]> git.ipfire.org Git - thirdparty/git.git/log
thirdparty/git.git
2 years agoMerge branch 'rs/apply-symlinks-use-strset'
Junio C Hamano [Sat, 5 Feb 2022 17:42:30 +0000 (09:42 -0800)] 
Merge branch 'rs/apply-symlinks-use-strset'

"git apply" (ab)used the util pointer of the string-list to keep
track of how each symbolic link needs to be handled, which has been
simplified by using strset.

* rs/apply-symlinks-use-strset:
  apply: use strsets to track symlinks

2 years agoMerge branch 'rs/grep-expr-cleanup'
Junio C Hamano [Sat, 5 Feb 2022 17:42:30 +0000 (09:42 -0800)] 
Merge branch 'rs/grep-expr-cleanup'

Code clean-up.

* rs/grep-expr-cleanup:
  grep: use grep_and_expr() in compile_pattern_and()
  grep: extract grep_binexp() from grep_or_expr()
  grep: use grep_not_expr() in compile_pattern_not()
  grep: use grep_or_expr() in compile_pattern_or()

2 years agoMerge branch 'jh/p4-spawning-external-commands-cleanup'
Junio C Hamano [Sat, 5 Feb 2022 17:42:29 +0000 (09:42 -0800)] 
Merge branch 'jh/p4-spawning-external-commands-cleanup'

* jh/p4-spawning-external-commands-cleanup:
  git-p4: don't print shell commands as python lists
  git-p4: pass command arguments as lists instead of using shell
  git-p4: don't select shell mode using the type of the command argument

2 years agoMerge branch 'jh/p4-fix-use-of-process-error-exception'
Junio C Hamano [Sat, 5 Feb 2022 17:42:29 +0000 (09:42 -0800)] 
Merge branch 'jh/p4-fix-use-of-process-error-exception'

* jh/p4-fix-use-of-process-error-exception:
  git-p4: fix instantiation of CalledProcessError

2 years agoMerge branch 'jc/find-header'
Junio C Hamano [Sat, 5 Feb 2022 17:42:29 +0000 (09:42 -0800)] 
Merge branch 'jc/find-header'

Code clean-up.

* jc/find-header:
  receive-pack.c: consolidate find header logic

2 years agoMerge branch 'pb/pull-rebase-autostash-fix'
Junio C Hamano [Sat, 5 Feb 2022 17:42:28 +0000 (09:42 -0800)] 
Merge branch 'pb/pull-rebase-autostash-fix'

"git pull --rebase" ignored the rebase.autostash configuration
variable when the remote history is a descendant of our history,
which has been corrected.

* pb/pull-rebase-autostash-fix:
  pull --rebase: honor rebase.autostash when fast-forwarding

2 years agoSync with Git 2.35.1
Junio C Hamano [Sat, 29 Jan 2022 00:58:04 +0000 (16:58 -0800)] 
Sync with Git 2.35.1

2 years agoName the next one 2.36 to prepare for 2.35.1
Junio C Hamano [Sat, 29 Jan 2022 00:57:10 +0000 (16:57 -0800)] 
Name the next one 2.36 to prepare for 2.35.1

Signed-off-by: Junio C Hamano <gitster@pobox.com>
2 years agoGit 2.35.1 v2.35.1
Junio C Hamano [Sat, 29 Jan 2022 00:48:42 +0000 (16:48 -0800)] 
Git 2.35.1

Signed-off-by: Junio C Hamano <gitster@pobox.com>
2 years agoMerge branch 'en/keep-cwd' into maint
Junio C Hamano [Sat, 29 Jan 2022 00:45:39 +0000 (16:45 -0800)] 
Merge branch 'en/keep-cwd' into maint

Fix a regression in 2.35 that roke the use of "rebase" and "stash"
in a secondary worktree.

* en/keep-cwd:
  sequencer, stash: fix running from worktree subdir

2 years agoMerge branch 'en/keep-cwd'
Junio C Hamano [Thu, 27 Jan 2022 06:22:24 +0000 (22:22 -0800)] 
Merge branch 'en/keep-cwd'

Fix a regression in 2.35 that roke the use of "rebase" and "stash"
in a secondary worktree.

* en/keep-cwd:
  sequencer, stash: fix running from worktree subdir

2 years agoStart post 2.35 cycle
Junio C Hamano [Thu, 27 Jan 2022 06:21:01 +0000 (22:21 -0800)] 
Start post 2.35 cycle

The tree is not open for new development yet, but let's mark the
beginning of the new cycle before we start merging down regression
fix topics.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
2 years agosequencer, stash: fix running from worktree subdir
Elijah Newren [Wed, 26 Jan 2022 01:43:45 +0000 (01:43 +0000)] 
sequencer, stash: fix running from worktree subdir

In commits bc3ae46b42 ("rebase: do not attempt to remove
startup_info->original_cwd", 2021-12-09) and 0fce211ccc ("stash: do not
attempt to remove startup_info->original_cwd", 2021-12-09), we wanted to
allow the subprocess to know which directory the parent process was
running from, so that the subprocess could protect it.  However...

When run from a non-main worktree, setup_git_directory() will note
that the discovered git directory
(/PATH/TO/.git/worktree/non-main-worktree) does not match
DEFAULT_GIT_DIR_ENVIRONMENT (see setup_discovered_git_dir()), and
decide to set GIT_DIR in the environment.  This matters because...

Whenever git is run with the GIT_DIR environment variable set, and
GIT_WORK_TREE not set, it presumes that '.' is the working tree.  So...

This combination results in the subcommand being very confused about
the working tree.  Fix it by also setting the GIT_WORK_TREE environment
variable along with setting cmd.dir.

A possibly more involved fix we could consider for later would be to
make setup.c set GIT_WORK_TREE whenever (a) it discovers both the git
directory and the working tree and (b) it decides to set GIT_DIR in the
environment.  I did not attempt that here as such would be too big of a
change for a 2.35.1 release.

Test-case-by: Glen Choo <chooglen@google.com>
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2 years agoGit 2.35 v2.35.0
Junio C Hamano [Mon, 24 Jan 2022 17:25:25 +0000 (09:25 -0800)] 
Git 2.35

Signed-off-by: Junio C Hamano <gitster@pobox.com>
2 years agoMerge branch 'ab/checkout-branch-info-leakfix'
Junio C Hamano [Mon, 24 Jan 2022 17:14:46 +0000 (09:14 -0800)] 
Merge branch 'ab/checkout-branch-info-leakfix'

We added an unrelated sanity checking that leads to a BUG() while
plugging a leak, which triggered in a repository with symrefs in
the local branch namespace that point at a ref outside.  Partially
revert the change to avoid triggering the BUG().

