1 @comment This file is included by both standards.texi and make.texinfo.
2 @comment It was broken out of standards.texi on 1/6/93 by roland.
4 @node Makefile Conventions
5 @chapter Makefile Conventions
6 @comment standards.texi does not print an index, but make.texinfo does.
7 @cindex makefile, conventions for
8 @cindex conventions for makefiles
9 @cindex standards for makefiles
11 @c Copyright 1992, 1993, 1994, 1995, 1996, 1997, 1998, 2000, 2001 Free
12 @c Software Foundation, Inc.
14 @c Permission is granted to copy, distribute and/or modify this document
15 @c under the terms of the GNU Free Documentation License, Version 1.1
16 @c or any later version published by the Free Software Foundation;
17 @c with no Invariant Sections, with no
18 @c Front-Cover Texts, and with no Back-Cover Texts.
19 @c A copy of the license is included in the section entitled ``GNU
20 @c Free Documentation License''.
34 describes conventions for writing the Makefiles for GNU programs.
35 Using Automake will help you write a Makefile that follows these
39 * Makefile Basics:: General Conventions for Makefiles
40 * Utilities in Makefiles:: Utilities in Makefiles
41 * Command Variables:: Variables for Specifying Commands
42 * Directory Variables:: Variables for Installation Directories
43 * Standard Targets:: Standard Targets for Users
44 * Install Command Categories:: Three categories of commands in the `install'
45 rule: normal, pre-install and post-install.
49 @section General Conventions for Makefiles
51 Every Makefile should contain this line:
58 to avoid trouble on systems where the @code{SHELL} variable might be
59 inherited from the environment. (This is never a problem with GNU
62 Different @code{make} programs have incompatible suffix lists and
63 implicit rules, and this sometimes creates confusion or misbehavior. So
64 it is a good idea to set the suffix list explicitly using only the
65 suffixes you need in the particular Makefile, like this:
73 The first line clears out the suffix list, the second introduces all
74 suffixes which may be subject to implicit rules in this Makefile.
76 Don't assume that @file{.} is in the path for command execution. When
77 you need to run programs that are a part of your package during the
78 make, please make sure that it uses @file{./} if the program is built as
79 part of the make or @file{$(srcdir)/} if the file is an unchanging part
80 of the source code. Without one of these prefixes, the current search
83 The distinction between @file{./} (the @dfn{build directory}) and
84 @file{$(srcdir)/} (the @dfn{source directory}) is important because
85 users can build in a separate directory using the @samp{--srcdir} option
86 to @file{configure}. A rule of the form:
89 foo.1 : foo.man sedscript
90 sed -e sedscript foo.man > foo.1
94 will fail when the build directory is not the source directory, because
95 @file{foo.man} and @file{sedscript} are in the source directory.
97 When using GNU @code{make}, relying on @samp{VPATH} to find the source
98 file will work in the case where there is a single dependency file,
99 since the @code{make} automatic variable @samp{$<} will represent the
100 source file wherever it is. (Many versions of @code{make} set @samp{$<}
101 only in implicit rules.) A Makefile target like
105 $(CC) -I. -I$(srcdir) $(CFLAGS) -c bar.c -o foo.o
109 should instead be written as
113 $(CC) -I. -I$(srcdir) $(CFLAGS) -c $< -o $@@
117 in order to allow @samp{VPATH} to work correctly. When the target has
118 multiple dependencies, using an explicit @samp{$(srcdir)} is the easiest
119 way to make the rule work well. For example, the target above for
120 @file{foo.1} is best written as:
123 foo.1 : foo.man sedscript
124 sed -e $(srcdir)/sedscript $(srcdir)/foo.man > $@@
127 GNU distributions usually contain some files which are not source
128 files---for example, Info files, and the output from Autoconf, Automake,
129 Bison or Flex. Since these files normally appear in the source
130 directory, they should always appear in the source directory, not in the
131 build directory. So Makefile rules to update them should put the
132 updated files in the source directory.
134 However, if a file does not appear in the distribution, then the
135 Makefile should not put it in the source directory, because building a
136 program in ordinary circumstances should not modify the source directory
139 Try to make the build and installation targets, at least (and all their
140 subtargets) work correctly with a parallel @code{make}.
142 @node Utilities in Makefiles
143 @section Utilities in Makefiles
145 Write the Makefile commands (and any shell scripts, such as
146 @code{configure}) to run in @code{sh}, not in @code{csh}. Don't use any
147 special features of @code{ksh} or @code{bash}.
