| commit | e8634851f46e1b54db17a45a9008b6e6284f34ce | |
| author | Shawn O. Pearce <spearce@spearce.org> | |
| Fri, 29 Aug 2008 05:54:10 +0000 (28 22:54 -0700) | ||
| committer | Robin Rosenberg <robin.rosenberg@dewire.com> | |
| Fri, 29 Aug 2008 09:17:36 +0000 (29 11:17 +0200) | ||
| tree | f3df5bdf61cb6144072f4d91d783a5d0c6f0ee56 | treesnapshot (tar.gz zip) |
| parent | 089eab66c415fc706d6947e401d9996a6c3b1ba1 | commitdiff |
index-pack: Avoid disk corruption in objects appended to thin packs
When we are appending onto a thin pack we need to make sure the whole
object we wrote into the pack is what is there when we re-read the
pack to compute the final footer.
Nicolas Pitre pointed out the need to perform this sort of compare
in C Git. We missed it in my first round of index-pack improvements.
We also now take advantage of "roughly" block aligned reads during
the fixup operation. 4k is a popular filesystem block size and by
arranging the code a little differently we can always perform reads
on 4k alignments within the file. This should make it easier for
the OS to move data to us.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Robin Rosenberg <robin.rosenberg@dewire.com>
When we are appending onto a thin pack we need to make sure the whole
object we wrote into the pack is what is there when we re-read the
pack to compute the final footer.
Nicolas Pitre pointed out the need to perform this sort of compare
in C Git. We missed it in my first round of index-pack improvements.
We also now take advantage of "roughly" block aligned reads during
the fixup operation. 4k is a popular filesystem block size and by
arranging the code a little differently we can always perform reads
on 4k alignments within the file. This should make it easier for
the OS to move data to us.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Robin Rosenberg <robin.rosenberg@dewire.com>
| org.spearce.jgit/src/org/spearce/jgit/transport/IndexPack.java | diffblobblamehistory |