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 February 20, 2004
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 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 World Wide Web server host in several
98 different formats: @uref{http://www.gnu.org/prep/standards.text},
99 @uref{http://www.gnu.org/prep/standards.info}, and
100 @uref{http://www.gnu.org/prep/standards.dvi}, as well as the
101 Texinfo ``source'' which is divided in two files:
102 @uref{http://www.gnu.org/prep/standards.texi} and
103 @uref{http://www.gnu.org/prep/make-stds.texi}. The GNU Coding
104 Standards are also available in HTML format starting at
105 @uref{http://www.gnu.org/prep/standards_toc.html}.
107 Corrections or suggestions for this document should be sent to
108 @email{bug-standards@@gnu.org}. If you make a suggestion, please include a
109 suggested new wording for it; our time is limited. We prefer a context
110 diff to the @file{standards.texi} or @file{make-stds.texi} files, but if
111 you don't have those files, please mail your suggestion anyway.
113 These standards cover the minimum of what is important when writing a
114 GNU package. Likely, the needs for additional standards will come up.
115 Sometimes, you might suggest that such standards be added to this
116 document. If you think your standards would be generally useful, please
119 You should also set standards for your package on many questions not
120 addressed or not firmly specified here. The most important point is to
121 be self-consistent---try to stick to the conventions you pick, and try
122 to document them as much as possible. That way, your program will be
123 more maintainable by others.
125 The GNU Hello program serves as an example of how to follow the GNU
126 coding standards for a trivial program which prints @samp{Hello,
127 world!}. @uref{http://www.gnu.org/software/hello/hello.html}.
130 @chapter Keeping Free Software Free
131 @cindex legal aspects
133 This @value{CHAPTER} discusses how you can make sure that GNU software
134 avoids legal difficulties, and other related issues.
137 * Reading Non-Free Code:: Referring to Proprietary Programs
138 * Contributions:: Accepting Contributions
139 * Trademarks:: How We Deal with Trademark Issues
142 @node Reading Non-Free Code
143 @section Referring to Proprietary Programs
144 @cindex proprietary programs
145 @cindex avoiding proprietary code
147 Don't in any circumstances refer to Unix source code for or during
148 your work on GNU! (Or to any other proprietary programs.)
150 If you have a vague recollection of the internals of a Unix program,
151 this does not absolutely mean you can't write an imitation of it, but
152 do try to organize the imitation internally along different lines,
153 because this is likely to make the details of the Unix version
154 irrelevant and dissimilar to your results.
156 For example, Unix utilities were generally optimized to minimize
157 memory use; if you go for speed instead, your program will be very
158 different. You could keep the entire input file in core and scan it
159 there instead of using stdio. Use a smarter algorithm discovered more
160 recently than the Unix program. Eliminate use of temporary files. Do
161 it in one pass instead of two (we did this in the assembler).
163 Or, on the contrary, emphasize simplicity instead of speed. For some
164 applications, the speed of today's computers makes simpler algorithms
167 Or go for generality. For example, Unix programs often have static
168 tables or fixed-size strings, which make for arbitrary limits; use
169 dynamic allocation instead. Make sure your program handles NULs and
170 other funny characters in the input files. Add a programming language
171 for extensibility and write part of the program in that language.
173 Or turn some parts of the program into independently usable libraries.
174 Or use a simple garbage collector instead of tracking precisely when
175 to free memory, or use a new GNU facility such as obstacks.
178 @section Accepting Contributions
180 @cindex accepting contributions
182 If the program you are working on is copyrighted by the Free Software
183 Foundation, then when someone else sends you a piece of code to add to
184 the program, we need legal papers to use it---just as we asked you to
185 sign papers initially. @emph{Each} person who makes a nontrivial
186 contribution to a program must sign some sort of legal papers in order
187 for us to have clear title to the program; the main author alone is not
190 So, before adding in any contributions from other people, please tell
191 us, so we can arrange to get the papers. Then wait until we tell you
192 that we have received the signed papers, before you actually use the
195 This applies both before you release the program and afterward. If
196 you receive diffs to fix a bug, and they make significant changes, we
197 need legal papers for that change.
199 This also applies to comments and documentation files. For copyright
200 law, comments and code are just text. Copyright applies to all kinds of
201 text, so we need legal papers for all kinds.
203 We know it is frustrating to ask for legal papers; it's frustrating for
204 us as well. But if you don't wait, you are going out on a limb---for
205 example, what if the contributor's employer won't sign a disclaimer?
206 You might have to take that code out again!
208 You don't need papers for changes of a few lines here or there, since
209 they are not significant for copyright purposes. Also, you don't need
210 papers if all you get from the suggestion is some ideas, not actual code
211 which you use. For example, if someone send you one implementation, but
212 you write a different implementation of the same idea, you don't need to
215 The very worst thing is if you forget to tell us about the other
216 contributor. We could be very embarrassed in court some day as a
219 We have more detailed advice for maintainers of programs; if you have
220 reached the stage of actually maintaining a program for GNU (whether
221 released or not), please ask us for a copy.
227 Please do not include any trademark acknowledgements in GNU software
228 packages or documentation.
230 Trademark acknowledgements are the statements that such-and-such is a
231 trademark of so-and-so. The GNU Project has no objection to the basic
232 idea of trademarks, but these acknowledgements feel like kowtowing,
233 and there is no legal requirement for them, so we don't use them.
235 What is legally required, as regards other people's trademarks, is to
236 avoid using them in ways which a reader might reasonably understand as
237 naming or labeling our own programs or activities. For example, since
238 ``Objective C'' is (or at least was) a trademark, we made sure to say
239 that we provide a ``compiler for the Objective C language'' rather
240 than an ``Objective C compiler''. The latter would have been meant as
241 a shorter way of saying the former, but it does not explicitly state
242 the relationship, so it could be misinterpreted as using ``Objective
243 C'' as a label for the compiler rather than for the language.
245 Please don't use ``win'' as an abbreviation for Microsoft Windows in
246 GNU software or documentation. In hacker terminology, calling
247 something a "win" is a form of praise. If you wish to praise
248 Microsoft Windows when speaking on your own, by all means do so, but
249 not in GNU software. Usually we write the word ``windows'' in full,
250 but when brevity is very important (as in file names and sometimes
251 symbol names), we abbreviate it to ``w''. For instance, the files and
252 functions in Emacs that deal with Windows start with @samp{w32}.
255 @chapter General Program Design
256 @cindex program design
258 This @value{CHAPTER} discusses some of the issues you should take into
259 account when designing your program.
261 @c Standard or ANSI C
263 @c In 1989 the American National Standards Institute (ANSI) standardized
264 @c C as standard X3.159-1989. In December of that year the
265 @c International Standards Organization ISO adopted the ANSI C standard
266 @c making minor changes. In 1990 ANSI then re-adopted ISO standard
267 @c C. This version of C is known as either ANSI C or Standard C.
269 @c A major revision of the C Standard appeared in 1999.
272 * Source Language:: Which languges to use.
273 * Compatibility:: Compatibility with other implementations
274 * Using Extensions:: Using non-standard features
275 * Standard C:: Using Standard C features
276 * Conditional Compilation:: Compiling Code Only If A Conditional is True
279 @node Source Language
280 @section Which Languages to Use
281 @cindex programming languges
283 When you want to use a language that gets compiled and runs at high
284 speed, the best language to use is C. Using another language is like
285 using a non-standard feature: it will cause trouble for users. Even if
286 GCC supports the other language, users may find it inconvenient to have
287 to install the compiler for that other language in order to build your
288 program. For example, if you write your program in C++, people will
289 have to install the GNU C++ compiler in order to compile your program.
291 C has one other advantage over C++ and other compiled languages: more
292 people know C, so more people will find it easy to read and modify the
293 program if it is written in C.
295 So in general it is much better to use C, rather than the
296 comparable alternatives.
298 But there are two exceptions to that conclusion:
302 It is no problem to use another language to write a tool specifically
303 intended for use with that language. That is because the only people
304 who want to build the tool will be those who have installed the other
308 If an application is of interest only to a narrow part of the community,
309 then the question of which language it is written in has less effect on
310 other people, so you may as well please yourself.
313 Many programs are designed to be extensible: they include an interpreter
314 for a language that is higher level than C. Often much of the program
315 is written in that language, too. The Emacs editor pioneered this
319 The standard extensibility interpreter for GNU software is GUILE, which
320 implements the language Scheme (an especially clean and simple dialect
321 of Lisp). @uref{http://www.gnu.org/software/guile/}. We don't reject
322 programs written in other ``scripting languages'' such as Perl and
323 Python, but using GUILE is very important for the overall consistency of
327 @section Compatibility with Other Implementations
328 @cindex compatibility with C and @sc{posix} standards
329 @cindex @sc{posix} compatibility
331 With occasional exceptions, utility programs and libraries for GNU
332 should be upward compatible with those in Berkeley Unix, and upward
333 compatible with Standard C if Standard C specifies their
334 behavior, and upward compatible with @sc{posix} if @sc{posix} specifies
337 When these standards conflict, it is useful to offer compatibility
338 modes for each of them.
340 @cindex options for compatibility
341 Standard C and @sc{posix} prohibit many kinds of extensions. Feel
342 free to make the extensions anyway, and include a @samp{--ansi},
343 @samp{--posix}, or @samp{--compatible} option to turn them off.
344 However, if the extension has a significant chance of breaking any real
345 programs or scripts, then it is not really upward compatible. So you
346 should try to redesign its interface to make it upward compatible.
348 @cindex @code{POSIXLY_CORRECT}, environment variable
349 Many GNU programs suppress extensions that conflict with @sc{posix} if the
350 environment variable @code{POSIXLY_CORRECT} is defined (even if it is
351 defined with a null value). Please make your program recognize this
352 variable if appropriate.
354 When a feature is used only by users (not by programs or command
355 files), and it is done poorly in Unix, feel free to replace it
356 completely with something totally different and better. (For example,
357 @code{vi} is replaced with Emacs.) But it is nice to offer a compatible
358 feature as well. (There is a free @code{vi} clone, so we offer it.)
360 Additional useful features are welcome regardless of whether
361 there is any precedent for them.
363 @node Using Extensions
364 @section Using Non-standard Features
365 @cindex non-standard extensions
367 Many GNU facilities that already exist support a number of convenient
368 extensions over the comparable Unix facilities. Whether to use these
369 extensions in implementing your program is a difficult question.
371 On the one hand, using the extensions can make a cleaner program.
372 On the other hand, people will not be able to build the program
373 unless the other GNU tools are available. This might cause the
374 program to work on fewer kinds of machines.
376 With some extensions, it might be easy to provide both alternatives.
377 For example, you can define functions with a ``keyword'' @code{INLINE}
378 and define that as a macro to expand into either @code{inline} or
379 nothing, depending on the compiler.
381 In general, perhaps it is best not to use the extensions if you can
382 straightforwardly do without them, but to use the extensions if they
383 are a big improvement.
385 An exception to this rule are the large, established programs (such as
386 Emacs) which run on a great variety of systems. Using GNU extensions in
387 such programs would make many users unhappy, so we don't do that.
389 Another exception is for programs that are used as part of compilation:
390 anything that must be compiled with other compilers in order to
391 bootstrap the GNU compilation facilities. If these require the GNU
392 compiler, then no one can compile them without having them installed
393 already. That would be extremely troublesome in certain cases.
396 @section Standard C and Pre-Standard C
397 @cindex @sc{ansi} C standard
399 1989 Standard C is widespread enough now that it is ok to use its
400 features in new programs. There is one exception: do not ever use the
401 ``trigraph'' feature of Standard C.
403 1999 Standard C is not widespread yet, so please do not require its
404 features in programs. It is ok to use its features if they are present.
406 However, it is easy to support pre-standard compilers in most programs,
407 so if you know how to do that, feel free. If a program you are
408 maintaining has such support, you should try to keep it working.
410 @cindex function prototypes
411 To support pre-standard C, instead of writing function definitions in
412 standard prototype form,
421 write the definition in pre-standard style like this,
431 and use a separate declaration to specify the argument prototype:
437 You need such a declaration anyway, in a header file, to get the benefit
438 of prototypes in all the files where the function is called. And once
439 you have the declaration, you normally lose nothing by writing the
440 function definition in the pre-standard style.
442 This technique does not work for integer types narrower than @code{int}.
443 If you think of an argument as being of a type narrower than @code{int},
444 declare it as @code{int} instead.
446 There are a few special cases where this technique is hard to use. For
447 example, if a function argument needs to hold the system type
448 @code{dev_t}, you run into trouble, because @code{dev_t} is shorter than
449 @code{int} on some machines; but you cannot use @code{int} instead,
450 because @code{dev_t} is wider than @code{int} on some machines. There
451 is no type you can safely use on all machines in a non-standard
452 definition. The only way to support non-standard C and pass such an
453 argument is to check the width of @code{dev_t} using Autoconf and choose
454 the argument type accordingly. This may not be worth the trouble.
456 In order to support pre-standard compilers that do not recognize
457 prototypes, you may want to use a preprocessor macro like this:
460 /* Declare the prototype for a general external function. */
461 #if defined (__STDC__) || defined (WINDOWSNT)
462 #define P_(proto) proto
468 @node Conditional Compilation
469 @section Conditional Compilation
471 When supporting configuration options already known when building your
472 program we prefer using @code{if (... )} over conditional compilation,
473 as in the former case the compiler is able to perform more extensive
474 checking of all possible code paths.
476 For example, please write
496 A modern compiler such as GCC will generate exactly the same code in
497 both cases, and we have been using similar techniques with good success
498 in several projects. Of course, the former method assumes that
499 @code{HAS_FOO} is defined as either 0 or 1.