149 The @code{configure} script and the Makefile rules for building and
150 installation should not use any utilities directly except these:
153 @c gunzip gzip md5sum
154 @c mkfifo mknod tee uname
157 cat cmp cp diff echo egrep expr false grep install-info
158 ln ls mkdir mv pwd rm rmdir sed sleep sort tar test touch true
161 The compression program @code{gzip} can be used in the @code{dist} rule.
163 Stick to the generally supported options for these programs. For
164 example, don't use @samp{mkdir -p}, convenient as it may be, because
165 most systems don't support it.
167 It is a good idea to avoid creating symbolic links in makefiles, since a
168 few systems don't support them.
170 The Makefile rules for building and installation can also use compilers
171 and related programs, but should do so via @code{make} variables so that the
172 user can substitute alternatives. Here are some of the programs we
176 ar bison cc flex install ld ldconfig lex
177 make makeinfo ranlib texi2dvi yacc
180 Use the following @code{make} variables to run those programs:
183 $(AR) $(BISON) $(CC) $(FLEX) $(INSTALL) $(LD) $(LDCONFIG) $(LEX)
184 $(MAKE) $(MAKEINFO) $(RANLIB) $(TEXI2DVI) $(YACC)
187 When you use @code{ranlib} or @code{ldconfig}, you should make sure
188 nothing bad happens if the system does not have the program in question.
189 Arrange to ignore an error from that command, and print a message before
190 the command to tell the user that failure of this command does not mean
191 a problem. (The Autoconf @samp{AC_PROG_RANLIB} macro can help with
194 If you use symbolic links, you should implement a fallback for systems
195 that don't have symbolic links.
197 Additional utilities that can be used via Make variables are:
200 chgrp chmod chown mknod
203 It is ok to use other utilities in Makefile portions (or scripts)
204 intended only for particular systems where you know those utilities
207 @node Command Variables
208 @section Variables for Specifying Commands
210 Makefiles should provide variables for overriding certain commands, options,
213 In particular, you should run most utility programs via variables.
214 Thus, if you use Bison, have a variable named @code{BISON} whose default
215 value is set with @samp{BISON = bison}, and refer to it with
216 @code{$(BISON)} whenever you need to use Bison.
218 File management utilities such as @code{ln}, @code{rm}, @code{mv}, and
219 so on, need not be referred to through variables in this way, since users
220 don't need to replace them with other programs.
222 Each program-name variable should come with an options variable that is
223 used to supply options to the program. Append @samp{FLAGS} to the
224 program-name variable name to get the options variable name---for
225 example, @code{BISONFLAGS}. (The names @code{CFLAGS} for the C
226 compiler, @code{YFLAGS} for yacc, and @code{LFLAGS} for lex, are
227 exceptions to this rule, but we keep them because they are standard.)
228 Use @code{CPPFLAGS} in any compilation command that runs the
229 preprocessor, and use @code{LDFLAGS} in any compilation command that
230 does linking as well as in any direct use of @code{ld}.
232 If there are C compiler options that @emph{must} be used for proper
233 compilation of certain files, do not include them in @code{CFLAGS}.
234 Users expect to be able to specify @code{CFLAGS} freely themselves.
235 Instead, arrange to pass the necessary options to the C compiler
236 independently of @code{CFLAGS}, by writing them explicitly in the
237 compilation commands or by defining an implicit rule, like this:
241 ALL_CFLAGS = -I. $(CFLAGS)
243 $(CC) -c $(CPPFLAGS) $(ALL_CFLAGS) $<
246 Do include the @samp{-g} option in @code{CFLAGS}, because that is not
247 @emph{required} for proper compilation. You can consider it a default
248 that is only recommended. If the package is set up so that it is
249 compiled with GCC by default, then you might as well include @samp{-O}
250 in the default value of @code{CFLAGS} as well.
252 Put @code{CFLAGS} last in the compilation command, after other variables
253 containing compiler options, so the user can use @code{CFLAGS} to
256 @code{CFLAGS} should be used in every invocation of the C compiler,
257 both those which do compilation and those which do linking.
259 Every Makefile should define the variable @code{INSTALL}, which is the
260 basic command for installing a file into the system.
