git.git
10 years agoUpdate RelNotes to 2.6 to describe leftover bits since -rc2
Junio C Hamano [Mon, 14 Sep 2015 22:00:41 +0000 (15:00 -0700)]
Update RelNotes to 2.6 to describe leftover bits since -rc2

Signed-off-by: Junio C Hamano <redacted>
10 years agoMerge branch 'js/maint-am-skip-performance-regression'
Junio C Hamano [Mon, 14 Sep 2015 21:59:07 +0000 (14:59 -0700)]
Merge branch 'js/maint-am-skip-performance-regression'

Recent versions of scripted "git am" has a performance regression in
"git am --skip" codepath, which no longer exists in the built-in
version on the 'master' front.  Fix the regression in the last
scripted version that appear in 2.5.x maintenance track and older.

* js/maint-am-skip-performance-regression:
  am --skip/--abort: merge HEAD/ORIG_HEAD tree into index

10 years agoMerge branch 'ah/show-ref-usage-string'
Junio C Hamano [Mon, 14 Sep 2015 21:59:06 +0000 (14:59 -0700)]
Merge branch 'ah/show-ref-usage-string'

Both "git show-ref -h" and "git show-ref --help" illustrated that the
"--exclude-existing" option makes the command read list of refs
from its standard input.  Change only the "show-ref -h" output to
have a pair of "<>" around the placeholder that designate an input
file, i.e. "git show-ref --exclude-existing < <ref-list>".

* ah/show-ref-usage-string:
  show-ref: place angle brackets around variables in usage string

10 years agoMerge branch 'sg/help-group'
Junio C Hamano [Mon, 14 Sep 2015 21:59:05 +0000 (14:59 -0700)]
Merge branch 'sg/help-group'

* sg/help-group:
  Makefile: use SHELL_PATH when running generate-cmdlist.sh

10 years agoMerge branch 'rt/help-strings-fix'
Junio C Hamano [Mon, 14 Sep 2015 21:59:04 +0000 (14:59 -0700)]
Merge branch 'rt/help-strings-fix'

* rt/help-strings-fix:
  tag, update-ref: improve description of option "create-reflog"
  pull: don't mark values for option "rebase" for translation

10 years agoGit 2.6-rc2
Junio C Hamano [Mon, 14 Sep 2015 20:17:56 +0000 (13:17 -0700)]
Git 2.6-rc2

Signed-off-by: Junio C Hamano <redacted>
10 years agopoll: honor the timeout on Win32
Edward Thomson [Sat, 12 Sep 2015 17:50:26 +0000 (17:50 +0000)]
poll: honor the timeout on Win32

Ensure that when passing a pipe, the gnulib poll replacement will not
return 0 before the timeout has passed.

Not obeying the timeout (and merely returning 0) causes pathological
behavior when preparing a packfile for a repository and taking a
long time to do so.  If poll were to return 0 immediately, this would
cause keep-alives to get sent as quickly as possible until the packfile
was created.  Such deviance from the standard would cause megabytes (or
more) of keep-alive packets to be sent.

GetTickCount is used as it is efficient, stable and monotonically
increasing.  (Neither GetSystemTime nor QueryPerformanceCounter have
all three of these properties.)

Signed-off-by: Edward Thomson <redacted>
Signed-off-by: Johannes Schindelin <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agodoc: show usage of branch description
Philip Oakley [Mon, 14 Sep 2015 14:10:53 +0000 (15:10 +0100)]
doc: show usage of branch description

The branch description will be included in 'git format-patch
--cover-letter' and in 'git pull-request' emails. It can also
be used in the automatic merge message. Tell the reader.

While here, clarify that the description may be a multi-line
explanation of the purpose of the branch's patch series.

Signed-off-by: Philip Oakley <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agoMerge git://ozlabs.org/~paulus/gitk
Junio C Hamano [Mon, 14 Sep 2015 18:50:21 +0000 (11:50 -0700)]
Merge git://ozlabs.org/~paulus/gitk

* git://ozlabs.org/~paulus/gitk:
  gitk: Accelerators for the main menu
  gitk: Adjust the menu line numbers to compensate for the new entry
  gitk: Add a "Copy commit summary" command
  gitk: Update Bulgarian translation (307t)
  gitk: Update .po files
  gitk: Update Bulgarian translation (304t)
  gitk: Use translated version of "Command line" in getcommitlines
  gitk: Make it easier to go quickly to a specific commit
  gitk: Show the current view's name in the window title
  gitk: Add mouse right-click options to copy path and branch name
  gitk: Remove mc parameter from proc show_error
  gitk: Fix error when changing colors after closing "List references" window
  gitk: Replace catch {unset foo} with unset -nocomplain foo
  gitk: Rearrange window title to be more conventional
  gitk: sv.po: Update Swedish translation (305t0f0u)
  gitk: Fix bad English grammar "Matches none Commit Info"

10 years agoMerge branch 'jk/pack-protocol-doc'
Junio C Hamano [Mon, 14 Sep 2015 18:46:59 +0000 (11:46 -0700)]
Merge branch 'jk/pack-protocol-doc'

Streamline documentation of the pkt-line protocol.

* jk/pack-protocol-doc:
  pack-protocol: clarify LF-handling in PKT-LINE()

10 years agoMerge branch 'mp/t7060-diff-index-test'
Junio C Hamano [Mon, 14 Sep 2015 18:46:28 +0000 (11:46 -0700)]
Merge branch 'mp/t7060-diff-index-test'

Fix an old test that was doing the same thing as another one.

* mp/t7060-diff-index-test:
  t7060: actually test "git diff-index --cached -M"

10 years agoMerge branch 'gb/apply-comment-typofix'
Junio C Hamano [Mon, 14 Sep 2015 18:44:43 +0000 (11:44 -0700)]
Merge branch 'gb/apply-comment-typofix'

* gb/apply-comment-typofix:
  apply: comment grammar fix

10 years agopath: implement common_dir handling in git_pathdup_submodule()
Max Kirillov [Sun, 13 Sep 2015 22:17:42 +0000 (01:17 +0300)]
path: implement common_dir handling in git_pathdup_submodule()

When submodule is a linked worktree, "git diff --submodule" and other
calls which directly access the submodule's object database do not correctly
calculate its path. Fix it by changing the git_pathdup_submodule() behavior,
to use either common or per-worktree directory.

Do it similarly as for parent repository, but ignore the GIT_COMMON_DIR
environment variable, because it would mean common directory for the parent
repository and does not make sense for submodule.

Also add test for functionality which uses this call.

Signed-off-by: Max Kirillov <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agosubmodule refactor: use strbuf_git_path_submodule() in add_submodule_odb()
Max Kirillov [Sun, 13 Sep 2015 22:17:41 +0000 (01:17 +0300)]
submodule refactor: use strbuf_git_path_submodule() in add_submodule_odb()

Functions which directly operate submodule's object database do not
handle the case when the submodule is linked worktree (which are
introduced in c7b3a3d2fe). Instead of fixing the path calculation use
already existing strbuf_git_path_submodule() function without changing
overall behaviour. Then it will be possible to modify only that function
whenever we need to change real location of submodule's repository
content.

Helped-by: Jeff King <redacted>
Signed-off-by: Max Kirillov <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agogitk: Accelerators for the main menu
Giuseppe Bilotta [Wed, 9 Sep 2015 13:20:53 +0000 (15:20 +0200)]
gitk: Accelerators for the main menu

This allows fast, keyboard-only usage of the menu (e.g. Alt+V, N to open a
new view).

Signed-off-by: Giuseppe Bilotta <redacted>
Signed-off-by: Paul Mackerras <redacted>
10 years agogitk: Adjust the menu line numbers to compensate for the new entry
Beat Bolli [Mon, 7 Sep 2015 23:16:37 +0000 (01:16 +0200)]
gitk: Adjust the menu line numbers to compensate for the new entry

Commit d835dbb9 ("gitk: Add a "Copy commit summary" command",
2015-08-13) in the upstream gitk repo added a new context menu entry.
Therefore, the line numbers of the entries below the new one need to be
adjusted when their text or state is changed.

Signed-off-by: Beat Bolli <redacted>
Cc: Paul Mackerras <redacted>
Signed-off-by: Paul Mackerras <redacted>
10 years agoDocumentation/config: fix formatting for branch.*.rebase and pull.rebase
Andreas Schwab [Sat, 12 Sep 2015 14:26:53 +0000 (16:26 +0200)]
Documentation/config: fix formatting for branch.*.rebase and pull.rebase

Don't format the second paragraph as a literal block.

Signed-off-by: Andreas Schwab <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agotag, update-ref: improve description of option "create-reflog"
Ralf Thielow [Fri, 11 Sep 2015 16:04:13 +0000 (18:04 +0200)]
tag, update-ref: improve description of option "create-reflog"

The description of option "create-reflog" is "create_reflog", which
is neither a good description, nor a sensible string to translate.
Change it to a more meaningful message.

Signed-off-by: Ralf Thielow <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agopull: don't mark values for option "rebase" for translation
Ralf Thielow [Fri, 11 Sep 2015 15:53:17 +0000 (17:53 +0200)]
pull: don't mark values for option "rebase" for translation

"false|true|preserve" are actual values for option "rebase"
of the "git-pull" command and should therefore not be marked
for translation.

Signed-off-by: Ralf Thielow <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agoMakefile: use SHELL_PATH when running generate-cmdlist.sh
Alejandro R. Sedeño [Thu, 10 Sep 2015 23:37:07 +0000 (19:37 -0400)]
Makefile: use SHELL_PATH when running generate-cmdlist.sh

Non-POSIX shells, such as /bin/sh on SunOS, do not support $((...))
arithmetic expansion or $(...) command substitution needed by
generate-cmdlist.sh.  Make sure that we use a POSIX compliant shell
$(SHELL_PATH) when running generate-cmdlist.sh.

Signed-off-by: Alejandro R. Sedeño <redacted>
Acked-by: Eric Sunshine <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agoDocumentation/git-rebase: fix --no-autostash formatting
John Keeping [Thu, 10 Sep 2015 22:30:52 +0000 (23:30 +0100)]
Documentation/git-rebase: fix --no-autostash formatting

All of the other "--option" and "--no-option" pairs in this file are
formatted as separate options.

Signed-off-by: John Keeping <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agorebase: support --no-autostash
John Keeping [Thu, 10 Sep 2015 22:30:51 +0000 (23:30 +0100)]
rebase: support --no-autostash

This is documented as an option but we don't actually accept it.
Support it so that it is possible to override the "rebase.autostash"
config variable.

Reported-by: Daniel Hahler <redacted>
Signed-off-by: John Keeping <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agoMakefile: allow $(ARFLAGS) specified from the command line
Junio C Hamano [Thu, 10 Sep 2015 21:27:21 +0000 (14:27 -0700)]
Makefile: allow $(ARFLAGS) specified from the command line

We can do this because we have a very simple needs and run "ar"
exactly the same way everywhere ;-).

Requested-by: Jeffrey Walton
Signed-off-by: Junio C Hamano <redacted>
10 years agoMerge git://bogomips.org/git-svn
Junio C Hamano [Thu, 10 Sep 2015 21:06:58 +0000 (14:06 -0700)]
Merge git://bogomips.org/git-svn

* git://bogomips.org/git-svn:
  git-svn: parse authors file more leniently

10 years agogit-svn: parse authors file more leniently
Michael J Gruber [Thu, 10 Sep 2015 12:32:13 +0000 (14:32 +0200)]
git-svn: parse authors file more leniently

Currently, git-svn parses an authors file using the perl regex

/^(.+?|\(no author\))\s*=\s*(.+?)\s*<(.+)>\s*$/

in order to extract svn user name, real name and e-mail.
This does not match an empty e-mail field like "<>". On the other hand,
the output of an authors-prog is parsed with the perl regex

/^\s*(.+?)\s*<(.*)>\s*$/

in order to extract real name and e-mail.

So, specifying a trivial file grep such as

grep "$1" /tmp/authors | head -n 1 | cut -d'=' -f2 | cut -c'2-'

as the authors prog gives different results compared to specifying
/tmp/authors as the authors file directly.

Instead, make git svn uses the perl regex

/^(.+?|\(no author\))\s*=\s*(.+?)\s*<(.*)>\s*$/

for parsing the authors file so that the same (slightly more lenient)
regex is used in both cases.

Reported-by: Till Schäfer <redacted>
Signed-off-by: Michael J Gruber <redacted>
Signed-off-by: Eric Wong <redacted>
10 years agoMakefile: fix MAKEFLAGS tests with multiple flags
John Keeping [Sat, 5 Sep 2015 12:22:10 +0000 (13:22 +0100)]
Makefile: fix MAKEFLAGS tests with multiple flags

findstring is defined as $(findstring FIND,IN) so if multiple flags are
set these tests do the wrong thing unless $(MAKEFLAGS) is the second
argument.

Signed-off-by: John Keeping <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agoMerge branch 'master' of github.com:jiangxin/git
Jiang Xin [Thu, 10 Sep 2015 15:14:16 +0000 (23:14 +0800)]
Merge branch 'master' of github.com:jiangxin/git

* 'master' of github.com:jiangxin/git:
  l10n: zh_CN: for git v2.6.0 l10n round 1
  l10n: zh_CN: Update Git Glossary: "commit message"
  l10n: zh_CN: Update Git Glossary: pickaxe
  l10n: zh_CN: Update Git Glossary: fork
  l10n: zh_CN: Update Git Glossary: tag
  l10n: zh_CN: Update Git Glossary: "dumb", "smart"
  l10n: zh_CN: Update Git Glossary: SHA-1
  l10n: zh_CN: Add Surrounding Spaces
  l10n: zh_CN: Add translations for Git glossary
  l10n: TEAMS: stash inactive zh_CN team members
  l10n: zh_CN: Update Translation of "tag"
  l10n: zh_CN: Unify Translation of "packfile"
  l10n: zh_CN: Update Translation: "tag object"

10 years agoSync with 2.5.2
Junio C Hamano [Wed, 9 Sep 2015 21:30:35 +0000 (14:30 -0700)]
Sync with 2.5.2

10 years agoam --skip/--abort: merge HEAD/ORIG_HEAD tree into index
Johannes Schindelin [Wed, 9 Sep 2015 09:10:07 +0000 (09:10 +0000)]
am --skip/--abort: merge HEAD/ORIG_HEAD tree into index

f8da6801 (am --skip: support skipping while on unborn branch,
2015-06-06) introduced a performance regression to "git am --skip",
where it used "read-tree" to reconstruct the index from scratch
without reusing the cached stat information.

This is a backport of the corresponding patch to the builtin am in 2.6:
3ecc704 (am --skip/--abort: merge HEAD/ORIG_HEAD tree into index,
2015-08-19).

Reportedly, it can make a huge difference on Windows, in one case a `git
rebase --skip` took 1m40s without, and 5s with, this patch.

cf. https://github.com/git-for-windows/git/issues/365

Reported-and-suggested-by: Kim Gybels <redacted>
Acked-by: Paul Tan <redacted>
Signed-off-by: Johannes Schindelin <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agol10n: sv.po: Update Swedish translation (2441t0f0u)
Peter Krefting [Wed, 9 Sep 2015 20:46:35 +0000 (21:46 +0100)]
l10n: sv.po: Update Swedish translation (2441t0f0u)

Signed-off-by: Peter Krefting <redacted>
10 years agoRelease Notes: typofix
Junio C Hamano [Wed, 9 Sep 2015 17:34:35 +0000 (10:34 -0700)]
Release Notes: typofix

Thanks to Andreas Schwab for careful reading.

Signed-off-by: Junio C Hamano <redacted>
10 years agoapply: comment grammar fix
Giuseppe Bilotta [Wed, 9 Sep 2015 13:33:18 +0000 (15:33 +0200)]
apply: comment grammar fix

Signed-off-by: Giuseppe Bilotta <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agosubmodule: rewrite `module_clone` shell function in C
Stefan Beller [Tue, 8 Sep 2015 18:57:45 +0000 (11:57 -0700)]
submodule: rewrite `module_clone` shell function in C

This reimplements the helper function `module_clone` in shell
in C as `clone`. This functionality is needed for converting
`git submodule update` later on, which we want to add threading
to.

Signed-off-by: Stefan Beller <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agoGit 2.6-rc1
Junio C Hamano [Tue, 8 Sep 2015 22:38:43 +0000 (15:38 -0700)]
Git 2.6-rc1

Signed-off-by: Junio C Hamano <redacted>
10 years agoMerge branch 'jc/builtin-am-signoff-regression-fix'
Junio C Hamano [Tue, 8 Sep 2015 22:34:57 +0000 (15:34 -0700)]
Merge branch 'jc/builtin-am-signoff-regression-fix'

Recent "git am" had regression when adding a Signed-off-by line
with its "-s" option by an unintended tightening of how an existing
trailer block is detected.

* jc/builtin-am-signoff-regression-fix:
  am: match --signoff to the original scripted version

10 years agogit_connect: clarify conn->use_shell flag
Jeff King [Tue, 8 Sep 2015 08:33:14 +0000 (04:33 -0400)]
git_connect: clarify conn->use_shell flag

When executing user-specified programs, we generally always
want to use a shell, for flexibility and consistency. One
big exception is executing $GIT_SSH, which for historical
reasons must not use a shell.

Once upon a time the logic in git_connect looked like:

  if (protocol == PROTO_SSH) {
  ... setup ssh ...
  } else {
  ... setup local connection ...
  conn->use_shell = 1;
  }

But over time the PROTO_SSH block has grown, and the "local"
block has shrunk so that it contains only conn->use_shell;
it's easy to miss at the end of the large block.  Moreover,
PROTO_SSH now also sometimes sets use_shell, when the new
GIT_SSH_COMMAND is used.

Let's just set conn->use_shell when we're setting up the "conn"
struct, and unset it (with a comment) in the historical GIT_SSH
case.  This will make the flow easier to follow.

Signed-off-by: Jeff King <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agointerpret-trailers: allow running outside a repository
John Keeping [Sat, 5 Sep 2015 13:39:24 +0000 (14:39 +0100)]
interpret-trailers: allow running outside a repository

It may be useful to run git-interpret-trailers without needing to be in
a repository.

Signed-off-by: John Keeping <redacted>
Acked-by: Christian Couder <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agocontrib/subtree: respect spaces in a repository path
Alexey Shumkin [Fri, 4 Sep 2015 22:24:11 +0000 (01:24 +0300)]
contrib/subtree: respect spaces in a repository path

Remote repository may have spaces in its path, so take it into account.

Also, as far as there are no tests for the `push` command, add them.

Signed-off-by: Alexey Shumkin <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agot7900-subtree: test the "space in a subdirectory name" case
Alexey Shumkin [Fri, 4 Sep 2015 22:24:10 +0000 (01:24 +0300)]
t7900-subtree: test the "space in a subdirectory name" case

In common case there can be spaces in a subdirectory name. Change tests
accorgingly to this statement.

Also, as far as a call to the `rejoin_msg` function (in `cmd_split`)
does not take into account such a case this patch fixes commit message
when `--rejoin` option is set .

Besides, as `fixnl` and `multiline` functions did not take into account
the "new" tested "space in a subdirectory name" case they become unused
and redundant, so they are removed.

Signed-off-by: Alexey Shumkin <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agotest-lib-functions: detect test_when_finished in subshell
John Keeping [Sat, 5 Sep 2015 13:12:49 +0000 (14:12 +0100)]
test-lib-functions: detect test_when_finished in subshell

test_when_finished does nothing in a subshell because the change to
test_cleanup does not affect the parent.

There is no POSIX way to detect that we are in a subshell ($$ and $PPID
are specified to remain unchanged), but we can detect it on Bash and
fall back to ignoring the bug on other shells.

Signed-off-by: John Keeping <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agot7800: don't use test_config in a subshell
John Keeping [Sat, 5 Sep 2015 13:12:48 +0000 (14:12 +0100)]
t7800: don't use test_config in a subshell

Use the new "-C" option to test_config to change the configuration in
the submodule from the top level of the test so that it can be unset
correctly when the test finishes.

Signed-off-by: John Keeping <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agotest-lib-functions: support "test_config -C <dir> ..."
John Keeping [Sat, 5 Sep 2015 13:12:47 +0000 (14:12 +0100)]
test-lib-functions: support "test_config -C <dir> ..."

If used in a subshell, test_config cannot unset variables at the end of
a test.  This is a problem when testing submodules because we do not
want to "cd" at to top level of a test script in order to run the
command inside the submodule.

Add a "-C" option to test_config (and test_unconfig) so that test_config
can be kept outside subshells and still affect subrepositories.

Signed-off-by: John Keeping <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agot5801: don't use test_when_finished in a subshell
John Keeping [Sat, 5 Sep 2015 13:12:46 +0000 (14:12 +0100)]
t5801: don't use test_when_finished in a subshell

test_when_finished has no effect in a subshell.  Since the cmp_marks
function is only used once, inline it at its call site and move the
test_when_finished invocation to the start of the test.

Signed-off-by: John Keeping <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agot7610: don't use test_config in a subshell
John Keeping [Sat, 5 Sep 2015 13:12:45 +0000 (14:12 +0100)]
t7610: don't use test_config in a subshell

test_config uses test_when_finished to reset the configuration after the
test, but this does not work inside a subshell.  This does not cause a
problem here because the first thing the next test does is to set this
config variable itself, but we are about to add a check that will
complain when test_when_finished is used in a subshell.

In this case, "subdir" not a submodule so test_config has the same
effect when run at the top level and can simply be moved out of the
subshell.

Signed-off-by: John Keeping <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agol10n: zh_CN: for git v2.6.0 l10n round 1
Jiang Xin [Sun, 23 Aug 2015 07:23:21 +0000 (15:23 +0800)]
l10n: zh_CN: for git v2.6.0 l10n round 1

Update 123 translations (2441t0f0u) for git v2.6.0-rc0.

Signed-off-by: Jiang Xin <redacted>
Reviewed-by: Ray Chen <redacted>
10 years agopath.c: delete an extra space
Nguyễn Thái Ngọc Duy [Sat, 22 Aug 2015 00:39:58 +0000 (07:39 +0700)]
path.c: delete an extra space

Signed-off-by: Nguyễn Thái Ngọc Duy <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agoadd and use a convenience macro ce_intent_to_add()
Nguyễn Thái Ngọc Duy [Sat, 22 Aug 2015 01:08:05 +0000 (08:08 +0700)]
add and use a convenience macro ce_intent_to_add()

Signed-off-by: Nguyễn Thái Ngọc Duy <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agoam: match --signoff to the original scripted version
Junio C Hamano [Sun, 6 Sep 2015 02:56:20 +0000 (19:56 -0700)]
am: match --signoff to the original scripted version

Linus noticed that the recently reimplemented "git am -s" defines
the trailer block too rigidly, resulting in an unnecessary blank
line between the existing sign-offs and his new sign-off.  An e-mail
submission sent to Linus in real life ends with mixture of sign-offs
and commentaries, e.g.

title here

message here

Signed-off-by: Original Author <redacted>
[rv: tweaked frotz and nitfol]
Signed-off-by: Re Viewer <redacted>
Signed-off-by: Other Reviewer <redacted>
---
patch here

Because the reimplementation reused append_signoff() helper that is
used by other codepaths, which is unaware that people intermix such
comments with their sign-offs in the trailer block, such a message
was judged to end with a non-trailer, resulting in an extra blank
line before adding a new sign-off.

The original scripted version of "git am" used a lot looser
definition, i.e. "if and only if there is no line that begins with
Signed-off-by:, add a blank line before adding a new sign-off".  For
the upcoming release, stop using the append_signoff() in "git am"
and reimplement the looser definition used by the scripted version
to use only in "git am" to fix this regression in "am" while
avoiding new regressions to other users of append_signoff().

In the longer term, we should look into loosening append_signoff()
so that other codepaths that add a new sign-off behave the same way
as "git am -s", but that is a task for post-release.