* ab/checkout-branch-info-leakfix:
  checkout: avoid BUG() when hitting a broken repository

2 years agoMerge tag 'l10n-2.35.0-rnd2' of git://github.com/git-l10n/git-po
Junio C Hamano [Mon, 24 Jan 2022 17:09:34 +0000 (09:09 -0800)] 
Merge tag 'l10n-2.35.0-rnd2' of git://github.com/git-l10n/git-po

l10n-2.35.0-rnd2

* tag 'l10n-2.35.0-rnd2' of git://github.com/git-l10n/git-po:
  l10n: Update Catalan translation
  l10n: zh_TW: v2.35.0 round 2 (0 untranslated)
  l10n: Update Catalan translation
  l10n: de.po: Update German translation
  l10n: de.po: Fix translation for "'%s' is aliased to '%s'"
  l10n: po-id for 2.35 (round 2)
  l10n: Update Catalan translation
  l10n: vi(5195t): Update for v2.35.0 round 2
  l10n: batch update to fix typo in branch.c
  l10n: git.pot: v2.35.0 round 2 (1 new, 1 removed)
  l10n: bg.po: Updated Bulgarian translation (5195t)
  l10n: zh_CN: v2.35.0 round 1
  l10n: fr: v2.35.0 round 1
  l10n: zh_TW: v2.35.0 round 1 (1 fuzzy)
  l10n: po-id for 2.35 (round 1)
  l10n: sv.po: Update Swedish translation (5196t0f0u)
  l10n: sv.po: Fix typo
  l10n: tr: v2.35.0 round 1
  l10n: git.pot: v2.35.0 round 1 (126 new, 142 removed)

2 years agol10n: Update Catalan translation
Jordi Mas [Sun, 23 Jan 2022 08:40:52 +0000 (09:40 +0100)] 
l10n: Update Catalan translation

Signed-off-by: Jordi Mas <jmas@softcatala.org>
2 years agoMerge branch 'l10n/zh_TW/220113' of github.com:l10n-tw/git-po
Jiang Xin [Sat, 22 Jan 2022 08:27:41 +0000 (16:27 +0800)] 
Merge branch 'l10n/zh_TW/220113' of github.com:l10n-tw/git-po

* 'l10n/zh_TW/220113' of github.com:l10n-tw/git-po:
  l10n: zh_TW: v2.35.0 round 2 (0 untranslated)
  l10n: zh_TW: v2.35.0 round 1 (1 fuzzy)

2 years agocheckout: avoid BUG() when hitting a broken repository
Junio C Hamano [Sat, 22 Jan 2022 00:58:30 +0000 (16:58 -0800)] 
checkout: avoid BUG() when hitting a broken repository

When 9081a421 (checkout: fix "branch info" memory leaks, 2021-11-16)
cleaned up existing memory leaks, we added an unrelated sanity check
to ensure that a local branch is truly local and not a symref to
elsewhere that dies with BUG() otherwise.  This was misguided in two
ways.  First of all, such a tightening did not belong to a leak-fix
patch.  And the condition it detected was *not* a bug in our program
but a problem in user data, where warning() or die() would have been
more appropriate.

As the condition is not fatal (the result of computing the local
branch name in the code that is involved in the faulty check is only
used as a textual label for the commit), let's revert the code to
the original state, i.e. strip "refs/heads/" to compute the local
branch name if possible, and otherwise leave it NULL.  The consumer
of the information in merge_working_tree() is prepared to see NULL
in there and act accordingly.

cf. https://bugzilla.redhat.com/show_bug.cgi?id=2042920

Reported-by: Petr Šplíchal <psplicha@redhat.com>
Reported-by: Todd Zullinger <tmz@pobox.com>
Helped-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2 years agol10n: zh_TW: v2.35.0 round 2 (0 untranslated)
Yi-Jyun Pan [Fri, 21 Jan 2022 23:06:36 +0000 (07:06 +0800)] 
l10n: zh_TW: v2.35.0 round 2 (0 untranslated)

Used 1 translation from zh_CN. Thanks to zh_CN translation team!

Signed-off-by: Yi-Jyun Pan <pan93412@gmail.com>
2 years agol10n: Update Catalan translation
Jordi Mas [Fri, 21 Jan 2022 06:56:02 +0000 (07:56 +0100)] 
l10n: Update Catalan translation

Signed-off-by: Jordi Mas <jmas@softcatala.org>
2 years agoMerge branch 'js/branch-track-inherit'
Junio C Hamano [Thu, 20 Jan 2022 23:25:38 +0000 (15:25 -0800)] 
Merge branch 'js/branch-track-inherit'

"git branch -h" incorrectly said "--track[=direct|inherit]",
implying that "--trackinherit" is a valid option, which has been
corrected.
source: <3de40324bea6a1dd9bca2654721471e3809e87d8.1642538935.git.steadmon@google.com>
source: <c3c26192-aee9-185a-e559-b8735139e49c@web.de>

* js/branch-track-inherit:
  branch,checkout: fix --track documentation

2 years agobranch,checkout: fix --track documentation
René Scharfe [Thu, 20 Jan 2022 12:35:54 +0000 (13:35 +0100)] 
branch,checkout: fix --track documentation

Document that the accepted variants of the --track option are --track,
--track=direct, and --track=inherit.  The equal sign in the latter two
cannot be replaced with whitespace; in general optional arguments need
to be attached firmly to their option.

Put "direct" consistently before "inherit", if only for the reasons
that the former is the default, explained first in the documentation,
and comes before the latter alphabetically.

Mention both modes in the short help so that readers don't have to look
them up in the full documentation.  They are literal strings and thus
untranslatable.  PARSE_OPT_LITERAL_ARGHELP is inferred due to the pipe
and parenthesis characters, so we don't have to provide that flag
explicitly.

Mention that -t has the same effect as --track and --track=direct.
There is no way to specify inherit mode using the short option, because
short options generally don't accept optional arguments.

Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2 years agol10n: de.po: Update German translation
Matthias Rüster [Sat, 15 Jan 2022 15:36:51 +0000 (16:36 +0100)] 
l10n: de.po: Update German translation

Signed-off-by: Matthias Rüster <matthias.ruester@gmail.com>
Reviewed-by: Ralf Thielow <ralf.thielow@gmail.com>
2 years agol10n: de.po: Fix translation for "'%s' is aliased to '%s'"
Jürgen Krämer [Thu, 18 Nov 2021 15:25:48 +0000 (16:25 +0100)] 
l10n: de.po: Fix translation for "'%s' is aliased to '%s'"

The German translation for "'%s' is aliased to '%s'" is incorrect. It
switches the order of alias name and alias definition.

A better translation would be "'%s' ist ein Alias für '%s'". (Full stop
removed intentionally, because the original does not use one either.)

Signed-off-by: Matthias Rüster <matthias.ruester@gmail.com>
2 years agoMerge branch 'po-id' of github.com:bagasme/git-po
Jiang Xin [Thu, 20 Jan 2022 02:40:08 +0000 (10:40 +0800)] 
Merge branch 'po-id' of github.com:bagasme/git-po

* 'po-id' of github.com:bagasme/git-po:
  l10n: po-id for 2.35 (round 2)

2 years agoGit 2.35-rc2 v2.35.0-rc2
Junio C Hamano [Wed, 19 Jan 2022 20:48:46 +0000 (12:48 -0800)] 
Git 2.35-rc2

Signed-off-by: Junio C Hamano <gitster@pobox.com>
2 years agogetcwd(mingw): handle the case when there is no cwd
Johannes Schindelin [Wed, 19 Jan 2022 18:56:01 +0000 (18:56 +0000)] 
getcwd(mingw): handle the case when there is no cwd

A recent upstream topic introduced checks for certain Git commands that
prevent them from deleting the current working directory, introducing
also a regression test that ensures that commands such as `git version`
_can_ run without a current working directory.

While technically not possible on Windows via the regular Win32 API, we
do run the regression tests in an MSYS2 Bash which uses a POSIX
emulation layer (the MSYS2/Cygwin runtime) where a really evil hack
_does_ allow to delete a directory even if it is the current working
directory.

Therefore, Git needs to be prepared for a missing working directory,
even on Windows.

This issue was not noticed in upstream Git because there was no caller
that tried to discover a Git directory with a deleted current working
directory in the test suite. But in the microsoft/git fork, we do want
to run `pre-command`/`post-command` hooks for every command, even for
`git version`, which means that we make precisely such a call. The bug
is not in that `pre-command`/`post-command` feature, though, but in
`mingw_getcwd()` and needs to be addressed there.

Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2 years agol10n: po-id for 2.35 (round 2)
Bagas Sanjaya [Tue, 18 Jan 2022 08:27:07 +0000 (15:27 +0700)] 
l10n: po-id for 2.35 (round 2)

Translate following new components:

  * advice.c
  * alias.c
  * sequencer.c
  * sparse-index.c
  * builtin/sparse-checkout.c

Signed-off-by: Bagas Sanjaya <bagasdotme@gmail.com>
2 years agol10n: Update Catalan translation
Jordi Mas [Wed, 19 Jan 2022 06:50:14 +0000 (07:50 +0100)] 
l10n: Update Catalan translation

Signed-off-by: Jordi Mas <jmas@softcatala.org>
2 years agoMerge branch 'js/branch-track-inherit'
Junio C Hamano [Wed, 19 Jan 2022 00:02:23 +0000 (16:02 -0800)] 
Merge branch 'js/branch-track-inherit'

"git branch -h" incorrectly said "--track[=direct|inherit]",
implying that "--trackinherit" is a valid option, which has been
corrected.

* js/branch-track-inherit:
  branch,checkout: fix --track usage strings

2 years agoMerge branch 'jc/freebsd-without-c99-only-build'
Junio C Hamano [Wed, 19 Jan 2022 00:02:23 +0000 (16:02 -0800)] 
Merge branch 'jc/freebsd-without-c99-only-build'

FreeBSD 13.0 headers have unconditional dependency on C11 language
features, and adding -std=gnu99 to DEVELOPER_CFLAGS would just
break the developer build.

* jc/freebsd-without-c99-only-build:
  Makefile: FreeBSD cannot do C99-or-below build

2 years agobranch,checkout: fix --track usage strings
Josh Steadmon [Tue, 18 Jan 2022 20:49:46 +0000 (12:49 -0800)] 
branch,checkout: fix --track usage strings

As Ævar pointed out in [1], the use of PARSE_OPT_LITERAL_ARGHELP with a
list of allowed parameters is not recommended. Both git-branch and
git-checkout were changed in d311566 (branch: add flags and config to
inherit tracking, 2021-12-20) to use this discouraged combination for
their --track flags.

Fix this by removing PARSE_OPT_LITERAL_ARGHELP, and changing the arghelp
to simply be "mode". Users may discover allowed values in the manual
pages.

[1]: https://lore.kernel.org/git/220111.86a6g3yqf9.gmgdl@evledraar.gmail.com/

Signed-off-by: Josh Steadmon <steadmon@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2 years agoMakefile: FreeBSD cannot do C99-or-below build
Junio C Hamano [Tue, 18 Jan 2022 17:47:39 +0000 (09:47 -0800)] 
Makefile: FreeBSD cannot do C99-or-below build

In "make DEVELOPER=YesPlease" builds, we try to help developers to
catch as many potential issues as they can by using -Wall and
turning compilation warnings into errors.  In the same spirit, we
recently started adding -std=gnu99 to their CFLAGS, so that they can
notice when they accidentally used language features beyond C99.

It however turns out that FreeBSD 13.0 mistakenly uses C11 extension
in its system header files regardless of what __STDC_VERSION__ says,
which means that the platform (unless we tweak their system headers)
cannot be used for this purpose.

It seems that -std=gnu99 is only added conditionally even in today's
config.mak.dev, so it is fine if we dropped -std=gnu99 from there.
Which means that developers on FreeBSD cannot participate in vetting
use of features beyond C99, but there are developers on other
platforms who will, so it's not too bad.

We might want a more "fundamental" fix to make the platform capable
of taking -std=gnu99, like working around the use of unconditional
C11 extension in its system header files by supplying a set of
"replacement" definitions in our header files.  We chose not to
pursue such an approach for two reasons at this point:

 (1) The fix belongs to the FreeBSD project, not this project, and
     such an upstream fix may happen hopefully in a not-too-distant
     future.

 (2) Fixing such a bug in system header files and working it around
     can lead to unexpected breakages (other parts of their system
     header files may not be expecting to see and do not work well
     with our "replacement" definitions).  This close to the final
     release of this cycle, we have no time for that.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
2 years agoMerge branch 'da/rhel7-lacks-uncompress2-and-c99'
Junio C Hamano [Mon, 17 Jan 2022 23:15:59 +0000 (15:15 -0800)] 
Merge branch 'da/rhel7-lacks-uncompress2-and-c99'

Adjust build on RHEL 7 to explicitly ask C99 support and use
the fallback implementation of uncompress2 we ship.

* da/rhel7-lacks-uncompress2-and-c99:
  build: centos/RHEL 7 ships with an older gcc and zlib

2 years agol10n: vi(5195t): Update for v2.35.0 round 2
Tran Ngoc Quan [Mon, 17 Jan 2022 07:15:31 +0000 (14:15 +0700)] 
l10n: vi(5195t): Update for v2.35.0 round 2

Signed-off-by: Tran Ngoc Quan <vnwildman@gmail.com>
2 years agol10n: batch update to fix typo in branch.c
Jiang Xin [Mon, 17 Jan 2022 00:50:41 +0000 (08:50 +0800)] 
l10n: batch update to fix typo in branch.c

In git 2.35 l10n round 1, a space between two words was missing in the
message from "branch.c", and it was fixed by commit 68d924e1de (branch:
missing space fix at line 313, 2022-01-11).

