Performance fix #696
Performance fix #696
Conversation
|
Crap, there are not tests around the commit message stuff. I broke it. Need to fix. |
|
Ok, should be fixed. If either of you can test with one of your repos which caused the perf issues that would be awesome. |
|
This weekend is pretty stuffed for me but will give it a go at Monday the latest!
|
JakeGinnivan
force-pushed
the
PerformanceFix
branch
from
df2c347 to
77f778e
Compare
|
It's looking good! From 13s to 10s running latest master against this pull request, Release mode, see snapshot diff below: |
| Until = baseCommit, | ||
| SortBy = CommitSortStrategies.Topological | CommitSortStrategies.Reverse | ||
| }; | ||
| if (commit.Sha == baseCommit.Sha) |
There was a problem hiding this comment.
@JakeGinnivan why are we comparing the Sha instead of the commit? We're using Sha for comparison over the whole code base so I'll guess it's the way to go, but I'm just curious about why?
There was a problem hiding this comment.
No reason, I assume libgit2sharp does implement equal and equality operators properly but I haven't checked.
|
@JakeGinnivan Sorry, I haven't had time to test this until now. But yea, this is a major improvement. Before: After: |
|
Still would be good to make better, but that is a pretty good start. From: Asbjørn Ulsberg [mailto:[email protected]] @JakeGinnivanhttps://github.com/JakeGinnivan Sorry, I haven't had time to test this until now. But yea, this is a major improvement. Before: 2015-10-19 08:14:13 INFO [10/19/15 8:14:13:18] End: Calculating base versions (Took: 24,395.41ms) After: 2015-10-19 08:15:24 INFO [10/19/15 8:15:24:99] End: Calculating base versions (Took: 5,154.25ms) — |
// @asbjornu @orjan can you guys test this with
commit-message-incrementing: Enabled. Hopefully it should fix the perf issues we talked about in #677