Reported-by: Linus Torvalds <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agol10n: Updated Vietnamese translation (2441t)
Tran Ngoc Quan [Mon, 7 Sep 2015 01:54:05 +0000 (08:54 +0700)]
l10n: Updated Vietnamese translation (2441t)

Signed-off-by: Tran Ngoc Quan <redacted>
10 years agoMerge branch 'master' of github.com:jiangxin/git into master
Jiang Xin [Sat, 5 Sep 2015 01:23:31 +0000 (09:23 +0800)]
Merge branch 'master' of github.com:jiangxin/git into master

* 'master' of github.com:jiangxin/git:
  l10n: zh_CN: Update Git Glossary: "commit message"
  l10n: zh_CN: Update Git Glossary: pickaxe
  l10n: zh_CN: Update Git Glossary: fork
  l10n: zh_CN: Update Git Glossary: tag
  l10n: zh_CN: Update Git Glossary: "dumb", "smart"
  l10n: zh_CN: Update Git Glossary: SHA-1
  l10n: zh_CN: Add Surrounding Spaces
  l10n: zh_CN: Add translations for Git glossary
  l10n: TEAMS: stash inactive zh_CN team members
  l10n: zh_CN: Update Translation of "tag"
  l10n: zh_CN: Unify Translation of "packfile"
  l10n: zh_CN: Update Translation: "tag object"

Signed-off-by: Jiang Xin <redacted>
10 years agol10n: git.pot: v2.6.0 round 1 (123 new, 41 removed)
Jiang Xin [Sat, 5 Sep 2015 01:21:10 +0000 (09:21 +0800)]
l10n: git.pot: v2.6.0 round 1 (123 new, 41 removed)

Generate po/git.pot from v2.6.0-rc0-24-gec371ff for git v2.6.0 l10n
round 1.

Signed-off-by: Jiang Xin <redacted>
10 years agogit_connect: clear GIT_* environment for ssh
Jeff King [Fri, 4 Sep 2015 22:40:08 +0000 (18:40 -0400)]
git_connect: clear GIT_* environment for ssh

When we "switch" to another local repository to run the server
side of a fetch or push, we must clear the variables in
local_repo_env so that our local $GIT_DIR, etc, do not
pollute the upload-pack or receive-pack that is executing in
the "remote" repository.

We have never done so for ssh connections. For the most
part, nobody has noticed because ssh will not pass unknown
environment variables by default. However, it is not out of
the question for a user to configure ssh to pass along GIT_*
variables using SendEnv/AcceptEnv.

We can demonstrate the problem by using "git -c" on a local
command and seeing its impact on a remote repository.  This
config ends up in $GIT_CONFIG_PARAMETERS. In the local case,
the config has no impact, but in the ssh transport, it does
(our test script has a fake ssh that passes through all
environment variables; this isn't normal, but does simulate
one possible setup).

Signed-off-by: Jeff King <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agopager: don't use unsafe functions in signal handlers
Takashi Iwai [Fri, 4 Sep 2015 09:35:57 +0000 (11:35 +0200)]
pager: don't use unsafe functions in signal handlers

Since the commit a3da8821208d (pager: do wait_for_pager on signal
death), we call wait_for_pager() in the pager's signal handler.  The
recent bug report revealed that this causes a deadlock in glibc at
aborting "git log" [*1*].  When this happens, git process is left
unterminated, and it can't be killed by SIGTERM but only by SIGKILL.

The problem is that wait_for_pager() function does more than waiting
for pager process's termination, but it does cleanups and printing
errors.  Unfortunately, the functions that may be used in a signal
handler are very limited [*2*].  Particularly, malloc(), free() and the
variants can't be used in a signal handler because they take a mutex
internally in glibc.  This was the cause of the deadlock above.  Other
than the direct calls of malloc/free, many functions calling
malloc/free can't be used.  strerror() is such one, either.

Also the usage of fflush() and printf() in a signal handler is bad,
although it seems working so far.  In a safer side, we should avoid
them, too.

This patch tries to reduce the calls of such functions in signal
handlers.  wait_for_signal() takes a flag and avoids the unsafe
calls.   Also, finish_command_in_signal() is introduced for the
same reason.  There the free() calls are removed, and only waits for
the children without whining at errors.

[*1*] https://bugzilla.opensuse.org/show_bug.cgi?id=942297
[*2*] http://pubs.opengroup.org/onlinepubs/9699919799/functions/V2_chap02.html#tag_15_04_03

Signed-off-by: Takashi Iwai <redacted>
Reviewed-by: Jeff King <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agoSync with maint
Junio C Hamano [Fri, 4 Sep 2015 21:34:57 +0000 (14:34 -0700)]
Sync with maint

* maint:

10 years agoGit 2.5.2
Junio C Hamano [Fri, 4 Sep 2015 17:46:00 +0000 (10:46 -0700)]
Git 2.5.2

Signed-off-by: Junio C Hamano <redacted>
10 years agoSync with 2.4.9
Junio C Hamano [Fri, 4 Sep 2015 17:43:23 +0000 (10:43 -0700)]
Sync with 2.4.9

10 years agoGit 2.4.9
Junio C Hamano [Fri, 4 Sep 2015 17:36:00 +0000 (10:36 -0700)]
Git 2.4.9

Signed-off-by: Junio C Hamano <redacted>
10 years agoSync with 2.3.9
Junio C Hamano [Fri, 4 Sep 2015 17:34:19 +0000 (10:34 -0700)]
Sync with 2.3.9

10 years agoGit 2.3.9
Junio C Hamano [Fri, 4 Sep 2015 17:31:34 +0000 (10:31 -0700)]
Git 2.3.9

Signed-off-by: Junio C Hamano <redacted>
10 years agoSync with 2.2.3
Junio C Hamano [Fri, 4 Sep 2015 17:29:28 +0000 (10:29 -0700)]
Sync with 2.2.3

10 years agoGit 2.2.3
Junio C Hamano [Fri, 4 Sep 2015 17:25:47 +0000 (10:25 -0700)]
Git 2.2.3

Signed-off-by: Junio C Hamano <redacted>
10 years agoMerge branch 'jk/long-paths' into maint-2.2
Junio C Hamano [Fri, 4 Sep 2015 17:25:23 +0000 (10:25 -0700)]
Merge branch 'jk/long-paths' into maint-2.2

10 years agoshow-branch: use a strbuf for reflog descriptions
Jeff King [Wed, 19 Aug 2015 18:12:48 +0000 (14:12 -0400)]
show-branch: use a strbuf for reflog descriptions

When we show "branch@{0}", we format into a fixed-size
buffer using sprintf. This can overflow if you have long
branch names. We can fix it by using a temporary strbuf.

