplicity instead of speed. For some
158 applications, the speed of today's computers makes simpler algorithms
161 Or go for generality. For example, Unix programs often have static
162 tables or fixed-size strings, which make for arbitrary limits; use
163 dynamic allocation instead. Make sure your program handles NULs and
164 other funny characters in the input files. Add a programming language
165 for extensibility and write part of the program in that language.
167 Or turn some parts of the program into independently usable libraries.
168 Or use a simple garbage collector instead of tracking precisely when
169 to free memory, or use a new GNU facility such as obstacks.
172 @section Accepting Contributions
174 @cindex accepting contributions
176 If the program you are working on is copyrighted by the Free Software
177 Foundation, then when someone else sends you a piece of code to add to
178 the program, we need legal papers to use it---just as we asked you to
179 sign papers initially. @emph{Each} person who makes a nontrivial
180 contribution to a program must sign some sort of legal papers in order
181 for us to have clear title to the program; the main author alone is not
184 So, before adding in any contributions from other people, please tell
185 us, so we can arrange to get the papers. Then wait until we tell you
186 that we have received the signed papers, before you actually use the
189 This applies both before you release the program and afterward. If
190 you receive diffs to fix a bug, and they make significant changes, we
191 need legal papers for that change.
193 This also applies to comments and documentation files. For copyright
194 law, comments and code are just text. Copyright applies to all kinds of
195 text, so we need legal papers for all kinds.
197 We know it is frustrating to ask for legal papers; it's frustrating for
198 us as well. But if you don't wait, you are going out on a limb---for
199 example, what if the contributor's employer won't sign a disclaimer?
200 You might have to take that code out again!
202 You don't need papers for changes of a few lines here or there, since
203 they are not significant for copyright purposes. Also, you don't need
204 papers if all you get from the suggestion is some ideas, not actual code
205 which you use. For example, if someone sent you one implementation, but
206 you write a different implementation of the same idea, you don't need to
209 The very worst thing is if you forget to tell us about the other
210 contributor. We could be very embarrassed in court some day as a
213 We have more detailed advice for maintainers of programs; if you have
214 reached the stage of actually maintaining a program for GNU (whether
215 released or not), please ask us for a copy. It is also available
216 online for your perusal: @uref{http://www.gnu.org/prep/maintain/}.
222 Please do not include any trademark acknowledgements in GNU software
223 packages or documentation.
225 Trademark acknowledgements are the statements that such-and-such is a
226 trademark of so-and-so. The GNU Project has no objection to the basic
227 idea of trademarks, but these acknowledgements feel like kowtowing,
228 and there is no legal requirement for them, so we don't use them.
230 What is legally required, as regards other people's trademarks, is to
231 avoid using them in ways which a reader might reasonably understand as
232 naming or labeling our own programs or activities. For example, since
233 ``Objective C'' is (or at least was) a trademark, we made sure to say
234 that we provide a ``compiler for the Objective C language'' rather
235 than an ``Objective C compiler''. The latter would have been meant as
236 a shorter way of saying the former, but it does not explicitly state
237 the relationship, so it could be misinterpreted as using ``Objective
238 C'' as a label for the compiler rather than for the language.
240 Please don't use ``win'' as an abbreviation for Microsoft Windows in
241 GNU software or documentation. In hacker terminology, calling
242 something a ``win'' is a form of praise. If you wish to praise
243 Microsoft Windows when speaking on your own, by all means do so, but
244 not in GNU software. Usually we write the name ``Windows'' in full,
245 but when brevity is very important (as in file names and sometimes
246 symbol names), we abbreviate it to ``w''. For instance, the files and
247 functions in Emacs that deal with Windows start with @samp{w32}.
250 @chapter General Program Design
251 @cindex program design
253 This chapter discusses some of the issues you should take into
254 account when designing your program.
256 @c Standard or ANSI C
258 @c In 1989 the American National Standards Institute (ANSI) standardized
259 @c C as standard X3.159-1989. In December of that year the
260 @c International Standards Organization ISO adopted the ANSI C standard
261 @c making minor changes. In 1990 ANSI then re-adopted ISO standard
262 @c C. This version of C is known as either ANSI C or Standard C.
264 @c A major revision of the C Standard appeared in 1999.
267 * Source Language:: Which languages to use.
268 * Compatibility:: Compatibility with other implementations
269 * Using Extensions:: Using non-standard features
270 * Standard C:: Using Standard C features
271 * Conditional Compilation:: Compiling Code Only If A Conditional is True
274 @node Source Language
275 @section Which Languages to Use
276 @cindex programming languages
278 When you want to use a language that gets compiled and runs at high
279 speed, the best language to use is C. Using another language is like
280 using a non-standard feature: it will cause trouble for users. Even if
281 GCC supports the other language, users may find it inconvenient to have
282 to install the compiler for that other language in order to build your
283 program. For example, if you write your program in C++, people will
284 have to install the GNU C++ compiler in order to compile your program.
286 C has one other advantage over C++ and other compiled languages: more
287 people know C, so more people will find it easy to read and modify the
288 program if it is written in C.
290 So in general it is much better to use C, rather than the
291 comparable alternatives.
293 But there are two exceptions to that conclusion:
297 It is no problem to use another language to write a tool specifically
298 intended for use with that language. That is because the only people
299 who want to build the tool will be those who have installed the other
303 If an application is of interest only to a narrow part of the community,
304 then the question of which language it is written in has less effect on
305 other people, so you may as well please yourself.
308 Many programs are designed to be extensible: they include an interpreter
309 for a language that is higher level than C. Often much of the program
310 is written in that language, too. The Emacs editor pioneered this
314 The standard extensibility interpreter for GNU software is GUILE
315 (@uref{http://www.gnu.org/software/guile/}), which implements the
316 language Scheme (an especially clean and simple dialect of Lisp). We
317 don't reject programs written in other ``scripting languages'' such as
318 Perl and Python, but using GUILE is very important for the overall
319 consistency of the GNU system.
322 @section Compatibility with Other Implementations
323 @cindex compatibility with C and @sc{posix} standards
324 @cindex @sc{posix} compatibility
326 With occasional exceptions, utility programs and libraries for GNU
327 should be upward compatible with those in Berkeley Unix, and upward
328 compatible with Standard C if Standard C specifies their
329 behavior, and upward compatible with @sc{posix} if @sc{posix} specifies
332 When these standards conflict, it is useful to offer compatibility
333 modes for each of them.
335 @cindex options for compatibility
336 Standard C and @sc{posix} prohibit many kinds of extensions. Feel
337 free to make the extensions anyway, and include a @samp{--ansi},
338 @samp{--posix}, or @samp{--compatible} option to turn them off.
339 However, if the extension has a significant chance of breaking any real
340 programs or scripts, then it is not really upward compatible. So you
341 should try to redesign its interface to make it upward compatible.
343 @cindex @code{POSIXLY_CORRECT}, environment variable
344 Many GNU programs suppress extensions that conflict with @sc{posix} if the
345 environment variable @code{POSIXLY_CORRECT} is defined (even if it is
346 defined with a null value). Please make your program recognize this
347 variable if appropriate.
349 When a feature is used only by users (not by programs or command
350 files), and it is done poorly in Unix, feel free to replace it
351 completely with something totally different and better. (For example,
352 @code{vi} is replaced with Emacs.) But it is nice to offer a compatible
353 feature as well. (There is a free @code{vi} clone, so we offer it.)
355 Additional useful features are welcome regardless of whether
356 there is any precedent for them.
358 @node Using Extensions
359 @section Using Non-standard Features
360 @cindex non-standard extensions
362 Many GNU facilities that already exist support a number of convenient
363 extensions over the comparable Unix facilities. Whether to use these
364 extensions in implementing your program is a difficult question.
366 On the one hand, using the extensions can make a cleaner program.
367 On the other hand, people will not be able to build the program
368 unless the other GNU tools are available. This might cause the
369 program to work on fewer kinds of machines.
371 With some extensions, it might be easy to provide both alternatives.
372 For example, you can define functions with a ``keyword'' @code{INLINE}
373 and define that as a macro to expand into either @code{inline} or
374 nothing, depending on the compiler.
376 In general, perhaps it is best not to use the extensions if you can
377 straightforwardly do without them, but to use the extensions if they
378 are a big improvement.
380 An exception to this rule are the large, established programs (such as
381 Emacs) which run on a great variety of systems. Using GNU extensions in
382 such programs would make many users unhappy, so we don't do that.
384 Another exception is for programs that are used as part of compilation:
385 anything that must be compiled with other compilers in order to
386 bootstrap the GNU compilation facilities. If these require the GNU
387 compiler, then no one can compile them without having them installed
388 already. That would be extremely troublesome in certain cases.
391 @section Standard C and Pre-Standard C
392 @cindex @sc{ansi} C standard
394 1989 Standard C is widespread enough now that it is ok to use its
395 features in new programs. There is one exception: do not ever use the
396 ``trigraph'' feature of Standard C.
398 1999 Standard C is not widespread yet, so please do not require its
399 features in programs. It is ok to use its features if they are present.
401 However, it is easy to support pre-standard compilers in most programs,
402 so if you know how to do that, feel free. If a program you are
403 maintaining has such support, you should try to keep it working.
405 @cindex function prototypes
406 To support pre-standard C, instead of writing function definitions in
407 standard prototype form,
416 write the definition in pre-standard style like this,
426 and use a separate declaration to specify the argument prototype:
432 You need such a declaration anyway, in a header file, to get the benefit
433 of prototypes in all the files where the function is called. And once
434 you have the declaration, you normally lose nothing by writing the
435 function definition in the pre-standard style.
437 This technique does not work for integer types narrower than @code{int}.
438 If you think of an argument as being of a type narrower than @code{int},
439 declare it as @code{int} instead.
441 There are a few special cases where this technique is hard to use. For
442 example, if a function argument needs to hold the system type
443 @code{dev_t}, you run into trouble, because @code{dev_t} is shorter than
444 @code{int} on some machines; but you cannot use @code{int} instead,
445 because @code{dev_t} is wider than @code{int} on some machines. There
446 is no type you can safely use on all machines in a non-standard
447 definition. The only way to support non-standard C and pass such an
448 argument is to check the width of @code{dev_t} using Autoconf and choose
449 the argument type accordingly. This may not be worth the trouble.
451 In order to support pre-standard compilers that do not recognize
452 prototypes, you may want to use a preprocessor macro like this:
455 /* Declare the prototype for a general external function. */
456 #if defined (__STDC__) || defined (WINDOWSNT)
457 #define P_(proto) proto
463 @node Conditional Compilation
464 @section Conditional Compilation
466 When supporting configuration options already known when building your
467 program we prefer using @code{if (... )} over conditional compilation,
468 as in the former case the compiler is able to perform more extensive
469 checking of all possible code paths.
471 For example, please write
491 A modern compiler such as GCC will generate exactly the same code in
492 both cases, and we have been using similar techniques with good success
493 in several projects. Of course, the former method assumes that
494 @code{HAS_FOO} is defined as either 0 or 1.
496 While this is not a silver bullet solving all portability problems,
497 and is not always appropriate, following this policy would have saved
498 GCC developers many hours, or even days, per year.
500 In the case of function-like macros like @code{REVERSIBLE_CC_MODE} in
501 GCC which cannot be simply used in @code{if( ...)} statements, there is
502 an easy workaround. Simply introduce another macro
503 @code{HAS_REVERSIBLE_CC_MODE} as in the following example:
506 #ifdef REVERSIBLE_CC_MODE
507 #define HAS_REVERSIBLE_CC_MODE 1
509 #define HAS_REVERSIBLE_CC_MODE 0
513 @node Program Behavior
514 @chapter Program Behavior for All Programs
516 This chapter describes conventions for writing robust
517 software. It also describes general standards for error messages, the
518 command line interface, and how libraries should behave.
521 * Non-GNU Standards:: We consider standards such as POSIX;
522 we don't "obey" them.
523 * Semantics:: Writing robust programs
524 * Libraries:: Library behavior
525 * Errors:: Formatting error messages
526 * User Interfaces:: Standards about interfaces generally
527 * Graphical Interfaces:: Standards for graphical interfaces
528 * Command-Line Interfaces:: Standards for command line interfaces
529 * Option Table:: Table of long options
530 * Memory Usage:: When and how to care about memory needs
531 * File Usage:: Which files to use, and where
534 @node Non-GNU Standards
535 @section Non-GNU Standards
537 The GNU Project regards standards published by other organizations as
538 suggestions, not orders. We consider those standards, but we do not
539 ``obey'' them. In developing a GNU program, you should implement
540 an outside standard's specifications when that makes the GNU system
541 better overall in an objective sense. When it doesn't, you shouldn't.
543 In most cases, following published standards is convenient for
544 users---it means that their programs or scripts will work more
545 portably. For instance, GCC implements nearly all the features of
546 Standard C as specified by that standard. C program developers would
547 be unhappy if it did not. And GNU utilities mostly follow
548 specifications of POSIX.2; shell script writers and users would be
549 unhappy if our programs were incompatible.
551 But we do not follow either of these specifications rigidly, and there
552 are specific points on which we decided not to follow them, so as to
553 make the GNU system better for users.
555 For instance, Standard C says that nearly all extensions to C are
556 prohibited. How silly! GCC implements many extensions, some of which
557 were later adopted as part of the standard. If you want these
558 constructs to give an error message as ``required'' by the standard,
559 you must specify @samp{--pedantic}, which was implemented only so that
560 we can say ``GCC is a 100% implementation of the standard,'' not
561 because there is any reason to actually use it.
563 POSIX.2 specifies that @samp{df} and @samp{du} must output sizes by
564 default in units of 512 bytes. What users want is units of 1k, so
565 that is what we do by default. If you want the ridiculous behavior
566 ``required'' by POSIX, you must set the environment variable
567 @samp{POSIXLY_CORRECT} (which was originally going to be named
568 @samp{POSIX_ME_HARDER}).
570 GNU utilities also depart from the letter of the POSIX.2 specification
571 when they support long-named command-line options, and intermixing
572 options with ordinary arguments. This minor incompatibility with
573 POSIX is never a problem in practice, and it is very useful.
575 In particular, don't reject a new feature, or remove an old one,
576 merely because a standard says it is ``forbidden'' or ``deprecated.''
