lint: cosmetics: use #-comments, not ##-comments

[automake.git] / HACKING

============================================================================
= 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.

============================================================================
= 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 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.

* 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.

      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.

      <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:

      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 this:

    make bootstrap && make check && make distcheck

  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:
    <http://validator.w3.org/checklink>

* 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) 2003-2013 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 <http://www.gnu.org/licenses/>.

Local Variables:
mode: text
End: