view hgext/sparse.py @ 39506:b66ea3fc3a86

sparse-revlog: set max delta chain length to on thousand The new snapshot system used in the sparse-revlog case gave us some small size benefit so far. However its most important property is to gracefully handle harder limit on delta chainlength. Long delta chain has a very detrimental impact on read (and write) performance in revlog. Being able to shorter them provide a great boost. However, shorting delta used to result significantly lower compression ratio. The intermediate snapshots effectively suppress most of this effect (even all in some case). # Effect on the test repository The repository we use for test is not "realistic" but can still show this in action using an unreasonably low chain limit. Limiting the chain length show a sizeable increase but stay under control: +6% for limit=15; +15% for limit=10. Without the snapshot system the increase is significantly bigger: +45% for limit=15; +80% for limit=10. Even slightly larger than without delta chain limit, the resulting size is still smaller than before we started doing snapshots. Here is a table for comparison. *Since the repository is not branchy, the initial sparse-revlog version does not bring much benefit compare to the non-sparse one): chain length limit | none | limit=15 | limit=10 | without sparse-revlog | 62 818 987 | 112 664 615 | 131 222 574 | without snapshot | 74 365 490 | 108 211 410 | 133 857 764 | with snapshot | 59 230 936 | 63 002 924 | 68 415 329 | # Effect On Real Life Repositories The series provides significant benefits on all kind of repositories. Using `hg debugupgraderepo -o redeltaparent --run`, we recomputed delta chain for various repositories with different settings: - delta chain length: unlimited or 1000 limit - sparse-revlog: enabled or disabled - this series: applied or not applied We can observe multiple types of effect: - On very branchy repositories: * The delta chain limit as low impact on the repo size. * Intermediate snapshot greatly reduces manifest size: - pypy: -80% - netbeans: -95% * The delta chain limit is effective, without a size impact: - netbeans average: 613 -> 282 - private #1 average: 1 068 -> 307 - On more linear repository: * Intermediate snapshot limit the impact of delta chain limit: - mozilla: without the series: +360% with the series: +25% * The delta chain limit provides large improvement: - mozilla's average chain length: unlimited: 15 338 limited: 469 * Despite the chain length limit, the manifest size is reduced: - mercurial: -25% - mozilla: -30% It is clear that the use of chains of intermediate snapshots provide large benefits both in storage size and delta chains quality. We should now switch our effort toward making sure the write performance are acceptable. Then, `sparse-revlog` will be a suitable format for all new repository. # Raw Statistic * no-sparse: general delta repository not using sparse-revlog * no-snapshot: sparse-revlog repository not using this series * snapshot: sparse-revlog repository using this series mercurial Manifest Size: limit | none | 1000 ------------|-------------|------------ no-sparse | 8 021 373 | 8 199 366 no-snapshot | 8 103 561 | 8 259 719 snapshot | 6 137 116 | 6 126 433 Manifest Chain length data limit || none || 1000 || value || average | max || average | max || ------------||---------|---------||---------|---------|| no-sparse || 307 | 1456 || 279 | 1000 || no-snapshot || 312 | 1456 || 283 | 1000 || snapshot || 248 | 1208 || 241 | 1000 || Full Store Size