Signed-off-by: Jeff King <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agoread_info_alternates: handle paths larger than PATH_MAX
Jeff King [Wed, 19 Aug 2015 18:12:45 +0000 (14:12 -0400)]
read_info_alternates: handle paths larger than PATH_MAX

This function assumes that the relative_base path passed
into it is no larger than PATH_MAX, and writes into a
fixed-size buffer. However, this path may not have actually
come from the filesystem; for example, add_submodule_odb
generates a path using a strbuf and passes it in. This is
hard to trigger in practice, though, because the long
submodule directory would have to exist on disk before we
would try to open its info/alternates file.

We can easily avoid the bug, though, by simply creating the
filename on the heap.

Signed-off-by: Jeff King <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agonotes: use a strbuf in add_non_note
Jeff King [Wed, 19 Aug 2015 18:12:41 +0000 (14:12 -0400)]
notes: use a strbuf in add_non_note

When we are loading a notes tree into our internal hash
table, we also collect any files that are clearly non-notes.
We format the name of the file into a PATH_MAX buffer, but
unlike true notes (which cannot be larger than a fanned-out
sha1 hash), these tree entries can be arbitrarily long,
overflowing our buffer.

We can fix this by switching to a strbuf. It doesn't even
cost us an extra allocation, as we can simply hand ownership
of the buffer over to the non-note struct.

This is of moderate security interest, as you might fetch
notes trees from an untrusted remote. However, we do not do
so by default, so you would have to manually fetch into the
notes namespace.

Signed-off-by: Jeff King <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agoverify_absent: allow filenames longer than PATH_MAX
Jeff King [Wed, 19 Aug 2015 18:12:37 +0000 (14:12 -0400)]
verify_absent: allow filenames longer than PATH_MAX

When unpack-trees wants to know whether a path will
overwrite anything in the working tree, we use lstat() to
see if there is anything there. But if we are going to write
"foo/bar", we can't just lstat("foo/bar"); we need to look
for leading prefixes (e.g., "foo"). So we use the lstat cache
to find the length of the leading prefix, and copy the
filename up to that length into a temporary buffer (since
the original name is const, we cannot just stick a NUL in
it).

The copy we make goes into a PATH_MAX-sized buffer, which
will overflow if the prefix is longer than PATH_MAX. How
this happens is a little tricky, since in theory PATH_MAX is
the biggest path we will have read from the filesystem. But
this can happen if:

  - the compiled-in PATH_MAX does not accurately reflect
    what the filesystem is capable of

  - the leading prefix is not _quite_ what is on disk; it
    contains the next element from the name we are checking.
    So if we want to write "aaa/bbb/ccc/ddd" and "aaa/bbb"
    exists, the prefix of interest is "aaa/bbb/ccc". If
    "aaa/bbb" approaches PATH_MAX, then "ccc" can overflow
    it.

So this can be triggered, but it's hard to do. In
particular, you cannot just "git clone" a bogus repo. The
verify_absent checks happen before unpack-trees writes
anything to the filesystem, so there are never any leading
prefixes during the initial checkout, and the bug doesn't
trigger. And by definition, these files are larger than
PATH_MAX, so writing them will fail, and clone will
complain (though it may write a partial path, which will
cause a subsequent "git checkout" to hit the bug).

We can fix it by creating the temporary path on the heap.
The extra malloc overhead is not important, as we are
already making at least one stat() call (and probably more
for the prefix discovery).

Signed-off-by: Jeff King <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agoMerge branch 'ee/clean-test-fixes' into maint
Junio C Hamano [Fri, 4 Sep 2015 02:18:05 +0000 (19:18 -0700)]
Merge branch 'ee/clean-test-fixes' into maint

* ee/clean-test-fixes:
  t7300: fix broken && chains

10 years agoMerge branch 'jk/log-missing-default-HEAD' into maint
Junio C Hamano [Fri, 4 Sep 2015 02:18:03 +0000 (19:18 -0700)]
Merge branch 'jk/log-missing-default-HEAD' into maint

"git init empty && git -C empty log" said "bad default revision 'HEAD'",
which was found to be a bit confusing to new users.

* jk/log-missing-default-HEAD:
  log: diagnose empty HEAD more clearly

10 years agoMerge branch 'cc/trailers-corner-case-fix' into maint
Junio C Hamano [Fri, 4 Sep 2015 02:18:03 +0000 (19:18 -0700)]
Merge branch 'cc/trailers-corner-case-fix' into maint

The "interpret-trailers" helper mistook a multi-paragraph title of
a commit log message with a colon in it as the end of the trailer
block.

* cc/trailers-corner-case-fix:
  trailer: support multiline title
  trailer: retitle a test and correct an in-comment message
  trailer: ignore first line of message

10 years agoMerge branch 'dt/commit-preserve-base-index-upon-opportunistic-cache-tree-update...
Junio C Hamano [Fri, 4 Sep 2015 02:18:02 +0000 (19:18 -0700)]
Merge branch 'dt/commit-preserve-base-index-upon-opportunistic-cache-tree-update' into maint

When re-priming the cache-tree opportunistically while committing
the in-core index as-is, we mistakenly invalidated the in-core
index too aggressively, causing the experimental split-index code
to unnecessarily rewrite the on-disk index file(s).

* dt/commit-preserve-base-index-upon-opportunistic-cache-tree-update:
  commit: don't rewrite shared index unnecessarily

10 years agoMerge branch 'rs/archive-zip-many' into maint
Junio C Hamano [Fri, 4 Sep 2015 02:18:01 +0000 (19:18 -0700)]
Merge branch 'rs/archive-zip-many' into maint

"git archive" did not use zip64 extension when creating an archive
with more than 64k entries, which nobody should need, right ;-)?

* rs/archive-zip-many:
  archive-zip: support more than 65535 entries
  archive-zip: use a local variable to store the creator version
  t5004: test ZIP archives with many entries

10 years agoMerge branch 'jc/calloc-pathspec' into maint
Junio C Hamano [Fri, 4 Sep 2015 02:18:00 +0000 (19:18 -0700)]
Merge branch 'jc/calloc-pathspec' into maint

Minor code cleanup.

* jc/calloc-pathspec:
  ps_matched: xcalloc() takes nmemb and then element size

10 years agoMerge branch 'ss/fix-config-fd-leak' into maint
Junio C Hamano [Fri, 4 Sep 2015 02:17:59 +0000 (19:17 -0700)]
Merge branch 'ss/fix-config-fd-leak' into maint

* ss/fix-config-fd-leak:
  config: close config file handle in case of error

10 years agoMerge branch 'sg/wt-status-header-inclusion' into maint
Junio C Hamano [Fri, 4 Sep 2015 02:17:57 +0000 (19:17 -0700)]
Merge branch 'sg/wt-status-header-inclusion' into maint

* sg/wt-status-header-inclusion:
  wt-status: move #include "pathspec.h" to the header

10 years agoMerge branch 'po/po-readme' into maint
Junio C Hamano [Fri, 4 Sep 2015 02:17:56 +0000 (19:17 -0700)]
Merge branch 'po/po-readme' into maint

Doc updates for i18n.

* po/po-readme:
  po/README: Update directions for l10n contributors

10 years agoMerge branch 'sg/t3020-typofix' into maint
Junio C Hamano [Fri, 4 Sep 2015 02:17:54 +0000 (19:17 -0700)]
Merge branch 'sg/t3020-typofix' into maint

* sg/t3020-typofix:
  t3020: fix typo in test description

10 years agoMerge branch 'as/docfix-reflog-expire-unreachable' into maint
Junio C Hamano [Fri, 4 Sep 2015 02:17:53 +0000 (19:17 -0700)]
Merge branch 'as/docfix-reflog-expire-unreachable' into maint

Docfix.

* as/docfix-reflog-expire-unreachable:
  Documentation/config: fix inconsistent label on gc.*.reflogExpireUnreachable

10 years agoMerge branch 'nd/fixup-linked-gitdir' into maint
Junio C Hamano [Fri, 4 Sep 2015 02:17:53 +0000 (19:17 -0700)]
Merge branch 'nd/fixup-linked-gitdir' into maint

The code in "multiple-worktree" support that attempted to recover
from an inconsistent state updated an incorrect file.

* nd/fixup-linked-gitdir:
  setup: update the right file in multiple checkouts

10 years agoMerge branch 'jk/rev-list-has-no-notes' into maint
Junio C Hamano [Fri, 4 Sep 2015 02:17:52 +0000 (19:17 -0700)]
Merge branch 'jk/rev-list-has-no-notes' into maint

"git rev-list" does not take "--notes" option, but did not complain
when one is given.

* jk/rev-list-has-no-notes:
  rev-list: make it obvious that we do not support notes

10 years agoMerge branch 'jk/fix-alias-pager-config-key-warnings' into maint
Junio C Hamano [Fri, 4 Sep 2015 02:17:52 +0000 (19:17 -0700)]
Merge branch 'jk/fix-alias-pager-config-key-warnings' into maint

Because the configuration system does not allow "alias.0foo" and
"pager.0foo" as the configuration key, the user cannot use '0foo'
as a custom command name anyway, but "git 0foo" tried to look these
keys up and emitted useless warnings before saying '0foo is not a
git command'.  These warning messages have been squelched.

* jk/fix-alias-pager-config-key-warnings:
  config: silence warnings for command names with invalid keys

10 years agoMerge branch 'nd/dwim-wildcards-as-pathspecs' into maint
Junio C Hamano [Fri, 4 Sep 2015 02:17:51 +0000 (19:17 -0700)]
Merge branch 'nd/dwim-wildcards-as-pathspecs' into maint

Test updates for Windows.

* nd/dwim-wildcards-as-pathspecs:
  t2019: skip test requiring '*' in a file name non Windows

10 years agoMerge branch 'sg/help-group' into maint
Junio C Hamano [Fri, 4 Sep 2015 02:17:51 +0000 (19:17 -0700)]
Merge branch 'sg/help-group' into maint

We rewrote one of the build scripts in Perl but this reimplements
in Bourne shell.

* sg/help-group:
  generate-cmdlist: re-implement as shell script

10 years agoMerge branch 'ps/t1509-chroot-test-fixup' into maint
Junio C Hamano [Fri, 4 Sep 2015 02:17:50 +0000 (19:17 -0700)]
Merge branch 'ps/t1509-chroot-test-fixup' into maint

t1509 test that requires a dedicated VM environment had some
bitrot, which has been corrected.

* ps/t1509-chroot-test-fixup:
  tests: fix cleanup after tests in t1509-root-worktree
  tests: fix broken && chains in t1509-root-worktree

10 years agoMerge branch 'jh/strbuf-read-use-read-in-full' into maint
Junio C Hamano [Fri, 4 Sep 2015 02:17:50 +0000 (19:17 -0700)]
Merge branch 'jh/strbuf-read-use-read-in-full' into maint

strbuf_read() used to have one extra iteration (and an unnecessary
strbuf_grow() of 8kB), which was eliminated.

* jh/strbuf-read-use-read-in-full:
  strbuf_read(): skip unnecessary strbuf_grow() at eof

10 years agoMerge branch 'jk/long-error-messages' into maint
Junio C Hamano [Fri, 4 Sep 2015 02:17:49 +0000 (19:17 -0700)]
Merge branch 'jk/long-error-messages' into maint

The codepath to produce error messages had a hard-coded limit to
the size of the message, primarily to avoid memory allocation while
calling die().

* jk/long-error-messages:
  vreportf: avoid intermediate buffer
  vreportf: report to arbitrary filehandles

10 years agoMerge branch 'cb/open-noatime-clear-errno' into maint
Junio C Hamano [Fri, 4 Sep 2015 02:17:49 +0000 (19:17 -0700)]
Merge branch 'cb/open-noatime-clear-errno' into maint

When trying to see that an object does not exist, a state errno
leaked from our "first try to open a packfile with O_NOATIME and
then if it fails retry without it" logic on a system that refuses
O_NOATIME.  This confused us and caused us to die, saying that the
packfile is unreadable, when we should have just reported that the
object does not exist in that packfile to the caller.

* cb/open-noatime-clear-errno:
  git_open_noatime: return with errno=0 on success

10 years agoMerge branch 'mh/get-remote-group-fix' into maint
Junio C Hamano [Fri, 4 Sep 2015 02:17:47 +0000 (19:17 -0700)]
Merge branch 'mh/get-remote-group-fix' into maint

An off-by-one error made "git remote" to mishandle a remote with a
single letter nickname.