579 @section Writing Robust Programs
581 @cindex arbitrary limits on data
582 Avoid arbitrary limits on the length or number of @emph{any} data
583 structure, including file names, lines, files, and symbols, by allocating
584 all data structures dynamically. In most Unix utilities, ``long lines
585 are silently truncated''. This is not acceptable in a GNU utility.
587 @cindex @code{NUL} characters
588 Utilities reading files should not drop NUL characters, or any other
589 nonprinting characters @emph{including those with codes above 0177}.
590 The only sensible exceptions would be utilities specifically intended
591 for interface to certain types of terminals or printers
592 that can't handle those characters.
593 Whenever possible, try to make programs work properly with
594 sequences of bytes that represent multibyte characters, using encodings
595 such as UTF-8 and others.
597 @cindex error messages
598 Check every system call for an error return, unless you know you wish to
599 ignore errors. Include the system error text (from @code{perror} or
600 equivalent) in @emph{every} error message resulting from a failing
601 system call, as well as the name of the file if any and the name of the
602 utility. Just ``cannot open foo.c'' or ``stat failed'' is not
605 @cindex @code{malloc} return value
606 @cindex memory allocation failure
607 Check every call to @code{malloc} or @code{realloc} to see if it
608 returned zero. Check @code{realloc} even if you are making the block
609 smaller; in a system that rounds block sizes to a power of 2,
610 @code{realloc} may get a different block if you ask for less space.
612 In Unix, @code{realloc} can destroy the storage block if it returns
613 zero. GNU @code{realloc} does not have this bug: if it fails, the
614 original block is unchanged. Feel free to assume the bug is fixed. If
615 you wish to run your program on Unix, and wish to avoid lossage in this
616 case, you can use the GNU @code{malloc}.
618 You must expect @code{free} to alter the contents of the block that was
619 freed. Anything you want to fetch from the block, you must fetch before
622 If @code{malloc} fails in a noninteractive program, make that a fatal
623 error. In an interactive program (one that reads commands from the
624 user), it is better to abort the command and return to the command
625 reader loop. This allows the user to kill other processes to free up
626 virtual memory, and then try the command again.
628 @cindex command-line arguments, decoding
629 Use @code{getopt_long} to decode arguments, unless the argument syntax
630 makes this unreasonable.
632 When static storage is to be written in during program execution, use
633 explicit C code to initialize it. Reserve C initialized declarations
634 for data that will not be changed.
637 Try to avoid low-level interfaces to obscure Unix data structures (such
638 as file directories, utmp, or the layout of kernel memory), since these
639 are less likely to work compatibly. If you need to find all the files
640 in a directory, use @code{readdir} or some other high-level interface.
641 These are supported compatibly by GNU.
643 @cindex signal handling
644 The preferred signal handling facilities are the BSD variant of
645 @code{signal}, and the @sc{posix} @code{sigaction} function; the
646 alternative USG @code{signal} interface is an inferior design.
648 Nowadays, using the @sc{posix} signal functions may be the easiest way
649 to make a program portable. If you use @code{signal}, then on GNU/Linux
650 systems running GNU libc version 1, you should include
651 @file{bsd/signal.h} instead of @file{signal.h}, so as to get BSD
652 behavior. It is up to you whether to support systems where
653 @code{signal} has only the USG behavior, or give up on them.
655 @cindex impossible conditions
656 In error checks that detect ``impossible'' conditions, just abort.
657 There is usually no point in printing any message. These checks
658 indicate the existence of bugs. Whoever wants to fix the bugs will have
659 to read the source code and run a debugger. So explain the problem with
660 comments in the source. The relevant data will be in variables, which
661 are easy to examine with the debugger, so there is no point moving them
664 Do not use a count of errors as the exit status for a program.
665 @emph{That does not work}, because exit status values are limited to 8
666 bits (0 through 255). A single run of the program might have 256
667 errors; if you try to return 256 as the exit status, the parent process
668 will see 0 as the status, and it will appear that the program succeeded.
670 @cindex temporary files
671 @cindex @code{TMPDIR} environment variable
672 If you make temporary files, check the @code{TMPDIR} environment
673 variable; if that variable is defined, use the specified directory
674 instead of @file{/tmp}.
676 In addition, be aware that there is a possible security problem when
677 creating temporary files in world-writable directories. In C, you can
678 avoid this problem by creating temporary files in this manner:
681 fd = open(filename, O_WRONLY | O_CREAT | O_EXCL, 0600);
685 or by using the @code{mkstemps} function from libiberty.
687 In bash, use @code{set -C} to avoid this problem.
690 @section Library Behavior
693 Try to make library functions reentrant. If they need to do dynamic
694 storage allocation, at least try to avoid any nonreentrancy aside from
695 that of @code{malloc} itself.
697 Here are certain name conventions for libraries, to avoid name
700 Choose a name prefix for the library, more than two characters long.
701 All external function and variable names should start with this
702 prefix. In addition, there should only be one of these in any given
703 library member. This usually means putting each one in a separate
706 An exception can be made when two external symbols are always used
707 together, so that no reasonable program could use one without the
708 other; then they can both go in the same file.
710 External symbols that are not documented entry points for the user
711 should have names beginning with @samp{_}. The @samp{_} should be
712 followed by the chosen name prefix for the library, to prevent
713 collisions with other libraries. These can go in the same files with
714 user entry points if you like.
716 Static functions and variables can be used as you like and need not
717 fit any naming convention.
720 @section Formatting Error Messages
721 @cindex formatting error messages
722 @cindex error messages, formatting
724 Error messages from compilers should look like this:
727 @var{source-file-name}:@var{lineno}: @var{message}
731 If you want to mention the column number, use one of these formats:
734 @var{source-file-name}:@var{lineno}:@var{column}: @var{message}
735 @var{source-file-name}:@var{lineno}.@var{column}: @var{message}
740 Line numbers should start from 1 at the beginning of the file, and
741 column numbers should start from 1 at the beginning of the line. (Both
742 of these conventions are chosen for compatibility.) Calculate column
743 numbers assuming that space and all ASCII printing characters have
744 equal width, and assuming tab stops every 8 columns.
746 The error message can also give both the starting and ending positions
747 of the erroneous text. There are several formats so that you can
748 avoid redundant information such as a duplicate line number.
749 Here are the possible formats:
752 @var{source-file-name}:@var{lineno-1}.@var{column-1}-@var{lineno-2}.@var{column-2}: @var{message}
753 @var{source-file-name}:@var{lineno-1}.@var{column-1}-@var{column-2}: @var{message}
754 @var{source-file-name}:@var{lineno-1}-@var{lineno-2}: @var{message}
758 When an error is spread over several files, you can use this format:
761 @var{file-1}:@var{lineno-1}.@var{column-1}-@var{file-2}:@var{lineno-2}.@var{column-2}: @var{message}
764 Error messages from other noninteractive programs should look like this:
767 @var{program}:@var{source-file-name}:@var{lineno}: @var{message}
771 when there is an appropriate source file, or like this:
774 @var{program}: @var{message}
778 when there is no relevant source file.
780 If you want to mention the column number, use this format:
783 @var{program}:@var{source-file-name}:@var{lineno}:@var{column}: @var{message}
786 In an interactive program (one that is reading commands from a
787 terminal), it is better not to include the program name in an error
788 message. The place to indicate which program is running is in the
789 prompt or with the screen layout. (When the same program runs with
790 input from a source other than a terminal, it is not interactive and
791 would do best to print error messages using the noninteractive style.)
793 The string @var{message} should not begin with a capital letter when
794 it follows a program name and/or file name, because that isn't the
795 beginning of a sentence. (The sentence conceptually starts at the
796 beginning of the line.) Also, it should not end with a period.
798 Error messages from interactive programs, and other messages such as
799 usage messages, should start with a capital letter. But they should not
802 @node User Interfaces
803 @section Standards for Interfaces Generally
805 @cindex program name and its behavior
806 @cindex behavior, dependent on program's name
807 Please don't make the behavior of a utility depend on the name used
808 to invoke it. It is useful sometimes to make a link to a utility
809 with a different name, and that should not change what it does.
811 Instead, use a run time option or a compilation switch or both
812 to select among the alternate behaviors.
814 @cindex output device and program's behavior
815 Likewise, please don't make the behavior of the program depend on the
816 type of output device it is used with. Device independence is an
817 important principle of the system's design; do not compromise it merely
818 to save someone from typing an option now and then. (Variation in error
819 message syntax when using a terminal is ok, because that is a side issue
820 that people do not depend on.)
822 If you think one behavior is most useful when the output is to a
823 terminal, and another is most useful when the output is a file or a
824 pipe, then it is usually best to make the default behavior the one that
825 is useful with output to a terminal, and have an option for the other
828 Compatibility requires certain programs to depend on the type of output
829 device. It would be disastrous if @code{ls} or @code{sh} did not do so
830 in the way all users expect. In some of these cases, we supplement the
831 program with a preferred alternate version that does not depend on the
832 output device type. For example, we provide a @code{dir} program much
833 like @code{ls} except that its default output format is always
836 @node Graphical Interfaces
837 @section Standards for Graphical Interfaces
838 @cindex graphical user interface
841 When you write a program that provides a graphical user interface,
842 please make it work with X Windows and the GTK+ toolkit unless the
843 functionality specifically requires some alternative (for example,
844 ``displaying jpeg images while in console mode'').
846 In addition, please provide a command-line interface to control the
847 functionality. (In many cases, the graphical user interface can be a
848 separate program which invokes the command-line program.) This is
849 so that the same jobs can be done from scripts.
853 Please also consider providing a CORBA interface (for use from GNOME), a
854 library interface (for use from C), and perhaps a keyboard-driven
855 console interface (for use by users from console mode). Once you are
856 doing the work to provide the functionality and the graphical interface,
857 these won't be much extra work.
859 @node Command-Line Interfaces
860 @section Standards for Command Line Interfaces
861 @cindex command-line interface
864 It is a good idea to follow the @sc{posix} guidelines for the
865 command-line options of a program. The easiest way to do this is to use
866 @code{getopt} to parse them. Note that the GNU version of @code{getopt}
867 will normally permit options anywhere among the arguments unless the
868 special argument @samp{--} is used. This is not what @sc{posix}
869 specifies; it is a GNU extension.
871 @cindex long-named options
872 Please define long-named options that are equivalent to the
873 single-letter Unix-style options. We hope to make GNU more user
874 friendly this way. This is easy to do with the GNU function
877 One of the advantages of long-named options is that they can be
878 consistent from program to program. For example, users should be able
879 to expect the ``verbose'' option of any GNU program which has one, to be
880 spelled precisely @samp{--verbose}. To achieve this uniformity, look at
881 the table of common long-option names when you choose the option names
882 for your program (@pxref{Option Table}).
884 It is usually a good idea for file names given as ordinary arguments to
885 be input files only; any output files would be specified using options
886 (preferably @samp{-o} or @samp{--output}). Even if you allow an output
887 file name as an ordinary argument for compatibility, try to provide an
888 option as another way to specify it. This will lead to more consistency
889 among GNU utilities, and fewer idiosyncracies for users to remember.
891 @cindex standard command-line options
892 @cindex options, standard command-line
893 @cindex CGI programs, standard options for
894 @cindex PATH_INFO, specifying standard options as
895 All programs should support two standard options: @samp{--version}
896 and @samp{--help}. CGI programs should accept these as command-line
897 options, and also if given as the @env{PATH_INFO}; for instance,
898 visiting @url{http://example.org/p.cgi/--help} in a browser should
899 output the same information as invoking @samp{p.cgi --help} from the
903 @cindex @samp{--version} option
905 This option should direct the program to print information about its name,
906 version, origin and legal status, all on standard output, and then exit
907 successfully. Other options and arguments should be ignored once this
908 is seen, and the program should not perform its normal function.
910 @cindex canonical name of a program
911 @cindex program's canonical name
912 The first line is meant to be easy for a program to parse; the version
913 number proper starts after the last space. In addition, it contains
914 the canonical name for this program, in this format:
921 The program's name should be a constant string; @emph{don't} compute it
922 from @code{argv[0]}. The idea is to state the standard or canonical
923 name for the program, not its file name. There are other ways to find
924 out the precise file name where a command is found in @code{PATH}.
926 If the program is a subsidiary part of a larger package, mention the
927 package name in parentheses, like this:
930 emacsserver (GNU Emacs) 19.30
934 If the package has a version number which is different from this
935 program's version number, you can mention the package version number
936 just before the close-parenthesis.
938 If you @strong{need} to mention the version numbers of libraries which
939 are distributed separately from the package which contains this program,
940 you can do so by printing an additional line of version info for each
941 library you want to mention. Use the same format for these lines as for
944 Please do not mention all of the libraries that the program uses ``just
945 for completeness''---that would produce a lot of unhelpful clutter.
946 Please mention library version numbers only if you find in practice that
947 they are very important to you in debugging.
949 The following line, after the version number line or lines, should be a
950 copyright notice. If more than one copyright notice is called for, put
951 each on a separate line.
953 Next should follow a brief statement that the program is free software,
954 and that users are free to copy and change it on certain conditions. If
955 the program is covered by the GNU GPL, say so here. Also mention that
956 there is no warranty, to the extent permitted by law.
958 It is ok to finish the output with a list of the major authors of the
959 program, as a way of giving credit.
961 Here's an example of output that follows these rules:
965 Copyright (C) 1996 Free Software Foundation, Inc.
966 GNU Emacs comes with NO WARRANTY,
967 to the extent permitted by law.
968 You may redistribute copies of GNU Emacs
969 under the terms of the GNU General Public License.
970 For more information about these matters,
971 see the files named COPYING.
974 You should adapt this to your program, of course, filling in the proper
975 year, copyright holder, name of program, and the references to
976 distribution terms, and changing the rest of the wording as necessary.