Do a batch update for teams (bg, fr, id, sv, tr and zh_CN) that have
already completed their works on l10n round 1.

Signed-off-by: Jiang Xin <worldhello.net@gmail.com>
2 years agol10n: git.pot: v2.35.0 round 2 (1 new, 1 removed)
Jiang Xin [Mon, 17 Jan 2022 00:32:09 +0000 (08:32 +0800)] 
l10n: git.pot: v2.35.0 round 2 (1 new, 1 removed)

Generate po/git.pot from v2.35.0-rc1 for git v2.35.0 l10n round 2.

Signed-off-by: Jiang Xin <worldhello.net@gmail.com>
2 years agoMerge tag 'v2.35.0-rc1'
Jiang Xin [Mon, 17 Jan 2022 00:30:45 +0000 (08:30 +0800)] 
Merge tag 'v2.35.0-rc1'

Git 2.35-rc1

* tag 'v2.35.0-rc1':
  Git 2.35-rc1
  reftable tests: avoid "int" overflow, use "uint64_t"
  reftable: avoid initializing structs from structs
  t1450-fsck: exec-bit is not needed to make loose object writable
  refs API: use "failure_errno", not "errno"
  Last minute fixes before -rc1
  build: NonStop ships with an older zlib
  packfile: fix off-by-one error in decoding logic
  t/gpg: simplify test for unknown key
  branch: missing space fix at line 313
  fmt-merge-msg: prevent use-after-free with signed tags
  cache.h: drop duplicate `ensure_full_index()` declaration
  lazyload: use correct calling conventions
  fetch: fix deadlock when cleaning up lockfiles in async signals

2 years agobuild: centos/RHEL 7 ships with an older gcc and zlib
David Aguilar [Sun, 16 Jan 2022 02:05:20 +0000 (18:05 -0800)] 
build: centos/RHEL 7 ships with an older gcc and zlib

GCC 4.8.5 is the default system compiler on centos7/RHEL7.
This version requires -std=c99 to enable c99 support.

zlib 1.2.7 on centos7/rhel7 lacks uncompress2().

Signed-off-by: David Aguilar <davvid@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2 years agol10n: bg.po: Updated Bulgarian translation (5195t)
Alexander Shopov [Sat, 15 Jan 2022 22:35:45 +0000 (23:35 +0100)] 
l10n: bg.po: Updated Bulgarian translation (5195t)

Signed-off-by: Alexander Shopov <ash@kambanaria.org>
2 years agoGit 2.35-rc1 v2.35.0-rc1
Junio C Hamano [Fri, 14 Jan 2022 23:26:53 +0000 (15:26 -0800)] 
Git 2.35-rc1

Signed-off-by: Junio C Hamano <gitster@pobox.com>
2 years agoMerge branch 'js/t1450-making-it-writable-does-not-need-full-posixperm'
Junio C Hamano [Fri, 14 Jan 2022 23:25:15 +0000 (15:25 -0800)] 
Merge branch 'js/t1450-making-it-writable-does-not-need-full-posixperm'

Test fix.

* js/t1450-making-it-writable-does-not-need-full-posixperm:
  t1450-fsck: exec-bit is not needed to make loose object writable

2 years agoMerge branch 'ab/reftable-build-fixes'
Junio C Hamano [Fri, 14 Jan 2022 23:25:15 +0000 (15:25 -0800)] 
Merge branch 'ab/reftable-build-fixes'

A few portability tweaks.

* ab/reftable-build-fixes:
  reftable tests: avoid "int" overflow, use "uint64_t"
  reftable: avoid initializing structs from structs

2 years agoMerge branch 'ab/refs-errno-cleanup'
Junio C Hamano [Fri, 14 Jan 2022 23:25:14 +0000 (15:25 -0800)] 
Merge branch 'ab/refs-errno-cleanup'

A brown-paper-bag fix on top of a topic that was merged during this
cycle.

* ab/refs-errno-cleanup:
  refs API: use "failure_errno", not "errno"

2 years agopull --rebase: honor rebase.autostash when fast-forwarding
Philippe Blain [Fri, 14 Jan 2022 03:14:51 +0000 (22:14 -0500)] 
pull --rebase: honor rebase.autostash when fast-forwarding

"pull --rebase" internally uses the merge machinery when the other
history is a descendant of ours (i.e. perform fast-forward).  This
came from [1], where the discussion was started from a feature
request to do so.  It is a bit hard to read the rationale behind it
in the discussion, but it seems that it was an established fact for
everybody involved that does not even need to be mentioned that
fast-forwarding done with "rebase" was much undesirable than done
with "merge", and more importantly, the result left by "merge" is as
good as (or better than) that by "rebase".

Except for one thing.  Because "git merge" does not (and should not)
honor rebase.autostash, "git pull" needs to read it and forward it
when we use "git merge" as a (hopefully better) substitute for "git
rebase" during the fast-forwarding.  But we forgot to do so (we only
add "--[no-]autostash" to the "git merge" command when "git pull" itself
was invoked with "--[no-]autostash" command line option.

Make sure "git merge" is run with "--autostash" when
rebase.autostash is set and used to fast-forward the history on
behalf of "git rebase".  Incidentally this change also takes care of
the case where

 - "git pull --rebase" (without other command line options) is run
 - "rebase.autostash" is not set
 - The history fast-forwards

In such a case, "git merge" is run with an explicit "--no-autostash"
to prevent it from honoring merge.autostash configuration, which is
what we want.  After all, we want the "git merge" to pretend as if
it is "git rebase" while being used for this purpose.

[1] https://lore.kernel.org/git/xmqqa8cfbkeq.fsf_-_@gitster.mtv.corp.google.com/

Reported-by: Tilman Vogel <tilman.vogel@web.de>
Helped-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Philippe Blain <levraiphilippeblain@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2 years agoreftable tests: avoid "int" overflow, use "uint64_t"
Ævar Arnfjörð Bjarmason [Tue, 11 Jan 2022 16:40:23 +0000 (17:40 +0100)] 
reftable tests: avoid "int" overflow, use "uint64_t"

Change code added in 1ae2b8cda84 (reftable: add merged table view,
2021-10-07) to consistently use the "uint64_t" type. These "min" and
"max" variables get passed in the body of this function to a function
whose prototype is:

    [...] reftable_writer_set_limits([...], uint64_t min, uint64_t max

This avoids the following warning on SunCC 12.5 on
gcc211.fsffrance.org:

    "reftable/merged_test.c", line 27: warning: initializer does not fit or is out of range: 0xffffffff

Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2 years agoreftable: avoid initializing structs from structs
Han-Wen Nienhuys [Thu, 13 Jan 2022 16:55:34 +0000 (16:55 +0000)] 
reftable: avoid initializing structs from structs

Apparently, the IBM xlc compiler doesn't like this.

Signed-off-by: Han-Wen Nienhuys <hanwen@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2 years agot1450-fsck: exec-bit is not needed to make loose object writable
Johannes Sixt [Thu, 13 Jan 2022 20:28:45 +0000 (21:28 +0100)] 
t1450-fsck: exec-bit is not needed to make loose object writable

A test case wants to append stuff to a loose object file to ensure
that this kind of corruption is detected. To make a read-only loose
object file writable with chmod, it is not necessary to also make
it executable. Replace the bitmask 755 with the instruction +w to
request only the write bit and to also heed the umask. And get rid
of a POSIXPERM prerequisite, which is unnecessary for the test.

Signed-off-by: Johannes Sixt <j6t@kdbg.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2 years agorefs API: use "failure_errno", not "errno"
Ævar Arnfjörð Bjarmason [Wed, 12 Jan 2022 12:36:46 +0000 (13:36 +0100)] 
refs API: use "failure_errno", not "errno"

Fix a logic error in refs_resolve_ref_unsafe() introduced in a recent
series of mine to abstract the refs API away from errno. See
96f6623ada0 (Merge branch 'ab/refs-errno-cleanup', 2021-11-29)for that
series.

In that series introduction of "failure_errno" to
refs_resolve_ref_unsafe came in ef18119dec8 (refs API: add a version
of refs_resolve_ref_unsafe() with "errno", 2021-10-16). There we'd set
"errno = 0" immediately before refs_read_raw_ref(), and then set
"failure_errno" to "errno" if errno was non-zero afterwards.

Then in the next commit 8b72fea7e91 (refs API: make
refs_read_raw_ref() not set errno, 2021-10-16) we started expecting
"refs_read_raw_ref()" to set "failure_errno". It would do that if
refs_read_raw_ref() failed, but it wouldn't be the same errno.

So we might set the "errno" here to any arbitrary bad value, and end
up e.g. returning NULL when we meant to return the refname from
refs_resolve_ref_unsafe(), or the other way around. Instrumenting this
code will reveal cases where refs_read_raw_ref() will fail, and
"errno" and "failure_errno" will be set to different values.

In practice I haven't found a case where this scary bug changed
anything in practice. The reason for that is that we'll not care about
the actual value of "errno" here per-se, but only whether:

 1. We have an errno
 2. If it's one of ENOENT, EISDIR or ENOTDIR. See the adjacent code
    added in a1c1d8170db (refs_resolve_ref_unsafe: handle d/f
    conflicts for writes, 2017-10-06)

I.e. if we clobber "failure_errno" with "errno", but it happened to be
one of those three, and we'll clobber it with another one of the three
we were OK.

Perhaps there are cases where the difference ended up mattering, but I
haven't found them. Instrumenting the test suite to fail if "errno"
and "failure_errno" are different shows a lot of failures, checking if
they're different *and* one is but not the other is outside that list
of three "errno" values yields no failures.

But let's fix the obvious bug. We should just stop paying attention to
"errno" in refs_resolve_ref_unsafe(). In addition let's change the
partial resetting of "errno" in files_read_raw_ref() to happen just
before the "return", to ensure that any such bug will be more easily
spotted in the future.

Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2 years agol10n: zh_CN: v2.35.0 round 1
Fangyi Zhou [Tue, 11 Jan 2022 16:42:06 +0000 (16:42 +0000)] 
l10n: zh_CN: v2.35.0 round 1

- Translate new messages
- Translate the word 'cone' instead of leaving it verbatim
  (in the context of sparse checkout)
- Make translations of 'failed to' consistent

Signed-off-by: Fangyi Zhou <me@fangyi.io>
Reviewed-by: Teng Long <dyroneteng@gmail.com>
Reviewed-by: 依云 <lilydjwg@gmail.com>
Reviewed-by: Jiang Xin <worldhello.net@gmail.com>
2 years agoMerge branch 'fr_2.35.0_rnd1' of github.com:jnavila/git
Jiang Xin [Thu, 13 Jan 2022 01:11:17 +0000 (09:11 +0800)] 
Merge branch 'fr_2.35.0_rnd1' of github.com:jnavila/git

* 'fr_2.35.0_rnd1' of github.com:jnavila/git:
  l10n: fr: v2.35.0 round 1

2 years agoLast minute fixes before -rc1
Junio C Hamano [Thu, 13 Jan 2022 00:27:08 +0000 (16:27 -0800)] 
Last minute fixes before -rc1

Signed-off-by: Junio C Hamano <gitster@pobox.com>
2 years agoMerge branch 'ps/lockfile-cleanup-fix'
Junio C Hamano [Wed, 12 Jan 2022 23:11:43 +0000 (15:11 -0800)] 
Merge branch 'ps/lockfile-cleanup-fix'

Some lockfile code called free() in signal-death code path, which
has been corrected.

* ps/lockfile-cleanup-fix:
  fetch: fix deadlock when cleaning up lockfiles in async signals

2 years agoMerge branch 'ma/header-dup-cleanup'
Junio C Hamano [Wed, 12 Jan 2022 23:11:43 +0000 (15:11 -0800)] 
Merge branch 'ma/header-dup-cleanup'

Code clean-up.

* ma/header-dup-cleanup:
  cache.h: drop duplicate `ensure_full_index()` declaration

2 years agoMerge branch 'fs/gpg-unknown-key-test-fix'
Junio C Hamano [Wed, 12 Jan 2022 23:11:42 +0000 (15:11 -0800)] 
Merge branch 'fs/gpg-unknown-key-test-fix'

Test simplification.

* fs/gpg-unknown-key-test-fix:
  t/gpg: simplify test for unknown key

2 years agoMerge branch 'ak/protect-any-current-branch'
Junio C Hamano [Wed, 12 Jan 2022 23:11:41 +0000 (15:11 -0800)] 
Merge branch 'ak/protect-any-current-branch'

* ak/protect-any-current-branch:
  branch: missing space fix at line 313

2 years agoMerge branch 'jt/pack-header-lshift-overflow'
Junio C Hamano [Wed, 12 Jan 2022 23:11:41 +0000 (15:11 -0800)] 
Merge branch 'jt/pack-header-lshift-overflow'

* jt/pack-header-lshift-overflow:
  packfile: fix off-by-one error in decoding logic

2 years agoMerge branch 'rb/nonstop-lacks-uncompress2'
Junio C Hamano [Wed, 12 Jan 2022 23:11:41 +0000 (15:11 -0800)] 
Merge branch 'rb/nonstop-lacks-uncompress2'

* rb/nonstop-lacks-uncompress2:
  build: NonStop ships with an older zlib

2 years agoMerge branch 'ma/windows-dynload-fix'
Junio C Hamano [Wed, 12 Jan 2022 23:11:41 +0000 (15:11 -0800)] 
Merge branch 'ma/windows-dynload-fix'

Fix calling dynamically loaded functions on Windows.

