copies: avoid materializing a full directory map during copy tracing
Materializing a full copy of every directory in a treemanifest repo can be quite
expensive, even with a narrow matcher. For flat manifest repos, this should be
equivalent - it will still materialize (and cache) a dict of all of the dirs
inside of the manifest object, we just don't get a copy of it.
In a repo I have here, this brings the time for a simple rebase from 11.197s to
4.609s.
Differential Revision: https://phab.mercurial-scm.org/D9503
#!/usr/bin/env python3
from __future__ import absolute_import, print_function
import sys
from mercurial import (
commands,
localrepo,
ui as uimod,
)
print_ = print
def print(*args, **kwargs):
"""print() wrapper that flushes stdout buffers to avoid py3 buffer issues
We could also just write directly to sys.stdout.buffer the way the
ui object will, but this was easier for porting the test.
"""
print_(*args, **kwargs)
sys.stdout.flush()
u = uimod.ui.load()
print('% creating repo')
repo = localrepo.instance(u, b'.', create=True)
f = open('test.py', 'w')
try:
f.write('foo\n')
finally:
f.close
print('% add and commit')
commands.add(u, repo, b'test.py')
commands.commit(u, repo, message=b'*')
commands.status(u, repo, clean=True)
print('% change')
f = open('test.py', 'w')
try:
f.write('bar\n')
finally:
f.close()
# this would return clean instead of changed before the fix
commands.status(u, repo, clean=True, modified=True)