##// END OF EJS Templates
copies: calculate mergecopies() based on pathcopies()...
copies: calculate mergecopies() based on pathcopies() When copies are stored in changesets, we need a changeset-centric version of mergecopies() just like we have a changeset-centric version of pathcopies(). I think the natural way of thinking about mergecopies() is in terms of pathcopies() from the base to each of the commits. So if we can rewrite mergecopies() based on two such pathcopies() calls, we'll get the changeset-centric version for free. That's what this patch does. A nice bonus is that it ends up being a lot simpler. mergecopies() has accumulated a lot of technical debt over time. One good example is the code for dealing with grafts (the "partial/incomplete/dirty" stuff). Since pathcopies() already deals with backwards renames and ping-pong renames, we get that for free. I've run tests with hard-coded debug logging for "fullcopy" and while I haven't looked at every difference it produces, all the ones I have looked at seemed reasonable to me. I'm a little surprised that no more tests fail when run with '--extra-config-opt experimental.copies.read-from=compatibility' compared to before this patch. This patch also fixes the broken cases in test-annotate.t and test-fastannotate.t. It also enables the part of test-copies.t that was previously disabled exactly because mergecopies() needed to get a changeset-centric version. One drawback of the rewritten code is that we may now make remotefilelog prefetch more files. We used to prefetch files that were unique to either side of the merge compared to the other. We now prefetch files that are unique to either side of the merge compared to the base. This means that if you added the same file to each side, we would not prefetch it before, but we would now. Such cases are probably quite rare, but one likely scenario where they happen is when moving from a commit to its successor (or the other way around). The user will probably already have the files in the cache in such cases, so it's probably not a big deal. Some timings for calculating mergecopies between two revisions (revisions shown on each line, all using the common ancestor as base): In the hg repo: 4.8 4.9: 0.21s -> 0.21s 4.0 4.8: 0.35s -> 0.63s In and old copy of the mozilla-unified repo: FIREFOX_BETA_60_BASE^ FIREFOX_BETA_60_BASE: 0.82s -> 0.82s FIREFOX_NIGHTLY_59_END FIREFOX_BETA_60_BASE: 2.5s -> 2.6s FIREFOX_BETA_59_END FIREFOX_BETA_60_BASE: 3.9s -> 4.1s FIREFOX_AURORA_50_BASE FIREFOX_BETA_60_BASE: 31s -> 33s So it's measurably slower in most cases. The most significant difference is in the hg repo between revisions 4.0 and 4.8. In that case it seems to come from the fact that pathcopies() uses fctx.isintroducedafter() (in _tracefile), while the old mergecopies() used fctx.linkrev() (in _checkcopies()). That results in a single call to filectx._adjustlinkrev(), which is responsible for the entire difference in time (in my repo). So we pay a performance penalty but we get more correct code (see change in test-mv-cp-st-diff.t). Deleting the "== f.filenode()" in _tracefile() recovers the lost performance in the hg repo. There were are few other optimizations in _checkcopies() that I could not measure any impact from. One was from the "seen" set. Another was from a "continue" when the file was not in the destination manifest (corresponding to "am" in _tracefile). Also note that merge copies are not calculated when updating with a clean working copy, which is probably the most common case. I therefore think the much simpler code is worth the slowdown. Differential Revision: https://phab.mercurial-scm.org/D6255

File last commit:

r42132:42e2c7c5 default
r42408:57203e02 default
Show More
split.py
177 lines | 6.3 KiB | text/x-python | PythonLexer
# split.py - split a changeset into smaller ones
#
# Copyright 2015 Laurent Charignon <lcharignon@fb.com>
# Copyright 2017 Facebook, Inc.
#
# This software may be used and distributed according to the terms of the
# GNU General Public License version 2 or any later version.
"""command to split a changeset into smaller ones (EXPERIMENTAL)"""
from __future__ import absolute_import
from mercurial.i18n import _
from mercurial.node import (
nullid,
short,
)
from mercurial import (
bookmarks,
cmdutil,
commands,
error,
hg,
obsolete,
phases,
pycompat,
registrar,
revsetlang,
scmutil,
)
# allow people to use split without explicitly enabling rebase extension
from . import (
rebase,
)
cmdtable = {}
command = registrar.command(cmdtable)
# Note for extension authors: ONLY specify testedwith = 'ships-with-hg-core' for
# extensions which SHIP WITH MERCURIAL. Non-mainline extensions should
# be specifying the version(s) of Mercurial they are tested with, or
# leave the attribute unspecified.
testedwith = 'ships-with-hg-core'
@command('split',
[('r', 'rev', '', _("revision to split"), _('REV')),
('', 'rebase', True, _('rebase descendants after split')),
] + cmdutil.commitopts2,
_('hg split [--no-rebase] [[-r] REV]'),
helpcategory=command.CATEGORY_CHANGE_MANAGEMENT, helpbasic=True)
def split(ui, repo, *revs, **opts):
"""split a changeset into smaller ones
Repeatedly prompt changes and commit message for new changesets until there
is nothing left in the original changeset.
