##// END OF EJS Templates
util: teach lrucachedict to enforce a max total cost...
util: teach lrucachedict to enforce a max total cost Now that lrucachedict entries can have a numeric cost associated with them and we can easily pop the oldest item in the cache, it now becomes relatively trivial to implement support for enforcing a high water mark on the total cost of items in the cache. This commit teaches lrucachedict instances to have a max cost associated with them. When items are inserted, we pop old items until enough "cost" frees up to make room for the new item. This feature is close to zero cost when not used (modulo the insertion regressed introduced by the previous commit): $ ./hg perflrucachedict --size 4 --gets 1000000 --sets 1000000 --mixed 1000000 ! gets ! wall 0.607444 comb 0.610000 user 0.610000 sys 0.000000 (best of 17) ! wall 0.601653 comb 0.600000 user 0.600000 sys 0.000000 (best of 17) ! inserts ! wall 0.678261 comb 0.680000 user 0.680000 sys 0.000000 (best of 14) ! wall 0.685042 comb 0.680000 user 0.680000 sys 0.000000 (best of 15) ! sets ! wall 0.808770 comb 0.800000 user 0.800000 sys 0.000000 (best of 13) ! wall 0.834241 comb 0.830000 user 0.830000 sys 0.000000 (best of 12) ! mixed ! wall 0.782441 comb 0.780000 user 0.780000 sys 0.000000 (best of 13) ! wall 0.803804 comb 0.800000 user 0.800000 sys 0.000000 (best of 13) $ hg perflrucachedict --size 1000 --gets 1000000 --sets 1000000 --mixed 1000000 ! init ! wall 0.006952 comb 0.010000 user 0.010000 sys 0.000000 (best of 418) ! gets ! wall 0.613350 comb 0.610000 user 0.610000 sys 0.000000 (best of 17) ! wall 0.617415 comb 0.620000 user 0.620000 sys 0.000000 (best of 17) ! inserts ! wall 0.701270 comb 0.700000 user 0.700000 sys 0.000000 (best of 15) ! wall 0.700516 comb 0.700000 user 0.700000 sys 0.000000 (best of 15) ! sets ! wall 0.825720 comb 0.830000 user 0.830000 sys 0.000000 (best of 13) ! wall 0.837946 comb 0.840000 user 0.830000 sys 0.010000 (best of 12) ! mixed ! wall 0.821644 comb 0.820000 user 0.820000 sys 0.000000 (best of 13) ! wall 0.850559 comb 0.850000 user 0.850000 sys 0.000000 (best of 12) I reckon the slight slowdown on insert is due to added if checks. For caches with total cost limiting enabled: $ hg perflrucachedict --size 4 --gets 1000000 --sets 1000000 --mixed 1000000 --costlimit 100 ! gets w/ cost limit ! wall 0.598737 comb 0.590000 user 0.590000 sys 0.000000 (best of 17) ! inserts w/ cost limit ! wall 1.694282 comb 1.700000 user 1.700000 sys 0.000000 (best of 6) ! mixed w/ cost limit ! wall 1.157655 comb 1.150000 user 1.150000 sys 0.000000 (best of 9) $ hg perflrucachedict --size 1000 --gets 1000000 --sets 1000000 --mixed 1000000 --costlimit 10000 ! gets w/ cost limit ! wall 0.598526 comb 0.600000 user 0.600000 sys 0.000000 (best of 17) ! inserts w/ cost limit ! wall 37.838315 comb 37.840000 user 37.840000 sys 0.000000 (best of 3) ! mixed w/ cost limit ! wall 18.060198 comb 18.060000 user 18.060000 sys 0.000000 (best of 3) $ hg perflrucachedict --size 1000 --gets 1000000 --sets 1000000 --mixed 1000000 --costlimit 10000 --mixedgetfreq 90 ! gets w/ cost limit ! wall 0.600024 comb 0.600000 user 0.600000 sys 0.000000 (best of 17) ! inserts w/ cost limit ! wall 37.154547 comb 37.120000 user 37.120000 sys 0.000000 (best of 3) ! mixed w/ cost limit ! wall 4.381602 comb 4.380000 user 4.370000 sys 0.010000 (best of 3) The functions we're benchmarking are slightly different, which could move numbers by a few milliseconds. But the slowdown on insert is too great to be explained by that. The slowness is due to insert heavy operations needing to call popoldest() repeatedly when the cache is at capacity. The next commit will address this. Differential Revision: https://phab.mercurial-scm.org/D4503

File last commit:

r37483:39e5e346 default
r39604:842cd0bd default
Show More
fancyopts.py
378 lines | 11.2 KiB | text/x-python | PythonLexer
# 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 abc
import functools
from .i18n import _
from . import (
error,
pycompat,
)
# Set of flags to not apply boolean negation logic on
nevernegate = {
# 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 _earlyoptarg(arg, shortlist, namelist):
"""Check if the given arg is a valid unabbreviated option
Returns (flag_str, has_embedded_value?, embedded_value, takes_value?)
