##// END OF EJS Templates
bdiff: replace hash algorithm...
bdiff: replace hash algorithm This patch replaces lyhash with the hash algorithm used by diffutils. The algorithm has its origins in Git commit 2e9d1410, which is all the way back from 1992. The license header in the code at that revision in GPL v2. I have not performed an extensive analysis of the distribution (and therefore buckets) of hash output. However, `hg perfbdiff` gives some clear wins. I'd like to think that if it is good enough for diffutils it is good enough for us? From the mozilla-unified repository: $ perfbdiff -m 3041e4d59df2 ! wall 0.053271 comb 0.060000 user 0.060000 sys 0.000000 (best of 100) ! wall 0.035827 comb 0.040000 user 0.040000 sys 0.000000 (best of 100) $ perfbdiff 0e9928989e9c --alldata --count 100 ! wall 6.204277 comb 6.200000 user 6.200000 sys 0.000000 (best of 3) ! wall 4.309710 comb 4.300000 user 4.300000 sys 0.000000 (best of 3) From the hg repo: $ perfbdiff 35000 --alldata --count 1000 ! wall 0.660358 comb 0.660000 user 0.660000 sys 0.000000 (best of 15) ! wall 0.534092 comb 0.530000 user 0.530000 sys 0.000000 (best of 19) Looking at the generated assembly and statistical profiler output from the kernel level, I believe there is room to make this function even faster. Namely, we're still consuming data character by character instead of at the word level. This translates to more loop iterations and more instructions. At this juncture though, the real performance killer is that we're hashing every line. We should get a significant speedup if we change the algorithm to find the longest prefix, longest suffix, treat those as single "lines" and then only do the line splitting and hashing on the parts that are different. That will require a lot of C code, however. I'm optimistic this approach could result in a ~2x speedup.

File last commit:

r29841:d5883fd0 default
r30318:e1d6aa0e default
Show More
win32text.py
183 lines | 5.9 KiB | text/x-python | PythonLexer
# win32text.py - LF <-> CRLF/CR translation utilities for Windows/Mac users
#
# Copyright 2005, 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.
'''perform automatic newline conversion (DEPRECATED)
Deprecation: The win32text extension requires each user to configure
the extension again and again for each clone since the configuration
is not copied when cloning.
We have therefore made the ``eol`` as an alternative. The ``eol``
uses a version controlled file for its configuration and each clone
will therefore use the right settings from the start.
To perform automatic newline conversion, use::
[extensions]
win32text =
[encode]
** = cleverencode:
# or ** = macencode:
[decode]
** = cleverdecode:
# or ** = macdecode:
If not doing conversion, to make sure you do not commit CRLF/CR by accident::
[hooks]
pretxncommit.crlf = python:hgext.win32text.forbidcrlf
# or pretxncommit.cr = python:hgext.win32text.forbidcr
To do the same check on a server to prevent CRLF/CR from being
pushed or pulled::
[hooks]
pretxnchangegroup.crlf = python:hgext.win32text.forbidcrlf
# or pretxnchangegroup.cr = python:hgext.win32text.forbidcr
'''
from __future__ import absolute_import
import re
from mercurial.i18n import _
from mercurial.node import (
short,
)
from mercurial import (
util,
)
# 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'
# regexp for single LF without CR preceding.
re_single_lf = re.compile('(^|[^\r])\n', re.MULTILINE)
newlinestr = {'\r\n': 'CRLF', '\r': 'CR'}
filterstr = {'\r\n': 'clever', '\r': 'mac'}
def checknewline(s, newline, ui=None, repo=None, filename=None):
# warn if already has 'newline' in repository.
# it might cause unexpected eol conversion.
# see issue 302:
# https://bz.mercurial-scm.org/302
if newline in s and ui and filename and repo:
ui.warn(_('WARNING: %s already has %s line endings\n'
'and does not need EOL conversion by the win32text plugin.\n'
'Before your next commit, please reconsider your '
'encode/decode settings in \nMercurial.ini or %s.\n') %
(filename, newlinestr[newline], repo.join('hgrc')))
def dumbdecode(s, cmd, **kwargs):
checknewline(s, '\r\n', **kwargs)
# replace single LF to CRLF
return re_single_lf.sub('\\1\r\n', s)
def dumbencode(s, cmd):
return s.replace('\r\n', '\n')
def macdumbdecode(s, cmd, **kwargs):
checknewline(s, '\r', **kwargs)
return s.replace('\n', '\r')
def macdumbencode(s, cmd):
return s.replace('\r', '\n')
def cleverdecode(s, cmd, **kwargs):
if not util.binary(s):
return dumbdecode(s, cmd, **kwargs)
return s
def cleverencode(s, cmd):
if not util.binary(s):
return dumbencode(s, cmd)
return s
def macdecode(s, cmd, **kwargs):
if not util.binary(s):
return macdumbdecode(s, cmd, **kwargs)
return s
def macencode(s, cmd):
if not util.binary(s):
return macdumbencode(s, cmd)
return s
_filters = {
'dumbdecode:': dumbdecode,
'dumbencode:': dumbencode,
'cleverdecode:': cleverdecode,
'cleverencode:': cleverencode,
'macdumbdecode:': macdumbdecode,
'macdumbencode:': macdumbencode,
'macdecode:': macdecode,
'macencode:': macencode,
}
def forbidnewline(ui, repo, hooktype, node, newline, **kwargs):
halt = False
seen = set()
# we try to walk changesets in reverse order from newest to
# oldest, so that if we see a file multiple times, we take the
# newest version as canonical. this prevents us from blocking a
# changegroup that contains an unacceptable commit followed later
# by a commit that fixes the problem.
tip = repo['tip']
for rev in xrange(len(repo) - 1, repo[node].rev() - 1, -1):
c = repo[rev]
for f in c.files():
if f in seen or f not in tip or f not in c:
continue
seen.add(f)
data = c[f].data()
if not util.binary(data) and newline in data:
if not halt:
ui.warn(_('attempt to commit or push text file(s) '
'using %s line endings\n') %
newlinestr[newline])
ui.warn(_('in %s: %s\n') % (short(c.node()), f))
halt = True
if halt and hooktype == 'pretxnchangegroup':
crlf = newlinestr[newline].lower()
filter = filterstr[newline]
ui.warn(_('\nTo prevent this mistake in your local repository,\n'
'add to Mercurial.ini or .hg/hgrc:\n'
'\n'
'[hooks]\n'
'pretxncommit.%s = python:hgext.win32text.forbid%s\n'
'\n'
'and also consider adding:\n'
'\n'
'[extensions]\n'
'win32text =\n'
'[encode]\n'
'** = %sencode:\n'
'[decode]\n'
'** = %sdecode:\n') % (crlf, crlf, filter, filter))
return halt
def forbidcrlf(ui, repo, hooktype, node, **kwargs):
return forbidnewline(ui, repo, hooktype, node, '\r\n', **kwargs)
def forbidcr(ui, repo, hooktype, node, **kwargs):
return forbidnewline(ui, repo, hooktype, node, '\r', **kwargs)
def reposetup(ui, repo):
if not repo.local():
return
for name, fn in _filters.iteritems():
repo.adddatafilter(name, fn)
def extsetup(ui):
# deprecated config: win32text.warn
if ui.configbool('win32text', 'warn', True):
ui.warn(_("win32text is deprecated: "
"https://mercurial-scm.org/wiki/Win32TextExtension\n"))