]> git.proxmox.com Git - libgit2.git/log
libgit2.git
10 years agoRename the ssh credentials
Carlos Martín Nieto [Wed, 23 Oct 2013 10:08:54 +0000 (12:08 +0200)]
Rename the ssh credentials

The names from libssh2 are somewhat obtuse for us. We can simplify the
usual key/passphrase credential's name, as well as make clearer what the
custom signature function is.

10 years agoMerge pull request #1897 from libgit2/split-patch-from-diff
Vicent Martí [Tue, 22 Oct 2013 18:55:54 +0000 (11:55 -0700)]
Merge pull request #1897 from libgit2/split-patch-from-diff

RFC: Proposed reworking of diff APIs

10 years agoFix typo
Russell Belfer [Tue, 22 Oct 2013 18:12:44 +0000 (11:12 -0700)]
Fix typo

10 years agoImplement `git_odb_object_dup`
Vicent Marti [Tue, 22 Oct 2013 14:05:47 +0000 (16:05 +0200)]
Implement `git_odb_object_dup`

10 years agoImplement patience and minimal diff flags
Russell Belfer [Mon, 21 Oct 2013 22:36:38 +0000 (15:36 -0700)]
Implement patience and minimal diff flags

It seems that to implement these options, we just have to pass
the appropriate flags through to the libxdiff code taken from
core git.  So let's do it (and add a test).

10 years agoFix warnings for win64
Russell Belfer [Mon, 21 Oct 2013 21:16:53 +0000 (14:16 -0700)]
Fix warnings for win64

10 years agoCreate git_diff_line and extend git_diff_hunk
Russell Belfer [Mon, 21 Oct 2013 20:42:42 +0000 (13:42 -0700)]
Create git_diff_line and extend git_diff_hunk

Instead of having functions with so very many parameters to pass
hunk and line data, this takes the existing git_diff_hunk struct
and extends it with more hunk data, plus adds a git_diff_line.
Those structs are used to pass back hunk and line data instead of
the old APIs that took tons of parameters.

Some work that was previously only being done for git_diff_patch
creation (scanning the diff content for exact line counts) is now
done for all callbacks, but the performance difference should not
be noticable.

10 years agoMerge pull request #1903 from ethomson/ssh
Vicent Martí [Mon, 21 Oct 2013 20:06:27 +0000 (13:06 -0700)]
Merge pull request #1903 from ethomson/ssh

Allowed credential types should be a bitfield

10 years agoAllowed credential types should be a bitfield
Edward Thomson [Mon, 21 Oct 2013 16:03:31 +0000 (11:03 -0500)]
Allowed credential types should be a bitfield

10 years agoTweak to git_diff_delta structure for nfiles
Russell Belfer [Mon, 21 Oct 2013 16:07:19 +0000 (09:07 -0700)]
Tweak to git_diff_delta structure for nfiles

While the base git_diff_delta structure always contains two files,
when we introduce conflict data, it will be helpful to have an
indicator when an additional file is involved.

10 years agoMerge pull request #1900 from haneefmubarak/patch-1
Vicent Martí [Fri, 18 Oct 2013 17:00:26 +0000 (10:00 -0700)]
Merge pull request #1900 from haneefmubarak/patch-1

license clarification + formatting all over

10 years agoslightly clarified the "license" portion of `README.md" + formatting all over
Haneef Mubarak [Thu, 17 Oct 2013 23:01:43 +0000 (16:01 -0700)]
slightly clarified the "license" portion of `README.md" + formatting all over

fixed typo `can can` -> `can`

10 years agoslightly clarified the "license" portion of `README.md" + formatting all over
Haneef Mubarak [Thu, 17 Oct 2013 08:52:20 +0000 (01:52 -0700)]
slightly clarified the "license" portion of `README.md" + formatting all over

added the second change suggested by @nulltoken (with two extra words :) )

10 years agoslightly clarified the "license" portion of `README.md" + formatting all over
Haneef Mubarak [Thu, 17 Oct 2013 07:28:08 +0000 (00:28 -0700)]
slightly clarified the "license" portion of `README.md" + formatting all over

There are commercial open source applications, hence proprietary seems to be a better term.