501 While this is not a silver bullet solving all portability problems,
502 and is not always appropriate, following this policy would have saved
503 GCC developers many hours, or even days, per year.
505 In the case of function-like macros like @code{REVERSIBLE_CC_MODE} in
506 GCC which cannot be simply used in @code{if( ...)} statements, there is
507 an easy workaround. Simply introduce another macro
508 @code{HAS_REVERSIBLE_CC_MODE} as in the following example:
511 #ifdef REVERSIBLE_CC_MODE
512 #define HAS_REVERSIBLE_CC_MODE 1
514 #define HAS_REVERSIBLE_CC_MODE 0
518 @node Program Behavior
519 @chapter Program Behavior for All Programs
521 This @value{CHAPTER} describes conventions for writing robust
522 software. It also describes general standards for error messages, the
523 command line interface, and how libraries should behave.
526 * Semantics:: Writing robust programs
527 * Libraries:: Library behavior
528 * Errors:: Formatting error messages
529 * User Interfaces:: Standards about interfaces generally
530 * Graphical Interfaces:: Standards for graphical interfaces
531 * Command-Line Interfaces:: Standards for command line interfaces
532 * Option Table:: Table of long options
533 * Memory Usage:: When and how to care about memory needs
534 * File Usage:: Which files to use, and where
538 @section Writing Robust Programs
540 @cindex arbitrary limits on data
541 Avoid arbitrary limits on the length or number of @emph{any} data
542 structure, including file names, lines, files, and symbols, by allocating
543 all data structures dynamically. In most Unix utilities, ``long lines
544 are silently truncated''. This is not acceptable in a GNU utility.
546 @cindex @code{NUL} characters
547 Utilities reading files should not drop NUL characters, or any other
548 nonprinting characters @emph{including those with codes above 0177}.
549 The only sensible exceptions would be utilities specifically intended
550 for interface to certain types of terminals or printers
551 that can't handle those characters.
552 Whenever possible, try to make programs work properly with
553 sequences of bytes that represent multibyte characters, using encodings
554 such as UTF-8 and others.
556 @cindex error messages
557 Check every system call for an error return, unless you know you wish to
558 ignore errors. Include the system error text (from @code{perror} or
559 equivalent) in @emph{every} error message resulting from a failing
560 system call, as well as the name of the file if any and the name of the
561 utility. Just ``cannot open foo.c'' or ``stat failed'' is not
564 @cindex @code{malloc} return value
565 @cindex memory allocation failure
566 Check every call to @code{malloc} or @code{realloc} to see if it
567 returned zero. Check @code{realloc} even if you are making the block
568 smaller; in a system that rounds block sizes to a power of 2,
569 @code{realloc} may get a different block if you ask for less space.
571 In Unix, @code{realloc} can destroy the storage block if it returns
572 zero. GNU @code{realloc} does not have this bug: if it fails, the
573 original block is unchanged. Feel free to assume the bug is fixed. If
574 you wish to run your program on Unix, and wish to avoid lossage in this
575 case, you can use the GNU @code{malloc}.
577 You must expect @code{free} to alter the contents of the block that was
578 freed. Anything you want to fetch from the block, you must fetch before
581 If @code{malloc} fails in a noninteractive program, make that a fatal
582 error. In an interactive program (one that reads commands from the
583 user), it is better to abort the command and return to the command
584 reader loop. This allows the user to kill other processes to free up
585 virtual memory, and then try the command again.
587 @cindex command-line arguments, decoding
588 Use @code{getopt_long} to decode arguments, unless the argument syntax
589 makes this unreasonable.
591 When static storage is to be written in during program execution, use
592 explicit C code to initialize it. Reserve C initialized declarations
593 for data that will not be changed.
596 Try to avoid low-level interfaces to obscure Unix data structures (such
597 as file directories, utmp, or the layout of kernel memory), since these
598 are less likely to work compatibly. If you need to find all the files
599 in a directory, use @code{readdir} or some other high-level interface.
600 These are supported compatibly by GNU.
602 @cindex signal handling
603 The preferred signal handling facilities are the BSD variant of
604 @code{signal}, and the @sc{posix} @code{sigaction} function; the
605 alternative USG @code{signal} interface is an inferior design.
607 Nowadays, using the @sc{posix} signal functions may be the easiest way
608 to make a program portable. If you use @code{signal}, then on GNU/Linux
609 systems running GNU libc version 1, you should include
610 @file{bsd/signal.h} instead of @file{signal.h}, so as to get BSD
611 behavior. It is up to you whether to support systems where
612 @code{signal} has only the USG behavior, or give up on them.
614 @cindex impossible conditions
615 In error checks that detect ``impossible'' conditions, just abort.
616 There is usually no point in printing any message. These checks
617 indicate the existence of bugs. Whoever wants to fix the bugs will have
618 to read the source code and run a debugger. So explain the problem with
619 comments in the source. The relevant data will be in variables, which
620 are easy to examine with the debugger, so there is no point moving them
623 Do not use a count of errors as the exit status for a program.
624 @emph{That does not work}, because exit status values are limited to 8
625 bits (0 through 255). A single run of the program might have 256
626 errors; if you try to return 256 as the exit status, the parent process
627 will see 0 as the status, and it will appear that the program succeeded.
629 @cindex temporary files
630 @cindex @code{TMPDIR} environment variable
631 If you make temporary files, check the @code{TMPDIR} environment
632 variable; if that variable is defined, use the specified directory
633 instead of @file{/tmp}.
635 In addition, be aware that there is a possible security problem when
636 creating temporary files in world-writable directories. In C, you can
637 avoid this problem by creating temporary files in this manner:
640 fd = open(filename, O_WRONLY | O_CREAT | O_EXCL, 0600);
644 or by using the @code{mkstemps} function from libiberty.
646 In bash, use @code{set -C} to avoid this problem.
649 @section Library Behavior
652 Try to make library functions reentrant. If they need to do dynamic
653 storage allocation, at least try to avoid any nonreentrancy aside from
654 that of @code{malloc} itself.
656 Here are certain name conventions for libraries, to avoid name
659 Choose a name prefix for the library, more than two characters long.
660 All external function and variable names should start with this
661 prefix. In addition, there should only be one of these in any given
662 library member. This usually means putting each one in a separate
665 An exception can be made when two external symbols are always used
666 together, so that no reasonable program could use one without the
667 other; then they can both go in the same file.
669 External symbols that are not documented entry points for the user
670 should have names beginning with @samp{_}. The @samp{_} should be
671 followed by the chosen name prefix for the library, to prevent
672 collisions with other libraries. These can go in the same files with
673 user entry points if you like.
675 Static functions and variables can be used as you like and need not
676 fit any naming convention.
679 @section Formatting Error Messages
680 @cindex formatting error messages
681 @cindex error messages, formatting
683 Error messages from compilers should look like this:
686 @var{source-file-name}:@var{lineno}: @var{message}
690 If you want to mention the column number, use one of these formats:
693 @var{source-file-name}:@var{lineno}:@var{column}: @var{message}
694 @var{source-file-name}:@var{lineno}.@var{column}: @var{message}
699 Line numbers should start from 1 at the beginning of the file, and
700 column numbers should start from 1 at the beginning of the line. (Both
701 of these conventions are chosen for compatibility.) Calculate column
702 numbers assuming that space and all ASCII printing characters have
703 equal width, and assuming tab stops every 8 columns.
705 The error message can also give both the starting and ending positions
706 of the erroneous text. There are several formats so that you can
707 avoid redundant information such as a duplicate line number.
708 Here are the possible formats:
711 @var{source-file-name}:@var{lineno-1}.@var{column-1}-@var{lineno-2}.@var{column-2}: @var{message}
712 @var{source-file-name}:@var{lineno-1}.@var{column-1}-@var{column-2}: @var{message}
713 @var{source-file-name}:@var{lineno-1}-@var{lineno-2}: @var{message}
717 When an error is spread over several files, you can use this format:
720 @var{file-1}:@var{lineno-1}.@var{column-1}-@var{file-2}:@var{lineno-2}.@var{column-2}: @var{message}
723 Error messages from other noninteractive programs should look like this:
726 @var{program}:@var{source-file-name}:@var{lineno}: @var{message}
730 when there is an appropriate source file, or like this:
733 @var{program}: @var{message}
737 when there is no relevant source file.
739 If you want to mention the column number, use this format:
742 @var{program}:@var{source-file-name}:@var{lineno}:@var{column}: @var{message}
745 In an interactive program (one that is reading commands from a
746 terminal), it is better not to include the program name in an error
747 message. The place to indicate which program is running is in the
748 prompt or with the screen layout. (When the same program runs with
749 input from a source other than a terminal, it is not interactive and
750 would do best to print error messages using the noninteractive style.)
752 The string @var{message} should not begin with a capital letter when
753 it follows a program name and/or file name, because that isn't the
754 beginning of a sentence. (The sentence conceptually starts at the
755 beginning of the line.) Also, it should not end with a period.
757 Error messages from interactive programs, and other messages such as
758 usage messages, should start with a capital letter. But they should not
761 @node User Interfaces
762 @section Standards for Interfaces Generally
764 @cindex program name and its behavior
765 @cindex behavior, dependent on program's name
766 Please don't make the behavior of a utility depend on the name used
767 to invoke it. It is useful sometimes to make a link to a utility
768 with a different name, and that should not change what it does.
770 Instead, use a run time option or a compilation switch or both
771 to select among the alternate behaviors.
773 @cindex output device and program's behavior
774 Likewise, please don't make the behavior of the program depend on the
775 type of output device it is used with. Device independence is an
776 important principle of the system's design; do not compromise it merely
777 to save someone from typing an option now and then. (Variation in error
778 message syntax when using a terminal is ok, because that is a side issue
779 that people do not depend on.)
781 If you think one behavior is most useful when the output is to a
782 terminal, and another is most useful when the output is a file or a
783 pipe, then it is usually best to make the default behavior the one that
784 is useful with output to a terminal, and have an option for the other
787 Compatibility requires certain programs to depend on the type of output
788 device. It would be disastrous if @code{ls} or @code{sh} did not do so
789 in the way all users expect. In some of these cases, we supplement the
790 program with a preferred alternate version that does not depend on the
791 output device type. For example, we provide a @code{dir} program much
792 like @code{ls} except that its default output format is always
795 @node Graphical Interfaces
796 @section Standards for Graphical Interfaces
797 @cindex graphical user interface
800 When you write a program that provides a graphical user interface,
801 please make it work with X Windows and the GTK+ toolkit unless the
802 functionality specifically requires some alternative (for example,
803 ``displaying jpeg images while in console mode'').
805 In addition, please provide a command-line interface to control the
806 functionality. (In many cases, the graphical user interface can be a
807 separate program which invokes the command-line program.) This is
808 so that the same jobs can be done from scripts.
812 Please also consider providing a CORBA interface (for use from GNOME), a
813 library interface (for use from C), and perhaps a keyboard-driven
814 console interface (for use by users from console mode). Once you are
815 doing the work to provide the functionality and the graphical interface,
816 these won't be much extra work.
818 @node Command-Line Interfaces
819 @section Standards for Command Line Interfaces
820 @cindex command-line interface
823 It is a good idea to follow the @sc{posix} guidelines for the
824 command-line options of a program. The easiest way to do this is to use
825 @code{getopt} to parse them. Note that the GNU version of @code{getopt}
826 will normally permit options anywhere among the arguments unless the
827 special argument @samp{--} is used. This is not what @sc{posix}
828 specifies; it is a GNU extension.
830 @cindex long-named options
831 Please define long-named options that are equivalent to the
832 single-letter Unix-style options. We hope to make GNU more user
833 friendly this way. This is easy to do with the GNU function
836 One of the advantages of long-named options is that they can be
837 consistent from program to program. For example, users should be able
838 to expect the ``verbose'' option of any GNU program which has one, to be
839 spelled precisely @samp{--verbose}. To achieve this uniformity, look at
840 the table of common long-option names when you choose the option names
841 for your program (@pxref{Option Table}).
843 It is usually a good idea for file names given as ordinary arguments to
844 be input files only; any output files would be specified using options
845 (preferably @samp{-o} or @samp{--output}). Even if you allow an output
846 file name as an ordinary argument for compatibility, try to provide an
847 option as another way to specify it. This will lead to more consistency
848 among GNU utilities, and fewer idiosyncracies for users to remember.
850 @cindex standard command-line options
851 @cindex options, standard command-line
852 @cindex CGI programs, standard options for
853 @cindex PATH_INFO, specifying standard options as
854 All programs should support two standard options: @samp{--version}
855 and @samp{--help}. CGI programs should accept these as command-line
856 options, and also if given as the @env{PATH_INFO}; for instance,
857 visiting @url{http://example.org/p.cgi/--help} in a browser should
858 output the same information as inokving @samp{p.cgi --help} from the
862 @cindex @samp{--version} option
864 This option should direct the program to print information about its name,
865 version, origin and legal status, all on standard output, and then exit
866 successfully. Other options and arguments should be ignored once this
867 is seen, and the program should not perform its normal function.
869 @cindex canonical name of a program
870 @cindex program's canonical name
871 The first line is meant to be easy for a program to parse; the version
872 number proper starts after the last space. In addition, it contains
873 the canonical name for this program, in this format:
880 The program's name should be a constant string; @emph{don't} compute it
881 from @code{argv[0]}. The idea is to state the standard or canonical
882 name for the program, not its file name. There are other ways to find
883 out the precise file name where a command is found in @code{PATH}.
885 If the program is a subsidiary part of a larger package, mention the
886 package name in parentheses, like this:
889 emacsserver (GNU Emacs) 19.30
893 If the package has a version number which is different from this
894 program's version number, you can mention the package version number
895 just before the close-parenthesis.
897 If you @strong{need} to mention the version numbers of libraries which
898 are distributed separately from the package which contains this program,
899 you can do so by printing an additional line of version info for each
900 library you want to mention. Use the same format for these lines as for
903 Please do not mention all of the libraries that the program uses ``just
904 for completeness''---that would produce a lot of unhelpful clutter.
905 Please mention library version numbers only if you find in practice that
906 they are very important to you in debugging.
908 The following line, after the version number line or lines, should be a
909 copyright notice. If more than one copyright notice is called for, put
910 each on a separate line.
912 Next should follow a brief statement that the program is free software,
913 and that users are free to copy and change it on certain conditions. If
914 the program is covered by the GNU GPL, say so here. Also mention that
915 there is no warranty, to the extent permitted by law.
917 It is ok to finish the output with a list of the major authors of the
918 program, as a way of giving credit.
920 Here's an example of output that follows these rules:
924 Copyright (C) 1996 Free Software Foundation, Inc.
925 GNU Emacs comes with NO WARRANTY,
926 to the extent permitted by law.
927 You may redistribute copies of GNU Emacs
928 under the terms of the GNU General Public License.
929 For more information about these matters,
930 see the files named COPYING.
933 You should adapt this to your program, of course, filling in the proper
934 year, copyright holder, name of program, and the references to
935 distribution terms, and changing the rest of the wording as necessary.
937 This copyright notice only needs to mention the most recent year in
938 which changes were made---there's no need to list the years for previous
939 versions' changes. You don't have to mention the name of the program in
940 these notices, if that is inconvenient, since it appeared in the first
943 Translations of the above lines must preserve the validity of the
944 copyright notices (@pxref{Internationalization}). If the translation's
945 character set supports it, the @samp{(C)} should be replaced with the
946 copyright symbol, as follows:
949 (the official copyright symbol, which is the letter C in a circle);
955 Write the word ``Copyright'' exactly like that, in English. Do not
956 translate it into another language. International treaties recognize
957 the English word ``Copyright''; translations into other languages do not
958 have legal significance.
961 @cindex @samp{--help} option
963 This option should output brief documentation for how to invoke the
964 program, on standard output, then exit successfully. Other options and
965 arguments should be ignored once this is seen, and the program should
966 not perform its normal function.
968 @cindex address for bug reports
970 Near the end of the @samp{--help} option's output there should be a line
971 that says where to mail bug reports. It should have this format:
974 Report bugs to @var{mailing-address}.
979 @section Table of Long Options
980 @cindex long option names
981 @cindex table of long options
983 Here is a table of long options used by GNU programs. It is surely
984 incomplete, but we aim to list all the options that a new program might
985 want to be compatible with. If you use names not already in the table,
986 please send @email{bug-standards@@gnu.org} a list of them, with their
987 meanings, so we can update the table.
989 @c Please leave newlines between items in this table; it's much easier
990 @c to update when it isn't completely squashed together and unreadable.
991 @c When there is more than one short option for a long option name, put
992 @c a semicolon between the lists of the programs that use them, not a
993 @c period. --friedman
997 @samp{-N} in @code{tar}.
1000 @samp{-a} in @code{du}, @code{ls}, @code{nm}, @code{stty}, @code{uname},
1001 and @code{unexpand}.
1004 @samp{-a} in @code{diff}.
1007 @samp{-A} in @code{ls}.
1010 @samp{-a} in @code{etags}, @code{tee}, @code{time};
1011 @samp{-r} in @code{tar}.
1014 @samp{-a} in @code{cp}.
1017 @samp{-n} in @code{shar}.
1020 @samp{-l} in @code{m4}.
1023 @samp{-a} in @code{diff}.
1026 @samp{-v} in @code{gawk}.
1035 @samp{-a} in @code{recode}.
1038 @samp{-a} in @code{wdiff}.
1040 @item auto-reference
1041 @samp{-A} in @code{ptx}.
1044 @samp{-n} in @code{wdiff}.
1047 For server programs, run in the background.
1049 @item backward-search
1050 @samp{-B} in @code{ctags}.
1053 @samp{-f} in @code{shar}.
1062 @samp{-b} in @code{tac}.
1065 @samp{-b} in @code{cpio} and @code{diff}.
1068 @samp{-b} in @code{shar}.
1071 Used in @code{cpio} and @code{tar}.
1074 @samp{-b} in @code{head} and @code{tail}.
1077 @samp{-b} in @code{ptx}.
1080 Used in various programs to make output shorter.
1083 @samp{-c} in @code{head}, @code{split}, and @code{tail}.
1086 @samp{-C} in @code{etags}.
1089 @samp{-A} in @code{tar}.
1092 Used in various programs to specify the directory to use.
1095 @samp{-c} in @code{chgrp} and @code{chown}.
1098 @samp{-F} in @code{ls}.
1101 @samp{-c} in @code{recode}.
1104 @samp{-c} in @code{su};
1108 @samp{-d} in @code{tar}.
1111 Used in @code{gawk}.
1114 @samp{-Z} in @code{tar} and @code{shar}.
1117 @samp{-A} in @code{tar}.
1120 @samp{-w} in @code{tar}.
1123 Used in @code{diff}.
1126 @samp{-W copyleft} in @code{gawk}.
1129 @samp{-C} in @code{ptx}, @code{recode}, and @code{wdiff};
1130 @samp{-W copyright} in @code{gawk}.
1136 @samp{-q} in @code{who}.
1139 @samp{-l} in @code{du}.
1142 Used in @code{tar} and @code{cpio}.
1145 @samp{-c} in @code{shar}.
1148 @samp{-x} in @code{ctags}.
1151 @samp{-d} in @code{touch}.
1154 @samp{-d} in Make and @code{m4};
1158 @samp{-D} in @code{m4}.
1161 @samp{-d} in Bison and @code{ctags}.
1164 @samp{-D} in @code{tar}.
1167 @samp{-L} in @code{chgrp}, @code{chown}, @code{cpio}, @code{du},
1168 @code{ls}, and @code{tar}.
1170 @item dereference-args
1171 @samp{-D} in @code{du}.
1174 Specify an I/O device (special file name).
1177 @samp{-d} in @code{recode}.
1179 @item dictionary-order
1180 @samp{-d} in @code{look}.
1183 @samp{-d} in @code{tar}.
1186 @samp{-n} in @code{csplit}.
1189 Specify the directory to use, in various programs. In @code{ls}, it
1190 means to show directories themselves rather than their contents. In
1191 @code{rm} and @code{ln}, it means to not treat links to directories
1195 @samp{-x} in @code{strip}.
1197 @item discard-locals
1198 @samp{-X} in @code{strip}.
1204 @samp{-e} in @code{diff}.
1206 @item elide-empty-files
1207 @samp{-z} in @code{csplit}.
1210 @samp{-x} in @code{wdiff}.
1213 @samp{-z} in @code{wdiff}.
1215 @item entire-new-file
1216 @samp{-N} in @code{diff}.
1218 @item environment-overrides
1222 @samp{-e} in @code{xargs}.
1228 Used in @code{makeinfo}.
1231 @samp{-o} in @code{m4}.
1234 @samp{-b} in @code{ls}.
1237 @samp{-X} in @code{tar}.
1243 @samp{-x} in @code{xargs}.
1246 @samp{-e} in @code{unshar}.
1249 @samp{-t} in @code{diff}.
1252 @samp{-e} in @code{sed}.
1255 @samp{-g} in @code{nm}.
1258 @samp{-i} in @code{cpio};
1259 @samp{-x} in @code{tar}.
1262 @samp{-f} in @code{finger}.
1265 @samp{-f} in @code{su}.
1267 @item fatal-warnings
1268 @samp{-E} in @code{m4}.
1271 @samp{-f} in @code{info}, @code{gawk}, Make, @code{mt}, and @code{tar};
1272 @samp{-n} in @code{sed};
1273 @samp{-r} in @code{touch}.
1275 @item field-separator
1276 @samp{-F} in @code{gawk}.
1282 @samp{-F} in @code{ls}.
1285 @samp{-T} in @code{tar}.
1288 Used in @code{makeinfo}.
1290 @item flag-truncation
1291 @samp{-F} in @code{ptx}.
1293 @item fixed-output-files
1297 @samp{-f} in @code{tail}.
1299 @item footnote-style
1300 Used in @code{makeinfo}.
1303 @samp{-f} in @code{cp}, @code{ln}, @code{mv}, and @code{rm}.
1306 @samp{-F} in @code{shar}.
1309 For server programs, run in the foreground;
1310 in other words, don't do anything special to run the server
1314 Used in @code{ls}, @code{time}, and @code{ptx}.
1317 @samp{-F} in @code{m4}.
1323 @samp{-g} in @code{ptx}.
1326 @samp{-x} in @code{tar}.
1329 @samp{-i} in @code{ul}.
1332 @samp{-g} in @code{recode}.
1335 @samp{-g} in @code{install}.
1338 @samp{-z} in @code{tar} and @code{shar}.
1341 @samp{-H} in @code{m4}.
1344 @samp{-h} in @code{objdump} and @code{recode}
1347 @samp{-H} in @code{who}.
1350 Used to ask for brief usage information.
1352 @item here-delimiter
1353 @samp{-d} in @code{shar}.
1355 @item hide-control-chars
1356 @samp{-q} in @code{ls}.
1359 In @code{makeinfo}, output HTML.
1362 @samp{-u} in @code{who}.
1365 @samp{-D} in @code{diff}.
1368 @samp{-I} in @code{ls};
1369 @samp{-x} in @code{recode}.
1371 @item ignore-all-space
1372 @samp{-w} in @code{diff}.
1374 @item ignore-backups
1375 @samp{-B} in @code{ls}.
1377 @item ignore-blank-lines
1378 @samp{-B} in @code{diff}.
1381 @samp{-f} in @code{look} and @code{ptx};
1382 @samp{-i} in @code{diff} and @code{wdiff}.
1388 @samp{-i} in @code{ptx}.
1390 @item ignore-indentation
1391 @samp{-I} in @code{etags}.
1393 @item ignore-init-file
1396 @item ignore-interrupts
1397 @samp{-i} in @code{tee}.
1399 @item ignore-matching-lines
1400 @samp{-I} in @code{diff}.
1402 @item ignore-space-change
1403 @samp{-b} in @code{diff}.
1406 @samp{-i} in @code{tar}.
1409 @samp{-i} in @code{etags};
1410 @samp{-I} in @code{m4}.
1416 @samp{-G} in @code{tar}.
1419 @samp{-i}, @samp{-l}, and @samp{-m} in Finger.
1422 In some programs, specify the name of the file to read as the user's
1426 @samp{-i} in @code{expand}.
1429 @samp{-T} in @code{diff}.
1432 @samp{-i} in @code{ls}.
1435 @samp{-i} in @code{cp}, @code{ln}, @code{mv}, @code{rm};
1436 @samp{-e} in @code{m4};
1437 @samp{-p} in @code{xargs};
1438 @samp{-w} in @code{tar}.
1441 @samp{-p} in @code{shar}.
1456 @samp{-k} in @code{csplit}.
1459 @samp{-k} in @code{du} and @code{ls}.
1462 @samp{-l} in @code{etags}.
1465 @samp{-l} in @code{wdiff}.
1467 @item level-for-gzip
1468 @samp{-g} in @code{shar}.
1471 @samp{-C} in @code{split}.
1474 Used in @code{split}, @code{head}, and @code{tail}.
1477 @samp{-l} in @code{cpio}.
1481 Used in @code{gawk}.
1484 @samp{-t} in @code{cpio};
1485 @samp{-l} in @code{recode}.
1488 @samp{-t} in @code{tar}.
1491 @samp{-N} in @code{ls}.
1500 No listing of which programs already use this;
1501 someone should check to
1502 see if any actually do, and tell @email{gnu@@gnu.org}.
1505 @samp{-M} in @code{ptx}.
1508 @samp{-m} in @code{hello} and @code{uname}.
1510 @item make-directories
1511 @samp{-d} in @code{cpio}.
1520 @samp{-n} in @code{xargs}.
1523 @samp{-n} in @code{xargs}.
1526 @samp{-l} in @code{xargs}.
1532 @samp{-P} in @code{xargs}.
1535 @samp{-T} in @code{who}.
1538 @samp{-T} in @code{who}.
1541 @samp{-d} in @code{diff}.
1543 @item mixed-uuencode
1544 @samp{-M} in @code{shar}.
1547 @samp{-m} in @code{install}, @code{mkdir}, and @code{mkfifo}.
1549 @item modification-time
1550 @samp{-m} in @code{tar}.
1553 @samp{-M} in @code{tar}.
1559 @samp{-L} in @code{m4}.
1562 @samp{-a} in @code{shar}.
1567 @item no-builtin-rules
1570 @item no-character-count
1571 @samp{-w} in @code{shar}.
1573 @item no-check-existing
1574 @samp{-x} in @code{shar}.
1577 @samp{-3} in @code{wdiff}.
1580 @samp{-c} in @code{touch}.
1583 @samp{-D} in @code{etags}.
1586 @samp{-1} in @code{wdiff}.
1588 @item no-dereference
1589 @samp{-d} in @code{cp}.
1592 @samp{-2} in @code{wdiff}.
1601 @samp{-P} in @code{shar}.
1604 @samp{-e} in @code{gprof}.
1607 @samp{-R} in @code{etags}.
1610 @samp{-p} in @code{nm}.
1613 Don't print a startup splash screen.
1616 Used in @code{makeinfo}.
1619 @samp{-a} in @code{gprof}.
1622 @samp{-E} in @code{gprof}.
1625 @samp{-m} in @code{shar}.
1628 Used in @code{makeinfo}.
1631 Used in @code{emacsclient}.
1634 Used in various programs to inhibit warnings.
1637 @samp{-n} in @code{info}.
