]> git.proxmox.com Git - libgit2.git/log
libgit2.git
11 years agomerge cleanup should actually cleanup and the test should actually test
Edward Thomson [Mon, 7 Jan 2013 00:38:29 +0000 (18:38 -0600)]
merge cleanup should actually cleanup and the test should actually test

11 years agoCorrect typos in documentation
Kevin Sawicki [Sun, 6 Jan 2013 18:39:35 +0000 (10:39 -0800)]
Correct typos in documentation

11 years agoRevert "Tab align value of GIT_FILEMODE_BLOB_EXECUTABLE"
Kevin Sawicki [Sun, 6 Jan 2013 18:12:05 +0000 (10:12 -0800)]
Revert "Tab align value of GIT_FILEMODE_BLOB_EXECUTABLE"

This reverts commit 47fbcbb5a8d4aad87371a381ebdadd04cde3fb2b.

11 years agoTab align value of GIT_FILEMODE_BLOB_EXECUTABLE
Kevin Sawicki [Sun, 6 Jan 2013 18:02:37 +0000 (10:02 -0800)]
Tab align value of GIT_FILEMODE_BLOB_EXECUTABLE

11 years agoMerge pull request #1198 from nulltoken/travis/run-valgrind-on-success
Vicent Martí [Sun, 6 Jan 2013 17:12:02 +0000 (09:12 -0800)]
Merge pull request #1198 from nulltoken/travis/run-valgrind-on-success

travis: Only run Valgrind when all tests pass

11 years agotravis: Only run Valgrind when all tests pass
nulltoken [Sun, 6 Jan 2013 16:21:37 +0000 (17:21 +0100)]
travis: Only run Valgrind when all tests pass

11 years agoMerge pull request #1196 from scunz/allow_clone_without_master_branch
Vicent Martí [Sun, 6 Jan 2013 15:56:45 +0000 (07:56 -0800)]
Merge pull request #1196 from scunz/allow_clone_without_master_branch

Allow to clone repositories that don't have a `master` branch

11 years agoMerge pull request #1197 from nulltoken/travis/run-online-tests
Vicent Martí [Sun, 6 Jan 2013 15:56:08 +0000 (07:56 -0800)]
Merge pull request #1197 from nulltoken/travis/run-online-tests

travis: Include the online suite when running against Travis

11 years agoreset: Fix a memory leak
nulltoken [Sun, 6 Jan 2013 14:26:54 +0000 (15:26 +0100)]
reset: Fix a memory leak

11 years agotests: Fix some memory leaks
nulltoken [Sun, 6 Jan 2013 14:18:00 +0000 (15:18 +0100)]
tests: Fix some memory leaks

11 years agoclone: Fix a memory leak
nulltoken [Sun, 6 Jan 2013 14:09:27 +0000 (15:09 +0100)]
clone: Fix a memory leak

11 years agotravis: Include the online suite when running against Travis
nulltoken [Sun, 6 Jan 2013 13:40:32 +0000 (14:40 +0100)]
travis: Include the online suite when running against Travis

11 years agoAllow to clone repositories that don't have a `master` branch
Sascha Cunz [Sun, 6 Jan 2013 13:30:52 +0000 (14:30 +0100)]
Allow to clone repositories that don't have a `master` branch

Before this, we error out from `reference_matches_remote_head` if the
reference we're searching for does not exist.
Since we explicitly check if master is existing in `update_head_to_remote`
and error out if it doesn't, a repository without master branch could
not be cloned.
In fact this was later clobbered by what is fixed in #1194.

However, this patch introduces a `found` member in `head_info` and sets
it accordingly. That also saves us from checking the string length of
`branchname` a few times.

11 years agoMerge pull request #1194 from libgit2/dont-clobber-errors
Vicent Martí [Sun, 6 Jan 2013 02:52:09 +0000 (18:52 -0800)]
Merge pull request #1194 from libgit2/dont-clobber-errors

git_mwindow_file_deregister() shouldn't return errors

11 years agogit_mwindow_file_deregister() shouldn't return errors
Scott J. Goldman [Sun, 6 Jan 2013 02:20:42 +0000 (18:20 -0800)]
git_mwindow_file_deregister() shouldn't return errors

As a function that appears to only be called on error paths, I don't
think it makes sense for it to return an error, or clobber the global
giterr. Note that no existing callsites actually check the return
code.

In my own application, there are errors where the real error ends
up being hidden, as git_mwindow_file_deregister() clobbers the
global giterr. I'm not sure this error is even relevant?

11 years agoMerge pull request #1131 from libgit2/correct-ahead-behind
Vicent Martí [Sat, 5 Jan 2013 23:44:19 +0000 (15:44 -0800)]
Merge pull request #1131 from libgit2/correct-ahead-behind

Fix an issue with ahead-behind for lopsided traversal

11 years agoMerge pull request #1193 from libgit2/no-newline-packed-refs
Vicent Martí [Sat, 5 Jan 2013 12:30:24 +0000 (04:30 -0800)]
Merge pull request #1193 from libgit2/no-newline-packed-refs

Handle packed refs with no trailing newline

11 years agoHandle packed refs with no trailing newline
Scott J. Goldman [Sat, 5 Jan 2013 08:46:39 +0000 (00:46 -0800)]
Handle packed refs with no trailing newline

I saw a repo in the wild today which had a master branch ref which was packed, but had no trailing newline. Git handled it fine, but libgit2 choked on it. Fix seems simple enough. If we don't see a newline, assume the end of the buffer is the end of the ref line.

11 years agoMerge pull request #1192 from arrbee/fix-win32-checkout-test
Vicent Martí [Sat, 5 Jan 2013 01:28:09 +0000 (17:28 -0800)]
Merge pull request #1192 from arrbee/fix-win32-checkout-test

Fix win32 checkout test

11 years agoActually fix win32 checkout test
Russell Belfer [Sat, 5 Jan 2013 01:17:37 +0000 (17:17 -0800)]
Actually fix win32 checkout test

It turns out that using REMOVE_UNTRACKED with checkout for this
particular test was causing the .gitattributes file to be removed
and so we do have to allow for the CRs in the created file...

11 years agoFix crlf issue with checkout tests
Russell Belfer [Sat, 5 Jan 2013 00:56:21 +0000 (16:56 -0800)]
Fix crlf issue with checkout tests

Move some checkout utility functions into a shared file and fix
some crlf filtering issues when verifying file contents.

11 years agoMerge pull request #1191 from maxpow4h/fix-format-string
Vicent Martí [Sat, 5 Jan 2013 00:49:46 +0000 (16:49 -0800)]
Merge pull request #1191 from maxpow4h/fix-format-string

Fixed size_t format string warning

11 years agoFixed size_t format string warning
Maxwell Swadling [Sat, 5 Jan 2013 00:34:19 +0000 (11:34 +1100)]
Fixed size_t format string warning

11 years agoMerge pull request #1118 from arrbee/fix-reset-hard
Vicent Martí [Sat, 5 Jan 2013 00:14:49 +0000 (16:14 -0800)]
Merge pull request #1118 from arrbee/fix-reset-hard

Fix checkout corner cases, index handling, and reset hard behavior

11 years agoFix checkout of index-only dirs and prefixed paths
Russell Belfer [Fri, 4 Jan 2013 00:56:27 +0000 (16:56 -0800)]
Fix checkout of index-only dirs and prefixed paths

There are a couple of checkout bugs fixed here.  One is with
untracked working directory entries that are prefixes of tree
entries but not in a meaningful way (i.e. "read" is a prefix of
"readme.txt" but doesn't interfere in any way).  The second bug
is actually a redo of 07edfa0fc640f85f95507c3101e77accd7d2bf0d
where directory entries in the index that are not in the diff
were not being removed correctly.  That fix remedied one case
but broke another.

11 years agoAdd index API to remove all files in a directory
Russell Belfer [Thu, 3 Jan 2013 23:48:52 +0000 (15:48 -0800)]
Add index API to remove all files in a directory

This adds the git_index_remove_directory API plus tests.

11 years agoFix oid tostr issue with NULL oid
Russell Belfer [Thu, 3 Jan 2013 22:21:25 +0000 (14:21 -0800)]
Fix oid tostr issue with NULL oid

I made a small change to the behavior of this code and apparently
got it wrong. Sigh.

11 years agoFix checkout bug rmv untracked trees from index
Russell Belfer [Thu, 3 Jan 2013 22:08:53 +0000 (14:08 -0800)]
Fix checkout bug rmv untracked trees from index

When checking out with the GIT_CHECKOUT_REMOVE_UNTRACKED option
and there was an entire tree in the working directory and in the
index that is not in the baseline nor target commit, the tree was
correctly(?) removed from the working directory but was not
successfully removed from the index.  This fixes that and adds a
test of the functionality.

11 years agoFix memory leak with checkout tree iterator
Russell Belfer [Thu, 3 Jan 2013 21:22:34 +0000 (13:22 -0800)]
Fix memory leak with checkout tree iterator

11 years agoFixing checkout UPDATE_ONLY and adding tests
Russell Belfer [Thu, 3 Jan 2013 18:51:18 +0000 (10:51 -0800)]
Fixing checkout UPDATE_ONLY and adding tests

This adds a bunch of new checkout tests and in the process I found
a bug in the GIT_CHECKOUT_UPDATE_ONLY flag which I fixed.

11 years agoOh yeah, bugs from my rebase
Russell Belfer [Thu, 3 Jan 2013 17:11:52 +0000 (09:11 -0800)]
Oh yeah, bugs from my rebase

11 years agoSimplify checkout documentation
Russell Belfer [Thu, 3 Jan 2013 01:14:00 +0000 (17:14 -0800)]
Simplify checkout documentation

This moves a lot of the detailed checkout documentation into a new
file (docs/checkout-internals.md) and simplifies the public docs
for the checkout API.

11 years agoClone should use GIT_CHECKOUT_SAFE_CREATE
Russell Belfer [Thu, 3 Jan 2013 01:12:45 +0000 (17:12 -0800)]
Clone should use GIT_CHECKOUT_SAFE_CREATE

For clone to work as expected, it should be using a SAFE_CREATE
checkout (i.e. create files that are missing, even if the target
tree matches the current HEAD).

11 years agoExtend tests for checkout with typechanges
Russell Belfer [Thu, 3 Jan 2013 01:10:56 +0000 (17:10 -0800)]
Extend tests for checkout with typechanges

Test a number of other cases, including intentionally forced
conflicts and deeper inspection that trees get created properly.

There is a still a bug in checkout because the first test here
(i.e. test_checkout_typechange__checkout_typechanges_safe) should
be able to pass with GIT_CHECKOUT_SAFE as a strategy, but it will
not because of some lingering submodule checkout issues.

11 years agoFix some submodule and typechange checkout cases
Russell Belfer [Thu, 3 Jan 2013 01:09:07 +0000 (17:09 -0800)]
Fix some submodule and typechange checkout cases

There were a bunch of small bugs in the checkout code where I was
assuming that a typechange was always from a tree to a blob or
vice versa.  This fixes up most of those cases.  Also, there were
circumstances where the submodule definitions were changed by the
checkout and the submodule data was not getting reloaded properly
before the new submodules were checked out.

11 years agoFix workdir notifications and removals
Russell Belfer [Thu, 3 Jan 2013 01:05:54 +0000 (17:05 -0800)]
Fix workdir notifications and removals

The notifications were broken from the various iterations over
this code and were not returning working dir item data correctly.
Also, workdir items that were alphabetically after the last item
in diff were not being processed.

11 years agoFix up spoolandsort iterator usage
Russell Belfer [Thu, 3 Jan 2013 01:01:34 +0000 (17:01 -0800)]
Fix up spoolandsort iterator usage

The spoolandsort iterator changes got sort-of cherry picked out of
this branch and so I dropped the commit when rebasing; however,
there were a few small changes that got dropped as well (since the
version merged upstream wasn't quite the same as what I dropped).

11 years agoSubmodule caching fix and location API
Russell Belfer [Fri, 21 Dec 2012 00:16:22 +0000 (16:16 -0800)]
Submodule caching fix and location API

This adds a new API to the submodule interface that just returns
where information about the submodule was found (e.g. config file
only or in the HEAD, index, or working directory).

Also, the old "refresh" call was potentially keeping some stale
submodule data around, so this simplfies that code and literally
discards the old cache, then reallocates.

11 years agoFix use of uninitialized variable
Russell Belfer [Fri, 21 Dec 2012 00:15:02 +0000 (16:15 -0800)]
Fix use of uninitialized variable

11 years agoImprove error propagation in stash
Russell Belfer [Wed, 19 Dec 2012 23:06:40 +0000 (15:06 -0800)]
Improve error propagation in stash

Stash was sometimes obscuring the actual error code, replacing it
with a -1 when there was more descriptive value.  This updates
stash to preserve the original error code more reliably along
with a variety of other error handling tweaks.

I believe this is an improvement, but arguably, preserving the
underlying error code may result in values that are harder to
interpret by the caller who does not understand the internals.
Discussion is welcome!

11 years agoMake git_oid_tostr use out buffer for NULL oid
Russell Belfer [Wed, 19 Dec 2012 23:06:14 +0000 (15:06 -0800)]
Make git_oid_tostr use out buffer for NULL oid

Previously a NULL oid was handled like an empty buffer and
returned a status empty string.  This makes git_oid_tostr()
set the output buffer to the empty string instead.

11 years agoAdd index updating to checkout
Russell Belfer [Tue, 18 Dec 2012 23:19:24 +0000 (15:19 -0800)]
Add index updating to checkout

Make checkout update entries in the index for all files that are
updated and/or removed, unless flag GIT_CHECKOUT_DONT_UPDATE_INDEX
is given.  To do this, iterators were extended to allow a little
more introspection into the index being iterated over, etc.

11 years agoMore checkout improvements
Russell Belfer [Mon, 10 Dec 2012 23:31:43 +0000 (15:31 -0800)]
More checkout improvements

This flips checkout back to be driven off the changes between
the baseline and the target trees.  This reinstates the complex
code for tracking the contents of the working directory, but
overall, I think the resulting logic is easier to follow.

11 years agoRework checkout internals (again)
Russell Belfer [Thu, 6 Dec 2012 21:36:17 +0000 (13:36 -0800)]
Rework checkout internals (again)

I've tried to map out the detailed behaviors of checkout and make
sure that we're handling the various cases correctly, along with
providing options to allow us to emulate "git checkout" and "git
checkout-index" with the various flags.  I've thrown away flags
in the checkout API that seemed like clutter and added some new
ones.  Also, I've converted the conflict callback to a general
notification callback so we can emulate "git checkout" output and
display "dirty" files.

As of this commit, the new behavior is not working 100% but some
of that is probably baked into tests that are not testing the
right thing.  This is a decent snapshot point, I think, along the
way to getting the update done.

11 years agoReorder operations in git reset
Russell Belfer [Tue, 27 Nov 2012 01:24:02 +0000 (17:24 -0800)]
Reorder operations in git reset

This corrects the order of operations in git reset so that the
checkout to reset the working directory content is done before
the HEAD is moved.  This allows us to use the HEAD and the index
content to know what files can / should safely be reset.

Unfortunately, there are still some cases where the behavior of
this revision differs from core git.  Notable, a file which has
been added to the index but is not present in the HEAD is
considered to be tracked by core git (and thus removable by a
reset command) whereas since this loads the target state into
the index prior to resetting, it will consider such a file to be
untracked and won't touch it.  That is a larger fix that I'll
defer to a future commit.

11 years agoFailing test on git_checkout_tree when removing directories
Jameson Miller [Fri, 30 Nov 2012 15:59:03 +0000 (10:59 -0500)]
Failing test on git_checkout_tree when removing directories

11 years agoAdd failing test to demonstrate wrong checkout behaviour
Marvin Gülker [Thu, 6 Dec 2012 15:12:21 +0000 (16:12 +0100)]
Add failing test to demonstrate wrong checkout behaviour

11 years agomissing error message is confusing
Russell Belfer [Tue, 18 Dec 2012 23:13:11 +0000 (15:13 -0800)]
missing error message is confusing

11 years agoClear error to avoid leaving invalid error behind
Russell Belfer [Tue, 18 Dec 2012 23:12:06 +0000 (15:12 -0800)]
Clear error to avoid leaving invalid error behind

11 years agoUpdate cl_git_pass to return more info
Russell Belfer [Wed, 19 Dec 2012 22:52:12 +0000 (14:52 -0800)]
Update cl_git_pass to return more info

This adds a failure reporting function that is called by
cl_git_pass which captures the actual error return code and
the error message if available in the failure report.

11 years agoclar: lolpython
Vicent Marti [Fri, 4 Jan 2013 19:28:33 +0000 (20:28 +0100)]
clar: lolpython

11 years agoclar: Corrupted pickles
Vicent Marti [Fri, 4 Jan 2013 19:25:10 +0000 (20:25 +0100)]
clar: Corrupted pickles

11 years agoMerge pull request #1174 from nulltoken/topic/soft_reset_with_index_conflicts
Vicent Martí [Fri, 4 Jan 2013 19:10:39 +0000 (11:10 -0800)]
Merge pull request #1174 from nulltoken/topic/soft_reset_with_index_conflicts

Prevent soft reset when index contains conflicts

11 years agoclar: make it compatible with python3
Carlos Martín Nieto [Fri, 4 Jan 2013 19:01:36 +0000 (19:01 +0000)]
clar: make it compatible with python3

11 years agoclar: haha
Vicent Marti [Fri, 4 Jan 2013 19:02:01 +0000 (20:02 +0100)]
clar: haha

11 years agoclar: fix warning on Windows
Vicent Marti [Fri, 4 Jan 2013 19:00:09 +0000 (20:00 +0100)]
clar: fix warning on Windows

11 years agoIgnore clar.suite.rule
nulltoken [Fri, 4 Jan 2013 18:54:45 +0000 (19:54 +0100)]
Ignore clar.suite.rule

11 years agoFix git_index sorting with core.ignorecase in git_index_read
Philip Kelley [Fri, 4 Jan 2013 18:48:08 +0000 (13:48 -0500)]
Fix git_index sorting with core.ignorecase in git_index_read

11 years agoreset: Cannot soft reset with a conflicted index
nulltoken [Thu, 27 Dec 2012 12:42:27 +0000 (13:42 +0100)]
reset: Cannot soft reset with a conflicted index

11 years agoMerge pull request #1189 from martinwoodward/tests-compliance
Vicent Martí [Fri, 4 Jan 2013 17:33:54 +0000 (09:33 -0800)]
Merge pull request #1189 from martinwoodward/tests-compliance

Add jGit license block to derrived tests

11 years agoAdd jGit license block to derrived tests
Martin Woodward [Fri, 4 Jan 2013 17:29:45 +0000 (17:29 +0000)]
Add jGit license block to derrived tests

Add the jGit license block to tests derrived from jGit as per the
terms of the BSD license.

11 years agoMerge branch 'clar2' into development
Vicent Marti [Fri, 4 Jan 2013 16:52:49 +0000 (17:52 +0100)]
Merge branch 'clar2' into development

Conflicts:
src/pqueue.c
src/pqueue.h
src/util.c
tests-clar/merge/setup.c

11 years agoclar: fix merge/setup.c
Vicent Marti [Fri, 4 Jan 2013 16:46:38 +0000 (17:46 +0100)]
clar: fix merge/setup.c

11 years agoRemove whitespace
Ted Nyman [Fri, 4 Jan 2013 00:15:52 +0000 (16:15 -0800)]
Remove whitespace

11 years agoAdd note in CONVENTIONS about inlined functions
Ted Nyman [Fri, 4 Jan 2013 00:14:23 +0000 (16:14 -0800)]
Add note in CONVENTIONS about inlined functions

11 years agoAdd Brian Downing to the hall of fame
Martin Woodward [Thu, 3 Jan 2013 23:42:04 +0000 (23:42 +0000)]
Add Brian Downing to the hall of fame

11 years agoMERGE_HEAD contents iterator
Edward Thomson [Thu, 3 Jan 2013 18:44:09 +0000 (12:44 -0600)]
MERGE_HEAD contents iterator

11 years agoGive proper license notice to code from Android
Martin Woodward [Thu, 3 Jan 2013 22:28:59 +0000 (22:28 +0000)]
Give proper license notice to code from Android

The usage of the Android derrived code contains a full notice
which must be provided with the source code as per the terms
given at:
https://android.googlesource.com/platform/bionic/+/android-4.0.3_r1.1/libc/bionic/dirname_r.c

11 years agoAdd full license notice to bsearch code
Martin Woodward [Thu, 3 Jan 2013 22:24:10 +0000 (22:24 +0000)]
Add full license notice to bsearch code

The original BSD glibc code contains the notice as given at
http://opensource.apple.com/source/gcc/gcc-5666.3/libiberty/bsearch.c
and should be given in full along with the code.

11 years agoFix bug in gen_pktline() for deletes of missing remote refs
Congyi Wu [Thu, 3 Jan 2013 18:26:11 +0000 (13:26 -0500)]
Fix bug in gen_pktline() for deletes of missing remote refs

* gen_pktline() in smart_protocol.c was skipping refspecs that deleted
  refs that were not advertised by the server.  The new behavior is to
  send a delete command with an old-id of zero, which matches the behavior
  of the official git client.
* Update test_network_push__delete() in reaction to above fix.
* Obviate messy logic that handles missing push_spec rrefs by canonicalizing
  push_spec.  After calculate_work(), loid, roid, and rref, are filled in with
  exactly what is sent to the server

11 years agoAdd Apache license header back to libpqueue files
Martin Woodward [Thu, 3 Jan 2013 22:16:37 +0000 (22:16 +0000)]
Add Apache license header back to libpqueue files

The original libpqueue file were licensed under Apache 2.0 so
therefore should retain their copyrights and header as per the
license terms at http://www.apache.org/licenses/LICENSE-2.0

11 years agoexpose merge metadata cleanup
Edward Thomson [Thu, 3 Jan 2013 21:53:50 +0000 (15:53 -0600)]
expose merge metadata cleanup

11 years agoadd option to allow git note overwrite
Nikolai Vladimirov [Thu, 3 Jan 2013 14:31:36 +0000 (16:31 +0200)]
add option to allow git note overwrite

11 years agoFix git__strncasecmp
Philip Kelley [Thu, 3 Jan 2013 13:45:09 +0000 (08:45 -0500)]
Fix git__strncasecmp

11 years agonotes.c - whitespace fix
Nikolai Vladimirov [Thu, 3 Jan 2013 13:43:51 +0000 (15:43 +0200)]
notes.c - whitespace fix

11 years agoMerge pull request #1184 from ethomson/mergehead_iterator
Vicent Martí [Fri, 4 Jan 2013 16:36:13 +0000 (08:36 -0800)]
Merge pull request #1184 from ethomson/mergehead_iterator

MERGE_HEAD contents iterator

11 years agoMerge pull request #1183 from congyiwu/push_delete_fix
Philip Kelley [Fri, 4 Jan 2013 15:33:31 +0000 (07:33 -0800)]
Merge pull request #1183 from congyiwu/push_delete_fix

Fix bug in gen_pktline() for deletes of missing remote refs

11 years agoMerge pull request #1188 from tnm/inline-convention
Vicent Martí [Fri, 4 Jan 2013 11:57:31 +0000 (03:57 -0800)]
Merge pull request #1188 from tnm/inline-convention

Inline convention

11 years agoRemove whitespace
Ted Nyman [Fri, 4 Jan 2013 00:15:52 +0000 (16:15 -0800)]
Remove whitespace

11 years agoAdd note in CONVENTIONS about inlined functions
Ted Nyman [Fri, 4 Jan 2013 00:14:23 +0000 (16:14 -0800)]
Add note in CONVENTIONS about inlined functions

11 years agoMerge pull request #1187 from martinwoodward/give-props
Vicent Martí [Thu, 3 Jan 2013 23:46:23 +0000 (15:46 -0800)]
Merge pull request #1187 from martinwoodward/give-props

Add Brian Downing to the hall of fame

11 years agoAdd Brian Downing to the hall of fame
Martin Woodward [Thu, 3 Jan 2013 23:42:04 +0000 (23:42 +0000)]
Add Brian Downing to the hall of fame

11 years agoMerge pull request #1186 from martinwoodward/compliance-audit
Vicent Martí [Thu, 3 Jan 2013 22:46:45 +0000 (14:46 -0800)]
Merge pull request #1186 from martinwoodward/compliance-audit

License Compliance Audit

11 years agoMERGE_HEAD contents iterator
Edward Thomson [Thu, 3 Jan 2013 18:44:09 +0000 (12:44 -0600)]
MERGE_HEAD contents iterator