Also:
 - Trimmed header `=`'s to match text length
 - made libgit2 stand out everywhere
 - pretty printing links
 - irc link for those with the correct setup

10 years agoMerge pull request #1702 from ethomson/checkout_merge
Vicent Martí [Wed, 16 Oct 2013 20:51:24 +0000 (13:51 -0700)]
Merge pull request #1702 from ethomson/checkout_merge

Checkout merge

10 years agoMove path prefixed help to path.h
Edward Thomson [Wed, 9 Oct 2013 22:26:42 +0000 (18:26 -0400)]
Move path prefixed help to path.h

10 years agoMove functions in checkout_conflicts to checkout.c
Edward Thomson [Mon, 23 Sep 2013 21:52:40 +0000 (17:52 -0400)]
Move functions in checkout_conflicts to checkout.c

It seemed exceptionally silly to have a split there
where no split needed to be.

10 years agoTwo-step conflict checkout (load / perform)
Edward Thomson [Mon, 23 Sep 2013 13:47:47 +0000 (09:47 -0400)]
Two-step conflict checkout (load / perform)

Move conflict handling into two steps: load the conflicts and
then apply the conflicts.  This is more compatible with the
existing checkout implementation and makes progress reporting
more sane.

10 years agoHonor UPDATE_ONLY bit when checking out conflicts
Edward Thomson [Sat, 10 Aug 2013 01:23:36 +0000 (20:23 -0500)]
Honor UPDATE_ONLY bit when checking out conflicts

10 years agoFix warning, fix memory leak
Edward Thomson [Tue, 6 Aug 2013 01:06:09 +0000 (20:06 -0500)]
Fix warning, fix memory leak

10 years agoDon't overwrite ~ files checking out conflicts
Edward Thomson [Thu, 8 Aug 2013 21:46:49 +0000 (16:46 -0500)]
Don't overwrite ~ files checking out conflicts

If a D/F conflict or rename 2->1 conflict occurs,
we write the file sides as filename~branchname.  If
a file with that name already exists in the working
directory, write as filename~branchname_0 instead.
(Incrementing 0 until a unique filename is found.)

10 years agoGet rid of some quick hacks
Edward Thomson [Thu, 8 Aug 2013 18:16:13 +0000 (13:16 -0500)]
Get rid of some quick hacks

10 years agocheckout rename conflicts
Edward Thomson [Tue, 23 Jul 2013 22:32:42 +0000 (15:32 -0700)]
checkout rename conflicts

rename conflict tests for checkout conflicts, don't suffix filenames
when checking out with USE_OURS or USE_THEIRS

10 years agoPrevent checkout_tree when conflicts exist, clear NAME on checkout tree
Edward Thomson [Mon, 22 Jul 2013 15:51:29 +0000 (08:51 -0700)]
Prevent checkout_tree when conflicts exist, clear NAME on checkout tree

Prevent checkout tree when unresolved changes exist (unless FORCE flag
is specified).  Clear NAME table when checking out, to avoid
checkout_conflicts from attempting to manipulate it.  Ensure that NAME
is also cleared at reset.

10 years agocheckout (from index) can write conflicts
Edward Thomson [Wed, 17 Jul 2013 14:49:56 +0000 (09:49 -0500)]
checkout (from index) can write conflicts

10 years agotests for checkout index with conflicts
Edward Thomson [Wed, 17 Jul 2013 14:43:32 +0000 (09:43 -0500)]
tests for checkout index with conflicts

10 years agoDiff API cleanup
Russell Belfer [Tue, 15 Oct 2013 22:10:07 +0000 (15:10 -0700)]
Diff API cleanup

This lays groundwork for separating formatting options from diff
creation options.  This groups the formatting flags separately
from the diff list creation flags and reorders the options.  This
also tweaks some APIs to further separate code that uses patches
from code that just looks at git_diffs.

