Page MenuHomePhabricator

tracing: add many more timedcm() calls throughout the codebase
Needs ReviewPublic

Authored by durin42 on Jun 12 2019, 5:06 PM.

Details

Reviewers
None
Group Reviewers
hg-reviewers
Summary

I've been trying to understand where time goes in hg status and this helped
me get a sense of things. I'm happy to try and split this up if it's too
confusing, but I felt like it was fine as-is since all the changes are the
same.

Diff Detail

Repository
rHG Mercurial
Lint
Lint Skipped
Unit
Unit Tests Skipped

Event Timeline

durin42 created this revision.Jun 12 2019, 5:06 PM
durin42 updated this revision to Diff 15601.Jun 19 2019, 4:25 PM
durin42 updated this revision to Diff 16102.Aug 1 2019, 2:35 PM
durin42 updated this revision to Diff 16300.Aug 23 2019, 5:05 PM

I've been trying to understand where time goes in hg status and this helped
me get a sense of things.

Do you remember which of these were slow? You have instrumented the matcher creation quite a bit and I wouldn't have thought that would take very long (except when the matcher is very complex).

test-check-*.t were unhappy with some of the changes in this stack (including in this patch). Can you fix those and update?

I've been trying to understand where time goes in hg status and this helped
me get a sense of things.

Do you remember which of these were slow? You have instrumented the matcher creation quite a bit and I wouldn't have thought that would take very long (except when the matcher is very complex).

I believe the narrowmatcher was taking noticeable time for some of my tests.

durin42 updated this revision to Diff 16594.Mon, Sep 23, 3:25 PM