978 This copyright notice only needs to mention the most recent year in
979 which changes were made---there's no need to list the years for previous
980 versions' changes. You don't have to mention the name of the program in
981 these notices, if that is inconvenient, since it appeared in the first
982 line. (The rules are different for copyright notices in source files;
983 see @ref{Copyright Notices,,,maintain,Information for GNU Maintainers}.)
985 Translations of the above lines must preserve the validity of the
986 copyright notices (@pxref{Internationalization}). If the translation's
987 character set supports it, the @samp{(C)} should be replaced with the
988 copyright symbol, as follows:
991 (the official copyright symbol, which is the letter C in a circle);
997 Write the word ``Copyright'' exactly like that, in English. Do not
998 translate it into another language. International treaties recognize
999 the English word ``Copyright''; translations into other languages do not
1000 have legal significance.
1003 @cindex @samp{--help} option
1005 This option should output brief documentation for how to invoke the
1006 program, on standard output, then exit successfully. Other options and
1007 arguments should be ignored once this is seen, and the program should
1008 not perform its normal function.
1010 @cindex address for bug reports
1012 Near the end of the @samp{--help} option's output there should be a line
1013 that says where to mail bug reports. It should have this format:
1016 Report bugs to @var{mailing-address}.
1021 @section Table of Long Options
1022 @cindex long option names
1023 @cindex table of long options
1025 Here is a table of long options used by GNU programs. It is surely
1026 incomplete, but we aim to list all the options that a new program might
1027 want to be compatible with. If you use names not already in the table,
1028 please send @email{bug-standards@@gnu.org} a list of them, with their
1029 meanings, so we can update the table.
1031 @c Please leave newlines between items in this table; it's much easier
1032 @c to update when it isn't completely squashed together and unreadable.
1033 @c When there is more than one short option for a long option name, put
1034 @c a semicolon between the lists of the programs that use them, not a
1035 @c period. --friedman
1039 @samp{-N} in @code{tar}.
1042 @samp{-a} in @code{du}, @code{ls}, @code{nm}, @code{stty}, @code{uname},
1043 and @code{unexpand}.
1046 @samp{-a} in @code{diff}.
1049 @samp{-A} in @code{ls}.
1052 @samp{-a} in @code{etags}, @code{tee}, @code{time};
1053 @samp{-r} in @code{tar}.
1056 @samp{-a} in @code{cp}.
1059 @samp{-n} in @code{shar}.
1062 @samp{-l} in @code{m4}.
1065 @samp{-a} in @code{diff}.
1068 @samp{-v} in @code{gawk}.
1077 @samp{-a} in @code{recode}.
1080 @samp{-a} in @code{wdiff}.
1082 @item auto-reference
1083 @samp{-A} in @code{ptx}.
1086 @samp{-n} in @code{wdiff}.
1089 For server programs, run in the background.
1091 @item backward-search
1092 @samp{-B} in @code{ctags}.
1095 @samp{-f} in @code{shar}.
1104 @samp{-b} in @code{tac}.
1107 @samp{-b} in @code{cpio} and @code{diff}.
1110 @samp{-b} in @code{shar}.
1113 Used in @code{cpio} and @code{tar}.
1116 @samp{-b} in @code{head} and @code{tail}.
1119 @samp{-b} in @code{ptx}.
1122 Used in various programs to make output shorter.
1125 @samp{-c} in @code{head}, @code{split}, and @code{tail}.
1128 @samp{-C} in @code{etags}.
1131 @samp{-A} in @code{tar}.
1134 Used in various programs to specify the directory to use.
1137 @samp{-c} in @code{chgrp} and @code{chown}.
1140 @samp{-F} in @code{ls}.
1143 @samp{-c} in @code{recode}.
1146 @samp{-c} in @code{su};
1150 @samp{-d} in @code{tar}.
1153 Used in @code{gawk}.
1156 @samp{-Z} in @code{tar} and @code{shar}.
1159 @samp{-A} in @code{tar}.
1162 @samp{-w} in @code{tar}.
1165 Used in @code{diff}.
1168 @samp{-W copyleft} in @code{gawk}.
1171 @samp{-C} in @code{ptx}, @code{recode}, and @code{wdiff};
1172 @samp{-W copyright} in @code{gawk}.
1178 @samp{-q} in @code{who}.
1181 @samp{-l} in @code{du}.
1184 Used in @code{tar} and @code{cpio}.
1187 @samp{-c} in @code{shar}.
1190 @samp{-x} in @code{ctags}.
1193 @samp{-d} in @code{touch}.
1196 @samp{-d} in Make and @code{m4};
1200 @samp{-D} in @code{m4}.
1203 @samp{-d} in Bison and @code{ctags}.
1206 @samp{-D} in @code{tar}.
1209 @samp{-L} in @code{chgrp}, @code{chown}, @code{cpio}, @code{du},
1210 @code{ls}, and @code{tar}.
1212 @item dereference-args
1213 @samp{-D} in @code{du}.
1216 Specify an I/O device (special file name).
1219 @samp{-d} in @code{recode}.
1221 @item dictionary-order
1222 @samp{-d} in @code{look}.
1225 @samp{-d} in @code{tar}.
1228 @samp{-n} in @code{csplit}.
1231 Specify the directory to use, in various programs. In @code{ls}, it
1232 means to show directories themselves rather than their contents. In
1233 @code{rm} and @code{ln}, it means to not treat links to directories
1237 @samp{-x} in @code{strip}.
1239 @item discard-locals
1240 @samp{-X} in @code{strip}.
1246 @samp{-e} in @code{diff}.
1248 @item elide-empty-files
1249 @samp{-z} in @code{csplit}.
1252 @samp{-x} in @code{wdiff}.
1255 @samp{-z} in @code{wdiff}.
1257 @item entire-new-file
1258 @samp{-N} in @code{diff}.
1260 @item environment-overrides
1264 @samp{-e} in @code{xargs}.
1270 Used in @code{makeinfo}.
1273 @samp{-o} in @code{m4}.
1276 @samp{-b} in @code{ls}.
1279 @samp{-X} in @code{tar}.
1285 @samp{-x} in @code{xargs}.
1288 @samp{-e} in @code{unshar}.
1291 @samp{-t} in @code{diff}.
1294 @samp{-e} in @code{sed}.
1297 @samp{-g} in @code{nm}.
1300 @samp{-i} in @code{cpio};
1301 @samp{-x} in @code{tar}.
1304 @samp{-f} in @code{finger}.
1307 @samp{-f} in @code{su}.
1309 @item fatal-warnings
1310 @samp{-E} in @code{m4}.
1313 @samp{-f} in @code{info}, @code{gawk}, Make, @code{mt}, and @code{tar};
1314 @samp{-n} in @code{sed};
1315 @samp{-r} in @code{touch}.
1317 @item field-separator
1318 @samp{-F} in @code{gawk}.
1324 @samp{-F} in @code{ls}.
1327 @samp{-T} in @code{tar}.
1330 Used in @code{makeinfo}.
1332 @item flag-truncation
1333 @samp{-F} in @code{ptx}.
1335 @item fixed-output-files
1339 @samp{-f} in @code{tail}.
1341 @item footnote-style
1342 Used in @code{makeinfo}.
1345 @samp{-f} in @code{cp}, @code{ln}, @code{mv}, and @code{rm}.
1348 @samp{-F} in @code{shar}.
1351 For server programs, run in the foreground;
1352 in other words, don't do anything special to run the server
1356 Used in @code{ls}, @code{time}, and @code{ptx}.
1359 @samp{-F} in @code{m4}.
1365 @samp{-g} in @code{ptx}.
1368 @samp{-x} in @code{tar}.
1371 @samp{-i} in @code{ul}.
1374 @samp{-g} in @code{recode}.
1377 @samp{-g} in @code{install}.
1380 @samp{-z} in @code{tar} and @code{shar}.
1383 @samp{-H} in @code{m4}.
1386 @samp{-h} in @code{objdump} and @code{recode}
1389 @samp{-H} in @code{who}.
1392 Used to ask for brief usage information.
1394 @item here-delimiter
1395 @samp{-d} in @code{shar}.
1397 @item hide-control-chars
1398 @samp{-q} in @code{ls}.
1401 In @code{makeinfo}, output HTML.
1404 @samp{-u} in @code{who}.
1407 @samp{-D} in @code{diff}.
1410 @samp{-I} in @code{ls};
1411 @samp{-x} in @code{recode}.
1413 @item ignore-all-space
1414 @samp{-w} in @code{diff}.
1416 @item ignore-backups
1417 @samp{-B} in @code{ls}.
1419 @item ignore-blank-lines
1420 @samp{-B} in @code{diff}.
1423 @samp{-f} in @code{look} and @code{ptx};
1424 @samp{-i} in @code{diff} and @code{wdiff}.
1430 @samp{-i} in @code{ptx}.
1432 @item ignore-indentation
1433 @samp{-I} in @code{etags}.
1435 @item ignore-init-file
1438 @item ignore-interrupts
1439 @samp{-i} in @code{tee}.
1441 @item ignore-matching-lines
1442 @samp{-I} in @code{diff}.
1444 @item ignore-space-change
1445 @samp{-b} in @code{diff}.
1448 @samp{-i} in @code{tar}.
1451 @samp{-i} in @code{etags};
1452 @samp{-I} in @code{m4}.
1458 @samp{-G} in @code{tar}.
1461 @samp{-i}, @samp{-l}, and @samp{-m} in Finger.
1464 In some programs, specify the name of the file to read as the user's
1468 @samp{-i} in @code{expand}.
1471 @samp{-T} in @code{diff}.
1474 @samp{-i} in @code{ls}.
1477 @samp{-i} in @code{cp}, @code{ln}, @code{mv}, @code{rm};
1478 @samp{-e} in @code{m4};
1479 @samp{-p} in @code{xargs};
1480 @samp{-w} in @code{tar}.
1483 @samp{-p} in @code{shar}.
1498 @samp{-k} in @code{csplit}.
1501 @samp{-k} in @code{du} and @code{ls}.
1504 @samp{-l} in @code{etags}.
1507 @samp{-l} in @code{wdiff}.
1509 @item level-for-gzip
1510 @samp{-g} in @code{shar}.
1513 @samp{-C} in @code{split}.
1516 Used in @code{split}, @code{head}, and @code{tail}.
1519 @samp{-l} in @code{cpio}.
1523 Used in @code{gawk}.
1526 @samp{-t} in @code{cpio};
1527 @samp{-l} in @code{recode}.
1530 @samp{-t} in @code{tar}.
1533 @samp{-N} in @code{ls}.
1542 Used in @code{uname}.
1545 @samp{-M} in @code{ptx}.
1548 @samp{-m} in @code{hello} and @code{uname}.
1550 @item make-directories
1551 @samp{-d} in @code{cpio}.
1560 @samp{-n} in @code{xargs}.
1563 @samp{-n} in @code{xargs}.
1566 @samp{-l} in @code{xargs}.
1572 @samp{-P} in @code{xargs}.
1575 @samp{-T} in @code{who}.
1578 @samp{-T} in @code{who}.
1581 @samp{-d} in @code{diff}.
1583 @item mixed-uuencode
1584 @samp{-M} in @code{shar}.
1587 @samp{-m} in @code{install}, @code{mkdir}, and @code{mkfifo}.
1589 @item modification-time
1590 @samp{-m} in @code{tar}.
1593 @samp{-M} in @code{tar}.
1599 @samp{-L} in @code{m4}.
1602 @samp{-a} in @code{shar}.
1607 @item no-builtin-rules
1610 @item no-character-count
1611 @samp{-w} in @code{shar}.
1613 @item no-check-existing
1614 @samp{-x} in @code{shar}.
1617 @samp{-3} in @code{wdiff}.
1620 @samp{-c} in @code{touch}.
1623 @samp{-D} in @code{etags}.
1626 @samp{-1} in @code{wdiff}.
1628 @item no-dereference
1629 @samp{-d} in @code{cp}.
1632 @samp{-2} in @code{wdiff}.
1641 @samp{-P} in @code{shar}.
1644 @samp{-e} in @code{gprof}.
1647 @samp{-R} in @code{etags}.
1650 @samp{-p} in @code{nm}.
1653 Don't print a startup splash screen.
1656 Used in @code{makeinfo}.
1659 @samp{-a} in @code{gprof}.
1662 @samp{-E} in @code{gprof}.
1665 @samp{-m} in @code{shar}.
1668 Used in @code{makeinfo}.
1671 Used in @code{emacsclient}.
1674 Used in various programs to inhibit warnings.
1677 @samp{-n} in @code{info}.
1680 @samp{-n} in @code{uname}.
1683 @samp{-f} in @code{cpio}.
1686 @samp{-n} in @code{objdump}.
1689 @samp{-0} in @code{xargs}.
1692 @samp{-n} in @code{cat}.
1694 @item number-nonblank
1695 @samp{-b} in @code{cat}.
1698 @samp{-n} in @code{nm}.
1700 @item numeric-uid-gid
1701 @samp{-n} in @code{cpio} and @code{ls}.
1707 @samp{-o} in @code{tar}.
1712 @item one-file-system
1713 @samp{-l} in @code{tar}, @code{cp}, and @code{du}.
1716 @samp{-o} in @code{ptx}.
1719 @samp{-f} in @code{gprof}.
1722 @samp{-F} in @code{gprof}.
1725 @samp{-o} in @code{getopt}, @code{fdlist}, @code{fdmount},
1726 @code{fdmountd}, and @code{fdumount}.
1729 In various programs, specify the output file name.
1732 @samp{-o} in @code{shar}.
1735 @samp{-o} in @code{rm}.
1738 @samp{-c} in @code{unshar}.
1741 @samp{-o} in @code{install}.
1744 @samp{-l} in @code{diff}.
1746 @item paragraph-indent
1747 Used in @code{makeinfo}.
1750 @samp{-p} in @code{mkdir} and @code{rmdir}.
1753 @samp{-p} in @code{ul}.
1756 @samp{-p} in @code{cpio}.
1759 @samp{-P} in @code{finger}.
1762 @samp{-c} in @code{cpio} and @code{tar}.
1765 Used in @code{gawk}.
1767 @item prefix-builtins
1768 @samp{-P} in @code{m4}.
1771 @samp{-f} in @code{csplit}.
1774 Used in @code{tar} and @code{cp}.
1776 @item preserve-environment
1777 @samp{-p} in @code{su}.
1779 @item preserve-modification-time
1780 @samp{-m} in @code{cpio}.
1782 @item preserve-order
1783 @samp{-s} in @code{tar}.
1785 @item preserve-permissions
1786 @samp{-p} in @code{tar}.
1789 @samp{-l} in @code{diff}.
1792 @samp{-L} in @code{cmp}.
1794 @item print-data-base
1797 @item print-directory
1800 @item print-file-name
1801 @samp{-o} in @code{nm}.
1804 @samp{-s} in @code{nm}.
1807 @samp{-p} in @code{wdiff}.
1810 @samp{-p} in @code{ed}.
1813 Specify an HTTP proxy.
1816 @samp{-X} in @code{shar}.
1822 Used in many programs to inhibit the usual output. Every
1823 program accepting @samp{--quiet} should accept @samp{--silent} as a
1827 @samp{-Q} in @code{shar}
1830 @samp{-Q} in @code{ls}.
1833 @samp{-n} in @code{diff}.
1836 Used in @code{gawk}.
1838 @item read-full-blocks
1839 @samp{-B} in @code{tar}.
1848 @samp{-R} in @code{tar}.
1851 Used in @code{chgrp}, @code{chown}, @code{cp}, @code{ls}, @code{diff},
1854 @item reference-limit
1855 Used in @code{makeinfo}.
1858 @samp{-r} in @code{ptx}.
1861 @samp{-r} in @code{tac} and @code{etags}.
1864 @samp{-r} in @code{uname}.
1867 @samp{-R} in @code{m4}.
1870 @samp{-r} in @code{objdump}.
1873 @samp{-r} in @code{cpio}.
1876 @samp{-i} in @code{xargs}.
1878 @item report-identical-files
1879 @samp{-s} in @code{diff}.
1881 @item reset-access-time
1882 @samp{-a} in @code{cpio}.
1885 @samp{-r} in @code{ls} and @code{nm}.
1888 @samp{-f} in @code{diff}.
1890 @item right-side-defs
1891 @samp{-R} in @code{ptx}.
1894 @samp{-s} in @code{tar}.
1896 @item same-permissions
1897 @samp{-p} in @code{tar}.
1900 @samp{-g} in @code{stty}.
1905 @item sentence-regexp
1906 @samp{-S} in @code{ptx}.
1909 @samp{-S} in @code{du}.
1912 @samp{-s} in @code{tac}.
1915 Used by @code{recode} to chose files or pipes for sequencing passes.
1918 @samp{-s} in @code{su}.
1921 @samp{-A} in @code{cat}.
1923 @item show-c-function
1924 @samp{-p} in @code{diff}.
1927 @samp{-E} in @code{cat}.
1929 @item show-function-line
1930 @samp{-F} in @code{diff}.
1933 @samp{-T} in @code{cat}.
1936 Used in many programs to inhibit the usual output.
1937 Every program accepting
1938 @samp{--silent} should accept @samp{--quiet} as a synonym.
1941 @samp{-s} in @code{ls}.
1944 Specify a file descriptor for a network server to use for its socket,
1945 instead of opening and binding a new socket. This provides a way to
1946 run, in a nonpriveledged process, a server that normally needs a
1947 reserved port number.
1953 @samp{-W source} in @code{gawk}.
1956 @samp{-S} in @code{tar}.
1958 @item speed-large-files
1959 @samp{-H} in @code{diff}.
1962 @samp{-E} in @code{unshar}.
1964 @item split-size-limit
1965 @samp{-L} in @code{shar}.
1968 @samp{-s} in @code{cat}.
1971 @samp{-w} in @code{wdiff}.
1974 @samp{-y} in @code{wdiff}.
1977 Used in @code{tar} and @code{diff} to specify which file within
1978 a directory to start processing with.
1981 @samp{-s} in @code{wdiff}.
1983 @item stdin-file-list
1984 @samp{-S} in @code{shar}.
1990 @samp{-s} in @code{recode}.
1993 @samp{-s} in @code{install}.
1996 @samp{-s} in @code{strip}.
1999 @samp{-S} in @code{strip}.
2002 @samp{-s} in @code{shar}.
2005 @samp{-S} in @code{cp}, @code{ln}, @code{mv}.
2008 @samp{-b} in @code{csplit}.
2011 @samp{-s} in @code{gprof}.
2014 @samp{-s} in @code{du}.
2017 @samp{-s} in @code{ln}.
2020 Used in GDB and @code{objdump}.
2023 @samp{-s} in @code{m4}.
2026 @samp{-s} in @code{uname}.
2029 @samp{-t} in @code{expand} and @code{unexpand}.
2032 @samp{-T} in @code{ls}.
2035 @samp{-T} in @code{tput} and @code{ul}.
2036 @samp{-t} in @code{wdiff}.
2039 @samp{-a} in @code{diff}.
2042 @samp{-T} in @code{shar}.
2045 Used in @code{ls} and @code{touch}.
2048 Specify how long to wait before giving up on some operation.
2051 @samp{-O} in @code{tar}.
2054 @samp{-c} in @code{du}.
2057 @samp{-t} in Make, @code{ranlib}, and @code{recode}.
2060 @samp{-t} in @code{m4}.
2063 @samp{-t} in @code{hello};
2064 @samp{-W traditional} in @code{gawk};
2065 @samp{-G} in @code{ed}, @code{m4}, and @code{ptx}.
2071 @samp{-t} in @code{ctags}.
2073 @item typedefs-and-c++
2074 @samp{-T} in @code{ctags}.
2077 @samp{-t} in @code{ptx}.
2080 @samp{-z} in @code{tar}.
2083 @samp{-u} in @code{cpio}.
2086 @samp{-U} in @code{m4}.
2088 @item undefined-only
2089 @samp{-u} in @code{nm}.
2092 @samp{-u} in @code{cp}, @code{ctags}, @code{mv}, @code{tar}.
2095 Used in @code{gawk}; same as @samp{--help}.
2098 @samp{-B} in @code{shar}.
2100 @item vanilla-operation
2101 @samp{-V} in @code{shar}.
2104 Print more information about progress. Many programs support this.
2107 @samp{-W} in @code{tar}.
2110 Print the version number.
2112 @item version-control
2113 @samp{-V} in @code{cp}, @code{ln}, @code{mv}.
2116 @samp{-v} in @code{ctags}.
2119 @samp{-V} in @code{tar}.
2124 @item whole-size-limit
2125 @samp{-l} in @code{shar}.
2128 @samp{-w} in @code{ls} and @code{ptx}.
2131 @samp{-W} in @code{ptx}.
2134 @samp{-T} in @code{who}.
2137 @samp{-z} in @code{gprof}.
2141 @section Memory Usage
2142 @cindex memory usage
2144 If a program typically uses just a few meg of memory, don't bother making any
2145 effort to reduce memory usage. For example, if it is impractical for
2146 other reasons to operate on files more than a few meg long, it is
2147 reasonable to read entire input files into core to operate on them.
2149 However, for programs such as @code{cat} or @code{tail}, that can
2150 usefully operate on very large files, it is important to avoid using a
2151 technique that would artificially limit the size of files it can handle.
2152 If a program works by lines and could be applied to arbitrary
2153 user-supplied input files, it should keep only a line in memory, because
2154 this is not very hard and users will want to be able to operate on input
2155 files that are bigger than will fit in core all at once.
2157 If your program creates complicated data structures, just make them in
2158 core and give a fatal error if @code{malloc} returns zero.
2164 Programs should be prepared to operate when @file{/usr} and @file{/etc}
2165 are read-only file systems. Thus, if the program manages log files,
2166 lock files, backup files, score files, or any other files which are
2167 modified for internal purposes, these files should not be stored in
2168 @file{/usr} or @file{/etc}.
2170 There are two exceptions. @file{/etc} is used to store system
2171 configuration information; it is reasonable for a program to modify
2172 files in @file{/etc} when its job is to update the system configuration.
2173 Also, if the user explicitly asks to modify one file in a directory, it
2174 is reasonable for the program to store other files in the same
2178 @chapter Making The Best Use of C
2180 This chapter provides advice on how best to use the C language
2181 when writing GNU software.
2184 * Formatting:: Formatting Your Source Code
2185 * Comments:: Commenting Your Work
2186 * Syntactic Conventions:: Clean Use of C Constructs
2187 * Names:: Naming Variables, Functions, and Files
2188 * System Portability:: Portability between different operating systems
2189 * CPU Portability:: Supporting the range of CPU types
2190 * System Functions:: Portability and ``standard'' library functions
2191 * Internationalization:: Techniques for internationalization
2192 * Character Set:: Use ASCII by default.
2193 * Quote Characters:: Use `...' in the C locale.
2194 * Mmap:: How you can safely use @code{mmap}.
2198 @section Formatting Your Source Code
2199 @cindex formatting source code
2202 @cindex braces, in C source
2203 It is important to put the open-brace that starts the body of a C
2204 function in column one, and avoid putting any other open-brace or
2205 open-parenthesis or open-bracket in column one. Several tools look
2206 for open-braces in column one to find the beginnings of C functions.
2207 These tools will not work on code not formatted that way.
2209 It is also important for function definitions to start the name of the
2210 function in column one. This helps people to search for function
2211 definitions, and may also help certain tools recognize them. Thus,
2212 using Standard C syntax, the format is this:
2216 concat (char *s1, char *s2)
2223 or, if you want to use traditional C syntax, format the definition like
2228 concat (s1, s2) /* Name starts in column one here */
2230 @{ /* Open brace in column one here */
2235 In Standard C, if the arguments don't fit nicely on one line,
2240 lots_of_args (int an_integer, long a_long, short a_short,
2241 double a_double, float a_float)
2245 The rest of this section gives our recommendations for other aspects of
2246 C formatting style, which is also the default style of the @code{indent}
2247 program in version 1.2 and newer. It corresponds to the options
2250 -nbad -bap -nbc -bbo -bl -bli2 -bls -ncdb -nce -cp1 -cs -di2
2251 -ndj -nfc1 -nfca -hnl -i2 -ip5 -lp -pcs -psl -nsc -nsob
2254 We don't think of these recommendations as requirements, because it
2255 causes no problems for users if two different programs have different
2258 But whatever style you use, please use it consistently, since a mixture
2259 of styles within one program tends to look ugly. If you are
2260 contributing changes to an existing program, please follow the style of
2263 For the body of the function, our recommended style looks like this:
2275 return ++x + bar ();
2279 @cindex spaces before open-paren
2280 We find it easier to read a program when it has spaces before the
2281 open-parentheses and after the commas. Especially after the commas.
2283 When you split an expression into multiple lines, split it
2284 before an operator, not after one. Here is the right way:
2286 @cindex expressions, splitting
2288 if (foo_this_is_long && bar > win (x, y, z)
2289 && remaining_condition)
2292 Try to avoid having two operators of different precedence at the same
2293 level of indentation. For example, don't write this:
2296 mode = (inmode[j] == VOIDmode
2297 || GET_MODE_SIZE (outmode[j]) > GET_MODE_SIZE (inmode[j])
2298 ? outmode[j] : inmode[j]);
2301 Instead, use extra parentheses so that the indentation shows the nesting:
2304 mode = ((inmode[j] == VOIDmode
2305 || (GET_MODE_SIZE (outmode[j]) > GET_MODE_SIZE (inmode[j])))
2306 ? outmode[j] : inmode[j]);
2309 Insert extra parentheses so that Emacs will indent the code properly.
2310 For example, the following indentation looks nice if you do it by hand,
2313 v = rup->ru_utime.tv_sec*1000 + rup->ru_utime.tv_usec/1000
2314 + rup->ru_stime.tv_sec*1000 + rup->ru_stime.tv_usec/1000;
2318 but Emacs would alter it. Adding a set of parentheses produces
2319 something that looks equally nice, and which Emacs will preserve:
2322 v = (rup->ru_utime.tv_sec*1000 + rup->ru_utime.tv_usec/1000
2323 + rup->ru_stime.tv_sec*1000 + rup->ru_stime.tv_usec/1000);
2326 Format do-while statements like this:
2338 Please use formfeed characters (control-L) to divide the program into
2339 pages at logical places (but not within a function). It does not matter
2340 just how long the pages are, since they do not have to fit on a printed
2341 page. The formfeeds should appear alone on lines by themselves.
2344 @section Commenting Your Work
2347 Every program should start with a comment saying briefly what it is for.
2348 Example: @samp{fmt - filter for simple filling of text}. This comment
2349 should be at the top of the source file containing the @samp{main}
2350 function of the program.
2352 Also, please write a brief comment at the start of each source file,
2353 with the file name and a line or two about the overall purpose of the
2356 Please write the comments in a GNU program in English, because English
2357 is the one language that nearly all programmers in all countries can
2358 read. If you do not write English well, please write comments in
2359 English as well as you can, then ask other people to help rewrite them.
2360 If you can't write comments in English, please find someone to work with
2361 you and translate your comments into English.
2363 Please put a comment on each function saying what the function does,
2364 what sorts of arguments it gets, and what the possible values of
2365 arguments mean and are used for. It is not necessary to duplicate in
2366 words the meaning of the C argument declarations, if a C type is being
2367 used in its customary fashion. If there is anything nonstandard about
2368 its use (such as an argument of type @code{char *} which is really the
2369 address of the second character of a string, not the first), or any
2370 possible values that would not work the way one would expect (such as,
2371 that strings containing newlines are not guaranteed to work), be sure
2374 Also explain the significance of the return value, if there is one.
2376 Please put two spaces after the end of a sentence in your comments, so
2377 that the Emacs sentence commands will work. Also, please write
2378 complete sentences and capitalize the first word. If a lower-case
2379 identifier comes at the beginning of a sentence, don't capitalize it!
2380 Changing the spelling makes it a different identifier. If you don't
2381 like starting a sentence with a lower case letter, write the sentence
2382 differently (e.g., ``The identifier lower-case is @dots{}'').