* ma/windows-dynload-fix:
  lazyload: use correct calling conventions

2 years agoMerge branch 'fs/ssh-signing-key-lifetime'
Junio C Hamano [Wed, 12 Jan 2022 23:11:41 +0000 (15:11 -0800)] 
Merge branch 'fs/ssh-signing-key-lifetime'

"git merge $signed_tag" started to drop the tag message from the
default merge message it uses by accident, which has been corrected.

* fs/ssh-signing-key-lifetime:
  fmt-merge-msg: prevent use-after-free with signed tags

2 years agobuild: NonStop ships with an older zlib
Randall S. Becker [Tue, 11 Jan 2022 03:51:39 +0000 (22:51 -0500)] 
build: NonStop ships with an older zlib

Notably, it lacks uncompress2(); use the fallback we ship in our
tree instead.

Signed-off-by: Randall S. Becker <rsbecker@nexbridge.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2 years agopackfile: fix off-by-one error in decoding logic
Junio C Hamano [Wed, 12 Jan 2022 20:11:42 +0000 (12:11 -0800)] 
packfile: fix off-by-one error in decoding logic

shift count being exactly at 7-bit smaller than the long is OK; on
32-bit architecture, shift count starts at 4 and goes through 11, 18
and 25, at which point the guard triggers one iteration too early.

Reported-by: Marc Strapetz <marc.strapetz@syntevo.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2 years agol10n: fr: v2.35.0 round 1
Jean-Noël Avila [Wed, 12 Jan 2022 20:14:45 +0000 (21:14 +0100)] 
l10n: fr: v2.35.0 round 1

Signed-off-by: Jean-Noël Avila <jn.avila@free.fr>
2 years agot/gpg: simplify test for unknown key
Fabian Stelzer [Wed, 12 Jan 2022 12:07:57 +0000 (13:07 +0100)] 
t/gpg: simplify test for unknown key

To test for a key that is completely unknown to the keyring we need one
to sign the commit with. This was done by generating a new key and not
add it into the keyring. To avoid the key generation overhead and
problems where GPG did hang in CI during it, switch GNUPGHOME to the
empty $GNUPGHOME_NOT_USED instead, therefore making all used keys unknown
for this single `verify-commit` call.

Reported-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Fabian Stelzer <fs@gigacodes.de>
Reviewed-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2 years agobranch: missing space fix at line 313
Bagas Sanjaya [Tue, 11 Jan 2022 12:36:27 +0000 (19:36 +0700)] 
branch: missing space fix at line 313

The message introduced by commit 593a2a5d06 (branch: protect branches
checked out in all worktrees, 2021-12-01) is missing a space in the
first line, add it.

Signed-off-by: Bagas Sanjaya <bagasdotme@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2 years agol10n: zh_TW: v2.35.0 round 1 (1 fuzzy)
Yi-Jyun Pan [Wed, 12 Jan 2022 17:10:46 +0000 (01:10 +0800)] 
l10n: zh_TW: v2.35.0 round 1 (1 fuzzy)

Signed-off-by: Yi-Jyun Pan <pan93412@gmail.com>
2 years agol10n: po-id for 2.35 (round 1)
Bagas Sanjaya [Tue, 11 Jan 2022 12:21:51 +0000 (19:21 +0700)] 
l10n: po-id for 2.35 (round 1)

Update following components:
  * apply.c
  * branch.c
  * builtin/add.c
  * builtin/am.c
  * builtin/fetch.c
  * builtin/ls-files.c
  * builtin/stash.c

Translate following new components:
  * archive-tar.c
  * archive-zip.c

Also clean up obsolete translations.

Signed-off-by: Bagas Sanjaya <bagasdotme@gmail.com>
2 years agoMerge branch 'master' of github.com:nafmo/git-l10n-sv
Jiang Xin [Wed, 12 Jan 2022 05:46:21 +0000 (13:46 +0800)] 
Merge branch 'master' of github.com:nafmo/git-l10n-sv

* 'master' of github.com:nafmo/git-l10n-sv:
  l10n: sv.po: Update Swedish translation (5196t0f0u)
  l10n: sv.po: Fix typo

2 years agol10n: sv.po: Update Swedish translation (5196t0f0u)
Peter Krefting [Tue, 11 Jan 2022 15:36:26 +0000 (16:36 +0100)] 
l10n: sv.po: Update Swedish translation (5196t0f0u)

Includes extra translation for branch.c:313 which has an
upstream whitespace error that is expected to be fixed.

Signed-off-by: Peter Krefting <peter@softwolves.pp.se>
2 years agol10n: sv.po: Fix typo
Peter Krefting [Sat, 8 Jan 2022 20:11:34 +0000 (21:11 +0100)] 
l10n: sv.po: Fix typo

Suggested-by: Anders Jonsson <anders.jonsson@norsjovallen.se>
Signed-off-by: Peter Krefting <peter@softwolves.pp.se>
2 years agol10n: tr: v2.35.0 round 1
Emir SARI [Tue, 11 Jan 2022 10:08:26 +0000 (13:08 +0300)] 
l10n: tr: v2.35.0 round 1

Signed-off-by: Emir SARI <emir_sari@icloud.com>
2 years agol10n: git.pot: v2.35.0 round 1 (126 new, 142 removed)
Jiang Xin [Tue, 11 Jan 2022 01:52:27 +0000 (09:52 +0800)] 
l10n: git.pot: v2.35.0 round 1 (126 new, 142 removed)

Generate po/git.pot from v2.35.0-rc0 for git v2.35.0 l10n round 1.

Signed-off-by: Jiang Xin <worldhello.net@gmail.com>
2 years agofmt-merge-msg: prevent use-after-free with signed tags
Taylor Blau [Mon, 10 Jan 2022 21:19:06 +0000 (16:19 -0500)] 
fmt-merge-msg: prevent use-after-free with signed tags

When merging a signed tag, fmt_merge_msg_sigs() is responsible for
populating the body of the merge message with the names of the signed
tags, their signatures, and the validity of those signatures.

In 02769437e1 (ssh signing: use sigc struct to pass payload,
2021-12-09), check_signature() was taught to pass the object payload via
the sigc struct instead of passing the payload buffer separately.

