-================================================================
+============================================================================
= 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
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
+ - 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.
+ 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
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.
+* Changes other than *trivial* bug fixes must be mentioned in NEWS.
-================================================================
+============================================================================
= 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.
+* 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_SUBST.
+ 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 `_'. Yay for them. I changed the target
- naming convention just to be safe.
+ variable begins with the underscore character. Yay for them.
+ I changed the target naming convention just to be safe.
-================================================================
-= Editing `.am' files
+============================================================================
+= Editing '.am' files
* Always use $(...) and not ${...}
-* Use `:', not `true'. Use `exit 1', not `false'.
+* Prefer ':' over 'true', mostly for consistency with existing code.
-* Use `##' comments liberally. Comment anything even remotely
- unusual.
+* Use '##' comments liberally. Comment anything even remotely unusual.
-* Never use basename or dirname. Instead use sed.
+* Never use basename or dirname. Instead, use sed.
-* Do not use `cd' within back-quotes, use `$(am__cd)' instead.
+* 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
+ 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.
+ 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
* 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.
-================================================================
+============================================================================
+= 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' script. It uses the code from the source
+ 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. 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
+ 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
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.
+* For merges, 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. It is usually a good idea to keep rewindable
+ branches in the 'experimental/*' "namespace"; i.e., give them names
+ like 'experimental/testsuite-work' or 'experimental/djgpp-for-WinNT'.
+
+============================================================================
+= 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")
+
+ <reference to relevant bugs, if any>
+
+ Here goes a more detailed explanation of why the commit is needed,
+ and a general overview of what it does, and how. This section is
+ optional, but you are expected to provide it more often than not.
+
+ And if the detailed explanation is quite long or detailed, you can
+ want to break it in more paragraphs.
-* 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.
+ 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.
-* 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.
+ 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.
+
+ <detailed list of touched files>
+
+* The <detailed list of touched files> is mandatory but for the most
+ trivial changes, and should follows 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:
-* 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.
+ 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.
-* Make sure each test file is executable.
+* Export the 'keep_testdirs' environment variable to "yes" to keep
+ test directories for successful tests also.
-* 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.
+* 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.
- Commit. Unfortunately you need an FSF account to do this.
- (You can also use `make fetch', but that is still woefully incomplete.)
+* 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.
-* Update NEWS.
+* 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.
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.
+ make bootstrap && make check && make distcheck
-* 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.
+ 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!
-* Update version number in configure.ac to next alpha number.
- Re-run ./bootstrap and commit.
+* 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)).
-* Don't forget to `git push' your changes so they appear in the public
- git tree.
+* 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
-* 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:
<http://validator.w3.org/checklink>
-* Send the announcement at least to <autotools-announce@gnu.org> and
- <automake@gnu.org>. If the release is a stable one, the announcement
- must also go to <info-gnu@gnu.org>; if it is an alpha or beta release,
- announcement should be sent also to <platform-testers@gnu.org>, to
- maximize the possibility of early testing on exotic or proprietary
- systems. Finally, copy the announcement into the NEWS feed at
+* 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
+ <autotools-announce@gnu.org> and <automake@gnu.org>. If the release
+ is a stable one, the announcement must also go to <info-gnu@gnu.org>;
+ if it is an alpha or beta release, announcement should be sent also
+ to <platform-testers@gnu.org>, 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:
<https://savannah.gnu.org/projects/automake>.
+ Be sure to link a version to the complete announcement (from
+ the version you sent to the automake list, as get archived on
+ <http://lists.gnu.org/archive/html/automake/>).
-----
-Copyright (C) 2003, 2007, 2008, 2010, 2011, 2012 Free Software
-Foundation, Inc.
+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
projects / platform / upstream / automake.git / blobdiff