2384 The comment on a function is much clearer if you use the argument
2385 names to speak about the argument values. The variable name itself
2386 should be lower case, but write it in upper case when you are speaking
2387 about the value rather than the variable itself. Thus, ``the inode
2388 number NODE_NUM'' rather than ``an inode''.
2390 There is usually no purpose in restating the name of the function in
2391 the comment before it, because the reader can see that for himself.
2392 There might be an exception when the comment is so long that the function
2393 itself would be off the bottom of the screen.
2395 There should be a comment on each static variable as well, like this:
2398 /* Nonzero means truncate lines in the display;
2399 zero means continue them. */
2403 @cindex conditionals, comments for
2404 @cindex @code{#endif}, commenting
2405 Every @samp{#endif} should have a comment, except in the case of short
2406 conditionals (just a few lines) that are not nested. The comment should
2407 state the condition of the conditional that is ending, @emph{including
2408 its sense}. @samp{#else} should have a comment describing the condition
2409 @emph{and sense} of the code that follows. For example:
2417 #endif /* not foo */
2427 but, by contrast, write the comments this way for a @samp{#ifndef}:
2440 #endif /* not foo */
2444 @node Syntactic Conventions
2445 @section Clean Use of C Constructs
2446 @cindex syntactic conventions
2448 @cindex implicit @code{int}
2449 @cindex function argument, declaring
2450 Please explicitly declare the types of all objects. For example, you
2451 should explicitly declare all arguments to functions, and you should
2452 declare functions to return @code{int} rather than omitting the
2455 @cindex compiler warnings
2456 @cindex @samp{-Wall} compiler option
2457 Some programmers like to use the GCC @samp{-Wall} option, and change the
2458 code whenever it issues a warning. If you want to do this, then do.
2459 Other programmers prefer not to use @samp{-Wall}, because it gives
2460 warnings for valid and legitimate code which they do not want to change.
2461 If you want to do this, then do. The compiler should be your servant,
2464 Declarations of external functions and functions to appear later in the
2465 source file should all go in one place near the beginning of the file
2466 (somewhere before the first function definition in the file), or else
2467 should go in a header file. Don't put @code{extern} declarations inside
2470 @cindex temporary variables
2471 It used to be common practice to use the same local variables (with
2472 names like @code{tem}) over and over for different values within one
2473 function. Instead of doing this, it is better to declare a separate local
2474 variable for each distinct purpose, and give it a name which is
2475 meaningful. This not only makes programs easier to understand, it also
2476 facilitates optimization by good compilers. You can also move the
2477 declaration of each local variable into the smallest scope that includes
2478 all its uses. This makes the program even cleaner.
2480 Don't use local variables or parameters that shadow global identifiers.
2482 @cindex multiple variables in a line
2483 Don't declare multiple variables in one declaration that spans lines.
2484 Start a new declaration on each line, instead. For example, instead
2510 (If they are global variables, each should have a comment preceding it
2513 When you have an @code{if}-@code{else} statement nested in another
2514 @code{if} statement, always put braces around the @code{if}-@code{else}.
2515 Thus, never write like this:
2538 If you have an @code{if} statement nested inside of an @code{else}
2539 statement, either write @code{else if} on one line, like this,
2549 with its @code{then}-part indented like the preceding @code{then}-part,
2550 or write the nested @code{if} within braces like this:
2562 Don't declare both a structure tag and variables or typedefs in the
2563 same declaration. Instead, declare the structure tag separately
2564 and then use it to declare the variables or typedefs.
2566 Try to avoid assignments inside @code{if}-conditions. For example,
2570 if ((foo = (char *) malloc (sizeof *foo)) == 0)
2571 fatal ("virtual memory exhausted");
2575 instead, write this:
2578 foo = (char *) malloc (sizeof *foo);
2580 fatal ("virtual memory exhausted");
2584 Don't make the program ugly to placate @code{lint}. Please don't insert any
2585 casts to @code{void}. Zero without a cast is perfectly fine as a null
2586 pointer constant, except when calling a varargs function.
2589 @section Naming Variables, Functions, and Files
2591 @cindex names of variables, functions, and files
2592 The names of global variables and functions in a program serve as
2593 comments of a sort. So don't choose terse names---instead, look for
2594 names that give useful information about the meaning of the variable or
2595 function. In a GNU program, names should be English, like other
2598 Local variable names can be shorter, because they are used only within
2599 one context, where (presumably) comments explain their purpose.
2601 Try to limit your use of abbreviations in symbol names. It is ok to
2602 make a few abbreviations, explain what they mean, and then use them
2603 frequently, but don't use lots of obscure abbreviations.
2605 Please use underscores to separate words in a name, so that the Emacs
2606 word commands can be useful within them. Stick to lower case; reserve
2607 upper case for macros and @code{enum} constants, and for name-prefixes
2608 that follow a uniform convention.
2610 For example, you should use names like @code{ignore_space_change_flag};
2611 don't use names like @code{iCantReadThis}.
2613 Variables that indicate whether command-line options have been
2614 specified should be named after the meaning of the option, not after
2615 the option-letter. A comment should state both the exact meaning of
2616 the option and its letter. For example,
2620 /* Ignore changes in horizontal whitespace (-b). */
2621 int ignore_space_change_flag;
2625 When you want to define names with constant integer values, use
2626 @code{enum} rather than @samp{#define}. GDB knows about enumeration
2629 @cindex file-name limitations
2631 You might want to make sure that none of the file names would conflict
2632 if the files were loaded onto an MS-DOS file system which shortens the
2633 names. You can use the program @code{doschk} to test for this.
2635 Some GNU programs were designed to limit themselves to file names of 14
2636 characters or less, to avoid file name conflicts if they are read into
2637 older System V systems. Please preserve this feature in the existing
2638 GNU programs that have it, but there is no need to do this in new GNU
2639 programs. @code{doschk} also reports file names longer than 14
2642 @node System Portability
2643 @section Portability between System Types
2644 @cindex portability, between system types
2646 In the Unix world, ``portability'' refers to porting to different Unix
2647 versions. For a GNU program, this kind of portability is desirable, but
2650 The primary purpose of GNU software is to run on top of the GNU kernel,
2651 compiled with the GNU C compiler, on various types of @sc{cpu}. So the
2652 kinds of portability that are absolutely necessary are quite limited.
2653 But it is important to support Linux-based GNU systems, since they
2654 are the form of GNU that is popular.
2656 Beyond that, it is good to support the other free operating systems
2657 (*BSD), and it is nice to support other Unix-like systems if you want
2658 to. Supporting a variety of Unix-like systems is desirable, although
2659 not paramount. It is usually not too hard, so you may as well do it.
2660 But you don't have to consider it an obligation, if it does turn out to
2664 The easiest way to achieve portability to most Unix-like systems is to
2665 use Autoconf. It's unlikely that your program needs to know more
2666 information about the host platform than Autoconf can provide, simply
2667 because most of the programs that need such knowledge have already been
2670 Avoid using the format of semi-internal data bases (e.g., directories)
2671 when there is a higher-level alternative (@code{readdir}).
2673 @cindex non-@sc{posix} systems, and portability
2674 As for systems that are not like Unix, such as MSDOS, Windows, VMS, MVS,
2675 and older Macintosh systems, supporting them is often a lot of work.
2676 When that is the case, it is better to spend your time adding features
2677 that will be useful on GNU and GNU/Linux, rather than on supporting
2678 other incompatible systems.
2680 If you do support Windows, please do not abbreviate it as ``win''. In
2681 hacker terminology, calling something a ``win'' is a form of praise.
2682 You're free to praise Microsoft Windows on your own if you want, but
2683 please don't do this in GNU packages. Instead of abbreviating
2684 ``Windows'' to ``un'', you can write it in full or abbreviate it to
2685 ``woe'' or ``w''. In GNU Emacs, for instance, we use @samp{w32} in
2686 file names of Windows-specific files, but the macro for Windows
2687 conditionals is called @code{WINDOWSNT}.
2689 It is a good idea to define the ``feature test macro''
2690 @code{_GNU_SOURCE} when compiling your C files. When you compile on GNU
2691 or GNU/Linux, this will enable the declarations of GNU library extension
2692 functions, and that will usually give you a compiler error message if
2693 you define the same function names in some other way in your program.
2694 (You don't have to actually @emph{use} these functions, if you prefer
2695 to make the program more portable to other systems.)
2697 But whether or not you use these GNU extensions, you should avoid
2698 using their names for any other meanings. Doing so would make it hard
2699 to move your code into other GNU programs.
2701 @node CPU Portability
2702 @section Portability between @sc{cpu}s
2704 @cindex data types, and portability
2705 @cindex portability, and data types
2706 Even GNU systems will differ because of differences among @sc{cpu}
2707 types---for example, difference in byte ordering and alignment
2708 requirements. It is absolutely essential to handle these differences.
2709 However, don't make any effort to cater to the possibility that an
2710 @code{int} will be less than 32 bits. We don't support 16-bit machines
2713 Similarly, don't make any effort to cater to the possibility that
2714 @code{long} will be smaller than predefined types like @code{size_t}.
2715 For example, the following code is ok:
2718 printf ("size = %lu\n", (unsigned long) sizeof array);
2719 printf ("diff = %ld\n", (long) (pointer2 - pointer1));
2722 1989 Standard C requires this to work, and we know of only one
2723 counterexample: 64-bit programs on Microsoft Windows. We will
2724 leave it to those who want to port GNU programs to that environment
2725 to figure out how to do it.
2727 Predefined file-size types like @code{off_t} are an exception: they are
2728 longer than @code{long} on many platforms, so code like the above won't
2729 work with them. One way to print an @code{off_t} value portably is to
2730 print its digits yourself, one by one.
2732 Don't assume that the address of an @code{int} object is also the
2733 address of its least-significant byte. This is false on big-endian
2734 machines. Thus, don't make the following mistake:
2739 while ((c = getchar()) != EOF)
2740 write(file_descriptor, &c, 1);
2743 It used to be ok to not worry about the difference between pointers
2744 and integers when passing arguments to functions. However, on most
2745 modern 64-bit machines pointers are wider than @code{int}.
2746 Conversely, integer types like @code{long long int} and @code{off_t}
2747 are wider than pointers on most modern 32-bit machines. Hence it's
2748 often better nowadays to use prototypes to define functions whose
2749 argument types are not trivial.
2751 In particular, if functions accept varying argument counts or types
2752 they should be declared using prototypes containing @samp{...} and
2753 defined using @file{stdarg.h}. For an example of this, please see the
2754 @uref{http://www.gnu.org/software/gnulib/, Gnulib} error module, which
2755 declares and defines the following function:
2758 /* Print a message with `fprintf (stderr, FORMAT, ...)';
2759 if ERRNUM is nonzero, follow it with ": " and strerror (ERRNUM).
2760 If STATUS is nonzero, terminate the program with `exit (STATUS)'. */
2762 void error (int status, int errnum, const char *format, ...);
2765 A simple way to use the Gnulib error module is to obtain the two
2766 source files @file{error.c} and @file{error.h} from the Gnulib library
2767 source code repository at
2768 @uref{http://savannah.gnu.org/cgi-bin/viewcvs/gnulib/gnulib/lib/}.
2769 Here's a sample use:
2776 char *program_name = "myprogram";
2779 xfopen (char const *name)
2781 FILE *fp = fopen (name, "r");
2783 error (1, errno, "cannot read %s", name);
2788 @cindex casting pointers to integers
2789 Avoid casting pointers to integers if you can. Such casts greatly
2790 reduce portability, and in most programs they are easy to avoid. In the
2791 cases where casting pointers to integers is essential---such as, a Lisp
2792 interpreter which stores type information as well as an address in one
2793 word---you'll have to make explicit provisions to handle different word
2794 sizes. You will also need to make provision for systems in which the
2795 normal range of addresses you can get from @code{malloc} starts far away
2798 @node System Functions
2799 @section Calling System Functions
2800 @cindex library functions, and portability
2801 @cindex portability, and library functions
2803 C implementations differ substantially. Standard C reduces but does
2804 not eliminate the incompatibilities; meanwhile, many GNU packages still
2805 support pre-standard compilers because this is not hard to do. This
2806 chapter gives recommendations for how to use the more-or-less standard C
2807 library functions to avoid unnecessary loss of portability.
2811 Don't use the return value of @code{sprintf}. It returns the number of
2812 characters written on some systems, but not on all systems.
2815 Be aware that @code{vfprintf} is not always available.
2818 @code{main} should be declared to return type @code{int}. It should
2819 terminate either by calling @code{exit} or by returning the integer
2820 status code; make sure it cannot ever return an undefined value.
2822 @cindex declaration for system functions
2824 Don't declare system functions explicitly.
2826 Almost any declaration for a system function is wrong on some system.
2827 To minimize conflicts, leave it to the system header files to declare
2828 system functions. If the headers don't declare a function, let it
2831 While it may seem unclean to use a function without declaring it, in
2832 practice this works fine for most system library functions on the
2833 systems where this really happens; thus, the disadvantage is only
2834 theoretical. By contrast, actual declarations have frequently caused
2838 If you must declare a system function, don't specify the argument types.
2839 Use an old-style declaration, not a Standard C prototype. The more you
2840 specify about the function, the more likely a conflict.
2843 In particular, don't unconditionally declare @code{malloc} or
2846 Most GNU programs use those functions just once, in functions
2847 conventionally named @code{xmalloc} and @code{xrealloc}. These
2848 functions call @code{malloc} and @code{realloc}, respectively, and
2851 Because @code{xmalloc} and @code{xrealloc} are defined in your program,
2852 you can declare them in other files without any risk of type conflict.
2854 On most systems, @code{int} is the same length as a pointer; thus, the
2855 calls to @code{malloc} and @code{realloc} work fine. For the few
2856 exceptional systems (mostly 64-bit machines), you can use
2857 @strong{conditionalized} declarations of @code{malloc} and
2858 @code{realloc}---or put these declarations in configuration files
2859 specific to those systems.
2861 @cindex string library functions
2863 The string functions require special treatment. Some Unix systems have
2864 a header file @file{string.h}; others have @file{strings.h}. Neither
2865 file name is portable. There are two things you can do: use Autoconf to
2866 figure out which file to include, or don't include either file.