In effect, 02769437e1 causes buf, and sigc.payload to point at the same
region in memory. This causes a problem for fmt_tag_signature(), which
wants to read from this location, since it is freed beforehand by
signature_check_clear() (which frees it via sigc's `payload` member).

That makes the subsequent use in fmt_tag_signature() a use-after-free.

As a result, merge messages did not contain the body of any signed tags.
Luckily, they tend not to contain garbage, either, since the result of
strstr()-ing the object buffer in fmt_tag_signature() is guarded:

    const char *tag_body = strstr(buf, "\n\n");
    if (tag_body) {
      tag_body += 2;
      strbuf_add(tagbuf, tag_body, buf + len - tag_body);
    }

Unfortunately, the tests in t6200 did not catch this at the time because
they do not search for the body of signed tags in fmt-merge-msg's
output.

Resolve this by waiting to call signature_check_clear() until after its
contents can be safely discarded. Harden ourselves against any future
regressions in this area by making sure we can find signed tag messages
in the output of fmt-merge-msg, too.

Reported-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2 years agoGit 2.35-rc0 v2.35.0-rc0
Junio C Hamano [Mon, 10 Jan 2022 19:28:51 +0000 (11:28 -0800)] 
Git 2.35-rc0

Signed-off-by: Junio C Hamano <gitster@pobox.com>
2 years agoMerge branch 'en/stash-df-fix'
Junio C Hamano [Mon, 10 Jan 2022 19:52:57 +0000 (11:52 -0800)] 
Merge branch 'en/stash-df-fix'

"git stash apply" forgot to attempt restoring untracked files when
it failed to restore changes to tracked ones.

* en/stash-df-fix:
  stash: do not return before restoring untracked files

2 years agoMerge branch 'ms/t-readme-typofix'
Junio C Hamano [Mon, 10 Jan 2022 19:52:56 +0000 (11:52 -0800)] 
Merge branch 'ms/t-readme-typofix'

Typofix.

* ms/t-readme-typofix:
  t/README: fix typo

2 years agoMerge branch 'ja/i18n-similar-messages'
Junio C Hamano [Mon, 10 Jan 2022 19:52:56 +0000 (11:52 -0800)] 
Merge branch 'ja/i18n-similar-messages'

Similar message templates have been consolidated so that
translators need to work on fewer number of messages.

* ja/i18n-similar-messages:
  i18n: turn even more messages into "cannot be used together" ones
  i18n: ref-filter: factorize "%(foo) atom used without %(bar) atom"
  i18n: factorize "--foo outside a repository"
  i18n: refactor "unrecognized %(foo) argument" strings
  i18n: factorize "no directory given for --foo"
  i18n: factorize "--foo requires --bar" and the like
  i18n: tag.c factorize i18n strings
  i18n: standardize "cannot open" and "cannot read"
  i18n: turn "options are incompatible" into "cannot be used together"
  i18n: refactor "%s, %s and %s are mutually exclusive"
  i18n: refactor "foo and bar are mutually exclusive"

2 years agoMerge branch 'en/merge-ort-renorm-with-rename-delete-conflict-fix'
Junio C Hamano [Mon, 10 Jan 2022 19:52:55 +0000 (11:52 -0800)] 
Merge branch 'en/merge-ort-renorm-with-rename-delete-conflict-fix'

A corner case bug in the ort merge strategy has been corrected.

* en/merge-ort-renorm-with-rename-delete-conflict-fix:
  merge-ort: fix bug with renormalization and rename/delete conflicts

2 years agoMerge branch 'jc/doc-submitting-patches-choice-of-base'
Junio C Hamano [Mon, 10 Jan 2022 19:52:55 +0000 (11:52 -0800)] 
Merge branch 'jc/doc-submitting-patches-choice-of-base'

Extend the guidance to choose the base commit to build your work
on, and hint/nudge contributors to read others' changes.

* jc/doc-submitting-patches-choice-of-base:
  SubmittingPatchs: clarify choice of base and testing

2 years agoMerge branch 'jl/subtree-check-parents-argument-passing-fix'
Junio C Hamano [Mon, 10 Jan 2022 19:52:54 +0000 (11:52 -0800)] 
Merge branch 'jl/subtree-check-parents-argument-passing-fix'

Fix performance-releated bug in "git subtree" (in contrib/).

* jl/subtree-check-parents-argument-passing-fix:
  subtree: fix argument handling in check_parents

2 years agoMerge branch 'lh/use-gnu-color-in-grep'
Junio C Hamano [Mon, 10 Jan 2022 19:52:54 +0000 (11:52 -0800)] 
Merge branch 'lh/use-gnu-color-in-grep'

The color palette used by "git grep" has been updated to match that
of GNU grep.

* lh/use-gnu-color-in-grep:
  grep: align default colors with GNU grep ones

2 years agoMerge branch 'js/branch-track-inherit'
Junio C Hamano [Mon, 10 Jan 2022 19:52:54 +0000 (11:52 -0800)] 
Merge branch 'js/branch-track-inherit'

"git -c branch.autosetupmerge=inherit branch new old" makes "new"
to have the same upstream as the "old" branch, instead of marking
"old" itself as its upstream.

* js/branch-track-inherit:
  config: require lowercase for branch.*.autosetupmerge
  branch: add flags and config to inherit tracking
  branch: accept multiple upstream branches for tracking

2 years agoMerge branch 'ab/usage-die-message'
Junio C Hamano [Mon, 10 Jan 2022 19:52:53 +0000 (11:52 -0800)] 
Merge branch 'ab/usage-die-message'

Code clean-up to hide vreportf() from public API.

* ab/usage-die-message:
  config API: use get_error_routine(), not vreportf()
  usage.c + gc: add and use a die_message_errno()
  gc: return from cmd_gc(), don't call exit()
  usage.c API users: use die_message() for error() + exit 128
  usage.c API users: use die_message() for "fatal :" + exit 128
  usage.c: add a die_message() routine

2 years agoMerge branch 'jz/apply-3-corner-cases'
Junio C Hamano [Mon, 10 Jan 2022 19:52:53 +0000 (11:52 -0800)] 
Merge branch 'jz/apply-3-corner-cases'

"git apply --3way" bypasses the attempt to do a three-way
application in more cases to address the regression caused by the
recent change to use direct application as a fallback.

* jz/apply-3-corner-cases:
  git-apply: skip threeway in add / rename cases

2 years agoMerge branch 'hn/reftable-fixes'
Junio C Hamano [Mon, 10 Jan 2022 19:52:52 +0000 (11:52 -0800)] 
Merge branch 'hn/reftable-fixes'

Assorted fixlets in reftable code.