262 Every Makefile should also define the variables @code{INSTALL_PROGRAM}
263 and @code{INSTALL_DATA}. (The default for @code{INSTALL_PROGRAM} should
264 be @code{$(INSTALL)}; the default for @code{INSTALL_DATA} should be
265 @code{$@{INSTALL@} -m 644}.) Then it should use those variables as the
266 commands for actual installation, for executables and nonexecutables
267 respectively. Use these variables as follows:
270 $(INSTALL_PROGRAM) foo $(bindir)/foo
271 $(INSTALL_DATA) libfoo.a $(libdir)/libfoo.a
274 Optionally, you may prepend the value of @code{DESTDIR} to the target
275 filename. Doing this allows the installer to create a snapshot of the
276 installation to be copied onto the real target filesystem later. Do not
277 set the value of @code{DESTDIR} in your Makefile, and do not include it
278 in any installed files. With support for @code{DESTDIR}, the above
282 $(INSTALL_PROGRAM) foo $(DESTDIR)$(bindir)/foo
283 $(INSTALL_DATA) libfoo.a $(DESTDIR)$(libdir)/libfoo.a
287 Always use a file name, not a directory name, as the second argument of
288 the installation commands. Use a separate command for each file to be
291 @node Directory Variables
292 @section Variables for Installation Directories
294 Installation directories should always be named by variables, so it is
295 easy to install in a nonstandard place. The standard names for these
296 variables are described below. They are based on a standard filesystem
297 layout; variants of it are used in SVR4, 4.4BSD, GNU/Linux, Ultrix v4,
298 and other modern operating systems.
300 These two variables set the root for the installation. All the other
301 installation directories should be subdirectories of one of these two,
302 and nothing should be directly installed into these two directories.
307 A prefix used in constructing the default values of the variables listed
308 below. The default value of @code{prefix} should be @file{/usr/local}.
309 When building the complete GNU system, the prefix will be empty and
310 @file{/usr} will be a symbolic link to @file{/}.
311 (If you are using Autoconf, write it as @samp{@@prefix@@}.)
313 Running @samp{make install} with a different value of @code{prefix} from
314 the one used to build the program should @emph{not} recompile the
319 A prefix used in constructing the default values of some of the
320 variables listed below. The default value of @code{exec_prefix} should
322 (If you are using Autoconf, write it as @samp{@@exec_prefix@@}.)
324 Generally, @code{$(exec_prefix)} is used for directories that contain
325 machine-specific files (such as executables and subroutine libraries),
326 while @code{$(prefix)} is used directly for other directories.
328 Running @samp{make install} with a different value of @code{exec_prefix}
329 from the one used to build the program should @emph{not} recompile the
333 Executable programs are installed in one of the following directories.
338 The directory for installing executable programs that users can run.
339 This should normally be @file{/usr/local/bin}, but write it as
340 @file{$(exec_prefix)/bin}.
341 (If you are using Autoconf, write it as @samp{@@bindir@@}.)
345 The directory for installing executable programs that can be run from
346 the shell, but are only generally useful to system administrators. This
347 should normally be @file{/usr/local/sbin}, but write it as
348 @file{$(exec_prefix)/sbin}.
349 (If you are using Autoconf, write it as @samp{@@sbindir@@}.)
353 @comment This paragraph adjusted to avoid overfull hbox --roland 5jul94
354 The directory for installing executable programs to be run by other
355 programs rather than by users. This directory should normally be
356 @file{/usr/local/libexec}, but write it as @file{$(exec_prefix)/libexec}.
357 (If you are using Autoconf, write it as @samp{@@libexecdir@@}.)
360 Data files used by the program during its execution are divided into
361 categories in two ways.
365 Some files are normally modified by programs; others are never normally
366 modified (though users may edit some of these).
369 Some files are architecture-independent and can be shared by all
370 machines at a site; some are architecture-dependent and can be shared
371 only by machines of the same kind and operating system; others may never
372 be shared between two machines.
375 This makes for six different possibilities. However, we want to
376 discourage the use of architecture-dependent files, aside from object
377 files and libraries. It is much cleaner to make other data files
378 architecture-independent, and it is generally not hard.
380 Here are the variables Makefiles should use to specify directories
381 to put these various kinds of files in:
385 The root of the directory tree for read-only architecture-independent
386 data files. This should normally be @file{/usr/local/share}, but
387 write it as @file{$(prefix)/share}. @samp{datadir}'s default value is
388 based on this variable; so are @samp{infodir}, @samp{mandir}, and others.
391 The directory for installing idiosyncratic read-only
392 architecture-independent data files for this program. This is usually
393 the same place as @samp{datarootdir}, but we use the two separate
394 variables so that you can move these idiosyncratic files without
395 altering the location for Info files, man pages, etc.
397 The default definition of @samp{datadir} should be
398 @file{$(datarootdir)}. (If you are using Autoconf, write it as
402 The directory for installing read-only data files that pertain to a
403 single machine--that is to say, files for configuring a host. Mailer
404 and network configuration files, @file{/etc/passwd}, and so forth belong
405 here. All the files in this directory should be ordinary ASCII text
406 files. This directory should normally be @file{/usr/local/etc}, but
407 write it as @file{$(prefix)/etc}.
408 (If you are using Autoconf, write it as @samp{@@sysconfdir@@}.)
410 Do not install executables here in this directory (they probably belong
411 in @file{$(libexecdir)} or @file{$(sbindir)}). Also do not install
412 files that are modified in the normal course of their use (programs
413 whose purpose is to change the configuration of the system excluded).
414 Those probably belong in @file{$(localstatedir)}.
417 The directory for installing architecture-independent data files which
418 the programs modify while they run. This should normally be
419 @file{/usr/local/com}, but write it as @file{$(prefix)/com}.
420 (If you are using Autoconf, write it as @samp{@@sharedstatedir@@}.)
423 The directory for installing data files which the programs modify while
424 they run, and that pertain to one specific machine. Users should never
425 need to modify files in this directory to configure the package's
426 operation; put such configuration information in separate files that go
427 in @file{$(datadir)} or @file{$(sysconfdir)}. @file{$(localstatedir)}
428 should normally be @file{/usr/local/var}, but write it as
429 @file{$(prefix)/var}.
430 (If you are using Autoconf, write it as @samp{@@localstatedir@@}.)
433 These variables specify the directory for installing certain specific
434 types of files, if your program has them. Every GNU package should
435 have Info files, so every program needs @samp{infodir}, but not all
436 need @samp{libdir} or @samp{lispdir}.
440 @c rewritten to avoid overfull hbox --roland
441 The directory for installing header files to be included by user
442 programs with the C @samp{#include} preprocessor directive. This
443 should normally be @file{/usr/local/include}, but write it as
444 @file{$(prefix)/include}.
445 (If you are using Autoconf, write it as @samp{@@includedir@@}.)
447 Most compilers other than GCC do not look for header files in directory
448 @file{/usr/local/include}. So installing the header files this way is
449 only useful with GCC. Sometimes this is not a problem because some
450 libraries are only really intended to work with GCC. But some libraries
451 are intended to work with other compilers. They should install their
452 header files in two places, one specified by @code{includedir} and one
453 specified by @code{oldincludedir}.
456 The directory for installing @samp{#include} header files for use with
457 compilers other than GCC. This should normally be @file{/usr/include}.
458 (If you are using Autoconf, you can write it as @samp{@@oldincludedir@@}.)
460 The Makefile commands should check whether the value of
461 @code{oldincludedir} is empty. If it is, they should not try to use
462 it; they should cancel the second installation of the header files.
464 A package should not replace an existing header in this directory unless
465 the header came from the same package. Thus, if your Foo package
466 provides a header file @file{foo.h}, then it should install the header
467 file in the @code{oldincludedir} directory if either (1) there is no
468 @file{foo.h} there or (2) the @file{foo.h} that exists came from the Foo
471 To tell whether @file{foo.h} came from the Foo package, put a magic
472 string in the file---part of a comment---and @code{grep} for that string.
475 The directory for installing documentation files (other than Info) for
476 this package. By default, it should be
477 @file{/usr/local/share/doc/@var{yourpkg}}, but it should be written as
478 @file{$(datarootdir)/doc/@var{yourpkg}}. (If you are using Autoconf,
479 write it as @samp{@@docdir@@}.) The @var{yourpkg} subdirectory, which
480 may include a version number, prevents collisions among files with
481 common names, such as @file{README}.
484 The directory for installing the Info files for this package. By
485 default, it should be @file{/usr/local/share/info}, but it should be
486 written as @file{$(datarootdir)/info}. (If you are using Autoconf,
487 write it as @samp{@@infodir@@}.) @code{infodir} is separate from
488 @code{docdir} for compatibility with existing practice.
494 Directories for installing documentation files in the particular
495 format. (It is not required to support documentation in all these
496 formats.) They should all be set to @code{$(docdir)} by default. (If
497 you are using Autoconf, write them as @samp{@@htmldir@@},
498 @samp{@@dvidir@@}, etc.) Packages which supply several translations
499 of their documentation should install them in
500 @samp{$(htmldir)/}@var{ll}, @samp{$(pdfdir)/}@var{ll}, etc. where
501 @var{ll} is a locale abbreviation such as @samp{en} or @samp{pt_BR}.
504 The directory for object files and libraries of object code. Do not
505 install executables here, they probably ought to go in @file{$(libexecdir)}
506 instead. The value of @code{libdir} should normally be
507 @file{/usr/local/lib}, but write it as @file{$(exec_prefix)/lib}.
