view mercurial/diffutil.py @ 38732:be4984261611

merge: mark file gets as not thread safe (issue5933)

In default installs, this has the effect of disabling the thread-based
worker on Windows when manifesting files in the working directory. My
measurements have shown that with revlog-based repositories, Mercurial
spends a lot of CPU time in revlog code resolving file data. This ends
up incurring a lot of context switching across threads and slows down
`hg update` operations when going from an empty working directory to
the tip of the repo.

On mozilla-unified (246,351 files) on an i7-6700K (4+4 CPUs):

before: 487s wall
after:  360s wall (equivalent to worker.enabled=false)
cpus=2: 379s wall

Even with only 2 threads, the thread pool is still slower.

The introduction of the thread-based worker (02b36e860e0b) states that
it resulted in a "~50%" speedup for `hg sparse --enable-profile` and
`hg sparse --disable-profile`. This disagrees with my measurement
above. I theorize a few reasons for this:

1) Removal of files from the working directory is I/O - not CPU - bound
   and should benefit from a thread pool (unless I/O is insanely fast
   and the GIL release is near instantaneous). So tests like `hg sparse
   --enable-profile` may exercise deletion throughput and aren't good
   benchmarks for worker tasks that are CPU heavy.
2) The patch was authored by someone at Facebook. The results were
   likely measured against a repository using remotefilelog. And I
   believe that revision retrieval during working directory updates with
   remotefilelog will often use a remote store, thus being I/O and not
   CPU bound. This probably resulted in an overstated performance gain.

Since there appears to be a need to enable the thread-based worker with
some stores, I've made the flagging of file gets as thread safe
configurable. I've made it experimental because I don't want to formalize
a boolean flag for this option and because this attribute is best
captured against the store implementation. But we don't have a proper
store API for this yet. I'd rather cross this bridge later.

It is possible there are revlog-based repositories that do benefit from
a thread-based worker. I didn't do very comprehensive testing. If there
are, we may want to devise a more proper algorithm for whether to use
the thread-based worker, including possibly config options to limit the
number of threads to use. But until I see evidence that justifies
complexity, simplicity wins.

Differential Revision: https://phab.mercurial-scm.org/D3963

# diffutil.py - utility functions related to diff and patch
#
# Copyright 2006 Brendan Cully <[e-mail]>
# Copyright 2007 Chris Mason <[e-mail]>
# Copyright 2018 Octobus <[e-mail]>
#
# 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

from .i18n import _

from . import (
    mdiff,
    pycompat,
)

def diffallopts(ui, opts=None, untrusted=False, section='diff'):
    '''return diffopts with all features supported and parsed'''
    return difffeatureopts(ui, opts=opts, untrusted=untrusted, section=section,
                           git=True, whitespace=True, formatchanging=True)

def difffeatureopts(ui, opts=None, untrusted=False, section='diff', git=False,
                    whitespace=False, formatchanging=False):
    '''return diffopts with only opted-in features parsed

    Features:
    - git: git-style diffs
    - whitespace: whitespace options like ignoreblanklines and ignorews
    - formatchanging: options that will likely break or cause correctness issues
      with most diff parsers
    '''
    def get(key, name=None, getter=ui.configbool, forceplain=None):
        if opts:
            v = opts.get(key)
            # diffopts flags are either None-default (which is passed
            # through unchanged, so we can identify unset values), or
            # some other falsey default (eg --unified, which defaults
            # to an empty string). We only want to override the config
            # entries from hgrc with command line values if they
            # appear to have been set, which is any truthy value,
            # True, or False.
            if v or isinstance(v, bool):
                return v
        if forceplain is not None and ui.plain():
            return forceplain
        return getter(section, name or key, untrusted=untrusted)

    # core options, expected to be understood by every diff parser
    buildopts = {
        'nodates': get('nodates'),
        'showfunc': get('show_function', 'showfunc'),
        'context': get('unified', getter=ui.config),
    }
    buildopts['xdiff'] = ui.configbool('experimental', 'xdiff')

    if git:
        buildopts['git'] = get('git')

        # since this is in the experimental section, we need to call
        # ui.configbool directory
        buildopts['showsimilarity'] = ui.configbool('experimental',
                                                    'extendedheader.similarity')

        # need to inspect the ui object instead of using get() since we want to
        # test for an int
        hconf = ui.config('experimental', 'extendedheader.index')
        if hconf is not None:
            hlen = None
            try:
                # the hash config could be an integer (for length of hash) or a
                # word (e.g. short, full, none)
                hlen = int(hconf)
                if hlen < 0 or hlen > 40:
                    msg = _("invalid length for extendedheader.index: '%d'\n")
                    ui.warn(msg % hlen)
            except ValueError:
                # default value
                if hconf == 'short' or hconf == '':
                    hlen = 12
                elif hconf == 'full':
                    hlen = 40
                elif hconf != 'none':
                    msg = _("invalid value for extendedheader.index: '%s'\n")
                    ui.warn(msg % hconf)
            finally:
                buildopts['index'] = hlen

    if whitespace:
        buildopts['ignorews'] = get('ignore_all_space', 'ignorews')
        buildopts['ignorewsamount'] = get('ignore_space_change',
                                          'ignorewsamount')
        buildopts['ignoreblanklines'] = get('ignore_blank_lines',
                                            'ignoreblanklines')
        buildopts['ignorewseol'] = get('ignore_space_at_eol', 'ignorewseol')
    if formatchanging:
        buildopts['text'] = opts and opts.get('text')
        binary = None if opts is None else opts.get('binary')
        buildopts['nobinary'] = (not binary if binary is not None
                                 else get('nobinary', forceplain=False))
        buildopts['noprefix'] = get('noprefix', forceplain=False)
        buildopts['worddiff'] = get('word_diff', 'word-diff', forceplain=False)

    return mdiff.diffopts(**pycompat.strkwargs(buildopts))