##// END OF EJS Templates
pathcopies: give up any optimization based on `introrev`...
pathcopies: give up any optimization based on `introrev` Between 8a0136f69027 and d98fb3f42f33, we sped up the search for the introduction revision during path copies. However, further checking show that finding the introduction revision is still expensive and that we are better off without it. So we simply drop it and only rely on the linkrev optimisation. I ran `perfpathcopies` on 6989 pair of revision in the pypy repository (`hg perfhelper-pathcopies`. The result is massively in favor of dropping this condition. The result of the copy tracing are unchanged. Attempt to use a smaller changes preserving linkrev usage were unsuccessful, it can return wrong result. The following changesets broke test-mv-cp-st-diff.t - if not f.isintroducedafter(limit): + if limit >= 0 and f.linkrev() < limit: return None Here are various numbers (before this changeset/after this changesets) source destination before after saved-time ratio worth cases e66f24650daf 695dfb0f493b 1.062843 1.246369 -0.183526 1.172675 c979853a3b6a 8d60fe293e79 1.036985 1.196414 -0.159429 1.153743 22349fa2fc33 fbb1c9fd86c0 0.879926 1.038682 -0.158756 1.180420 682b98f3e672 a4878080a536 0.909952 1.063801 -0.153849 1.169074 5adabc9b9848 920958a93997 0.993622 1.147452 -0.153830 1.154817 worse 1% dbfbfcf077e9 aea8f2fd3593 1.016595 1.082999 -0.066404 1.065320 worse 5% c95f1ced15f2 7d29d5e39734 0.453694 0.471156 -0.017462 1.038488 worse 10% 3e144ed1d5b7 2aef0e942480 0.035140 0.037535 -0.002395 1.068156 worse 25% 321fc60db035 801748ba582a 0.009267 0.009325 -0.000058 1.006259 median 2088ce763fc2 e6991321d78b 0.000665 0.000651 0.000014 0.978947 best 25% 915631a97de6 385b31354be6 0.040743 0.040363 0.000380 0.990673 best 10% ad495c36a765 19c10384d3e7 0.431658 0.411490 0.020168 0.953278 best 5% d13ae7d283ae 813c99f810ac 1.141404 1.075346 0.066058 0.942126 best 1% 81593cb4a496 99ae11866969 1.833297 0.063823 1.769474 0.034813 best cases c3b14617fbd7 743a0fcaa4eb 1101.811740 2.735970 1099.075770 0.002483 c3b14617fbd7 9ba6ab77fd29 1116.753953 2.800729 1113.953224 0.002508 058b99d6e81f 57e249b7a3ea 1246.128485 3.042762 1243.085723 0.002442 9a8c361aab49 0354a250d371 1253.111894 3.085796 1250.026098 0.002463 442dbbc53c68 3ec1002a818c 1261.786294 3.138607 1258.647687 0.002487 As one can see, the average case is not really impacted. However, the worth case we get after this changeset are much better than the one we had before it. We have 30 pairs where improvements are above 10 minutes. This reflect in the combined time for all pairs before: 26256s after: 1300s (-95%) If we remove these pathological 30 cases, we still see a significant improvements: before: 1631s after: 1245s (-24%)

File last commit:

r43347:687b865b default
r43469:c16fe77e default
Show More
darcs.py
242 lines | 8.4 KiB | text/x-python | PythonLexer
# darcs.py - darcs support for the convert extension
#
# Copyright 2007-2009 Matt Mackall <mpm@selenic.com> and others
#
# This software may be used and distributed according to the terms of the
# GNU General Public License version 2 or any later version.
from __future__ import absolute_import
import errno
import os
import re
import shutil
from mercurial.i18n import _
from mercurial import (
error,
pycompat,
util,
)
from mercurial.utils import dateutil
from . import common
NoRepo = common.NoRepo
# The naming drift of ElementTree is fun!
try:
import xml.etree.cElementTree.ElementTree as ElementTree
import xml.etree.cElementTree.XMLParser as XMLParser
except ImportError:
try:
import xml.etree.ElementTree.ElementTree as ElementTree
import xml.etree.ElementTree.XMLParser as XMLParser
except ImportError:
try:
import elementtree.cElementTree.ElementTree as ElementTree
import elementtree.cElementTree.XMLParser as XMLParser
except ImportError:
try:
import elementtree.ElementTree.ElementTree as ElementTree
import elementtree.ElementTree.XMLParser as XMLParser
except ImportError:
pass
class darcs_source(common.converter_source, common.commandline):
def __init__(self, ui, repotype, path, revs=None):
common.converter_source.__init__(self, ui, repotype, path, revs=revs)
common.commandline.__init__(self, ui, b'darcs')
# check for _darcs, ElementTree so that we can easily skip
# test-convert-darcs if ElementTree is not around
if not os.path.exists(os.path.join(path, b'_darcs')):
raise NoRepo(_(b"%s does not look like a darcs repository") % path)
common.checktool(b'darcs')
version = self.run0(b'--version').splitlines()[0].strip()
if version < b'2.1':
raise error.Abort(
_(b'darcs version 2.1 or newer needed (found %r)') % version
)
if b"ElementTree" not in globals():
raise error.Abort(_(b"Python ElementTree module is not available"))
self.path = os.path.realpath(path)
self.lastrev = None
self.changes = {}
self.parents = {}
self.tags = {}
# Check darcs repository format
format = self.format()
if format:
if format in (b'darcs-1.0', b'hashed'):
raise NoRepo(
_(
b"%s repository format is unsupported, "
b"please upgrade"
)
% format
)
else:
self.ui.warn(_(b'failed to detect repository format!'))