508 (If you are using Autoconf, write it as @samp{@@libdir@@}.)
511 The directory for installing any Emacs Lisp files in this package. By
512 default, it should be @file{/usr/local/share/emacs/site-lisp}, but it
513 should be written as @file{$(datarootdir)/emacs/site-lisp}.
515 If you are using Autoconf, write the default as @samp{@@lispdir@@}.
516 In order to make @samp{@@lispdir@@} work, you need the following lines
517 in your @file{configure.in} file:
520 lispdir='$@{datarootdir@}/emacs/site-lisp'
525 The directory for installing locale-specific message catalogs for this
526 package. By default, it should be @file{/usr/local/share/locale}, but
527 it should be written as @file{$(datarootdir)/locale}. (If you are
528 using Autoconf, write it as @samp{@@localedir@@}.)
531 Unix-style man pages are installed in one of the following:
535 The top-level directory for installing the man pages (if any) for this
536 package. It will normally be @file{/usr/local/share/man}, but you
537 should write it as @file{$(datarootdir)/man}. (If you are using
538 Autoconf, write it as @samp{@@mandir@@}.)
541 The directory for installing section 1 man pages. Write it as
542 @file{$(mandir)/man1}.
544 The directory for installing section 2 man pages. Write it as
545 @file{$(mandir)/man2}
548 @strong{Don't make the primary documentation for any GNU software be a
549 man page. Write a manual in Texinfo instead. Man pages are just for
550 the sake of people running GNU software on Unix, which is a secondary
554 The file name extension for the installed man page. This should contain
555 a period followed by the appropriate digit; it should normally be @samp{.1}.
558 The file name extension for installed section 1 man pages.
560 The file name extension for installed section 2 man pages.
562 Use these names instead of @samp{manext} if the package needs to install man
563 pages in more than one section of the manual.
566 And finally, you should set the following variable:
570 The directory for the sources being compiled. The value of this
571 variable is normally inserted by the @code{configure} shell script.
572 (If you are using Autconf, use @samp{srcdir = @@srcdir@@}.)
578 @c I have changed some of the comments here slightly to fix an overfull
579 @c hbox, so the make manual can format correctly. --roland
580 # Common prefix for installation directories.
581 # NOTE: This directory must exist when you start the install.
583 datarootdir = $(prefix)/share
584 datadir = $(datarootdir)
585 exec_prefix = $(prefix)
586 # Where to put the executable for the command `gcc'.
587 bindir = $(exec_prefix)/bin
588 # Where to put the directories used by the compiler.
589 libexecdir = $(exec_prefix)/libexec
590 # Where to put the Info files.
591 infodir = $(datarootdir)/info
594 If your program installs a large number of files into one of the
595 standard user-specified directories, it might be useful to group them
596 into a subdirectory particular to that program. If you do this, you
597 should write the @code{install} rule to create these subdirectories.
599 Do not expect the user to include the subdirectory name in the value of
600 any of the variables listed above. The idea of having a uniform set of
601 variable names for installation directories is to enable the user to
602 specify the exact same values for several different GNU packages. In
603 order for this to be useful, all the packages must be designed so that
604 they will work sensibly when the user does so.
606 @node Standard Targets
607 @section Standard Targets for Users
609 All GNU programs should have the following targets in their Makefiles:
613 Compile the entire program. This should be the default target. This
614 target need not rebuild any documentation files; Info files should
615 normally be included in the distribution, and DVI files should be made
616 only when explicitly asked for.
618 By default, the Make rules should compile and link with @samp{-g}, so
619 that executable programs have debugging symbols. Users who don't mind
620 being helpless can strip the executables later if they wish.
623 Compile the program and copy the executables, libraries, and so on to
624 the file names where they should reside for actual use. If there is a
625 simple test to verify that a program is properly installed, this target
626 should run that test.
628 Do not strip executables when installing them. Devil-may-care users can
629 use the @code{install-strip} target to do that.
631 If possible, write the @code{install} target rule so that it does not
632 modify anything in the directory where the program was built, provided
633 @samp{make all} has just been done. This is convenient for building the
634 program under one user name and installing it under another.
636 The commands should create all the directories in which files are to be
637 installed, if they don't already exist. This includes the directories
638 specified as the values of the variables @code{prefix} and
639 @code{exec_prefix}, as well as all subdirectories that are needed.
640 One way to do this is by means of an @code{installdirs} target
643 Use @samp{-} before any command for installing a man page, so that
644 @code{make} will ignore any errors. This is in case there are systems
645 that don't have the Unix man page documentation system installed.