1640 @samp{-n} in @code{uname}.
1643 @samp{-f} in @code{cpio}.
1646 @samp{-n} in @code{objdump}.
1649 @samp{-0} in @code{xargs}.
1652 @samp{-n} in @code{cat}.
1654 @item number-nonblank
1655 @samp{-b} in @code{cat}.
1658 @samp{-n} in @code{nm}.
1660 @item numeric-uid-gid
1661 @samp{-n} in @code{cpio} and @code{ls}.
1667 @samp{-o} in @code{tar}.
1672 @item one-file-system
1673 @samp{-l} in @code{tar}, @code{cp}, and @code{du}.
1676 @samp{-o} in @code{ptx}.
1679 @samp{-f} in @code{gprof}.
1682 @samp{-F} in @code{gprof}.
1685 @samp{-o} in @code{getopt}, @code{fdlist}, @code{fdmount},
1686 @code{fdmountd}, and @code{fdumount}.
1689 In various programs, specify the output file name.
1692 @samp{-o} in @code{shar}.
1695 @samp{-o} in @code{rm}.
1698 @samp{-c} in @code{unshar}.
1701 @samp{-o} in @code{install}.
1704 @samp{-l} in @code{diff}.
1706 @item paragraph-indent
1707 Used in @code{makeinfo}.
1710 @samp{-p} in @code{mkdir} and @code{rmdir}.
1713 @samp{-p} in @code{ul}.
1716 @samp{-p} in @code{cpio}.
1719 @samp{-P} in @code{finger}.
1722 @samp{-c} in @code{cpio} and @code{tar}.
1725 Used in @code{gawk}.
1727 @item prefix-builtins
1728 @samp{-P} in @code{m4}.
1731 @samp{-f} in @code{csplit}.
1734 Used in @code{tar} and @code{cp}.
1736 @item preserve-environment
1737 @samp{-p} in @code{su}.
1739 @item preserve-modification-time
1740 @samp{-m} in @code{cpio}.
1742 @item preserve-order
1743 @samp{-s} in @code{tar}.
1745 @item preserve-permissions
1746 @samp{-p} in @code{tar}.
1749 @samp{-l} in @code{diff}.
1752 @samp{-L} in @code{cmp}.
1754 @item print-data-base
1757 @item print-directory
1760 @item print-file-name
1761 @samp{-o} in @code{nm}.
1764 @samp{-s} in @code{nm}.
1767 @samp{-p} in @code{wdiff}.
1770 @samp{-p} in @code{ed}.
1773 Specify an HTTP proxy.
1776 @samp{-X} in @code{shar}.
1782 Used in many programs to inhibit the usual output. Every
1783 program accepting @samp{--quiet} should accept @samp{--silent} as a
1787 @samp{-Q} in @code{shar}
1790 @samp{-Q} in @code{ls}.
1793 @samp{-n} in @code{diff}.
1796 Used in @code{gawk}.
1798 @item read-full-blocks
1799 @samp{-B} in @code{tar}.
1808 @samp{-R} in @code{tar}.
1811 Used in @code{chgrp}, @code{chown}, @code{cp}, @code{ls}, @code{diff},
1814 @item reference-limit
1815 Used in @code{makeinfo}.
1818 @samp{-r} in @code{ptx}.
1821 @samp{-r} in @code{tac} and @code{etags}.
1824 @samp{-r} in @code{uname}.
1827 @samp{-R} in @code{m4}.
1830 @samp{-r} in @code{objdump}.
1833 @samp{-r} in @code{cpio}.
1836 @samp{-i} in @code{xargs}.
1838 @item report-identical-files
1839 @samp{-s} in @code{diff}.
1841 @item reset-access-time
1842 @samp{-a} in @code{cpio}.
1845 @samp{-r} in @code{ls} and @code{nm}.
1848 @samp{-f} in @code{diff}.
1850 @item right-side-defs
1851 @samp{-R} in @code{ptx}.
1854 @samp{-s} in @code{tar}.
1856 @item same-permissions
1857 @samp{-p} in @code{tar}.
1860 @samp{-g} in @code{stty}.
1865 @item sentence-regexp
1866 @samp{-S} in @code{ptx}.
1869 @samp{-S} in @code{du}.
1872 @samp{-s} in @code{tac}.
1875 Used by @code{recode} to chose files or pipes for sequencing passes.
1878 @samp{-s} in @code{su}.
1881 @samp{-A} in @code{cat}.
1883 @item show-c-function
1884 @samp{-p} in @code{diff}.
1887 @samp{-E} in @code{cat}.
1889 @item show-function-line
1890 @samp{-F} in @code{diff}.
1893 @samp{-T} in @code{cat}.
1896 Used in many programs to inhibit the usual output.
1897 Every program accepting
1898 @samp{--silent} should accept @samp{--quiet} as a synonym.
1901 @samp{-s} in @code{ls}.
1904 Specify a file descriptor for a network server to use for its socket,
1905 instead of opening and binding a new socket. This provides a way to
1906 run, in a nonpriveledged process, a server that normally needs a
1907 reserved port number.
1913 @samp{-W source} in @code{gawk}.
1916 @samp{-S} in @code{tar}.
1918 @item speed-large-files
1919 @samp{-H} in @code{diff}.
1922 @samp{-E} in @code{unshar}.
1924 @item split-size-limit
1925 @samp{-L} in @code{shar}.
1928 @samp{-s} in @code{cat}.
1931 @samp{-w} in @code{wdiff}.
1934 @samp{-y} in @code{wdiff}.
1937 Used in @code{tar} and @code{diff} to specify which file within
1938 a directory to start processing with.
1941 @samp{-s} in @code{wdiff}.
1943 @item stdin-file-list
1944 @samp{-S} in @code{shar}.
1950 @samp{-s} in @code{recode}.
1953 @samp{-s} in @code{install}.
1956 @samp{-s} in @code{strip}.
1959 @samp{-S} in @code{strip}.
1962 @samp{-s} in @code{shar}.
1965 @samp{-S} in @code{cp}, @code{ln}, @code{mv}.
1968 @samp{-b} in @code{csplit}.
1971 @samp{-s} in @code{gprof}.
1974 @samp{-s} in @code{du}.
1977 @samp{-s} in @code{ln}.
1980 Used in GDB and @code{objdump}.
1983 @samp{-s} in @code{m4}.
1986 @samp{-s} in @code{uname}.
1989 @samp{-t} in @code{expand} and @code{unexpand}.
1992 @samp{-T} in @code{ls}.
1995 @samp{-T} in @code{tput} and @code{ul}.
1996 @samp{-t} in @code{wdiff}.
1999 @samp{-a} in @code{diff}.
2002 @samp{-T} in @code{shar}.
2005 Used in @code{ls} and @code{touch}.
2008 Specify how long to wait before giving up on some operation.
2011 @samp{-O} in @code{tar}.
2014 @samp{-c} in @code{du}.
2017 @samp{-t} in Make, @code{ranlib}, and @code{recode}.
2020 @samp{-t} in @code{m4}.
2023 @samp{-t} in @code{hello};
2024 @samp{-W traditional} in @code{gawk};
2025 @samp{-G} in @code{ed}, @code{m4}, and @code{ptx}.
2031 @samp{-t} in @code{ctags}.
2033 @item typedefs-and-c++
2034 @samp{-T} in @code{ctags}.
2037 @samp{-t} in @code{ptx}.
2040 @samp{-z} in @code{tar}.
2043 @samp{-u} in @code{cpio}.
2046 @samp{-U} in @code{m4}.
2048 @item undefined-only
2049 @samp{-u} in @code{nm}.
2052 @samp{-u} in @code{cp}, @code{ctags}, @code{mv}, @code{tar}.
2055 Used in @code{gawk}; same as @samp{--help}.
2058 @samp{-B} in @code{shar}.
2060 @item vanilla-operation
2061 @samp{-V} in @code{shar}.
2064 Print more information about progress. Many programs support this.
2067 @samp{-W} in @code{tar}.
2070 Print the version number.
2072 @item version-control
2073 @samp{-V} in @code{cp}, @code{ln}, @code{mv}.
2076 @samp{-v} in @code{ctags}.
2079 @samp{-V} in @code{tar}.
2084 @item whole-size-limit
2085 @samp{-l} in @code{shar}.
2088 @samp{-w} in @code{ls} and @code{ptx}.
2091 @samp{-W} in @code{ptx}.
2094 @samp{-T} in @code{who}.
2097 @samp{-z} in @code{gprof}.
2101 @section Memory Usage
2102 @cindex memory usage
2104 If a program typically uses just a few meg of memory, don't bother making any
2105 effort to reduce memory usage. For example, if it is impractical for
2106 other reasons to operate on files more than a few meg long, it is
2107 reasonable to read entire input files into core to operate on them.
2109 However, for programs such as @code{cat} or @code{tail}, that can
2110 usefully operate on very large files, it is important to avoid using a
2111 technique that would artificially limit the size of files it can handle.
2112 If a program works by lines and could be applied to arbitrary
2113 user-supplied input files, it should keep only a line in memory, because
2114 this is not very hard and users will want to be able to operate on input
2115 files that are bigger than will fit in core all at once.
2117 If your program creates complicated data structures, just make them in
2118 core and give a fatal error if @code{malloc} returns zero.
2124 Programs should be prepared to operate when @file{/usr} and @file{/etc}
2125 are read-only file systems. Thus, if the program manages log files,
2126 lock files, backup files, score files, or any other files which are
2127 modified for internal purposes, these files should not be stored in
2128 @file{/usr} or @file{/etc}.
2130 There are two exceptions. @file{/etc} is used to store system
2131 configuration information; it is reasonable for a program to modify
2132 files in @file{/etc} when its job is to update the system configuration.
2133 Also, if the user explicitly asks to modify one file in a directory, it
2134 is reasonable for the program to store other files in the same
2138 @chapter Making The Best Use of C
2140 This @value{CHAPTER} provides advice on how best to use the C language
2141 when writing GNU software.
2144 * Formatting:: Formatting Your Source Code
2145 * Comments:: Commenting Your Work
2146 * Syntactic Conventions:: Clean Use of C Constructs
2147 * Names:: Naming Variables, Functions, and Files
2148 * System Portability:: Portability between different operating systems
2149 * CPU Portability:: Supporting the range of CPU types
2150 * System Functions:: Portability and ``standard'' library functions
2151 * Internationalization:: Techniques for internationalization
2152 * Mmap:: How you can safely use @code{mmap}.
2156 @section Formatting Your Source Code
2157 @cindex formatting source code
2160 @cindex braces, in C source
2161 It is important to put the open-brace that starts the body of a C
2162 function in column zero, and avoid putting any other open-brace or
2163 open-parenthesis or open-bracket in column zero. Several tools look
2164 for open-braces in column zero to find the beginnings of C functions.
2165 These tools will not work on code not formatted that way.
2167 It is also important for function definitions to start the name of the
2168 function in column zero. This helps people to search for function
2169 definitions, and may also help certain tools recognize them. Thus,
2170 the proper format is this:
2174 concat (s1, s2) /* Name starts in column zero here */
2176 @{ /* Open brace in column zero here */
2182 or, if you want to use Standard C syntax, format the definition like
2187 concat (char *s1, char *s2)
2193 In Standard C, if the arguments don't fit nicely on one line,
2198 lots_of_args (int an_integer, long a_long, short a_short,
2199 double a_double, float a_float)
2203 The rest of this section gives our recommendations for other aspects of
2204 C formatting style, which is also the default style of the @code{indent}
2205 program in version 1.2 and newer. It corresponds to the options
2208 -nbad -bap -nbc -bbo -bl -bli2 -bls -ncdb -nce -cp1 -cs -di2
2209 -ndj -nfc1 -nfca -hnl -i2 -ip5 -lp -pcs -psl -nsc -nsob
2212 We don't think of these recommendations as requirements, because it
2213 causes no problems for users if two different programs have different
2216 But whatever style you use, please use it consistently, since a mixture
2217 of styles within one program tends to look ugly. If you are
2218 contributing changes to an existing program, please follow the style of
2221 For the body of the function, our recommended style looks like this:
2233 return ++x + bar ();
2237 @cindex spaces before open-paren
2238 We find it easier to read a program when it has spaces before the
2239 open-parentheses and after the commas. Especially after the commas.
2241 When you split an expression into multiple lines, split it
2242 before an operator, not after one. Here is the right way:
2244 @cindex expressions, splitting
2246 if (foo_this_is_long && bar > win (x, y, z)
2247 && remaining_condition)
2250 Try to avoid having two operators of different precedence at the same
2251 level of indentation. For example, don't write this:
2254 mode = (inmode[j] == VOIDmode
2255 || GET_MODE_SIZE (outmode[j]) > GET_MODE_SIZE (inmode[j])
2256 ? outmode[j] : inmode[j]);
2259 Instead, use extra parentheses so that the indentation shows the nesting:
2262 mode = ((inmode[j] == VOIDmode
2263 || (GET_MODE_SIZE (outmode[j]) > GET_MODE_SIZE (inmode[j])))
2264 ? outmode[j] : inmode[j]);
2267 Insert extra parentheses so that Emacs will indent the code properly.
2268 For example, the following indentation looks nice if you do it by hand,
2271 v = rup->ru_utime.tv_sec*1000 + rup->ru_utime.tv_usec/1000
2272 + rup->ru_stime.tv_sec*1000 + rup->ru_stime.tv_usec/1000;
2276 but Emacs would alter it. Adding a set of parentheses produces
2277 something that looks equally nice, and which Emacs will preserve:
2280 v = (rup->ru_utime.tv_sec*1000 + rup->ru_utime.tv_usec/1000
2281 + rup->ru_stime.tv_sec*1000 + rup->ru_stime.tv_usec/1000);
2284 Format do-while statements like this:
2296 Please use formfeed characters (control-L) to divide the program into
2297 pages at logical places (but not within a function). It does not matter
2298 just how long the pages are, since they do not have to fit on a printed
2299 page. The formfeeds should appear alone on lines by themselves.
