| commit | f0d28143d6191a44e13eaf9c6507101d487506c4 | |
| author | Shawn O. Pearce <spearce@spearce.org> | |
| Fri, 12 Dec 2008 23:19:36 +0000 (12 15:19 -0800) | ||
| committer | Robin Rosenberg <robin.rosenberg@dewire.com> | |
| Sat, 13 Dec 2008 02:13:28 +0000 (13 03:13 +0100) | ||
| tree | 2668b0b972e93ed3785c98114f85c5914350a757 | treesnapshot (tar.gz zip) |
| parent | d9996907ad7bb1f4efc48f857a48e34bba3d1a08 | commitdiff |
Add support for parsing "diff --cc" style patches
Even though the diff --cc format used by Git is only meant to be
read by humans, JGit needs to be able to parse these to get the
patch metadata so it can be shown in a user interface to facilitate
code review processes.
Patches are parsed into the specialized CombinedFileHeader and
CombinedHunkHeader classes, where the old image information is
augmented with additional fields for the arbitrary number of parents
that can appear in such patches. These cost more in terms of memory,
but "diff --cc" style patches tend to occur very infrequently as
they only occur during a merge conflict resolution.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Robin Rosenberg <robin.rosenberg@dewire.com>
Even though the diff --cc format used by Git is only meant to be
read by humans, JGit needs to be able to parse these to get the
patch metadata so it can be shown in a user interface to facilitate
code review processes.
Patches are parsed into the specialized CombinedFileHeader and
CombinedHunkHeader classes, where the old image information is
augmented with additional fields for the arbitrary number of parents
that can appear in such patches. These cost more in terms of memory,
but "diff --cc" style patches tend to occur very infrequently as
they only occur during a merge conflict resolution.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Robin Rosenberg <robin.rosenberg@dewire.com>