phases: update the phase set as we go during retract boundary
Apparently iterating over the `changed_revs` dictionary is very expensive.
On mozilla-try-2019-02-18, a perf::unbundle call with a 10 000 changesets
bundle gives give use the following timing.
e57d4b868a3e: 4.6 seconds
ac1c75188440: 102.5 seconds
prev-changeset: 30.0 seconds
this-changeset: 4.6 seconds
So, the performance regression is gone.
Once again: thanks to marvelous Python!
This test doesn't yet work due to the way fsmonitor is integrated with test runner
$ exit 80
test sparse interaction with other extensions
$ hg init myrepo
$ cd myrepo
$ cat > .hg/hgrc <<EOF
> [extensions]
> sparse=
> strip=
> EOF
Test fsmonitor integration (if available)
TODO: make fully isolated integration test a'la https://github.com/facebook/watchman/blob/master/tests/integration/WatchmanInstance.py
(this one is using the systemwide watchman instance)
$ touch .watchmanconfig
$ echo "ignoredir1/" >> .hgignore
$ hg commit -Am ignoredir1
adding .hgignore
$ echo "ignoredir2/" >> .hgignore
$ hg commit -m ignoredir2
$ hg sparse --reset
$ hg sparse -I ignoredir1 -I ignoredir2 -I dir1
$ mkdir ignoredir1 ignoredir2 dir1
$ touch ignoredir1/file ignoredir2/file dir1/file
Run status twice to compensate for a condition in fsmonitor where it will check
ignored files the second time it runs, regardless of previous state (ask @sid0)
$ hg status --config extensions.fsmonitor=
? dir1/file
$ hg status --config extensions.fsmonitor=
? dir1/file
Test that fsmonitor ignore hash check updates when .hgignore changes
$ hg up -q ".^"
$ hg status --config extensions.fsmonitor=
? dir1/file
? ignoredir2/file