============================================================================ = 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 that fixes or exposes the bug, and 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 *trivial* bug fixes must be mentioned in NEWS. * Changes which are potentially controversial, require a non-trivial plan, or must be implemented gradually with a roadmap spanning several releases (either minor or major) should be discussed on the list, and have a proper entry in the PLANS directory. This entry should be always committed in the "maint" branch, even if the change it deals with is only for the master branch, or a topic branch. Usually, in addition to this, it is useful to open a "wishlist" report on the Automake debbugs tracker, to keep the idea more visible, and have the discussions surrounding it easily archived in a central place. ============================================================================ = Naming * We've adopted the convention that internal AC_SUBSTs and make variables 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_SUBSTs. However, it turns out that NEWS-OS 4.2R complains if a Makefile variable begins with the underscore character. Yay for them. I changed the target naming convention just to be safe. ============================================================================ = Editing '.am' files * Always use $(...) and not ${...} * Prefer ':' over 'true', mostly for consistency with existing code. * Use '##' comments liberally. Comment anything even remotely unusual. * Never use basename or dirname. Instead, use sed. * Do not use 'cd' within back-quotes, use '$(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. * Ensure install rules do not create any installation directory where nothing is to be actually installed. See automake bug#11030. ============================================================================ = 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 really required. The use of & prevents prototypes from being checked. ============================================================================ = Automake versioning and compatibility scheme * There are three kinds of automake releases: - new major releases (e.g., 2.0, 5.0) - new minor releases (e.g., 1.14, 2.1) - micro a.k.a. "bug-fixing" releases (e.g., 1.13.2, 2.0.1, 3.5.17). A new major release should have the major version number bumped, and the minor and micro version numbers reset to zero. A new minor release should have the major version number unchanged, the minor version number bumped, and the micro version number reset to zero. Finally, a new micro version should have the major and minor version numbers unchanged, and the micro version number bumped. For example, the first minor version after 1.13.2 will be 1.14; the first bug-fixing version after 1.14 that will be 1.14.1; the first new major version after all such releases will be 2.0; the first bug-fixing version after 2.0 will be 2.0.1; and a further bug-fixing version after 2.0.1 will be 2.0.2. * Micro releases should be just bug-fixing releases; no new features should be added, and ideally, only trivial bugs, recent regressions, or documentation issues should be addressed by them. On the other hand, it's OK to include testsuite work and even testsuite refactoring in a micro version, since a regression there is not going to annoy or inconvenience Automake users, but only the Automake developers. * Minor releases can introduce new "safe" features, do non-trivial but mostly safe code clean-ups, and even add new runtime warnings (rigorously non-fatal). But they shouldn't include any backward incompatible change, nor contain any potentially destabilizing refactoring or sweeping change, nor introduce new features whose implementation might be liable to cause bugs or regressions in existing code. However, it might be acceptable to introduce very limited and localized backward-incompatibilities, *only* if that is necessary to fix non-trivial bugs, address serious performance issues, or greatly enhance usability. But please, do this sparsely and rarely! * Major releases can introduce backward-incompatibilities (albeit such incompatibilities should be announced well in advance, and a smooth transition plan prepared for them), and try more risking and daring refactorings and code cleanups. * For more information, refer to the extensive discussion associated with automake bug#13578. ============================================================================ = Working with git * To regenerate dependent files created by aclocal and automake, use the 'bootstrap.sh' 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 and available early in your PATH. * The Automake git tree currently carries three basic branches: 'micro', 'maint' and 'master'. * The 'micro' branch, reserved to changes that should go into the next micro release; so it will just see fixes for regressions, trivial bugs, or documentation issues, and no "active" development whatsoever. Since emergency regression-fixing or security releases could be cut from this branch at any time, it should always be kept in a releasable state. * The 'maint' branch is where the development of the next minor release takes place. It should be kept in a stable, almost-releasable state, to simplify testing and deploying of new minor version. Note that this is not a hard rule, and such "stability" is not expected to be absolute (emergency releases are cut from the 'micro' branch anyway). * The 'master' branch is reserved for the development of the next major release. Experimenting a little is OK here, but don't let the branch grow too unstable; if you need to do exploratory programming or over-arching change, you should use a dedicated topic branch, and only merge that back once it is reasonably stable. * The 'micro' branch should be kept regularly merged into the 'maint' branch, and the 'maint' branch into the 'master' branch. It is advisable to merge only after a set of related commits have been applied, to avoid introducing too much noise in the history. * 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. * After a new minor release is done, the 