phases: rework the logic of _pushdiscoveryphase to bound complexity
This rework the various graph traversal in _pushdiscoveryphase to keep the
complexity in check.
This is done though a couple of things:
- first, limiting the space we have to explore, for example, if we are not in
publishing push, we don't need to consider remote draft roots that are also
draft locally, as there is nothing to be moved there.
- avoid unbounded descendant computation, and use the faster "rev between"
computation.
This provide a massive boost to performance when exchanging with repository with
a massive amount of draft, like mozilla-try:
### data-env-vars.name = mozilla-try-2023-03-22-zstd-sparse-revlog
# benchmark.name = hg.command.push
# bin-env-vars.hg.flavor = default
# bin-env-vars.hg.py-re2-module = default
# benchmark.variants.explicit-rev = all-out-heads
# benchmark.variants.issue6528 = disabled
# benchmark.variants.protocol = ssh
# benchmark.variants.reuse-external-delta-parent = default
## benchmark.variants.revs = any-1-extra-rev
before: 20.346590 seconds
after: 11.232059 seconds (-38.15%, -7.48 seconds)
## benchmark.variants.revs = any-100-extra-rev
before: 24.752051 seconds
after: 15.367412 seconds (-37.91%, -9.38 seconds)
After this changes, the push operation is still quite too slow. Some of this
can be attributed to general phases slowness (reading all the roots from disk
for example) and other know slowness (not using persistent-nodemap, branchmap,
tags, etc. We are also working on them, but with this series, phase discovery
during push no longer showing up in profile and this is a pretty nice and bit
low-hanging fruit out of the way.
### (same case as the above)
# benchmark.variants.revs = any-1-extra-rev
pre-%ln-change: 44.235070
this-changeset: 11.232059 seconds (-74.61%, -33.00 seconds)
# benchmark.variants.revs = any-100-extra-rev
pre-%ln-change: 49.234697
this-changeset: 15.367412 seconds (-68.79%, -33.87 seconds)
Note that with this change, the `hg push` performance is now much closer to the
`hg pull` performance, even it still lagging behind a bit. (and the overall
performance are still too slow).
### data-env-vars.name = mozilla-try-2023-03-22-ds2-pnm
# benchmark.variants.explicit-rev = all-out-heads
# benchmark.variants.issue6528 = disabled
# benchmark.variants.protocol = ssh
# benchmark.variants.pulled-delta-reuse-policy = default
# bin-env-vars.hg.flavor = rust
## benchmark.variants.revs = any-1-extra-rev
hg.command.pull: 6.517450
hg.command.push: 11.219888
## benchmark.variants.revs = any-100-extra-rev
hg.command.pull: 10.160991
hg.command.push: 14.251107
### data-env-vars.name = mozilla-try-2023-03-22-zstd-sparse-revlog
# bin-env-vars.hg.py-re2-module = default
# benchmark.variants.explicit-rev = all-out-heads
# benchmark.variants.issue6528 = disabled
# benchmark.variants.protocol = ssh
# benchmark.variants.pulled-delta-reuse-policy = default
## bin-env-vars.hg.flavor = default
## benchmark.variants.revs = any-1-extra-rev
hg.command.pull: 8.577772
hg.command.push: 11.232059
## bin-env-vars.hg.flavor = default
## benchmark.variants.revs = any-100-extra-rev
hg.command.pull: 13.152976
hg.command.push: 15.367412
## bin-env-vars.hg.flavor = rust
## benchmark.variants.revs = any-1-extra-rev
hg.command.pull: 8.731982
hg.command.push: 11.178751
## bin-env-vars.hg.flavor = rust
## benchmark.variants.revs = any-100-extra-rev
hg.command.pull: 13.184236
hg.command.push: 15.620843
A python hook for "hg fix" that prints out the number of files and revisions
that were affected, along with which fixer tools were applied. Also checks how
many times it sees a specific key generated by one of the fixer tools defined
below.
$ cat >> $TESTTMP/postfixhook.py <<EOF
> import collections
> def file(ui, repo, rev=None, path=b'', metadata=None, **kwargs):
> ui.status(b'fixed %s in revision %d using %s\n' %
> (path, rev, b', '.join(metadata.keys())))
> def summarize(ui, repo, replacements=None, wdirwritten=False,
> metadata=None, **kwargs):
> counts = collections.defaultdict(int)
> keys = 0
> for fixername, metadatalist in metadata.items():
> for metadata in metadatalist:
> if metadata is None:
> continue
> counts[fixername] += 1
> if 'key' in metadata:
> keys += 1
> ui.status(b'saw "key" %d times\n' % (keys,))
> for name, count in sorted(counts.items()):
> ui.status(b'fixed %d files with %s\n' % (count, name))
> if replacements:
> ui.status(b'fixed %d revisions\n' % (len(replacements),))
> if wdirwritten:
> ui.status(b'fixed the working copy\n')
> EOF
Some mock output for fixer tools that demonstrate what could go wrong with
expecting the metadata output format.
$ printf 'new content\n' > $TESTTMP/missing
$ printf 'not valid json\0new content\n' > $TESTTMP/invalid
$ printf '{"key": "value"}\0new content\n' > $TESTTMP/valid
Configure some fixer tools based on the output defined above, and enable the
hooks defined above. Disable parallelism to make output of the parallel file
processing phase stable.
$ cat >> $HGRCPATH <<EOF
> [extensions]
> fix =
> [fix]
> metadatafalse:command=cat $TESTTMP/missing
> metadatafalse:pattern=metadatafalse
> metadatafalse:metadata=false
> missing:command=cat $TESTTMP/missing
> missing:pattern=missing
> missing:metadata=true
> invalid:command=cat $TESTTMP/invalid
> invalid:pattern=invalid
> invalid:metadata=true
> valid:command=cat $TESTTMP/valid
> valid:pattern=valid
> valid:metadata=true
> [hooks]
> postfixfile = python:$TESTTMP/postfixhook.py:file
> postfix = python:$TESTTMP/postfixhook.py:summarize
> [worker]
> enabled=false
> EOF
See what happens when we execute each of the fixer tools. Some print warnings,
some write back to the file.
$ hg init repo
$ cd repo
$ printf "old content\n" > metadatafalse
$ printf "old content\n" > invalid
$ printf "old content\n" > missing
$ printf "old content\n" > valid
$ hg add -q
$ hg fix -w
ignored invalid output from fixer tool: invalid
fixed metadatafalse in revision 2147483647 using metadatafalse
ignored invalid output from fixer tool: missing
fixed valid in revision 2147483647 using valid
saw "key" 1 times
fixed 1 files with valid
fixed the working copy
$ cat metadatafalse
new content
$ cat missing
old content
$ cat invalid
old content
$ cat valid
new content
$ cd ..