10 years agoRename diff objects and split patch.h
Russell Belfer [Fri, 11 Oct 2013 21:51:54 +0000 (14:51 -0700)]
Rename diff objects and split patch.h

This makes no functional change to diff but renames a couple of
the objects and splits the new git_patch (formerly git_diff_patch)
into a new header file.

10 years agoMerge pull request #1562 from libgit2/cmn/refs-namespace-lookup
Vicent Martí [Fri, 11 Oct 2013 16:47:05 +0000 (09:47 -0700)]
Merge pull request #1562 from libgit2/cmn/refs-namespace-lookup

Provide the user with a more useful error code when a looking up a reference which name points to a namepace

10 years agofutils: return GIT_ENOTFOUND when trying to read a directory
Carlos Martín Nieto [Sat, 17 Aug 2013 05:11:31 +0000 (07:11 +0200)]
futils: return GIT_ENOTFOUND when trying to read a directory

This lets the reference code return not-found when the user asks to
look up a reference when in fact they pass a namespace.

10 years agoMerge pull request #1898 from arthurschreiber/patch-2
Ben Straub [Thu, 10 Oct 2013 22:29:57 +0000 (15:29 -0700)]
Merge pull request #1898 from arthurschreiber/patch-2

Fix a typo in the rpm build instructions.

10 years agoFix a typo in the rpm build instructions.
Arthur Schreiber [Thu, 10 Oct 2013 22:06:08 +0000 (00:06 +0200)]
Fix a typo in the rpm build instructions.

10 years agoAfter iconv init reset ref normalize error
Russell Belfer [Wed, 9 Oct 2013 18:57:03 +0000 (11:57 -0700)]
After iconv init reset ref normalize error

The iconv init was accidentally clearing the default error state
during reference normalization. This resets so that normalization
errors will be detected correctly.

10 years agoMerge pull request #1886 from libgit2/precompose-utf8
Vicent Martí [Wed, 9 Oct 2013 00:03:12 +0000 (17:03 -0700)]
Merge pull request #1886 from libgit2/precompose-utf8

Add support for core.precomposeunicode on Mac

10 years agoRename new fn to git_repository_reinit_filesystem
Russell Belfer [Tue, 8 Oct 2013 23:59:59 +0000 (16:59 -0700)]
Rename new fn to git_repository_reinit_filesystem

10 years agoMake reference lookups apply precomposeunicode
Russell Belfer [Tue, 8 Oct 2013 23:35:57 +0000 (16:35 -0700)]
Make reference lookups apply precomposeunicode

Before these changes, looking up a reference would return the
same precomposed or decomposed form of the reference name that
was used to look it up, so on MacOS which ignores the difference
between the two, a single reference could be looked up either way
and git_reference_name would return the form of the name that was
used to look it up!  This change makes lookup always return the
precomposed name if core.precomposeunicode is set regardless of
which version was used to look it up.  The reference iterator was
already returning the precomposed form from earlier work.

This also updates the CMakeLists.txt rules for enabling iconv
usage because the clar tests for this code were actually not being
activated properly with the old version.

Finally, this moves git_repository_reset_filesystem from include/
git2/repository.h to include/git2/sys/repository.h since it is not
really a function that normal library users should have to think
about very often.

10 years agoMerge pull request #1895 from libgit2/winhttp-accept
Vicent Martí [Tue, 8 Oct 2013 23:32:06 +0000 (16:32 -0700)]
Merge pull request #1895 from libgit2/winhttp-accept

WinHTTP: set Accept header for POSTs

10 years agoCombine WinHTTP API calls
Ben Straub [Tue, 8 Oct 2013 22:52:37 +0000 (15:52 -0700)]
Combine WinHTTP API calls

10 years agoCase sensitivity issues on Linux
Russell Belfer [Tue, 8 Oct 2013 21:41:32 +0000 (14:41 -0700)]
Case sensitivity issues on Linux

A couple of tests were actually dealing incorrectly with case
sensitivity issues on Linux because they were relying on having
core.ignorecase set to true.  Now that the fixture initialization
sets the case sensitivity to be accurate for the platform, it
exposed bugs in these tests.