2869 If you don't include either strings file, you can't get declarations for
2870 the string functions from the header file in the usual way.
2872 That causes less of a problem than you might think. The newer standard
2873 string functions should be avoided anyway because many systems still
2874 don't support them. The string functions you can use are these:
2877 strcpy strncpy strcat strncat
2878 strlen strcmp strncmp
2882 The copy and concatenate functions work fine without a declaration as
2883 long as you don't use their values. Using their values without a
2884 declaration fails on systems where the width of a pointer differs from
2885 the width of @code{int}, and perhaps in other cases. It is trivial to
2886 avoid using their values, so do that.
2888 The compare functions and @code{strlen} work fine without a declaration
2889 on most systems, possibly all the ones that GNU software runs on.
2890 You may find it necessary to declare them @strong{conditionally} on a
2893 The search functions must be declared to return @code{char *}. Luckily,
2894 there is no variation in the data type they return. But there is
2895 variation in their names. Some systems give these functions the names
2896 @code{index} and @code{rindex}; other systems use the names
2897 @code{strchr} and @code{strrchr}. Some systems support both pairs of
2898 names, but neither pair works on all systems.
2900 You should pick a single pair of names and use it throughout your
2901 program. (Nowadays, it is better to choose @code{strchr} and
2902 @code{strrchr} for new programs, since those are the standard
2903 names.) Declare both of those names as functions returning @code{char
2904 *}. On systems which don't support those names, define them as macros
2905 in terms of the other pair. For example, here is what to put at the
2906 beginning of your file (or in a header) if you want to use the names
2907 @code{strchr} and @code{strrchr} throughout:
2911 #define strchr index
2913 #ifndef HAVE_STRRCHR
2914 #define strrchr rindex
2922 Here we assume that @code{HAVE_STRCHR} and @code{HAVE_STRRCHR} are
2923 macros defined in systems where the corresponding functions exist.
2924 One way to get them properly defined is to use Autoconf.
2926 @node Internationalization
2927 @section Internationalization
2928 @cindex internationalization
2931 GNU has a library called GNU gettext that makes it easy to translate the
2932 messages in a program into various languages. You should use this
2933 library in every program. Use English for the messages as they appear
2934 in the program, and let gettext provide the way to translate them into
2937 Using GNU gettext involves putting a call to the @code{gettext} macro
2938 around each string that might need translation---like this:
2941 printf (gettext ("Processing file `%s'..."));
2945 This permits GNU gettext to replace the string @code{"Processing file
2946 `%s'..."} with a translated version.
2948 Once a program uses gettext, please make a point of writing calls to
2949 @code{gettext} when you add new strings that call for translation.
2951 Using GNU gettext in a package involves specifying a @dfn{text domain
2952 name} for the package. The text domain name is used to separate the
2953 translations for this package from the translations for other packages.
2954 Normally, the text domain name should be the same as the name of the
2955 package---for example, @samp{fileutils} for the GNU file utilities.
2957 @cindex message text, and internationalization
2958 To enable gettext to work well, avoid writing code that makes
2959 assumptions about the structure of words or sentences. When you want
2960 the precise text of a sentence to vary depending on the data, use two or
2961 more alternative string constants each containing a complete sentences,
2962 rather than inserting conditionalized words or phrases into a single
2965 Here is an example of what not to do:
2968 printf ("%d file%s processed", nfiles,
2969 nfiles != 1 ? "s" : "");
2973 The problem with that example is that it assumes that plurals are made
2974 by adding `s'. If you apply gettext to the format string, like this,
2977 printf (gettext ("%d file%s processed"), nfiles,
2978 nfiles != 1 ? "s" : "");
2982 the message can use different words, but it will still be forced to use
2983 `s' for the plural. Here is a better way:
2986 printf ((nfiles != 1 ? "%d files processed"
2987 : "%d file processed"),
2992 This way, you can apply gettext to each of the two strings
2996 printf ((nfiles != 1 ? gettext ("%d files processed")
2997 : gettext ("%d file processed")),
3002 This can be any method of forming the plural of the word for ``file'', and
3003 also handles languages that require agreement in the word for
3006 A similar problem appears at the level of sentence structure with this
3010 printf ("# Implicit rule search has%s been done.\n",
3011 f->tried_implicit ? "" : " not");
3015 Adding @code{gettext} calls to this code cannot give correct results for
3016 all languages, because negation in some languages requires adding words
3017 at more than one place in the sentence. By contrast, adding
3018 @code{gettext} calls does the job straightfowardly if the code starts
3022 printf (f->tried_implicit
3023 ? "# Implicit rule search has been done.\n",
3024 : "# Implicit rule search has not been done.\n");
3029 @section Character Set
3030 @cindex character set
3032 @cindex ASCII characters
3033 @cindex non-ASCII characters
3035 Sticking to the ASCII character set (plain text, 7-bit characters) is
3036 preferred in GNU source code comments, text documents, and other
3037 contexts, unless there is good reason to do something else because of
3038 the application domain. For example, if source code deals with the
3039 French Revolutionary calendar, it is OK if its literal strings contain
3040 accented characters in month names like ``Flor@'eal''. Also, it is OK
3041 to use non-ASCII characters to represent proper names of contributors in
3042 change logs (@pxref{Change Logs}).
3044 If you need to use non-ASCII characters, you should normally stick with
3045 one encoding, as one cannot in general mix encodings reliably.
3048 @node Quote Characters
3049 @section Quote Characters
3050 @cindex quote characters
3051 @cindex locale-specific quote characters
3053 @cindex grave accent
3055 In the C locale, GNU programs should stick to plain ASCII for quotation
3056 characters in messages to users: preferably 0x60 (@samp{`}) for left
3057 quotes and 0x27 (@samp{'}) for right quotes. It is ok, but not
3058 required, to use locale-specific quotes in other locales.
3060 The @uref{http://www.gnu.org/software/gnulib/, Gnulib} @code{quote} and
3061 @code{quotearg} modules provide a reasonably straightforward way to
3062 support locale-specific quote characters, as well as taking care of
3063 other issues, such as quoting a filename that itself contains a quote
3064 character. See the Gnulib documentation for usage details.
3066 In any case, the documentation for your program should clearly specify
3067 how it does quoting, if different than the preferred method of @samp{`}
3068 and @samp{'}. This is especially important if the output of your
3069 program is ever likely to be parsed by another program.
3071 Quotation characters are a difficult area in the computing world at
3072 this time: there are no true left or right quote characters in Latin1;
3073 the @samp{`} character we use was standardized there as a grave
3074 accent. Moreover, Latin1 is still not universally usable.
3076 Unicode contains the unambiguous quote characters required, and its
3077 common encoding UTF-8 is upward compatible with Latin1. However,
3078 Unicode and UTF-8 are not universally well-supported, either.
3080 This may change over the next few years, and then we will revisit
3088 Don't assume that @code{mmap} either works on all files or fails
3089 for all files. It may work on some files and fail on others.
3091 The proper way to use @code{mmap} is to try it on the specific file for
3092 which you want to use it---and if @code{mmap} doesn't work, fall back on
3093 doing the job in another way using @code{read} and @code{write}.
3095 The reason this precaution is needed is that the GNU kernel (the HURD)
3096 provides a user-extensible file system, in which there can be many
3097 different kinds of ``ordinary files.'' Many of them support
3098 @code{mmap}, but some do not. It is important to make programs handle
3099 all these kinds of files.
3102 @chapter Documenting Programs
3103 @cindex documentation
3105 A GNU program should ideally come with full free documentation, adequate
3106 for both reference and tutorial purposes. If the package can be
3107 programmed or extended, the documentation should cover programming or
3108 extending it, as well as just using it.
3111 * GNU Manuals:: Writing proper manuals.
3112 * Doc Strings and Manuals:: Compiling doc strings doesn't make a manual.
3113 * Manual Structure Details:: Specific structure conventions.
3114 * License for Manuals:: Writing the distribution terms for a manual.
3115 * Manual Credits:: Giving credit to documentation contributors.
3116 * Printed Manuals:: Mentioning the printed manual.
3117 * NEWS File:: NEWS files supplement manuals.
3118 * Change Logs:: Recording Changes
3119 * Man Pages:: Man pages are secondary.
3120 * Reading other Manuals:: How far you can go in learning
3125 @section GNU Manuals
3127 The preferred document format for the GNU system is the Texinfo
3128 formatting language. Every GNU package should (ideally) have
3129 documentation in Texinfo both for reference and for learners. Texinfo
3130 makes it possible to produce a good quality formatted book, using
3131 @TeX{}, and to generate an Info file. It is also possible to generate
3132 HTML output from Texinfo source. See the Texinfo manual, either the
3133 hardcopy, or the on-line version available through @code{info} or the
3134 Emacs Info subsystem (@kbd{C-h i}).
3136 Nowadays some other formats such as Docbook and Sgmltexi can be
3137 converted automatically into Texinfo. It is ok to produce the Texinfo
3138 documentation by conversion this way, as long as it gives good results.
3140 Programmers often find it most natural to structure the documentation
3141 following the structure of the implementation, which they know. But
3142 this structure is not necessarily good for explaining how to use the
3143 program; it may be irrelevant and confusing for a user.
3145 At every level, from the sentences in a paragraph to the grouping of
3146 topics into separate manuals, the right way to structure documentation
3147 is according to the concepts and questions that a user will have in mind
3148 when reading it. Sometimes this structure of ideas matches the
3149 structure of the implementation of the software being documented---but
3150 often they are different. Often the most important part of learning to
3151 write good documentation is learning to notice when you are structuring
3152 the documentation like the implementation, and think about better
3155 For example, each program in the GNU system probably ought to be
3156 documented in one manual; but this does not mean each program should
3157 have its own manual. That would be following the structure of the
3158 implementation, rather than the structure that helps the user
3161 Instead, each manual should cover a coherent @emph{topic}. For example,
3162 instead of a manual for @code{diff} and a manual for @code{diff3}, we
3163 have one manual for ``comparison of files'' which covers both of those
3164 programs, as well as @code{cmp}. By documenting these programs
3165 together, we can make the whole subject clearer.
3167 The manual which discusses a program should certainly document all of
3168 the program's command-line options and all of its commands. It should
3169 give examples of their use. But don't organize the manual as a list
3170 of features. Instead, organize it logically, by subtopics. Address
3171 the questions that a user will ask when thinking about the job that
3172 the program does. Don't just tell the reader what each feature can
3173 do---say what jobs it is good for, and show how to use it for those
3174 jobs. Explain what is recommended usage, and what kinds of usage
3177 In general, a GNU manual should serve both as tutorial and reference.
3178 It should be set up for convenient access to each topic through Info,
3179 and for reading straight through (appendixes aside). A GNU manual
3180 should give a good introduction to a beginner reading through from the
3181 start, and should also provide all the details that hackers want.
3182 The Bison manual is a good example of this---please take a look at it
3183 to see what we mean.
3185 That is not as hard as it first sounds. Arrange each chapter as a
3186 logical breakdown of its topic, but order the sections, and write their
3187 text, so that reading the chapter straight through makes sense. Do
3188 likewise when structuring the book into chapters, and when structuring a
3189 section into paragraphs. The watchword is, @emph{at each point, address
3190 the most fundamental and important issue raised by the preceding text.}
3192 If necessary, add extra chapters at the beginning of the manual which
3193 are purely tutorial and cover the basics of the subject. These provide
3194 the framework for a beginner to understand the rest of the manual. The
3195 Bison manual provides a good example of how to do this.
3197 To serve as a reference, a manual should have an Index that list all the
3198 functions, variables, options, and important concepts that are part of
3199 the program. One combined Index should do for a short manual, but
3200 sometimes for a complex package it is better to use multiple indices.
3201 The Texinfo manual includes advice on preparing good index entries, see
3202 @ref{Index Entries, , Making Index Entries, texinfo, GNU Texinfo}, and
3203 see @ref{Indexing Commands, , Defining the Entries of an
3204 Index, texinfo, GNU Texinfo}.
3206 Don't use Unix man pages as a model for how to write GNU documentation;
3207 most of them are terse, badly structured, and give inadequate
3208 explanation of the underlying concepts. (There are, of course, some
3209 exceptions.) Also, Unix man pages use a particular format which is
3210 different from what we use in GNU manuals.
3212 Please include an email address in the manual for where to report
3213 bugs @emph{in the text of the manual}.
3215 Please do not use the term ``pathname'' that is used in Unix
3216 documentation; use ``file name'' (two words) instead. We use the term
3217 ``path'' only for search paths, which are lists of directory names.
3219 Please do not use the term ``illegal'' to refer to erroneous input to
3220 a computer program. Please use ``invalid'' for this, and reserve the
3221 term ``illegal'' for activities prohibited by law.
3223 @node Doc Strings and Manuals
3224 @section Doc Strings and Manuals
3226 Some programming systems, such as Emacs, provide a documentation string
3227 for each function, command or variable. You may be tempted to write a
3228 reference manual by compiling the documentation strings and writing a
3229 little additional text to go around them---but you must not do it. That
3230 approach is a fundamental mistake. The text of well-written
3231 documentation strings will be entirely wrong for a manual.
3233 A documentation string needs to stand alone---when it appears on the
3234 screen, there will be no other text to introduce or explain it.
3235 Meanwhile, it can be rather informal in style.
3237 The text describing a function or variable in a manual must not stand
3238 alone; it appears in the context of a section or subsection. Other text
3239 at the beginning of the section should explain some of the concepts, and
3240 should often make some general points that apply to several functions or
3241 variables. The previous descriptions of functions and variables in the
3242 section will also have given information about the topic. A description
3243 written to stand alone would repeat some of that information; this
3244 redundance looks bad. Meanwhile, the informality that is acceptable in
3245 a documentation string is totally unacceptable in a manual.
3247 The only good way to use documentation strings in writing a good manual
3248 is to use them as a source of information for writing good text.
