| commit | f66d3c283c03fc34f0a16994385a1b5e98e11104 | |
| author | Shawn O. Pearce <spearce@spearce.org> | |
| Mon, 18 Aug 2008 23:53:15 +0000 (18 16:53 -0700) | ||
| committer | Robin Rosenberg <robin.rosenberg@dewire.com> | |
| Wed, 20 Aug 2008 21:07:32 +0000 (20 23:07 +0200) | ||
| tree | 5c1d48727518bbf5e5747433871e3a0eaea26b5a | treesnapshot (tar.gz zip) |
| parent | 4a91b92f1314a4219c465ed6166986afe65b1ff7 | commitdiff |
Micro-optimize AbstractTreeIterator.pathCompare
We were doing far too much work in pathCompare to handle
cases that just cannot ever happen, such as if the paths
were the same length but had different "last path char"
and then somehow had different lengths.
We also had the JVM doing a lot of comparsion ops just
to return -1/0/1 when really we can get away with the
non-zero result returned to the caller. Issuing just
the subtraction and one comparsion to 0 is much quicker,
JIT or not.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Robin Rosenberg <robin.rosenberg@dewire.com>
We were doing far too much work in pathCompare to handle
cases that just cannot ever happen, such as if the paths
were the same length but had different "last path char"
and then somehow had different lengths.
We also had the JVM doing a lot of comparsion ops just
to return -1/0/1 when really we can get away with the
non-zero result returned to the caller. Issuing just
the subtraction and one comparsion to 0 is much quicker,
JIT or not.
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/treewalk/AbstractTreeIterator.java | diffblobblamehistory |