urls: remove deprecated APIs

Differential Revision: https://phab.mercurial-scm.org/D11950

mercurial/hg.py

@@ -1,1608 +1,1601 b''
 # hg.py - repository classes for mercurial
 #
 # Copyright 2005-2007 Olivia Mackall <[e-mail]>
 # Copyright 2006 Vadim Gelfer <[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
 
 import errno
 import os
 import shutil
 import stat
 import weakref
 
 from .i18n import _
 from .node import (
     hex,
     sha1nodeconstants,
     short,
 )
 from .pycompat import getattr
 
 from . import (
     bookmarks,
     bundlerepo,
     cmdutil,
     destutil,
     discovery,
     error,
     exchange,
     extensions,
     graphmod,
     httppeer,
     localrepo,
     lock,
     logcmdutil,
     logexchange,
     merge as mergemod,
     mergestate as mergestatemod,
     narrowspec,
     phases,
     requirements,
     scmutil,
     sshpeer,
     statichttprepo,
     ui as uimod,
     unionrepo,
     url,
     util,
     verify as verifymod,
     vfs as vfsmod,
 )
 from .interfaces import repository as repositorymod
 from .utils import (
     hashutil,
     stringutil,
     urlutil,
 )
 
 
 release = lock.release
 
 # shared features
 sharedbookmarks = b'bookmarks'
 
 
 def _local(path):
     path = util.expandpath(urlutil.urllocalpath(path))
 
     try:
         # we use os.stat() directly here instead of os.path.isfile()
         # because the latter started returning `False` on invalid path
         # exceptions starting in 3.8 and we care about handling
         # invalid paths specially here.
         st = os.stat(path)
         isfile = stat.S_ISREG(st.st_mode)
     # Python 2 raises TypeError, Python 3 ValueError.
     except (TypeError, ValueError) as e:
         raise error.Abort(
             _(b'invalid path %s: %s') % (path, stringutil.forcebytestr(e))
         )
     except OSError:
         isfile = False
 
     return isfile 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(b'branchmap'):
         if branches:
             raise error.Abort(_(b"remote branch lookup not supported"))
         revs.append(hashbranch)
         return revs, revs[0]
 
     with peer.commandexecutor() as e:
         branchmap = e.callcommand(b'branchmap', {}).result()
 
     def primary(branch):
         if branch == b'.':
             if not lrepo:
                 raise error.Abort(_(b"dirstate branch not accessible"))
             branch = lrepo.dirstate.branch()
         if branch in branchmap:
             revs.extend(hex(r) for r in reversed(branchmap[branch]))
             return True
         else:
             return False
 
     for branch in branches:
         if not primary(branch):
             raise error.RepoLookupError(_(b"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))'''
-    msg = b'parseurl(...) moved to mercurial.utils.urlutil'
-    util.nouideprecwarn(msg, b'6.0', stacklevel=2)
-    return urlutil.parseurl(path, branches=branches)
-
-
 schemes = {
     b'bundle': bundlerepo,
     b'union': unionrepo,
     b'file': _local,
     b'http': httppeer,
     b'https': httppeer,
     b'ssh': sshpeer,
     b'static-http': statichttprepo,
 }
 
 
 def _peerlookup(path):
     u = urlutil.url(path)
     scheme = u.scheme or b'file'
     thing = schemes.get(scheme) or schemes[b'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, b'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, sendaccept=True):
     '''open path with open if local, url.open if remote'''
     pathurl = urlutil.url(path, parsequery=False, parsefragment=False)
     if pathurl.islocal():
         return util.posixfile(pathurl.localpath(), b'rb')
     else:
         return url.open(ui, path, sendaccept=sendaccept)
 
 
 # a list of (ui, repo) functions called for wire peer initialization
 wirepeersetupfuncs = []
 
 
 def _peerorrepo(
     ui, path, create=False, presetupfuncs=None, intents=None, createopts=None
 ):
     """return a repository object for the specified path"""
     obj = _peerlookup(path).instance(
         ui, path, create, intents=intents, createopts=createopts
     )
     ui = getattr(obj, "ui", ui)
     for f in presetupfuncs or []:
         f(ui, obj)
     ui.log(b'extension', b'- executing reposetup hooks\n')
     with util.timedcm('all reposetup') as allreposetupstats:
         for name, module in extensions.extensions(ui):
             ui.log(b'extension', b'  - running reposetup for %s\n', name)
             hook = getattr(module, 'reposetup', None)
             if hook:
                 with util.timedcm('reposetup %r', name) as stats:
                     hook(ui, obj)
                 ui.log(
                     b'extension', b'  > reposetup for %s took %s\n', name, stats
                 )
     ui.log(b'extension', b'> all reposetup took %s\n', allreposetupstats)
     if not obj.local():
         for f in wirepeersetupfuncs:
             f(ui, obj)
     return obj
 
 
 def repository(
     ui,
     path=b'',
     create=False,
     presetupfuncs=None,
     intents=None,
     createopts=None,
 ):
     """return a repository object for the specified path"""
     peer = _peerorrepo(
         ui,
         path,
         create,
         presetupfuncs=presetupfuncs,
         intents=intents,
         createopts=createopts,
     )
     repo = peer.local()
     if not repo:
         raise error.Abort(
             _(b"repository '%s' is not local") % (path or peer.url())
         )
     return repo.filtered(b'visible')
 
 
 def peer(uiorrepo, opts, path, create=False, intents=None, createopts=None):
     '''return a repository peer for the specified path'''
     rui = remoteui(uiorrepo, opts)
     return _peerorrepo(
         rui, path, create, intents=intents, createopts=createopts
     ).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 = urlutil.url(source).path
     if not path:
         return b''
     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, b'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 = urlutil.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(_(b'can only share local repositories'))
 
     if not dest:
         dest = defaultdest(source)
     else:
         dest = urlutil.get_clone_path(ui, dest)[1]
 
     if isinstance(source, bytes):
         origsource, source, branches = urlutil.get_clone_path(ui, source)
         srcrepo = repository(ui, source)
         rev, checkout = addbranchrevs(srcrepo, srcrepo, branches, None)
     else:
         srcrepo = source.local()
         checkout = None
 
     shareditems = set()
     if bookmarks:
         shareditems.add(sharedbookmarks)
 
     r = repository(
         ui,
         dest,
         create=True,
         createopts={
             b'sharedrepo': srcrepo,
             b'sharedrelative': relative,
             b'shareditems': shareditems,
         },
     )
 
     postshare(srcrepo, r, defaultpath=defaultpath)
     r = repository(ui, dest)
     _postshareupdate(r, update, checkout=checkout)
     return r
 
 
 def _prependsourcehgrc(repo):
     """copies the source repo config and prepend it in current repo .hg/hgrc
     on unshare. This is only done if the share was perfomed using share safe
     method where we share config of source in shares"""
     srcvfs = vfsmod.vfs(repo.sharedpath)
     dstvfs = vfsmod.vfs(repo.path)
 
     if not srcvfs.exists(b'hgrc'):
         return
 
     currentconfig = b''
     if dstvfs.exists(b'hgrc'):
         currentconfig = dstvfs.read(b'hgrc')
 
     with dstvfs(b'hgrc', b'wb') as fp:
         sourceconfig = srcvfs.read(b'hgrc')
         fp.write(b"# Config copied from shared source\n")
         fp.write(sourceconfig)
         fp.write(b'\n')
         fp.write(currentconfig)
 
 
 def unshare(ui, repo):
     """convert a shared repository to a normal one
 
     Copy the store data to the repo and remove the sharedpath data.
 
     Returns a new repository object representing the unshared repository.
 
     The passed repository object is not usable after this function is
     called.
     """
 
     with repo.lock():
         # 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)
         with destlock or util.nullcontextmanager():
             if requirements.SHARESAFE_REQUIREMENT in repo.requirements:
                 # we were sharing .hg/hgrc of the share source with the current
                 # repo. We need to copy that while unsharing otherwise it can
                 # disable hooks and other checks
                 _prependsourcehgrc(repo)
 
             sharefile = repo.vfs.join(b'sharedpath')
             util.rename(sharefile, sharefile + b'.old')
 
             repo.requirements.discard(requirements.SHARED_REQUIREMENT)
             repo.requirements.discard(requirements.RELATIVE_SHARED_REQUIREMENT)
             scmutil.writereporequirements(repo)
 
     # Removing share changes some fundamental properties of the repo instance.
     # So we instantiate a new repo object and operate on it rather than
     # try to keep the existing repo usable.
     newrepo = repository(repo.baseui, repo.root, create=False)
 
     # TODO: figure out how to access subrepos that exist, but were previously
     #       removed from .hgsub
     c = newrepo[b'.']
     subs = c.substate
     for s in sorted(subs):
         c.sub(s).unshare()
 
     localrepo.poisonrepository(repo)
 
     return newrepo
 
 
 def postshare(sourcerepo, destrepo, 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(b'paths', b'default')
     if default:
         template = b'[paths]\ndefault = %s\n'
         destrepo.vfs.write(b'hgrc', util.tonativeeol(template % default))
     if requirements.NARROW_REQUIREMENT in sourcerepo.requirements:
         with destrepo.wlock():
             narrowspec.copytoworkingcopy(destrepo)
 
 
 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(_(b"updating working directory\n"))
     if update is not True:
         checkout = update
     for test in (checkout, b'default', b'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 = _(b'linking') if hardlink else _(b'copying')
         with ui.makeprogress(topic, unit=_(b'files')) 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(b'phaseroots'):
                     continue
                 dstbase = os.path.dirname(f)
                 if dstbase and not dstvfs.exists(dstbase):
                     dstvfs.mkdir(dstbase)
                 if srcvfs.exists(f):
                     if f.endswith(b'data'):
                         # 'dstbase' may be empty (e.g. revlog format 0)
                         lockfile = os.path.join(dstbase, b"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(b"linked %d files\n" % num)
             else:
                 ui.debug(b"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(b'lookup'):
             raise error.Abort(
                 _(
                     b"src repository does not support "
                     b"revision lookup and so doesn't "
                     b"support clone by revision"
                 )
             )
 
         # TODO this is batchable.
         remoterevs = []
         for r in rev:
             with srcpeer.commandexecutor() as e:
                 remoterevs.append(
                     e.callcommand(
                         b'lookup',
                         {
                             b'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, b'%s.lock' % basename):
         if os.path.exists(sharepath):
             ui.status(
                 _(b'(sharing from existing pooled repository %s)\n') % basename
             )
         else:
             ui.status(
                 _(b'(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 = util.abspath(urlutil.urllocalpath(source))
     else:
         defaultpath = source
 
     sharerepo = repository(ui, path=sharepath)
     destrepo = 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.
     exchange.pull(destrepo, srcpeer, heads=revs)
 
     _postshareupdate(destrepo, update)
 
     return srcpeer, peer(ui, peeropts, dest)
 
 
 # Recomputing caches is often slow on big repos, so copy them.
 def _copycache(srcrepo, dstcachedir, fname):
     """copy a cache from srcrepo to destcachedir (if it exists)"""
     srcfname = srcrepo.cachevfs.join(fname)
     dstfname = os.path.join(dstcachedir, fname)
     if os.path.exists(srcfname):
         if not os.path.exists(dstcachedir):
             os.mkdir(dstcachedir)
         util.copyfile(srcfname, dstfname)
 
 
 def clone(
     ui,
     peeropts,
     source,
     dest=None,
     pull=False,
     revs=None,
     update=True,
     stream=False,
     branch=None,
     shareopts=None,
     storeincludepats=None,
     storeexcludepats=None,
     depth=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."
 
     storeincludepats and storeexcludepats: sets of file patterns to include and
     exclude in the repository copy, respectively. If not defined, all files
     will be included (a "full" clone). Otherwise a "narrow" clone containing
     only the requested files will be performed. If ``storeincludepats`` is not
     defined but ``storeexcludepats`` is, ``storeincludepats`` is assumed to be
     ``path:.``. If both are empty sets, no files will be cloned.
     """
 
     if isinstance(source, bytes):
         src = urlutil.get_clone_path(ui, source, branch)
         origsource, source, branches = src
         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()
     srclock = destlock = destwlock = cleandir = None
     destpeer = None
     try:
         revs, checkout = addbranchrevs(srcpeer, srcpeer, branches, revs)
 
         if dest is None:
             dest = defaultdest(source)
             if dest:
                 ui.status(_(b"destination directory: %s\n") % dest)
         else:
             dest = urlutil.get_clone_path(ui, dest)[0]
 
         dest = urlutil.urllocalpath(dest)
         source = urlutil.urllocalpath(source)
 
         if not dest:
             raise error.InputError(_(b"empty destination path is not valid"))
 
         destvfs = vfsmod.vfs(dest, expandpath=True)
         if destvfs.lexists():
             if not destvfs.isdir():
                 raise error.InputError(
                     _(b"destination '%s' already exists") % dest
                 )
             elif destvfs.listdir():
                 raise error.InputError(
                     _(b"destination '%s' is not empty") % dest
                 )
 
         createopts = {}
         narrow = False
 
         if storeincludepats is not None:
             narrowspec.validatepatterns(storeincludepats)
             narrow = True
 
         if storeexcludepats is not None:
             narrowspec.validatepatterns(storeexcludepats)
             narrow = True
 
         if narrow:
             # Include everything by default if only exclusion patterns defined.
             if storeexcludepats and not storeincludepats:
                 storeincludepats = {b'path:.'}
 
             createopts[b'narrowfiles'] = True
 
         if depth:
             createopts[b'shallowfilestore'] = True
 
         if srcpeer.capable(b'lfs-serve'):
             # Repository creation honors the config if it disabled the extension, so
             # we can't just announce that lfs will be enabled.  This check avoids
             # saying that lfs will be enabled, and then saying it's an unknown
             # feature.  The lfs creation option is set in either case so that a
             # requirement is added.  If the extension is explicitly disabled but the
             # requirement is set, the clone aborts early, before transferring any
             # data.
             createopts[b'lfs'] = True
 
             if extensions.disabled_help(b'lfs'):
                 ui.status(
                     _(
                         b'(remote is using large file support (lfs), but it is '
                         b'explicitly disabled in the local configuration)\n'
                     )
                 )
             else:
                 ui.status(
                     _(
                         b'(remote is using large file support (lfs); lfs will '
                         b'be enabled for this repository)\n'
                     )
                 )
 
         shareopts = shareopts or {}
         sharepool = shareopts.get(b'pool')
         sharenamemode = shareopts.get(b'mode')
         if sharepool and islocal(dest):
             sharepath = None
             if sharenamemode == b'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(
                             b'lookup',
                             {
                                 b'key': b'0',
                             },
                         ).result()
 
                     if rootnode != sha1nodeconstants.nullid:
                         sharepath = os.path.join(sharepool, hex(rootnode))
                     else:
                         ui.status(
                             _(
                                 b'(not using pooled storage: '
                                 b'remote appears to be empty)\n'
                             )
                         )
                 except error.RepoLookupError:
                     ui.status(
                         _(
                             b'(not using pooled storage: '
                             b'unable to resolve identity of remote)\n'
                         )
                     )
             elif sharenamemode == b'remote':
                 sharepath = os.path.join(
                     sharepool, hex(hashutil.sha1(source).digest())
                 )
             else:
                 raise error.Abort(
                     _(b'unknown share naming mode: %s') % sharenamemode
                 )
 
             # TODO this is a somewhat arbitrary restriction.
             if narrow:
                 ui.status(
                     _(b'(pooled storage not supported for narrow clones)\n')
                 )
                 sharepath = None
 
             if sharepath:
                 return clonewithshare(
                     ui,
                     peeropts,
                     sharepath,
                     source,
                     srcpeer,
                     dest,
                     pull=pull,
                     rev=revs,
                     update=update,
                     stream=stream,
                 )
 
         srcrepo = srcpeer.local()
 
         abspath = origsource
         if islocal(origsource):
             abspath = util.abspath(urlutil.urllocalpath(origsource))
 
         if islocal(dest):
             if os.path.exists(dest):
                 # only clean up directories we create ourselves
                 hgdir = os.path.realpath(os.path.join(dest, b".hg"))
                 cleandir = hgdir
             else:
                 cleandir = dest
 
         copy = False
         if (
             srcrepo
             and srcrepo.cancopy()
             and islocal(dest)
             and not phases.hassecret(srcrepo)
         ):
             copy = not pull and not revs
 