11 years agoGive proper license notice to code from Android
Martin Woodward [Thu, 3 Jan 2013 22:28:59 +0000 (22:28 +0000)]
Give proper license notice to code from Android

The usage of the Android derrived code contains a full notice
which must be provided with the source code as per the terms
given at:
https://android.googlesource.com/platform/bionic/+/android-4.0.3_r1.1/libc/bionic/dirname_r.c

11 years agoAdd full license notice to bsearch code
Martin Woodward [Thu, 3 Jan 2013 22:24:10 +0000 (22:24 +0000)]
Add full license notice to bsearch code

The original BSD glibc code contains the notice as given at
http://opensource.apple.com/source/gcc/gcc-5666.3/libiberty/bsearch.c
and should be given in full along with the code.

11 years agoFix bug in gen_pktline() for deletes of missing remote refs
Congyi Wu [Thu, 3 Jan 2013 18:26:11 +0000 (13:26 -0500)]
Fix bug in gen_pktline() for deletes of missing remote refs

* gen_pktline() in smart_protocol.c was skipping refspecs that deleted
  refs that were not advertised by the server.  The new behavior is to
  send a delete command with an old-id of zero, which matches the behavior
  of the official git client.
* Update test_network_push__delete() in reaction to above fix.
* Obviate messy logic that handles missing push_spec rrefs by canonicalizing
  push_spec.  After calculate_work(), loid, roid, and rref, are filled in with
  exactly what is sent to the server

11 years agoAdd Apache license header back to libpqueue files
Martin Woodward [Thu, 3 Jan 2013 22:16:37 +0000 (22:16 +0000)]
Add Apache license header back to libpqueue files

The original libpqueue file were licensed under Apache 2.0 so
therefore should retain their copyrights and header as per the
license terms at http://www.apache.org/licenses/LICENSE-2.0

11 years agoMerge pull request #1185 from ethomson/cleanup_merge
Vicent Martí [Thu, 3 Jan 2013 21:57:17 +0000 (13:57 -0800)]
Merge pull request #1185 from ethomson/cleanup_merge

expose merge metadata cleanup

11 years agoexpose merge metadata cleanup
Edward Thomson [Thu, 3 Jan 2013 21:53:50 +0000 (15:53 -0600)]
expose merge metadata cleanup

11 years agoFix core::env cleanup code
Russell Belfer [Thu, 3 Jan 2013 19:04:03 +0000 (11:04 -0800)]
Fix core::env cleanup code

Mark fake home directories that failed to be created, so we won't
try to remove them and have cleanup just use p_rmdir.

11 years agoStatus tests...
Vicent Marti [Thu, 3 Jan 2013 18:38:29 +0000 (19:38 +0100)]
Status tests...

11 years agoEven more cleanups
Vicent Marti [Thu, 3 Jan 2013 18:17:07 +0000 (19:17 +0100)]
Even more cleanups

11 years agoProper cleanup jeez
Vicent Marti [Thu, 3 Jan 2013 18:07:41 +0000 (19:07 +0100)]
Proper cleanup jeez

11 years agoMove test cleanup into cleanup functions
Ben Straub [Thu, 3 Jan 2013 17:10:38 +0000 (09:10 -0800)]
Move test cleanup into cleanup functions

11 years agoCleanup after tests
Ben Straub [Thu, 3 Jan 2013 16:38:00 +0000 (08:38 -0800)]
Cleanup after tests

11 years agoTests should clean up after themselves
Ben Straub [Thu, 3 Jan 2013 15:47:51 +0000 (07:47 -0800)]
Tests should clean up after themselves

11 years agoMerge pull request #1181 from nvloff/allow_note_overwrite
Vicent Martí [Thu, 3 Jan 2013 15:43:27 +0000 (07:43 -0800)]
Merge pull request #1181 from nvloff/allow_note_overwrite

Allow note overwrite

11 years agoadd option to allow git note overwrite
Nikolai Vladimirov [Thu, 3 Jan 2013 14:31:36 +0000 (16:31 +0200)]
add option to allow git note overwrite

11 years agoFix git__strncasecmp
Philip Kelley [Thu, 3 Jan 2013 13:45:09 +0000 (08:45 -0500)]
Fix git__strncasecmp