If --rev was not given, split the working directory parent.
By default, rebase connected non-obsoleted descendants onto the new
changeset. Use --no-rebase to avoid the rebase.
"""
opts = pycompat.byteskwargs(opts)
revlist = []
if opts.get('rev'):
revlist.append(opts.get('rev'))
revlist.extend(revs)
with repo.wlock(), repo.lock(), repo.transaction('split') as tr:
revs = scmutil.revrange(repo, revlist or ['.'])
if len(revs) > 1:
raise error.Abort(_('cannot split multiple revisions'))
rev = revs.first()
ctx = repo[rev]
if rev is None or ctx.node() == nullid:
ui.status(_('nothing to split\n'))
return 1
if ctx.node() is None:
raise error.Abort(_('cannot split working directory'))
# rewriteutil.precheck is not very useful here because:
# 1. null check is done above and it's more friendly to return 1
# instead of abort
# 2. mergestate check is done below by cmdutil.bailifchanged
# 3. unstable check is more complex here because of --rebase
#
# So only "public" check is useful and it's checked directly here.
if ctx.phase() == phases.public:
raise error.Abort(_('cannot split public changeset'),
hint=_("see 'hg help phases' for details"))
descendants = list(repo.revs('(%d::) - (%d)', rev, rev))
alloworphaned = obsolete.isenabled(repo, obsolete.allowunstableopt)
if opts.get('rebase'):
# Skip obsoleted descendants and their descendants so the rebase
# won't cause conflicts for sure.
torebase = list(repo.revs('%ld - (%ld & obsolete())::',
descendants, descendants))
if not alloworphaned and len(torebase) != len(descendants):
raise error.Abort(_('split would leave orphaned changesets '
'behind'))
else:
if not alloworphaned and descendants:
raise error.Abort(
_('cannot split changeset with children without rebase'))
torebase = ()
if len(ctx.parents()) > 1:
raise error.Abort(_('cannot split a merge changeset'))
cmdutil.bailifchanged(repo)
# Deactivate bookmark temporarily so it won't get moved unintentionally
bname = repo._activebookmark
if bname and repo._bookmarks[bname] != ctx.node():
bookmarks.deactivate(repo)
wnode = repo['.'].node()
top = None
try:
top = dosplit(ui, repo, tr, ctx, opts)
finally:
# top is None: split failed, need update --clean recovery.
# wnode == ctx.node(): wnode split, no need to update.
if top is None or wnode != ctx.node():
hg.clean(repo, wnode, show_stats=False)
if bname:
bookmarks.activate(repo, bname)
if torebase and top:
dorebase(ui, repo, torebase, top)
def dosplit(ui, repo, tr, ctx, opts):
committed = [] # [ctx]
# Set working parent to ctx.p1(), and keep working copy as ctx's content
if ctx.node() != repo.dirstate.p1():
hg.clean(repo, ctx.node(), show_stats=False)
with repo.dirstate.parentchange():
scmutil.movedirstate(repo, ctx.p1())
# Any modified, added, removed, deleted result means split is incomplete
incomplete = lambda repo: any(repo.status()[:4])
# Main split loop
while incomplete(repo):
if committed:
header = (_('HG: Splitting %s. So far it has been split into:\n')
% short(ctx.node()))
for c in committed:
firstline = c.description().split('\n', 1)[0]
header += _('HG: - %s: %s\n') % (short(c.node()), firstline)
header += _('HG: Write commit message for the next split '
'changeset.\n')
else:
header = _('HG: Splitting %s. Write commit message for the '
'first split changeset.\n') % short(ctx.node())
opts.update({
'edit': True,
'interactive': True,
'message': header + ctx.description(),
})
commands.commit(ui, repo, **pycompat.strkwargs(opts))
newctx = repo['.']
committed.append(newctx)
if not committed:
raise error.Abort(_('cannot split an empty revision'))
scmutil.cleanupnodes(repo, {ctx.node(): [c.node() for c in committed]},
operation='split', fixphase=True)
return committed[-1]
def dorebase(ui, repo, src, destctx):
rebase.rebase(ui, repo, rev=[revsetlang.formatspec('%ld', src)],
dest=revsetlang.formatspec('%d', destctx.rev()))