* mh/get-remote-group-fix:
  get_remote_group(): use skip_prefix()
  get_remote_group(): eliminate superfluous call to strcspn()
  get_remote_group(): rename local variable "space" to "wordlen"
  get_remote_group(): handle remotes with single-character names

10 years agot6300: add tests for "-local" date formats
John Keeping [Thu, 3 Sep 2015 21:49:01 +0000 (22:49 +0100)]
t6300: add tests for "-local" date formats

Signed-off-by: John Keeping <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agot6300: make UTC and local dates different
John Keeping [Thu, 3 Sep 2015 21:49:00 +0000 (22:49 +0100)]
t6300: make UTC and local dates different

By setting the UTC time to 23:18:43 the date in +0200 is the following
day, 2006-07-04.  This will ensure that the test for "short-local" to be
added in the following patch tests for different output from the "short"
format.

Signed-off-by: John Keeping <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agodate: make "local" orthogonal to date format
Jeff King [Thu, 3 Sep 2015 21:48:59 +0000 (22:48 +0100)]
date: make "local" orthogonal to date format

Most of our "--date" modes are about the format of the date:
which items we show and in what order. But "--date=local" is
a bit of an oddball. It means "show the date in the normal
format, but using the local timezone". The timezone we use
is orthogonal to the actual format, and there is no reason
we could not have "localized iso8601", etc.

This patch adds a "local" boolean field to "struct
date_mode", and drops the DATE_LOCAL element from the
date_mode_type enum (it's now just DATE_NORMAL plus
local=1). The new feature is accessible to users by adding
"-local" to any date mode (e.g., "iso-local"), and we retain
"local" as an alias for "default-local" for backwards
compatibility.

Signed-off-by: Jeff King <redacted>
Signed-off-by: John Keeping <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agodate: check for "local" before anything else
John Keeping [Thu, 3 Sep 2015 21:48:58 +0000 (22:48 +0100)]
date: check for "local" before anything else

In a following commit we will make "local" orthogonal to the format.
Although this will not apply to "relative", which does not use the
timezone, it applies to all other formats so move the timezone
conversion to the start of the function.

Signed-off-by: John Keeping <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agot6300: add test for "raw" date format
John Keeping [Thu, 3 Sep 2015 21:48:57 +0000 (22:48 +0100)]
t6300: add test for "raw" date format

Signed-off-by: John Keeping <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agot6300: introduce test_date() helper
John Keeping [Thu, 3 Sep 2015 21:48:56 +0000 (22:48 +0100)]
t6300: introduce test_date() helper

This moves the setup of the "expected" file inside the test case.  The
helper function has the advantage that we can use SQ in the file content
without needing to escape the quotes.

Signed-off-by: John Keeping <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agofast-import: switch crash-report date to iso8601
Jeff King [Thu, 3 Sep 2015 21:48:55 +0000 (22:48 +0100)]
fast-import: switch crash-report date to iso8601

When fast-import emits a crash report, it does so in the
user's local timezone. But because we omit the timezone
completely for DATE_LOCAL, a reader of the report does not
immediately know which time zone was used. Let's switch this
to ISO8601 instead, which includes the time zone.

This does mean we will show the time in UTC, but that's not
a big deal. A crash report like this will either be looked
at immediately (in which case nobody even looks at the
timestamp), or it will be passed along to a developer to
debug, in which case the original timezone is less likely to
be of interest.

Signed-off-by: Jeff King <redacted>
Signed-off-by: John Keeping <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agoDocumentation/rev-list: don't list date formats
John Keeping [Thu, 3 Sep 2015 21:48:54 +0000 (22:48 +0100)]
Documentation/rev-list: don't list date formats

We are about to add several new date formats which will make this list
too long to display in a single line.

Signed-off-by: John Keeping <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agoDocumentation/git-for-each-ref: don't list date formats
John Keeping [Thu, 3 Sep 2015 21:48:53 +0000 (22:48 +0100)]
Documentation/git-for-each-ref: don't list date formats

We are about to add a new set of supported date formats and do not want
to have to maintain the same list in several different bits of
documentation.  Refer to git-rev-list(1) which contains the full list of
supported formats.

Signed-off-by: John Keeping <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agoDocumentation/config: don't list date formats
John Keeping [Thu, 3 Sep 2015 21:48:52 +0000 (22:48 +0100)]
Documentation/config: don't list date formats

This list is already incomplete (missing "raw") and we're about to add
new formats.  Since this option sets a default for git-log's --date
option, just refer to git-log(1).

Signed-off-by: John Keeping <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agoDocumentation/blame-options: don't list date formats
John Keeping [Thu, 3 Sep 2015 21:48:51 +0000 (22:48 +0100)]
Documentation/blame-options: don't list date formats

This list is already incomplete (missing "raw") and we're about to add
new formats.  Remove it and refer to the canonical documentation in
git-log(1).

Signed-off-by: John Keeping <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agopack-protocol: clarify LF-handling in PKT-LINE()
Jeff King [Thu, 3 Sep 2015 08:24:09 +0000 (04:24 -0400)]
pack-protocol: clarify LF-handling in PKT-LINE()

The spec is very inconsistent about which PKT-LINE() parts
of the grammar include a LF. On top of that, the code is not
consistent, either (e.g., send-pack does not put newlines
into the ref-update commands it sends).

Let's make explicit the long-standing expectation that we
generally expect pkt-lines to end in a newline, but that
receivers should be lenient. This makes the spec consistent,
and matches what git already does (though it does not always
fulfill the SHOULD).

We do make an exception for the push-cert, where the
receiving code is currently a bit pickier. This is a
reasonable way to be, as the data needs to be byte-for-byte
compatible with what was signed. We _could_ make up some
rules about signing a canonicalized version including
newlines, but that would require a code change, and is out
of scope for this patch.

Signed-off-by: Jeff King <redacted>
Signed-off-by: Junio C Hamano <redacted>
10 years agoMerge branch 'jk/am-rerere-lock-fix'
Junio C Hamano [Thu, 3 Sep 2015 21:14:01 +0000 (14:14 -0700)]
Merge branch 'jk/am-rerere-lock-fix'

Recent "git am" introduced a double-locking failure when used with
the "--3way" option that invokes rerere machinery.

* jk/am-rerere-lock-fix:
  rerere: release lockfile in non-writing functions

git clone https://git.99rst.org/PROJECT