10 years agoAdd Assembla unit test
Ben Straub [Tue, 8 Oct 2013 20:59:43 +0000 (13:59 -0700)]
Add Assembla unit test

10 years agoWinHTTP: set Accept header for POSTs
Ben Straub [Tue, 8 Oct 2013 20:40:52 +0000 (13:40 -0700)]
WinHTTP: set Accept header for POSTs

10 years agoMore filemode cleanups for FAT on MacOS
Russell Belfer [Tue, 8 Oct 2013 19:45:43 +0000 (12:45 -0700)]
More filemode cleanups for FAT on MacOS

This cleans up some additional issues.  The main change is that
on a filesystem that doesn't support mode bits, libgit2 will now
create new blobs with GIT_FILEMODE_BLOB always instead of being
at the mercy to the filesystem driver to report executable or not.
This means that if "core.filemode" lies and claims that filemode
is not supported, then we will ignore the executable bit from the
filesystem.  Previously we would have allowed it.

This adds an option to the new git_repository_reset_filesystem to
recurse through submodules if desired.  There may be other types
of APIs that would like a "recurse submodules" option, but this
one is particularly useful.

This also has a number of cleanups, etc., for related things
including trying to give better error messages when problems come
up from the filesystem.  For example, the FAT filesystem driver on
MacOS appears to return errno EINVAL if you attempt to write a
filename with invalid UTF-8 in it.  We try to capture that with a
better error message now.

10 years agoMerge pull request #1894 from libgit2/cmn/tree-raw
Vicent Martí [Tue, 8 Oct 2013 11:31:26 +0000 (04:31 -0700)]
Merge pull request #1894 from libgit2/cmn/tree-raw

tree: allow retrieval of raw attributes

10 years agotree: allow retrieval of raw attributes
Carlos Martín Nieto [Mon, 15 Apr 2013 07:07:57 +0000 (09:07 +0200)]
tree: allow retrieval of raw attributes

When a tool needs to recreate the tree object (for example an
interface to another VCS), it needs to use the raw attributes,
forgoing any normalization.

10 years agoMerge pull request #1890 from libgit2/multiple-init
Vicent Martí [Sat, 5 Oct 2013 14:56:03 +0000 (07:56 -0700)]
Merge pull request #1890 from libgit2/multiple-init

Multiple init

10 years agoDon't use git_atomic as an integer
Ben Straub [Sat, 5 Oct 2013 02:35:32 +0000 (19:35 -0700)]
Don't use git_atomic as an integer

10 years agoUse InterlockedCompareExchange for the lock
Philip Kelley [Fri, 4 Oct 2013 22:38:37 +0000 (18:38 -0400)]
Use InterlockedCompareExchange for the lock

10 years agoPosix synchronized init, prototype win32 version
Ben Straub [Sat, 5 Oct 2013 02:33:48 +0000 (19:33 -0700)]
Posix synchronized init, prototype win32 version

10 years agoAdd test for multiple thread init/shutdown
Ben Straub [Thu, 3 Oct 2013 19:44:34 +0000 (12:44 -0700)]
Add test for multiple thread init/shutdown

10 years agoAdd git_repository_reset_filesystem and fix tests
Russell Belfer [Fri, 4 Oct 2013 23:32:16 +0000 (16:32 -0700)]
Add git_repository_reset_filesystem and fix tests

When a repository is transferred from one file system to another,
many of the config settings that represent the properties of the
file system may be wrong.  This adds a new public API that will
refresh the config settings of the repository to account for the
change of file system.  This doesn't do a full "reinitialize" and
operates on a existing git_repository object refreshing the config
when done.

This commit then makes use of the new API in clar as each test
repository is set up.

This commit also has a number of other clar test fixes where we
were making assumptions about the type of filesystem, either based
on outdated config data or based on the OS instead of the FS.