* hn/reftable-fixes:
  reftable: support preset file mode for writing
  reftable: signal overflow
  reftable: fix typo in header

2 years agoMerge branch 'ab/reflog-prep'
Junio C Hamano [Mon, 10 Jan 2022 19:52:52 +0000 (11:52 -0800)] 
Merge branch 'ab/reflog-prep'

Code refactoring in the reflog part of refs API.

* ab/reflog-prep:
  reflog + refs-backend: move "verbose" out of the backend
  refs files-backend: assume cb->newlog if !EXPIRE_REFLOGS_DRY_RUN
  reflog: reduce scope of "struct rev_info"
  reflog expire: don't use lookup_commit_reference_gently()
  reflog expire: refactor & use "tip_commit" only for UE_NORMAL
  reflog expire: use "switch" over enum values
  reflog: change one->many worktree->refnames to use a string_list
  reflog expire: narrow scope of "cb" in cmd_reflog_expire()
  reflog delete: narrow scope of "cmd" passed to count_reflog_ent()

2 years agoMerge branch 'ab/do-not-limit-stash-help-to-push'
Junio C Hamano [Mon, 10 Jan 2022 19:52:52 +0000 (11:52 -0800)] 
Merge branch 'ab/do-not-limit-stash-help-to-push'

"git stash" by default triggers its "push" action, but its
implementation also made "git stash -h" to show short help only for
"git stash push", which has been corrected.

* ab/do-not-limit-stash-help-to-push:
  stash: don't show "git stash push" usage on bad "git stash" usage

2 years agoMerge branch 'ab/makefile-hook-list-dependency-fix'
Junio C Hamano [Mon, 10 Jan 2022 19:52:52 +0000 (11:52 -0800)] 
Merge branch 'ab/makefile-hook-list-dependency-fix'

Fix dependency rules to generate hook-list.h header file.

* ab/makefile-hook-list-dependency-fix:
  Makefile: correct the dependency graph of hook-list.h

2 years agoMerge branch 'ab/makefile-pager-env-is-used-only-by-pager.c'
Junio C Hamano [Mon, 10 Jan 2022 19:52:51 +0000 (11:52 -0800)] 
Merge branch 'ab/makefile-pager-env-is-used-only-by-pager.c'

* ab/makefile-pager-env-is-used-only-by-pager.c:
  Makefile: move -DPAGER_ENV from BASIC_CFLAGS to EXTRA_CPPFLAGS

2 years agoMerge branch 'ab/makefile-msgfmt-wo-stats'
Junio C Hamano [Mon, 10 Jan 2022 19:52:51 +0000 (11:52 -0800)] 
Merge branch 'ab/makefile-msgfmt-wo-stats'

Make the recipe that runs msgfmt less noisy.

* ab/makefile-msgfmt-wo-stats:
  Makefile: don't invoke msgfmt with --statistics

2 years agoMerge branch 'hn/refs-debug-update'
Junio C Hamano [Mon, 10 Jan 2022 19:52:51 +0000 (11:52 -0800)] 
Merge branch 'hn/refs-debug-update'

Debugging support for refs API.

* hn/refs-debug-update:
  refs: centralize initialization of the base ref_store.
  refs: print error message in debug output
  refs: pass gitdir to packed_ref_store_create

2 years agoMerge branch 'ds/fetch-pull-with-sparse-index'
Junio C Hamano [Mon, 10 Jan 2022 19:52:50 +0000 (11:52 -0800)] 
Merge branch 'ds/fetch-pull-with-sparse-index'

"git fetch" and "git pull" are now declared sparse-index clean.
Also "git ls-files" learns the "--sparse" option to help debugging.

* ds/fetch-pull-with-sparse-index:
  test-read-cache: remove --table, --expand options
  t1091/t3705: remove 'test-tool read-cache --table'
  t1092: replace 'read-cache --table' with 'ls-files --sparse'
  ls-files: add --sparse option
  fetch/pull: use the sparse index

2 years agoMerge branch 'hn/ref-api-tests-update'
Junio C Hamano [Mon, 10 Jan 2022 19:52:50 +0000 (11:52 -0800)] 
Merge branch 'hn/ref-api-tests-update'

Test updates.

* hn/ref-api-tests-update:
  t7004: use "test-tool ref-store" for reflog inspection
  t7004: create separate tags for different tests
  t5550: require REFFILES
  t5540: require REFFILES

2 years agoMerge branch 'jh/p4-remove-unused'
Junio C Hamano [Mon, 10 Jan 2022 19:52:50 +0000 (11:52 -0800)] 
Merge branch 'jh/p4-remove-unused'

Remove a few commands from "git p4" that aren't very useful.

* jh/p4-remove-unused:
  git-p4: remove "rollback" verb
  git-p4: remove "debug" verb

2 years agoMerge branch 'ja/perf-use-specified-shell'
Junio C Hamano [Mon, 10 Jan 2022 19:52:50 +0000 (11:52 -0800)] 
Merge branch 'ja/perf-use-specified-shell'

Perf tests were run with end-user's shell, but it has been
corrected to use the shell specified by $TEST_SHELL_PATH.

* ja/perf-use-specified-shell:
  t/perf: do not run tests in user's $SHELL

2 years agoMerge branch 'hn/test-ref-store-show-hash-algo'
Junio C Hamano [Mon, 10 Jan 2022 19:52:50 +0000 (11:52 -0800)] 
Merge branch 'hn/test-ref-store-show-hash-algo'

Debugging support for refs API.

* hn/test-ref-store-show-hash-algo:
  test-ref-store: print hash algorithm

2 years agoMerge branch 'ws/fast-export-with-revision-options'
Junio C Hamano [Mon, 10 Jan 2022 19:52:50 +0000 (11:52 -0800)] 
Merge branch 'ws/fast-export-with-revision-options'

Use of certain "git rev-list" options with "git fast-export"
created nonsense results (the worst two of which being "--reverse"
and "--invert-grep --grep=<foo>").  The use of "--first-parent" is
made to behave a bit more sensible than before.

* ws/fast-export-with-revision-options:
  fast-export: fix surprising behavior with --first-parent

2 years agoMerge branch 'js/l10n-mention-ngettext-early-in-readme'
Junio C Hamano [Mon, 10 Jan 2022 19:52:49 +0000 (11:52 -0800)] 
Merge branch 'js/l10n-mention-ngettext-early-in-readme'

Localization doc update.

* js/l10n-mention-ngettext-early-in-readme:
  l10n: README: call more attention to plural strings

2 years agoMerge branch 'jh/p4-human-unit-numbers'
Junio C Hamano [Mon, 10 Jan 2022 19:52:49 +0000 (11:52 -0800)] 
Merge branch 'jh/p4-human-unit-numbers'

The way "git p4" shows file sizes in its output has been updated to
use human-readable units.

* jh/p4-human-unit-numbers:
  git-p4: show progress as an integer
  git-p4: print size values in appropriate units