647 The way to install Info files is to copy them into @file{$(infodir)}
648 with @code{$(INSTALL_DATA)} (@pxref{Command Variables}), and then run
649 the @code{install-info} program if it is present. @code{install-info}
650 is a program that edits the Info @file{dir} file to add or update the
651 menu entry for the given Info file; it is part of the Texinfo package.
652 Here is a sample rule to install an Info file:
654 @comment This example has been carefully formatted for the Make manual.
655 @comment Please do not reformat it without talking to roland@gnu.ai.mit.edu.
657 $(DESTDIR)$(infodir)/foo.info: foo.info
659 # There may be a newer info file in . than in srcdir.
660 -if test -f foo.info; then d=.; \
661 else d=$(srcdir); fi; \
662 $(INSTALL_DATA) $$d/foo.info $(DESTDIR)$@@; \
663 # Run install-info only if it exists.
664 # Use `if' instead of just prepending `-' to the
665 # line so we notice real errors from install-info.
666 # We use `$(SHELL) -c' because some shells do not
667 # fail gracefully when there is an unknown command.
668 if $(SHELL) -c 'install-info --version' \
669 >/dev/null 2>&1; then \
670 install-info --dir-file=$(DESTDIR)$(infodir)/dir \
671 $(DESTDIR)$(infodir)/foo.info; \
675 When writing the @code{install} target, you must classify all the
676 commands into three categories: normal ones, @dfn{pre-installation}
677 commands and @dfn{post-installation} commands. @xref{Install Command
684 These targets install documentation in formats other than Info;
685 they're intended to be called explicitly by the person installing the
686 package, if that format is desired. GNU prefers Info files, so these
687 must be installed by the @code{install} target.
689 When you have many documentation files to install, we recommend that
690 you avoid collisions and clutter by arranging for these targets to
691 install in subdirectories of the appropriate installation directory,
692 such as @code{htmldir}. As one example, if your package has multiple
693 manuals, and you wish to install HTML documentation with many files
694 (such as the ``split'' mode output by @code{makeinfo --html}), you'll
695 certainly want to use subdirectories, or two nodes with the same name
696 in different manuals will overwrite each other.
699 Delete all the installed files---the copies that the @samp{install}
702 This rule should not modify the directories where compilation is done,
703 only the directories where files are installed.
705 The uninstallation commands are divided into three categories, just like
706 the installation commands. @xref{Install Command Categories}.
709 Like @code{install}, but strip the executable files while installing
710 them. In simple cases, this target can use the @code{install} target in
715 $(MAKE) INSTALL_PROGRAM='$(INSTALL_PROGRAM) -s' \
719 But if the package installs scripts as well as real executables, the
720 @code{install-strip} target can't just refer to the @code{install}
721 target; it has to strip the executables but not the scripts.
723 @code{install-strip} should not strip the executables in the build
724 directory which are being copied for installation. It should only strip
725 the copies that are installed.
727 Normally we do not recommend stripping an executable unless you are sure
728 the program has no bugs. However, it can be reasonable to install a
729 stripped executable for actual execution while saving the unstripped
730 executable elsewhere in case there is a bug.
732 @comment The gratuitous blank line here is to make the table look better
733 @comment in the printed Make manual. Please leave it in.
736 Delete all files in the current directory that are normally created by
737 building the program. Also delete files in other directories if they
738 are created by this makefile. However, don't delete the files that
739 record the configuration. Also preserve files that could be made by
740 building, but normally aren't because the distribution comes with
741 them. There is no need to delete parent directories that were created
742 with @samp{mkdir -p}, since they could have existed anyway.
744 Delete @file{.dvi} files here if they are not part of the distribution.
747 Delete all files in the current directory (or created by this
748 makefile) that are created by configuring or building the program. If
749 you have unpacked the source and built the program without creating
750 any other files, @samp{make distclean} should leave only the files
751 that were in the distribution. However, there is no need to delete
752 parent directories that were created with @samp{mkdir -p}, since they
753 could have existed anyway.
756 Like @samp{clean}, but may refrain from deleting a few files that people
757 normally don't want to recompile. For example, the @samp{mostlyclean}
758 target for GCC does not delete @file{libgcc.a}, because recompiling it
759 is rarely necessary and takes a lot of time.
