1 ============================================================================
4 * This file attempts to describe the rules to use when hacking
7 ============================================================================
10 * The correct response to most actual bugs is to write a new test case
11 which demonstrates the bug. Then fix the bug, re-run the test suite,
12 and check everything in.
14 * If you incorporate a change from somebody on the net:
15 - First, if it is a large change, you must make sure they have
16 signed the appropriate paperwork.
17 - Second, be sure to add their name and email address to THANKS.
19 * If a change fixes a test, mention the test in the commit message.
20 If a change fixes a bug registered in the Automake debbugs tracker,
21 mention the bug number in the commit message.
23 * If somebody reports a new bug, mention his name in the commit message
24 that fixes or exposes the bug, and put him into THANKS.
26 * When documenting a non-trivial idiom or example in the manual, be
27 sure to add a test case for it, and to reference such test case from
28 a proper Texinfo comment.
30 * Some files in the automake package are not owned by automake; these
31 files are listed in the $(FETCHFILES) variable in Makefile.am. They
32 should never be edited here. Almost all of them can be updated from
33 respective upstreams with "make fetch" (this should be done especially
34 before releases). The only exception is the 'lib/COPYING' (from FSF),
35 which should be updated by hand whenever the GPL gets updated (which
36 shouldn't happen that often anyway :-)
38 * Changes other than *trivial* bug fixes must be mentioned in NEWS.
40 ============================================================================
43 * We've adopted the convention that internal AC_SUBSTs and make variables
44 should be named with a leading 'am__', and internally generated targets
45 should be named with a leading 'am--'. This convention, although in
46 place from at least February 2001, isn't yet universally used.
47 But all new code should use it.
49 We used to use '_am_' as the prefix for an internal AC_SUBSTs.
50 However, it turns out that NEWS-OS 4.2R complains if a Makefile
51 variable begins with the underscore character. Yay for them.
52 I changed the target naming convention just to be safe.
54 ============================================================================
57 * Always use $(...) and not ${...}
59 * Prefer ':' over 'true', mostly for consistency with existing code.
61 * Use '##' comments liberally. Comment anything even remotely unusual.
63 * Never use basename or dirname. Instead, use sed.
65 * Do not use 'cd' within back-quotes, use '$(am__cd)' instead.
66 Otherwise the directory name may be printed, depending on CDPATH.
67 More generally, do not ever use plain 'cd' together with a relative
68 directory that does not start with a dot, or you might end up in one
71 * For install and uninstall rules, if a loop is required, it should be
72 silent. Then the body of the loop itself should print each "important"
73 command it runs. The printed commands should be preceded by a single
76 * Ensure install rules do not create any installation directory where
77 nothing is to be actually installed. See automake bug#11030.
79 ============================================================================
80 = Editing automake.in and aclocal.in
82 * Indent using GNU style. For historical reasons, the perl code
83 contains portions indented using Larry Wall's style (perl-mode's
84 default), and other portions using the GNU style (cperl-mode's
85 default). Write new code using GNU style.
87 * Don't use & for function calls, unless required.
88 The use of & prevents prototypes from being checked.
90 ============================================================================
91 = Automake versioning and compatibility scheme
93 * There are three kinds of automake releases:
95 - new major releases (e.g., 2.0, 5.0)
96 - new minor releases (e.g., 1.14, 2.1)
97 - micro a.k.a. "bug-fixing" releases (e.g., 1.13.2, 2.0.1, 3.5.17).
99 A new major release should have the major version number bumped, and
100 the minor and micro version numbers reset to zero. A new minor release
101 should have the major version number unchanged, the minor version number
102 bumped, and the micro version number reset to zero. Finally, a new
103 micro version should have the major and minor version numbers unchanged,
104 and the micro version number bumped.
106 For example, the first minor version after 1.13.2 will be 1.14; the
107 first bug-fixing version after 1.14 that will be 1.14.1; the first
108 new major version after all such releases will be 2.0; the first
109 bug-fixing version after 2.0 will be 2.0.1; and a further bug-fixing
110 version after 2.0.1 will be 2.0.2.
112 * Micro releases should be just bug-fixing releases; no new features
113 should be added, and ideally, only trivial bugs, recent regressions,
114 or documentation issues should be addressed by them. On the other
115 hand, it's OK to include testsuite work and even testsuite refactoring
116 in a micro version, since a regression there is not going to annoy or
117 inconvenience Automake users, but only the Automake developers.