3250 @node Manual Structure Details
3251 @section Manual Structure Details
3252 @cindex manual structure
3254 The title page of the manual should state the version of the programs or
3255 packages documented in the manual. The Top node of the manual should
3256 also contain this information. If the manual is changing more
3257 frequently than or independent of the program, also state a version
3258 number for the manual in both of these places.
3260 Each program documented in the manual should have a node named
3261 @samp{@var{program} Invocation} or @samp{Invoking @var{program}}. This
3262 node (together with its subnodes, if any) should describe the program's
3263 command line arguments and how to run it (the sort of information people
3264 would look for in a man page). Start with an @samp{@@example}
3265 containing a template for all the options and arguments that the program
3268 Alternatively, put a menu item in some menu whose item name fits one of
3269 the above patterns. This identifies the node which that item points to
3270 as the node for this purpose, regardless of the node's actual name.
3272 The @samp{--usage} feature of the Info reader looks for such a node
3273 or menu item in order to find the relevant text, so it is essential
3274 for every Texinfo file to have one.
3276 If one manual describes several programs, it should have such a node for
3277 each program described in the manual.
3279 @node License for Manuals
3280 @section License for Manuals
3281 @cindex license for manuals
3283 Please use the GNU Free Documentation License for all GNU manuals that
3284 are more than a few pages long. Likewise for a collection of short
3285 documents---you only need one copy of the GNU FDL for the whole
3286 collection. For a single short document, you can use a very permissive
3287 non-copyleft license, to avoid taking up space with a long license.
3289 See @uref{http://www.gnu.org/copyleft/fdl-howto.html} for more explanation
3290 of how to employ the GFDL.
3292 Note that it is not obligatory to include a copy of the GNU GPL or GNU
3293 LGPL in a manual whose license is neither the GPL nor the LGPL. It can
3294 be a good idea to include the program's license in a large manual; in a
3295 short manual, whose size would be increased considerably by including
3296 the program's license, it is probably better not to include it.
3298 @node Manual Credits
3299 @section Manual Credits
3300 @cindex credits for manuals
3302 Please credit the principal human writers of the manual as the authors,
3303 on the title page of the manual. If a company sponsored the work, thank
3304 the company in a suitable place in the manual, but do not cite the
3305 company as an author.
3307 @node Printed Manuals
3308 @section Printed Manuals
3310 The FSF publishes some GNU manuals in printed form. To encourage sales
3311 of these manuals, the on-line versions of the manual should mention at
3312 the very start that the printed manual is available and should point at
3313 information for getting it---for instance, with a link to the page
3314 @url{http://www.gnu.org/order/order.html}. This should not be included
3315 in the printed manual, though, because there it is redundant.
3317 It is also useful to explain in the on-line forms of the manual how the
3318 user can print out the manual from the sources.
3321 @section The NEWS File
3322 @cindex @file{NEWS} file
3324 In addition to its manual, the package should have a file named
3325 @file{NEWS} which contains a list of user-visible changes worth
3326 mentioning. In each new release, add items to the front of the file and
3327 identify the version they pertain to. Don't discard old items; leave
3328 them in the file after the newer items. This way, a user upgrading from
3329 any previous version can see what is new.
3331 If the @file{NEWS} file gets very long, move some of the older items
3332 into a file named @file{ONEWS} and put a note at the end referring the
3336 @section Change Logs
3339 Keep a change log to describe all the changes made to program source
3340 files. The purpose of this is so that people investigating bugs in the
3341 future will know about the changes that might have introduced the bug.
3342 Often a new bug can be found by looking at what was recently changed.
3343 More importantly, change logs can help you eliminate conceptual
3344 inconsistencies between different parts of a program, by giving you a
3345 history of how the conflicting concepts arose and who they came from.
3348 * Change Log Concepts::
3349 * Style of Change Logs::
3351 * Conditional Changes::
3352 * Indicating the Part Changed::
3355 @node Change Log Concepts
3356 @subsection Change Log Concepts
3358 You can think of the change log as a conceptual ``undo list'' which
3359 explains how earlier versions were different from the current version.
3360 People can see the current version; they don't need the change log
3361 to tell them what is in it. What they want from a change log is a
3362 clear explanation of how the earlier version differed.
3364 The change log file is normally called @file{ChangeLog} and covers an
3365 entire directory. Each directory can have its own change log, or a
3366 directory can use the change log of its parent directory--it's up to
3369 Another alternative is to record change log information with a version
3370 control system such as RCS or CVS. This can be converted automatically
3371 to a @file{ChangeLog} file using @code{rcs2log}; in Emacs, the command
3372 @kbd{C-x v a} (@code{vc-update-change-log}) does the job.
3374 There's no need to describe the full purpose of the changes or how they
3375 work together. If you think that a change calls for explanation, you're
3376 probably right. Please do explain it---but please put the explanation
3377 in comments in the code, where people will see it whenever they see the
3378 code. For example, ``New function'' is enough for the change log when
3379 you add a function, because there should be a comment before the
3380 function definition to explain what it does.
3382 In the past, we recommended not mentioning changes in non-software
3383 files (manuals, help files, etc.) in change logs. However, we've been
3384 advised that it is a good idea to include them, for the sake of
3387 However, sometimes it is useful to write one line to describe the
3388 overall purpose of a batch of changes.
3390 The easiest way to add an entry to @file{ChangeLog} is with the Emacs
3391 command @kbd{M-x add-change-log-entry}. An entry should have an
3392 asterisk, the name of the changed file, and then in parentheses the name
3393 of the changed functions, variables or whatever, followed by a colon.
3394 Then describe the changes you made to that function or variable.
3396 @node Style of Change Logs
3397 @subsection Style of Change Logs
3398 @cindex change logs, style
3400 Here are some simple examples of change log entries, starting with the
3401 header line that says who made the change and when it was installed,
3402 followed by descriptions of specific changes. (These examples are
3403 drawn from Emacs and GCC.)
3406 1998-08-17 Richard Stallman <rms@@gnu.org>
3408 * register.el (insert-register): Return nil.
3409 (jump-to-register): Likewise.
3411 * sort.el (sort-subr): Return nil.
3413 * tex-mode.el (tex-bibtex-file, tex-file, tex-region):
3414 Restart the tex shell if process is gone or stopped.
3415 (tex-shell-running): New function.
3417 * expr.c (store_one_arg): Round size up for move_block_to_reg.
3418 (expand_call): Round up when emitting USE insns.
3419 * stmt.c (assign_parms): Round size up for move_block_from_reg.
3422 It's important to name the changed function or variable in full. Don't
3423 abbreviate function or variable names, and don't combine them.
3424 Subsequent maintainers will often search for a function name to find all
3425 the change log entries that pertain to it; if you abbreviate the name,
3426 they won't find it when they search.
3428 For example, some people are tempted to abbreviate groups of function
3429 names by writing @samp{* register.el (@{insert,jump-to@}-register)};
3430 this is not a good idea, since searching for @code{jump-to-register} or
3431 @code{insert-register} would not find that entry.
3433 Separate unrelated change log entries with blank lines. When two
3434 entries represent parts of the same change, so that they work together,
3435 then don't put blank lines between them. Then you can omit the file
3436 name and the asterisk when successive entries are in the same file.
3438 Break long lists of function names by closing continued lines with
3439 @samp{)}, rather than @samp{,}, and opening the continuation with
3440 @samp{(} as in this example:
3443 * keyboard.c (menu_bar_items, tool_bar_items)
3444 (Fexecute_extended_command): Deal with `keymap' property.
3447 When you install someone else's changes, put the contributor's name in
3448 the change log entry rather than in the text of the entry. In other
3452 2002-07-14 John Doe <jdoe@@gnu.org>
3454 * sewing.c: Make it sew.
3461 2002-07-14 Usual Maintainer <usual@@gnu.org>
3463 * sewing.c: Make it sew. Patch by jdoe@@gnu.org.
3466 As for the date, that should be the date you applied the change.
3468 @node Simple Changes
3469 @subsection Simple Changes
3471 Certain simple kinds of changes don't need much detail in the change
3474 When you change the calling sequence of a function in a simple fashion,
3475 and you change all the callers of the function to use the new calling
3476 sequence, there is no need to make individual entries for all the
3477 callers that you changed. Just write in the entry for the function
3478 being called, ``All callers changed''---like this:
3481 * keyboard.c (Fcommand_execute): New arg SPECIAL.
3482 All callers changed.
3485 When you change just comments or doc strings, it is enough to write an
3486 entry for the file, without mentioning the functions. Just ``Doc
3487 fixes'' is enough for the change log.
3489 There's no technical need to make change log entries for documentation
3490 files. This is because documentation is not susceptible to bugs that
3491 are hard to fix. Documentation does not consist of parts that must
3492 interact in a precisely engineered fashion. To correct an error, you
3493 need not know the history of the erroneous passage; it is enough to
3494 compare what the documentation says with the way the program actually
3497 However, you should keep change logs for documentation files when the
3498 project gets copyright assignments from its contributors, so as to
3499 make the records of authorship more accurate.
3501 @node Conditional Changes
3502 @subsection Conditional Changes
3503 @cindex conditional changes, and change logs
3504 @cindex change logs, conditional changes
3506 C programs often contain compile-time @code{#if} conditionals. Many
3507 changes are conditional; sometimes you add a new definition which is
3508 entirely contained in a conditional. It is very useful to indicate in
3509 the change log the conditions for which the change applies.
3511 Our convention for indicating conditional changes is to use square
3512 brackets around the name of the condition.
3514 Here is a simple example, describing a change which is conditional but
3515 does not have a function or entity name associated with it:
3518 * xterm.c [SOLARIS2]: Include string.h.
3521 Here is an entry describing a new definition which is entirely
3522 conditional. This new definition for the macro @code{FRAME_WINDOW_P} is
3523 used only when @code{HAVE_X_WINDOWS} is defined:
3526 * frame.h [HAVE_X_WINDOWS] (FRAME_WINDOW_P): Macro defined.
3529 Here is an entry for a change within the function @code{init_display},
3530 whose definition as a whole is unconditional, but the changes themselves
3531 are contained in a @samp{#ifdef HAVE_LIBNCURSES} conditional:
3534 * dispnew.c (init_display) [HAVE_LIBNCURSES]: If X, call tgetent.
3537 Here is an entry for a change that takes affect only when
3538 a certain macro is @emph{not} defined:
3541 (gethostname) [!HAVE_SOCKETS]: Replace with winsock version.
3544 @node Indicating the Part Changed
3545 @subsection Indicating the Part Changed
3547 Indicate the part of a function which changed by using angle brackets
3548 enclosing an indication of what the changed part does. Here is an entry
3549 for a change in the part of the function @code{sh-while-getopts} that
3550 deals with @code{sh} commands:
3553 * progmodes/sh-script.el (sh-while-getopts) <sh>: Handle case that
3554 user-specified option string is empty.
3562 In the GNU project, man pages are secondary. It is not necessary or
3563 expected for every GNU program to have a man page, but some of them do.
3564 It's your choice whether to include a man page in your program.
3566 When you make this decision, consider that supporting a man page
3567 requires continual effort each time the program is changed. The time
3568 you spend on the man page is time taken away from more useful work.
3570 For a simple program which changes little, updating the man page may be
3571 a small job. Then there is little reason not to include a man page, if
3574 For a large program that changes a great deal, updating a man page may
3575 be a substantial burden. If a user offers to donate a man page, you may
3576 find this gift costly to accept. It may be better to refuse the man
3577 page unless the same person agrees to take full responsibility for
3578 maintaining it---so that you can wash your hands of it entirely. If
3579 this volunteer later ceases to do the job, then don't feel obliged to
3580 pick it up yourself; it may be better to withdraw the man page from the
3581 distribution until someone else agrees to update it.
3583 When a program changes only a little, you may feel that the
3584 discrepancies are small enough that the man page remains useful without
3585 updating. If so, put a prominent note near the beginning of the man
3586 page explaining that you don't maintain it and that the Texinfo manual
3587 is more authoritative. The note should say how to access the Texinfo
3590 Be sure that man pages include a copyright statement and free
3591 license. The simple all-permissive license is appropriate for simple
3595 Copying and distribution of this file, with or without modification,
3596 are permitted in any medium without royalty provided the copyright
3597 notice and this notice are preserved.
3600 For long man pages, with enough explanation and documentation that
3601 they can be considered true manuals, use the GFDL (@pxref{License for
3604 Finally, the GNU help2man program
3605 (@uref{http://www.gnu.org/software/help2man/}) is one way to automate
3606 generation of a man page, in this case from @option{--help} output.
3607 This is sufficient in many cases.
3609 @node Reading other Manuals
3610 @section Reading other Manuals
3612 There may be non-free books or documentation files that describe the
3613 program you are documenting.
3615 It is ok to use these documents for reference, just as the author of a
3616 new algebra textbook can read other books on algebra. A large portion
3617 of any non-fiction book consists of facts, in this case facts about how
3618 a certain program works, and these facts are necessarily the same for
3619 everyone who writes about the subject. But be careful not to copy your
3620 outline structure, wording, tables or examples from preexisting non-free
3621 documentation. Copying from free documentation may be ok; please check
3622 with the FSF about the individual case.
3624 @node Managing Releases
3625 @chapter The Release Process
3628 Making a release is more than just bundling up your source files in a
3629 tar file and putting it up for FTP. You should set up your software so
3630 that it can be configured to run on a variety of systems. Your Makefile
3631 should conform to the GNU standards described below, and your directory
3632 layout should also conform to the standards discussed below. Doing so
3633 makes it easy to include your package into the larger framework of
3637 * Configuration:: How Configuration Should Work
3638 * Makefile Conventions:: Makefile Conventions
3639 * Releases:: Making Releases
3643 @section How Configuration Should Work
3644 @cindex program configuration
3647 Each GNU distribution should come with a shell script named
3648 @code{configure}. This script is given arguments which describe the
3649 kind of machine and system you want to compile the program for.
3651 The @code{configure} script must record the configuration options so
3652 that they affect compilation.
3654 One way to do this is to make a link from a standard name such as
3655 @file{config.h} to the proper configuration file for the chosen system.
3656 If you use this technique, the distribution should @emph{not} contain a
3657 file named @file{config.h}. This is so that people won't be able to
3658 build the program without configuring it first.