>>> def opt(arg):
... return _earlyoptarg(arg, b'R:q', [b'cwd=', b'debugger'])
long form:
>>> opt(b'--cwd')
('--cwd', False, '', True)
>>> opt(b'--cwd=')
('--cwd', True, '', True)
>>> opt(b'--cwd=foo')
('--cwd', True, 'foo', True)
>>> opt(b'--debugger')
('--debugger', False, '', False)
>>> opt(b'--debugger=') # invalid but parsable
('--debugger', True, '', False)
short form:
>>> opt(b'-R')
('-R', False, '', True)
>>> opt(b'-Rfoo')
('-R', True, 'foo', True)
>>> opt(b'-q')
('-q', False, '', False)
>>> opt(b'-qfoo') # invalid but parsable
('-q', True, 'foo', False)
unknown or invalid:
>>> opt(b'--unknown')
('', False, '', False)
>>> opt(b'-u')
('', False, '', False)
>>> opt(b'-ufoo')
('', False, '', False)
>>> opt(b'--')
('', False, '', False)
>>> opt(b'-')
('', False, '', False)
>>> opt(b'-:')
('', False, '', False)
>>> opt(b'-:foo')
('', False, '', False)
"""
if arg.startswith('--'):
flag, eq, val = arg.partition('=')
if flag[2:] in namelist:
return flag, bool(eq), val, False
if flag[2:] + '=' in namelist:
return flag, bool(eq), val, True
elif arg.startswith('-') and arg != '-' and not arg.startswith('-:'):
flag, val = arg[:2], arg[2:]
i = shortlist.find(flag[1:])
if i >= 0:
return flag, bool(val), val, shortlist.startswith(':', i + 1)
return '', False, '', False
def earlygetopt(args, shortlist, namelist, gnu=False, keepsep=False):
"""Parse options like getopt, but ignores unknown options and abbreviated
forms
If gnu=False, this stops processing options as soon as a non/unknown-option
argument is encountered. Otherwise, option and non-option arguments may be
intermixed, and unknown-option arguments are taken as non-option.
If keepsep=True, '--' won't be removed from the list of arguments left.
This is useful for stripping early options from a full command arguments.
>>> def get(args, gnu=False, keepsep=False):
... return earlygetopt(args, b'R:q', [b'cwd=', b'debugger'],
... gnu=gnu, keepsep=keepsep)
default parsing rules for early options:
>>> get([b'x', b'--cwd', b'foo', b'-Rbar', b'-q', b'y'], gnu=True)
([('--cwd', 'foo'), ('-R', 'bar'), ('-q', '')], ['x', 'y'])
>>> get([b'x', b'--cwd=foo', b'y', b'-R', b'bar', b'--debugger'], gnu=True)
([('--cwd', 'foo'), ('-R', 'bar'), ('--debugger', '')], ['x', 'y'])
>>> get([b'--unknown', b'--cwd=foo', b'--', '--debugger'], gnu=True)
([('--cwd', 'foo')], ['--unknown', '--debugger'])
restricted parsing rules (early options must come first):
>>> get([b'--cwd', b'foo', b'-Rbar', b'x', b'-q', b'y'], gnu=False)
([('--cwd', 'foo'), ('-R', 'bar')], ['x', '-q', 'y'])
>>> get([b'--cwd=foo', b'x', b'y', b'-R', b'bar', b'--debugger'], gnu=False)
([('--cwd', 'foo')], ['x', 'y', '-R', 'bar', '--debugger'])
>>> get([b'--unknown', b'--cwd=foo', b'--', '--debugger'], gnu=False)
([], ['--unknown', '--cwd=foo', '--', '--debugger'])
stripping early options (without loosing '--'):
>>> get([b'x', b'-Rbar', b'--', '--debugger'], gnu=True, keepsep=True)[1]
['x', '--', '--debugger']
last argument:
>>> get([b'--cwd'])
([], ['--cwd'])
>>> get([b'--cwd=foo'])
([('--cwd', 'foo')], [])
>>> get([b'-R'])
([], ['-R'])
>>> get([b'-Rbar'])
([('-R', 'bar')], [])
>>> get([b'-q'])
([('-q', '')], [])
>>> get([b'-q', b'--'])
([('-q', '')], [])
'--' may be a value:
>>> get([b'-R', b'--', b'x'])
([('-R', '--')], ['x'])
>>> get([b'--cwd', b'--', b'x'])
([('--cwd', '--')], ['x'])
value passed to bool options:
>>> get([b'--debugger=foo', b'x'])
([], ['--debugger=foo', 'x'])
>>> get([b'-qfoo', b'x'])
([], ['-qfoo', 'x'])
short option isn't separated with '=':
>>> get([b'-R=bar'])
([('-R', '=bar')], [])
':' may be in shortlist, but shouldn't be taken as an option letter:
>>> get([b'-:', b'y'])
([], ['-:', 'y'])
'-' is a valid non-option argument:
>>> get([b'-', b'y'])
([], ['-', 'y'])
"""
parsedopts = []
parsedargs = []
pos = 0
while pos < len(args):
arg = args[pos]
if arg == '--':
pos += not keepsep
break