119 * Minor releases can introduce new "safe" features, do non-trivial but
120 mostly safe code clean-ups, and even add new runtime warnings (rigorously
121 non-fatal). But they shouldn't include any backward incompatible change,
122 nor contain any potentially destabilizing refactoring or sweeping change,
123 nor introduce new features whose implementation might be liable to cause
124 bugs or regressions in existing code. However, it might be acceptable to
125 introduce very limited and localized backward-incompatibilities, *only*
126 if that is necessary to fix non-trivial bugs, address serious performance
127 issues, or greatly enhance usability. But please, do this sparsely and
130 * Major releases can introduce backward-incompatibilities (albeit such
131 incompatibilities should be announced well in advance, and a smooth
132 transition plan prepared for them), and try more risking and daring
133 refactorings and code cleanups.
135 * For more information, refer to the extensive discussion associated
136 with automake bug#13578.
138 ============================================================================
141 * To regenerate dependent files created by aclocal and automake,
142 use the 'bootstrap.sh' script. It uses the code from the source
143 tree, so the resulting files (aclocal.m4 and Makefile.in) should
144 be the same as you would get if you install this version of
145 automake and use it to generate those files. Be sure to have the
146 latest stable version of Autoconf installed and available early
149 * The Automake git tree currently carries three basic branches: 'micro',
150 'maint' and 'master'.
152 * The 'micro' branch, reserved to changes that should go into the next
153 micro release; so it will just see fixes for regressions, trivial
154 bugs, or documentation issues, and no "active" development whatsoever.
155 Since emergency regression-fixing or security releases could be cut
156 from this branch at any time, it should always be kept in a releasable
159 * The 'maint' branch is where the development of the next minor release
160 takes place. It should be kept in a stable, almost-releasable state,
161 to simplify testing and deploying of new minor version. Note that
162 this is not a hard rule, and such "stability" is not expected to be
163 absolute (emergency releases are cut from the 'micro' branch anyway).
165 * The 'master' branch is reserved for the development of the next major
166 release. Experimenting a little is OK here, but don't let the branch
167 grow too unstable; if you need to do exploratory programming or
168 over-arching change, you should use a dedicated topic branch, and
169 only merge that back once it is reasonably stable.
171 * The 'micro' branch should be kept regularly merged into the 'maint'
172 branch, and the 'maint' branch into the 'master' branch. It is advisable
173 to merge only after a set of related commits have been applied, to avoid
174 introducing too much noise in the history.
176 * There may be a number of longer-lived feature branches for new
177 developments. They should be based off of a common ancestor of all
178 active branches to which the feature should or might be merged later.
180 * After a new minor release is done, the 'maint' branch is to be merged
181 into the 'micro' branch, and then a "new" 'maint' branch created
182 stemming from the resulting commit.
183 Similarly, after a new major release is done, the 'master' branch is to
184 be merged into both the 'micro' and 'maint' branches, and then "new"
185 'master' branch created stemming from the resulting commit.
187 * When fixing a bug (especially a long-standing one), it may be useful
188 to commit the fix to a new temporary branch based off the commit that
189 introduced the bug. Then this "bugfix branch" can be merged into all
190 the active branches descending from the buggy commit. This offers a
191 simple way to fix the bug consistently and effectively.
193 * For merges, prefer 'git merge --log' over plain 'git merge', so that
194 a later 'git log' gives an indication of which actual patches were
195 merged even when they don't appear early in the list.
197 * The 'master', 'maint' and 'micro' branches should not be rewound, i.e.,
198 should always fast-forward, except maybe for privacy issues. For
199 feature branches, the announcement for the branch should document
200 the rewinding policy. It is usually a good idea to keep rewindable
201 branches in the 'experimental/*' "namespace"; i.e., give them names
202 like 'experimental/testsuite-work' or 'experimental/djgpp-for-WinNT'.
204 ============================================================================
205 = Writing a good commit message
207 * Here is the general format that Automake's commit messages are expected
208 to follow. See the further points below for clarifications and minor
211 topic: brief description (this is the "summary line")
213 <reference to relevant bugs, if any>
215 Here goes a more detailed explanation of why the commit is needed,
216 and a general overview of what it does, and how. This section is
217 optional, but you are expected to provide it more often than not.
219 And if the detailed explanation is quite long or detailed, you can
220 want to break it in more paragraphs.
222 Then you can add references to relevant mailing list discussions
223 (if any), with proper links. But don't take this as an excuse for
224 writing incomplete commit messages! The "distilled" conclusions
225 reached in such discussions should have been placed in the
228 Finally, here you can thank people that motivated or helped the
229 change. So, thanks to John Doe for bringing up the issue, and to
230 J. Random Hacker for providing suggestions and testing the patch.
232 <detailed list of touched files>
234 * The <detailed list of touched files> is mandatory but for the most
235 trivial changes, and should follows the GNU guidelines for ChangeLog
236 entries (described explicitly in the GNU Coding Standards); it might
237 be something of this sort:
239 * some/file (func1): Improved frobnication.
240 (func2): Adjusted accordingly.
241 * another/file (foo, bar): Likewise.