def before(self):
self.tmppath = pycompat.mkdtemp(
prefix=b'convert-' + os.path.basename(self.path) + b'-'
)
output, status = self.run(b'init', repodir=self.tmppath)
self.checkexit(status)
tree = self.xml(
b'changes', xml_output=True, summary=True, repodir=self.path
)
tagname = None
child = None
for elt in tree.findall(b'patch'):
node = elt.get(b'hash')
name = elt.findtext(b'name', b'')
if name.startswith(b'TAG '):
tagname = name[4:].strip()
elif tagname is not None:
self.tags[tagname] = node
tagname = None
self.changes[node] = elt
self.parents[child] = [node]
child = node
self.parents[child] = []
def after(self):
self.ui.debug(b'cleaning up %s\n' % self.tmppath)
shutil.rmtree(self.tmppath, ignore_errors=True)
def recode(self, s, encoding=None):
if isinstance(s, pycompat.unicode):
# XMLParser returns unicode objects for anything it can't
# encode into ASCII. We convert them back to str to get
# recode's normal conversion behavior.
s = s.encode('latin-1')
return super(darcs_source, self).recode(s, encoding)
def xml(self, cmd, **kwargs):
# NOTE: darcs is currently encoding agnostic and will print
# patch metadata byte-for-byte, even in the XML changelog.
etree = ElementTree()
# While we are decoding the XML as latin-1 to be as liberal as
# possible, etree will still raise an exception if any
# non-printable characters are in the XML changelog.
parser = XMLParser(encoding=b'latin-1')
p = self._run(cmd, **kwargs)
etree.parse(p.stdout, parser=parser)
p.wait()
self.checkexit(p.returncode)
return etree.getroot()
def format(self):
output, status = self.run(b'show', b'repo', repodir=self.path)
self.checkexit(status)
m = re.search(r'^\s*Format:\s*(.*)$', output, re.MULTILINE)
if not m:
return None
return b','.join(sorted(f.strip() for f in m.group(1).split(b',')))
def manifest(self):
man = []
output, status = self.run(
b'show', b'files', no_directories=True, repodir=self.tmppath
)
self.checkexit(status)
for line in output.split(b'\n'):
path = line[2:]
if path:
man.append(path)
return man
def getheads(self):
return self.parents[None]
def getcommit(self, rev):
elt = self.changes[rev]
dateformat = b'%a %b %d %H:%M:%S %Z %Y'
date = dateutil.strdate(elt.get(b'local_date'), dateformat)
desc = elt.findtext(b'name') + b'\n' + elt.findtext(b'comment', b'')
# etree can return unicode objects for name, comment, and author,
# so recode() is used to ensure str objects are emitted.
newdateformat = b'%Y-%m-%d %H:%M:%S %1%2'
return common.commit(
author=self.recode(elt.get(b'author')),
date=dateutil.datestr(date, newdateformat),
desc=self.recode(desc).strip(),
parents=self.parents[rev],
)
def pull(self, rev):
output, status = self.run(
b'pull',
self.path,
all=True,
match=b'hash %s' % rev,
no_test=True,
no_posthook=True,
external_merge=b'/bin/false',
repodir=self.tmppath,
)
if status:
if output.find(b'We have conflicts in') == -1:
self.checkexit(status, output)
output, status = self.run(b'revert', all=True, repodir=self.tmppath)
self.checkexit(status, output)
def getchanges(self, rev, full):
if full:
raise error.Abort(_(b"convert from darcs does not support --full"))
copies = {}
changes = []
man = None
for elt in self.changes[rev].find(b'summary').getchildren():
if elt.tag in (b'add_directory', b'remove_directory'):
continue
if elt.tag == b'move':
if man is None:
man = self.manifest()
source, dest = elt.get(b'from'), elt.get(b'to')
if source in man:
# File move
changes.append((source, rev))
changes.append((dest, rev))
copies[dest] = source
else:
# Directory move, deduce file moves from manifest
source = source + b'/'
for f in man:
if not f.startswith(source):
continue
fdest = dest + b'/' + f[len(source) :]
changes.append((f, rev))
changes.append((fdest, rev))
copies[fdest] = f
else:
changes.append((elt.text.strip(), rev))
self.pull(rev)
self.lastrev = rev
return sorted(changes), copies, set()
def getfile(self, name, rev):
if rev != self.lastrev:
raise error.Abort(_(b'internal calling inconsistency'))
path = os.path.join(self.tmppath, name)
try:
data = util.readfile(path)
mode = os.lstat(path).st_mode
except IOError as inst:
if inst.errno == errno.ENOENT:
return None, None
raise
mode = (mode & 0o111) and b'x' or b''
return data, mode
def gettags(self):
return self.tags