flag, hasval, val, takeval = _earlyoptarg(arg, shortlist, namelist)
if not hasval and takeval and pos + 1 >= len(args):
# missing last argument
break
if not flag or hasval and not takeval:
# non-option argument or -b/--bool=INVALID_VALUE
if gnu:
parsedargs.append(arg)
pos += 1
else:
break
elif hasval == takeval:
# -b/--bool or -s/--str=VALUE
parsedopts.append((flag, val))
pos += 1
else:
# -s/--str VALUE
parsedopts.append((flag, args[pos + 1]))
pos += 2
parsedargs.extend(args[pos:])
return parsedopts, parsedargs
class customopt(object):
"""Manage defaults and mutations for any type of opt."""
__metaclass__ = abc.ABCMeta
def __init__(self, defaultvalue):
self._defaultvalue = defaultvalue
def _isboolopt(self):
return False
def getdefaultvalue(self):
"""Returns the default value for this opt.
Subclasses should override this to return a new value if the value type
is mutable."""
return self._defaultvalue
@abc.abstractmethod
def newstate(self, oldstate, newparam, abort):
"""Adds newparam to oldstate and returns the new state.
On failure, abort can be called with a string error message."""
class _simpleopt(customopt):
def _isboolopt(self):
return isinstance(self._defaultvalue, (bool, type(None)))
def newstate(self, oldstate, newparam, abort):
return newparam
class _callableopt(customopt):
def __init__(self, callablefn):
self.callablefn = callablefn
super(_callableopt, self).__init__(None)
def newstate(self, oldstate, newparam, abort):
return self.callablefn(newparam)
class _listopt(customopt):
def getdefaultvalue(self):
return self._defaultvalue[:]
def newstate(self, oldstate, newparam, abort):
oldstate.append(newparam)
return oldstate
class _intopt(customopt):
def newstate(self, oldstate, newparam, abort):
try:
return int(newparam)
except ValueError:
abort(_('expected int'))
def _defaultopt(default):
"""Returns a default opt implementation, given a default value."""
if isinstance(default, customopt):
return default
elif callable(default):
return _callableopt(default)
elif isinstance(default, list):
return _listopt(default[:])
elif type(default) is type(1):
return _intopt(default)
else:
return _simpleopt(default)
def fancyopts(args, options, state, gnu=False, early=False, optaliases=None):
"""
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
customopt - subclass of 'customopt'
optaliases is a mapping from a canonical option name to a list of
additional long options. This exists for preserving backward compatibility
of early options. If we want to use it extensively, please consider moving
the functionality to the options table (e.g separate long options by '|'.)
non-option args are returned
"""
if optaliases is None:
optaliases = {}
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
onames = [name]
onames.extend(optaliases.get(name, []))
name = name.replace('-', '_')
argmap['-' + short] = name
for n in onames:
argmap['--' + n] = name
defmap[name] = _defaultopt(default)
# copy defaults to state
state[name] = defmap[name].getdefaultvalue()
# does it take a parameter?
if not defmap[name]._isboolopt():
if short:
short += ':'
onames = [n + '=' for n in onames]
elif name not in nevernegate:
for n in onames:
if n.startswith('no-'):
insert = n[3:]
else:
insert = 'no-' + n
# 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 ('--' + n) not in negations
negations['--' + insert] = '--' + n
namelist.append(insert)
if short:
shortlist += short
if name:
namelist.extend(onames)
# parse arguments
if early:
parse = functools.partial(earlygetopt, gnu=gnu)
elif gnu:
parse = pycompat.gnugetoptb
else:
parse = pycompat.getoptb
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]
if obj._isboolopt():
state[name] = boolval
else:
def abort(s):
raise error.Abort(_('invalid value %r for option %s, %s')
% (pycompat.maybebytestr(val), opt, s))
state[name] = defmap[name].newstate(state[name], val, abort)
# return unparsed args
return args