Signed-off-by: NBrandon Casey <casey@nrlssc.navy.mil>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

We only care about getting what should be an empty string and
sending it to a file, without a trailing LF, so the empty string
translates into a 0 byte file.  Earlier when I originally wrote
these lines Mac OS X allowed the format string of printf to be
the empty string, but more recent versions appear to have been
'improved' with error messages if the format is not given.

This may cause problems if we ever wind up with changes to the hook
tests.  A minor cleanup makes the test more safe on all systems,
by conforming to accepted printf conventions.
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

One of the first things filter-branch does is to create a temporary
directory. This directory is eventually removed by the script during
normal operation, but is not removed if the script encounters an error.

Set a trap to remove it when the script terminates for any reason.
Signed-off-by: NBrandon Casey <casey@nrlssc.navy.mil>
Acked-by: NJohannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

git-relink is intended to search for packs and loose objects in
common between two repositories and to replace the one set with
hard links to the other. Files other than packs and loose objects
should not be touched, so add the "info" sub-directory to the
pattern of directory excludes.
Signed-off-by: NBrandon Casey <casey@nrlssc.navy.mil>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NBruno Ribas <ribas@c3sl.ufpr.br>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Those two configuration variables are important enough that it is
worth to explicitely write about them in the "Gitweb config file
variables" section even if they are usually set during build by
GITWEB_PROJECTROOT and GITWEB_LIST build (Makefile) configuration
variables.
Signed-off-by: NJakub Narebski <jnareb@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NJim Meyering <meyering@redhat.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

"git reflog expire --all" opened a directory in $GIT_DIR/logs/,
read reflog files in there readdir(3), and rewrote the file by
creating a new file and renaming it back inside the loop.  This
code structure can cause the newly created file to be returned
by subsequent call to readdir(3), and fall into an infinite loop
in the worst case.

This separates the processing to two phase.  Running
for_each_reflog() to find out and collect all refs, and then
iterate over them, calling expire_reflog().  This way, the
program would behave exactly the same way as if all the refs
were given by the user from the command line.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

If the commit message, or commit author contains non-ascii, it must be
converted from Perl internal representation to utf-8, to follow what
got declared in HTTP header.  Use to_utf8() to do the conversion.

This necessarily replaces here-doc with "print" statements.
Signed-off-by: NYasushi SHOJI <yashi@atmark-techno.com>
Acked-by: Nİsmail Dönmez <ismail@pardus.org.tr>
Acked-by: NJakub Narebski <jnareb@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NChristian Couder <chriscool@tuxfamily.org>
Acked-by: NEric Wong <normalperson@yhbt.net>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NChristian Couder <chriscool@tuxfamily.org>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NChristian Couder <chriscool@tuxfamily.org>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

This accompanies c5b09feb (Avoid
update hook during git-rebase --interactive) to make sure that
any regression to make Debian's Bug#458782 (git-core: git-rebase
doesn't work when trying to squash changes into commits created
with --no-verify) resurface will be caught.
Signed-off-by: NJohannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

The exit value of some commands was not being used for the
test output.
Signed-off-by: NJeff King <peff@peff.net>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Hopefully the last rc before the final...
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

When the upstream branch is tracked, we can detect if that branch
was rebased since it was last fetched.  Teach git to use that
information to rebase from the old remote head onto the new remote head.
Signed-off-by: NJohannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Git histories may have multiple roots, which can cause
git merge-base to fail and this caused git cvsserver to die.

This commit teaches git cvsserver to handle a failing git
merge-base gracefully, and modifies the test case to verify this.
All the test cases now use a history with two roots.
Signed-off-by: NSteffen Prohaska <prohaska@zib.de>

 git-cvsserver.perl              |    9 ++++++++-
 t/t9400-git-cvsserver-server.sh |   10 +++++++++-
 2 files changed, 17 insertions(+), 2 deletions(-)
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

It is preferable to have the test setup in a test case.  The
setup itself may fail and having it as a test case handles this
situation more gracefully.
Signed-off-by: NSteffen Prohaska <prohaska@zib.de>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

While there is information about this in the configuration section, it was
missing in the options section.
Signed-off-by: NMike Hommey <mh@glandium.org>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

If options are aggregated, and that the whole token is an exact
prefix of a long option that is longer than 2 letters, reject
it.  This is to prevent a common typo:

	$ git commit -amend

to get interpreted as "commit all with message 'end'".

The typo check isn't performed if there is no aggregation,
because the stuck form is the recommended one.  If we have `-o`
being a valid short option that takes an argument, and --option
a long one, then we _MUST_ accept -option as "'o' option with
argument 'ption'", which is our official recommended form.
Signed-off-by: NPierre Habouzit <madcoder@debian.org>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NMike Hommey <mh@glandium.org>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

The -n option stands for --no-summary in git pull

[jes: reworded the description to avoid mentioning 'git-fetch';
 also exclude '-n' conditional on git-pull -- ugly because of
 the missing "else" statement in asciidoc]
Signed-off-by: NMiklos Vajna <vmiklos@frugalware.org>
Signed-off-by: NJohannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

git-svn expects its references under refs/remotes/*; but these will
not be copied or set by "git clone"; put in this man page the manual
fiddling that is required with current git-svn to get this to work.
Signed-off-by: NSam Vilain <sam.vilain@catalyst.net.nz>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Pre-POSIX.1-2001 systems don't have <sys/select.h>, but select(2)
is declared in <sys/time.h>, which git-compat-util.h includes.
Signed-off-by: NJakub Narebski <jnareb@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NRobert Schiele <rschiele@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

POSIX.1-2001 has declaration of select(2) in <sys/select.h>, but
in the previous version of SUS, it was declared in <sys/time.h>
(which is already included in git-compat-util.h).

This introduces NO_SYS_SELECT_H macro in the Makefile to be set
on older systems, to skip inclusion of <sys/select.h> that does
not exist on them.

We could check _POSIX_VERSION with 200112L and do this
automatically, but earlier it was reported that the approach
does not work well on some vintage of HP-UX.  Other systems may
get _POSIX_VERSION itself wrong.  At least for now, this manual
configuration is safer.
Signed-off-by: NRobert Schiele <rschiele@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

* git://repo.or.cz/git-gui:
  git-gui: Correctly cleanup msgfmt '1 message untranslated' output
  git-gui: Make the statistics of po2msg match those of msgfmt
  git-gui: Fallback to Tcl based po2msg.sh if msgfmt isn't available
  git-gui: Work around random missing scrollbar in revision list

In certain rare cases, the creation of the commit object
and update of HEAD can succeed, but then installing the
updated index will fail. This is most likely caused by a
full disk or exceeded disk quota. When this happens the
new index file will be removed, and the repository will
be left with the original now-out-of-sync index. The
user can recover with a "git reset HEAD" once the disk
space issue is resolved.

We should detect this failure and offer the user some
helpful guidance.
Signed-off-by: NBrandon Casey <casey@nrlssc.navy.mil>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

There is a scenario when using git clone -s and git gc --prune togother is
dangerous. Document this.
Signed-off-by: NMiklos Vajna <vmiklos@frugalware.org>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

In the multiple message case we remove the word "messages" from the
statistics output of msgfmt as it looks cleaner on the tty when you
are watching the build process.  However we failed to strip the word
"message" when only 1 message was found to be untranslated or fuzzy,
as msgfmt does not produce the 's' suffix.
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

The strings we were showing from po2msg didn't exactly match those
of msgfmt's --statistics output so we didn't show quite the same
results when building git-gui's message files.  Now we're closer
to what msgfmt shows (at least for an en_US locale) so the make
output matches.

I noticed that the fuzzy translation count is off by one for the
current po/zh_cn.po file.  Not sure why and I'm not going to try
and debug it at this time as the po2msg is strictly a fallback,
users building from source really should prefer msgfmt.
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

If msgfmt fails with exit code 127 that typically means the program
is not found in the user's PATH and thus cannot be executed by make.
In such a case we can try to fallback to the Tcl based po2msg program
that we distributed with git-gui, as it does a "good enough" job.

We still don't default to po2msg.sh however as it does not perform
a lot of the sanity checks that msgfmt does, and quite a few of
those are too useful to give up.
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

If the horizontal scrollbar isn't currently visible (because it has
not been needed) but we get an update to the scroll port we may find
the scrollbar window exists but the Tcl command doesn't.  Apparently
it is possible for Tk to have partially destroyed the scrollbar by
removing the Tcl procedure name but still leaving the widget name in
the window registry.
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

This should reduce disk space usage when doing large imports.
We'll be switching to "gc --auto" post-1.5.4 to handle
repacking for us.
Signed-off-by: NEric Wong <normalperson@yhbt.net>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NJunio C Hamano <gitster@pobox.com>

When partitioning the work amongst threads, dividing the number of
objects by the number of threads may return 0 when there are less
objects than threads; this will cause the subsequent code to segfault
when accessing list[sub_size-1].  Allow some threads to have
zero objects to work on instead of barfing, while letting others
to have more.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

The test 'creating too deep nesting' can fail even when cloning the repos,
but is not its main purpose (it has to prepare nested repos and ensure
the last one is invalid). So split the test into the creation and
invalidity checking parts.
Signed-off-by: NAlex Riesen <raa.lkml@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

This fixes the subtile bug in git send-email that was introduced into
git send-email with aa54892f (send-email:
detect invocation errors earlier), which caused no patches to be sent
out if the --compose flag was used.
Signed-off-by: NGustaf Hendeby <hendeby@isy.liu.se>
Tested-by: NSeth Falcon <seth@userprimary.net>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Junio pointed out this part of fast-import wasn't very clear on
initial read, and it took some time for someone who was new to
fast-import's "dirty little tricks" to understand how this was
even working.  So a little bit of commentary in the proper place
may help future readers.
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

We now use the configured pack.compression and pack.depth values
within fast-import, as like builtin-pack-objects fast-import is
generating a packfile for consumption by the Git tools.

We use the same behavior as builtin-pack-objects does for these
options, allowing core.compression to supply the default value
for pack.compression.

The default setting for pack.depth within fast-import is still 10
as users will generally repack fast-import generated packfiles by
`repack -f`.  A large delta depth within the fast-import packfile
can significantly slow down such a later repack.
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>