2302 @section Commenting Your Work
2305 Every program should start with a comment saying briefly what it is for.
2306 Example: @samp{fmt - filter for simple filling of text}.
2308 Please write the comments in a GNU program in English, because English
2309 is the one language that nearly all programmers in all countries can
2310 read. If you do not write English well, please write comments in
2311 English as well as you can, then ask other people to help rewrite them.
2312 If you can't write comments in English, please find someone to work with
2313 you and translate your comments into English.
2315 Please put a comment on each function saying what the function does,
2316 what sorts of arguments it gets, and what the possible values of
2317 arguments mean and are used for. It is not necessary to duplicate in
2318 words the meaning of the C argument declarations, if a C type is being
2319 used in its customary fashion. If there is anything nonstandard about
2320 its use (such as an argument of type @code{char *} which is really the
2321 address of the second character of a string, not the first), or any
2322 possible values that would not work the way one would expect (such as,
2323 that strings containing newlines are not guaranteed to work), be sure
2326 Also explain the significance of the return value, if there is one.
2328 Please put two spaces after the end of a sentence in your comments, so
2329 that the Emacs sentence commands will work. Also, please write
2330 complete sentences and capitalize the first word. If a lower-case
2331 identifier comes at the beginning of a sentence, don't capitalize it!
2332 Changing the spelling makes it a different identifier. If you don't
2333 like starting a sentence with a lower case letter, write the sentence
2334 differently (e.g., ``The identifier lower-case is @dots{}'').
2336 The comment on a function is much clearer if you use the argument
2337 names to speak about the argument values. The variable name itself
2338 should be lower case, but write it in upper case when you are speaking
2339 about the value rather than the variable itself. Thus, ``the inode
2340 number NODE_NUM'' rather than ``an inode''.
2342 There is usually no purpose in restating the name of the function in
2343 the comment before it, because the reader can see that for himself.
2344 There might be an exception when the comment is so long that the function
2345 itself would be off the bottom of the screen.
2347 There should be a comment on each static variable as well, like this:
2350 /* Nonzero means truncate lines in the display;
2351 zero means continue them. */
2355 @cindex conditionals, comments for
2356 @cindex @code{#endif}, commenting
2357 Every @samp{#endif} should have a comment, except in the case of short
2358 conditionals (just a few lines) that are not nested. The comment should
2359 state the condition of the conditional that is ending, @emph{including
2360 its sense}. @samp{#else} should have a comment describing the condition
2361 @emph{and sense} of the code that follows. For example:
2369 #endif /* not foo */
2379 but, by contrast, write the comments this way for a @samp{#ifndef}:
2392 #endif /* not foo */
2396 @node Syntactic Conventions
2397 @section Clean Use of C Constructs
2398 @cindex syntactic conventions
2400 @cindex implicit @code{int}
2401 @cindex function argument, declaring
2402 Please explicitly declare the types of all objects. For example, you
2403 should explicitly declare all arguments to functions, and you should
2404 declare functions to return @code{int} rather than omitting the
2407 @cindex compiler warnings
2408 @cindex @samp{-Wall} compiler option
2409 Some programmers like to use the GCC @samp{-Wall} option, and change the
2410 code whenever it issues a warning. If you want to do this, then do.
2411 Other programmers prefer not to use @samp{-Wall}, because it gives
2412 warnings for valid and legitimate code which they do not want to change.
2413 If you want to do this, then do. The compiler should be your servant,
2416 Declarations of external functions and functions to appear later in the
2417 source file should all go in one place near the beginning of the file
2418 (somewhere before the first function definition in the file), or else
2419 should go in a header file. Don't put @code{extern} declarations inside
2422 @cindex temporary variables
2423 It used to be common practice to use the same local variables (with
2424 names like @code{tem}) over and over for different values within one
2425 function. Instead of doing this, it is better to declare a separate local
2426 variable for each distinct purpose, and give it a name which is
2427 meaningful. This not only makes programs easier to understand, it also
2428 facilitates optimization by good compilers. You can also move the
2429 declaration of each local variable into the smallest scope that includes
2430 all its uses. This makes the program even cleaner.
2432 Don't use local variables or parameters that shadow global identifiers.
2434 @cindex multiple variables in a line
2435 Don't declare multiple variables in one declaration that spans lines.
2436 Start a new declaration on each line, instead. For example, instead
2462 (If they are global variables, each should have a comment preceding it
2465 When you have an @code{if}-@code{else} statement nested in another
2466 @code{if} statement, always put braces around the @code{if}-@code{else}.
2467 Thus, never write like this:
2490 If you have an @code{if} statement nested inside of an @code{else}
2491 statement, either write @code{else if} on one line, like this,
2501 with its @code{then}-part indented like the preceding @code{then}-part,
2502 or write the nested @code{if} within braces like this:
2514 Don't declare both a structure tag and variables or typedefs in the
2515 same declaration. Instead, declare the structure tag separately
2516 and then use it to declare the variables or typedefs.
2518 Try to avoid assignments inside @code{if}-conditions. For example,
2522 if ((foo = (char *) malloc (sizeof *foo)) == 0)
2523 fatal ("virtual memory exhausted");
2527 instead, write this:
2530 foo = (char *) malloc (sizeof *foo);
2532 fatal ("virtual memory exhausted");
2536 Don't make the program ugly to placate @code{lint}. Please don't insert any
2537 casts to @code{void}. Zero without a cast is perfectly fine as a null
2538 pointer constant, except when calling a varargs function.
2541 @section Naming Variables, Functions, and Files
2543 @cindex names of variables, functions, and files
2544 The names of global variables and functions in a program serve as
2545 comments of a sort. So don't choose terse names---instead, look for
2546 names that give useful information about the meaning of the variable or
2547 function. In a GNU program, names should be English, like other
2550 Local variable names can be shorter, because they are used only within
2551 one context, where (presumably) comments explain their purpose.
2553 Try to limit your use of abbreviations in symbol names. It is ok to
2554 make a few abbreviations, explain what they mean, and then use them
2555 frequently, but don't use lots of obscure abbreviations.
2557 Please use underscores to separate words in a name, so that the Emacs
2558 word commands can be useful within them. Stick to lower case; reserve
2559 upper case for macros and @code{enum} constants, and for name-prefixes
2560 that follow a uniform convention.
2562 For example, you should use names like @code{ignore_space_change_flag};
2563 don't use names like @code{iCantReadThis}.
2565 Variables that indicate whether command-line options have been
2566 specified should be named after the meaning of the option, not after
2567 the option-letter. A comment should state both the exact meaning of
2568 the option and its letter. For example,
2572 /* Ignore changes in horizontal whitespace (-b). */
2573 int ignore_space_change_flag;
2577 When you want to define names with constant integer values, use
2578 @code{enum} rather than @samp{#define}. GDB knows about enumeration
2581 @cindex file-name limitations
2583 You might want to make sure that none of the file names would conflict
2584 the files were loaded onto an MS-DOS file system which shortens the
2585 names. You can use the program @code{doschk} to test for this.
2587 Some GNU programs were designed to limit themselves to file names of 14
2588 characters or less, to avoid file name conflicts if they are read into
2589 older System V systems. Please preserve this feature in the existing
2590 GNU programs that have it, but there is no need to do this in new GNU
2591 programs. @code{doschk} also reports file names longer than 14
2594 @node System Portability
2595 @section Portability between System Types
2596 @cindex portability, between system types
2598 In the Unix world, ``portability'' refers to porting to different Unix
2599 versions. For a GNU program, this kind of portability is desirable, but
2602 The primary purpose of GNU software is to run on top of the GNU kernel,
2603 compiled with the GNU C compiler, on various types of @sc{cpu}. So the
2604 kinds of portability that are absolutely necessary are quite limited.
2605 But it is important to support Linux-based GNU systems, since they
2606 are the form of GNU that is popular.
2608 Beyond that, it is good to support the other free operating systems
2609 (*BSD), and it is nice to support other Unix-like systems if you want
2610 to. Supporting a variety of Unix-like systems is desirable, although
2611 not paramount. It is usually not too hard, so you may as well do it.
2612 But you don't have to consider it an obligation, if it does turn out to
2616 The easiest way to achieve portability to most Unix-like systems is to
2617 use Autoconf. It's unlikely that your program needs to know more
2618 information about the host platform than Autoconf can provide, simply
2619 because most of the programs that need such knowledge have already been
2622 Avoid using the format of semi-internal data bases (e.g., directories)
2623 when there is a higher-level alternative (@code{readdir}).
2625 @cindex non-@sc{posix} systems, and portability
2626 As for systems that are not like Unix, such as MSDOS, Windows, the
2627 Macintosh, VMS, and MVS, supporting them is often a lot of work. When
2628 that is the case, it is better to spend your time adding features that
2629 will be useful on GNU and GNU/Linux, rather than on supporting other
2630 incompatible systems.
2632 If you do support Windows, please do not abbreviate it as ``win''. In
2633 hacker terminology, calling something a ``win'' is a form of praise.
2634 You're free to praise Microsoft Windows on your own if you want, but
2635 please don't do this in GNU packages. Instead of abbreviating
2636 ``Windows'' to ``un'', you can write it in full or abbreviate it to
2637 ``woe'' or ``w''. In GNU Emacs, for instance, we use @samp{w32} in
2638 file names of Windows-specific files, but the macro for Windows
2639 conditionals is called @code{WINDOWSNT}.
2641 It is a good idea to define the ``feature test macro''
2642 @code{_GNU_SOURCE} when compiling your C files. When you compile on GNU
2643 or GNU/Linux, this will enable the declarations of GNU library extension
2644 functions, and that will usually give you a compiler error message if
2645 you define the same function names in some other way in your program.
2646 (You don't have to actually @emph{use} these functions, if you prefer
2647 to make the program more portable to other systems.)
2649 But whether or not you use these GNU extensions, you should avoid
2650 using their names for any other meanings. Doing so would make it hard
2651 to move your code into other GNU programs.
2653 @node CPU Portability
2654 @section Portability between @sc{cpu}s
2656 @cindex data types, and portability
2657 @cindex portability, and data types
2658 Even GNU systems will differ because of differences among @sc{cpu}
2659 types---for example, difference in byte ordering and alignment
2660 requirements. It is absolutely essential to handle these differences.
2661 However, don't make any effort to cater to the possibility that an
2662 @code{int} will be less than 32 bits. We don't support 16-bit machines
2665 Similarly, don't make any effort to cater to the possibility that
2666 @code{long} will be smaller than predefined types like @code{size_t}.
2667 For example, the following code is ok:
2670 printf ("size = %lu\n", (unsigned long) sizeof array);
2671 printf ("diff = %ld\n", (long) (pointer2 - pointer1));
2674 1989 Standard C requires this to work, and we know of only one
2675 counterexample: 64-bit programs on Microsoft Windows IA-64. We will
2676 leave it to those who want to port GNU programs to that environment
2677 to figure out how to do it.
2679 Predefined file-size types like @code{off_t} are an exception: they are
2680 longer than @code{long} on many platforms, so code like the above won't
2681 work with them. One way to print an @code{off_t} value portably is to
2682 print its digits yourself, one by one.
2684 Don't assume that the address of an @code{int} object is also the
2685 address of its least-significant byte. This is false on big-endian
2686 machines. Thus, don't make the following mistake:
2691 while ((c = getchar()) != EOF)
2692 write(file_descriptor, &c, 1);
2695 When calling functions, you need not worry about the difference between
2696 pointers of various types, or between pointers and integers. On most
2697 machines, there's no difference anyway. As for the few machines where
2698 there is a difference, all of them support Standard C prototypes, so you can
2699 use prototypes (perhaps conditionalized to be active only in Standard C)
2700 to make the code work on those systems.
2702 In certain cases, it is ok to pass integer and pointer arguments
2703 indiscriminately to the same function, and use no prototype on any
2704 system. For example, many GNU programs have error-reporting functions
2705 that pass their arguments along to @code{printf} and friends:
2708 error (s, a1, a2, a3)
2712 fprintf (stderr, "error: ");
2713 fprintf (stderr, s, a1, a2, a3);
2718 In practice, this works on all machines, since a pointer is generally
2719 the widest possible kind of argument; it is much simpler than any
2720 ``correct'' alternative. Be sure @emph{not} to use a prototype for such
2723 If you have decided to use Standard C, then you can instead define
2724 @code{error} using @file{stdarg.h}, and pass the arguments along to
2727 @cindex casting pointers to integers
2728 Avoid casting pointers to integers if you can. Such casts greatly
2729 reduce portability, and in most programs they are easy to avoid. In the
2730 cases where casting pointers to integers is essential---such as, a Lisp
2731 interpreter which stores type information as well as an address in one
2732 word---you'll have to make explicit provisions to handle different word
2733 sizes. You will also need to make provision for systems in which the
2734 normal range of addresses you can get from @code{malloc} starts far away
2737 @node System Functions
2738 @section Calling System Functions
2739 @cindex library functions, and portability
2740 @cindex portability, and library functions
2742 C implementations differ substantially. Standard C reduces but does
2743 not eliminate the incompatibilities; meanwhile, many GNU packages still
2744 support pre-standard compilers because this is not hard to do. This
2745 chapter gives recommendations for how to use the more-or-less standard C
2746 library functions to avoid unnecessary loss of portability.
2750 Don't use the return value of @code{sprintf}. It returns the number of
2751 characters written on some systems, but not on all systems.
