================================================================ = This file * This file attempts to describe the rules to use when hacking automake. ================================================================ = Administrivia * The correct response to most actual bugs is to write a new test case which demonstrates the bug. Then fix the bug, re-run the test suite, and check everything in. * If you incorporate a change from somebody on the net: First, if it is a large change, you must make sure they have signed the appropriate paperwork. Second, be sure to add their name and email address to THANKS * If a change fixes a test, mention the test in the commit message. If a change fixes a bug registered in the Automake debbugs tracker, mention the bug number in the commit message. * If somebody reports a new bug, mention his name in the commit message and in the test case you write. Put him into THANKS. * When documenting a non-trivial idiom or example in the manual, be sure to add a test case for it, and to reference such test case from a proper Texinfo comment. * Some files in the automake package are not owned by automake; these files are listed in the $(FETCHFILES) variable in Makefile.am. They should never be edited here. Almost all of them can be updated from respective upstreams with "make fetch" (this should be done especially before releases). The only exception is the 'lib/COPYING' (from FSF), which should be updated by hand whenever the GPL gets updated (which shouldn't happen that often anyway :-) * Changes other than bug fixes must be mentioned in NEWS. Important bug fixes should be mentioned in NEWS, too. ================================================================ = Naming * We've adopted the convention that internal AC_SUBSTs should be named with a leading `am__', and internally generated targets should be named with a leading `am--'. This convention, although in place from at least February 2001, isn't yet universally used. But all new code should use it. We used to use `_am_' as the prefix for an internal AC_SUBST. However, it turns out that NEWS-OS 4.2R complains if a Makefile variable begins with `_'. Yay for them. I changed the target naming convention just to be safe. ================================================================ = Editing `.am' files * Always use $(...) and not ${...} * Use `:', not `true'. Use `exit 1', not `false'. * Use `##' comments liberally. Comment anything even remotely unusual. * Never use basename or dirname. Instead use sed. * Do not use `cd' within back-quotes, use `$(am__cd)' instead. Otherwise the directory name may be printed, depending on CDPATH. More generally, do not ever use plain `cd' together with a relative directory that does not start with a dot, or you might end up in one computed with CDPATH. * For install and uninstall rules, if a loop is required, it should be silent. Then the body of the loop itself should print each "important" command it runs. The printed commands should be preceded by a single space. ================================================================ = Editing automake.in and aclocal.in * Indent using GNU style. For historical reasons, the perl code contains portions indented using Larry Wall's style (perl-mode's default), and other portions using the GNU style (cperl-mode's default). Write new code using GNU style. * Don't use & for function calls, unless required. The use of & prevents prototypes from being checked. Just as above, don't change massively all the code to strip the &, just convert the old code as you work on it, and write new code without. ================================================================ = Working with git * To regenerate dependent files created by aclocal and automake, use the `bootstrap' script. It uses the code from the source tree, so the resulting files (aclocal.m4 and Makefile.in) should be the same as you would get if you install this version of automake and use it to generate those files. Be sure to have the latest stable version of Autoconf installed. If such version is not installed as "autoconf", pass it explicitly (along with the accompanying "autom4te") when calling `bootstrap' and `configure'. For example: $ AUTOCONF=autoconf2.68 AUTOM4TE=autom4te2.68 ./bootstrap $ ./configure AUTOCONF=autoconf2.68 AUTOM4TE=autom4te2.68 * Dependent files aclocal.m4, configure and Makefile.in in all directories should be up to date in the git repository, so that the changes in them can be easily noticed and analyzed. * The git tree currently carries a number of branches: master for the current development, and release branches named branch-X.Y. The maint branch serves as common ground for both master and the active release branches. Changes intended for both should be applied to maint, which should then be merged to release branches and master, of course after suitable testing. It is advisable to merge only after a set of related commits have been applied. * Example work flow for patches to maint: # 1. Checkout the "maint" branch: git checkout maint # 2. Apply the patch(es) with "git am" (or create them with $EDITOR): git am -3 0*.patch # 2a. Run required tests, if any ... # 3. Merge maint into branch-1.11: git checkout branch-1.11 git merge maint # 3a. Run required tests, if any ... # 4. Redo steps 3 and 3a for master: git checkout master git merge maint # testing ... # 5. Push the maint and master branches: git push --dry-run origin maint branch-1.11 master # if all seems ok, then actually push: git push origin maint branch-1.11 master * When fixing a bug (especially a long-standing one), it may be useful to commit the fix to a new temporary branch based off the commit that introduced the bug. Then this "bugfix branch" can be merged into all the active branches descending from the buggy commit. This offers a simple way to fix the bug consistently and effectively. * When referring to older commits, use 'git describe' output as pointer. * There may be a number of longer-lived feature branches for new developments. They should be based off of a common ancestor of all active branches to which the feature should or might be merged later. The next branch may serve as common ground for feature merging and testing, should they not be ready for master yet. * For merges from branches other than maint, prefer 'git merge --log' over plain 'git merge', so that a later 'git log' gives an indication of which actual patches were merged even when they don't appear early in the list. * master and release branches should not be rewound, i.e., should always fast-forward, except maybe for privacy issues. The maint branch should not be rewound except maybe after retiring a release branch or a new stable release. For next, and for feature branches, the announcement for the branch should document rewinding policy. ================================================================ = Test suite * Use "make check" and "make maintainer-check" liberally. * Make sure each test file is executable. * Use `keep_testdirs=yes' to keep test directories for successful tests also. * Use perl coverage information to ensure your new code is thoroughly tested by your new tests. * See file `tests/README' for more information. ================================================================ = Release procedure * The steps outlined here are meant to be followed for alpha and stable releases as well. Where differences are expected, they will be explicitly described. * Fetch new versions of the files that are maintained by the FSF. Commit. Unfortunately you need an FSF account to do this. (You can also use `make fetch', but that is still woefully incomplete.) * Update NEWS. * Update the version number in configure.ac. (The idea is that every other alpha number will be a net release. The repository will always have its own "odd" number so we can easily distinguish net and repo versions.) * Run this: ./bootstrap && ./configure && make && make check && make distcheck * Run `make release-stats' if release statistics in doc/automake.texi have not been updated yet. * Run `make git-release'. This will run "make dist" to create the tarballs, commit the last changes to NEWS, configure.ac and m4/amversion.m4, tag the repository, sign the tarballs, and upload them. Use `make GNUPLOADFLAGS="--user key" git-release' to sign with a non-default key. * Update version number in configure.ac to next alpha number. Re-run ./bootstrap and commit. * Don't forget to `git push' your changes so they appear in the public git tree. * For stable releases, update the manuals at www.gnu.org: - Generate manuals: cd doc wget "http://savannah.gnu.org/cgi-bin/viewcvs/~checkout~/texinfo/texinfo/util/gendocs.sh" wget "http://savannah.gnu.org/cgi-bin/viewcvs/~checkout~/texinfo/texinfo/util/gendocs_template" sh ./gendocs.sh --email bug-automake@gnu.org automake "GNU Automake" - copy manuals recursively to web cvs, - commit. - Check for link errors, fix them, recheck until convergence: * Send the announcement at least to and . If the release is a stable one, the announcement must also go to ; if it is an alpha or beta release, announcement should be sent also to , to maximize the possibility of early testing on exotic or proprietary systems. Finally, copy the announcement into the NEWS feed at . ----- Copyright (C) 2003, 2007, 2008, 2010, 2011, 2012 Free Software Foundation, Inc. This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2, or (at your option) any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program. If not, see . Local Variables: mode: text End: