##// 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:

r40105:fd4d59cf default
r42408:57203e02 default
Show More
remotenames.py
403 lines | 13.5 KiB | text/x-python | PythonLexer
# remotenames.py - extension to display remotenames
#
# Copyright 2017 Augie Fackler <raf@durin42.com>
# Copyright 2017 Sean Farley <sean@farley.io>
#
# This software may be used and distributed according to the terms of the
# GNU General Public License version 2 or any later version.
""" showing remotebookmarks and remotebranches in UI (EXPERIMENTAL)
By default both remotebookmarks and remotebranches are turned on. Config knob to
control the individually are as follows.
Config options to tweak the default behaviour:
remotenames.bookmarks
Boolean value to enable or disable showing of remotebookmarks (default: True)
remotenames.branches
Boolean value to enable or disable showing of remotebranches (default: True)
remotenames.hoistedpeer
Name of the peer whose remotebookmarks should be hoisted into the top-level
namespace (default: 'default')
"""
from __future__ import absolute_import
from mercurial.i18n import _
from mercurial.node import (
bin,
)
from mercurial import (
bookmarks,
error,
extensions,
logexchange,
namespaces,
pycompat,
registrar,
revsetlang,
smartset,
templateutil,
util,
)
from mercurial.utils import (
stringutil,
)
if pycompat.ispy3:
import collections.abc
mutablemapping = collections.abc.MutableMapping
else:
import collections
mutablemapping = collections.MutableMapping
# 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'
configtable = {}
configitem = registrar.configitem(configtable)
templatekeyword = registrar.templatekeyword()
revsetpredicate = registrar.revsetpredicate()
configitem('remotenames', 'bookmarks',
default=True,
)
configitem('remotenames', 'branches',
default=True,
)
configitem('remotenames', 'hoistedpeer',
default='default',
)
class lazyremotenamedict(mutablemapping):
"""
Read-only dict-like Class to lazily resolve remotename entries
We are doing that because remotenames startup was slow.
We lazily read the remotenames file once to figure out the potential entries
and store them in self.potentialentries. Then when asked to resolve an
entry, if it is not in self.potentialentries, then it isn't there, if it
is in self.potentialentries we resolve it and store the result in
self.cache. We cannot be lazy is when asked all the entries (keys).
"""
def __init__(self, kind, repo):
self.cache = {}
self.potentialentries = {}
self._kind = kind # bookmarks or branches
self._repo = repo
self.loaded = False
def _load(self):
""" Read the remotenames file, store entries matching selected kind """
self.loaded = True
repo = self._repo
for node, rpath, rname in logexchange.readremotenamefile(repo,
self._kind):
name = rpath + '/' + rname
self.potentialentries[name] = (node, rpath, name)
def _resolvedata(self, potentialentry):
""" Check that the node for potentialentry exists and return it """
if not potentialentry in self.potentialentries:
return None
node, remote, name = self.potentialentries[potentialentry]
repo = self._repo
binnode = bin(node)
# if the node doesn't exist, skip it
try:
repo.changelog.rev(binnode)
except LookupError:
return None
# Skip closed branches
if (self._kind == 'branches' and repo[binnode].closesbranch()):
return None
return [binnode]
def __getitem__(self, key):
if not self.loaded:
self._load()
val = self._fetchandcache(key)
if val is not None:
return val
else:
raise KeyError()
def __iter__(self):
return iter(self.potentialentries)
def __len__(self):
return len(self.potentialentries)
def __setitem__(self):
raise NotImplementedError
def __delitem__(self):
raise NotImplementedError
def _fetchandcache(self, key):
if key in self.cache:
return self.cache[key]
val = self._resolvedata(key)
if val is not None:
self.cache[key] = val
return val
else:
return None
def keys(self):
""" Get a list of bookmark or branch names """
if not self.loaded:
self._load()
return self.potentialentries.keys()
def iteritems(self):
""" Iterate over (name, node) tuples """
if not self.loaded:
self._load()
for k, vtup in self.potentialentries.iteritems():
yield (k, [bin(vtup[0])])
class remotenames(object):
"""
This class encapsulates all the remotenames state. It also contains
methods to access that state in convenient ways. Remotenames are lazy
loaded. Whenever client code needs to ensure the freshest copy of
remotenames, use the `clearnames` method to force an eventual load.
"""
def __init__(self, repo, *args):
self._repo = repo
self.clearnames()
def clearnames(self):
""" Clear all remote names state """
self.bookmarks = lazyremotenamedict("bookmarks", self._repo)
self.branches = lazyremotenamedict("branches", self._repo)
self._invalidatecache()
def _invalidatecache(self):
self._nodetobmarks = None
self._nodetobranch = None
self._hoisttonodes = None
self._nodetohoists = None
def bmarktonodes(self):
return self.bookmarks
def nodetobmarks(self):
if not self._nodetobmarks:
bmarktonodes = self.bmarktonodes()
self._nodetobmarks = {}
for name, node in bmarktonodes.iteritems():
self._nodetobmarks.setdefault(node[0], []).append(name)
return self._nodetobmarks
def branchtonodes(self):
return self.branches
def nodetobranch(self):
if not self._nodetobranch:
branchtonodes = self.branchtonodes()
self._nodetobranch = {}
for name, nodes in branchtonodes.iteritems():
for node in nodes:
self._nodetobranch.setdefault(node, []).append(name)
return self._nodetobranch
def hoisttonodes(self, hoist):
if not self._hoisttonodes:
marktonodes = self.bmarktonodes()
self._hoisttonodes = {}
hoist += '/'
for name, node in marktonodes.iteritems():
if name.startswith(hoist):
name = name[len(hoist):]
self._hoisttonodes[name] = node
return self._hoisttonodes
def nodetohoists(self, hoist):
if not self._nodetohoists:
marktonodes = self.bmarktonodes()
self._nodetohoists = {}
hoist += '/'
for name, node in marktonodes.iteritems():
if name.startswith(hoist):
name = name[len(hoist):]
self._nodetohoists.setdefault(node[0], []).append(name)
return self._nodetohoists
def wrapprintbookmarks(orig, ui, repo, fm, bmarks):
if 'remotebookmarks' not in repo.names:
return
ns = repo.names['remotebookmarks']
for name in ns.listnames(repo):
nodes = ns.nodes(repo, name)
if not nodes:
continue
node = nodes[0]
bmarks[name] = (node, ' ', '')
return orig(ui, repo, fm, bmarks)
def extsetup(ui):
extensions.wrapfunction(bookmarks, '_printbookmarks', wrapprintbookmarks)
def reposetup(ui, repo):
# set the config option to store remotenames
repo.ui.setconfig('experimental', 'remotenames', True, 'remotenames-ext')
if not repo.local():
return
repo._remotenames = remotenames(repo)
ns = namespaces.namespace
if ui.configbool('remotenames', 'bookmarks'):
remotebookmarkns = ns(
'remotebookmarks',
templatename='remotebookmarks',
colorname='remotebookmark',
logfmt='remote bookmark: %s\n',
listnames=lambda repo: repo._remotenames.bmarktonodes().keys(),
namemap=lambda repo, name:
repo._remotenames.bmarktonodes().get(name, []),
nodemap=lambda repo, node:
repo._remotenames.nodetobmarks().get(node, []))
repo.names.addnamespace(remotebookmarkns)
# hoisting only works if there are remote bookmarks
hoist = ui.config('remotenames', 'hoistedpeer')
if hoist:
hoistednamens = ns(
'hoistednames',
templatename='hoistednames',
colorname='hoistedname',
logfmt='hoisted name: %s\n',
listnames = lambda repo:
repo._remotenames.hoisttonodes(hoist).keys(),
namemap = lambda repo, name:
repo._remotenames.hoisttonodes(hoist).get(name, []),
nodemap = lambda repo, node:
repo._remotenames.nodetohoists(hoist).get(node, []))
repo.names.addnamespace(hoistednamens)
if ui.configbool('remotenames', 'branches'):
remotebranchns = ns(
'remotebranches',
templatename='remotebranches',
colorname='remotebranch',
logfmt='remote branch: %s\n',
listnames = lambda repo: repo._remotenames.branchtonodes().keys(),
namemap = lambda repo, name:
repo._remotenames.branchtonodes().get(name, []),
nodemap = lambda repo, node:
repo._remotenames.nodetobranch().get(node, []))
repo.names.addnamespace(remotebranchns)
@templatekeyword('remotenames', requires={'repo', 'ctx'})
def remotenameskw(context, mapping):
"""List of strings. Remote names associated with the changeset."""
repo = context.resource(mapping, 'repo')
ctx = context.resource(mapping, 'ctx')
remotenames = []
if 'remotebookmarks' in repo.names:
remotenames = repo.names['remotebookmarks'].names(repo, ctx.node())
if 'remotebranches' in repo.names:
remotenames += repo.names['remotebranches'].names(repo, ctx.node())
return templateutil.compatlist(context, mapping, 'remotename', remotenames,
plural='remotenames')
@templatekeyword('remotebookmarks', requires={'repo', 'ctx'})
def remotebookmarkskw(context, mapping):
"""List of strings. Remote bookmarks associated with the changeset."""
repo = context.resource(mapping, 'repo')
ctx = context.resource(mapping, 'ctx')
remotebmarks = []
if 'remotebookmarks' in repo.names:
remotebmarks = repo.names['remotebookmarks'].names(repo, ctx.node())
return templateutil.compatlist(context, mapping, 'remotebookmark',
remotebmarks, plural='remotebookmarks')
@templatekeyword('remotebranches', requires={'repo', 'ctx'})
def remotebrancheskw(context, mapping):
"""List of strings. Remote branches associated with the changeset."""
repo = context.resource(mapping, 'repo')
ctx = context.resource(mapping, 'ctx')
remotebranches = []
if 'remotebranches' in repo.names:
remotebranches = repo.names['remotebranches'].names(repo, ctx.node())
return templateutil.compatlist(context, mapping, 'remotebranch',
remotebranches, plural='remotebranches')
def _revsetutil(repo, subset, x, rtypes):
"""utility function to return a set of revs based on the rtypes"""
args = revsetlang.getargs(x, 0, 1, _('only one argument accepted'))
if args:
kind, pattern, matcher = stringutil.stringmatcher(
revsetlang.getstring(args[0], _('argument must be a string')))
else:
kind = pattern = None
matcher = util.always
nodes = set()
cl = repo.changelog
for rtype in rtypes:
if rtype in repo.names:
ns = repo.names[rtype]
for name in ns.listnames(repo):
if not matcher(name):
continue
nodes.update(ns.nodes(repo, name))
if kind == 'literal' and not nodes:
raise error.RepoLookupError(_("remote name '%s' does not exist")
% pattern)
revs = (cl.rev(n) for n in nodes if cl.hasnode(n))
return subset & smartset.baseset(revs)
@revsetpredicate('remotenames([name])')
def remotenamesrevset(repo, subset, x):
"""All changesets which have a remotename on them. If `name` is
specified, only remotenames of matching remote paths are considered.
Pattern matching is supported for `name`. See :hg:`help revisions.patterns`.
"""
return _revsetutil(repo, subset, x, ('remotebookmarks', 'remotebranches'))
@revsetpredicate('remotebranches([name])')
def remotebranchesrevset(repo, subset, x):
"""All changesets which are branch heads on remotes. If `name` is
specified, only remotenames of matching remote paths are considered.
Pattern matching is supported for `name`. See :hg:`help revisions.patterns`.
"""
return _revsetutil(repo, subset, x, ('remotebranches',))
@revsetpredicate('remotebookmarks([name])')
def remotebmarksrevset(repo, subset, x):
"""All changesets which have bookmarks on remotes. If `name` is
specified, only remotenames of matching remote paths are considered.
Pattern matching is supported for `name`. See :hg:`help revisions.patterns`.
"""
return _revsetutil(repo, subset, x, ('remotebookmarks',))