         # TODO this is a somewhat arbitrary restriction.
         if narrow:
             copy = False
 
         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(b'preoutgoing', throw=True, source=b'clone')
 
             destrootpath = urlutil.urllocalpath(dest)
             dest_reqs = localrepo.clone_requirements(ui, createopts, srcrepo)
             localrepo.createrepository(
                 ui,
                 destrootpath,
                 requirements=dest_reqs,
             )
             destrepo = localrepo.makelocalrepository(ui, destrootpath)
 
             destwlock = destrepo.wlock()
             destlock = destrepo.lock()
             from . import streamclone  # avoid cycle
 
             streamclone.local_copy(srcrepo, destrepo)
 
             # we need to re-init the repo after manually copying the data
             # into it
             destpeer = peer(srcrepo, peeropts, dest)
 
             # make the peer aware that is it already locked
             #
             # important:
             #
             #    We still need to release that lock at the end of the function
             destpeer.local()._lockref = weakref.ref(destlock)
             destpeer.local()._wlockref = weakref.ref(destwlock)
             # dirstate also needs to be copied because `_wlockref` has a reference
             # to it: this dirstate is saved to disk when the wlock is released
             destpeer.local().dirstate = destrepo.dirstate
 
             srcrepo.hook(
                 b'outgoing', source=b'clone', node=srcrepo.nodeconstants.nullhex
             )
         else:
             try:
                 # only pass ui when no srcrepo
                 destpeer = peer(
                     srcrepo or ui,
                     peeropts,
                     dest,
                     create=True,
                     createopts=createopts,
                 )
             except OSError as inst:
                 if inst.errno == errno.EEXIST:
                     cleandir = None
                     raise error.Abort(
                         _(b"destination '%s' already exists") % dest
                     )
                 raise
 
             if revs:
                 if not srcpeer.capable(b'lookup'):
                     raise error.Abort(
                         _(
                             b"src repository does not support "
                             b"revision lookup and so doesn't "
                             b"support clone by revision"
                         )
                     )
 
                 # TODO this is batchable.
                 remoterevs = []
                 for rev in revs:
                     with srcpeer.commandexecutor() as e:
                         remoterevs.append(
                             e.callcommand(
                                 b'lookup',
                                 {
                                     b'key': rev,
                                 },
                             ).result()
                         )
                 revs = remoterevs
 
                 checkout = revs[0]
             else:
                 revs = None
             local = destpeer.local()
             if local:
                 if narrow:
                     with local.wlock(), local.lock():
                         local.setnarrowpats(storeincludepats, storeexcludepats)
                         narrowspec.copytoworkingcopy(local)
 
                 u = urlutil.url(abspath)
                 defaulturl = bytes(u)
                 local.ui.setconfig(b'paths', b'default', defaulturl, b'clone')
                 if not stream:
                     if pull:
                         stream = False
                     else:
                         stream = None
                 # internal config: ui.quietbookmarkmove
                 overrides = {(b'ui', b'quietbookmarkmove'): True}
                 with local.ui.configoverride(overrides, b'clone'):
                     exchange.pull(
                         local,
                         srcpeer,
                         heads=revs,
                         streamclonerequested=stream,
                         includepats=storeincludepats,
                         excludepats=storeexcludepats,
                         depth=depth,
                     )
             elif srcrepo:
                 # TODO lift restriction once exchange.push() accepts narrow
                 # push.
                 if narrow:
                     raise error.Abort(
                         _(
                             b'narrow clone not available for '
                             b'remote destinations'
                         )
                     )
 
                 exchange.push(
                     srcrepo,
                     destpeer,
                     revs=revs,
                     bookmarks=srcrepo._bookmarks.keys(),
                 )
             else:
                 raise error.Abort(
                     _(b"clone from remote to remote not supported")
                 )
 
         cleandir = None
 
         destrepo = destpeer.local()
         if destrepo:
             template = uimod.samplehgrcs[b'cloned']
             u = urlutil.url(abspath)
             u.passwd = None
             defaulturl = bytes(u)
             destrepo.vfs.write(b'hgrc', util.tonativeeol(template % defaulturl))
             destrepo.ui.setconfig(b'paths', b'default', defaulturl, b'clone')
 
             if ui.configbool(b'experimental', b'remotenames'):
                 logexchange.pullremotenames(destrepo, srcpeer)
 
             if update:
                 if update is not True:
                     with srcpeer.commandexecutor() as e:
                         checkout = e.callcommand(
                             b'lookup',
                             {
                                 b'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:
                         if destrepo._activebookmark:
                             uprev = destrepo.lookup(destrepo._activebookmark)
                             update = destrepo._activebookmark
                         else:
                             uprev = destrepo._bookmarks[b'@']
                             update = b'@'
                         bn = destrepo[uprev].branch()
                         if bn == b'default':
                             status = _(b"updating to bookmark %s\n" % update)
                         else:
                             status = (
                                 _(b"updating to bookmark %s on branch %s\n")
                             ) % (update, bn)
                     except KeyError:
                         try:
                             uprev = destrepo.branchtip(b'default')
                         except error.RepoLookupError:
                             uprev = destrepo.lookup(b'tip')
                 if not status:
                     bn = destrepo[uprev].branch()
                     status = _(b"updating to branch %s\n") % bn
                 destrepo.ui.status(status)
                 _update(destrepo, uprev)
                 if update in destrepo._bookmarks:
                     bookmarks.activate(destrepo, update)
             if destlock is not None:
                 release(destlock)
             if destwlock is not None:
                 release(destlock)
             # here is a tiny windows were someone could end up writing the
             # repository before the cache are sure to be warm. This is "fine"
             # as the only "bad" outcome would be some slowness. That potential
             # slowness already affect reader.
             with destrepo.lock():
                 destrepo.updatecaches(caches=repositorymod.CACHES_POST_CLONE)
     finally:
         release(srclock, destlock, destwlock)
         if cleandir is not None:
             shutil.rmtree(cleandir, True)
         if srcpeer is not None:
             srcpeer.close()
         if destpeer and destpeer.local() is None:
             destpeer.close()
     return srcpeer, destpeer
 
 
 def _showstats(repo, stats, quietempty=False):
     if quietempty and stats.isempty():
         return
     repo.ui.status(
         _(
             b"%d files updated, %d files merged, "
             b"%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)"""
     repo.ui.deprecwarn(
         b'prefer merge.update() or merge.clean_update() over hg.updaterepo()',
         b'5.7',
     )
     return mergemod._update(
         repo,
         node,
         branchmerge=False,
         force=overwrite,
         labels=[b'working copy', b'destination'],
         updatecheck=updatecheck,
     )
 
 
 def update(repo, node, quietempty=False, updatecheck=None):
     """update the working directory to node"""
     stats = mergemod.update(repo[node], updatecheck=updatecheck)
     _showstats(repo, stats, quietempty)
     if stats.unresolvedcount:
         repo.ui.status(_(b"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 = mergemod.clean_update(repo[node])
     assert stats.unresolvedcount == 0
     if show_stats:
         _showstats(repo, stats, quietempty)
     return False
 
 
 # naming conflict in updatetotally()
 _clean = clean
 
 _VALID_UPDATECHECKS = {
     mergemod.UPDATECHECK_ABORT,
     mergemod.UPDATECHECK_NONE,
     mergemod.UPDATECHECK_LINEAR,
     mergemod.UPDATECHECK_NO_CONFLICT,
 }
 
 
 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 the UPDATECHECK_* constants
     defined in the merge module. Passing `None` will result in using the
     configured default.
 
      * 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
      * NO_CONFLICT: 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(b'commands', b'update.check')
         if updatecheck not in _VALID_UPDATECHECKS:
             # If not configured, or invalid value configured
             updatecheck = mergemod.UPDATECHECK_LINEAR
     if updatecheck not in _VALID_UPDATECHECKS:
         raise ValueError(
             r'Invalid updatecheck value %r (can accept %r)'
             % (updatecheck, _VALID_UPDATECHECKS)
         )
     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 == mergemod.UPDATECHECK_ABORT:
                 cmdutil.bailifchanged(repo, merge=False)
                 updatecheck = mergemod.UPDATECHECK_NONE
             ret = _update(repo, checkout, updatecheck=updatecheck)
 
         if not ret and movemarkfrom:
             if movemarkfrom == repo[b'.'].node():
                 pass  # no-op update
             elif bookmarks.update(repo, [movemarkfrom], repo[b'.'].node()):
                 b = ui.label(repo._activebookmark, b'bookmarks.active')
                 ui.status(_(b"updating bookmark %s\n") % b)
             else:
                 # this can happen with a non-linear update
                 b = ui.label(repo._activebookmark, b'bookmarks')
                 ui.status(_(b"(leaving bookmark %s)\n") % b)
                 bookmarks.deactivate(repo)
         elif brev in repo._bookmarks:
             if brev != repo._activebookmark:
                 b = ui.label(brev, b'bookmarks.active')
                 ui.status(_(b"(activating bookmark %s)\n") % b)
             bookmarks.activate(repo, brev)
         elif brev:
             if repo._activebookmark:
                 b = ui.label(repo._activebookmark, b'bookmarks')
                 ui.status(_(b"(leaving bookmark %s)\n") % b)
             bookmarks.deactivate(repo)
 
         if warndest:
             destutil.statusotherdests(ui, repo)
 
     return ret
 
 
 def merge(
     ctx,
     force=False,
     remind=True,
     labels=None,
 ):
     """Branch merge with node, resolving changes. Return true if any
     unresolved conflicts."""
     repo = ctx.repo()
     stats = mergemod.merge(ctx, force=force, labels=labels)
     _showstats(repo, stats)
     if stats.unresolvedcount:
         repo.ui.status(
             _(
                 b"use 'hg resolve' to retry unresolved file merges "
                 b"or 'hg merge --abort' to abandon\n"
             )
         )
     elif remind:
         repo.ui.status(_(b"(branch merge, don't forget to commit)\n"))
     return stats.unresolvedcount > 0
 
 
 def abortmerge(ui, repo):
     ms = mergestatemod.mergestate.read(repo)
     if ms.active():
         # there were conflicts
         node = ms.localctx.hex()
     else:
         # there were no conficts, mergestate was not stored
         node = repo[b'.'].hex()
 
     repo.ui.status(_(b"aborting the merge, updating back to %s\n") % node[:12])
     stats = mergemod.clean_update(repo[node])
     assert stats.unresolvedcount == 0
     _showstats(repo, stats)
 
 
 def _incoming(
     displaychlist,
     subreporecurse,
     ui,
     repo,
     source,
     opts,
     buffered=False,
     subpath=None,
 ):
     """
     Helper for incoming / gincoming.
     displaychlist gets called with
         (remoterepo, incomingchangesetlist, displayer) parameters,
     and is supposed to contain only code that can't be unified.
     """
     srcs = urlutil.get_pull_paths(repo, ui, [source])
     srcs = list(srcs)
     if len(srcs) != 1:
         msg = _(b'for now, incoming supports only a single source, %d provided')
         msg %= len(srcs)
         raise error.Abort(msg)
     path = srcs[0]
     source, branches = urlutil.parseurl(path.rawloc, opts.get(b'branch'))
     if subpath is not None:
         subpath = urlutil.url(subpath)
         if subpath.isabs():
             source = bytes(subpath)
         else:
             p = urlutil.url(source)
             p.path = os.path.normpath(b'%s/%s' % (p.path, subpath))
             source = bytes(p)
     other = peer(repo, opts, source)
     cleanupfn = other.close
     try:
         ui.status(_(b'comparing with %s\n') % urlutil.hidepassword(source))
         revs, checkout = addbranchrevs(repo, other, branches, opts.get(b'rev'))
 
         if revs:
             revs = [other.lookup(rev) for rev in revs]
         other, chlist, cleanupfn = bundlerepo.getremotechanges(
             ui, repo, other, revs, opts.get(b"bundle"), opts.get(b"force")
         )
 
         if not chlist:
             ui.status(_(b"no changes found\n"))
             return subreporecurse()
         ui.pager(b'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, subpath=None):
     def subreporecurse():
         ret = 1
         if opts.get(b'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(b'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 != repo.nullid
             ]
             if opts.get(b'no_merges') and len(parents) == 2:
                 continue
             count += 1
             displayer.show(other[n])
 
     return _incoming(
         display, subreporecurse, ui, repo, source, opts, subpath=subpath
     )
 
 
 def _outgoing(ui, repo, dests, opts, subpath=None):
     out = set()
     others = []
     for path in urlutil.get_push_paths(repo, ui, dests):
         dest = path.pushloc or path.loc
         if subpath is not None:
             subpath = urlutil.url(subpath)
             if subpath.isabs():
                 dest = bytes(subpath)
             else:
                 p = urlutil.url(dest)
                 p.path = os.path.normpath(b'%s/%s' % (p.path, subpath))
                 dest = bytes(p)
         branches = path.branch, opts.get(b'branch') or []
 
         ui.status(_(b'comparing with %s\n') % urlutil.hidepassword(dest))
         revs, checkout = addbranchrevs(repo, repo, branches, opts.get(b'rev'))
         if revs:
             revs = [repo[rev].node() for rev in logcmdutil.revrange(repo, revs)]
 
         other = peer(repo, opts, dest)
         try:
             outgoing = discovery.findcommonoutgoing(
                 repo, other, revs, force=opts.get(b'force')
             )
             o = outgoing.missing
             out.update(o)
             if not o:
                 scmutil.nochangesfound(repo.ui, repo, outgoing.excluded)
             others.append(other)
         except:  # re-raises
             other.close()
             raise
     # make sure this is ordered by revision number
     outgoing_revs = list(out)
     cl = repo.changelog
     outgoing_revs.sort(key=cl.rev)
     return outgoing_revs, others
 
 
 def _outgoing_recurse(ui, repo, dests, opts):
     ret = 1
     if opts.get(b'subrepos'):
         ctx = repo[None]
         for subpath in sorted(ctx.substate):
             sub = ctx.sub(subpath)
             ret = min(ret, sub.outgoing(ui, dests, opts))
     return ret
 
 
 def _outgoing_filter(repo, revs, opts):
     """apply revision filtering/ordering option for outgoing"""
     limit = logcmdutil.getlimit(opts)
     no_merges = opts.get(b'no_merges')
     if opts.get(b'newest_first'):
         revs.reverse()
     if limit is None and not no_merges:
         for r in revs:
             yield r
         return
 
     count = 0
     cl = repo.changelog
     for n in revs:
         if limit is not None and count >= limit:
             break
         parents = [p for p in cl.parents(n) if p != repo.nullid]
         if no_merges and len(parents) == 2:
             continue
         count += 1
         yield n
 
 
 def outgoing(ui, repo, dests, opts, subpath=None):
     if opts.get(b'graph'):
         logcmdutil.checkunsupportedgraphflags([], opts)
     o, others = _outgoing(ui, repo, dests, opts, subpath=subpath)
     ret = 1
     try:
         if o:
             ret = 0
 
             if opts.get(b'graph'):
                 revdag = logcmdutil.graphrevs(repo, o, opts)
                 ui.pager(b'outgoing')
                 displayer = logcmdutil.changesetdisplayer(
                     ui, repo, opts, buffered=True
                 )
                 logcmdutil.displaygraph(
                     ui, repo, revdag, displayer, graphmod.asciiedges
                 )
             else:
                 ui.pager(b'outgoing')
                 displayer = logcmdutil.changesetdisplayer(ui, repo, opts)
                 for n in _outgoing_filter(repo, o, opts):
                     displayer.show(repo[n])
                 displayer.close()
         for oth in others:
             cmdutil.outgoinghooks(ui, repo, oth, opts, o)
             ret = min(ret, _outgoing_recurse(ui, repo, dests, opts))
         return ret  # exit code is zero since we found outgoing changes
     finally:
         for oth in others:
             oth.close()
 
 
 def verify(repo, level=None):
     """verify the consistency of a repository"""
     ret = verifymod.verify(repo, level=level)
 
     # 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(
         b"filelog(%s)", util.pathto(repo.root, repo.getcwd(), b'.hgsubstate')
     )
 
     if revs:
         repo.ui.status(_(b'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(b'%d: %s\n' % (rev, e))
             except Exception:
                 repo.ui.warn(
                     _(b'.hgsubstate is corrupt in revision %s\n')
                     % short(ctx.node())
                 )
 
     return ret
 
 
 def remoteui(src, opts):
     """build a remote ui from ui or repo and opts"""
     if util.safehasattr(src, b'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 b'ssh', b'remotecmd':
         v = opts.get(o) or src.config(b'ui', o)
         if v:
             dst.setconfig(b"ui", o, v, b'copied')
 
     # copy bundle-specific options
     r = src.config(b'bundle', b'mainreporoot')
     if r:
         dst.setconfig(b'bundle', b'mainreporoot', r, b'copied')
 
     # copy selected local settings to the remote ui
     for sect in (b'auth', b'hostfingerprints', b'hostsecurity', b'http_proxy'):
         for key, val in src.configitems(sect):
             dst.setconfig(sect, key, val, b'copied')
     v = src.config(b'web', b'cacerts')
     if v:
         dst.setconfig(b'web', b'cacerts', util.expandpath(v), b'copied')
 
     return dst
 
 
 # Files of interest
 # Used to check if the repository has changed looking at mtime and size of
 # these files.
 foi = [
     (b'spath', b'00changelog.i'),
     (b'spath', b'phaseroots'),  # ! phase can change content at the same size
     (b'spath', b'obsstore'),
     (b'path', b'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

mercurial/util.py

@@ -1,3410 +1,3361 b''
 # util.py - Mercurial utility functions and platform specific implementations
 #
 #  Copyright 2005 K. Thananchayan <[e-mail]>
 #  Copyright 2005-2007 Olivia Mackall <[e-mail]>
 #  Copyright 2006 Vadim Gelfer <[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.
 
 """Mercurial utility functions and platform specific implementations.
 
 This contains helper routines that are independent of the SCM core and
 hide platform-specific details from the core.
 """
 
 from __future__ import absolute_import, print_function
 
 import abc
 import collections
 import contextlib
 import errno
 import gc
 import hashlib
 import itertools
 import locale
 import mmap
 import os
 import platform as pyplatform
 import re as remod
 import shutil
 import stat
 import sys
 import time
 import traceback
 import warnings
 
 from .node import hex
 from .thirdparty import attr
 from .pycompat import (
     delattr,
     getattr,
     open,
     setattr,
 )
 from .node import hex
 from hgdemandimport import tracing
 from . import (
     encoding,
     error,
     i18n,
     policy,
     pycompat,
     urllibcompat,
 )
 from .utils import (
     compression,
     hashutil,
     procutil,
     stringutil,
-    urlutil,
 )
 
 if pycompat.TYPE_CHECKING:
     from typing import (
         Iterator,
         List,
         Optional,
         Tuple,
     )
 
 
 base85 = policy.importmod('base85')
 osutil = policy.importmod('osutil')
 
 b85decode = base85.b85decode
 b85encode = base85.b85encode
 
 cookielib = pycompat.cookielib
 httplib = pycompat.httplib
 pickle = pycompat.pickle
 safehasattr = pycompat.safehasattr
 socketserver = pycompat.socketserver
 bytesio = pycompat.bytesio
 # TODO deprecate stringio name, as it is a lie on Python 3.
 stringio = bytesio
 xmlrpclib = pycompat.xmlrpclib
 
 httpserver = urllibcompat.httpserver
 urlerr = urllibcompat.urlerr
 urlreq = urllibcompat.urlreq
 
 # workaround for win32mbcs
 _filenamebytestr = pycompat.bytestr
 
 if pycompat.iswindows:
     from . import windows as platform
 else:
     from . import posix as platform
 
 _ = i18n._
 
 abspath = platform.abspath
 bindunixsocket = platform.bindunixsocket
 cachestat = platform.cachestat
 checkexec = platform.checkexec
 checklink = platform.checklink
 copymode = platform.copymode
 expandglobs = platform.expandglobs
 getfsmountpoint = platform.getfsmountpoint
 getfstype = platform.getfstype
 get_password = platform.get_password
 groupmembers = platform.groupmembers
 groupname = platform.groupname
 isexec = platform.isexec
 isowner = platform.isowner
 listdir = osutil.listdir
 localpath = platform.localpath
 lookupreg = platform.lookupreg
 makedir = platform.makedir
 nlinks = platform.nlinks
 normpath = platform.normpath
 normcase = platform.normcase
 normcasespec = platform.normcasespec
 normcasefallback = platform.normcasefallback
 openhardlinks = platform.openhardlinks
 oslink = platform.oslink
 parsepatchoutput = platform.parsepatchoutput
 pconvert = platform.pconvert
 poll = platform.poll
 posixfile = platform.posixfile
 readlink = platform.readlink
 rename = platform.rename
 removedirs = platform.removedirs
 samedevice = platform.samedevice
 samefile = platform.samefile
 samestat = platform.samestat
 setflags = platform.setflags
 split = platform.split
 statfiles = getattr(osutil, 'statfiles', platform.statfiles)
 statisexec = platform.statisexec
 statislink = platform.statislink
 umask = platform.umask
 unlink = platform.unlink
 username = platform.username
 
 
 def setumask(val):
     # type: (int) -> None
     '''updates the umask. used by chg server'''
     if pycompat.iswindows:
         return
     os.umask(val)
     global umask
     platform.umask = umask = val & 0o777
 
 
 # small compat layer
 compengines = compression.compengines
 SERVERROLE = compression.SERVERROLE
 CLIENTROLE = compression.CLIENTROLE
 
 try:
     recvfds = osutil.recvfds
 except AttributeError:
     pass
 
 # Python compatibility
 
 _notset = object()
 
 
 def bitsfrom(container):
     bits = 0
     for bit in container:
         bits |= bit
     return bits
 
 
 # python 2.6 still have deprecation warning enabled by default. We do not want
 # to display anything to standard user so detect if we are running test and
 # only use python deprecation warning in this case.
 _dowarn = bool(encoding.environ.get(b'HGEMITWARNINGS'))
 if _dowarn:
     # explicitly unfilter our warning for python 2.7
     #
     # The option of setting PYTHONWARNINGS in the test runner was investigated.
     # However, module name set through PYTHONWARNINGS was exactly matched, so
     # we cannot set 'mercurial' and have it match eg: 'mercurial.scmutil'. This
     # makes the whole PYTHONWARNINGS thing useless for our usecase.
     warnings.filterwarnings('default', '', DeprecationWarning, 'mercurial')
     warnings.filterwarnings('default', '', DeprecationWarning, 'hgext')
     warnings.filterwarnings('default', '', DeprecationWarning, 'hgext3rd')
 if _dowarn and pycompat.ispy3:
     # silence warning emitted by passing user string to re.sub()
     warnings.filterwarnings(
         'ignore', 'bad escape', DeprecationWarning, 'mercurial'
     )
     warnings.filterwarnings(
         'ignore', 'invalid escape sequence', DeprecationWarning, 'mercurial'
     )
     # TODO: reinvent imp.is_frozen()
     warnings.filterwarnings(
         'ignore',
         'the imp module is deprecated',
         DeprecationWarning,
         'mercurial',
     )
 
 
 def nouideprecwarn(msg, version, stacklevel=1):
     """Issue an python native deprecation warning
 
     This is a noop outside of tests, use 'ui.deprecwarn' when possible.
     """
     if _dowarn:
         msg += (
             b"\n(compatibility will be dropped after Mercurial-%s,"
             b" update your code.)"
         ) % version
         warnings.warn(pycompat.sysstr(msg), DeprecationWarning, stacklevel + 1)
         # on python 3 with chg, we will need to explicitly flush the output
         sys.stderr.flush()
 
 
 DIGESTS = {
     b'md5': hashlib.md5,
     b'sha1': hashutil.sha1,
     b'sha512': hashlib.sha512,
 }
 # List of digest types from strongest to weakest
 DIGESTS_BY_STRENGTH = [b'sha512', b'sha1', b'md5']
 
 for k in DIGESTS_BY_STRENGTH:
     assert k in DIGESTS
 
 
 class digester(object):
     """helper to compute digests.
 
     This helper can be used to compute one or more digests given their name.
 
     >>> d = digester([b'md5', b'sha1'])
     >>> d.update(b'foo')
     >>> [k for k in sorted(d)]
     ['md5', 'sha1']
     >>> d[b'md5']
     'acbd18db4cc2f85cedef654fccc4a4d8'
     >>> d[b'sha1']
     '0beec7b5ea3f0fdbc95d0dd47f3c5bc275da8a33'
     >>> digester.preferred([b'md5', b'sha1'])
     'sha1'
     """
 
     def __init__(self, digests, s=b''):
         self._hashes = {}
         for k in digests:
             if k not in DIGESTS:
                 raise error.Abort(_(b'unknown digest type: %s') % k)
             self._hashes[k] = DIGESTS[k]()
         if s:
             self.update(s)
 
     def update(self, data):
         for h in self._hashes.values():
             h.update(data)
 
     def __getitem__(self, key):
         if key not in DIGESTS:
             raise error.Abort(_(b'unknown digest type: %s') % k)
         return hex(self._hashes[key].digest())
 
     def __iter__(self):
         return iter(self._hashes)
 
     @staticmethod
     def preferred(supported):
         """returns the strongest digest type in both supported and DIGESTS."""
 
         for k in DIGESTS_BY_STRENGTH:
             if k in supported:
                 return k
         return None
 
 
 class digestchecker(object):
     """file handle wrapper that additionally checks content against a given
     size and digests.
 
         d = digestchecker(fh, size, {'md5': '...'})
 
     When multiple digests are given, all of them are validated.
     """
 
     def __init__(self, fh, size, digests):
         self._fh = fh
         self._size = size
         self._got = 0
         self._digests = dict(digests)
         self._digester = digester(self._digests.keys())
 
     def read(self, length=-1):
         content = self._fh.read(length)
         self._digester.update(content)
         self._got += len(content)
         return content
 
     def validate(self):
         if self._size != self._got:
             raise error.Abort(
                 _(b'size mismatch: expected %d, got %d')
                 % (self._size, self._got)
             )
         for k, v in self._digests.items():
             if v != self._digester[k]:
                 # i18n: first parameter is a digest name
                 raise error.Abort(
                     _(b'%s mismatch: expected %s, got %s')
                     % (k, v, self._digester[k])
                 )
 
 
 try:
     buffer = buffer  # pytype: disable=name-error
 except NameError:
 
     def buffer(sliceable, offset=0, length=None):
         if length is not None:
             return memoryview(sliceable)[offset : offset + length]
         return memoryview(sliceable)[offset:]
 
 
 _chunksize = 4096
 
 
 class bufferedinputpipe(object):
     """a manually buffered input pipe
 
     Python will not let us use buffered IO and lazy reading with 'polling' at
     the same time. We cannot probe the buffer state and select will not detect
     that data are ready to read if they are already buffered.
 
     This class let us work around that by implementing its own buffering
     (allowing efficient readline) while offering a way to know if the buffer is
     empty from the output (allowing collaboration of the buffer with polling).
 
     This class lives in the 'util' module because it makes use of the 'os'
     module from the python stdlib.
     """
 
     def __new__(cls, fh):
         # If we receive a fileobjectproxy, we need to use a variation of this
         # class that notifies observers about activity.
         if isinstance(fh, fileobjectproxy):
             cls = observedbufferedinputpipe
 
         return super(bufferedinputpipe, cls).__new__(cls)
 
     def __init__(self, input):
         self._input = input
         self._buffer = []
         self._eof = False
         self._lenbuf = 0
 
     @property
     def hasbuffer(self):
         """True is any data is currently buffered
 
         This will be used externally a pre-step for polling IO. If there is
         already data then no polling should be set in place."""
         return bool(self._buffer)
 
     @property
     def closed(self):
         return self._input.closed
 
     def fileno(self):
         return self._input.fileno()
 
     def close(self):
         return self._input.close()
 
     def read(self, size):
         while (not self._eof) and (self._lenbuf < size):
             self._fillbuffer()
         return self._frombuffer(size)
 
     def unbufferedread(self, size):
         if not self._eof and self._lenbuf == 0:
             self._fillbuffer(max(size, _chunksize))
         return self._frombuffer(min(self._lenbuf, size))
 
     def readline(self, *args, **kwargs):
         if len(self._buffer) > 1:
             # this should not happen because both read and readline end with a
             # _frombuffer call that collapse it.
             self._buffer = [b''.join(self._buffer)]
             self._lenbuf = len(self._buffer[0])
         lfi = -1
         if self._buffer:
             lfi = self._buffer[-1].find(b'\n')
         while (not self._eof) and lfi < 0:
             self._fillbuffer()
             if self._buffer:
                 lfi = self._buffer[-1].find(b'\n')
         size = lfi + 1
         if lfi < 0:  # end of file
             size = self._lenbuf
         elif len(self._buffer) > 1:
             # we need to take previous chunks into account
             size += self._lenbuf - len(self._buffer[-1])
         return self._frombuffer(size)
 
     def _frombuffer(self, size):
         """return at most 'size' data from the buffer
 
         The data are removed from the buffer."""
         if size == 0 or not self._buffer:
             return b''
         buf = self._buffer[0]
         if len(self._buffer) > 1:
             buf = b''.join(self._buffer)
 
         data = buf[:size]
         buf = buf[len(data) :]
         if buf:
             self._buffer = [buf]
             self._lenbuf = len(buf)
         else:
             self._buffer = []
             self._lenbuf = 0
         return data
 
     def _fillbuffer(self, size=_chunksize):
         """read data to the buffer"""
         data = os.read(self._input.fileno(), size)
         if not data:
             self._eof = True
         else:
             self._lenbuf += len(data)
             self._buffer.append(data)
 
         return data
 
 
 def mmapread(fp, size=None):
     if size == 0:
         # size of 0 to mmap.mmap() means "all data"
         # rather than "zero bytes", so special case that.
         return b''
     elif size is None:
         size = 0
     fd = getattr(fp, 'fileno', lambda: fp)()
     try:
         return mmap.mmap(fd, size, access=mmap.ACCESS_READ)
     except ValueError:
         # Empty files cannot be mmapped, but mmapread should still work.  Check
         # if the file is empty, and if so, return an empty buffer.
         if os.fstat(fd).st_size == 0:
             return b''
         raise
 
 
 class fileobjectproxy(object):
     """A proxy around file objects that tells a watcher when events occur.
 
     This type is intended to only be used for testing purposes. Think hard
     before using it in important code.
     """
 
     __slots__ = (
         '_orig',
         '_observer',
     )
 
     def __init__(self, fh, observer):
         object.__setattr__(self, '_orig', fh)
         object.__setattr__(self, '_observer', observer)
 
     def __getattribute__(self, name):
         ours = {
             '_observer',
             # IOBase
             'close',
             # closed if a property
             'fileno',
             'flush',
             'isatty',
             'readable',
             'readline',
             'readlines',
             'seek',
             'seekable',
             'tell',
             'truncate',
             'writable',
             'writelines',
             # RawIOBase
             'read',
             'readall',
             'readinto',
             'write',
             # BufferedIOBase
             # raw is a property
             'detach',
             # read defined above
             'read1',
             # readinto defined above
             # write defined above
         }
 
         # We only observe some methods.
         if name in ours:
             return object.__getattribute__(self, name)
 
         return getattr(object.__getattribute__(self, '_orig'), name)
 
     def __nonzero__(self):
         return bool(object.__getattribute__(self, '_orig'))
 
     __bool__ = __nonzero__
 
     def __delattr__(self, name):
         return delattr(object.__getattribute__(self, '_orig'), name)
 
     def __setattr__(self, name, value):
         return setattr(object.__getattribute__(self, '_orig'), name, value)
 
     def __iter__(self):
         return object.__getattribute__(self, '_orig').__iter__()
 
     def _observedcall(self, name, *args, **kwargs):
         # Call the original object.
         orig = object.__getattribute__(self, '_orig')
         res = getattr(orig, name)(*args, **kwargs)
 
         # Call a method on the observer of the same name with arguments
         # so it can react, log, etc.
         observer = object.__getattribute__(self, '_observer')
         fn = getattr(observer, name, None)
         if fn:
             fn(res, *args, **kwargs)
 
         return res
 
     def close(self, *args, **kwargs):
         return object.__getattribute__(self, '_observedcall')(
             'close', *args, **kwargs
         )
 
     def fileno(self, *args, **kwargs):
         return object.__getattribute__(self, '_observedcall')(
             'fileno', *args, **kwargs
         )
 
     def flush(self, *args, **kwargs):
         return object.__getattribute__(self, '_observedcall')(
             'flush', *args, **kwargs
         )
 
     def isatty(self, *args, **kwargs):
         return object.__getattribute__(self, '_observedcall')(
             'isatty', *args, **kwargs
         )
 
     def readable(self, *args, **kwargs):
         return object.__getattribute__(self, '_observedcall')(
             'readable', *args, **kwargs
         )
 
     def readline(self, *args, **kwargs):
         return object.__getattribute__(self, '_observedcall')(
             'readline', *args, **kwargs
         )
 
     def readlines(self, *args, **kwargs):
         return object.__getattribute__(self, '_observedcall')(
             'readlines', *args, **kwargs
         )
 
     def seek(self, *args, **kwargs):
         return object.__getattribute__(self, '_observedcall')(
             'seek', *args, **kwargs
         )
 
     def seekable(self, *args, **kwargs):
         return object.__getattribute__(self, '_observedcall')(
             'seekable', *args, **kwargs
         )
 
     def tell(self, *args, **kwargs):
         return object.__getattribute__(self, '_observedcall')(
             'tell', *args, **kwargs
         )
 
     def truncate(self, *args, **kwargs):
         return object.__getattribute__(self, '_observedcall')(
             'truncate', *args, **kwargs
         )
 
     def writable(self, *args, **kwargs):
         return object.__getattribute__(self, '_observedcall')(
             'writable', *args, **kwargs
         )
 
     def writelines(self, *args, **kwargs):
         return object.__getattribute__(self, '_observedcall')(
             'writelines', *args, **kwargs
         )
 
     def read(self, *args, **kwargs):
         return object.__getattribute__(self, '_observedcall')(
             'read', *args, **kwargs
         )
 
     def readall(self, *args, **kwargs):
         return object.__getattribute__(self, '_observedcall')(
             'readall', *args, **kwargs
         )
 
     def readinto(self, *args, **kwargs):
         return object.__getattribute__(self, '_observedcall')(
             'readinto', *args, **kwargs
         )
 
     def write(self, *args, **kwargs):
         return object.__getattribute__(self, '_observedcall')(
             'write', *args, **kwargs
         )
 
     def detach(self, *args, **kwargs):
         return object.__getattribute__(self, '_observedcall')(
             'detach', *args, **kwargs
         )
 
     def read1(self, *args, **kwargs):
         return object.__getattribute__(self, '_observedcall')(
             'read1', *args, **kwargs
         )
 
 
 class observedbufferedinputpipe(bufferedinputpipe):
     """A variation of bufferedinputpipe that is aware of fileobjectproxy.
 
     ``bufferedinputpipe`` makes low-level calls to ``os.read()`` that
     bypass ``fileobjectproxy``. Because of this, we need to make
     ``bufferedinputpipe`` aware of these operations.
 
     This variation of ``bufferedinputpipe`` can notify observers about
     ``os.read()`` events. It also re-publishes other events, such as
     ``read()`` and ``readline()``.
     """
 
     def _fillbuffer(self):
         res = super(observedbufferedinputpipe, self)._fillbuffer()
 
         fn = getattr(self._input._observer, 'osread', None)
         if fn:
             fn(res, _chunksize)
 
         return res
 
     # We use different observer methods because the operation isn't
     # performed on the actual file object but on us.
     def read(self, size):
         res = super(observedbufferedinputpipe, self).read(size)
 
         fn = getattr(self._input._observer, 'bufferedread', None)
         if fn:
             fn(res, size)
 
         return res
 
     def readline(self, *args, **kwargs):
         res = super(observedbufferedinputpipe, self).readline(*args, **kwargs)
 
         fn = getattr(self._input._observer, 'bufferedreadline', None)
         if fn:
             fn(res)
 
         return res
 
 
 PROXIED_SOCKET_METHODS = {
     'makefile',
     'recv',
     'recvfrom',
     'recvfrom_into',
     'recv_into',
     'send',
     'sendall',
     'sendto',
     'setblocking',
     'settimeout',
     'gettimeout',
     'setsockopt',
 }
 
 
 class socketproxy(object):
     """A proxy around a socket that tells a watcher when events occur.
 
     This is like ``fileobjectproxy`` except for sockets.
 
     This type is intended to only be used for testing purposes. Think hard
     before using it in important code.
     """
 
     __slots__ = (
         '_orig',
         '_observer',
     )
 
     def __init__(self, sock, observer):
         object.__setattr__(self, '_orig', sock)
         object.__setattr__(self, '_observer', observer)
 
     def __getattribute__(self, name):
         if name in PROXIED_SOCKET_METHODS:
             return object.__getattribute__(self, name)
 
         return getattr(object.__getattribute__(self, '_orig'), name)
 
     def __delattr__(self, name):
         return delattr(object.__getattribute__(self, '_orig'), name)
 
     def __setattr__(self, name, value):
         return setattr(object.__getattribute__(self, '_orig'), name, value)
 
     def __nonzero__(self):
         return bool(object.__getattribute__(self, '_orig'))
 
     __bool__ = __nonzero__
 
     def _observedcall(self, name, *args, **kwargs):
         # Call the original object.
         orig = object.__getattribute__(self, '_orig')
         res = getattr(orig, name)(*args, **kwargs)
 
         # Call a method on the observer of the same name with arguments
         # so it can react, log, etc.
         observer = object.__getattribute__(self, '_observer')
         fn = getattr(observer, name, None)
         if fn:
             fn(res, *args, **kwargs)
 
         return res
 
     def makefile(self, *args, **kwargs):
         res = object.__getattribute__(self, '_observedcall')(
             'makefile', *args, **kwargs
         )
 
         # The file object may be used for I/O. So we turn it into a
         # proxy using our observer.
         observer = object.__getattribute__(self, '_observer')
         return makeloggingfileobject(
             observer.fh,
             res,
             observer.name,
             reads=observer.reads,
             writes=observer.writes,
             logdata=observer.logdata,
             logdataapis=observer.logdataapis,
         )
 
     def recv(self, *args, **kwargs):
         return object.__getattribute__(self, '_observedcall')(
             'recv', *args, **kwargs
         )
 
     def recvfrom(self, *args, **kwargs):
         return object.__getattribute__(self, '_observedcall')(
             'recvfrom', *args, **kwargs
         )
 
     def recvfrom_into(self, *args, **kwargs):
         return object.__getattribute__(self, '_observedcall')(
             'recvfrom_into', *args, **kwargs
         )
 
     def recv_into(self, *args, **kwargs):
         return object.__getattribute__(self, '_observedcall')(
             'recv_info', *args, **kwargs
         )
 
     def send(self, *args, **kwargs):
         return object.__getattribute__(self, '_observedcall')(
             'send', *args, **kwargs
         )
 
     def sendall(self, *args, **kwargs):
         return object.__getattribute__(self, '_observedcall')(
             'sendall', *args, **kwargs
         )
 
     def sendto(self, *args, **kwargs):
         return object.__getattribute__(self, '_observedcall')(
             'sendto', *args, **kwargs
         )
 
     def setblocking(self, *args, **kwargs):
         return object.__getattribute__(self, '_observedcall')(
             'setblocking', *args, **kwargs
         )
 
     def settimeout(self, *args, **kwargs):
         return object.__getattribute__(self, '_observedcall')(
             'settimeout', *args, **kwargs
         )
 
     def gettimeout(self, *args, **kwargs):
         return object.__getattribute__(self, '_observedcall')(
             'gettimeout', *args, **kwargs
         )
 
     def setsockopt(self, *args, **kwargs):
         return object.__getattribute__(self, '_observedcall')(
             'setsockopt', *args, **kwargs
         )
 
 
 class baseproxyobserver(object):
     def __init__(self, fh, name, logdata, logdataapis):
         self.fh = fh
         self.name = name
         self.logdata = logdata
         self.logdataapis = logdataapis
 
     def _writedata(self, data):
         if not self.logdata:
             if self.logdataapis:
                 self.fh.write(b'\n')
                 self.fh.flush()
             return
 
         # Simple case writes all data on a single line.
         if b'\n' not in data:
             if self.logdataapis:
                 self.fh.write(b': %s\n' % stringutil.escapestr(data))
             else:
                 self.fh.write(
                     b'%s>     %s\n' % (self.name, stringutil.escapestr(data))
                 )
             self.fh.flush()
             return
 
         # Data with newlines is written to multiple lines.
         if self.logdataapis:
             self.fh.write(b':\n')
 
         lines = data.splitlines(True)
         for line in lines:
             self.fh.write(
                 b'%s>     %s\n' % (self.name, stringutil.escapestr(line))
             )
         self.fh.flush()
 
 
 class fileobjectobserver(baseproxyobserver):
     """Logs file object activity."""
 
     def __init__(
         self, fh, name, reads=True, writes=True, logdata=False, logdataapis=True
     ):
         super(fileobjectobserver, self).__init__(fh, name, logdata, logdataapis)
         self.reads = reads
         self.writes = writes
 
     def read(self, res, size=-1):
         if not self.reads:
             return
         # Python 3 can return None from reads at EOF instead of empty strings.
         if res is None:
             res = b''
 
         if size == -1 and res == b'':
             # Suppress pointless read(-1) calls that return
             # nothing. These happen _a lot_ on Python 3, and there
             # doesn't seem to be a better workaround to have matching
             # Python 2 and 3 behavior. :(
             return
 
         if self.logdataapis:
             self.fh.write(b'%s> read(%d) -> %d' % (self.name, size, len(res)))
 
         self._writedata(res)
 
     def readline(self, res, limit=-1):
         if not self.reads:
             return
 
         if self.logdataapis:
             self.fh.write(b'%s> readline() -> %d' % (self.name, len(res)))
 
         self._writedata(res)
 
     def readinto(self, res, dest):
         if not self.reads:
             return
 
         if self.logdataapis:
             self.fh.write(
                 b'%s> readinto(%d) -> %r' % (self.name, len(dest), res)
             )
 
         data = dest[0:res] if res is not None else b''
 
         # _writedata() uses "in" operator and is confused by memoryview because
         # characters are ints on Python 3.
         if isinstance(data, memoryview):
             data = data.tobytes()
 
         self._writedata(data)
 
     def write(self, res, data):
         if not self.writes:
             return
 
         # Python 2 returns None from some write() calls. Python 3 (reasonably)
         # returns the integer bytes written.
         if res is None and data:
             res = len(data)
 
         if self.logdataapis:
             self.fh.write(b'%s> write(%d) -> %r' % (self.name, len(data), res))
 
         self._writedata(data)
 
     def flush(self, res):
         if not self.writes:
             return
 
         self.fh.write(b'%s> flush() -> %r\n' % (self.name, res))
 
     # For observedbufferedinputpipe.
     def bufferedread(self, res, size):
         if not self.reads:
             return
 
         if self.logdataapis:
             self.fh.write(
                 b'%s> bufferedread(%d) -> %d' % (self.name, size, len(res))
             )
 
         self._writedata(res)
 
     def bufferedreadline(self, res):
         if not self.reads:
             return
 
         if self.logdataapis:
             self.fh.write(
                 b'%s> bufferedreadline() -> %d' % (self.name, len(res))
             )
 
         self._writedata(res)
 
 
 def makeloggingfileobject(
     logh, fh, name, reads=True, writes=True, logdata=False, logdataapis=True
 ):
     """Turn a file object into a logging file object."""
 
     observer = fileobjectobserver(
         logh,
         name,
         reads=reads,
         writes=writes,
         logdata=logdata,
         logdataapis=logdataapis,
     )
     return fileobjectproxy(fh, observer)
 
 
 class socketobserver(baseproxyobserver):
     """Logs socket activity."""
 
     def __init__(
         self,
         fh,
         name,
         reads=True,
         writes=True,
         states=True,
         logdata=False,
         logdataapis=True,
     ):
         super(socketobserver, self).__init__(fh, name, logdata, logdataapis)
         self.reads = reads
         self.writes = writes
         self.states = states
 
     def makefile(self, res, mode=None, bufsize=None):
         if not self.states:
             return
 
         self.fh.write(b'%s> makefile(%r, %r)\n' % (self.name, mode, bufsize))
 
     def recv(self, res, size, flags=0):
         if not self.reads:
             return
 
         if self.logdataapis:
             self.fh.write(
                 b'%s> recv(%d, %d) -> %d' % (self.name, size, flags, len(res))
             )
         self._writedata(res)
 
     def recvfrom(self, res, size, flags=0):
         if not self.reads:
             return
 
         if self.logdataapis:
             self.fh.write(
                 b'%s> recvfrom(%d, %d) -> %d'
                 % (self.name, size, flags, len(res[0]))
             )
 
         self._writedata(res[0])
 
     def recvfrom_into(self, res, buf, size, flags=0):
         if not self.reads:
             return
 
         if self.logdataapis:
             self.fh.write(
                 b'%s> recvfrom_into(%d, %d) -> %d'
                 % (self.name, size, flags, res[0])
             )
 
         self._writedata(buf[0 : res[0]])
 
     def recv_into(self, res, buf, size=0, flags=0):
         if not self.reads:
             return
 
         if self.logdataapis:
             self.fh.write(
                 b'%s> recv_into(%d, %d) -> %d' % (self.name, size, flags, res)
             )
 
         self._writedata(buf[0:res])
 
     def send(self, res, data, flags=0):
         if not self.writes:
             return
 
         self.fh.write(
             b'%s> send(%d, %d) -> %d' % (self.name, len(data), flags, len(res))
         )
         self._writedata(data)
 
     def sendall(self, res, data, flags=0):
         if not self.writes:
             return
 
         if self.logdataapis:
             # Returns None on success. So don't bother reporting return value.
             self.fh.write(
                 b'%s> sendall(%d, %d)' % (self.name, len(data), flags)
             )
 
         self._writedata(data)
 
     def sendto(self, res, data, flagsoraddress, address=None):
         if not self.writes:
             return
 
         if address:
             flags = flagsoraddress
         else:
             flags = 0
 
         if self.logdataapis:
             self.fh.write(
                 b'%s> sendto(%d, %d, %r) -> %d'
                 % (self.name, len(data), flags, address, res)
             )
 
         self._writedata(data)
 
     def setblocking(self, res, flag):
         if not self.states:
             return
 
         self.fh.write(b'%s> setblocking(%r)\n' % (self.name, flag))
 
     def settimeout(self, res, value):
         if not self.states:
             return
 
         self.fh.write(b'%s> settimeout(%r)\n' % (self.name, value))
 
     def gettimeout(self, res):
         if not self.states:
             return
 
         self.fh.write(b'%s> gettimeout() -> %f\n' % (self.name, res))
 
     def setsockopt(self, res, level, optname, value):
         if not self.states:
             return
 
         self.fh.write(
             b'%s> setsockopt(%r, %r, %r) -> %r\n'
             % (self.name, level, optname, value, res)
         )
 
 
 def makeloggingsocket(
     logh,
     fh,
     name,
     reads=True,
     writes=True,
     states=True,
     logdata=False,
     logdataapis=True,
 ):
     """Turn a socket into a logging socket."""
 
     observer = socketobserver(
         logh,
         name,
         reads=reads,
         writes=writes,
         states=states,
         logdata=logdata,
         logdataapis=logdataapis,
     )
     return socketproxy(fh, observer)
 
 
 def version():
     """Return version information if available."""
     try:
         from . import __version__
 
         return __version__.version
     except ImportError:
         return b'unknown'
 
 
 def versiontuple(v=None, n=4):
     """Parses a Mercurial version string into an N-tuple.
 
     The version string to be parsed is specified with the ``v`` argument.
     If it isn't defined, the current Mercurial version string will be parsed.
 
     ``n`` can be 2, 3, or 4. Here is how some version strings map to
     returned values:
 
     >>> v = b'3.6.1+190-df9b73d2d444'
     >>> versiontuple(v, 2)
     (3, 6)
     >>> versiontuple(v, 3)
     (3, 6, 1)
     >>> versiontuple(v, 4)
     (3, 6, 1, '190-df9b73d2d444')
 
     >>> versiontuple(b'3.6.1+190-df9b73d2d444+20151118')
     (3, 6, 1, '190-df9b73d2d444+20151118')
 
     >>> v = b'3.6'
     >>> versiontuple(v, 2)
     (3, 6)
     >>> versiontuple(v, 3)
     (3, 6, None)
     >>> versiontuple(v, 4)
     (3, 6, None, None)
 
     >>> v = b'3.9-rc'
     >>> versiontuple(v, 2)
     (3, 9)
     >>> versiontuple(v, 3)
     (3, 9, None)
     >>> versiontuple(v, 4)
     (3, 9, None, 'rc')
 
     >>> v = b'3.9-rc+2-02a8fea4289b'
     >>> versiontuple(v, 2)
     (3, 9)
     >>> versiontuple(v, 3)
     (3, 9, None)
     >>> versiontuple(v, 4)
     (3, 9, None, 'rc+2-02a8fea4289b')
 
     >>> versiontuple(b'4.6rc0')
     (4, 6, None, 'rc0')
     >>> versiontuple(b'4.6rc0+12-425d55e54f98')
     (4, 6, None, 'rc0+12-425d55e54f98')
     >>> versiontuple(b'.1.2.3')
     (None, None, None, '.1.2.3')
     >>> versiontuple(b'12.34..5')
     (12, 34, None, '..5')
     >>> versiontuple(b'1.2.3.4.5.6')
     (1, 2, 3, '.4.5.6')
     """
     if not v:
         v = version()
     m = remod.match(br'(\d+(?:\.\d+){,2})[+-]?(.*)', v)
     if not m:
         vparts, extra = b'', v
     elif m.group(2):
         vparts, extra = m.groups()
     else:
         vparts, extra = m.group(1), None
 
     assert vparts is not None  # help pytype
 
     vints = []
     for i in vparts.split(b'.'):
         try:
             vints.append(int(i))
         except ValueError:
             break
     # (3, 6) -> (3, 6, None)
     while len(vints) < 3:
         vints.append(None)
 
     if n == 2:
         return (vints[0], vints[1])
     if n == 3:
         return (vints[0], vints[1], vints[2])
     if n == 4:
         return (vints[0], vints[1], vints[2], extra)
 
     raise error.ProgrammingError(b"invalid version part request: %d" % n)
 
 
 def cachefunc(func):
     '''cache the result of function calls'''
     # XXX doesn't handle keywords args
     if func.__code__.co_argcount == 0:
         listcache = []
 
         def f():
             if len(listcache) == 0:
                 listcache.append(func())
             return listcache[0]
 
         return f
     cache = {}
     if func.__code__.co_argcount == 1:
         # we gain a small amount of time because
         # we don't need to pack/unpack the list
         def f(arg):
             if arg not in cache:
                 cache[arg] = func(arg)
             return cache[arg]
 
     else:
 
         def f(*args):
             if args not in cache:
                 cache[args] = func(*args)
             return cache[args]
 
     return f
 
 
 class cow(object):
     """helper class to make copy-on-write easier
 
     Call preparewrite before doing any writes.
     """
 
     def preparewrite(self):
         """call this before writes, return self or a copied new object"""
         if getattr(self, '_copied', 0):
             self._copied -= 1
             # Function cow.__init__ expects 1 arg(s), got 2 [wrong-arg-count]
             return self.__class__(self)  # pytype: disable=wrong-arg-count
         return self
 
     def copy(self):
         """always do a cheap copy"""
         self._copied = getattr(self, '_copied', 0) + 1
         return self
 
 
 class sortdict(collections.OrderedDict):
     """a simple sorted dictionary
 
     >>> d1 = sortdict([(b'a', 0), (b'b', 1)])
     >>> d2 = d1.copy()
     >>> d2
     sortdict([('a', 0), ('b', 1)])
     >>> d2.update([(b'a', 2)])
     >>> list(d2.keys()) # should still be in last-set order
     ['b', 'a']
     >>> d1.insert(1, b'a.5', 0.5)
     >>> d1
     sortdict([('a', 0), ('a.5', 0.5), ('b', 1)])
     """
 
     def __setitem__(self, key, value):
         if key in self:
             del self[key]
         super(sortdict, self).__setitem__(key, value)
 
     if pycompat.ispypy:
         # __setitem__() isn't called as of PyPy 5.8.0
         def update(self, src, **f):
             if isinstance(src, dict):
                 src = pycompat.iteritems(src)
             for k, v in src:
                 self[k] = v
             for k in f:
                 self[k] = f[k]
 
     def insert(self, position, key, value):
         for (i, (k, v)) in enumerate(list(self.items())):
             if i == position:
                 self[key] = value
             if i >= position:
                 del self[k]
                 self[k] = v
 
 
 class cowdict(cow, dict):
     """copy-on-write dict
 
     Be sure to call d = d.preparewrite() before writing to d.
 
     >>> a = cowdict()
     >>> a is a.preparewrite()
     True
     >>> b = a.copy()
     >>> b is a
     True
     >>> c = b.copy()
     >>> c is a
     True
     >>> a = a.preparewrite()
     >>> b is a
     False
     >>> a is a.preparewrite()
     True
     >>> c = c.preparewrite()
     >>> b is c
     False
     >>> b is b.preparewrite()
     True
     """
 
 
 class cowsortdict(cow, sortdict):
     """copy-on-write sortdict
 
     Be sure to call d = d.preparewrite() before writing to d.
     """
 
 
 class transactional(object):  # pytype: disable=ignored-metaclass
     """Base class for making a transactional type into a context manager."""
 
     __metaclass__ = abc.ABCMeta
 
     @abc.abstractmethod
     def close(self):
         """Successfully closes the transaction."""
 
     @abc.abstractmethod
     def release(self):
         """Marks the end of the transaction.
 
         If the transaction has not been closed, it will be aborted.
         """
 
     def __enter__(self):
         return self
 
     def __exit__(self, exc_type, exc_val, exc_tb):
         try:
             if exc_type is None:
                 self.close()
         finally:
             self.release()
 
 
 @contextlib.contextmanager
 def acceptintervention(tr=None):
     """A context manager that closes the transaction on InterventionRequired
 
     If no transaction was provided, this simply runs the body and returns
     """
     if not tr:
         yield
         return
     try:
         yield
         tr.close()
     except error.InterventionRequired:
         tr.close()
         raise
     finally:
         tr.release()
 
 
 @contextlib.contextmanager
 def nullcontextmanager(enter_result=None):
     yield enter_result
 
 
 class _lrucachenode(object):
     """A node in a doubly linked list.
 
     Holds a reference to nodes on either side as well as a key-value
     pair for the dictionary entry.
     """
 
     __slots__ = ('next', 'prev', 'key', 'value', 'cost')
 
     def __init__(self):
         self.next = self
         self.prev = self
 
         self.key = _notset
         self.value = None
         self.cost = 0
 
     def markempty(self):
         """Mark the node as emptied."""
         self.key = _notset
         self.value = None
         self.cost = 0
 
 
 class lrucachedict(object):
     """Dict that caches most recent accesses and sets.
 
     The dict consists of an actual backing dict - indexed by original
     key - and a doubly linked circular list defining the order of entries in
     the cache.
 
     The head node is the newest entry in the cache. If the cache is full,
     we recycle head.prev and make it the new head. Cache accesses result in
     the node being moved to before the existing head and being marked as the
     new head node.
 
     Items in the cache can be inserted with an optional "cost" value. This is
     simply an integer that is specified by the caller. The cache can be queried
     for the total cost of all items presently in the cache.
 
     The cache can also define a maximum cost. If a cache insertion would
     cause the total cost of the cache to go beyond the maximum cost limit,
     nodes will be evicted to make room for the new code. This can be used
     to e.g. set a max memory limit and associate an estimated bytes size
     cost to each item in the cache. By default, no maximum cost is enforced.
     """
 
     def __init__(self, max, maxcost=0):
         self._cache = {}
 
         self._head = _lrucachenode()
         self._size = 1
         self.capacity = max
         self.totalcost = 0
         self.maxcost = maxcost
 
     def __len__(self):
         return len(self._cache)
 
     def __contains__(self, k):
         return k in self._cache
 
     def __iter__(self):
         # We don't have to iterate in cache order, but why not.
         n = self._head
         for i in range(len(self._cache)):
             yield n.key
             n = n.next
 
     def __getitem__(self, k):
         node = self._cache[k]
         self._movetohead(node)
         return node.value
 
     def insert(self, k, v, cost=0):
         """Insert a new item in the cache with optional cost value."""
         node = self._cache.get(k)
         # Replace existing value and mark as newest.
         if node is not None:
             self.totalcost -= node.cost
             node.value = v
             node.cost = cost
             self.totalcost += cost
             self._movetohead(node)
 
             if self.maxcost:
                 self._enforcecostlimit()
 
             return
 
         if self._size < self.capacity:
             node = self._addcapacity()
         else:
             # Grab the last/oldest item.
             node = self._head.prev
 
         # At capacity. Kill the old entry.
         if node.key is not _notset:
             self.totalcost -= node.cost
             del self._cache[node.key]
 
         node.key = k
         node.value = v
         node.cost = cost
         self.totalcost += cost
         self._cache[k] = node
         # And mark it as newest entry. No need to adjust order since it
         # is already self._head.prev.
         self._head = node
 
         if self.maxcost:
             self._enforcecostlimit()
 
     def __setitem__(self, k, v):
         self.insert(k, v)
 
     def __delitem__(self, k):
         self.pop(k)
 
     def pop(self, k, default=_notset):
         try:
             node = self._cache.pop(k)
         except KeyError:
             if default is _notset:
                 raise
             return default
 
         assert node is not None  # help pytype
         value = node.value
         self.totalcost -= node.cost
         node.markempty()
 
         # Temporarily mark as newest item before re-adjusting head to make
         # this node the oldest item.
         self._movetohead(node)
         self._head = node.next
 
         return value
 
     # Additional dict methods.
 
     def get(self, k, default=None):
         try:
             return self.__getitem__(k)
         except KeyError:
             return default
 
     def peek(self, k, default=_notset):
         """Get the specified item without moving it to the head
 
         Unlike get(), this doesn't mutate the internal state. But be aware
         that it doesn't mean peek() is thread safe.
         """
         try:
             node = self._cache[k]
             assert node is not None  # help pytype
             return node.value
         except KeyError:
             if default is _notset:
                 raise
             return default
 
     def clear(self):
         n = self._head
         while n.key is not _notset:
             self.totalcost -= n.cost
             n.markempty()
             n = n.next
 
         self._cache.clear()
 
     def copy(self, capacity=None, maxcost=0):
         """Create a new cache as a copy of the current one.
 
         By default, the new cache has the same capacity as the existing one.
         But, the cache capacity can be changed as part of performing the
         copy.
 
         Items in the copy have an insertion/access order matching this
         instance.
         """
 
         capacity = capacity or self.capacity
         maxcost = maxcost or self.maxcost
         result = lrucachedict(capacity, maxcost=maxcost)
 
         # We copy entries by iterating in oldest-to-newest order so the copy
         # has the correct ordering.
 
         # Find the first non-empty entry.
         n = self._head.prev
         while n.key is _notset and n is not self._head:
             n = n.prev
 
         # We could potentially skip the first N items when decreasing capacity.
         # But let's keep it simple unless it is a performance problem.
         for i in range(len(self._cache)):
             result.insert(n.key, n.value, cost=n.cost)
             n = n.prev
 
         return result
 
     def popoldest(self):
         """Remove the oldest item from the cache.
 
         Returns the (key, value) describing the removed cache entry.
         """
         if not self._cache:
             return
 
         # Walk the linked list backwards starting at tail node until we hit
         # a non-empty node.
         n = self._head.prev
 
         assert n is not None  # help pytype
 
         while n.key is _notset:
             n = n.prev
 
         assert n is not None  # help pytype
 
         key, value = n.key, n.value
 
         # And remove it from the cache and mark it as empty.
         del self._cache[n.key]
         self.totalcost -= n.cost
         n.markempty()
 
         return key, value
 
     def _movetohead(self, node):
         """Mark a node as the newest, making it the new head.
 
         When a node is accessed, it becomes the freshest entry in the LRU
         list, which is denoted by self._head.
 
         Visually, let's make ``N`` the new head node (* denotes head):
 
             previous/oldest <-> head <-> next/next newest
 
             ----<->--- A* ---<->-----
             |                       |
             E <-> D <-> N <-> C <-> B
 
         To:
 
             ----<->--- N* ---<->-----
             |                       |
             E <-> D <-> C <-> B <-> A
 
         This requires the following moves:
 
            C.next = D  (node.prev.next = node.next)
            D.prev = C  (node.next.prev = node.prev)
            E.next = N  (head.prev.next = node)
            N.prev = E  (node.prev = head.prev)
            N.next = A  (node.next = head)
            A.prev = N  (head.prev = node)
         """
         head = self._head
         # C.next = D
         node.prev.next = node.next
         # D.prev = C
         node.next.prev = node.prev
         # N.prev = E
         node.prev = head.prev
         # N.next = A
         # It is tempting to do just "head" here, however if node is
         # adjacent to head, this will do bad things.
         node.next = head.prev.next
         # E.next = N
         node.next.prev = node
         # A.prev = N
         node.prev.next = node
 
         self._head = node
 
     def _addcapacity(self):
         """Add a node to the circular linked list.
 
         The new node is inserted before the head node.
         """
         head = self._head
         node = _lrucachenode()
         head.prev.next = node
         node.prev = head.prev
         node.next = head
         head.prev = node
         self._size += 1
         return node
 
     def _enforcecostlimit(self):
         # This should run after an insertion. It should only be called if total
         # cost limits are being enforced.
         # The most recently inserted node is never evicted.
         if len(self) <= 1 or self.totalcost <= self.maxcost:
             return
 
         # This is logically equivalent to calling popoldest() until we
         # free up enough cost. We don't do that since popoldest() needs
         # to walk the linked list and doing this in a loop would be
         # quadratic. So we find the first non-empty node and then
         # walk nodes until we free up enough capacity.
         #
         # If we only removed the minimum number of nodes to free enough
         # cost at insert time, chances are high that the next insert would
         # also require pruning. This would effectively constitute quadratic
         # behavior for insert-heavy workloads. To mitigate this, we set a
         # target cost that is a percentage of the max cost. This will tend
         # to free more nodes when the high water mark is reached, which
         # lowers the chances of needing to prune on the subsequent insert.
         targetcost = int(self.maxcost * 0.75)
 
         n = self._head.prev
         while n.key is _notset:
             n = n.prev
 
         while len(self) > 1 and self.totalcost > targetcost:
             del self._cache[n.key]
             self.totalcost -= n.cost
             n.markempty()
             n = n.prev
 
 
 def lrucachefunc(func):
     '''cache most recent results of function calls'''
     cache = {}
     order = collections.deque()
     if func.__code__.co_argcount == 1:
 
         def f(arg):
             if arg not in cache:
                 if len(cache) > 20:
                     del cache[order.popleft()]
                 cache[arg] = func(arg)
             else:
                 order.remove(arg)
             order.append(arg)
             return cache[arg]
 
     else:
 
         def f(*args):
             if args not in cache:
                 if len(cache) > 20:
                     del cache[order.popleft()]
                 cache[args] = func(*args)
             else:
                 order.remove(args)
             order.append(args)
             return cache[args]
 
     return f
 
 
 class propertycache(object):
     def __init__(self, func):
         self.func = func
         self.name = func.__name__
 
     def __get__(self, obj, type=None):
         result = self.func(obj)
         self.cachevalue(obj, result)
         return result
 
     def cachevalue(self, obj, value):
         # __dict__ assignment required to bypass __setattr__ (eg: repoview)
         obj.__dict__[self.name] = value
 
 
 def clearcachedproperty(obj, prop):
     '''clear a cached property value, if one has been set'''
     prop = pycompat.sysstr(prop)
     if prop in obj.__dict__:
         del obj.__dict__[prop]
 
 
 def increasingchunks(source, min=1024, max=65536):
     """return no less than min bytes per chunk while data remains,
     doubling min after each chunk until it reaches max"""
 
     def log2(x):
         if not x:
             return 0
         i = 0
         while x:
             x >>= 1
             i += 1
         return i - 1
 
     buf = []
     blen = 0
     for chunk in source:
         buf.append(chunk)
         blen += len(chunk)
         if blen >= min:
             if min < max:
                 min = min << 1
                 nmin = 1 << log2(blen)
                 if nmin > min:
                     min = nmin
                 if min > max:
                     min = max
             yield b''.join(buf)
             blen = 0
             buf = []
     if buf:
         yield b''.join(buf)
 
 
 def always(fn):
     return True
 
 
 def never(fn):
     return False
 
 
 def nogc(func):
     """disable garbage collector
 
     Python's garbage collector triggers a GC each time a certain number of
     container objects (the number being defined by gc.get_threshold()) are
     allocated even when marked not to be tracked by the collector. Tracking has
     no effect on when GCs are triggered, only on what objects the GC looks
     into. As a workaround, disable GC while building complex (huge)
     containers.
 
     This garbage collector issue have been fixed in 2.7. But it still affect
     CPython's performance.
     """
 
     def wrapper(*args, **kwargs):
         gcenabled = gc.isenabled()
         gc.disable()
         try:
             return func(*args, **kwargs)
         finally:
             if gcenabled:
                 gc.enable()
 
     return wrapper
 
 
 if pycompat.ispypy:
     # PyPy runs slower with gc disabled
     nogc = lambda x: x
 
 
 def pathto(root, n1, n2):
     # type: (bytes, bytes, bytes) -> bytes
     """return the relative path from one place to another.
     root should use os.sep to separate directories
     n1 should use os.sep to separate directories
     n2 should use "/" to separate directories
     returns an os.sep-separated path.
 
     If n1 is a relative path, it's assumed it's
     relative to root.
     n2 should always be relative to root.
     """
     if not n1:
         return localpath(n2)
     if os.path.isabs(n1):
         if os.path.splitdrive(root)[0] != os.path.splitdrive(n1)[0]:
             return os.path.join(root, localpath(n2))
         n2 = b'/'.join((pconvert(root), n2))
     a, b = splitpath(n1), n2.split(b'/')
     a.reverse()
     b.reverse()
     while a and b and a[-1] == b[-1]:
         a.pop()
         b.pop()
     b.reverse()
     return pycompat.ossep.join(([b'..'] * len(a)) + b) or b'.'
 
 
 def checksignature(func, depth=1):
     '''wrap a function with code to check for calling errors'''
 
     def check(*args, **kwargs):
         try:
             return func(*args, **kwargs)
         except TypeError:
             if len(traceback.extract_tb(sys.exc_info()[2])) == depth:
                 raise error.SignatureError
             raise
 
     return check
 
 
 # a whilelist of known filesystems where hardlink works reliably
 _hardlinkfswhitelist = {
     b'apfs',
     b'btrfs',
     b'ext2',
     b'ext3',
     b'ext4',
     b'hfs',
     b'jfs',
     b'NTFS',
     b'reiserfs',
     b'tmpfs',
     b'ufs',
     b'xfs',
     b'zfs',
 }
 
 
 def copyfile(
     src,
     dest,
     hardlink=False,
     copystat=False,
     checkambig=False,
     nb_bytes=None,
     no_hardlink_cb=None,
     check_fs_hardlink=True,
 ):
     """copy a file, preserving mode and optionally other stat info like
     atime/mtime
 
     checkambig argument is used with filestat, and is useful only if
     destination file is guarded by any lock (e.g. repo.lock or
     repo.wlock).
 
     copystat and checkambig should be exclusive.
 
     nb_bytes: if set only copy the first `nb_bytes` of the source file.
     """
     assert not (copystat and checkambig)
     oldstat = None
     if os.path.lexists(dest):
         if checkambig:
             oldstat = checkambig and filestat.frompath(dest)
         unlink(dest)
     if hardlink and check_fs_hardlink:
         # Hardlinks are problematic on CIFS (issue4546), do not allow hardlinks
         # unless we are confident that dest is on a whitelisted filesystem.
         try:
             fstype = getfstype(os.path.dirname(dest))
         except OSError:
             fstype = None
         if fstype not in _hardlinkfswhitelist:
             if no_hardlink_cb is not None:
                 no_hardlink_cb()
             hardlink = False
     if hardlink:
         try:
             oslink(src, dest)
             if nb_bytes is not None:
                 m = "the `nb_bytes` argument is incompatible with `hardlink`"
                 raise error.ProgrammingError(m)
             return
         except (IOError, OSError) as exc:
             if exc.errno != errno.EEXIST and no_hardlink_cb is not None:
                 no_hardlink_cb()
             # fall back to normal copy
     if os.path.islink(src):
         os.symlink(os.readlink(src), dest)
         # copytime is ignored for symlinks, but in general copytime isn't needed
         # for them anyway
         if nb_bytes is not None:
             m = "cannot use `nb_bytes` on a symlink"
             raise error.ProgrammingError(m)
     else:
         try:
             shutil.copyfile(src, dest)
             if copystat:
                 # copystat also copies mode
                 shutil.copystat(src, dest)
             else:
                 shutil.copymode(src, dest)
                 if oldstat and oldstat.stat:
                     newstat = filestat.frompath(dest)
                     if newstat.isambig(oldstat):
                         # stat of copied file is ambiguous to original one
                         advanced = (
                             oldstat.stat[stat.ST_MTIME] + 1
                         ) & 0x7FFFFFFF
                         os.utime(dest, (advanced, advanced))
             # We could do something smarter using `copy_file_range` call or similar
             if nb_bytes is not None:
                 with open(dest, mode='r+') as f:
                     f.truncate(nb_bytes)
         except shutil.Error as inst:
             raise error.Abort(stringutil.forcebytestr(inst))
 
 
 def copyfiles(src, dst, hardlink=None, progress=None):
     """Copy a directory tree using hardlinks if possible."""
     num = 0
 
     def settopic():
         if progress:
             progress.topic = _(b'linking') if hardlink else _(b'copying')
 
     if os.path.isdir(src):
         if hardlink is None:
             hardlink = (
                 os.stat(src).st_dev == os.stat(os.path.dirname(dst)).st_dev
             )
         settopic()
         os.mkdir(dst)
         for name, kind in listdir(src):
             srcname = os.path.join(src, name)
             dstname = os.path.join(dst, name)
             hardlink, n = copyfiles(srcname, dstname, hardlink, progress)
             num += n
     else:
         if hardlink is None:
             hardlink = (
                 os.stat(os.path.dirname(src)).st_dev
                 == os.stat(os.path.dirname(dst)).st_dev
             )
         settopic()
 
         if hardlink:
             try:
                 oslink(src, dst)
             except (IOError, OSError) as exc:
                 if exc.errno != errno.EEXIST:
                     hardlink = False
                 # XXX maybe try to relink if the file exist ?
                 shutil.copy(src, dst)
         else:
             shutil.copy(src, dst)
         num += 1
         if progress:
             progress.increment()
 
     return hardlink, num
 
 
 _winreservednames = {
     b'con',
     b'prn',
     b'aux',
     b'nul',
     b'com1',
     b'com2',
     b'com3',
     b'com4',
     b'com5',
     b'com6',
     b'com7',
     b'com8',
     b'com9',
     b'lpt1',
     b'lpt2',
     b'lpt3',
     b'lpt4',
     b'lpt5',
     b'lpt6',
     b'lpt7',
     b'lpt8',
     b'lpt9',
 }
 _winreservedchars = b':*?"<>|'
 
 
 def checkwinfilename(path):
     # type: (bytes) -> Optional[bytes]
     r"""Check that the base-relative path is a valid filename on Windows.
     Returns None if the path is ok, or a UI string describing the problem.
 
     >>> checkwinfilename(b"just/a/normal/path")
     >>> checkwinfilename(b"foo/bar/con.xml")
     "filename contains 'con', which is reserved on Windows"
     >>> checkwinfilename(b"foo/con.xml/bar")
     "filename contains 'con', which is reserved on Windows"
     >>> checkwinfilename(b"foo/bar/xml.con")
     >>> checkwinfilename(b"foo/bar/AUX/bla.txt")
     "filename contains 'AUX', which is reserved on Windows"
     >>> checkwinfilename(b"foo/bar/bla:.txt")
     "filename contains ':', which is reserved on Windows"
     >>> checkwinfilename(b"foo/bar/b\07la.txt")
     "filename contains '\\x07', which is invalid on Windows"
     >>> checkwinfilename(b"foo/bar/bla ")
     "filename ends with ' ', which is not allowed on Windows"
     >>> checkwinfilename(b"../bar")
     >>> checkwinfilename(b"foo\\")
     "filename ends with '\\', which is invalid on Windows"
     >>> checkwinfilename(b"foo\\/bar")
     "directory name ends with '\\', which is invalid on Windows"
     """
     if path.endswith(b'\\'):
         return _(b"filename ends with '\\', which is invalid on Windows")
     if b'\\/' in path:
         return _(b"directory name ends with '\\', which is invalid on Windows")
     for n in path.replace(b'\\', b'/').split(b'/'):
         if not n:
             continue
         for c in _filenamebytestr(n):
             if c in _winreservedchars:
                 return (
                     _(
                         b"filename contains '%s', which is reserved "
                         b"on Windows"
                     )
                     % c
                 )
             if ord(c) <= 31:
                 return _(
                     b"filename contains '%s', which is invalid on Windows"
                 ) % stringutil.escapestr(c)
         base = n.split(b'.')[0]
         if base and base.lower() in _winreservednames:
             return (
                 _(b"filename contains '%s', which is reserved on Windows")
                 % base
             )
         t = n[-1:]
         if t in b'. ' and n not in b'..':
             return (
                 _(
                     b"filename ends with '%s', which is not allowed "
                     b"on Windows"
                 )
                 % t
             )
 
 
 timer = getattr(time, "perf_counter", None)
 
 if pycompat.iswindows:
     checkosfilename = checkwinfilename
     if not timer:
         timer = time.clock
 else:
     # mercurial.windows doesn't have platform.checkosfilename
     checkosfilename = platform.checkosfilename  # pytype: disable=module-attr
     if not timer:
         timer = time.time
 
 
 def makelock(info, pathname):
     """Create a lock file atomically if possible
 
     This may leave a stale lock file if symlink isn't supported and signal
     interrupt is enabled.
     """
     try:
         return os.symlink(info, pathname)
     except OSError as why:
         if why.errno == errno.EEXIST:
             raise
     except AttributeError:  # no symlink in os
         pass
 
     flags = os.O_CREAT | os.O_WRONLY | os.O_EXCL | getattr(os, 'O_BINARY', 0)
     ld = os.open(pathname, flags)
     os.write(ld, info)
     os.close(ld)
 
 
 def readlock(pathname):
     # type: (bytes) -> bytes
     try:
         return readlink(pathname)
     except OSError as why:
         if why.errno not in (errno.EINVAL, errno.ENOSYS):
             raise
     except AttributeError:  # no symlink in os
         pass
     with posixfile(pathname, b'rb') as fp:
         return fp.read()
 
 
 def fstat(fp):
     '''stat file object that may not have fileno method.'''
     try:
         return os.fstat(fp.fileno())
     except AttributeError:
         return os.stat(fp.name)
 
 
 # File system features
 
 
 def fscasesensitive(path):
     # type: (bytes) -> bool
     """
     Return true if the given path is on a case-sensitive filesystem
 
     Requires a path (like /foo/.hg) ending with a foldable final
     directory component.
     """
     s1 = os.lstat(path)
     d, b = os.path.split(path)
     b2 = b.upper()
     if b == b2:
         b2 = b.lower()
         if b == b2:
             return True  # no evidence against case sensitivity
     p2 = os.path.join(d, b2)
     try:
         s2 = os.lstat(p2)
         if s2 == s1:
             return False
         return True
     except OSError:
         return True
 
 
 _re2_input = lambda x: x
 try:
     import re2  # pytype: disable=import-error
 
     _re2 = None
 except ImportError:
     _re2 = False
 
 
 class _re(object):
     def _checkre2(self):
         global _re2
         global _re2_input
 
         check_pattern = br'\[([^\[]+)\]'
         check_input = b'[ui]'
         try:
             # check if match works, see issue3964
             _re2 = bool(re2.match(check_pattern, check_input))
         except ImportError:
             _re2 = False
         except TypeError:
             # the `pyre-2` project provides a re2 module that accept bytes
             # the `fb-re2` project provides a re2 module that acccept sysstr
             check_pattern = pycompat.sysstr(check_pattern)
             check_input = pycompat.sysstr(check_input)
             _re2 = bool(re2.match(check_pattern, check_input))
             _re2_input = pycompat.sysstr
 
     def compile(self, pat, flags=0):
         """Compile a regular expression, using re2 if possible
 
         For best performance, use only re2-compatible regexp features. The
         only flags from the re module that are re2-compatible are
         IGNORECASE and MULTILINE."""
         if _re2 is None:
             self._checkre2()
         if _re2 and (flags & ~(remod.IGNORECASE | remod.MULTILINE)) == 0:
             if flags & remod.IGNORECASE:
                 pat = b'(?i)' + pat
             if flags & remod.MULTILINE:
                 pat = b'(?m)' + pat
             try:
                 return re2.compile(_re2_input(pat))
             except re2.error:
                 pass
         return remod.compile(pat, flags)
 
     @propertycache
     def escape(self):
         """Return the version of escape corresponding to self.compile.
 
         This is imperfect because whether re2 or re is used for a particular
         function depends on the flags, etc, but it's the best we can do.
         """
         global _re2
         if _re2 is None:
             self._checkre2()
         if _re2:
             return re2.escape
         else:
             return remod.escape
 
 
 re = _re()
 
 _fspathcache = {}
 
 
 def fspath(name, root):
     # type: (bytes, bytes) -> bytes
     """Get name in the case stored in the filesystem
 
     The name should be relative to root, and be normcase-ed for efficiency.
 
     Note that this function is unnecessary, and should not be
     called, for case-sensitive filesystems (simply because it's expensive).
 
     The root should be normcase-ed, too.
     """
 
     def _makefspathcacheentry(dir):
         return {normcase(n): n for n in os.listdir(dir)}
 
     seps = pycompat.ossep
     if pycompat.osaltsep:
         seps = seps + pycompat.osaltsep
     # Protect backslashes. This gets silly very quickly.
     seps.replace(b'\\', b'\\\\')
     pattern = remod.compile(br'([^%s]+)|([%s]+)' % (seps, seps))
     dir = os.path.normpath(root)
     result = []
     for part, sep in pattern.findall(name):
         if sep:
             result.append(sep)
             continue
 
         if dir not in _fspathcache:
             _fspathcache[dir] = _makefspathcacheentry(dir)
         contents = _fspathcache[dir]
 
         found = contents.get(part)
         if not found:
             # retry "once per directory" per "dirstate.walk" which
             # may take place for each patches of "hg qpush", for example
             _fspathcache[dir] = contents = _makefspathcacheentry(dir)
             found = contents.get(part)
 
         result.append(found or part)
         dir = os.path.join(dir, part)
 
     return b''.join(result)
 
 
 def checknlink(testfile):
     # type: (bytes) -> bool
     '''check whether hardlink count reporting works properly'''
 
     # testfile may be open, so we need a separate file for checking to
     # work around issue2543 (or testfile may get lost on Samba shares)
     f1, f2, fp = None, None, None
     try:
         fd, f1 = pycompat.mkstemp(
             prefix=b'.%s-' % os.path.basename(testfile),
             suffix=b'1~',
             dir=os.path.dirname(testfile),
         )
         os.close(fd)
         f2 = b'%s2~' % f1[:-2]
 
         oslink(f1, f2)
         # nlinks() may behave differently for files on Windows shares if
         # the file is open.
         fp = posixfile(f2)
         return nlinks(f2) > 1
     except OSError:
         return False
     finally:
         if fp is not None:
             fp.close()
         for f in (f1, f2):
             try:
                 if f is not None:
                     os.unlink(f)
             except OSError:
                 pass
 
 
 def endswithsep(path):
     # type: (bytes) -> bool
     '''Check path ends with os.sep or os.altsep.'''
     return bool(  # help pytype
         path.endswith(pycompat.ossep)
         or pycompat.osaltsep
         and path.endswith(pycompat.osaltsep)
     )
 
 
 def splitpath(path):
     # type: (bytes) -> List[bytes]
     """Split path by os.sep.
     Note that this function does not use os.altsep because this is
     an alternative of simple "xxx.split(os.sep)".
     It is recommended to use os.path.normpath() before using this
     function if need."""
     return path.split(pycompat.ossep)
 
 
 def mktempcopy(name, emptyok=False, createmode=None, enforcewritable=False):
     """Create a temporary file with the same contents from name
 
     The permission bits are copied from the original file.
 
     If the temporary file is going to be truncated immediately, you
     can use emptyok=True as an optimization.
 
     Returns the name of the temporary file.
     """
     d, fn = os.path.split(name)
     fd, temp = pycompat.mkstemp(prefix=b'.%s-' % fn, suffix=b'~', dir=d)
     os.close(fd)
     # Temporary files are created with mode 0600, which is usually not
     # what we want.  If the original file already exists, just copy
     # its mode.  Otherwise, manually obey umask.
     copymode(name, temp, createmode, enforcewritable)
 
     if emptyok:
         return temp
     try:
         try:
             ifp = posixfile(name, b"rb")
         except IOError as inst:
             if inst.errno == errno.ENOENT:
                 return temp
             if not getattr(inst, 'filename', None):
                 inst.filename = name
             raise
         ofp = posixfile(temp, b"wb")
         for chunk in filechunkiter(ifp):
             ofp.write(chunk)
         ifp.close()
         ofp.close()
     except:  # re-raises
         try:
             os.unlink(temp)
         except OSError:
             pass
         raise
     return temp
 
 
 class filestat(object):
     """help to exactly detect change of a file
 
     'stat' attribute is result of 'os.stat()' if specified 'path'
     exists. Otherwise, it is None. This can avoid preparative
     'exists()' examination on client side of this class.
     """
 
     def __init__(self, stat):
         self.stat = stat
 
     @classmethod
     def frompath(cls, path):
         try:
             stat = os.stat(path)
         except OSError as err:
             if err.errno != errno.ENOENT:
                 raise
             stat = None
         return cls(stat)
 
     @classmethod
     def fromfp(cls, fp):
         stat = os.fstat(fp.fileno())
         return cls(stat)
 
     __hash__ = object.__hash__
 
     def __eq__(self, old):
         try:
             # if ambiguity between stat of new and old file is
             # avoided, comparison of size, ctime and mtime is enough
             # to exactly detect change of a file regardless of platform
             return (
                 self.stat.st_size == old.stat.st_size
                 and self.stat[stat.ST_CTIME] == old.stat[stat.ST_CTIME]
                 and self.stat[stat.ST_MTIME] == old.stat[stat.ST_MTIME]
             )
         except AttributeError:
             pass
         try:
             return self.stat is None and old.stat is None
         except AttributeError:
             return False
 
     def isambig(self, old):
         """Examine whether new (= self) stat is ambiguous against old one
 
         "S[N]" below means stat of a file at N-th change:
 
         - S[n-1].ctime  < S[n].ctime: can detect change of a file
         - S[n-1].ctime == S[n].ctime
           - S[n-1].ctime  < S[n].mtime: means natural advancing (*1)
           - S[n-1].ctime == S[n].mtime: is ambiguous (*2)
           - S[n-1].ctime  > S[n].mtime: never occurs naturally (don't care)
         - S[n-1].ctime  > S[n].ctime: never occurs naturally (don't care)
 
         Case (*2) above means that a file was changed twice or more at
         same time in sec (= S[n-1].ctime), and comparison of timestamp
         is ambiguous.
 
         Base idea to avoid such ambiguity is "advance mtime 1 sec, if
         timestamp is ambiguous".
 
         But advancing mtime only in case (*2) doesn't work as
         expected, because naturally advanced S[n].mtime in case (*1)
         might be equal to manually advanced S[n-1 or earlier].mtime.
 
         Therefore, all "S[n-1].ctime == S[n].ctime" cases should be
         treated as ambiguous regardless of mtime, to avoid overlooking
         by confliction between such mtime.
 
         Advancing mtime "if isambig(oldstat)" ensures "S[n-1].mtime !=
         S[n].mtime", even if size of a file isn't changed.
         """
         try:
             return self.stat[stat.ST_CTIME] == old.stat[stat.ST_CTIME]
         except AttributeError:
             return False
 
     def avoidambig(self, path, old):
         """Change file stat of specified path to avoid ambiguity
 
         'old' should be previous filestat of 'path'.
 
         This skips avoiding ambiguity, if a process doesn't have
         appropriate privileges for 'path'. This returns False in this
         case.
 
         Otherwise, this returns True, as "ambiguity is avoided".
         """
         advanced = (old.stat[stat.ST_MTIME] + 1) & 0x7FFFFFFF
         try:
             os.utime(path, (advanced, advanced))
         except OSError as inst:
             if inst.errno == errno.EPERM:
                 # utime() on the file created by another user causes EPERM,
                 # if a process doesn't have appropriate privileges
                 return False
             raise
         return True
 
     def __ne__(self, other):
         return not self == other
 
 
 class atomictempfile(object):
     """writable file object that atomically updates a file
 
     All writes will go to a temporary copy of the original file. Call
     close() when you are done writing, and atomictempfile will rename
     the temporary copy to the original name, making the changes
     visible. If the object is destroyed without being closed, all your
     writes are discarded.
 
     checkambig argument of constructor is used with filestat, and is
     useful only if target file is guarded by any lock (e.g. repo.lock
     or repo.wlock).
     """
 
     def __init__(self, name, mode=b'w+b', createmode=None, checkambig=False):
         self.__name = name  # permanent name
         self._tempname = mktempcopy(
             name,
             emptyok=(b'w' in mode),
             createmode=createmode,
             enforcewritable=(b'w' in mode),
         )
 
         self._fp = posixfile(self._tempname, mode)
         self._checkambig = checkambig
 
         # delegated methods
         self.read = self._fp.read
         self.write = self._fp.write
         self.seek = self._fp.seek
         self.tell = self._fp.tell
         self.fileno = self._fp.fileno
 
     def close(self):
         if not self._fp.closed:
             self._fp.close()
             filename = localpath(self.__name)
             oldstat = self._checkambig and filestat.frompath(filename)
             if oldstat and oldstat.stat:
                 rename(self._tempname, filename)
                 newstat = filestat.frompath(filename)
                 if newstat.isambig(oldstat):
                     # stat of changed file is ambiguous to original one
                     advanced = (oldstat.stat[stat.ST_MTIME] + 1) & 0x7FFFFFFF
                     os.utime(filename, (advanced, advanced))
             else:
                 rename(self._tempname, filename)
 
     def discard(self):
         if not self._fp.closed:
             try:
                 os.unlink(self._tempname)
             except OSError:
                 pass
             self._fp.close()
 
     def __del__(self):
         if safehasattr(self, '_fp'):  # constructor actually did something
             self.discard()
 
     def __enter__(self):
         return self
 
     def __exit__(self, exctype, excvalue, traceback):
         if exctype is not None:
             self.discard()
         else:
             self.close()
 
 
 def unlinkpath(f, ignoremissing=False, rmdir=True):
     # type: (bytes, bool, bool) -> None
     """unlink and remove the directory if it is empty"""
     if ignoremissing:
         tryunlink(f)
     else:
         unlink(f)
     if rmdir:
         # try removing directories that might now be empty
         try:
             removedirs(os.path.dirname(f))
         except OSError:
             pass
 
 
 def tryunlink(f):
     # type: (bytes) -> None
     """Attempt to remove a file, ignoring ENOENT errors."""
     try:
         unlink(f)
     except OSError as e:
         if e.errno != errno.ENOENT:
             raise
 
 
 def makedirs(name, mode=None, notindexed=False):
     # type: (bytes, Optional[int], bool) -> None
     """recursive directory creation with parent mode inheritance
 
     Newly created directories are marked as "not to be indexed by
     the content indexing service", if ``notindexed`` is specified
     for "write" mode access.
     """
     try:
         makedir(name, notindexed)
     except OSError as err:
         if err.errno == errno.EEXIST:
             return
         if err.errno != errno.ENOENT or not name:
             raise
         parent = os.path.dirname(abspath(name))
         if parent == name:
             raise
         makedirs(parent, mode, notindexed)
         try:
             makedir(name, notindexed)
         except OSError as err:
             # Catch EEXIST to handle races
             if err.errno == errno.EEXIST:
                 return
             raise
     if mode is not None:
         os.chmod(name, mode)
 
 
 def readfile(path):
     # type: (bytes) -> bytes
     with open(path, b'rb') as fp:
         return fp.read()
 
 
 def writefile(path, text):
     # type: (bytes, bytes) -> None
     with open(path, b'wb') as fp:
         fp.write(text)
 
 
 def appendfile(path, text):
     # type: (bytes, bytes) -> None
     with open(path, b'ab') as fp:
         fp.write(text)
 
 
 class chunkbuffer(object):
     """Allow arbitrary sized chunks of data to be efficiently read from an
     iterator over chunks of arbitrary size."""
 
     def __init__(self, in_iter):
         """in_iter is the iterator that's iterating over the input chunks."""
 
         def splitbig(chunks):
             for chunk in chunks:
                 if len(chunk) > 2 ** 20:
                     pos = 0
                     while pos < len(chunk):
                         end = pos + 2 ** 18
                         yield chunk[pos:end]
                         pos = end
                 else:
                     yield chunk
 
         self.iter = splitbig(in_iter)
         self._queue = collections.deque()
         self._chunkoffset = 0
 
     def read(self, l=None):
         """Read L bytes of data from the iterator of chunks of data.
         Returns less than L bytes if the iterator runs dry.
 
         If size parameter is omitted, read everything"""
         if l is None:
             return b''.join(self.iter)
 
         left = l
         buf = []
         queue = self._queue
         while left > 0:
             # refill the queue
             if not queue:
                 target = 2 ** 18
                 for chunk in self.iter:
                     queue.append(chunk)
                     target -= len(chunk)
                     if target <= 0:
                         break
                 if not queue:
                     break
 
             # The easy way to do this would be to queue.popleft(), modify the
             # chunk (if necessary), then queue.appendleft(). However, for cases
             # where we read partial chunk content, this incurs 2 dequeue
             # mutations and creates a new str for the remaining chunk in the
             # queue. Our code below avoids this overhead.
 
             chunk = queue[0]
             chunkl = len(chunk)
             offset = self._chunkoffset
 
             # Use full chunk.
             if offset == 0 and left >= chunkl:
                 left -= chunkl
                 queue.popleft()
                 buf.append(chunk)
                 # self._chunkoffset remains at 0.
                 continue
 
             chunkremaining = chunkl - offset
 
             # Use all of unconsumed part of chunk.
             if left >= chunkremaining:
                 left -= chunkremaining
                 queue.popleft()
                 # offset == 0 is enabled by block above, so this won't merely
                 # copy via ``chunk[0:]``.
                 buf.append(chunk[offset:])
                 self._chunkoffset = 0
 
             # Partial chunk needed.
             else:
                 buf.append(chunk[offset : offset + left])
                 self._chunkoffset += left
                 left -= chunkremaining
 
         return b''.join(buf)
 
 
 def filechunkiter(f, size=131072, limit=None):
     """Create a generator that produces the data in the file size
     (default 131072) bytes at a time, up to optional limit (default is
     to read all data).  Chunks may be less than size bytes if the
     chunk is the last chunk in the file, or the file is a socket or
     some other type of file that sometimes reads less data than is
     requested."""
     assert size >= 0
     assert limit is None or limit >= 0
     while True:
         if limit is None:
             nbytes = size
         else:
             nbytes = min(limit, size)
         s = nbytes and f.read(nbytes)
         if not s:
             break
         if limit:
             limit -= len(s)
         yield s
 
 
 class cappedreader(object):
     """A file object proxy that allows reading up to N bytes.
 
     Given a source file object, instances of this type allow reading up to
     N bytes from that source file object. Attempts to read past the allowed
     limit are treated as EOF.
 
     It is assumed that I/O is not performed on the original file object
     in addition to I/O that is performed by this instance. If there is,
     state tracking will get out of sync and unexpected results will ensue.
     """
 
     def __init__(self, fh, limit):
         """Allow reading up to <limit> bytes from <fh>."""
         self._fh = fh
         self._left = limit
 
     def read(self, n=-1):
         if not self._left:
             return b''
 
         if n < 0:
             n = self._left
 
         data = self._fh.read(min(n, self._left))
         self._left -= len(data)
         assert self._left >= 0
 
         return data
 
     def readinto(self, b):
         res = self.read(len(b))
         if res is None:
             return None
 
         b[0 : len(res)] = res
         return len(res)
 
 
 def unitcountfn(*unittable):
     '''return a function that renders a readable count of some quantity'''
 
     def go(count):
         for multiplier, divisor, format in unittable:
             if abs(count) >= divisor * multiplier:
                 return format % (count / float(divisor))
         return unittable[-1][2] % count
 
     return go
 
 
 def processlinerange(fromline, toline):
     # type: (int, int) -> Tuple[int, int]
     """Check that linerange <fromline>:<toline> makes sense and return a
     0-based range.
 
     >>> processlinerange(10, 20)
     (9, 20)
     >>> processlinerange(2, 1)
     Traceback (most recent call last):
         ...
     ParseError: line range must be positive
     >>> processlinerange(0, 5)
     Traceback (most recent call last):
         ...
     ParseError: fromline must be strictly positive
     """
     if toline - fromline < 0:
         raise error.ParseError(_(b"line range must be positive"))
     if fromline < 1:
         raise error.ParseError(_(b"fromline must be strictly positive"))
     return fromline - 1, toline
 
 
 bytecount = unitcountfn(
     (100, 1 << 30, _(b'%.0f GB')),
     (10, 1 << 30, _(b'%.1f GB')),
     (1, 1 << 30, _(b'%.2f GB')),
     (100, 1 << 20, _(b'%.0f MB')),
     (10, 1 << 20, _(b'%.1f MB')),
     (1, 1 << 20, _(b'%.2f MB')),
     (100, 1 << 10, _(b'%.0f KB')),
     (10, 1 << 10, _(b'%.1f KB')),
     (1, 1 << 10, _(b'%.2f KB')),
     (1, 1, _(b'%.0f bytes')),
 )
 
 
 class transformingwriter(object):
     """Writable file wrapper to transform data by function"""
 
     def __init__(self, fp, encode):
         self._fp = fp
         self._encode = encode
 
     def close(self):
         self._fp.close()
 
     def flush(self):
         self._fp.flush()
 
     def write(self, data):
         return self._fp.write(self._encode(data))
 
 
 # Matches a single EOL which can either be a CRLF where repeated CR
 # are removed or a LF. We do not care about old Macintosh files, so a
 # stray CR is an error.
 _eolre = remod.compile(br'\r*\n')
 
 
 def tolf(s):
     # type: (bytes) -> bytes
     return _eolre.sub(b'\n', s)
 
 
 def tocrlf(s):
     # type: (bytes) -> bytes
     return _eolre.sub(b'\r\n', s)
 
 
 def _crlfwriter(fp):
     return transformingwriter(fp, tocrlf)
 
 
 if pycompat.oslinesep == b'\r\n':
     tonativeeol = tocrlf
     fromnativeeol = tolf
     nativeeolwriter = _crlfwriter
 else:
     tonativeeol = pycompat.identity
     fromnativeeol = pycompat.identity
     nativeeolwriter = pycompat.identity
 
 if pyplatform.python_implementation() == b'CPython' and sys.version_info < (
     3,
     0,
 ):
     # There is an issue in CPython that some IO methods do not handle EINTR
     # correctly. The following table shows what CPython version (and functions)
     # are affected (buggy: has the EINTR bug, okay: otherwise):
     #
     #                | < 2.7.4 | 2.7.4 to 2.7.12 | >= 3.0
     #   --------------------------------------------------
     #    fp.__iter__ | buggy   | buggy           | okay
     #    fp.read*    | buggy   | okay [1]        | okay
     #
     # [1]: fixed by changeset 67dc99a989cd in the cpython hg repo.
     #
     # Here we workaround the EINTR issue for fileobj.__iter__. Other methods
     # like "read*" work fine, as we do not support Python < 2.7.4.
     #
     # Although we can workaround the EINTR issue for fp.__iter__, it is slower:
     # "for x in fp" is 4x faster than "for x in iter(fp.readline, '')" in
     # CPython 2, because CPython 2 maintains an internal readahead buffer for
     # fp.__iter__ but not other fp.read* methods.
     #
     # On modern systems like Linux, the "read" syscall cannot be interrupted
     # when reading "fast" files like on-disk files. So the EINTR issue only
     # affects things like pipes, sockets, ttys etc. We treat "normal" (S_ISREG)
     # files approximately as "fast" files and use the fast (unsafe) code path,
     # to minimize the performance impact.
 