10 years agoMerge pull request #1892 from libgit2/ntk/topic/index_read
Vicent Martí [Fri, 4 Oct 2013 15:50:44 +0000 (08:50 -0700)]
Merge pull request #1892 from libgit2/ntk/topic/index_read

Make git_index_read() cope with external additions and removals of the index file

10 years agoindex: Make _read() cope with index file creation
nulltoken [Fri, 4 Oct 2013 12:03:12 +0000 (14:03 +0200)]
index: Make _read() cope with index file creation

10 years agoindex: Enhance documentation
nulltoken [Fri, 4 Oct 2013 11:49:26 +0000 (13:49 +0200)]
index: Enhance documentation

10 years agoTabify indentations
nulltoken [Fri, 4 Oct 2013 09:36:40 +0000 (11:36 +0200)]
Tabify indentations

10 years agoMore cleanups to remove WIN assumptions
Russell Belfer [Thu, 3 Oct 2013 22:16:06 +0000 (15:16 -0700)]
More cleanups to remove WIN assumptions

This cleans up more of the test suite to check actual filesystem
behavior instead of relying on Windows vs. Mac vs. Linux to test.

10 years agoUpdate repo init with fewer platform assumptions
Russell Belfer [Thu, 3 Oct 2013 21:42:37 +0000 (14:42 -0700)]
Update repo init with fewer platform assumptions

The repo init code was assuming Windows == no filemode, and
Mac or Windows == no case sensitivity.  Those assumptions are not
consistently true depending on the mounted file system.  This is a
first step to removing those assumptions.  It focuses on the repo
init code and the tests of that code.  There are still many other
tests that are broken when those assumptions don't hold true, but
this clears up one area of the code.

Also, this moves the core.precomposeunicode logic to be closer to
the current logic in core Git where it will be set to true on any
filesystem where composed unicode is decomposed when read back.

10 years agoClean up annoying warnings
Russell Belfer [Wed, 2 Oct 2013 21:13:11 +0000 (14:13 -0700)]
Clean up annoying warnings

The indexer code was generating warnings on Windows 64-bit.  I
looked closely at the logic and was able to simplify it a bit.

Also this fixes some other Windows and Linux warnings.

10 years agoWrap iconv stuff and write tests
Russell Belfer [Wed, 2 Oct 2013 19:06:26 +0000 (12:06 -0700)]
Wrap iconv stuff and write tests

This adds a simple wrapper around the iconv APIs and uses it
instead of the old code that was inlining the iconv stuff.  This
makes it possible for me to test the iconv logic in isolation.
A "no iconv" version of the API was defined with macros so that
I could have fewer ifdefs in the code itself.

10 years agoSimplify git_path_is_empty_dir implementation
Russell Belfer [Wed, 2 Oct 2013 18:07:18 +0000 (11:07 -0700)]
Simplify git_path_is_empty_dir implementation

This simplifies git_path_is_empty_dir on both Windows (getting rid
of git_buf allocation inside the function) and other platforms (by
just using git_path_direach), and adds tests for the function, and
uses the function to simplify some existing tests.

10 years agoAdd to Git authors who have agreed to relicense
Russell Belfer [Tue, 1 Oct 2013 23:41:07 +0000 (16:41 -0700)]
Add to Git authors who have agreed to relicense

In email, Torsten Bögershausen agreed that we could use his code
from core Git in libgit2 under the modified license.  Also, since
his work is the basis for much of the precompose unicode support,
I have added him to the AUTHORS file as well.

10 years agoInitial iconv hookup for precomposed unicode
Russell Belfer [Tue, 1 Oct 2013 23:12:15 +0000 (16:12 -0700)]
Initial iconv hookup for precomposed unicode

This hooks up git_path_direach and git_path_dirload so that they
will take a flag indicating if directory entry names should be
tested and converted from decomposed unicode to precomposed form.
This code will only come into play on the Apple platform and even
then, only when certain types of filesystems are used.

This involved adding a flag to these functions which involved
changing a lot of places in the code.