3660 Another thing that @code{configure} can do is to edit the Makefile. If
3661 you do this, the distribution should @emph{not} contain a file named
3662 @file{Makefile}. Instead, it should include a file @file{Makefile.in} which
3663 contains the input used for editing. Once again, this is so that people
3664 won't be able to build the program without configuring it first.
3666 If @code{configure} does write the @file{Makefile}, then @file{Makefile}
3667 should have a target named @file{Makefile} which causes @code{configure}
3668 to be rerun, setting up the same configuration that was set up last
3669 time. The files that @code{configure} reads should be listed as
3670 dependencies of @file{Makefile}.
3672 All the files which are output from the @code{configure} script should
3673 have comments at the beginning explaining that they were generated
3674 automatically using @code{configure}. This is so that users won't think
3675 of trying to edit them by hand.
3677 The @code{configure} script should write a file named @file{config.status}
3678 which describes which configuration options were specified when the
3679 program was last configured. This file should be a shell script which,
3680 if run, will recreate the same configuration.
3682 The @code{configure} script should accept an option of the form
3683 @samp{--srcdir=@var{dirname}} to specify the directory where sources are found
3684 (if it is not the current directory). This makes it possible to build
3685 the program in a separate directory, so that the actual source directory
3688 If the user does not specify @samp{--srcdir}, then @code{configure} should
3689 check both @file{.} and @file{..} to see if it can find the sources. If
3690 it finds the sources in one of these places, it should use them from
3691 there. Otherwise, it should report that it cannot find the sources, and
3692 should exit with nonzero status.
3694 Usually the easy way to support @samp{--srcdir} is by editing a
3695 definition of @code{VPATH} into the Makefile. Some rules may need to
3696 refer explicitly to the specified source directory. To make this
3697 possible, @code{configure} can add to the Makefile a variable named
3698 @code{srcdir} whose value is precisely the specified directory.
3700 The @code{configure} script should also take an argument which specifies the
3701 type of system to build the program for. This argument should look like
3705 @var{cpu}-@var{company}-@var{system}
3708 For example, an Athlon-based GNU/Linux system might be
3709 @samp{i686-pc-linux-gnu}.
3711 The @code{configure} script needs to be able to decode all plausible
3712 alternatives for how to describe a machine. Thus,
3713 @samp{athlon-pc-gnu/linux} would be a valid alias. There is a shell
3715 @uref{http://savannah.gnu.org/@/cgi-bin/@/viewcvs/@/*checkout*/@/config/@/config/@/config.sub,
3716 @file{config.sub}} that you can use as a subroutine to validate system
3717 types and canonicalize aliases.
3719 The @code{configure} script should also take the option
3720 @option{--build=@var{buildtype}}, which should be equivalent to a
3721 plain @var{buildtype} argument. For example, @samp{configure
3722 --build=i686-pc-linux-gnu} is equivalent to @samp{configure
3723 i686-pc-linux-gnu}. When the build type is not specified by an option
3724 or argument, the @code{configure} script should normally guess it using
3726 @uref{http://savannah.gnu.org/@/cgi-bin/@/viewcvs/@/*checkout*/@/config/@/config/@/config.guess,
3727 @file{config.guess}}.
3729 @cindex optional features, configure-time
3730 Other options are permitted to specify in more detail the software
3731 or hardware present on the machine, and include or exclude optional
3732 parts of the package:
3735 @item --enable-@var{feature}@r{[}=@var{parameter}@r{]}
3736 Configure the package to build and install an optional user-level
3737 facility called @var{feature}. This allows users to choose which
3738 optional features to include. Giving an optional @var{parameter} of
3739 @samp{no} should omit @var{feature}, if it is built by default.
3741 No @samp{--enable} option should @strong{ever} cause one feature to
3742 replace another. No @samp{--enable} option should ever substitute one
3743 useful behavior for another useful behavior. The only proper use for
3744 @samp{--enable} is for questions of whether to build part of the program
3747 @item --with-@var{package}
3748 @c @r{[}=@var{parameter}@r{]}
3749 The package @var{package} will be installed, so configure this package
3750 to work with @var{package}.
3752 @c Giving an optional @var{parameter} of
3753 @c @samp{no} should omit @var{package}, if it is used by default.
3755 Possible values of @var{package} include
3756 @samp{gnu-as} (or @samp{gas}), @samp{gnu-ld}, @samp{gnu-libc},
3762 Do not use a @samp{--with} option to specify the file name to use to
3763 find certain files. That is outside the scope of what @samp{--with}
3767 All @code{configure} scripts should accept all of these ``detail''
3768 options, whether or not they make any difference to the particular
3769 package at hand. In particular, they should accept any option that
3770 starts with @samp{--with-} or @samp{--enable-}. This is so users will
3771 be able to configure an entire GNU source tree at once with a single set
3774 You will note that the categories @samp{--with-} and @samp{--enable-}
3775 are narrow: they @strong{do not} provide a place for any sort of option
3776 you might think of. That is deliberate. We want to limit the possible
3777 configuration options in GNU software. We do not want GNU programs to
3778 have idiosyncratic configuration options.
3780 Packages that perform part of the compilation process may support
3781 cross-compilation. In such a case, the host and target machines for the
3782 program may be different.
3784 The @code{configure} script should normally treat the specified type of
3785 system as both the host and the target, thus producing a program which
3786 works for the same type of machine that it runs on.
3788 To compile a program to run on a host type that differs from the build
3789 type, use the configure option @option{--host=@var{hosttype}}, where
3790 @var{hosttype} uses the same syntax as @var{buildtype}. The host type
3791 normally defaults to the build type.
3793 To configure a cross-compiler, cross-assembler, or what have you, you
3794 should specify a target different from the host, using the configure
3795 option @samp{--target=@var{targettype}}. The syntax for
3796 @var{targettype} is the same as for the host type. So the command would
3800 ./configure --host=@var{hosttype} --target=@var{targettype}
3803 The target type normally defaults to the host type.
3804 Programs for which cross-operation is not meaningful need not accept the
3805 @samp{--target} option, because configuring an entire operating system for
3806 cross-operation is not a meaningful operation.
3808 Some programs have ways of configuring themselves automatically. If
3809 your program is set up to do this, your @code{configure} script can simply
3810 ignore most of its arguments.
3812 @comment The makefile standards are in a separate file that is also
3813 @comment included by make.texinfo. Done by roland@gnu.ai.mit.edu on 1/6/93.
3814 @comment For this document, turn chapters into sections, etc.
3816 @include make-stds.texi
3820 @section Making Releases
3823 You should identify each release with a pair of version numbers, a
3824 major version and a minor. We have no objection to using more than
3825 two numbers, but it is very unlikely that you really need them.
3827 Package the distribution of @code{Foo version 69.96} up in a gzipped tar
3828 file with the name @file{foo-69.96.tar.gz}. It should unpack into a
3829 subdirectory named @file{foo-69.96}.
3831 Building and installing the program should never modify any of the files
3832 contained in the distribution. This means that all the files that form
3833 part of the program in any way must be classified into @dfn{source
3834 files} and @dfn{non-source files}. Source files are written by humans
3835 and never changed automatically; non-source files are produced from
3836 source files by programs under the control of the Makefile.
3838 @cindex @file{README} file
3839 The distribution should contain a file named @file{README} which gives
3840 the name of the package, and a general description of what it does. It
3841 is also good to explain the purpose of each of the first-level
3842 subdirectories in the package, if there are any. The @file{README} file
3843 should either state the version number of the package, or refer to where
3844 in the package it can be found.
3846 The @file{README} file should refer to the file @file{INSTALL}, which
3847 should contain an explanation of the installation procedure.
3849 The @file{README} file should also refer to the file which contains the
3850 copying conditions. The GNU GPL, if used, should be in a file called
3851 @file{COPYING}. If the GNU LGPL is used, it should be in a file called
3854 Naturally, all the source files must be in the distribution. It is okay
3855 to include non-source files in the distribution, provided they are
3856 up-to-date and machine-independent, so that building the distribution
3857 normally will never modify them. We commonly include non-source files
3858 produced by Bison, @code{lex}, @TeX{}, and @code{makeinfo}; this helps avoid
3859 unnecessary dependencies between our distributions, so that users can
3860 install whichever packages they want to install.
3862 Non-source files that might actually be modified by building and
3863 installing the program should @strong{never} be included in the
3864 distribution. So if you do distribute non-source files, always make
3865 sure they are up to date when you make a new distribution.
3867 Make sure that the directory into which the distribution unpacks (as
3868 well as any subdirectories) are all world-writable (octal mode 777).
3869 This is so that old versions of @code{tar} which preserve the
3870 ownership and permissions of the files from the tar archive will be
3871 able to extract all the files even if the user is unprivileged.
3873 Make sure that all the files in the distribution are world-readable.
3875 Don't include any symbolic links in the distribution itself. If the tar
3876 file contains symbolic links, then people cannot even unpack it on
3877 systems that don't support symbolic links. Also, don't use multiple
3878 names for one file in different directories, because certain file
3879 systems cannot handle this and that prevents unpacking the
3882 Try to make sure that all the file names will be unique on MS-DOS. A
3883 name on MS-DOS consists of up to 8 characters, optionally followed by a
3884 period and up to three characters. MS-DOS will truncate extra
3885 characters both before and after the period. Thus,
3886 @file{foobarhacker.c} and @file{foobarhacker.o} are not ambiguous; they
3887 are truncated to @file{foobarha.c} and @file{foobarha.o}, which are
3890 @cindex @file{texinfo.tex}, in a distribution
3891 Include in your distribution a copy of the @file{texinfo.tex} you used
3892 to test print any @file{*.texinfo} or @file{*.texi} files.
3894 Likewise, if your program uses small GNU software packages like regex,
3895 getopt, obstack, or termcap, include them in the distribution file.
3896 Leaving them out would make the distribution file a little smaller at
3897 the expense of possible inconvenience to a user who doesn't know what
3901 @chapter References to Non-Free Software and Documentation
3902 @cindex references to non-free material
3904 A GNU program should not recommend use of any non-free program. We
3905 can't stop some people from writing proprietary programs, or stop
3906 other people from using them, but we can and should refuse to
3907 advertise them to new potential customers. Proprietary software is a
3908 social and ethical problem, and the point of GNU is to solve that
3911 The GNU definition of free software is found on the GNU web site at
3912 @url{http://www.gnu.org/philosophy/free-sw.html}, and the definition
3913 of free documentation is found at
3914 @url{http://www.gnu.org/philosophy/free-doc.html}. A list of
3915 important licenses and whether they qualify as free is in
3916 @url{http://www.gnu.org/@/licenses/@/license-list.html}. The terms
3917 ``free'' and ``non-free'', used in this document, refer to that
3918 definition. If it is not clear whether a license qualifies as free
3919 under this definition, please ask the GNU Project by writing to
3920 @email{licensing@@gnu.org}. We will answer, and if the license is an
3921 important one, we will add it to the list.
3923 When a non-free program or system is well known, you can mention it in
3924 passing---that is harmless, since users who might want to use it
3925 probably already know about it. For instance, it is fine to explain
3926 how to build your package on top of some widely used non-free
3927 operating system, or how to use it together with some widely used
3930 However, you should give only the necessary information to help those
3931 who already use the non-free program to use your program with
3932 it---don't give, or refer to, any further information about the
3933 proprietary program, and don't imply that the proprietary program
3934 enhances your program, or that its existence is in any way a good
3935 thing. The goal should be that people already using the proprietary
3936 program will get the advice they need about how to use your free
3937 program with it, while people who don't already use the proprietary
3938 program will not see anything to lead them to take an interest in it.
3940 If a non-free program or system is obscure in your program's domain,
3941 your program should not mention or support it at all, since doing so
3942 would tend to popularize the non-free program more than it popularizes
3943 your program. (You cannot hope to find many additional users among
3944 the users of Foobar if the users of Foobar are few.)
3946 Sometimes a program is free software in itself but depends on a
3947 non-free platform in order to run. For instance, many Java programs
3948 depend on Sun's Java implementation, and won't run on the GNU Java
3949 Compiler (which does not yet have all the features) or won't run with
3950 the GNU Java libraries. To recommend that program is inherently to
3951 recommend the non-free platform as well; if you should not do the
3952 latter, then don't do the former.
3954 A GNU package should not refer the user to any non-free documentation
3955 for free software. Free documentation that can be included in free
3956 operating systems is essential for completing the GNU system, or any
3957 free operating system, so it is a major focus of the GNU Project; to
3958 recommend use of documentation that we are not allowed to use in GNU
3959 would weaken the impetus for the community to produce documentation
3960 that we can include. So GNU packages should never recommend non-free
3963 By contrast, it is ok to refer to journal articles and textbooks in
3964 the comments of a program for explanation of how it functions, even
3965 though they be non-free. This is because we don't include such things
3966 in the GNU system even if we are allowed to--they are outside the
3967 scope of an operating system project.
3969 Referring to a web site that describes or recommends a non-free
3970 program is in effect promoting that software, so please do not make
3971 links (or mention by name) web sites that contain such material. This
3972 policy is relevant particularly for the web pages for a GNU package.
3974 Following links from nearly any web site can lead to non-free
3975 software; this is an inescapable aspect of the nature of the web, and
3976 in itself is no objection to linking to a site. As long as the site
3977 does not itself recommend a non-free program, there is no need be
3978 concerned about the sites it links to for other reasons.
3980 Thus, for example, you should not make a link to AT&T's web site,
3981 because that recommends AT&T's non-free software packages; you should
3982 not make a link to a site that links to AT&T's site saying it is a
3983 place to get a non-free program; but if a site you want to link to
3984 refers to AT&T's web site in some other context (such as long-distance
3985 telephone service), that is not a problem.
3987 @node Copying This Manual
3988 @appendix Copying This Manual
3991 * GNU Free Documentation License:: License for copying this manual
4003 eval: (add-hook 'write-file-hooks 'time-stamp)
4004 time-stamp-start: "@set lastupdate "
4006 time-stamp-format: "%:b %:d, %:y"
4007 compile-command: "make just-standards"