limit | none | 1000 ------------|-------------|------------ no-sparse | 51 013 198 | 51 201 574 no-snapshot | 50 930 795 | 51 141 006 snapshot | 48 072 037 | 48 093 572 pypy Manifest Size: limit | none | 1000 ------------|-------------|------------ no-sparse | 193 987 784 | 193 987 784 no-snapshot | 163 171 745 | 163 312 229 snapshot | 34 605 900 | 34 600 750 Manifest Chain length data limit || none || 1000 || value || average | max || average | max || ------------||---------|---------||---------|---------|| no-sparse || 101 | 692 || 101 | 692 || no-snapshot || 151 | 1307 || 148 | 1000 || snapshot || 128 | 1309 || 125 | 1000 || Full Store Size limit | none | 1000 ------------|-------------|------------ no-sparse | 495 931 473 | 495 931 473 no-snapshot | 465 441 017 | 465 581 501 snapshot | 355 467 301 | 355 472 451 Mozilla Manifest Size: limit | none | 1000 ------------|----------------|--------------- no-sparse | 416 757 148 | 1 869 009 668 no-snapshot | 401 592 370 | 1 843 493 795 snapshot | 224 359 521 | 284 615 500 Manifest Chain length data limit || none || 1000 || value || average | max || average | max || ------------||---------|---------||---------|---------|| no-sparse || 15 333 | 58 980 || 468 | 1 000 || no-snapshot || 15 336 | 58 980 || 469 | 1 000 || snapshot || 15 338 | 58 983 || 469 | 1 000 || Full Store Size limit | none | 1000 ------------|----------------|--------------- no-sparse | 2 712 477 887 | 4 164 995 451 no-snapshot | 2 698 887 835 | 4 141 054 304 snapshot | 2 518 130 385 | 2 578 587 596 Netbeans Manifest Size: limit | none | 1000 ------------|----------------|--------------- no-sparse | 4 766 794 101 | 4 870 642 687 no-snapshot | 4 334 806 082 | 4 428 681 309 snapshot | 232 659 666 | 240 330 665 Manifest Chain length data limit || none || 1000 || value || average | max || average | max || ------------||---------|---------||---------|---------|| no-sparse || 597 | 6802 || 254 | 1 000 || no-snapshot || 648 | 6 802 || 305 | 1 000 || snapshot || 613 | 6 804 || 282 | 1 000 || Full Store Size limit | none | 1000 ------------|----------------|--------------- no-sparse | 5 807 347 998 | 5 911 196 584 no-snapshot | 5 375 398 602 | 5 469 273 829 snapshot | 1 282 519 928 | 1 290 190 927 Private repo #1 Manifest Size: limit | none | 1000 ------------|-----------------|--------------- no-sparse | 41 389 010 840 | 41 398 162 091 no-snapshot | 9 737 319 435 | 10 223 773 150 snapshot | 744 215 807 | 747 961 822 Manifest Chain length data limit || none || 1000 || value || average | max || average | max || ------------||---------|---------||---------|---------|| no-sparse || 245 | 8 885 || 81 | 1 000 || no-snapshot || 1 225 | 8 885 || 336 | 1 000 || snapshot || 1 068 | 7 909 || 307 | 1 000 || Full Store Size limit | none | 1000 ------------|----------------|--------------- no-sparse | 49 646 065 126 | 49 655 216 377 no-snapshot | 17 924 862 856 | 18 411 316 571 snapshot | 9 009 024 710 | 9 012 770 725 Private repo #2 We currently have less data available for that repository. * Before is a sparse-revlog repository without this series * After is a sparse-revlog repository with this series + 1000 chain limit Manifest Size: Before: 1 531 485 040 bytes After: 1 091 422 451 bytes Manifest Chain: Before: 2 218 avg; 6 575 Max After: 442 avg; 1 000 Max Full Store Size Before: 15 203 955 615 after: 8 207 180 693
author Boris Feld <boris.feld@octobus.net>
date Fri, 07 Sep 2018 11:18:45 -0400
parents d7cecea0b254
children fa88170c10bb
line wrap: on
line source

# sparse.py - allow sparse checkouts of the working directory
#
# Copyright 2014 Facebook, Inc.
#
# This software may be used and distributed according to the terms of the
# GNU General Public License version 2 or any later version.

"""allow sparse checkouts of the working directory (EXPERIMENTAL)

(This extension is not yet protected by backwards compatibility
guarantees. Any aspect may break in future releases until this
notice is removed.)

This extension allows the working directory to only consist of a
subset of files for the revision. This allows specific files or
directories to be explicitly included or excluded. Many repository
operations have performance proportional to the number of files in
the working directory. So only realizing a subset of files in the
working directory can improve performance.

Sparse Config Files
-------------------

The set of files that are part of a sparse checkout are defined by
a sparse config file. The file defines 3 things: includes (files to
include in the sparse checkout), excludes (files to exclude from the
sparse checkout), and profiles (links to other config files).

The file format is newline delimited. Empty lines and lines beginning
with ``#`` are ignored.

Lines beginning with ``%include `` denote another sparse config file
to include. e.g. ``%include tests.sparse``. The filename is relative
to the repository root.

The special lines ``[include]`` and ``[exclude]`` denote the section
for includes and excludes that follow, respectively. It is illegal to
have ``[include]`` after ``[exclude]``.

Non-special lines resemble file patterns to be added to either includes
or excludes. The syntax of these lines is documented by :hg:`help patterns`.
Patterns are interpreted as ``glob:`` by default and match against the
root of the repository.

Exclusion patterns take precedence over inclusion patterns. So even
if a file is explicitly included, an ``[exclude]`` entry can remove it.

For example, say you have a repository with 3 directories, ``frontend/``,
``backend/``, and ``tools/``. ``frontend/`` and ``backend/`` correspond
to different projects and it is uncommon for someone working on one
to need the files for the other. But ``tools/`` contains files shared
between both projects. Your sparse config files may resemble::

  # frontend.sparse
  frontend/**
  tools/**

  # backend.sparse
  backend/**
  tools/**

Say the backend grows in size. Or there's a directory with thousands
of files you wish to exclude. You can modify the profile to exclude
certain files::

  [include]
  backend/**
  tools/**

  [exclude]
  tools/tests/**
"""

from __future__ import absolute_import

from mercurial.i18n import _
from mercurial import (
    commands,
    dirstate,
    error,
    extensions,
    hg,
    logcmdutil,
    match as matchmod,
    pycompat,
    registrar,
    sparse,
    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'

cmdtable = {}
command = registrar.command(cmdtable)

def extsetup(ui):
    sparse.enabled = True

    _setupclone(ui)
    _setuplog(ui)
    _setupadd(ui)
    _setupdirstate(ui)

def replacefilecache(cls, propname, replacement):
    """Replace a filecache property with a new class. This allows changing the
    cache invalidation condition."""
    origcls = cls
    assert callable(replacement)
    while cls is not object:
        if propname in cls.__dict__:
            orig = cls.__dict__[propname]
            setattr(cls, propname, replacement(orig))
            break
        cls = cls.__bases__[0]

    if cls is object:
        raise AttributeError(_("type '%s' has no property '%s'") % (origcls,
                             propname))

def _setuplog(ui):
    entry = commands.table['^log|history']
    entry[1].append(('', 'sparse', None,
        "limit to changesets affecting the sparse checkout"))

    def _initialrevs(orig, repo, opts):
        revs = orig(repo, opts)
        if opts.get('sparse'):
            sparsematch = sparse.matcher(repo)
            def ctxmatch(rev):
                ctx = repo[rev]
                return any(f for f in ctx.files() if sparsematch(f))
            revs = revs.filter(ctxmatch)
        return revs
    extensions.wrapfunction(logcmdutil, '_initialrevs', _initialrevs)

def _clonesparsecmd(orig, ui, repo, *args, **opts):
    include_pat = opts.get(r'include')
    exclude_pat = opts.get(r'exclude')
    enableprofile_pat = opts.get(r'enable_profile')
    include = exclude = enableprofile = False
    if include_pat:
        pat = include_pat
        include = True
    if exclude_pat:
        pat = exclude_pat
        exclude = True
    if enableprofile_pat:
        pat = enableprofile_pat
        enableprofile = True
    if sum([include, exclude, enableprofile]) > 1:
        raise error.Abort(_("too many flags specified."))
    if include or exclude or enableprofile:
        def clonesparse(orig, self, node, overwrite, *args, **kwargs):
            sparse.updateconfig(self.unfiltered(), pat, {}, include=include,
                                exclude=exclude, enableprofile=enableprofile,
                                usereporootpaths=True)
            return orig(self, node, overwrite, *args, **kwargs)
        extensions.wrapfunction(hg, 'updaterepo', clonesparse)
    return orig(ui, repo, *args, **opts)

