bdiff: early pruning of common prefix before doing expensive computations
It seems quite common that files don't change completely. New lines are often
pretty much appended, and modifications will often only change a small section
of the file which on average will be in the middle.
There can thus be a big win by pruning a common prefix before starting the more
expensive search for longest common substrings.
Worst case, it will scan through a long sequence of similar bytes without
encountering a newline. Splitlines will then have to do the same again ...
twice for each side. If similar lines are found, splitlines will save the
double iteration and hashing of the lines ... plus there will be less lines to
find common substrings in.
This change might in some cases make the algorith pick shorter or less optimal
common substrings. We can't have the cake and eat it.
This make hg --time bundle --base null -r 4.0 go from 14.5 to 15 s - a 3%
increase.
On mozilla-unified:
perfbdiff -m 3041e4d59df2
! wall 0.053088 comb 0.060000 user 0.060000 sys 0.000000 (best of 100) to
! wall 0.024618 comb 0.020000 user 0.020000 sys 0.000000 (best of 116)
perfbdiff 0e9928989e9c --alldata --count 10
! wall 0.702075 comb 0.700000 user 0.700000 sys 0.000000 (best of 15) to
! wall 0.579235 comb 0.580000 user 0.580000 sys 0.000000 (best of 18)
# fancyopts.py - better command line parsing
#
# Copyright 2005-2009 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 getopt
from .i18n import _
from . import error
# Set of flags to not apply boolean negation logic on
nevernegate = set([
# avoid --no-noninteractive
'noninteractive',
# These two flags are special because they cause hg to do one
# thing and then exit, and so aren't suitable for use in things
# like aliases anyway.
'help',
'version',
])
def gnugetopt(args, options, longoptions):
"""Parse options mostly like getopt.gnu_getopt.
This is different from getopt.gnu_getopt in that an argument of - will
become an argument of - instead of vanishing completely.
"""
extraargs = []
if '--' in args:
stopindex = args.index('--')
extraargs = args[stopindex + 1:]
args = args[:stopindex]
opts, parseargs = getopt.getopt(args, options, longoptions)
args = []
while parseargs:
arg = parseargs.pop(0)
if arg and arg[0] == '-' and len(arg) > 1:
parseargs.insert(0, arg)
topts, newparseargs = getopt.getopt(parseargs, options, longoptions)
opts = opts + topts
parseargs = newparseargs
else:
args.append(arg)
args.extend(extraargs)
return opts, args
def fancyopts(args, options, state, gnu=False):
"""
read args, parse options, and store options in state
each option is a tuple of:
short option or ''
long option
default value
description
option value label(optional)
option types include:
boolean or none - option sets variable in state to true
string - parameter string is stored in state
list - parameter string is added to a list
integer - parameter strings is stored as int
function - call function with parameter
non-option args are returned
"""
namelist = []
shortlist = ''
argmap = {}
defmap = {}
negations = {}
alllong = set(o[1] for o in options)
for option in options:
if len(option) == 5:
short, name, default, comment, dummy = option
else:
short, name, default, comment = option
# convert opts to getopt format
oname = name
name = name.replace('-', '_')
argmap['-' + short] = argmap['--' + oname] = name
defmap[name] = default
# copy defaults to state
if isinstance(default, list):
state[name] = default[:]
elif callable(default):
state[name] = None
else:
state[name] = default
# does it take a parameter?
if not (default is None or default is True or default is False):
if short:
short += ':'
if oname:
oname += '='
elif oname not in nevernegate:
if oname.startswith('no-'):
insert = oname[3:]
else:
insert = 'no-' + oname
# backout (as a practical example) has both --commit and
# --no-commit options, so we don't want to allow the
# negations of those flags.
if insert not in alllong:
assert ('--' + oname) not in negations
negations['--' + insert] = '--' + oname
namelist.append(insert)
if short:
shortlist += short
if name:
namelist.append(oname)
# parse arguments
if gnu:
parse = gnugetopt
else:
parse = getopt.getopt
opts, args = parse(args, shortlist, namelist)
# transfer result to state
for opt, val in opts:
boolval = True
negation = negations.get(opt, False)
if negation:
opt = negation
boolval = False
name = argmap[opt]
obj = defmap[name]
t = type(obj)
if callable(obj):
state[name] = defmap[name](val)
elif t is type(1):
try:
state[name] = int(val)
except ValueError:
raise error.Abort(_('invalid value %r for option %s, '
'expected int') % (val, opt))
elif t is type(''):
state[name] = val
elif t is type([]):
state[name].append(val)
elif t is type(None) or t is type(False):
state[name] = boolval
# return unparsed args
return args