242 * tests/foo.tap: New test.
243 * tests/Makefile.am (TESTS): Add it.
245 * If your commit fixes an automake bug registered in the tracker (say
246 numbered 1234), you should put the following line after the summary
249 This change fixes automake bug#1234.
251 * If your commit is just related to the given bug report, but does not
252 fix it, you might want to add a line like this instead:
254 This change is related to automake bug#1234.
256 * When referring to older commits, use 'git describe' output as pointer.
257 But also try to identify the given commit by date and/or summary line
258 if possible. Examples:
260 Since yesterday's commit, v1.11-2019-g4d2bf42, ...
262 ... removed in commit 'v1.11-1674-g02e9072' of 01-01-2012,
263 "dist: ditch support for lzma"...
265 ============================================================================
268 * Use "make check" and "make maintainer-check" liberally.
270 * Export the 'keep_testdirs' environment variable to "yes" to keep
271 test directories for successful tests also.
273 * Use perl coverage information to ensure your new code is thoroughly
274 tested by your new tests.
276 * See file 't/README' for more information.
278 ============================================================================
281 * The steps outlined here are meant to be followed for alpha and stable
282 releases as well. Where differences are expected, they will be
283 explicitly described.
285 * Fetch new versions of the files that are maintained by the FSF by
286 running "make fetch". In case any file in the automake repository
287 has been updated, commit and re-run the testsuite.
289 * Ensure that the copyright notices of the distributed files is up to
290 date. The maintainer-only target "update-copyright" can help with
293 * Check NEWS; in particular, ensure that all the relevant differences
294 with the last release are actually reported.
296 * Update the version number in configure.ac.
297 (The idea is that every other alpha number will be a net release.
298 The repository will always have its own "odd" number so we can easily
299 distinguish net and repo versions.)
303 make bootstrap && make check && make distcheck
305 It is also advised to run "git clean -fdx" before invoking the
306 bootstrap, to ensure a really clean rebuild. However, it must
307 be done carefully, because that command will remove *all* the
308 files that are not tracked by git!
310 * Run "make git-tag-release".
311 This will run the maintainer checks, verify that the local git
312 repository and working tree are clean and up-to-date, and create
313 a proper signed git tag for the release (based on the contents
316 * Run "make git-upload-release".
317 This will first verify that you are releasing from a tagged version
318 and that the local git repository and working tree are clean and
319 up-to-date, and will then run "make dist" to create the tarballs,
320 and invoke the 'gnupload' script sign and upload them to the correct
321 locations. In case you need to sign with a non-default key, you can
322 use "make GNUPLOADFLAGS='--user KEY' git-upload-release".
324 * For stable releases you'll have to update the manuals at www.gnu.org.
326 - Generate manuals (with the help of the standard gendocs.sh script):
330 The ready-to-be-uploaded manuals (in several formats) will be left
331 in the 'doc/web-manuals' directory.
333 - Commit the updated manuals to web CVS:
335 make web-manual-update
337 If your local username is different from your username at Savannah,
338 you'll have to override the 'CVS_USER' make variable accordingly;
341 make web-manual-update CVS_USER=slattarini
343 - Check for link errors, fix them, recheck until convergence:
344 <http://validator.w3.org/checklink>
346 * Create an announcement message with "make announcement". Edit the
347 generated 'announcement' file appropriately, in particularly filling
348 in by hand any "TODO" left in there.
350 * Update version number in configure.ac to next alpha number.
351 Re-run ./bootstrap.sh and commit.
353 * Don't forget to "git push" your changes so they appear in the public
356 * Send the announcement generated in the earlier steps at least to
357 <autotools-announce@gnu.org> and <automake@gnu.org>. If the release
358 is a stable one, the announcement must also go to <info-gnu@gnu.org>;
359 if it is an alpha or beta release, announcement should be sent also
360 to <platform-testers@gnu.org>, to maximize the possibility of early
361 testing on exotic or proprietary systems. Finally, copy an abridged
362 version of the announcement into the NEWS feed at:
363 <https://savannah.gnu.org/projects/automake>.
364 Be sure to link a version to the complete announcement (from
365 the version you sent to the automake list, as get archived on
366 <http://lists.gnu.org/archive/html/automake/>).
370 Copyright (C) 2003-2013 Free Software Foundation, Inc.
372 This program is free software; you can redistribute it and/or modify
373 it under the terms of the GNU General Public License as published by
374 the Free Software Foundation; either version 2, or (at your option)
377 This program is distributed in the hope that it will be useful,
378 but WITHOUT ANY WARRANTY; without even the implied warranty of
379 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
380 GNU General Public License for more details.
382 You should have received a copy of the GNU General Public License
383 along with this program. If not, see <http://www.gnu.org/licenses/>.