This was an opportunity to do a bit of code cleanup here and there,
for example, getting rid of the git_futils_cleanupdir_r function in
favor of a simple flag to git_futils_rmdir_r to not remove the top
level entry.  That ended up adding depth tracking during rmdir_r
which led to a safety check for infinite directory recursion.  Yay.

This hasn't actually been tested on the Mac filesystems where the
issue occurs.  I still need to get test environment for that.

10 years agoPut hooks in place for precompose in dirload fn
Russell Belfer [Mon, 30 Sep 2013 23:58:33 +0000 (16:58 -0700)]
Put hooks in place for precompose in dirload fn

This doesn't actual do string precompose but it puts the hooks in
place into the iterators and the git_path_dirload function so that
the actual precompose work is ready to go.

10 years agoAdd check if we need to precompose unicode on Mac
Russell Belfer [Mon, 30 Sep 2013 23:13:53 +0000 (16:13 -0700)]
Add check if we need to precompose unicode on Mac

This adds initialization of core.precomposeunicode to repo init
on Mac.  This is necessary because when a Mac accesses a repo on
a VFAT or SAMBA file system, it will return directory entries in
decomposed unicode even if the filesystem entry is precomposed.

This also removes caching of a number of repo properties from the
repo init pipeline because these are properties of the specific
filesystem on which the repo is created, not of the system as a
whole.

10 years agoMerge pull request #1888 from jamill/network_cancellation
Vicent Martí [Thu, 3 Oct 2013 15:18:41 +0000 (08:18 -0700)]
Merge pull request #1888 from jamill/network_cancellation

network cancellation improvements

10 years agoStyle tweaks and changes for code review feedback
Jameson Miller [Wed, 2 Oct 2013 19:39:18 +0000 (15:39 -0400)]
Style tweaks and changes for code review feedback

10 years agoMerge pull request #1887 from libgit2/ntk/topic/git_message_raw
Vicent Martí [Thu, 3 Oct 2013 11:36:29 +0000 (04:36 -0700)]
Merge pull request #1887 from libgit2/ntk/topic/git_message_raw

commit: Introduce git_commit_message_raw()

10 years agocommit: Introduce git_commit_message_raw()
nulltoken [Wed, 2 Oct 2013 10:42:41 +0000 (12:42 +0200)]
commit: Introduce git_commit_message_raw()

10 years agoMerge pull request #1889 from libgit2/relative-redirect
Russell Belfer [Wed, 2 Oct 2013 22:26:20 +0000 (15:26 -0700)]
Merge pull request #1889 from libgit2/relative-redirect

HTTP: handle "relative" redirects

10 years agoHTTP: handle "relative" redirects
Ben Straub [Wed, 2 Oct 2013 21:45:57 +0000 (14:45 -0700)]
HTTP: handle "relative" redirects

10 years agoFix broken online push tests
Jameson Miller [Wed, 2 Oct 2013 19:10:39 +0000 (15:10 -0400)]
Fix broken online push tests

10 years agoSupport cancellation in push operation
Jameson Miller [Wed, 2 Oct 2013 17:45:32 +0000 (13:45 -0400)]
Support cancellation in push operation

This commit adds cancellation for the push operation. This work consists of:

1) Support cancellation during push operation
    - During object counting phase
    - During network transfer phase
        - Propagate GIT_EUSER error code out to caller
2) Improve cancellation support during fetch
    - Handle cancellation request during network transfer phase
    - Clear error string when cancelled during indexing
3) Fix error handling in git_smart__download_pack

Cancellation during push is still only handled in the pack building and
network transfer stages of push (and not during packbuilding).

10 years agoMerge pull request #1855 from libgit2/cmn/clone-into
Vicent Martí [Wed, 2 Oct 2013 12:56:37 +0000 (05:56 -0700)]
Merge pull request #1855 from libgit2/cmn/clone-into

An alternative for 'clone' with options

10 years agoclone: put the callbacks struct directly in the clone options
Carlos Martín Nieto [Wed, 2 Oct 2013 12:04:44 +0000 (14:04 +0200)]
clone: put the callbacks struct directly in the clone options

There's no need for this to be a pointer to somewhere else.

