##// END OF EJS Templates
sparse-revlog: set max delta chain length to on thousand...
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

File last commit:

r39425:ddfd8002 default
r39542:b66ea3fc default
Show More
hg.py
1176 lines | 41.0 KiB | text/x-python | PythonLexer
# hg.py - repository classes for mercurial
#
# Copyright 2005-2007 Matt Mackall <mpm@selenic.com>
# Copyright 2006 Vadim Gelfer <vadim.gelfer@gmail.com>
#
# 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 errno
import functools
import hashlib
import os
import shutil
import stat
from .i18n import _
from .node import (
nullid,
)
from . import (
bookmarks,
bundlerepo,
cacheutil,
cmdutil,
destutil,
discovery,
error,
exchange,
extensions,
httppeer,
localrepo,
lock,
logcmdutil,
logexchange,
merge as mergemod,
node,
phases,
scmutil,
sshpeer,
statichttprepo,
ui as uimod,
unionrepo,
url,
util,
verify as verifymod,
vfs as vfsmod,
)
from .utils import (
stringutil,
)
release = lock.release
# shared features
sharedbookmarks = 'bookmarks'
def _local(path):
path = util.expandpath(util.urllocalpath(path))
return (os.path.isfile(path) and bundlerepo or localrepo)
def addbranchrevs(lrepo, other, branches, revs):
peer = other.peer() # a courtesy to callers using a localrepo for other
hashbranch, branches = branches
if not hashbranch and not branches:
x = revs or None
if revs:
y = revs[0]
else:
y = None
return x, y
if revs:
revs = list(revs)
else:
revs = []
if not peer.capable('branchmap'):
if branches:
raise error.Abort(_("remote branch lookup not supported"))
revs.append(hashbranch)
return revs, revs[0]
with peer.commandexecutor() as e:
branchmap = e.callcommand('branchmap', {}).result()
def primary(branch):
if branch == '.':
if not lrepo:
raise error.Abort(_("dirstate branch not accessible"))
branch = lrepo.dirstate.branch()
if branch in branchmap:
revs.extend(node.hex(r) for r in reversed(branchmap[branch]))
return True
else:
return False
for branch in branches:
if not primary(branch):
raise error.RepoLookupError(_("unknown branch '%s'") % branch)
if hashbranch:
if not primary(hashbranch):
revs.append(hashbranch)
return revs, revs[0]
def parseurl(path, branches=None):
'''parse url#branch, returning (url, (branch, branches))'''
u = util.url(path)
branch = None
if u.fragment:
branch = u.fragment
u.fragment = None
return bytes(u), (branch, branches or [])
schemes = {
'bundle': bundlerepo,
'union': unionrepo,
'file': _local,
'http': httppeer,
'https': httppeer,
'ssh': sshpeer,
'static-http': statichttprepo,
}
def _peerlookup(path):
u = util.url(path)
scheme = u.scheme or 'file'
thing = schemes.get(scheme) or schemes['file']
try:
return thing(path)
except TypeError:
# we can't test callable(thing) because 'thing' can be an unloaded
# module that implements __call__
if not util.safehasattr(thing, 'instance'):
raise
return thing
def islocal(repo):
'''return true if repo (or path pointing to repo) is local'''
if isinstance(repo, bytes):
try:
return _peerlookup(repo).islocal(repo)
except AttributeError:
return False
return repo.local()
def openpath(ui, path):
'''open path with open if local, url.open if remote'''
pathurl = util.url(path, parsequery=False, parsefragment=False)
if pathurl.islocal():
return util.posixfile(pathurl.localpath(), 'rb')
else:
return url.open(ui, path)
# a list of (ui, repo) functions called for wire peer initialization
wirepeersetupfuncs = []
def _peerorrepo(ui, path, create=False, presetupfuncs=None,
intents=None):
"""return a repository object for the specified path"""
obj = _peerlookup(path).instance(ui, path, create, intents=intents)
ui = getattr(obj, "ui", ui)
if ui.configbool('devel', 'debug.extensions'):
log = functools.partial(
ui.debug, 'debug.extensions: ', label='debug.extensions')
else:
log = lambda *a, **kw: None
for f in presetupfuncs or []:
f(ui, obj)
log('- executing reposetup hooks\n')
for name, module in extensions.extensions(ui):
log(' - running reposetup for %s\n' % (name,))
hook = getattr(module, 'reposetup', None)
if hook:
hook(ui, obj)
if not obj.local():
for f in wirepeersetupfuncs:
f(ui, obj)
return obj
def repository(ui, path='', create=False, presetupfuncs=None, intents=None):
"""return a repository object for the specified path"""
peer = _peerorrepo(ui, path, create, presetupfuncs=presetupfuncs,
intents=intents)
repo = peer.local()
if not repo:
raise error.Abort(_("repository '%s' is not local") %
(path or peer.url()))
return repo.filtered('visible')
def peer(uiorrepo, opts, path, create=False, intents=None):
'''return a repository peer for the specified path'''
rui = remoteui(uiorrepo, opts)
return _peerorrepo(rui, path, create, intents=intents).peer()
def defaultdest(source):
'''return default destination of clone if none is given
>>> defaultdest(b'foo')
'foo'
>>> defaultdest(b'/foo/bar')
'bar'
>>> defaultdest(b'/')
''
>>> defaultdest(b'')
''
>>> defaultdest(b'http://example.org/')
''
>>> defaultdest(b'http://example.org/foo/')
'foo'
'''
path = util.url(source).path
if not path:
return ''
return os.path.basename(os.path.normpath(path))
def sharedreposource(repo):
"""Returns repository object for source repository of a shared repo.
If repo is not a shared repository, returns None.
"""
if repo.sharedpath == repo.path:
return None
if util.safehasattr(repo, 'srcrepo') and repo.srcrepo:
return repo.srcrepo
# the sharedpath always ends in the .hg; we want the path to the repo
source = repo.vfs.split(repo.sharedpath)[0]
srcurl, branches = parseurl(source)
srcrepo = repository(repo.ui, srcurl)
repo.srcrepo = srcrepo
return srcrepo
def share(ui, source, dest=None, update=True, bookmarks=True, defaultpath=None,
relative=False):
'''create a shared repository'''
if not islocal(source):
raise error.Abort(_('can only share local repositories'))
if not dest:
dest = defaultdest(source)
else:
dest = ui.expandpath(dest)
if isinstance(source, bytes):
origsource = ui.expandpath(source)
source, branches = parseurl(origsource)
srcrepo = repository(ui, source)
rev, checkout = addbranchrevs(srcrepo, srcrepo, branches, None)
else:
srcrepo = source.local()
origsource = source = srcrepo.url()
checkout = None
sharedpath = srcrepo.sharedpath # if our source is already sharing
destwvfs = vfsmod.vfs(dest, realpath=True)
destvfs = vfsmod.vfs(os.path.join(destwvfs.base, '.hg'), realpath=True)
if destvfs.lexists():
raise error.Abort(_('destination already exists'))
if not destwvfs.isdir():
destwvfs.makedirs()
destvfs.makedir()
requirements = ''
try:
requirements = srcrepo.vfs.read('requires')
except IOError as inst:
if inst.errno != errno.ENOENT:
raise
if relative:
try:
sharedpath = os.path.relpath(sharedpath, destvfs.base)
requirements += 'relshared\n'
except (IOError, ValueError) as e:
# ValueError is raised on Windows if the drive letters differ on
# each path
raise error.Abort(_('cannot calculate relative path'),
hint=stringutil.forcebytestr(e))
else:
requirements += 'shared\n'
destvfs.write('requires', requirements)
destvfs.write('sharedpath', sharedpath)
r = repository(ui, destwvfs.base)
postshare(srcrepo, r, bookmarks=bookmarks, defaultpath=defaultpath)
_postshareupdate(r, update, checkout=checkout)
return r
def unshare(ui, repo):
"""convert a shared repository to a normal one
Copy the store data to the repo and remove the sharedpath data.
"""
destlock = lock = None
lock = repo.lock()
try:
# we use locks here because if we race with commit, we
# can end up with extra data in the cloned revlogs that's
# not pointed to by changesets, thus causing verify to
# fail
destlock = copystore(ui, repo, repo.path)
sharefile = repo.vfs.join('sharedpath')
util.rename(sharefile, sharefile + '.old')
repo.requirements.discard('shared')
repo.requirements.discard('relshared')
repo._writerequirements()
finally:
destlock and destlock.release()
lock and lock.release()
# update store, spath, svfs and sjoin of repo
repo.unfiltered().__init__(repo.baseui, repo.root)
# TODO: figure out how to access subrepos that exist, but were previously
# removed from .hgsub
c = repo['.']
subs = c.substate
for s in sorted(subs):
c.sub(s).unshare()
def postshare(sourcerepo, destrepo, bookmarks=True, defaultpath=None):
"""Called after a new shared repo is created.
The new repo only has a requirements file and pointer to the source.
This function configures additional shared data.
Extensions can wrap this function and write additional entries to
destrepo/.hg/shared to indicate additional pieces of data to be shared.
"""
default = defaultpath or sourcerepo.ui.config('paths', 'default')
if default:
template = ('[paths]\n'
'default = %s\n')
destrepo.vfs.write('hgrc', util.tonativeeol(template % default))
with destrepo.wlock():
if bookmarks:
destrepo.vfs.write('shared', sharedbookmarks + '\n')
def _postshareupdate(repo, update, checkout=None):
"""Maybe perform a working directory update after a shared repo is created.
``update`` can be a boolean or a revision to update to.
"""
if not update:
return
repo.ui.status(_("updating working directory\n"))
if update is not True:
checkout = update
for test in (checkout, 'default', 'tip'):
if test is None:
continue
try:
uprev = repo.lookup(test)
break
except error.RepoLookupError:
continue
_update(repo, uprev)
def copystore(ui, srcrepo, destpath):
'''copy files from store of srcrepo in destpath
returns destlock
'''
destlock = None
try:
hardlink = None
topic = _('linking') if hardlink else _('copying')
with ui.makeprogress(topic) as progress:
num = 0
srcpublishing = srcrepo.publishing()
srcvfs = vfsmod.vfs(srcrepo.sharedpath)
dstvfs = vfsmod.vfs(destpath)
for f in srcrepo.store.copylist():
if srcpublishing and f.endswith('phaseroots'):
continue
dstbase = os.path.dirname(f)
if dstbase and not dstvfs.exists(dstbase):
dstvfs.mkdir(dstbase)
if srcvfs.exists(f):
if f.endswith('data'):
# 'dstbase' may be empty (e.g. revlog format 0)
lockfile = os.path.join(dstbase, "lock")
# lock to avoid premature writing to the target
destlock = lock.lock(dstvfs, lockfile)
hardlink, n = util.copyfiles(srcvfs.join(f), dstvfs.join(f),
hardlink, progress)
num += n
if hardlink:
ui.debug("linked %d files\n" % num)
else:
ui.debug("copied %d files\n" % num)
return destlock
except: # re-raises
release(destlock)
raise
def clonewithshare(ui, peeropts, sharepath, source, srcpeer, dest, pull=False,
rev=None, update=True, stream=False):
"""Perform a clone using a shared repo.
The store for the repository will be located at <sharepath>/.hg. The
specified revisions will be cloned or pulled from "source". A shared repo
will be created at "dest" and a working copy will be created if "update" is
True.
"""
revs = None
if rev:
if not srcpeer.capable('lookup'):
raise error.Abort(_("src repository does not support "
"revision lookup and so doesn't "
"support clone by revision"))
# TODO this is batchable.
remoterevs = []
for r in rev:
with srcpeer.commandexecutor() as e:
remoterevs.append(e.callcommand('lookup', {
'key': r,
}).result())
revs = remoterevs
# Obtain a lock before checking for or cloning the pooled repo otherwise
# 2 clients may race creating or populating it.
pooldir = os.path.dirname(sharepath)
# lock class requires the directory to exist.
try:
util.makedir(pooldir, False)
except OSError as e:
if e.errno != errno.EEXIST:
raise
poolvfs = vfsmod.vfs(pooldir)
basename = os.path.basename(sharepath)
with lock.lock(poolvfs, '%s.lock' % basename):
if os.path.exists(sharepath):
ui.status(_('(sharing from existing pooled repository %s)\n') %
basename)
else:
ui.status(_('(sharing from new pooled repository %s)\n') % basename)
# Always use pull mode because hardlinks in share mode don't work
# well. Never update because working copies aren't necessary in
# share mode.
clone(ui, peeropts, source, dest=sharepath, pull=True,
revs=rev, update=False, stream=stream)
# Resolve the value to put in [paths] section for the source.
if islocal(source):
defaultpath = os.path.abspath(util.urllocalpath(source))
else:
defaultpath = source
sharerepo = repository(ui, path=sharepath)
share(ui, sharerepo, dest=dest, update=False, bookmarks=False,
defaultpath=defaultpath)
# We need to perform a pull against the dest repo to fetch bookmarks
# and other non-store data that isn't shared by default. In the case of
# non-existing shared repo, this means we pull from the remote twice. This
# is a bit weird. But at the time it was implemented, there wasn't an easy
# way to pull just non-changegroup data.
destrepo = repository(ui, path=dest)
exchange.pull(destrepo, srcpeer, heads=revs)
_postshareupdate(destrepo, update)
return srcpeer, peer(ui, peeropts, dest)
# Recomputing branch cache might be slow on big repos,
# so just copy it
def _copycache(srcrepo, dstcachedir, fname):
"""copy a cache from srcrepo to destcachedir (if it exists)"""
srcbranchcache = srcrepo.vfs.join('cache/%s' % fname)
dstbranchcache = os.path.join(dstcachedir, fname)
if os.path.exists(srcbranchcache):
if not os.path.exists(dstcachedir):
os.mkdir(dstcachedir)
util.copyfile(srcbranchcache, dstbranchcache)
def clone(ui, peeropts, source, dest=None, pull=False, revs=None,
update=True, stream=False, branch=None, shareopts=None):
"""Make a copy of an existing repository.
Create a copy of an existing repository in a new directory. The
source and destination are URLs, as passed to the repository
function. Returns a pair of repository peers, the source and
newly created destination.
The location of the source is added to the new repository's
.hg/hgrc file, as the default to be used for future pulls and
pushes.
If an exception is raised, the partly cloned/updated destination
repository will be deleted.
Arguments:
source: repository object or URL
dest: URL of destination repository to create (defaults to base
name of source repository)
pull: always pull from source repository, even in local case or if the
server prefers streaming
stream: stream raw data uncompressed from repository (fast over
LAN, slow over WAN)
revs: revision to clone up to (implies pull=True)
update: update working directory after clone completes, if
destination is local repository (True means update to default rev,
anything else is treated as a revision)
branch: branches to clone
shareopts: dict of options to control auto sharing behavior. The "pool" key
activates auto sharing mode and defines the directory for stores. The
"mode" key determines how to construct the directory name of the shared
repository. "identity" means the name is derived from the node of the first
changeset in the repository. "remote" means the name is derived from the
remote's path/URL. Defaults to "identity."
"""
if isinstance(source, bytes):
origsource = ui.expandpath(source)
source, branches = parseurl(origsource, branch)
srcpeer = peer(ui, peeropts, source)
else:
srcpeer = source.peer() # in case we were called with a localrepo
branches = (None, branch or [])
origsource = source = srcpeer.url()
revs, checkout = addbranchrevs(srcpeer, srcpeer, branches, revs)
if dest is None:
dest = defaultdest(source)
if dest:
ui.status(_("destination directory: %s\n") % dest)
else:
dest = ui.expandpath(dest)
dest = util.urllocalpath(dest)
source = util.urllocalpath(source)
if not dest:
raise error.Abort(_("empty destination path is not valid"))
destvfs = vfsmod.vfs(dest, expandpath=True)
if destvfs.lexists():
if not destvfs.isdir():
raise error.Abort(_("destination '%s' already exists") % dest)
elif destvfs.listdir():
raise error.Abort(_("destination '%s' is not empty") % dest)
shareopts = shareopts or {}
sharepool = shareopts.get('pool')
sharenamemode = shareopts.get('mode')
if sharepool and islocal(dest):
sharepath = None
if sharenamemode == 'identity':
# Resolve the name from the initial changeset in the remote
# repository. This returns nullid when the remote is empty. It
# raises RepoLookupError if revision 0 is filtered or otherwise
# not available. If we fail to resolve, sharing is not enabled.
try:
with srcpeer.commandexecutor() as e:
rootnode = e.callcommand('lookup', {
'key': '0',
}).result()
if rootnode != node.nullid:
sharepath = os.path.join(sharepool, node.hex(rootnode))
else:
ui.status(_('(not using pooled storage: '
'remote appears to be empty)\n'))
except error.RepoLookupError:
ui.status(_('(not using pooled storage: '
'unable to resolve identity of remote)\n'))
elif sharenamemode == 'remote':
sharepath = os.path.join(
sharepool, node.hex(hashlib.sha1(source).digest()))
else:
raise error.Abort(_('unknown share naming mode: %s') %
sharenamemode)
if sharepath:
return clonewithshare(ui, peeropts, sharepath, source, srcpeer,
dest, pull=pull, rev=revs, update=update,
stream=stream)
srclock = destlock = cleandir = None
srcrepo = srcpeer.local()
try:
abspath = origsource
if islocal(origsource):
abspath = os.path.abspath(util.urllocalpath(origsource))
if islocal(dest):
cleandir = dest
copy = False
if (srcrepo and srcrepo.cancopy() and islocal(dest)
and not phases.hassecret(srcrepo)):
copy = not pull and not revs
if copy:
try:
# we use a lock here because if we race with commit, we
# can end up with extra data in the cloned revlogs that's
# not pointed to by changesets, thus causing verify to
# fail
srclock = srcrepo.lock(wait=False)
except error.LockError:
copy = False
if copy:
srcrepo.hook('preoutgoing', throw=True, source='clone')
hgdir = os.path.realpath(os.path.join(dest, ".hg"))
if not os.path.exists(dest):
util.makedirs(dest)
else:
# only clean up directories we create ourselves
cleandir = hgdir
try:
destpath = hgdir
util.makedir(destpath, notindexed=True)
except OSError as inst:
if inst.errno == errno.EEXIST:
cleandir = None
raise error.Abort(_("destination '%s' already exists")
% dest)
raise
destlock = copystore(ui, srcrepo, destpath)
# copy bookmarks over
srcbookmarks = srcrepo.vfs.join('bookmarks')
dstbookmarks = os.path.join(destpath, 'bookmarks')
if os.path.exists(srcbookmarks):
util.copyfile(srcbookmarks, dstbookmarks)
dstcachedir = os.path.join(destpath, 'cache')
for cache in cacheutil.cachetocopy(srcrepo):
_copycache(srcrepo, dstcachedir, cache)
# we need to re-init the repo after manually copying the data
# into it
destpeer = peer(srcrepo, peeropts, dest)
srcrepo.hook('outgoing', source='clone',
node=node.hex(node.nullid))
else:
try:
destpeer = peer(srcrepo or ui, peeropts, dest, create=True)
# only pass ui when no srcrepo
except OSError as inst:
if inst.errno == errno.EEXIST:
cleandir = None
raise error.Abort(_("destination '%s' already exists")
% dest)
raise
if revs:
if not srcpeer.capable('lookup'):
raise error.Abort(_("src repository does not support "
"revision lookup and so doesn't "
"support clone by revision"))
# TODO this is batchable.
remoterevs = []
for rev in revs:
with srcpeer.commandexecutor() as e:
remoterevs.append(e.callcommand('lookup', {
'key': rev,
}).result())
revs = remoterevs
checkout = revs[0]
else:
revs = None
local = destpeer.local()
if local:
u = util.url(abspath)
defaulturl = bytes(u)
local.ui.setconfig('paths', 'default', defaulturl, 'clone')
if not stream:
if pull:
stream = False
else:
stream = None
# internal config: ui.quietbookmarkmove
overrides = {('ui', 'quietbookmarkmove'): True}
with local.ui.configoverride(overrides, 'clone'):
exchange.pull(local, srcpeer, revs,
streamclonerequested=stream)
elif srcrepo:
exchange.push(srcrepo, destpeer, revs=revs,
bookmarks=srcrepo._bookmarks.keys())
else:
raise error.Abort(_("clone from remote to remote not supported")
)
cleandir = None
destrepo = destpeer.local()
if destrepo:
template = uimod.samplehgrcs['cloned']
u = util.url(abspath)
u.passwd = None
defaulturl = bytes(u)
destrepo.vfs.write('hgrc', util.tonativeeol(template % defaulturl))
destrepo.ui.setconfig('paths', 'default', defaulturl, 'clone')
if ui.configbool('experimental', 'remotenames'):
logexchange.pullremotenames(destrepo, srcpeer)
if update:
if update is not True:
with srcpeer.commandexecutor() as e:
checkout = e.callcommand('lookup', {
'key': update,
}).result()
uprev = None
status = None
if checkout is not None:
# Some extensions (at least hg-git and hg-subversion) have
# a peer.lookup() implementation that returns a name instead
# of a nodeid. We work around it here until we've figured
# out a better solution.
if len(checkout) == 20 and checkout in destrepo:
uprev = checkout
elif scmutil.isrevsymbol(destrepo, checkout):
uprev = scmutil.revsymbol(destrepo, checkout).node()
else:
if update is not True:
try:
uprev = destrepo.lookup(update)
except error.RepoLookupError:
pass
if uprev is None:
try:
uprev = destrepo._bookmarks['@']
update = '@'
bn = destrepo[uprev].branch()
if bn == 'default':
status = _("updating to bookmark @\n")
else:
status = (_("updating to bookmark @ on branch %s\n")
% bn)
except KeyError:
try:
uprev = destrepo.branchtip('default')
except error.RepoLookupError:
uprev = destrepo.lookup('tip')
if not status:
bn = destrepo[uprev].branch()
status = _("updating to branch %s\n") % bn
destrepo.ui.status(status)
_update(destrepo, uprev)
if update in destrepo._bookmarks:
bookmarks.activate(destrepo, update)
finally:
release(srclock, destlock)
if cleandir is not None:
shutil.rmtree(cleandir, True)
if srcpeer is not None:
srcpeer.close()
return srcpeer, destpeer
def _showstats(repo, stats, quietempty=False):
if quietempty and stats.isempty():
return
repo.ui.status(_("%d files updated, %d files merged, "
"%d files removed, %d files unresolved\n") % (
stats.updatedcount, stats.mergedcount,
stats.removedcount, stats.unresolvedcount))
def updaterepo(repo, node, overwrite, updatecheck=None):
"""Update the working directory to node.
When overwrite is set, changes are clobbered, merged else
returns stats (see pydoc mercurial.merge.applyupdates)"""
return mergemod.update(repo, node, False, overwrite,
labels=['working copy', 'destination'],
updatecheck=updatecheck)
def update(repo, node, quietempty=False, updatecheck=None):
"""update the working directory to node"""
stats = updaterepo(repo, node, False, updatecheck=updatecheck)
_showstats(repo, stats, quietempty)
if stats.unresolvedcount:
repo.ui.status(_("use 'hg resolve' to retry unresolved file merges\n"))
return stats.unresolvedcount > 0
# naming conflict in clone()
_update = update
def clean(repo, node, show_stats=True, quietempty=False):
"""forcibly switch the working directory to node, clobbering changes"""
stats = updaterepo(repo, node, True)
repo.vfs.unlinkpath('graftstate', ignoremissing=True)
if show_stats:
_showstats(repo, stats, quietempty)
return stats.unresolvedcount > 0
# naming conflict in updatetotally()
_clean = clean
def updatetotally(ui, repo, checkout, brev, clean=False, updatecheck=None):
"""Update the working directory with extra care for non-file components
This takes care of non-file components below:
:bookmark: might be advanced or (in)activated
This takes arguments below:
:checkout: to which revision the working directory is updated
:brev: a name, which might be a bookmark to be activated after updating
:clean: whether changes in the working directory can be discarded
:updatecheck: how to deal with a dirty working directory
Valid values for updatecheck are (None => linear):
* abort: abort if the working directory is dirty
* none: don't check (merge working directory changes into destination)
* linear: check that update is linear before merging working directory
changes into destination
* noconflict: check that the update does not result in file merges
This returns whether conflict is detected at updating or not.
"""
if updatecheck is None:
updatecheck = ui.config('commands', 'update.check')
if updatecheck not in ('abort', 'none', 'linear', 'noconflict'):
# If not configured, or invalid value configured
updatecheck = 'linear'
with repo.wlock():
movemarkfrom = None
warndest = False
if checkout is None:
updata = destutil.destupdate(repo, clean=clean)
checkout, movemarkfrom, brev = updata
warndest = True
if clean:
ret = _clean(repo, checkout)
else:
if updatecheck == 'abort':
cmdutil.bailifchanged(repo, merge=False)
updatecheck = 'none'
ret = _update(repo, checkout, updatecheck=updatecheck)
if not ret and movemarkfrom:
if movemarkfrom == repo['.'].node():
pass # no-op update
elif bookmarks.update(repo, [movemarkfrom], repo['.'].node()):
b = ui.label(repo._activebookmark, 'bookmarks.active')
ui.status(_("updating bookmark %s\n") % b)
else:
# this can happen with a non-linear update
b = ui.label(repo._activebookmark, 'bookmarks')
ui.status(_("(leaving bookmark %s)\n") % b)
bookmarks.deactivate(repo)
elif brev in repo._bookmarks:
if brev != repo._activebookmark:
b = ui.label(brev, 'bookmarks.active')
ui.status(_("(activating bookmark %s)\n") % b)
bookmarks.activate(repo, brev)
elif brev:
if repo._activebookmark:
b = ui.label(repo._activebookmark, 'bookmarks')
ui.status(_("(leaving bookmark %s)\n") % b)
bookmarks.deactivate(repo)
if warndest:
destutil.statusotherdests(ui, repo)
return ret
def merge(repo, node, force=None, remind=True, mergeforce=False, labels=None,
abort=False):
"""Branch merge with node, resolving changes. Return true if any
unresolved conflicts."""
if not abort:
stats = mergemod.update(repo, node, True, force, mergeforce=mergeforce,
labels=labels)
else:
ms = mergemod.mergestate.read(repo)
if ms.active():
# there were conflicts
node = ms.localctx.hex()
else:
# there were no conficts, mergestate was not stored
node = repo['.'].hex()
repo.ui.status(_("aborting the merge, updating back to"
" %s\n") % node[:12])
stats = mergemod.update(repo, node, branchmerge=False, force=True,
labels=labels)
_showstats(repo, stats)
if stats.unresolvedcount:
repo.ui.status(_("use 'hg resolve' to retry unresolved file merges "
"or 'hg merge --abort' to abandon\n"))
elif remind and not abort:
repo.ui.status(_("(branch merge, don't forget to commit)\n"))
return stats.unresolvedcount > 0
def _incoming(displaychlist, subreporecurse, ui, repo, source,
opts, buffered=False):
"""
Helper for incoming / gincoming.
displaychlist gets called with
(remoterepo, incomingchangesetlist, displayer) parameters,
and is supposed to contain only code that can't be unified.
"""
source, branches = parseurl(ui.expandpath(source), opts.get('branch'))
other = peer(repo, opts, source)
ui.status(_('comparing with %s\n') % util.hidepassword(source))
revs, checkout = addbranchrevs(repo, other, branches, opts.get('rev'))
if revs:
revs = [other.lookup(rev) for rev in revs]
other, chlist, cleanupfn = bundlerepo.getremotechanges(ui, repo, other,
revs, opts["bundle"], opts["force"])
try:
if not chlist:
ui.status(_("no changes found\n"))
return subreporecurse()
ui.pager('incoming')
displayer = logcmdutil.changesetdisplayer(ui, other, opts,
buffered=buffered)
displaychlist(other, chlist, displayer)
displayer.close()
finally:
cleanupfn()
subreporecurse()
return 0 # exit code is zero since we found incoming changes
def incoming(ui, repo, source, opts):
def subreporecurse():
ret = 1
if opts.get('subrepos'):
ctx = repo[None]
for subpath in sorted(ctx.substate):
sub = ctx.sub(subpath)
ret = min(ret, sub.incoming(ui, source, opts))
return ret
def display(other, chlist, displayer):
limit = logcmdutil.getlimit(opts)
if opts.get('newest_first'):
chlist.reverse()
count = 0
for n in chlist:
if limit is not None and count >= limit:
break
parents = [p for p in other.changelog.parents(n) if p != nullid]
if opts.get('no_merges') and len(parents) == 2:
continue
count += 1
displayer.show(other[n])
return _incoming(display, subreporecurse, ui, repo, source, opts)
def _outgoing(ui, repo, dest, opts):
path = ui.paths.getpath(dest, default=('default-push', 'default'))
if not path:
raise error.Abort(_('default repository not configured!'),
hint=_("see 'hg help config.paths'"))
dest = path.pushloc or path.loc
branches = path.branch, opts.get('branch') or []
ui.status(_('comparing with %s\n') % util.hidepassword(dest))
revs, checkout = addbranchrevs(repo, repo, branches, opts.get('rev'))
if revs:
revs = [repo[rev].node() for rev in scmutil.revrange(repo, revs)]
other = peer(repo, opts, dest)
outgoing = discovery.findcommonoutgoing(repo, other, revs,
force=opts.get('force'))
o = outgoing.missing
if not o:
scmutil.nochangesfound(repo.ui, repo, outgoing.excluded)
return o, other
def outgoing(ui, repo, dest, opts):
def recurse():
ret = 1
if opts.get('subrepos'):
ctx = repo[None]
for subpath in sorted(ctx.substate):
sub = ctx.sub(subpath)
ret = min(ret, sub.outgoing(ui, dest, opts))
return ret
limit = logcmdutil.getlimit(opts)
o, other = _outgoing(ui, repo, dest, opts)
if not o:
cmdutil.outgoinghooks(ui, repo, other, opts, o)
return recurse()
if opts.get('newest_first'):
o.reverse()
ui.pager('outgoing')
displayer = logcmdutil.changesetdisplayer(ui, repo, opts)
count = 0
for n in o:
if limit is not None and count >= limit:
break
parents = [p for p in repo.changelog.parents(n) if p != nullid]
if opts.get('no_merges') and len(parents) == 2:
continue
count += 1
displayer.show(repo[n])
displayer.close()
cmdutil.outgoinghooks(ui, repo, other, opts, o)
recurse()
return 0 # exit code is zero since we found outgoing changes
def verify(repo):
"""verify the consistency of a repository"""
ret = verifymod.verify(repo)
# Broken subrepo references in hidden csets don't seem worth worrying about,
# since they can't be pushed/pulled, and --hidden can be used if they are a
# concern.
# pathto() is needed for -R case
revs = repo.revs("filelog(%s)",
util.pathto(repo.root, repo.getcwd(), '.hgsubstate'))
if revs:
repo.ui.status(_('checking subrepo links\n'))
for rev in revs:
ctx = repo[rev]
try:
for subpath in ctx.substate:
try:
ret = (ctx.sub(subpath, allowcreate=False).verify()
or ret)
except error.RepoError as e:
repo.ui.warn(('%d: %s\n') % (rev, e))
except Exception:
repo.ui.warn(_('.hgsubstate is corrupt in revision %s\n') %
node.short(ctx.node()))
return ret
def remoteui(src, opts):
'build a remote ui from ui or repo and opts'
if util.safehasattr(src, 'baseui'): # looks like a repository
dst = src.baseui.copy() # drop repo-specific config
src = src.ui # copy target options from repo
else: # assume it's a global ui object
dst = src.copy() # keep all global options
# copy ssh-specific options
for o in 'ssh', 'remotecmd':
v = opts.get(o) or src.config('ui', o)
if v:
dst.setconfig("ui", o, v, 'copied')
# copy bundle-specific options
r = src.config('bundle', 'mainreporoot')
if r:
dst.setconfig('bundle', 'mainreporoot', r, 'copied')
# copy selected local settings to the remote ui
for sect in ('auth', 'hostfingerprints', 'hostsecurity', 'http_proxy'):
for key, val in src.configitems(sect):
dst.setconfig(sect, key, val, 'copied')
v = src.config('web', 'cacerts')
if v:
dst.setconfig('web', 'cacerts', util.expandpath(v), 'copied')
return dst
# Files of interest
# Used to check if the repository has changed looking at mtime and size of
# these files.
foi = [('spath', '00changelog.i'),
('spath', 'phaseroots'), # ! phase can change content at the same size
('spath', 'obsstore'),
('path', 'bookmarks'), # ! bookmark can change content at the same size
]
class cachedlocalrepo(object):
"""Holds a localrepository that can be cached and reused."""
def __init__(self, repo):
"""Create a new cached repo from an existing repo.
We assume the passed in repo was recently created. If the
repo has changed between when it was created and when it was
turned into a cache, it may not refresh properly.
"""
assert isinstance(repo, localrepo.localrepository)
self._repo = repo
self._state, self.mtime = self._repostate()
self._filtername = repo.filtername
def fetch(self):
"""Refresh (if necessary) and return a repository.
If the cached instance is out of date, it will be recreated
automatically and returned.
Returns a tuple of the repo and a boolean indicating whether a new
repo instance was created.
"""
# We compare the mtimes and sizes of some well-known files to
# determine if the repo changed. This is not precise, as mtimes
# are susceptible to clock skew and imprecise filesystems and
# file content can change while maintaining the same size.
state, mtime = self._repostate()
if state == self._state:
return self._repo, False
repo = repository(self._repo.baseui, self._repo.url())
if self._filtername:
self._repo = repo.filtered(self._filtername)
else:
self._repo = repo.unfiltered()
self._state = state
self.mtime = mtime
return self._repo, True
def _repostate(self):
state = []
maxmtime = -1
for attr, fname in foi:
prefix = getattr(self._repo, attr)
p = os.path.join(prefix, fname)
try:
st = os.stat(p)
except OSError:
st = os.stat(prefix)
state.append((st[stat.ST_MTIME], st.st_size))
maxmtime = max(maxmtime, st[stat.ST_MTIME])
return tuple(state), maxmtime
def copy(self):
"""Obtain a copy of this class instance.
A new localrepository instance is obtained. The new instance should be
completely independent of the original.
"""
repo = repository(self._repo.baseui, self._repo.origroot)
if self._filtername:
repo = repo.filtered(self._filtername)
else:
repo = repo.unfiltered()
c = cachedlocalrepo(repo)
c._state = self._state
c.mtime = self.mtime
return c