761 @item maintainer-clean
762 Delete almost everything that can be reconstructed with this Makefile.
763 This typically includes everything deleted by @code{distclean}, plus
764 more: C source files produced by Bison, tags tables, Info files, and
767 The reason we say ``almost everything'' is that running the command
768 @samp{make maintainer-clean} should not delete @file{configure} even
769 if @file{configure} can be remade using a rule in the Makefile. More
770 generally, @samp{make maintainer-clean} should not delete anything
771 that needs to exist in order to run @file{configure} and then begin to
772 build the program. Also, there is no need to delete parent
773 directories that were created with @samp{mkdir -p}, since they could
774 have existed anyway. These are the only exceptions;
775 @code{maintainer-clean} should delete everything else that can be
778 The @samp{maintainer-clean} target is intended to be used by a maintainer of
779 the package, not by ordinary users. You may need special tools to
780 reconstruct some of the files that @samp{make maintainer-clean} deletes.
781 Since these files are normally included in the distribution, we don't
782 take care to make them easy to reconstruct. If you find you need to
783 unpack the full distribution again, don't blame us.
785 To help make users aware of this, the commands for the special
786 @code{maintainer-clean} target should start with these two:
789 @@echo 'This command is intended for maintainers to use; it'
790 @@echo 'deletes files that may need special tools to rebuild.'
794 Update a tags table for this program.
798 Generate any Info files needed. The best way to write the rules is as
804 foo.info: foo.texi chap1.texi chap2.texi
805 $(MAKEINFO) $(srcdir)/foo.texi
809 You must define the variable @code{MAKEINFO} in the Makefile. It should
810 run the @code{makeinfo} program, which is part of the Texinfo
813 Normally a GNU distribution comes with Info files, and that means the
814 Info files are present in the source directory. Therefore, the Make
815 rule for an info file should update it in the source directory. When
816 users build the package, ordinarily Make will not update the Info files
817 because they will already be up to date.
823 Generate documentation files in the given format, if possible.
824 Here's an example rule for generating DVI files from Texinfo:
829 foo.dvi: foo.texi chap1.texi chap2.texi
830 $(TEXI2DVI) $(srcdir)/foo.texi
834 You must define the variable @code{TEXI2DVI} in the Makefile. It should
835 run the program @code{texi2dvi}, which is part of the Texinfo
836 distribution.@footnote{@code{texi2dvi} uses @TeX{} to do the real work
837 of formatting. @TeX{} is not distributed with Texinfo.} Alternatively,
838 write just the dependencies, and allow GNU @code{make} to provide the command.
840 Here's another example, this one for generating HTML from Texinfo:
845 foo.html: foo.texi chap1.texi chap2.texi
846 $(TEXI2HTML) $(srcdir)/foo.texi
850 Again, you would define the variable @code{TEXI2HTML} in the Makefile;
851 for example, it might run @code{makeinfo --no-split --html}
852 (@command{makeinfo} is part of the Texinfo distribution).
855 Create a distribution tar file for this program. The tar file should be
856 set up so that the file names in the tar file start with a subdirectory
857 name which is the name of the package it is a distribution for. This
858 name can include the version number.
860 For example, the distribution tar file of GCC version 1.40 unpacks into
861 a subdirectory named @file{gcc-1.40}.
863 The easiest way to do this is to create a subdirectory appropriately
864 named, use @code{ln} or @code{cp} to install the proper files in it, and
865 then @code{tar} that subdirectory.
867 Compress the tar file with @code{gzip}. For example, the actual
868 distribution file for GCC version 1.40 is called @file{gcc-1.40.tar.gz}.
870 The @code{dist} target should explicitly depend on all non-source files
871 that are in the distribution, to make sure they are up to date in the
874 @xref{Releases, , Making Releases}.
877 @xref{Releases, , Making Releases, standards, GNU Coding Standards}.
881 Perform self-tests (if any). The user must build the program before
882 running the tests, but need not install the program; you should write
883 the self-tests so that they work when the program is built but not
887 The following targets are suggested as conventional names, for programs
888 in which they are useful.
892 Perform installation tests (if any). The user must build and install
893 the program before running the tests. You should not assume that
894 @file{$(bindir)} is in the search path.
897 It's useful to add a target named @samp{installdirs} to create the
898 directories where files are installed, and their parent directories.
899 There is a script called @file{mkinstalldirs} which is convenient for
900 this; you can find it in the Texinfo package.
901 @c It's in /gd/gnu/lib/mkinstalldirs.
902 You can use a rule like this:
904 @comment This has been carefully formatted to look decent in the Make manual.
905 @comment Please be sure not to make it extend any further to the right.--roland
907 # Make sure all installation directories (e.g. $(bindir))
908 # actually exist by making them if necessary.
909 installdirs: mkinstalldirs
910 $(srcdir)/mkinstalldirs $(bindir) $(datadir) \
911 $(libdir) $(infodir) \
916 or, if you wish to support @env{DESTDIR},
919 # Make sure all installation directories (e.g. $(bindir))
920 # actually exist by making them if necessary.