def _setupclone(ui):
    entry = commands.table['^clone']
    entry[1].append(('', 'enable-profile', [],
                    'enable a sparse profile'))
    entry[1].append(('', 'include', [],
                    'include sparse pattern'))
    entry[1].append(('', 'exclude', [],
                    'exclude sparse pattern'))
    extensions.wrapcommand(commands.table, 'clone', _clonesparsecmd)

def _setupadd(ui):
    entry = commands.table['^add']
    entry[1].append(('s', 'sparse', None,
                    'also include directories of added files in sparse config'))

    def _add(orig, ui, repo, *pats, **opts):
        if opts.get(r'sparse'):
            dirs = set()
            for pat in pats:
                dirname, basename = util.split(pat)
                dirs.add(dirname)
            sparse.updateconfig(repo, list(dirs), opts, include=True)
        return orig(ui, repo, *pats, **opts)

    extensions.wrapcommand(commands.table, 'add', _add)

def _setupdirstate(ui):
    """Modify the dirstate to prevent stat'ing excluded files,
    and to prevent modifications to files outside the checkout.
    """

    def walk(orig, self, match, subrepos, unknown, ignored, full=True):
        # hack to not exclude explicitly-specified paths so that they can
        # be warned later on e.g. dirstate.add()
        em = matchmod.exact(match._root, match._cwd, match.files())
        sm = matchmod.unionmatcher([self._sparsematcher, em])
        match = matchmod.intersectmatchers(match, sm)
        return orig(self, match, subrepos, unknown, ignored, full)

    extensions.wrapfunction(dirstate.dirstate, 'walk', walk)

    # dirstate.rebuild should not add non-matching files
    def _rebuild(orig, self, parent, allfiles, changedfiles=None):
        matcher = self._sparsematcher
        if not matcher.always():
            allfiles = allfiles.matches(matcher)
            if changedfiles:
                changedfiles = [f for f in changedfiles if matcher(f)]

            if changedfiles is not None:
                # In _rebuild, these files will be deleted from the dirstate
                # when they are not found to be in allfiles
                dirstatefilestoremove = set(f for f in self if not matcher(f))
                changedfiles = dirstatefilestoremove.union(changedfiles)

        return orig(self, parent, allfiles, changedfiles)
    extensions.wrapfunction(dirstate.dirstate, 'rebuild', _rebuild)

    # Prevent adding files that are outside the sparse checkout
    editfuncs = ['normal', 'add', 'normallookup', 'copy', 'remove', 'merge']
    hint = _('include file with `hg debugsparse --include <pattern>` or use ' +
             '`hg add -s <file>` to include file directory while adding')
    for func in editfuncs:
        def _wrapper(orig, self, *args):
            sparsematch = self._sparsematcher
            if not sparsematch.always():
                for f in args:
                    if (f is not None and not sparsematch(f) and
                        f not in self):
                        raise error.Abort(_("cannot add '%s' - it is outside "
                                            "the sparse checkout") % f,
                                          hint=hint)
            return orig(self, *args)
        extensions.wrapfunction(dirstate.dirstate, func, _wrapper)