2754 Be aware that @code{vfprintf} is not always available.
2757 @code{main} should be declared to return type @code{int}. It should
2758 terminate either by calling @code{exit} or by returning the integer
2759 status code; make sure it cannot ever return an undefined value.
2761 @cindex declaration for system functions
2763 Don't declare system functions explicitly.
2765 Almost any declaration for a system function is wrong on some system.
2766 To minimize conflicts, leave it to the system header files to declare
2767 system functions. If the headers don't declare a function, let it
2770 While it may seem unclean to use a function without declaring it, in
2771 practice this works fine for most system library functions on the
2772 systems where this really happens; thus, the disadvantage is only
2773 theoretical. By contrast, actual declarations have frequently caused
2777 If you must declare a system function, don't specify the argument types.
2778 Use an old-style declaration, not a Standard C prototype. The more you
2779 specify about the function, the more likely a conflict.
2782 In particular, don't unconditionally declare @code{malloc} or
2785 Most GNU programs use those functions just once, in functions
2786 conventionally named @code{xmalloc} and @code{xrealloc}. These
2787 functions call @code{malloc} and @code{realloc}, respectively, and
2790 Because @code{xmalloc} and @code{xrealloc} are defined in your program,
2791 you can declare them in other files without any risk of type conflict.
2793 On most systems, @code{int} is the same length as a pointer; thus, the
2794 calls to @code{malloc} and @code{realloc} work fine. For the few
2795 exceptional systems (mostly 64-bit machines), you can use
2796 @strong{conditionalized} declarations of @code{malloc} and
2797 @code{realloc}---or put these declarations in configuration files
2798 specific to those systems.
2800 @cindex string library functions
2802 The string functions require special treatment. Some Unix systems have
2803 a header file @file{string.h}; others have @file{strings.h}. Neither
2804 file name is portable. There are two things you can do: use Autoconf to
2805 figure out which file to include, or don't include either file.
2808 If you don't include either strings file, you can't get declarations for
2809 the string functions from the header file in the usual way.
2811 That causes less of a problem than you might think. The newer standard
2812 string functions should be avoided anyway because many systems still
2813 don't support them. The string functions you can use are these:
2816 strcpy strncpy strcat strncat
2817 strlen strcmp strncmp
2821 The copy and concatenate functions work fine without a declaration as
2822 long as you don't use their values. Using their values without a
2823 declaration fails on systems where the width of a pointer differs from
2824 the width of @code{int}, and perhaps in other cases. It is trivial to
2825 avoid using their values, so do that.
2827 The compare functions and @code{strlen} work fine without a declaration
2828 on most systems, possibly all the ones that GNU software runs on.
2829 You may find it necessary to declare them @strong{conditionally} on a
2832 The search functions must be declared to return @code{char *}. Luckily,
2833 there is no variation in the data type they return. But there is
2834 variation in their names. Some systems give these functions the names
2835 @code{index} and @code{rindex}; other systems use the names
2836 @code{strchr} and @code{strrchr}. Some systems support both pairs of
2837 names, but neither pair works on all systems.
2839 You should pick a single pair of names and use it throughout your
2840 program. (Nowadays, it is better to choose @code{strchr} and
2841 @code{strrchr} for new programs, since those are the standard
2842 names.) Declare both of those names as functions returning @code{char
2843 *}. On systems which don't support those names, define them as macros
2844 in terms of the other pair. For example, here is what to put at the
2845 beginning of your file (or in a header) if you want to use the names
2846 @code{strchr} and @code{strrchr} throughout:
2850 #define strchr index
2852 #ifndef HAVE_STRRCHR
2853 #define strrchr rindex
2861 Here we assume that @code{HAVE_STRCHR} and @code{HAVE_STRRCHR} are
2862 macros defined in systems where the corresponding functions exist.
2863 One way to get them properly defined is to use Autoconf.
2865 @node Internationalization
2866 @section Internationalization
2867 @cindex internationalization
2870 GNU has a library called GNU gettext that makes it easy to translate the
2871 messages in a program into various languages. You should use this
2872 library in every program. Use English for the messages as they appear
2873 in the program, and let gettext provide the way to translate them into
2876 Using GNU gettext involves putting a call to the @code{gettext} macro
2877 around each string that might need translation---like this:
2880 printf (gettext ("Processing file `%s'..."));
2884 This permits GNU gettext to replace the string @code{"Processing file
2885 `%s'..."} with a translated version.
2887 Once a program uses gettext, please make a point of writing calls to
2888 @code{gettext} when you add new strings that call for translation.
2890 Using GNU gettext in a package involves specifying a @dfn{text domain
2891 name} for the package. The text domain name is used to separate the
2892 translations for this package from the translations for other packages.
2893 Normally, the text domain name should be the same as the name of the
2894 package---for example, @samp{fileutils} for the GNU file utilities.
2896 @cindex message text, and internationalization
2897 To enable gettext to work well, avoid writing code that makes
2898 assumptions about the structure of words or sentences. When you want
2899 the precise text of a sentence to vary depending on the data, use two or
2900 more alternative string constants each containing a complete sentences,
2901 rather than inserting conditionalized words or phrases into a single
2904 Here is an example of what not to do:
2907 printf ("%d file%s processed", nfiles,
2908 nfiles != 1 ? "s" : "");
2912 The problem with that example is that it assumes that plurals are made
2913 by adding `s'. If you apply gettext to the format string, like this,
2916 printf (gettext ("%d file%s processed"), nfiles,
2917 nfiles != 1 ? "s" : "");
2921 the message can use different words, but it will still be forced to use
2922 `s' for the plural. Here is a better way:
2925 printf ((nfiles != 1 ? "%d files processed"
2926 : "%d file processed"),
2931 This way, you can apply gettext to each of the two strings
2935 printf ((nfiles != 1 ? gettext ("%d files processed")
2936 : gettext ("%d file processed")),
2941 This can be any method of forming the plural of the word for ``file'', and
2942 also handles languages that require agreement in the word for
2945 A similar problem appears at the level of sentence structure with this
2949 printf ("# Implicit rule search has%s been done.\n",
2950 f->tried_implicit ? "" : " not");
2954 Adding @code{gettext} calls to this code cannot give correct results for
2955 all languages, because negation in some languages requires adding words
2956 at more than one place in the sentence. By contrast, adding
2957 @code{gettext} calls does the job straightfowardly if the code starts
2961 printf (f->tried_implicit
2962 ? "# Implicit rule search has been done.\n",
2963 : "# Implicit rule search has not been done.\n");
2970 Don't assume that @code{mmap} either works on all files or fails
2971 for all files. It may work on some files and fail on others.
2973 The proper way to use @code{mmap} is to try it on the specific file for
2974 which you want to use it---and if @code{mmap} doesn't work, fall back on
2975 doing the job in another way using @code{read} and @code{write}.
2977 The reason this precaution is needed is that the GNU kernel (the HURD)
2978 provides a user-extensible file system, in which there can be many
2979 different kinds of ``ordinary files.'' Many of them support
2980 @code{mmap}, but some do not. It is important to make programs handle
2981 all these kinds of files.
2984 @chapter Documenting Programs
2985 @cindex documentation
2987 A GNU program should ideally come with full free documentation, adequate
2988 for both reference and tutorial purposes. If the package can be
2989 programmed or extended, the documentation should cover programming or
2990 extending it, as well as just using it.
2993 * GNU Manuals:: Writing proper manuals.
2994 * Doc Strings and Manuals:: Compiling doc strings doesn't make a manual.
2995 * Manual Structure Details:: Specific structure conventions.
2996 * License for Manuals:: Writing the distribution terms for a manual.
2997 * Manual Credits:: Giving credit to documentation contributors.
2998 * Printed Manuals:: Mentioning the printed manual.
2999 * NEWS File:: NEWS files supplement manuals.
3000 * Change Logs:: Recording Changes
3001 * Man Pages:: Man pages are secondary.
3002 * Reading other Manuals:: How far you can go in learning
3007 @section GNU Manuals
3009 The preferred document format for the GNU system is the Texinfo
3010 formatting language. Every GNU package should (ideally) have
3011 documentation in Texinfo both for reference and for learners. Texinfo
3012 makes it possible to produce a good quality formatted book, using
3013 @TeX{}, and to generate an Info file. It is also possible to generate
3014 HTML output from Texinfo source. See the Texinfo manual, either the
3015 hardcopy, or the on-line version available through @code{info} or the
3016 Emacs Info subsystem (@kbd{C-h i}).
3018 Nowadays some other formats such as Docbook and Sgmltexi can be
3019 converted automatically into Texinfo. It is ok to produce the Texinfo
3020 documentation by conversion this way, as long as it gives good results.
3022 Programmers often find it most natural to structure the documentation
3023 following the structure of the implementation, which they know. But
3024 this structure is not necessarily good for explaining how to use the
3025 program; it may be irrelevant and confusing for a user.
3027 At every level, from the sentences in a paragraph to the grouping of
3028 topics into separate manuals, the right way to structure documentation
3029 is according to the concepts and questions that a user will have in mind
3030 when reading it. Sometimes this structure of ideas matches the
3031 structure of the implementation of the software being documented---but
3032 often they are different. Often the most important part of learning to
3033 write good documentation is learning to notice when you are structuring
3034 the documentation like the implementation, and think about better
3037 For example, each program in the GNU system probably ought to be
3038 documented in one manual; but this does not mean each program should
3039 have its own manual. That would be following the structure of the
3040 implementation, rather than the structure that helps the user
3043 Instead, each manual should cover a coherent @emph{topic}. For example,
3044 instead of a manual for @code{diff} and a manual for @code{diff3}, we
3045 have one manual for ``comparison of files'' which covers both of those
3046 programs, as well as @code{cmp}. By documenting these programs
3047 together, we can make the whole subject clearer.
3049 The manual which discusses a program should certainly document all of
3050 the program's command-line options and all of its commands. It should
3051 give examples of their use. But don't organize the manual as a list
3052 of features. Instead, organize it logically, by subtopics. Address
3053 the questions that a user will ask when thinking about the job that
3054 the program does. Don't just tell the reader what each feature can
3055 do---say what jobs it is good for, and show how to use it for those
3056 jobs. Explain what is recommended usage, and what kinds of usage
3059 In general, a GNU manual should serve both as tutorial and reference.
3060 It should be set up for convenient access to each topic through Info,
3061 and for reading straight through (appendixes aside). A GNU manual
3062 should give a good introduction to a beginner reading through from the
3063 start, and should also provide all the details that hackers want.
3064 The Bison manual is a good example of this---please take a look at it
3065 to see what we mean.
3067 That is not as hard as it first sounds. Arrange each chapter as a
3068 logical breakdown of its topic, but order the sections, and write their
3069 text, so that reading the chapter straight through makes sense. Do
3070 likewise when structuring the book into chapters, and when structuring a
3071 section into paragraphs. The watchword is, @emph{at each point, address
3072 the most fundamental and important issue raised by the preceding text.}
3074 If necessary, add extra chapters at the beginning of the manual which
3075 are purely tutorial and cover the basics of the subject. These provide
3076 the framework for a beginner to understand the rest of the manual. The
3077 Bison manual provides a good example of how to do this.
3079 To serve as a reference, a manual should have an Index that list all the
3080 functions, variables, options, and important concepts that are part of
3081 the program. One combined Index should do for a short manual, but
3082 sometimes for a complex package it is better to use multiple indices.
3083 The Texinfo manual includes advice on preparing good index entries, see
3084 @ref{Index Entries, , Making Index Entries, texinfo, The GNU Texinfo
3085 Manual}, and see @ref{Indexing Commands, , Defining the Entries of an
3086 Index, texinfo, The GNU Texinfo manual}.
3088 Don't use Unix man pages as a model for how to write GNU documentation;
3089 most of them are terse, badly structured, and give inadequate
3090 explanation of the underlying concepts. (There are, of course, some
3091 exceptions.) Also, Unix man pages use a particular format which is
3092 different from what we use in GNU manuals.
3094 Please include an email address in the manual for where to report
3095 bugs @emph{in the text of the manual}.
3097 Please do not use the term ``pathname'' that is used in Unix
3098 documentation; use ``file name'' (two words) instead. We use the term
3099 ``path'' only for search paths, which are lists of directory names.
3101 Please do not use the term ``illegal'' to refer to erroneous input to
3102 a computer program. Please use ``invalid'' for this, and reserve the
3103 term ``illegal'' for activities prohibited by law.
3105 @node Doc Strings and Manuals
3106 @section Doc Strings and Manuals
3108 Some programming systems, such as Emacs, provide a documentation string
3109 for each function, command or variable. You may be tempted to write a
3110 reference manual by compiling the documentation strings and writing a
3111 little additional text to go around them---but you must not do it. That
3112 approach is a fundamental mistake. The text of well-written
3113 documentation strings will be entirely wrong for a manual.
3115 A documentation string needs to stand alone---when it appears on the
3116 screen, there will be no other text to introduce or explain it.
3117 Meanwhile, it can be rather informal in style.
3119 The text describing a function or variable in a manual must not stand
3120 alone; it appears in the context of a section or subsection. Other text
3121 at the beginning of the section should explain some of the concepts, and
3122 should often make some general points that apply to several functions or
3123 variables. The previous descriptions of functions and variables in the
3124 section will also have given information about the topic. A description
3125 written to stand alone would repeat some of that information; this
3126 redundance looks bad. Meanwhile, the informality that is acceptable in
3127 a documentation string is totally unacceptable in a manual.
3129 The only good way to use documentation strings in writing a good manual
3130 is to use them as a source of information for writing good text.