     def iterfile(fp):
         fastpath = True
         if type(fp) is file:
             fastpath = stat.S_ISREG(os.fstat(fp.fileno()).st_mode)
         if fastpath:
             return fp
         else:
             # fp.readline deals with EINTR correctly, use it as a workaround.
             return iter(fp.readline, b'')
 
 
 else:
     # PyPy and CPython 3 do not have the EINTR issue thus no workaround needed.
     def iterfile(fp):
         return fp
 
 
 def iterlines(iterator):
     # type: (Iterator[bytes]) -> Iterator[bytes]
     for chunk in iterator:
         for line in chunk.splitlines():
             yield line
 
 
 def expandpath(path):
     # type: (bytes) -> bytes
     return os.path.expanduser(os.path.expandvars(path))
 
 
 def interpolate(prefix, mapping, s, fn=None, escape_prefix=False):
     """Return the result of interpolating items in the mapping into string s.
 
     prefix is a single character string, or a two character string with
     a backslash as the first character if the prefix needs to be escaped in
     a regular expression.
 
     fn is an optional function that will be applied to the replacement text
     just before replacement.
 
     escape_prefix is an optional flag that allows using doubled prefix for
     its escaping.
     """
     fn = fn or (lambda s: s)
     patterns = b'|'.join(mapping.keys())
     if escape_prefix:
         patterns += b'|' + prefix
         if len(prefix) > 1:
             prefix_char = prefix[1:]
         else:
             prefix_char = prefix
         mapping[prefix_char] = prefix_char
     r = remod.compile(br'%s(%s)' % (prefix, patterns))
     return r.sub(lambda x: fn(mapping[x.group()[1:]]), s)
 
 
-def getport(*args, **kwargs):
-    msg = b'getport(...) moved to mercurial.utils.urlutil'
-    nouideprecwarn(msg, b'6.0', stacklevel=2)
-    return urlutil.getport(*args, **kwargs)
-
-
-def url(*args, **kwargs):
-    msg = b'url(...) moved to mercurial.utils.urlutil'
-    nouideprecwarn(msg, b'6.0', stacklevel=2)
-    return urlutil.url(*args, **kwargs)
-
-
-def hasscheme(*args, **kwargs):
-    msg = b'hasscheme(...) moved to mercurial.utils.urlutil'
-    nouideprecwarn(msg, b'6.0', stacklevel=2)
-    return urlutil.hasscheme(*args, **kwargs)
-
-
-def hasdriveletter(*args, **kwargs):
-    msg = b'hasdriveletter(...) moved to mercurial.utils.urlutil'
-    nouideprecwarn(msg, b'6.0', stacklevel=2)
-    return urlutil.hasdriveletter(*args, **kwargs)
-
-
-def urllocalpath(*args, **kwargs):
-    msg = b'urllocalpath(...) moved to mercurial.utils.urlutil'
-    nouideprecwarn(msg, b'6.0', stacklevel=2)
-    return urlutil.urllocalpath(*args, **kwargs)
-
-
-def checksafessh(*args, **kwargs):
-    msg = b'checksafessh(...) moved to mercurial.utils.urlutil'
-    nouideprecwarn(msg, b'6.0', stacklevel=2)
-    return urlutil.checksafessh(*args, **kwargs)
-
-
-def hidepassword(*args, **kwargs):
-    msg = b'hidepassword(...) moved to mercurial.utils.urlutil'
-    nouideprecwarn(msg, b'6.0', stacklevel=2)
-    return urlutil.hidepassword(*args, **kwargs)
-
-
-def removeauth(*args, **kwargs):
-    msg = b'removeauth(...) moved to mercurial.utils.urlutil'
-    nouideprecwarn(msg, b'6.0', stacklevel=2)
-    return urlutil.removeauth(*args, **kwargs)
-
-
 timecount = unitcountfn(
     (1, 1e3, _(b'%.0f s')),
     (100, 1, _(b'%.1f s')),
     (10, 1, _(b'%.2f s')),
     (1, 1, _(b'%.3f s')),
     (100, 0.001, _(b'%.1f ms')),
     (10, 0.001, _(b'%.2f ms')),
     (1, 0.001, _(b'%.3f ms')),
     (100, 0.000001, _(b'%.1f us')),
     (10, 0.000001, _(b'%.2f us')),
     (1, 0.000001, _(b'%.3f us')),
     (100, 0.000000001, _(b'%.1f ns')),
     (10, 0.000000001, _(b'%.2f ns')),
     (1, 0.000000001, _(b'%.3f ns')),
 )
 
 
 @attr.s
 class timedcmstats(object):
     """Stats information produced by the timedcm context manager on entering."""
 
     # the starting value of the timer as a float (meaning and resulution is
     # platform dependent, see util.timer)
     start = attr.ib(default=attr.Factory(lambda: timer()))
     # the number of seconds as a floating point value; starts at 0, updated when
     # the context is exited.
     elapsed = attr.ib(default=0)
     # the number of nested timedcm context managers.
     level = attr.ib(default=1)
 
     def __bytes__(self):
         return timecount(self.elapsed) if self.elapsed else b'<unknown>'
 
     __str__ = encoding.strmethod(__bytes__)
 
 
 @contextlib.contextmanager
 def timedcm(whencefmt, *whenceargs):
     """A context manager that produces timing information for a given context.
 
     On entering a timedcmstats instance is produced.
 
     This context manager is reentrant.
 
     """
     # track nested context managers
     timedcm._nested += 1
     timing_stats = timedcmstats(level=timedcm._nested)
     try:
         with tracing.log(whencefmt, *whenceargs):
             yield timing_stats
     finally:
         timing_stats.elapsed = timer() - timing_stats.start
         timedcm._nested -= 1
 
 
 timedcm._nested = 0
 
 
 def timed(func):
     """Report the execution time of a function call to stderr.
 
     During development, use as a decorator when you need to measure
     the cost of a function, e.g. as follows:
 
     @util.timed
     def foo(a, b, c):
         pass
     """
 
     def wrapper(*args, **kwargs):
         with timedcm(pycompat.bytestr(func.__name__)) as time_stats:
             result = func(*args, **kwargs)
         stderr = procutil.stderr
         stderr.write(
             b'%s%s: %s\n'
             % (
                 b' ' * time_stats.level * 2,
                 pycompat.bytestr(func.__name__),
                 time_stats,
             )
         )
         return result
 
     return wrapper
 
 
 _sizeunits = (
     (b'm', 2 ** 20),
     (b'k', 2 ** 10),
     (b'g', 2 ** 30),
     (b'kb', 2 ** 10),
     (b'mb', 2 ** 20),
     (b'gb', 2 ** 30),
     (b'b', 1),
 )
 
 
 def sizetoint(s):
     # type: (bytes) -> int
     """Convert a space specifier to a byte count.
 
     >>> sizetoint(b'30')
     30
     >>> sizetoint(b'2.2kb')
     2252
     >>> sizetoint(b'6M')
     6291456
     """
     t = s.strip().lower()
     try:
         for k, u in _sizeunits:
             if t.endswith(k):
                 return int(float(t[: -len(k)]) * u)
         return int(t)
     except ValueError:
         raise error.ParseError(_(b"couldn't parse size: %s") % s)
 
 
 class hooks(object):
     """A collection of hook functions that can be used to extend a
     function's behavior. Hooks are called in lexicographic order,
     based on the names of their sources."""
 
     def __init__(self):
         self._hooks = []
 
     def add(self, source, hook):
         self._hooks.append((source, hook))
 
     def __call__(self, *args):
         self._hooks.sort(key=lambda x: x[0])
         results = []
         for source, hook in self._hooks:
             results.append(hook(*args))
         return results
 
 
 def getstackframes(skip=0, line=b' %-*s in %s\n', fileline=b'%s:%d', depth=0):
     """Yields lines for a nicely formatted stacktrace.
     Skips the 'skip' last entries, then return the last 'depth' entries.
     Each file+linenumber is formatted according to fileline.
     Each line is formatted according to line.
     If line is None, it yields:
       length of longest filepath+line number,
       filepath+linenumber,
       function
 
     Not be used in production code but very convenient while developing.
     """
     entries = [
         (fileline % (pycompat.sysbytes(fn), ln), pycompat.sysbytes(func))
         for fn, ln, func, _text in traceback.extract_stack()[: -skip - 1]
     ][-depth:]
     if entries:
         fnmax = max(len(entry[0]) for entry in entries)
         for fnln, func in entries:
             if line is None:
                 yield (fnmax, fnln, func)
             else:
                 yield line % (fnmax, fnln, func)
 
 
 def debugstacktrace(
     msg=b'stacktrace',
     skip=0,
     f=procutil.stderr,
     otherf=procutil.stdout,
     depth=0,
     prefix=b'',
 ):
     """Writes a message to f (stderr) with a nicely formatted stacktrace.
     Skips the 'skip' entries closest to the call, then show 'depth' entries.
     By default it will flush stdout first.
     It can be used everywhere and intentionally does not require an ui object.
     Not be used in production code but very convenient while developing.
     """
     if otherf:
         otherf.flush()
     f.write(b'%s%s at:\n' % (prefix, msg.rstrip()))
     for line in getstackframes(skip + 1, depth=depth):
         f.write(prefix + line)
     f.flush()
 
 
 # convenient shortcut
 dst = debugstacktrace
 
 
 def safename(f, tag, ctx, others=None):
     """
     Generate a name that it is safe to rename f to in the given context.
 
     f:      filename to rename
     tag:    a string tag that will be included in the new name
     ctx:    a context, in which the new name must not exist
     others: a set of other filenames that the new name must not be in
 
     Returns a file name of the form oldname~tag[~number] which does not exist
     in the provided context and is not in the set of other names.
     """
     if others is None:
         others = set()
 
     fn = b'%s~%s' % (f, tag)
     if fn not in ctx and fn not in others:
         return fn
     for n in itertools.count(1):
         fn = b'%s~%s~%s' % (f, tag, n)
         if fn not in ctx and fn not in others:
             return fn
 
 
 def readexactly(stream, n):
     '''read n bytes from stream.read and abort if less was available'''
     s = stream.read(n)
     if len(s) < n:
         raise error.Abort(
             _(b"stream ended unexpectedly (got %d bytes, expected %d)")
             % (len(s), n)
         )
     return s
 
 
 def uvarintencode(value):
     """Encode an unsigned integer value to a varint.
 
     A varint is a variable length integer of 1 or more bytes. Each byte
     except the last has the most significant bit set. The lower 7 bits of
     each byte store the 2's complement representation, least significant group
     first.
 
     >>> uvarintencode(0)
     '\\x00'
     >>> uvarintencode(1)
     '\\x01'
     >>> uvarintencode(127)
     '\\x7f'
     >>> uvarintencode(1337)
     '\\xb9\\n'
     >>> uvarintencode(65536)
     '\\x80\\x80\\x04'
     >>> uvarintencode(-1)
     Traceback (most recent call last):
         ...
     ProgrammingError: negative value for uvarint: -1
     """
     if value < 0:
         raise error.ProgrammingError(b'negative value for uvarint: %d' % value)
     bits = value & 0x7F
     value >>= 7
     bytes = []
     while value:
         bytes.append(pycompat.bytechr(0x80 | bits))
         bits = value & 0x7F
         value >>= 7
     bytes.append(pycompat.bytechr(bits))
 
     return b''.join(bytes)
 
 
 def uvarintdecodestream(fh):
     """Decode an unsigned variable length integer from a stream.
 
     The passed argument is anything that has a ``.read(N)`` method.
 
     >>> try:
     ...     from StringIO import StringIO as BytesIO
     ... except ImportError:
     ...     from io import BytesIO
     >>> uvarintdecodestream(BytesIO(b'\\x00'))
     0
     >>> uvarintdecodestream(BytesIO(b'\\x01'))
     1
     >>> uvarintdecodestream(BytesIO(b'\\x7f'))
     127
     >>> uvarintdecodestream(BytesIO(b'\\xb9\\n'))
     1337
     >>> uvarintdecodestream(BytesIO(b'\\x80\\x80\\x04'))
     65536
     >>> uvarintdecodestream(BytesIO(b'\\x80'))
     Traceback (most recent call last):
         ...
     Abort: stream ended unexpectedly (got 0 bytes, expected 1)
     """
     result = 0
     shift = 0
     while True:
         byte = ord(readexactly(fh, 1))
         result |= (byte & 0x7F) << shift
         if not (byte & 0x80):
             return result
         shift += 7
 
 
 # Passing the '' locale means that the locale should be set according to the
 # user settings (environment variables).
 # Python sometimes avoids setting the global locale settings. When interfacing
 # with C code (e.g. the curses module or the Subversion bindings), the global
 # locale settings must be initialized correctly. Python 2 does not initialize
 # the global locale settings on interpreter startup. Python 3 sometimes
 # initializes LC_CTYPE, but not consistently at least on Windows. Therefore we
 # explicitly initialize it to get consistent behavior if it's not already
 # initialized. Since CPython commit 177d921c8c03d30daa32994362023f777624b10d,
 # LC_CTYPE is always initialized. If we require Python 3.8+, we should re-check
 # if we can remove this code.
 @contextlib.contextmanager
 def with_lc_ctype():
     oldloc = locale.setlocale(locale.LC_CTYPE, None)
     if oldloc == 'C':
         try:
             try:
                 locale.setlocale(locale.LC_CTYPE, '')
             except locale.Error:
                 # The likely case is that the locale from the environment
                 # variables is unknown.
                 pass
             yield
         finally:
             locale.setlocale(locale.LC_CTYPE, oldloc)
     else:
         yield
 
 
 def _estimatememory():
     # type: () -> Optional[int]
     """Provide an estimate for the available system memory in Bytes.
 
     If no estimate can be provided on the platform, returns None.
     """
     if pycompat.sysplatform.startswith(b'win'):
         # On Windows, use the GlobalMemoryStatusEx kernel function directly.
         from ctypes import c_long as DWORD, c_ulonglong as DWORDLONG
         from ctypes.wintypes import (  # pytype: disable=import-error
             Structure,
             byref,
             sizeof,
             windll,
         )
 
         class MEMORYSTATUSEX(Structure):
             _fields_ = [
                 ('dwLength', DWORD),
                 ('dwMemoryLoad', DWORD),
                 ('ullTotalPhys', DWORDLONG),
                 ('ullAvailPhys', DWORDLONG),
                 ('ullTotalPageFile', DWORDLONG),
                 ('ullAvailPageFile', DWORDLONG),
                 ('ullTotalVirtual', DWORDLONG),
                 ('ullAvailVirtual', DWORDLONG),
                 ('ullExtendedVirtual', DWORDLONG),
             ]
 
         x = MEMORYSTATUSEX()
         x.dwLength = sizeof(x)
         windll.kernel32.GlobalMemoryStatusEx(byref(x))
         return x.ullAvailPhys
 
     # On newer Unix-like systems and Mac OSX, the sysconf interface
     # can be used. _SC_PAGE_SIZE is part of POSIX; _SC_PHYS_PAGES
     # seems to be implemented on most systems.
     try:
         pagesize = os.sysconf(os.sysconf_names['SC_PAGE_SIZE'])
         pages = os.sysconf(os.sysconf_names['SC_PHYS_PAGES'])
         return pagesize * pages
     except OSError:  # sysconf can fail
         pass
     except KeyError:  # unknown parameter
         pass