11 years agoremote: add some comments to the callback struct
Carlos Martín Nieto [Fri, 20 Sep 2013 21:23:42 +0000 (23:23 +0200)]
remote: add some comments to the callback struct

Hopefully clear up what they're for.

11 years agoclone: mention clone_into in the clone documentation
Carlos Martín Nieto [Fri, 20 Sep 2013 21:14:52 +0000 (23:14 +0200)]
clone: mention clone_into in the clone documentation

Make the difference more explicit.

11 years agoclone: bring back NULL as defaults
Carlos Martín Nieto [Fri, 20 Sep 2013 21:14:12 +0000 (23:14 +0200)]
clone: bring back NULL as defaults

This wasremoved as part of the large culling a few commits ago.

11 years agoclone: re-allow using a custom remote name
Carlos Martín Nieto [Fri, 20 Sep 2013 20:57:01 +0000 (22:57 +0200)]
clone: re-allow using a custom remote name

This is a small thing that by itself doesn't quite justify making the
user use clone_into.

11 years agoclone: const-ify checkout options
Carlos Martín Nieto [Fri, 20 Sep 2013 20:49:20 +0000 (22:49 +0200)]
clone: const-ify checkout options

The removal of many options which lead to the direct usage of the
user's checkout options means we should make sure they remain const.

11 years agoclone: re-add a way to ignore certificate errors
Carlos Martín Nieto [Fri, 20 Sep 2013 20:46:32 +0000 (22:46 +0200)]
clone: re-add a way to ignore certificate errors

This used to be done via transport flags, which was removed in a
previous commit.

11 years agoclone: remove more options from basic clone
Carlos Martín Nieto [Fri, 20 Sep 2013 20:34:05 +0000 (22:34 +0200)]
clone: remove more options from basic clone

The basic clone function is there to make it easy to create a "normal"
clone. Remove a bunch of options that are about changing the remote's
configuration.

11 years agoclone: implement git_clone on top of git_clone_into
Carlos Martín Nieto [Tue, 17 Sep 2013 03:31:34 +0000 (05:31 +0200)]
clone: implement git_clone on top of git_clone_into

Unify the code bases.

11 years agoclone: remove the autotag option
Carlos Martín Nieto [Mon, 16 Sep 2013 16:42:53 +0000 (18:42 +0200)]
clone: remove the autotag option

Downloading all tags is part of what makes it a clone instead of
simply a fetch.

11 years agoremote: add a convenience 'fetch' function.
Carlos Martín Nieto [Mon, 16 Sep 2013 14:54:37 +0000 (16:54 +0200)]
remote: add a convenience 'fetch' function.

11 years agoclone: implement git_clone_into
Carlos Martín Nieto [Mon, 16 Sep 2013 03:10:55 +0000 (05:10 +0200)]
clone: implement git_clone_into

This allows you to set up the repository and remote as you which to
have them before performing the clone operation.

11 years agoremote: move the credentials callback to the struct
Carlos Martín Nieto [Mon, 16 Sep 2013 03:02:25 +0000 (05:02 +0200)]
remote: move the credentials callback to the struct

Move this one as well, letting us have a single way of setting the
callbacks for the remote, and removing fields from the clone options.

11 years agoremote: put the _download() callback with the others
Carlos Martín Nieto [Mon, 16 Sep 2013 02:20:05 +0000 (04:20 +0200)]
remote: put the _download() callback with the others

The text progress and update_tips callbacks are already part of the
struct, which was meant to unify the callback setup, but the download
one was left out.

11 years agoMerge pull request #1885 from ethomson/transport_register
Vicent Martí [Tue, 1 Oct 2013 21:00:30 +0000 (14:00 -0700)]
Merge pull request #1885 from ethomson/transport_register

Add git_transport_register, git_transport_unregister

11 years agoAdd git_transport_register, git_transport_unregister
Philip Kelley [Fri, 4 Jan 2013 19:39:05 +0000 (14:39 -0500)]
Add git_transport_register, git_transport_unregister

11 years agoMark git__timer as inline on OSX
Ben Straub [Tue, 1 Oct 2013 19:56:47 +0000 (12:56 -0700)]
Mark git__timer as inline on OSX

11 years agoMerge pull request #1882 from linquize/config-subsection-fix
Vicent Martí [Tue, 1 Oct 2013 12:54:54 +0000 (05:54 -0700)]
Merge pull request #1882 from linquize/config-subsection-fix

Config subsection name should allow to have ']' and '\\' should allow to escape any characters

11 years agoAdd negative test cases for config header with invalid characters
Linquize [Tue, 1 Oct 2013 01:46:56 +0000 (09:46 +0800)]
Add negative test cases for config header with invalid characters

11 years agoAdd test case to test ']' and '\\' characters in config subsection
Linquize [Mon, 30 Sep 2013 15:58:58 +0000 (23:58 +0800)]
Add test case to test ']' and '\\' characters in config subsection

11 years agoConfig subsection name should allow to have ']' and '\\' should allow to escape any...
Linquize [Mon, 30 Sep 2013 15:38:22 +0000 (23:38 +0800)]
Config subsection name should allow to have ']' and '\\' should allow to escape any characters

11 years agoMerge pull request #1879 from libgit2/redir-refactor
Vicent Martí [Mon, 30 Sep 2013 22:03:35 +0000 (15:03 -0700)]
Merge pull request #1879 from libgit2/redir-refactor

Redir refactor

11 years agoMerge pull request #1412 from jamill/push_progress
Vicent Martí [Mon, 30 Sep 2013 21:58:45 +0000 (14:58 -0700)]
Merge pull request #1412 from jamill/push_progress

Initial Implementation of progress reports during push

11 years agoMerge pull request #1881 from libgit2/ignore-submodules-in-stash
Vicent Martí [Mon, 30 Sep 2013 21:57:48 +0000 (14:57 -0700)]
Merge pull request #1881 from libgit2/ignore-submodules-in-stash

Never consider submodules for stashing

11 years agoEnsure submodule repos and indices are freed
Ben Straub [Mon, 30 Sep 2013 21:47:56 +0000 (14:47 -0700)]
Ensure submodule repos and indices are freed

...before the helper's cleanup method tries to delete their files.

11 years agoMerge pull request #1878 from libgit2/ntk/fix/warnings
Vicent Martí [Mon, 30 Sep 2013 19:18:19 +0000 (12:18 -0700)]
Merge pull request #1878 from libgit2/ntk/fix/warnings

Fix x86/x64 size_t related warnings

11 years agoInitial Implementation of progress reports during push
Jameson Miller [Thu, 19 Sep 2013 18:52:57 +0000 (14:52 -0400)]
Initial Implementation of progress reports during push

This adds the basics of progress reporting during push. While progress
for all aspects of a push operation are not reported with this change,
it lays the foundation to add these later. Push progress reporting
can be improved in the future - and consumers of the API should
just get more accurate information at that point.

The main areas where this is lacking are:

1) packbuilding progress: does not report progress during deltafication,
   as this involves coordinating progress from multiple threads.

2) network progress: reports progress as objects and bytes are going
   to be written to the subtransport (instead of as client gets
   confirmation that they have been received by the server) and leaves
   out some of the bytes that are transfered as part of the push protocol.
   Basically, this reports the pack bytes that are written to the
   subtransport. It does not report the bytes sent on the wire that
   are received by the server. This should be a good estimate of
   progress (and an improvement over no progress).

11 years agoMerge pull request #1883 from libgit2/ntk/fix/empty_first_commit_line
Vicent Martí [Mon, 30 Sep 2013 16:40:15 +0000 (09:40 -0700)]
Merge pull request #1883 from libgit2/ntk/fix/empty_first_commit_line

Remove leading newlines from commit message