3132 @node Manual Structure Details
3133 @section Manual Structure Details
3134 @cindex manual structure
3136 The title page of the manual should state the version of the programs or
3137 packages documented in the manual. The Top node of the manual should
3138 also contain this information. If the manual is changing more
3139 frequently than or independent of the program, also state a version
3140 number for the manual in both of these places.
3142 Each program documented in the manual should have a node named
3143 @samp{@var{program} Invocation} or @samp{Invoking @var{program}}. This
3144 node (together with its subnodes, if any) should describe the program's
3145 command line arguments and how to run it (the sort of information people
3146 would look in a man page for). Start with an @samp{@@example}
3147 containing a template for all the options and arguments that the program
3150 Alternatively, put a menu item in some menu whose item name fits one of
3151 the above patterns. This identifies the node which that item points to
3152 as the node for this purpose, regardless of the node's actual name.
3154 The @samp{--usage} feature of the Info reader looks for such a node
3155 or menu item in order to find the relevant text, so it is essential
3156 for every Texinfo file to have one.
3158 If one manual describes several programs, it should have such a node for
3159 each program described in the manual.
3161 @node License for Manuals
3162 @section License for Manuals
3163 @cindex license for manuals
3165 Please use the GNU Free Documentation License for all GNU manuals that
3166 are more than a few pages long. Likewise for a collection of short
3167 documents---you only need one copy of the GNU FDL for the whole
3168 collection. For a single short document, you can use a very permissive
3169 non-copyleft license, to avoid taking up space with a long license.
3171 See @uref{http://www.gnu.org/copyleft/fdl-howto.html} for more explanation
3172 of how to employ the GFDL.
3174 Note that it is not obligatory to include a copy of the GNU GPL or GNU
3175 LGPL in a manual whose license is neither the GPL nor the LGPL. It can
3176 be a good idea to include the program's license in a large manual; in a
3177 short manual, whose size would be increased considerably by including
3178 the program's license, it is probably better not to include it.
3180 @node Manual Credits
3181 @section Manual Credits
3182 @cindex credits for manuals
3184 Please credit the principal human writers of the manual as the authors,
3185 on the title page of the manual. If a company sponsored the work, thank
3186 the company in a suitable place in the manual, but do not cite the
3187 company as an author.
3189 @node Printed Manuals
3190 @section Printed Manuals
3192 The FSF publishes some GNU manuals in printed form. To encourage sales
3193 of these manuals, the on-line versions of the manual should mention at
3194 the very start that the printed manual is available and should point at
3195 information for getting it---for instance, with a link to the page
3196 @url{http://www.gnu.org/order/order.html}. This should not be included
3197 in the printed manual, though, because there it is redundant.
3199 It is also useful to explain in the on-line forms of the manual how the
3200 user can print out the manual from the sources.
3203 @section The NEWS File
3204 @cindex @file{NEWS} file
3206 In addition to its manual, the package should have a file named
3207 @file{NEWS} which contains a list of user-visible changes worth
3208 mentioning. In each new release, add items to the front of the file and
3209 identify the version they pertain to. Don't discard old items; leave
3210 them in the file after the newer items. This way, a user upgrading from
3211 any previous version can see what is new.
3213 If the @file{NEWS} file gets very long, move some of the older items
3214 into a file named @file{ONEWS} and put a note at the end referring the
3218 @section Change Logs
3221 Keep a change log to describe all the changes made to program source
3222 files. The purpose of this is so that people investigating bugs in the
3223 future will know about the changes that might have introduced the bug.
3224 Often a new bug can be found by looking at what was recently changed.
3225 More importantly, change logs can help you eliminate conceptual
3226 inconsistencies between different parts of a program, by giving you a
3227 history of how the conflicting concepts arose and who they came from.
3230 * Change Log Concepts::
3231 * Style of Change Logs::
3233 * Conditional Changes::
3234 * Indicating the Part Changed::
3237 @node Change Log Concepts
3238 @subsection Change Log Concepts
3240 You can think of the change log as a conceptual ``undo list'' which
3241 explains how earlier versions were different from the current version.
3242 People can see the current version; they don't need the change log
3243 to tell them what is in it. What they want from a change log is a
3244 clear explanation of how the earlier version differed.
3246 The change log file is normally called @file{ChangeLog} and covers an
3247 entire directory. Each directory can have its own change log, or a
3248 directory can use the change log of its parent directory--it's up to
3251 Another alternative is to record change log information with a version
3252 control system such as RCS or CVS. This can be converted automatically
3253 to a @file{ChangeLog} file using @code{rcs2log}; in Emacs, the command
3254 @kbd{C-x v a} (@code{vc-update-change-log}) does the job.
3256 There's no need to describe the full purpose of the changes or how they
3257 work together. If you think that a change calls for explanation, you're
3258 probably right. Please do explain it---but please put the explanation
3259 in comments in the code, where people will see it whenever they see the
3260 code. For example, ``New function'' is enough for the change log when
3261 you add a function, because there should be a comment before the
3262 function definition to explain what it does.
3264 In the past, we recommended not mentioning changes in non-software
3265 files (manuals, help files, etc.) in change logs. However, we've been
3266 advised that it is a good idea to include them, for the sake of
3269 However, sometimes it is useful to write one line to describe the
3270 overall purpose of a batch of changes.
3272 The easiest way to add an entry to @file{ChangeLog} is with the Emacs
3273 command @kbd{M-x add-change-log-entry}. An entry should have an
3274 asterisk, the name of the changed file, and then in parentheses the name
3275 of the changed functions, variables or whatever, followed by a colon.
3276 Then describe the changes you made to that function or variable.
3278 @node Style of Change Logs
3279 @subsection Style of Change Logs
3280 @cindex change logs, style
3282 Here are some simple examples of change log entries, starting with the
3283 header line that says who made the change and when it was installed,
3284 followed by descriptions of specific changes. (These examples are
3285 drawn from Emacs and GCC.)
3288 1998-08-17 Richard Stallman <rms@@gnu.org>
3290 * register.el (insert-register): Return nil.
3291 (jump-to-register): Likewise.
3293 * sort.el (sort-subr): Return nil.
3295 * tex-mode.el (tex-bibtex-file, tex-file, tex-region):
3296 Restart the tex shell if process is gone or stopped.
3297 (tex-shell-running): New function.
3299 * expr.c (store_one_arg): Round size up for move_block_to_reg.
3300 (expand_call): Round up when emitting USE insns.
3301 * stmt.c (assign_parms): Round size up for move_block_from_reg.
3304 It's important to name the changed function or variable in full. Don't
3305 abbreviate function or variable names, and don't combine them.
3306 Subsequent maintainers will often search for a function name to find all
3307 the change log entries that pertain to it; if you abbreviate the name,
3308 they won't find it when they search.
3310 For example, some people are tempted to abbreviate groups of function
3311 names by writing @samp{* register.el (@{insert,jump-to@}-register)};
3312 this is not a good idea, since searching for @code{jump-to-register} or
3313 @code{insert-register} would not find that entry.
3315 Separate unrelated change log entries with blank lines. When two
3316 entries represent parts of the same change, so that they work together,
3317 then don't put blank lines between them. Then you can omit the file
3318 name and the asterisk when successive entries are in the same file.
3320 Break long lists of function names by closing continued lines with
3321 @samp{)}, rather than @samp{,}, and opening the continuation with
3322 @samp{(} as in this example:
3325 * keyboard.c (menu_bar_items, tool_bar_items)
3326 (Fexecute_extended_command): Deal with `keymap' property.
3329 When you install someone else's changes, put the contributor's name in
3330 the change log entry rather than in the text of the entry. In other
3334 2002-07-14 John Doe <jdoe@@gnu.org>
3336 * sewing.c: Make it sew.
3343 2002-07-14 Usual Maintainer <usual@@gnu.org>
3345 * sewing.c: Make it sew. Patch by jdoe@@gnu.org.
3348 As for the date, that should be the date you applied the change.
3350 @node Simple Changes
3351 @subsection Simple Changes
3353 Certain simple kinds of changes don't need much detail in the change
3356 When you change the calling sequence of a function in a simple fashion,
3357 and you change all the callers of the function to use the new calling
3358 sequence, there is no need to make individual entries for all the
3359 callers that you changed. Just write in the entry for the function
3360 being called, ``All callers changed''---like this:
3363 * keyboard.c (Fcommand_execute): New arg SPECIAL.
3364 All callers changed.
3367 When you change just comments or doc strings, it is enough to write an
3368 entry for the file, without mentioning the functions. Just ``Doc
3369 fixes'' is enough for the change log.
3371 There's no technical need to make change log entries for documentation
3372 files. This is because documentation is not susceptible to bugs that
3373 are hard to fix. Documentation does not consist of parts that must
3374 interact in a precisely engineered fashion. To correct an error, you
3375 need not know the history of the erroneous passage; it is enough to
3376 compare what the documentation says with the way the program actually
3379 However, you should keep change logs for documentation files when the
3380 project gets copyright assignments from its contributors, so as to
3381 make the records of authorship more accurate.
3383 @node Conditional Changes
3384 @subsection Conditional Changes
3385 @cindex conditional changes, and change logs
3386 @cindex change logs, conditional changes
3388 C programs often contain compile-time @code{#if} conditionals. Many
3389 changes are conditional; sometimes you add a new definition which is
3390 entirely contained in a conditional. It is very useful to indicate in
3391 the change log the conditions for which the change applies.
3393 Our convention for indicating conditional changes is to use square
3394 brackets around the name of the condition.
3396 Here is a simple example, describing a change which is conditional but
3397 does not have a function or entity name associated with it:
3400 * xterm.c [SOLARIS2]: Include string.h.
3403 Here is an entry describing a new definition which is entirely
3404 conditional. This new definition for the macro @code{FRAME_WINDOW_P} is
3405 used only when @code{HAVE_X_WINDOWS} is defined:
3408 * frame.h [HAVE_X_WINDOWS] (FRAME_WINDOW_P): Macro defined.
3411 Here is an entry for a change within the function @code{init_display},
3412 whose definition as a whole is unconditional, but the changes themselves
3413 are contained in a @samp{#ifdef HAVE_LIBNCURSES} conditional:
3416 * dispnew.c (init_display) [HAVE_LIBNCURSES]: If X, call tgetent.
3419 Here is an entry for a change that takes affect only when
3420 a certain macro is @emph{not} defined:
3423 (gethostname) [!HAVE_SOCKETS]: Replace with winsock version.
3426 @node Indicating the Part Changed
3427 @subsection Indicating the Part Changed
3429 Indicate the part of a function which changed by using angle brackets
3430 enclosing an indication of what the changed part does. Here is an entry
3431 for a change in the part of the function @code{sh-while-getopts} that
3432 deals with @code{sh} commands:
3435 * progmodes/sh-script.el (sh-while-getopts) <sh>: Handle case that
3436 user-specified option string is empty.
3444 In the GNU project, man pages are secondary. It is not necessary or
3445 expected for every GNU program to have a man page, but some of them do.
3446 It's your choice whether to include a man page in your program.
3448 When you make this decision, consider that supporting a man page
3449 requires continual effort each time the program is changed. The time
3450 you spend on the man page is time taken away from more useful work.
3452 For a simple program which changes little, updating the man page may be
3453 a small job. Then there is little reason not to include a man page, if
3456 For a large program that changes a great deal, updating a man page may
3457 be a substantial burden. If a user offers to donate a man page, you may
3458 find this gift costly to accept. It may be better to refuse the man
3459 page unless the same person agrees to take full responsibility for
3460 maintaining it---so that you can wash your hands of it entirely. If
3461 this volunteer later ceases to do the job, then don't feel obliged to
3462 pick it up yourself; it may be better to withdraw the man page from the
3463 distribution until someone else agrees to update it.
3465 When a program changes only a little, you may feel that the
3466 discrepancies are small enough that the man page remains useful without
3467 updating. If so, put a prominent note near the beginning of the man
3468 page explaining that you don't maintain it and that the Texinfo manual
3469 is more authoritative. The note should say how to access the Texinfo
3472 @node Reading other Manuals
3473 @section Reading other Manuals
3475 There may be non-free books or documentation files that describe the
3476 program you are documenting.
3478 It is ok to use these documents for reference, just as the author of a
3479 new algebra textbook can read other books on algebra. A large portion
3480 of any non-fiction book consists of facts, in this case facts about how
3481 a certain program works, and these facts are necessarily the same for
3482 everyone who writes about the subject. But be careful not to copy your
3483 outline structure, wording, tables or examples from preexisting non-free
3484 documentation. Copying from free documentation may be ok; please check
3485 with the FSF about the individual case.
3487 @node Managing Releases
3488 @chapter The Release Process
3491 Making a release is more than just bundling up your source files in a
3492 tar file and putting it up for FTP. You should set up your software so
3493 that it can be configured to run on a variety of systems. Your Makefile
3494 should conform to the GNU standards described below, and your directory
3495 layout should also conform to the standards discussed below. Doing so
3496 makes it easy to include your package into the larger framework of
3500 * Configuration:: How Configuration Should Work
3501 * Makefile Conventions:: Makefile Conventions
3502 * Releases:: Making Releases
3506 @section How Configuration Should Work
3507 @cindex program configuration
3510 Each GNU distribution should come with a shell script named
3511 @code{configure}. This script is given arguments which describe the
3512 kind of machine and system you want to compile the program for.
3514 The @code{configure} script must record the configuration options so
3515 that they affect compilation.
3517 One way to do this is to make a link from a standard name such as
3518 @file{config.h} to the proper configuration file for the chosen system.
3519 If you use this technique, the distribution should @emph{not} contain a
3520 file named @file{config.h}. This is so that people won't be able to
3521 build the program without configuring it first.