@command('^debugsparse', [
    ('I', 'include', False, _('include files in the sparse checkout')),
    ('X', 'exclude', False, _('exclude files in the sparse checkout')),
    ('d', 'delete', False, _('delete an include/exclude rule')),
    ('f', 'force', False, _('allow changing rules even with pending changes')),
    ('', 'enable-profile', False, _('enables the specified profile')),
    ('', 'disable-profile', False, _('disables the specified profile')),
    ('', 'import-rules', False, _('imports rules from a file')),
    ('', 'clear-rules', False, _('clears local include/exclude rules')),
    ('', 'refresh', False, _('updates the working after sparseness changes')),
    ('', 'reset', False, _('makes the repo full again')),
    ] + commands.templateopts,
    _('[--OPTION] PATTERN...'))
def debugsparse(ui, repo, *pats, **opts):
    """make the current checkout sparse, or edit the existing checkout

    The sparse command is used to make the current checkout sparse.
    This means files that don't meet the sparse condition will not be
    written to disk, or show up in any working copy operations. It does
    not affect files in history in any way.

    Passing no arguments prints the currently applied sparse rules.

    --include and --exclude are used to add and remove files from the sparse
    checkout. The effects of adding an include or exclude rule are applied
    immediately. If applying the new rule would cause a file with pending
    changes to be added or removed, the command will fail. Pass --force to
    force a rule change even with pending changes (the changes on disk will
    be preserved).

    --delete removes an existing include/exclude rule. The effects are
    immediate.

    --refresh refreshes the files on disk based on the sparse rules. This is
    only necessary if .hg/sparse was changed by hand.

    --enable-profile and --disable-profile accept a path to a .hgsparse file.
    This allows defining sparse checkouts and tracking them inside the
    repository. This is useful for defining commonly used sparse checkouts for
    many people to use. As the profile definition changes over time, the sparse
    checkout will automatically be updated appropriately, depending on which
    changeset is checked out. Changes to .hgsparse are not applied until they
    have been committed.

    --import-rules accepts a path to a file containing rules in the .hgsparse
    format, allowing you to add --include, --exclude and --enable-profile rules
    in bulk. Like the --include, --exclude and --enable-profile switches, the
    changes are applied immediately.

    --clear-rules removes all local include and exclude rules, while leaving
    any enabled profiles in place.

    Returns 0 if editing the sparse checkout succeeds.
    """
    opts = pycompat.byteskwargs(opts)
    include = opts.get('include')
    exclude = opts.get('exclude')
    force = opts.get('force')
    enableprofile = opts.get('enable_profile')
    disableprofile = opts.get('disable_profile')
    importrules = opts.get('import_rules')
    clearrules = opts.get('clear_rules')
    delete = opts.get('delete')
    refresh = opts.get('refresh')
    reset = opts.get('reset')
    count = sum([include, exclude, enableprofile, disableprofile, delete,
                 importrules, refresh, clearrules, reset])
    if count > 1:
        raise error.Abort(_("too many flags specified"))

    if count == 0:
        if repo.vfs.exists('sparse'):
            ui.status(repo.vfs.read("sparse") + "\n")
            temporaryincludes = sparse.readtemporaryincludes(repo)
            if temporaryincludes:
                ui.status(_("Temporarily Included Files (for merge/rebase):\n"))
                ui.status(("\n".join(temporaryincludes) + "\n"))
        else:
            ui.status(_('repo is not sparse\n'))
        return

    if include or exclude or delete or reset or enableprofile or disableprofile:
        sparse.updateconfig(repo, pats, opts, include=include, exclude=exclude,
                            reset=reset, delete=delete,
                            enableprofile=enableprofile,
                            disableprofile=disableprofile, force=force)

    if importrules:
        sparse.importfromfiles(repo, opts, pats, force=force)

    if clearrules:
        sparse.clearrules(repo, force=force)

    if refresh:
        try:
            wlock = repo.wlock()
            fcounts = map(
                len,
                sparse.refreshwdir(repo, repo.status(), sparse.matcher(repo),
                                   force=force))
            sparse.printchanges(ui, opts, added=fcounts[0], dropped=fcounts[1],
                                conflicting=fcounts[2])
        finally:
            wlock.release()