921 installdirs: mkinstalldirs
922 $(srcdir)/mkinstalldirs \
923 $(DESTDIR)$(bindir) $(DESTDIR)$(datadir) \
924 $(DESTDIR)$(libdir) $(DESTDIR)$(infodir) \
928 This rule should not modify the directories where compilation is done.
929 It should do nothing but create installation directories.
932 @node Install Command Categories
933 @section Install Command Categories
935 @cindex pre-installation commands
936 @cindex post-installation commands
937 When writing the @code{install} target, you must classify all the
938 commands into three categories: normal ones, @dfn{pre-installation}
939 commands and @dfn{post-installation} commands.
941 Normal commands move files into their proper places, and set their
942 modes. They may not alter any files except the ones that come entirely
943 from the package they belong to.
945 Pre-installation and post-installation commands may alter other files;
946 in particular, they can edit global configuration files or data bases.
948 Pre-installation commands are typically executed before the normal
949 commands, and post-installation commands are typically run after the
952 The most common use for a post-installation command is to run
953 @code{install-info}. This cannot be done with a normal command, since
954 it alters a file (the Info directory) which does not come entirely and
955 solely from the package being installed. It is a post-installation
956 command because it needs to be done after the normal command which
957 installs the package's Info files.
959 Most programs don't need any pre-installation commands, but we have the
960 feature just in case it is needed.
962 To classify the commands in the @code{install} rule into these three
963 categories, insert @dfn{category lines} among them. A category line
964 specifies the category for the commands that follow.
966 A category line consists of a tab and a reference to a special Make
967 variable, plus an optional comment at the end. There are three
968 variables you can use, one for each category; the variable name
969 specifies the category. Category lines are no-ops in ordinary execution
970 because these three Make variables are normally undefined (and you
971 @emph{should not} define them in the makefile).
973 Here are the three possible category lines, each with a comment that
974 explains what it means:
977 $(PRE_INSTALL) # @r{Pre-install commands follow.}
978 $(POST_INSTALL) # @r{Post-install commands follow.}
979 $(NORMAL_INSTALL) # @r{Normal commands follow.}
982 If you don't use a category line at the beginning of the @code{install}
983 rule, all the commands are classified as normal until the first category
984 line. If you don't use any category lines, all the commands are
985 classified as normal.
987 These are the category lines for @code{uninstall}:
990 $(PRE_UNINSTALL) # @r{Pre-uninstall commands follow.}
991 $(POST_UNINSTALL) # @r{Post-uninstall commands follow.}
992 $(NORMAL_UNINSTALL) # @r{Normal commands follow.}
995 Typically, a pre-uninstall command would be used for deleting entries
996 from the Info directory.
998 If the @code{install} or @code{uninstall} target has any dependencies
999 which act as subroutines of installation, then you should start
1000 @emph{each} dependency's commands with a category line, and start the
1001 main target's commands with a category line also. This way, you can
1002 ensure that each command is placed in the right category regardless of
1003 which of the dependencies actually run.
1005 Pre-installation and post-installation commands should not run any
1006 programs except for these:
1009 [ basename bash cat chgrp chmod chown cmp cp dd diff echo
1010 egrep expand expr false fgrep find getopt grep gunzip gzip
1011 hostname install install-info kill ldconfig ln ls md5sum
1012 mkdir mkfifo mknod mv printenv pwd rm rmdir sed sort tee
1013 test touch true uname xargs yes
1016 @cindex binary packages
1017 The reason for distinguishing the commands in this way is for the sake
1018 of making binary packages. Typically a binary package contains all the
1019 executables and other files that need to be installed, and has its own
1020 method of installing them---so it does not need to run the normal
1021 installation commands. But installing the binary package does need to
1022 execute the pre-installation and post-installation commands.
1024 Programs to build binary packages work by extracting the
1025 pre-installation and post-installation commands. Here is one way of
1026 extracting the pre-installation commands:
1029 make -n install -o all \
1030 PRE_INSTALL=pre-install \
1031 POST_INSTALL=post-install \
1032 NORMAL_INSTALL=normal-install \
1033 | gawk -f pre-install.awk
1037 where the file @file{pre-install.awk} could contain this:
1040 $0 ~ /^\t[ \t]*(normal_install|post_install)[ \t]*$/ @{on = 0@}
1042 $0 ~ /^\t[ \t]*pre_install[ \t]*$/ @{on = 1@}
1045 The resulting file of pre-installation commands is executed as a shell
1046 script as part of installing the binary package.