3523 Another thing that @code{configure} can do is to edit the Makefile. If
3524 you do this, the distribution should @emph{not} contain a file named
3525 @file{Makefile}. Instead, it should include a file @file{Makefile.in} which
3526 contains the input used for editing. Once again, this is so that people
3527 won't be able to build the program without configuring it first.
3529 If @code{configure} does write the @file{Makefile}, then @file{Makefile}
3530 should have a target named @file{Makefile} which causes @code{configure}
3531 to be rerun, setting up the same configuration that was set up last
3532 time. The files that @code{configure} reads should be listed as
3533 dependencies of @file{Makefile}.
3535 All the files which are output from the @code{configure} script should
3536 have comments at the beginning explaining that they were generated
3537 automatically using @code{configure}. This is so that users won't think
3538 of trying to edit them by hand.
3540 The @code{configure} script should write a file named @file{config.status}
3541 which describes which configuration options were specified when the
3542 program was last configured. This file should be a shell script which,
3543 if run, will recreate the same configuration.
3545 The @code{configure} script should accept an option of the form
3546 @samp{--srcdir=@var{dirname}} to specify the directory where sources are found
3547 (if it is not the current directory). This makes it possible to build
3548 the program in a separate directory, so that the actual source directory
3551 If the user does not specify @samp{--srcdir}, then @code{configure} should
3552 check both @file{.} and @file{..} to see if it can find the sources. If
3553 it finds the sources in one of these places, it should use them from
3554 there. Otherwise, it should report that it cannot find the sources, and
3555 should exit with nonzero status.
3557 Usually the easy way to support @samp{--srcdir} is by editing a
3558 definition of @code{VPATH} into the Makefile. Some rules may need to
3559 refer explicitly to the specified source directory. To make this
3560 possible, @code{configure} can add to the Makefile a variable named
3561 @code{srcdir} whose value is precisely the specified directory.
3563 The @code{configure} script should also take an argument which specifies the
3564 type of system to build the program for. This argument should look like
3568 @var{cpu}-@var{company}-@var{system}
3571 For example, an Athlon-based GNU/Linux system might be
3572 @samp{i686-pc-linux-gnu}.
3574 The @code{configure} script needs to be able to decode all plausible
3575 alternatives for how to describe a machine. Thus,
3576 @samp{athlon-pc-gnu/linux} would be a valid alias.
3577 There is a shell script called
3578 @uref{ftp://ftp.gnu.org/gnu/config/config.sub, @file{config.sub}}
3580 as a subroutine to validate system types and canonicalize aliases.
3582 The @code{configure} script should also take the option
3583 @option{--build=@var{buildtype}}, which should be equivalent to a
3584 plain @var{buildtype} argument. For example, @samp{configure
3585 --build=i686-pc-linux-gnu} is equivalent to @samp{configure
3586 i686-pc-linux-gnu}. When the build type is not specified by an option
3587 or argument, the @code{configure} script should normally guess it
3588 using the shell script
3589 @uref{ftp://ftp.gnu.org/gnu/config/config.guess, @file{config.guess}}.
3591 @cindex optional features, configure-time
3592 Other options are permitted to specify in more detail the software
3593 or hardware present on the machine, and include or exclude optional
3594 parts of the package:
3597 @item --enable-@var{feature}@r{[}=@var{parameter}@r{]}
3598 Configure the package to build and install an optional user-level
3599 facility called @var{feature}. This allows users to choose which
3600 optional features to include. Giving an optional @var{parameter} of
3601 @samp{no} should omit @var{feature}, if it is built by default.
3603 No @samp{--enable} option should @strong{ever} cause one feature to
3604 replace another. No @samp{--enable} option should ever substitute one
3605 useful behavior for another useful behavior. The only proper use for
3606 @samp{--enable} is for questions of whether to build part of the program
3609 @item --with-@var{package}
3610 @c @r{[}=@var{parameter}@r{]}
3611 The package @var{package} will be installed, so configure this package
3612 to work with @var{package}.
3614 @c Giving an optional @var{parameter} of
3615 @c @samp{no} should omit @var{package}, if it is used by default.
3617 Possible values of @var{package} include
3618 @samp{gnu-as} (or @samp{gas}), @samp{gnu-ld}, @samp{gnu-libc},
3624 Do not use a @samp{--with} option to specify the file name to use to
3625 find certain files. That is outside the scope of what @samp{--with}
3629 All @code{configure} scripts should accept all of these ``detail''
3630 options, whether or not they make any difference to the particular
3631 package at hand. In particular, they should accept any option that
3632 starts with @samp{--with-} or @samp{--enable-}. This is so users will
3633 be able to configure an entire GNU source tree at once with a single set
3636 You will note that the categories @samp{--with-} and @samp{--enable-}
3637 are narrow: they @strong{do not} provide a place for any sort of option
3638 you might think of. That is deliberate. We want to limit the possible
3639 configuration options in GNU software. We do not want GNU programs to
3640 have idiosyncratic configuration options.
3642 Packages that perform part of the compilation process may support
3643 cross-compilation. In such a case, the host and target machines for the
3644 program may be different.
3646 The @code{configure} script should normally treat the specified type of
3647 system as both the host and the target, thus producing a program which
3648 works for the same type of machine that it runs on.
3650 To compile a program to run on a host type that differs from the build
3651 type, use the configure option @option{--host=@var{hosttype}}, where
3652 @var{hosttype} uses the same syntax as @var{buildtype}. The host type
3653 normally defaults to the build type.
3655 To configure a cross-compiler, cross-assembler, or what have you, you
3656 should specify a target different from the host, using the configure
3657 option @samp{--target=@var{targettype}}. The syntax for
3658 @var{targettype} is the same as for the host type. So the command would
3662 ./configure --host=@var{hosttype} --target=@var{targettype}
3665 The target type normally defaults to the host type.
3666 Programs for which cross-operation is not meaningful need not accept the
3667 @samp{--target} option, because configuring an entire operating system for
3668 cross-operation is not a meaningful operation.
3670 Some programs have ways of configuring themselves automatically. If
3671 your program is set up to do this, your @code{configure} script can simply
3672 ignore most of its arguments.
3674 @comment The makefile standards are in a separate file that is also
3675 @comment included by make.texinfo. Done by roland@gnu.ai.mit.edu on 1/6/93.
3676 @comment For this document, turn chapters into sections, etc.
3678 @include make-stds.texi
3682 @section Making Releases
3685 You should identify each release with a pair of version numbers, a
3686 major version and a minor. We have no objection to using more than
3687 two numbers, but it is very unlikely that you really need them.
3689 Package the distribution of @code{Foo version 69.96} up in a gzipped tar
3690 file with the name @file{foo-69.96.tar.gz}. It should unpack into a
3691 subdirectory named @file{foo-69.96}.
3693 Building and installing the program should never modify any of the files
3694 contained in the distribution. This means that all the files that form
3695 part of the program in any way must be classified into @dfn{source
3696 files} and @dfn{non-source files}. Source files are written by humans
3697 and never changed automatically; non-source files are produced from
3698 source files by programs under the control of the Makefile.
3700 @cindex @file{README} file
3701 The distribution should contain a file named @file{README} which gives
3702 the name of the package, and a general description of what it does. It
3703 is also good to explain the purpose of each of the first-level
3704 subdirectories in the package, if there are any. The @file{README} file
3705 should either state the version number of the package, or refer to where
3706 in the package it can be found.
3708 The @file{README} file should refer to the file @file{INSTALL}, which
3709 should contain an explanation of the installation procedure.
3711 The @file{README} file should also refer to the file which contains the
3712 copying conditions. The GNU GPL, if used, should be in a file called
3713 @file{COPYING}. If the GNU LGPL is used, it should be in a file called
3716 Naturally, all the source files must be in the distribution. It is okay
3717 to include non-source files in the distribution, provided they are
3718 up-to-date and machine-independent, so that building the distribution
3719 normally will never modify them. We commonly include non-source files
3720 produced by Bison, @code{lex}, @TeX{}, and @code{makeinfo}; this helps avoid
3721 unnecessary dependencies between our distributions, so that users can
3722 install whichever packages they want to install.
3724 Non-source files that might actually be modified by building and
3725 installing the program should @strong{never} be included in the
3726 distribution. So if you do distribute non-source files, always make
3727 sure they are up to date when you make a new distribution.
3729 Make sure that the directory into which the distribution unpacks (as
3730 well as any subdirectories) are all world-writable (octal mode 777).
3731 This is so that old versions of @code{tar} which preserve the
3732 ownership and permissions of the files from the tar archive will be
3733 able to extract all the files even if the user is unprivileged.
3735 Make sure that all the files in the distribution are world-readable.
3737 Don't include any symbolic links in the distribution itself. If the tar
3738 file contains symbolic links, then people cannot even unpack it on
3739 systems that don't support symbolic links. Also, don't use multiple
3740 names for one file in different directories, because certain file
3741 systems cannot handle this and that prevents unpacking the
3744 Try to make sure that all the file names will be unique on MS-DOS. A
3745 name on MS-DOS consists of up to 8 characters, optionally followed by a
3746 period and up to three characters. MS-DOS will truncate extra
3747 characters both before and after the period. Thus,
3748 @file{foobarhacker.c} and @file{foobarhacker.o} are not ambiguous; they
3749 are truncated to @file{foobarha.c} and @file{foobarha.o}, which are
3752 @cindex @file{texinfo.tex}, in a distribution
3753 Include in your distribution a copy of the @file{texinfo.tex} you used
3754 to test print any @file{*.texinfo} or @file{*.texi} files.
3756 Likewise, if your program uses small GNU software packages like regex,
3757 getopt, obstack, or termcap, include them in the distribution file.
3758 Leaving them out would make the distribution file a little smaller at
3759 the expense of possible inconvenience to a user who doesn't know what
3763 @chapter References to Non-Free Software and Documentation
3764 @cindex references to non-free material
3766 A GNU program should not recommend use of any non-free program. We
3767 can't stop some people from writing proprietary programs, or stop
3768 other people from using them, but we can and should refuse to
3769 advertise them to new potential customers. Proprietary software is a
3770 social and ethical problem, and the point of GNU is to solve that
3773 The GNU definition of free software is found in
3774 @url{http://www.gnu.org/philosophy/free-sw.html}, with a list of
3775 important licenses and whether they qualify as free in
3776 @url{http://www.gnu.org/licenses/license-list.html}. The terms
3777 ``free'' and ``non-free'', used in this document, refer to that
3778 definition. If it is not clear whether a license qualifies as free
3779 under this definition, please ask the GNU Project by writing to
3780 @email{licensing@@gnu.org}. We will answer, and if the license is an
3781 important one, we will add it to the list.
3783 When a non-free program or system is well known, you can mention it in
3784 passing---that is harmless, since users who might want to use it
3785 probably already know about it. For instance, it is fine to explain
3786 how to build your package on top of some widely used non-free
3787 operating system, or how to use it together with some widely used
3790 However, you should give only the necessary information to help those
3791 who already use the non-free program to use your program with
3792 it---don't give, or refer to, any further information about the
3793 proprietary program, and don't imply that the proprietary program
3794 enhances your program, or that its existence is in any way a good
3795 thing. The goal should be that people already using the proprietary
3796 program will get the advice they need about how to use your free
3797 program with it, while people who don't already use the proprietary
3798 program will not see anything to lead them to take an interest in it.
3800 If a non-free program or system is obscure in your program's domain,
3801 your program should not mention or support it at all, since doing so
3802 would tend to popularize the non-free program more than it popularizes
3803 your program. (You cannot hope to find many additional users among
3804 the users of Foobar if the users of Foobar are few.)
3806 Sometimes a program is free software in itself but depends on a
3807 non-free platform in order to run. For instance, many Java programs
3808 depend on Sun's Java implementation, and won't run on the GNU Java
3809 Compiler (which does not yet have all the features) or won't run with
3810 the GNU Java libraries. To recommend that program is inherently to
3811 recommend the non-free platform as well; if you should not do the
3812 latter, then don't do the former.
3814 A GNU package should not refer the user to any non-free documentation
3815 for free software. Free documentation that can be included in free
3816 operating systems is essential for completing the GNU system, or any
3817 free operating system, so it is a major focus of the GNU Project; to
3818 recommend use of documentation that we are not allowed to use in GNU
3819 would weaken the impetus for the community to produce documentation
3820 that we can include. So GNU packages should never recommend non-free
3823 By contrast, it is ok to refer to journal articles and textbooks in
3824 the comments of a program for explanation of how it functions, even
3825 though they be non-free. This is because we don't include such things
3826 in the GNU system even if we are allowed to--they are outside the
3827 scope of an operating system project.
3829 Referring to a web site that describes or recommends a non-free
3830 program is in effect promoting that software, so please do not make
3831 links (or mention by name) web sites that contain such material. This
3832 policy is relevant particulary for the web pages for a GNU package.
3834 Following links from nearly any web site can lead to non-free
3835 software; this is an inescapable aspect of the nature of the web, and
3836 in itself is no objection to linking to a site. As long as the site
3837 does not itself recommend a non-free program, there is no need be
3838 concerned about the sites it links to for other reasons.
3840 Thus, for example, you should not make a link to AT&T's web site,
3841 because that recommends AT&T's non-free software packages; you should
3842 not make a link to a site that links to AT&T's site saying it is a
3843 place to get a non-free program; but if a site you want to link to
3844 refers to AT&T's web site in some other context (such as long-distance
3845 telephone service), that is not a problem.
3847 @node Copying This Manual
3848 @appendix Copying This Manual
3851 * GNU Free Documentation License:: License for copying this manual
3863 eval: (add-hook 'write-file-hooks 'time-stamp)
3864 time-stamp-start: "@set lastupdate "
3866 time-stamp-format: "%:b %:d, %:y"
3867 compile-command: "make just-standards"