'maint' branch is to be merged into the 'micro' branch, and then a "new" 'maint' branch created stemming from the resulting commit. Similarly, after a new major release is done, the 'master' branch is to be merged into both the 'micro' and 'maint' branches, and then "new" 'master' branch created stemming from the resulting commit. * 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 merging, 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. * The 'master', 'maint' and 'micro' branches should not be rewound, i.e., should always fast-forward, except maybe for privacy issues. For feature branches, the announcement for the branch should document the rewinding policy. If a topic branch is expected to be rewound, it is good practice to put it in the 'experimental/*' namespace; for example, a rewindable branch dealing with Vala support could be named like "experimental/vala-work". ============================================================================ = Writing a good commit message * Here is the general format that Automake's commit messages are expected to follow. See the further points below for clarifications and minor corrections. topic: brief description (this is the "summary line") Here goes a more detailed explanation of why the commit is needed, and a general overview of what it does, and how. This section should almost always be provided, possibly only with the expection of obvious fixes or very trivial changes. And if the detailed explanation is quite long or detailed, you can want to break it in more paragraphs. Then you can add references to relevant mailing list discussions (if any), with proper links. But don't take this as an excuse for writing incomplete commit messages! The "distilled" conclusions reached in such discussions should have been placed in the paragraphs above. Finally, here you can thank people that motivated or helped the change. So, thanks to John Doe for bringing up the issue, and to J. Random Hacker for providing suggestions and testing the patch. * The should usually be provided (but for short or trivial changes), and should follow the GNU guidelines for ChangeLog entries (described explicitly in the GNU Coding Standards); it might be something of this sort: * some/file (func1): Improved frobnication. (func2): Adjusted accordingly. * another/file (foo, bar): Likewise. * tests/foo.tap: New test. * tests/Makefile.am (TESTS): Add it. * If your commit fixes an automake bug registered in the tracker (say numbered 1234), you should put the following line after the summary line: This change fixes automake bug#1234. * If your commit is just related to the given bug report, but does not fix it, you might want to add a line like this instead: This change is related to automake bug#1234. * When referring to older commits, use 'git describe' output as pointer. But also try to identify the given commit by date and/or summary line if possible. Examples: Since yesterday's commit, v1.11-2019-g4d2bf42, ... ... removed in commit 'v1.11-1674-g02e9072' of 01-01-2012, "dist: ditch support for lzma"... ============================================================================ = Test suite * Use "make check" and "make maintainer-check" liberally. * Export the 'keep_testdirs' environment variable to "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 't/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 by running "make fetch". In case any file in the automake repository has been updated, commit and re-run the testsuite. * Ensure that the copyright notices of the distributed files is up to date. The maintainer-only target "update-copyright" can help with this. * Check NEWS; in particular, ensure that all the relevant differences with the last release are actually reported. * 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 these commands, in this order: make bootstrap make check keep_testdirs=yes make maintainer-check make distcheck make check-no-trailing-backslash-in-recipes make check-cc-no-c-o It is also advised to run "git clean -fdx" before invoking the bootstrap, to ensure a really clean rebuild. However, it must be done carefully, because that command will remove *all* the files that are not tracked by git! * Run "make git-tag-release". This will run the maintainer checks, verify that the local git repository and working tree are clean and up-to-date, and create a proper signed git tag for the release (based on the contents of $(VERSION)). * Run "make git-upload-release". This will first verify that you are releasing from a tagged version and that the local git repository and working tree are clean and up-to-date, and will then run "make dist" to create the tarballs, and invoke the 'gnupload' script sign and upload them to the correct locations. In case you need to sign with a non-default key, you can use "make GNUPLOADFLAGS='--user KEY' git-upload-release". * For stable releases you'll have to update the manuals at www.gnu.org. - Generate manuals (with the help of the standard gendocs.sh script): make web-manual The ready-to-be-uploaded manuals (in several formats) will be left in the 'doc/web-manuals' directory. - Commit the updated manuals to web CVS: make web-manual-update If your local username is different from your username at Savannah, you'll have to override the 'CVS_USER' make variable accordingly; for example: make web-manual-update CVS_USER=slattarini - Check for link errors, fix them, recheck until convergence: * Create an announcement message with "make announcement". Edit the generated 'announcement' file appropriately, in particularly filling in by hand any "TODO" left in there. * Update version number in configure.ac to next alpha number. Re-run ./bootstrap.sh and commit. * Don't forget to "git push" your changes so they appear in the public git tree. * Send the announcement generated in the earlier steps 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 an abridged version of the announcement into the NEWS feed at: . Be sure to link a version to the complete announcement (from the version you sent to the automake list, as get archived on ). ----- Copyright (C) 2003-2013 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: