merge with stable

mercurial/context.py

@@ -1,2748 +1,2749 b''
 # context.py - changeset and file context objects for mercurial
 #
 # Copyright 2006, 2007 Matt Mackall <[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 filecmp
 import os
 import re
 import stat
 
 from .i18n import _
 from .node import (
     addednodeid,
     bin,
     hex,
     modifiednodeid,
     nullid,
     nullrev,
     short,
     wdirid,
     wdirnodes,
     wdirrev,
 )
 from .thirdparty import (
     attr,
 )
 from . import (
     encoding,
     error,
     fileset,
     match as matchmod,
     mdiff,
     obsolete as obsmod,
     obsutil,
     patch,
     pathutil,
     phases,
     pycompat,
     repoview,
     revlog,
     scmutil,
     sparse,
     subrepo,
     subrepoutil,
     util,
 )
 from .utils import dateutil
 
 propertycache = util.propertycache
 
 nonascii = re.compile(r'[^\x21-\x7f]').search
 
 class basectx(object):
     """A basectx object represents the common logic for its children:
     changectx: read-only context that is already present in the repo,
     workingctx: a context that represents the working directory and can
                 be committed,
     memctx: a context that represents changes in-memory and can also
             be committed."""
     def __new__(cls, repo, changeid='', *args, **kwargs):
         if isinstance(changeid, basectx):
             return changeid
 
         o = super(basectx, cls).__new__(cls)
 
         o._repo = repo
         o._rev = nullrev
         o._node = nullid
 
         return o
 
     def __bytes__(self):
         return short(self.node())
 
     __str__ = encoding.strmethod(__bytes__)
 
     def __repr__(self):
         return r"<%s %s>" % (type(self).__name__, str(self))
 
     def __eq__(self, other):
         try:
             return type(self) == type(other) and self._rev == other._rev
         except AttributeError:
             return False
 
     def __ne__(self, other):
         return not (self == other)
 
     def __contains__(self, key):
         return key in self._manifest
 
     def __getitem__(self, key):
         return self.filectx(key)
 
     def __iter__(self):
         return iter(self._manifest)
 
     def _buildstatusmanifest(self, status):
         """Builds a manifest that includes the given status results, if this is
         a working copy context. For non-working copy contexts, it just returns
         the normal manifest."""
         return self.manifest()
 
     def _matchstatus(self, other, match):
         """This internal method provides a way for child objects to override the
         match operator.
         """
         return match
 
     def _buildstatus(self, other, s, match, listignored, listclean,
                      listunknown):
         """build a status with respect to another context"""
         # Load earliest manifest first for caching reasons. More specifically,
         # if you have revisions 1000 and 1001, 1001 is probably stored as a
         # delta against 1000. Thus, if you read 1000 first, we'll reconstruct
         # 1000 and cache it so that when you read 1001, we just need to apply a
         # delta to what's in the cache. So that's one full reconstruction + one
         # delta application.
         mf2 = None
         if self.rev() is not None and self.rev() < other.rev():
             mf2 = self._buildstatusmanifest(s)
         mf1 = other._buildstatusmanifest(s)
         if mf2 is None:
             mf2 = self._buildstatusmanifest(s)
 
         modified, added = [], []
         removed = []
         clean = []
         deleted, unknown, ignored = s.deleted, s.unknown, s.ignored
         deletedset = set(deleted)
         d = mf1.diff(mf2, match=match, clean=listclean)
         for fn, value in d.iteritems():
             if fn in deletedset:
                 continue
             if value is None:
                 clean.append(fn)
                 continue
             (node1, flag1), (node2, flag2) = value
             if node1 is None:
                 added.append(fn)
             elif node2 is None:
                 removed.append(fn)
             elif flag1 != flag2:
                 modified.append(fn)
             elif node2 not in wdirnodes:
                 # When comparing files between two commits, we save time by
                 # not comparing the file contents when the nodeids differ.
                 # Note that this means we incorrectly report a reverted change
                 # to a file as a modification.
                 modified.append(fn)
             elif self[fn].cmp(other[fn]):
                 modified.append(fn)
             else:
                 clean.append(fn)
 
         if removed:
             # need to filter files if they are already reported as removed
             unknown = [fn for fn in unknown if fn not in mf1 and
                                                (not match or match(fn))]
             ignored = [fn for fn in ignored if fn not in mf1 and
                                                (not match or match(fn))]
             # if they're deleted, don't report them as removed
             removed = [fn for fn in removed if fn not in deletedset]
 
         return scmutil.status(modified, added, removed, deleted, unknown,
                               ignored, clean)
 
     @propertycache
     def substate(self):
         return subrepoutil.state(self, self._repo.ui)
 
     def subrev(self, subpath):
         return self.substate[subpath][1]
 
     def rev(self):
         return self._rev
     def node(self):
         return self._node
     def hex(self):
         return hex(self.node())
     def manifest(self):
         return self._manifest
     def manifestctx(self):
         return self._manifestctx
     def repo(self):
         return self._repo
     def phasestr(self):
         return phases.phasenames[self.phase()]
     def mutable(self):
         return self.phase() > phases.public
 
     def getfileset(self, expr):
         return fileset.getfileset(self, expr)
 
     def obsolete(self):
         """True if the changeset is obsolete"""
         return self.rev() in obsmod.getrevs(self._repo, 'obsolete')
 
     def extinct(self):
         """True if the changeset is extinct"""
         return self.rev() in obsmod.getrevs(self._repo, 'extinct')
 
     def orphan(self):
         """True if the changeset is not obsolete but it's ancestor are"""
         return self.rev() in obsmod.getrevs(self._repo, 'orphan')
 
     def phasedivergent(self):
         """True if the changeset try to be a successor of a public changeset
 
         Only non-public and non-obsolete changesets may be bumped.
         """
         return self.rev() in obsmod.getrevs(self._repo, 'phasedivergent')
 
     def contentdivergent(self):
         """Is a successors of a changeset with multiple possible successors set
 
         Only non-public and non-obsolete changesets may be divergent.
         """
         return self.rev() in obsmod.getrevs(self._repo, 'contentdivergent')
 
     def isunstable(self):
         """True if the changeset is either unstable, bumped or divergent"""
         return self.orphan() or self.phasedivergent() or self.contentdivergent()
 
     def instabilities(self):
         """return the list of instabilities affecting this changeset.
 
         Instabilities are returned as strings. possible values are:
         - orphan,
         - phase-divergent,
         - content-divergent.
         """
         instabilities = []
         if self.orphan():
             instabilities.append('orphan')
         if self.phasedivergent():
             instabilities.append('phase-divergent')
         if self.contentdivergent():
             instabilities.append('content-divergent')
         return instabilities
 
     def parents(self):
         """return contexts for each parent changeset"""
         return self._parents
 
     def p1(self):
         return self._parents[0]
 
     def p2(self):
         parents = self._parents
         if len(parents) == 2:
             return parents[1]
         return changectx(self._repo, nullrev)
 
     def _fileinfo(self, path):
         if r'_manifest' in self.__dict__:
             try:
                 return self._manifest[path], self._manifest.flags(path)
             except KeyError:
                 raise error.ManifestLookupError(self._node, path,
                                                 _('not found in manifest'))
         if r'_manifestdelta' in self.__dict__ or path in self.files():
             if path in self._manifestdelta:
                 return (self._manifestdelta[path],
                         self._manifestdelta.flags(path))
         mfl = self._repo.manifestlog
         try:
             node, flag = mfl[self._changeset.manifest].find(path)
         except KeyError:
             raise error.ManifestLookupError(self._node, path,
                                             _('not found in manifest'))
 
         return node, flag
 
     def filenode(self, path):
         return self._fileinfo(path)[0]
 
     def flags(self, path):
         try:
             return self._fileinfo(path)[1]
         except error.LookupError:
             return ''
 
     def sub(self, path, allowcreate=True):
         '''return a subrepo for the stored revision of path, never wdir()'''
         return subrepo.subrepo(self, path, allowcreate=allowcreate)
 
     def nullsub(self, path, pctx):
         return subrepo.nullsubrepo(self, path, pctx)
 
     def workingsub(self, path):
         '''return a subrepo for the stored revision, or wdir if this is a wdir
         context.
         '''
         return subrepo.subrepo(self, path, allowwdir=True)
 
     def match(self, pats=None, include=None, exclude=None, default='glob',
               listsubrepos=False, badfn=None):
         r = self._repo
         return matchmod.match(r.root, r.getcwd(), pats,
                               include, exclude, default,
                               auditor=r.nofsauditor, ctx=self,
                               listsubrepos=listsubrepos, badfn=badfn)
 
     def diff(self, ctx2=None, match=None, **opts):
         """Returns a diff generator for the given contexts and matcher"""
         if ctx2 is None:
             ctx2 = self.p1()
         if ctx2 is not None:
             ctx2 = self._repo[ctx2]
         diffopts = patch.diffopts(self._repo.ui, pycompat.byteskwargs(opts))
         return patch.diff(self._repo, ctx2, self, match=match, opts=diffopts)
 
     def dirs(self):
         return self._manifest.dirs()
 
     def hasdir(self, dir):
         return self._manifest.hasdir(dir)
 
     def status(self, other=None, match=None, listignored=False,
                listclean=False, listunknown=False, listsubrepos=False):
         """return status of files between two nodes or node and working
         directory.
 
         If other is None, compare this node with working directory.
 
         returns (modified, added, removed, deleted, unknown, ignored, clean)
         """
 
         ctx1 = self
         ctx2 = self._repo[other]
 
         # This next code block is, admittedly, fragile logic that tests for
         # reversing the contexts and wouldn't need to exist if it weren't for
         # the fast (and common) code path of comparing the working directory
         # with its first parent.
         #
         # What we're aiming for here is the ability to call:
         #
         # workingctx.status(parentctx)
         #
         # If we always built the manifest for each context and compared those,
         # then we'd be done. But the special case of the above call means we
         # just copy the manifest of the parent.
         reversed = False
         if (not isinstance(ctx1, changectx)
             and isinstance(ctx2, changectx)):
             reversed = True
             ctx1, ctx2 = ctx2, ctx1
 
         match = match or matchmod.always(self._repo.root, self._repo.getcwd())
         match = ctx2._matchstatus(ctx1, match)
         r = scmutil.status([], [], [], [], [], [], [])
         r = ctx2._buildstatus(ctx1, r, match, listignored, listclean,
                               listunknown)
 
         if reversed:
             # Reverse added and removed. Clear deleted, unknown and ignored as
             # these make no sense to reverse.
             r = scmutil.status(r.modified, r.removed, r.added, [], [], [],
                                r.clean)
 
         if listsubrepos:
             for subpath, sub in scmutil.itersubrepos(ctx1, ctx2):
                 try:
                     rev2 = ctx2.subrev(subpath)
                 except KeyError:
                     # A subrepo that existed in node1 was deleted between
                     # node1 and node2 (inclusive). Thus, ctx2's substate
                     # won't contain that subpath. The best we can do ignore it.
                     rev2 = None
                 submatch = matchmod.subdirmatcher(subpath, match)
                 s = sub.status(rev2, match=submatch, ignored=listignored,
                                clean=listclean, unknown=listunknown,
                                listsubrepos=True)
                 for rfiles, sfiles in zip(r, s):
                     rfiles.extend("%s/%s" % (subpath, f) for f in sfiles)
 
         for l in r:
             l.sort()
 
         return r
 
 def _filterederror(repo, changeid):
     """build an exception to be raised about a filtered changeid
 
     This is extracted in a function to help extensions (eg: evolve) to
     experiment with various message variants."""
     if repo.filtername.startswith('visible'):
 
         # Check if the changeset is obsolete
         unfilteredrepo = repo.unfiltered()
         ctx = unfilteredrepo[changeid]
 
         # If the changeset is obsolete, enrich the message with the reason
         # that made this changeset not visible
         if ctx.obsolete():
             msg = obsutil._getfilteredreason(repo, changeid, ctx)
         else:
             msg = _("hidden revision '%s'") % changeid
 
         hint = _('use --hidden to access hidden revisions')
 
         return error.FilteredRepoLookupError(msg, hint=hint)
     msg = _("filtered revision '%s' (not in '%s' subset)")
     msg %= (changeid, repo.filtername)
     return error.FilteredRepoLookupError(msg)
 
 class changectx(basectx):
     """A changecontext object makes access to data related to a particular
     changeset convenient. It represents a read-only context already present in
     the repo."""
     def __init__(self, repo, changeid=''):
         """changeid is a revision number, node, or tag"""
 
         # since basectx.__new__ already took care of copying the object, we
         # don't need to do anything in __init__, so we just exit here
         if isinstance(changeid, basectx):
             return
 
         if changeid == '':
             changeid = '.'
         self._repo = repo
 
         try:
             if isinstance(changeid, int):
                 self._node = repo.changelog.node(changeid)
                 self._rev = changeid
                 return
             if not pycompat.ispy3 and isinstance(changeid, long):
                 changeid = str(changeid)
             if changeid == 'null':
                 self._node = nullid
                 self._rev = nullrev
                 return
             if changeid == 'tip':
                 self._node = repo.changelog.tip()
                 self._rev = repo.changelog.rev(self._node)
                 return
             if (changeid == '.'
                 or repo.local() and changeid == repo.dirstate.p1()):
                 # this is a hack to delay/avoid loading obsmarkers
                 # when we know that '.' won't be hidden
                 self._node = repo.dirstate.p1()
                 self._rev = repo.unfiltered().changelog.rev(self._node)
                 return
             if len(changeid) == 20:
                 try:
                     self._node = changeid
                     self._rev = repo.changelog.rev(changeid)
                     return
                 except error.FilteredRepoLookupError:
                     raise
                 except LookupError:
                     pass
 
             try:
                 r = int(changeid)
                 if '%d' % r != changeid:
                     raise ValueError
                 l = len(repo.changelog)
                 if r < 0:
                     r += l
                 if r < 0 or r >= l and r != wdirrev:
                     raise ValueError
                 self._rev = r
                 self._node = repo.changelog.node(r)
                 return
             except error.FilteredIndexError:
                 raise
             except (ValueError, OverflowError, IndexError):
                 pass
 
             if len(changeid) == 40:
                 try:
                     self._node = bin(changeid)
                     self._rev = repo.changelog.rev(self._node)
                     return
                 except error.FilteredLookupError:
                     raise
                 except (TypeError, LookupError):
                     pass
 
             # lookup bookmarks through the name interface
             try:
                 self._node = repo.names.singlenode(repo, changeid)
                 self._rev = repo.changelog.rev(self._node)
                 return
             except KeyError:
                 pass
             except error.FilteredRepoLookupError:
                 raise
             except error.RepoLookupError:
                 pass
 
             self._node = repo.unfiltered().changelog._partialmatch(changeid)
             if self._node is not None:
                 self._rev = repo.changelog.rev(self._node)
                 return
 
             # lookup failed
             # check if it might have come from damaged dirstate
             #
             # XXX we could avoid the unfiltered if we had a recognizable
             # exception for filtered changeset access
             if (repo.local()
                 and changeid in repo.unfiltered().dirstate.parents()):
                 msg = _("working directory has unknown parent '%s'!")
                 raise error.Abort(msg % short(changeid))
             try:
                 if len(changeid) == 20 and nonascii(changeid):
                     changeid = hex(changeid)
             except TypeError:
                 pass
         except (error.FilteredIndexError, error.FilteredLookupError,
                 error.FilteredRepoLookupError):
             raise _filterederror(repo, changeid)
         except IndexError:
             pass
         raise error.RepoLookupError(
             _("unknown revision '%s'") % changeid)
 
     def __hash__(self):
         try:
             return hash(self._rev)
         except AttributeError:
             return id(self)
 
     def __nonzero__(self):
         return self._rev != nullrev
 
     __bool__ = __nonzero__
 
     @propertycache
     def _changeset(self):
         return self._repo.changelog.changelogrevision(self.rev())
 
     @propertycache
     def _manifest(self):
         return self._manifestctx.read()
 
     @property
     def _manifestctx(self):
         return self._repo.manifestlog[self._changeset.manifest]
 
     @propertycache
     def _manifestdelta(self):
         return self._manifestctx.readdelta()
 
     @propertycache
     def _parents(self):
         repo = self._repo
         p1, p2 = repo.changelog.parentrevs(self._rev)
         if p2 == nullrev:
             return [changectx(repo, p1)]
         return [changectx(repo, p1), changectx(repo, p2)]
 
     def changeset(self):
         c = self._changeset
         return (
             c.manifest,
             c.user,
             c.date,
             c.files,
             c.description,
             c.extra,
         )
     def manifestnode(self):
         return self._changeset.manifest
 
     def user(self):
         return self._changeset.user
     def date(self):
         return self._changeset.date
     def files(self):
         return self._changeset.files
     def description(self):
         return self._changeset.description
     def branch(self):
         return encoding.tolocal(self._changeset.extra.get("branch"))
     def closesbranch(self):
         return 'close' in self._changeset.extra
     def extra(self):
         """Return a dict of extra information."""
         return self._changeset.extra
     def tags(self):
         """Return a list of byte tag names"""
         return self._repo.nodetags(self._node)
     def bookmarks(self):
         """Return a list of byte bookmark names."""
         return self._repo.nodebookmarks(self._node)
     def phase(self):
         return self._repo._phasecache.phase(self._repo, self._rev)
     def hidden(self):
         return self._rev in repoview.filterrevs(self._repo, 'visible')
 
     def isinmemory(self):
         return False
 
     def children(self):
         """return list of changectx contexts for each child changeset.
 
         This returns only the immediate child changesets. Use descendants() to
         recursively walk children.
         """
         c = self._repo.changelog.children(self._node)
         return [changectx(self._repo, x) for x in c]
 
     def ancestors(self):
         for a in self._repo.changelog.ancestors([self._rev]):
             yield changectx(self._repo, a)
 
     def descendants(self):
         """Recursively yield all children of the changeset.
 
         For just the immediate children, use children()
         """
         for d in self._repo.changelog.descendants([self._rev]):
             yield changectx(self._repo, d)
 
     def filectx(self, path, fileid=None, filelog=None):
         """get a file context from this changeset"""
         if fileid is None:
             fileid = self.filenode(path)
         return filectx(self._repo, path, fileid=fileid,
                        changectx=self, filelog=filelog)
 
     def ancestor(self, c2, warn=False):
         """return the "best" ancestor context of self and c2
 
         If there are multiple candidates, it will show a message and check
         merge.preferancestor configuration before falling back to the
         revlog ancestor."""
         # deal with workingctxs
         n2 = c2._node
         if n2 is None:
             n2 = c2._parents[0]._node
         cahs = self._repo.changelog.commonancestorsheads(self._node, n2)
         if not cahs:
             anc = nullid
         elif len(cahs) == 1:
             anc = cahs[0]
         else:
             # experimental config: merge.preferancestor
             for r in self._repo.ui.configlist('merge', 'preferancestor'):
                 try:
                     ctx = changectx(self._repo, r)
                 except error.RepoLookupError:
                     continue
                 anc = ctx.node()
                 if anc in cahs:
                     break
             else:
                 anc = self._repo.changelog.ancestor(self._node, n2)
             if warn:
                 self._repo.ui.status(
                     (_("note: using %s as ancestor of %s and %s\n") %
                      (short(anc), short(self._node), short(n2))) +
                     ''.join(_("      alternatively, use --config "
                               "merge.preferancestor=%s\n") %
                             short(n) for n in sorted(cahs) if n != anc))
         return changectx(self._repo, anc)
 
     def descendant(self, other):
         """True if other is descendant of this changeset"""
         return self._repo.changelog.descendant(self._rev, other._rev)
 
     def walk(self, match):
         '''Generates matching file names.'''
 
         # Wrap match.bad method to have message with nodeid
         def bad(fn, msg):
             # The manifest doesn't know about subrepos, so don't complain about
             # paths into valid subrepos.
             if any(fn == s or fn.startswith(s + '/')
                    for s in self.substate):
                 return
             match.bad(fn, _('no such file in rev %s') % self)
 
         m = matchmod.badmatch(match, bad)
         return self._manifest.walk(m)
 
     def matches(self, match):
         return self.walk(match)
 
 class basefilectx(object):
     """A filecontext object represents the common logic for its children:
     filectx: read-only access to a filerevision that is already present
              in the repo,
     workingfilectx: a filecontext that represents files from the working
                     directory,
     memfilectx: a filecontext that represents files in-memory,
     overlayfilectx: duplicate another filecontext with some fields overridden.
     """
     @propertycache
     def _filelog(self):
         return self._repo.file(self._path)
 
     @propertycache
     def _changeid(self):
         if r'_changeid' in self.__dict__:
             return self._changeid
         elif r'_changectx' in self.__dict__:
             return self._changectx.rev()
         elif r'_descendantrev' in self.__dict__:
             # this file context was created from a revision with a known
             # descendant, we can (lazily) correct for linkrev aliases
             return self._adjustlinkrev(self._descendantrev)
         else:
             return self._filelog.linkrev(self._filerev)
 
     @propertycache
     def _filenode(self):
         if r'_fileid' in self.__dict__:
             return self._filelog.lookup(self._fileid)
         else:
             return self._changectx.filenode(self._path)
 
     @propertycache
     def _filerev(self):
         return self._filelog.rev(self._filenode)
 
     @propertycache
     def _repopath(self):
         return self._path
 
     def __nonzero__(self):
         try:
             self._filenode
             return True
         except error.LookupError:
             # file is missing
             return False
 
     __bool__ = __nonzero__
 
     def __bytes__(self):
         try:
             return "%s@%s" % (self.path(), self._changectx)
         except error.LookupError:
             return "%s@???" % self.path()
 
     __str__ = encoding.strmethod(__bytes__)
 
     def __repr__(self):
         return "<%s %s>" % (type(self).__name__, str(self))
 
     def __hash__(self):
         try:
             return hash((self._path, self._filenode))
         except AttributeError:
             return id(self)
 
     def __eq__(self, other):
         try:
             return (type(self) == type(other) and self._path == other._path
                     and self._filenode == other._filenode)
         except AttributeError:
             return False
 
     def __ne__(self, other):
         return not (self == other)
 
     def filerev(self):
         return self._filerev
     def filenode(self):
         return self._filenode
     @propertycache
     def _flags(self):
         return self._changectx.flags(self._path)
     def flags(self):
         return self._flags
     def filelog(self):
         return self._filelog
     def rev(self):
         return self._changeid
     def linkrev(self):
         return self._filelog.linkrev(self._filerev)
     def node(self):
         return self._changectx.node()
     def hex(self):
         return self._changectx.hex()
     def user(self):
         return self._changectx.user()
     def date(self):
         return self._changectx.date()
     def files(self):
         return self._changectx.files()
     def description(self):
         return self._changectx.description()
     def branch(self):
         return self._changectx.branch()
     def extra(self):
         return self._changectx.extra()
     def phase(self):
         return self._changectx.phase()
     def phasestr(self):
         return self._changectx.phasestr()
     def obsolete(self):
         return self._changectx.obsolete()
     def instabilities(self):
         return self._changectx.instabilities()
     def manifest(self):
         return self._changectx.manifest()
     def changectx(self):
         return self._changectx
     def renamed(self):
         return self._copied
     def repo(self):
         return self._repo
     def size(self):
         return len(self.data())
 
     def path(self):
         return self._path
 
     def isbinary(self):
         try:
             return util.binary(self.data())
         except IOError:
             return False
     def isexec(self):
         return 'x' in self.flags()
     def islink(self):
         return 'l' in self.flags()
 
     def isabsent(self):
         """whether this filectx represents a file not in self._changectx
 
         This is mainly for merge code to detect change/delete conflicts. This is
         expected to be True for all subclasses of basectx."""
         return False
 
     _customcmp = False
     def cmp(self, fctx):
         """compare with other file context
 
         returns True if different than fctx.
         """
         if fctx._customcmp:
             return fctx.cmp(self)
 
         if (fctx._filenode is None
             and (self._repo._encodefilterpats
                  # if file data starts with '\1\n', empty metadata block is
                  # prepended, which adds 4 bytes to filelog.size().
                  or self.size() - 4 == fctx.size())
             or self.size() == fctx.size()):
             return self._filelog.cmp(self._filenode, fctx.data())
 
         return True
 
     def _adjustlinkrev(self, srcrev, inclusive=False):
         """return the first ancestor of <srcrev> introducing <fnode>
 
         If the linkrev of the file revision does not point to an ancestor of
         srcrev, we'll walk down the ancestors until we find one introducing
         this file revision.
 
         :srcrev: the changeset revision we search ancestors from
         :inclusive: if true, the src revision will also be checked
         """
         repo = self._repo
         cl = repo.unfiltered().changelog
         mfl = repo.manifestlog
         # fetch the linkrev
         lkr = self.linkrev()
         # hack to reuse ancestor computation when searching for renames
         memberanc = getattr(self, '_ancestrycontext', None)
         iteranc = None
         if srcrev is None:
             # wctx case, used by workingfilectx during mergecopy
             revs = [p.rev() for p in self._repo[None].parents()]
             inclusive = True # we skipped the real (revless) source
         else:
             revs = [srcrev]
         if memberanc is None:
             memberanc = iteranc = cl.ancestors(revs, lkr,
                                                inclusive=inclusive)
         # check if this linkrev is an ancestor of srcrev
         if lkr not in memberanc:
             if iteranc is None:
                 iteranc = cl.ancestors(revs, lkr, inclusive=inclusive)
             fnode = self._filenode
             path = self._path
             for a in iteranc:
                 ac = cl.read(a) # get changeset data (we avoid object creation)
                 if path in ac[3]: # checking the 'files' field.
                     # The file has been touched, check if the content is
                     # similar to the one we search for.
                     if fnode == mfl[ac[0]].readfast().get(path):
                         return a
             # In theory, we should never get out of that loop without a result.
             # But if manifest uses a buggy file revision (not children of the
             # one it replaces) we could. Such a buggy situation will likely
             # result is crash somewhere else at to some point.
         return lkr
 
     def introrev(self):
         """return the rev of the changeset which introduced this file revision
 
         This method is different from linkrev because it take into account the
         changeset the filectx was created from. It ensures the returned
         revision is one of its ancestors. This prevents bugs from
         'linkrev-shadowing' when a file revision is used by multiple
         changesets.
         """
         lkr = self.linkrev()
         attrs = vars(self)
         noctx = not (r'_changeid' in attrs or r'_changectx' in attrs)
         if noctx or self.rev() == lkr:
             return self.linkrev()
         return self._adjustlinkrev(self.rev(), inclusive=True)
 
     def introfilectx(self):
         """Return filectx having identical contents, but pointing to the
         changeset revision where this filectx was introduced"""
         introrev = self.introrev()
         if self.rev() == introrev:
             return self
         return self.filectx(self.filenode(), changeid=introrev)
 
     def _parentfilectx(self, path, fileid, filelog):
         """create parent filectx keeping ancestry info for _adjustlinkrev()"""
         fctx = filectx(self._repo, path, fileid=fileid, filelog=filelog)
         if r'_changeid' in vars(self) or r'_changectx' in vars(self):
             # If self is associated with a changeset (probably explicitly
             # fed), ensure the created filectx is associated with a
             # changeset that is an ancestor of self.changectx.
             # This lets us later use _adjustlinkrev to get a correct link.
             fctx._descendantrev = self.rev()
             fctx._ancestrycontext = getattr(self, '_ancestrycontext', None)
         elif r'_descendantrev' in vars(self):
             # Otherwise propagate _descendantrev if we have one associated.
             fctx._descendantrev = self._descendantrev
             fctx._ancestrycontext = getattr(self, '_ancestrycontext', None)
         return fctx
 
     def parents(self):
         _path = self._path
         fl = self._filelog
         parents = self._filelog.parents(self._filenode)
         pl = [(_path, node, fl) for node in parents if node != nullid]
 
         r = fl.renamed(self._filenode)
         if r:
             # - In the simple rename case, both parent are nullid, pl is empty.
             # - In case of merge, only one of the parent is null id and should
             # be replaced with the rename information. This parent is -always-
             # the first one.
             #
             # As null id have always been filtered out in the previous list
             # comprehension, inserting to 0 will always result in "replacing
             # first nullid parent with rename information.
             pl.insert(0, (r[0], r[1], self._repo.file(r[0])))
 
         return [self._parentfilectx(path, fnode, l) for path, fnode, l in pl]
 
     def p1(self):
         return self.parents()[0]
 
     def p2(self):
         p = self.parents()
         if len(p) == 2:
             return p[1]
         return filectx(self._repo, self._path, fileid=-1, filelog=self._filelog)
 
     def annotate(self, follow=False, linenumber=False, skiprevs=None,
                  diffopts=None):
         '''returns a list of tuples of ((ctx, number), line) for each line
         in the file, where ctx is the filectx of the node where
         that line was last changed; if linenumber parameter is true, number is
         the line number at the first appearance in the managed file, otherwise,
         number has a fixed value of False.
         '''
 
         def lines(text):
             if text.endswith("\n"):
                 return text.count("\n")
             return text.count("\n") + int(bool(text))
 
         if linenumber:
             def decorate(text, rev):
                 return ([annotateline(fctx=rev, lineno=i)
                          for i in xrange(1, lines(text) + 1)], text)
         else:
             def decorate(text, rev):
                 return ([annotateline(fctx=rev)] * lines(text), text)
 
         getlog = util.lrucachefunc(lambda x: self._repo.file(x))
 
         def parents(f):
             # Cut _descendantrev here to mitigate the penalty of lazy linkrev
             # adjustment. Otherwise, p._adjustlinkrev() would walk changelog
             # from the topmost introrev (= srcrev) down to p.linkrev() if it
             # isn't an ancestor of the srcrev.
             f._changeid
             pl = f.parents()
 
             # Don't return renamed parents if we aren't following.
             if not follow:
                 pl = [p for p in pl if p.path() == f.path()]
 
             # renamed filectx won't have a filelog yet, so set it
             # from the cache to save time
             for p in pl:
                 if not r'_filelog' in p.__dict__:
                     p._filelog = getlog(p.path())
 
             return pl
 
         # use linkrev to find the first changeset where self appeared
         base = self.introfilectx()
         if getattr(base, '_ancestrycontext', None) is None:
             cl = self._repo.changelog
             if base.rev() is None:
                 # wctx is not inclusive, but works because _ancestrycontext
                 # is used to test filelog revisions
                 ac = cl.ancestors([p.rev() for p in base.parents()],
                                   inclusive=True)
             else:
                 ac = cl.ancestors([base.rev()], inclusive=True)
             base._ancestrycontext = ac
 
         # This algorithm would prefer to be recursive, but Python is a
         # bit recursion-hostile. Instead we do an iterative
         # depth-first search.
 
         # 1st DFS pre-calculates pcache and needed
         visit = [base]
         pcache = {}
         needed = {base: 1}
         while visit:
             f = visit.pop()
             if f in pcache:
                 continue
             pl = parents(f)
             pcache[f] = pl
             for p in pl:
                 needed[p] = needed.get(p, 0) + 1
                 if p not in pcache:
                     visit.append(p)
 
         # 2nd DFS does the actual annotate
         visit[:] = [base]
         hist = {}
         while visit:
             f = visit[-1]
             if f in hist:
                 visit.pop()
                 continue
 
             ready = True
             pl = pcache[f]
             for p in pl:
                 if p not in hist:
                     ready = False
                     visit.append(p)
             if ready:
                 visit.pop()
                 curr = decorate(f.data(), f)
                 skipchild = False
                 if skiprevs is not None:
                     skipchild = f._changeid in skiprevs
                 curr = _annotatepair([hist[p] for p in pl], f, curr, skipchild,
                                      diffopts)
                 for p in pl:
                     if needed[p] == 1:
                         del hist[p]
                         del needed[p]
                     else:
                         needed[p] -= 1
 
                 hist[f] = curr
                 del pcache[f]
 
-        return pycompat.ziplist(hist[base][0], hist[base][1].splitlines(True))
+        lineattrs, text = hist[base]
+        return pycompat.ziplist(lineattrs, mdiff.splitnewlines(text))
 
     def ancestors(self, followfirst=False):
         visit = {}
         c = self
         if followfirst:
             cut = 1
         else:
             cut = None
 
         while True:
             for parent in c.parents()[:cut]:
                 visit[(parent.linkrev(), parent.filenode())] = parent
             if not visit:
                 break
             c = visit.pop(max(visit))
             yield c
 
     def decodeddata(self):
         """Returns `data()` after running repository decoding filters.
 
         This is often equivalent to how the data would be expressed on disk.
         """
         return self._repo.wwritedata(self.path(), self.data())
 
 @attr.s(slots=True, frozen=True)
 class annotateline(object):
     fctx = attr.ib()
     lineno = attr.ib(default=False)
     # Whether this annotation was the result of a skip-annotate.
     skip = attr.ib(default=False)
 
 def _annotatepair(parents, childfctx, child, skipchild, diffopts):
     r'''
     Given parent and child fctxes and annotate data for parents, for all lines
     in either parent that match the child, annotate the child with the parent's
     data.
 
     Additionally, if `skipchild` is True, replace all other lines with parent
     annotate data as well such that child is never blamed for any lines.
 
     See test-annotate.py for unit tests.
     '''
     pblocks = [(parent, mdiff.allblocks(parent[1], child[1], opts=diffopts))
                for parent in parents]
 
     if skipchild:
         # Need to iterate over the blocks twice -- make it a list
         pblocks = [(p, list(blocks)) for (p, blocks) in pblocks]
     # Mercurial currently prefers p2 over p1 for annotate.
     # TODO: change this?
     for parent, blocks in pblocks:
         for (a1, a2, b1, b2), t in blocks:
             # Changed blocks ('!') or blocks made only of blank lines ('~')
             # belong to the child.
             if t == '=':
                 child[0][b1:b2] = parent[0][a1:a2]
 
     if skipchild:
         # Now try and match up anything that couldn't be matched,
         # Reversing pblocks maintains bias towards p2, matching above
         # behavior.
         pblocks.reverse()
 
         # The heuristics are:
         # * Work on blocks of changed lines (effectively diff hunks with -U0).
         # This could potentially be smarter but works well enough.
         # * For a non-matching section, do a best-effort fit. Match lines in
         #   diff hunks 1:1, dropping lines as necessary.
         # * Repeat the last line as a last resort.
 
         # First, replace as much as possible without repeating the last line.
         remaining = [(parent, []) for parent, _blocks in pblocks]
         for idx, (parent, blocks) in enumerate(pblocks):
             for (a1, a2, b1, b2), _t in blocks:
                 if a2 - a1 >= b2 - b1:
                     for bk in xrange(b1, b2):
                         if child[0][bk].fctx == childfctx:
                             ak = min(a1 + (bk - b1), a2 - 1)
                             child[0][bk] = attr.evolve(parent[0][ak], skip=True)
                 else:
                     remaining[idx][1].append((a1, a2, b1, b2))
 
         # Then, look at anything left, which might involve repeating the last
         # line.
         for parent, blocks in remaining:
             for a1, a2, b1, b2 in blocks:
                 for bk in xrange(b1, b2):
                     if child[0][bk].fctx == childfctx:
                         ak = min(a1 + (bk - b1), a2 - 1)
                         child[0][bk] = attr.evolve(parent[0][ak], skip=True)
     return child
 
 class filectx(basefilectx):
     """A filecontext object makes access to data related to a particular
        filerevision convenient."""
     def __init__(self, repo, path, changeid=None, fileid=None,
                  filelog=None, changectx=None):
         """changeid can be a changeset revision, node, or tag.
            fileid can be a file revision or node."""
         self._repo = repo
         self._path = path
 
         assert (changeid is not None
                 or fileid is not None
                 or changectx is not None), \
                 ("bad args: changeid=%r, fileid=%r, changectx=%r"
                  % (changeid, fileid, changectx))
 
         if filelog is not None:
             self._filelog = filelog
 
         if changeid is not None:
             self._changeid = changeid
         if changectx is not None:
             self._changectx = changectx
         if fileid is not None:
             self._fileid = fileid
 
     @propertycache
     def _changectx(self):
         try:
             return changectx(self._repo, self._changeid)
         except error.FilteredRepoLookupError:
             # Linkrev may point to any revision in the repository.  When the
             # repository is filtered this may lead to `filectx` trying to build
             # `changectx` for filtered revision. In such case we fallback to
             # creating `changectx` on the unfiltered version of the reposition.
             # This fallback should not be an issue because `changectx` from
             # `filectx` are not used in complex operations that care about
             # filtering.
             #
             # This fallback is a cheap and dirty fix that prevent several
             # crashes. It does not ensure the behavior is correct. However the
             # behavior was not correct before filtering either and "incorrect
             # behavior" is seen as better as "crash"
             #
             # Linkrevs have several serious troubles with filtering that are
             # complicated to solve. Proper handling of the issue here should be
             # considered when solving linkrev issue are on the table.
             return changectx(self._repo.unfiltered(), self._changeid)
 
     def filectx(self, fileid, changeid=None):
         '''opens an arbitrary revision of the file without
         opening a new filelog'''
         return filectx(self._repo, self._path, fileid=fileid,
                        filelog=self._filelog, changeid=changeid)
 
     def rawdata(self):
         return self._filelog.revision(self._filenode, raw=True)
 
     def rawflags(self):
         """low-level revlog flags"""
         return self._filelog.flags(self._filerev)
 
     def data(self):
         try:
             return self._filelog.read(self._filenode)
         except error.CensoredNodeError:
             if self._repo.ui.config("censor", "policy") == "ignore":
                 return ""
             raise error.Abort(_("censored node: %s") % short(self._filenode),
                              hint=_("set censor.policy to ignore errors"))
 
     def size(self):
         return self._filelog.size(self._filerev)
 
     @propertycache
     def _copied(self):
         """check if file was actually renamed in this changeset revision
 
         If rename logged in file revision, we report copy for changeset only
         if file revisions linkrev points back to the changeset in question
         or both changeset parents contain different file revisions.
         """
 
         renamed = self._filelog.renamed(self._filenode)
         if not renamed:
             return renamed
 
         if self.rev() == self.linkrev():
             return renamed
 
         name = self.path()
         fnode = self._filenode
         for p in self._changectx.parents():
             try:
                 if fnode == p.filenode(name):
                     return None
             except error.LookupError:
                 pass
         return renamed
 
     def children(self):
         # hard for renames
         c = self._filelog.children(self._filenode)
         return [filectx(self._repo, self._path, fileid=x,
                         filelog=self._filelog) for x in c]
 
 class committablectx(basectx):
     """A committablectx object provides common functionality for a context that
     wants the ability to commit, e.g. workingctx or memctx."""
     def __init__(self, repo, text="", user=None, date=None, extra=None,
                  changes=None):
         self._repo = repo
         self._rev = None
         self._node = None
         self._text = text
         if date:
             self._date = dateutil.parsedate(date)
         if user:
             self._user = user
         if changes:
             self._status = changes
 
         self._extra = {}
         if extra:
             self._extra = extra.copy()
         if 'branch' not in self._extra:
             try:
                 branch = encoding.fromlocal(self._repo.dirstate.branch())
             except UnicodeDecodeError:
                 raise error.Abort(_('branch name not in UTF-8!'))
             self._extra['branch'] = branch
         if self._extra['branch'] == '':
             self._extra['branch'] = 'default'
 
     def __bytes__(self):
         return bytes(self._parents[0]) + "+"
 
     __str__ = encoding.strmethod(__bytes__)
 
     def __nonzero__(self):
         return True
 
     __bool__ = __nonzero__
 
     def _buildflagfunc(self):
         # Create a fallback function for getting file flags when the
         # filesystem doesn't support them
 
         copiesget = self._repo.dirstate.copies().get
         parents = self.parents()
         if len(parents) < 2:
             # when we have one parent, it's easy: copy from parent
             man = parents[0].manifest()
             def func(f):
                 f = copiesget(f, f)
                 return man.flags(f)
         else:
             # merges are tricky: we try to reconstruct the unstored
             # result from the merge (issue1802)
             p1, p2 = parents
             pa = p1.ancestor(p2)
             m1, m2, ma = p1.manifest(), p2.manifest(), pa.manifest()
 
             def func(f):
                 f = copiesget(f, f) # may be wrong for merges with copies
                 fl1, fl2, fla = m1.flags(f), m2.flags(f), ma.flags(f)
                 if fl1 == fl2:
                     return fl1
                 if fl1 == fla:
                     return fl2
                 if fl2 == fla:
                     return fl1
                 return '' # punt for conflicts
 
         return func
 
     @propertycache
     def _flagfunc(self):
         return self._repo.dirstate.flagfunc(self._buildflagfunc)
 
     @propertycache
     def _status(self):
         return self._repo.status()
 
     @propertycache
     def _user(self):
         return self._repo.ui.username()
 
     @propertycache
     def _date(self):
         ui = self._repo.ui
         date = ui.configdate('devel', 'default-date')
         if date is None:
             date = dateutil.makedate()
         return date
 
     def subrev(self, subpath):
         return None
 
     def manifestnode(self):
         return None
     def user(self):
         return self._user or self._repo.ui.username()
     def date(self):
         return self._date
     def description(self):
         return self._text
     def files(self):
         return sorted(self._status.modified + self._status.added +
                       self._status.removed)
 
     def modified(self):
         return self._status.modified
     def added(self):
         return self._status.added
     def removed(self):
         return self._status.removed
     def deleted(self):
         return self._status.deleted
     def branch(self):
         return encoding.tolocal(self._extra['branch'])
     def closesbranch(self):
         return 'close' in self._extra
     def extra(self):
         return self._extra
 
     def isinmemory(self):
         return False
 
     def tags(self):
         return []
 
     def bookmarks(self):
         b = []
         for p in self.parents():
             b.extend(p.bookmarks())
         return b
 
     def phase(self):
         phase = phases.draft # default phase to draft
         for p in self.parents():
             phase = max(phase, p.phase())
         return phase
 
     def hidden(self):
         return False
 
     def children(self):
         return []
 
     def flags(self, path):
         if r'_manifest' in self.__dict__:
             try:
                 return self._manifest.flags(path)
             except KeyError:
                 return ''
 
         try:
             return self._flagfunc(path)
         except OSError:
             return ''
 
     def ancestor(self, c2):
         """return the "best" ancestor context of self and c2"""
         return self._parents[0].ancestor(c2) # punt on two parents for now
 
     def walk(self, match):
         '''Generates matching file names.'''
         return sorted(self._repo.dirstate.walk(match,
                                                subrepos=sorted(self.substate),
                                                unknown=True, ignored=False))
 
     def matches(self, match):
         return sorted(self._repo.dirstate.matches(match))
 
     def ancestors(self):
         for p in self._parents:
             yield p
         for a in self._repo.changelog.ancestors(
             [p.rev() for p in self._parents]):
             yield changectx(self._repo, a)
 
     def markcommitted(self, node):
         """Perform post-commit cleanup necessary after committing this ctx
 
         Specifically, this updates backing stores this working context
         wraps to reflect the fact that the changes reflected by this
         workingctx have been committed.  For example, it marks
         modified and added files as normal in the dirstate.
 
         """
 
         with self._repo.dirstate.parentchange():
             for f in self.modified() + self.added():
                 self._repo.dirstate.normal(f)
             for f in self.removed():
                 self._repo.dirstate.drop(f)
             self._repo.dirstate.setparents(node)
 
         # write changes out explicitly, because nesting wlock at
         # runtime may prevent 'wlock.release()' in 'repo.commit()'
         # from immediately doing so for subsequent changing files
         self._repo.dirstate.write(self._repo.currenttransaction())
 
     def dirty(self, missing=False, merge=True, branch=True):
         return False
 
 class workingctx(committablectx):
     """A workingctx object makes access to data related to
     the current working directory convenient.
     date - any valid date string or (unixtime, offset), or None.
     user - username string, or None.
     extra - a dictionary of extra values, or None.
     changes - a list of file lists as returned by localrepo.status()
                or None to use the repository status.
     """
     def __init__(self, repo, text="", user=None, date=None, extra=None,
                  changes=None):
         super(workingctx, self).__init__(repo, text, user, date, extra, changes)
 
     def __iter__(self):
         d = self._repo.dirstate
         for f in d:
             if d[f] != 'r':
                 yield f
 
     def __contains__(self, key):
         return self._repo.dirstate[key] not in "?r"
 
     def hex(self):
         return hex(wdirid)
 
     @propertycache
     def _parents(self):
         p = self._repo.dirstate.parents()
         if p[1] == nullid:
             p = p[:-1]
         return [changectx(self._repo, x) for x in p]
 
     def filectx(self, path, filelog=None):
         """get a file context from the working directory"""
         return workingfilectx(self._repo, path, workingctx=self,
                               filelog=filelog)
 
     def dirty(self, missing=False, merge=True, branch=True):
         "check whether a working directory is modified"
         # check subrepos first
         for s in sorted(self.substate):
             if self.sub(s).dirty(missing=missing):
                 return True
         # check current working dir
         return ((merge and self.p2()) or
                 (branch and self.branch() != self.p1().branch()) or
                 self.modified() or self.added() or self.removed() or
                 (missing and self.deleted()))
 
     def add(self, list, prefix=""):
         with self._repo.wlock():
             ui, ds = self._repo.ui, self._repo.dirstate
             uipath = lambda f: ds.pathto(pathutil.join(prefix, f))
             rejected = []
             lstat = self._repo.wvfs.lstat
             for f in list:
                 # ds.pathto() returns an absolute file when this is invoked from
                 # the keyword extension.  That gets flagged as non-portable on
                 # Windows, since it contains the drive letter and colon.
                 scmutil.checkportable(ui, os.path.join(prefix, f))
                 try:
                     st = lstat(f)
                 except OSError:
                     ui.warn(_("%s does not exist!\n") % uipath(f))
                     rejected.append(f)
                     continue
                 if st.st_size > 10000000:
                     ui.warn(_("%s: up to %d MB of RAM may be required "
                               "to manage this file\n"
                               "(use 'hg revert %s' to cancel the "
                               "pending addition)\n")
                             % (f, 3 * st.st_size // 1000000, uipath(f)))
                 if not (stat.S_ISREG(st.st_mode) or stat.S_ISLNK(st.st_mode)):
                     ui.warn(_("%s not added: only files and symlinks "
                               "supported currently\n") % uipath(f))
                     rejected.append(f)
                 elif ds[f] in 'amn':
                     ui.warn(_("%s already tracked!\n") % uipath(f))
                 elif ds[f] == 'r':
                     ds.normallookup(f)
                 else:
                     ds.add(f)
             return rejected
 
     def forget(self, files, prefix=""):
         with self._repo.wlock():
             ds = self._repo.dirstate
             uipath = lambda f: ds.pathto(pathutil.join(prefix, f))
             rejected = []
             for f in files:
                 if f not in self._repo.dirstate:
                     self._repo.ui.warn(_("%s not tracked!\n") % uipath(f))
                     rejected.append(f)
                 elif self._repo.dirstate[f] != 'a':
                     self._repo.dirstate.remove(f)
                 else:
                     self._repo.dirstate.drop(f)
             return rejected
 
     def undelete(self, list):
         pctxs = self.parents()
         with self._repo.wlock():
             ds = self._repo.dirstate
             for f in list:
                 if self._repo.dirstate[f] != 'r':
                     self._repo.ui.warn(_("%s not removed!\n") % ds.pathto(f))
                 else:
                     fctx = f in pctxs[0] and pctxs[0][f] or pctxs[1][f]
                     t = fctx.data()
                     self._repo.wwrite(f, t, fctx.flags())
                     self._repo.dirstate.normal(f)
 
     def copy(self, source, dest):
         try:
             st = self._repo.wvfs.lstat(dest)
         except OSError as err:
             if err.errno != errno.ENOENT:
                 raise
             self._repo.ui.warn(_("%s does not exist!\n")
                                % self._repo.dirstate.pathto(dest))
             return
         if not (stat.S_ISREG(st.st_mode) or stat.S_ISLNK(st.st_mode)):
             self._repo.ui.warn(_("copy failed: %s is not a file or a "
                                  "symbolic link\n")
                                % self._repo.dirstate.pathto(dest))
         else:
             with self._repo.wlock():
                 if self._repo.dirstate[dest] in '?':
                     self._repo.dirstate.add(dest)
                 elif self._repo.dirstate[dest] in 'r':
                     self._repo.dirstate.normallookup(dest)
                 self._repo.dirstate.copy(source, dest)
 
     def match(self, pats=None, include=None, exclude=None, default='glob',
               listsubrepos=False, badfn=None):
         r = self._repo
 
         # Only a case insensitive filesystem needs magic to translate user input
         # to actual case in the filesystem.
         icasefs = not util.fscasesensitive(r.root)
         return matchmod.match(r.root, r.getcwd(), pats, include, exclude,
                               default, auditor=r.auditor, ctx=self,
                               listsubrepos=listsubrepos, badfn=badfn,
                               icasefs=icasefs)
 
     def _filtersuspectsymlink(self, files):
         if not files or self._repo.dirstate._checklink:
             return files
 
         # Symlink placeholders may get non-symlink-like contents
         # via user error or dereferencing by NFS or Samba servers,
         # so we filter out any placeholders that don't look like a
         # symlink
         sane = []
         for f in files:
             if self.flags(f) == 'l':
                 d = self[f].data()
                 if d == '' or len(d) >= 1024 or '\n' in d or util.binary(d):
                     self._repo.ui.debug('ignoring suspect symlink placeholder'
                                         ' "%s"\n' % f)
                     continue
             sane.append(f)
         return sane
 
     def _checklookup(self, files):
         # check for any possibly clean files
         if not files:
             return [], [], []
 
         modified = []
         deleted = []
         fixup = []
         pctx = self._parents[0]
         # do a full compare of any files that might have changed
         for f in sorted(files):
             try:
                 # This will return True for a file that got replaced by a
                 # directory in the interim, but fixing that is pretty hard.
                 if (f not in pctx or self.flags(f) != pctx.flags(f)
                     or pctx[f].cmp(self[f])):
                     modified.append(f)
                 else:
                     fixup.append(f)
             except (IOError, OSError):
                 # A file become inaccessible in between? Mark it as deleted,
                 # matching dirstate behavior (issue5584).
                 # The dirstate has more complex behavior around whether a
                 # missing file matches a directory, etc, but we don't need to
                 # bother with that: if f has made it to this point, we're sure
                 # it's in the dirstate.
                 deleted.append(f)
 
         return modified, deleted, fixup
 
     def _poststatusfixup(self, status, fixup):
         """update dirstate for files that are actually clean"""
         poststatus = self._repo.postdsstatus()
         if fixup or poststatus:
             try:
                 oldid = self._repo.dirstate.identity()
 
                 # updating the dirstate is optional
                 # so we don't wait on the lock
                 # wlock can invalidate the dirstate, so cache normal _after_
                 # taking the lock
                 with self._repo.wlock(False):
                     if self._repo.dirstate.identity() == oldid:
                         if fixup:
                             normal = self._repo.dirstate.normal
                             for f in fixup:
                                 normal(f)
                             # write changes out explicitly, because nesting
                             # wlock at runtime may prevent 'wlock.release()'
                             # after this block from doing so for subsequent
                             # changing files
                             tr = self._repo.currenttransaction()
                             self._repo.dirstate.write(tr)
 
                         if poststatus:
                             for ps in poststatus:
                                 ps(self, status)
                     else:
                         # in this case, writing changes out breaks
                         # consistency, because .hg/dirstate was
                         # already changed simultaneously after last
                         # caching (see also issue5584 for detail)
                         self._repo.ui.debug('skip updating dirstate: '
                                             'identity mismatch\n')
             except error.LockError:
                 pass
             finally:
                 # Even if the wlock couldn't be grabbed, clear out the list.
                 self._repo.clearpostdsstatus()
 
     def _dirstatestatus(self, match, ignored=False, clean=False, unknown=False):
         '''Gets the status from the dirstate -- internal use only.'''
         subrepos = []
         if '.hgsub' in self:
             subrepos = sorted(self.substate)
         cmp, s = self._repo.dirstate.status(match, subrepos, ignored=ignored,
                                             clean=clean, unknown=unknown)
 
         # check for any possibly clean files
         fixup = []
         if cmp:
             modified2, deleted2, fixup = self._checklookup(cmp)
             s.modified.extend(modified2)
             s.deleted.extend(deleted2)
 
             if fixup and clean:
                 s.clean.extend(fixup)
 
         self._poststatusfixup(s, fixup)
 
         if match.always():
             # cache for performance
             if s.unknown or s.ignored or s.clean:
                 # "_status" is cached with list*=False in the normal route
                 self._status = scmutil.status(s.modified, s.added, s.removed,
                                               s.deleted, [], [], [])
             else:
                 self._status = s
 
         return s
 
     @propertycache
     def _manifest(self):
         """generate a manifest corresponding to the values in self._status
 
         This reuse the file nodeid from parent, but we use special node
         identifiers for added and modified files. This is used by manifests
         merge to see that files are different and by update logic to avoid
         deleting newly added files.
         """
         return self._buildstatusmanifest(self._status)
 
     def _buildstatusmanifest(self, status):
         """Builds a manifest that includes the given status results."""
         parents = self.parents()
 
         man = parents[0].manifest().copy()
 
         ff = self._flagfunc
         for i, l in ((addednodeid, status.added),
                      (modifiednodeid, status.modified)):
             for f in l:
                 man[f] = i
                 try:
                     man.setflag(f, ff(f))
                 except OSError:
                     pass
 
         for f in status.deleted + status.removed:
             if f in man:
                 del man[f]
 
         return man
 
     def _buildstatus(self, other, s, match, listignored, listclean,
                      listunknown):
         """build a status with respect to another context
 
         This includes logic for maintaining the fast path of status when
         comparing the working directory against its parent, which is to skip
         building a new manifest if self (working directory) is not comparing
         against its parent (repo['.']).
         """
         s = self._dirstatestatus(match, listignored, listclean, listunknown)
         # Filter out symlinks that, in the case of FAT32 and NTFS filesystems,
         # might have accidentally ended up with the entire contents of the file
         # they are supposed to be linking to.
         s.modified[:] = self._filtersuspectsymlink(s.modified)
         if other != self._repo['.']:
             s = super(workingctx, self)._buildstatus(other, s, match,
                                                      listignored, listclean,
                                                      listunknown)
         return s
 
     def _matchstatus(self, other, match):
         """override the match method with a filter for directory patterns
 
         We use inheritance to customize the match.bad method only in cases of
         workingctx since it belongs only to the working directory when
         comparing against the parent changeset.
 
         If we aren't comparing against the working directory's parent, then we
         just use the default match object sent to us.
         """
         if other != self._repo['.']:
             def bad(f, msg):
                 # 'f' may be a directory pattern from 'match.files()',
                 # so 'f not in ctx1' is not enough
                 if f not in other and not other.hasdir(f):
                     self._repo.ui.warn('%s: %s\n' %
                                        (self._repo.dirstate.pathto(f), msg))
             match.bad = bad
         return match
 
     def markcommitted(self, node):
         super(workingctx, self).markcommitted(node)
 
         sparse.aftercommit(self._repo, node)
 
 class committablefilectx(basefilectx):
     """A committablefilectx provides common functionality for a file context
     that wants the ability to commit, e.g. workingfilectx or memfilectx."""
     def __init__(self, repo, path, filelog=None, ctx=None):
         self._repo = repo
         self._path = path
         self._changeid = None
         self._filerev = self._filenode = None
 
         if filelog is not None:
             self._filelog = filelog
         if ctx:
             self._changectx = ctx
 
     def __nonzero__(self):
         return True
 
     __bool__ = __nonzero__
 
     def linkrev(self):
         # linked to self._changectx no matter if file is modified or not
         return self.rev()
 
     def parents(self):
         '''return parent filectxs, following copies if necessary'''
         def filenode(ctx, path):
             return ctx._manifest.get(path, nullid)
 
         path = self._path
         fl = self._filelog
         pcl = self._changectx._parents
         renamed = self.renamed()
 
         if renamed:
             pl = [renamed + (None,)]
         else:
             pl = [(path, filenode(pcl[0], path), fl)]
 
         for pc in pcl[1:]:
             pl.append((path, filenode(pc, path), fl))
 
         return [self._parentfilectx(p, fileid=n, filelog=l)
                 for p, n, l in pl if n != nullid]
 
     def children(self):
         return []
 
 class workingfilectx(committablefilectx):
     """A workingfilectx object makes access to data related to a particular
        file in the working directory convenient."""
     def __init__(self, repo, path, filelog=None, workingctx=None):
         super(workingfilectx, self).__init__(repo, path, filelog, workingctx)
 
     @propertycache
     def _changectx(self):
         return workingctx(self._repo)
 
     def data(self):
         return self._repo.wread(self._path)
     def renamed(self):
         rp = self._repo.dirstate.copied(self._path)
         if not rp:
             return None
         return rp, self._changectx._parents[0]._manifest.get(rp, nullid)
 
     def size(self):
         return self._repo.wvfs.lstat(self._path).st_size
     def date(self):
         t, tz = self._changectx.date()
         try:
             return (self._repo.wvfs.lstat(self._path).st_mtime, tz)
         except OSError as err:
             if err.errno != errno.ENOENT:
                 raise
             return (t, tz)
 
     def exists(self):
         return self._repo.wvfs.exists(self._path)
 
     def lexists(self):
         return self._repo.wvfs.lexists(self._path)
 
     def audit(self):
         return self._repo.wvfs.audit(self._path)
 
     def cmp(self, fctx):
         """compare with other file context
 
         returns True if different than fctx.
         """
         # fctx should be a filectx (not a workingfilectx)
         # invert comparison to reuse the same code path
         return fctx.cmp(self)
 
     def remove(self, ignoremissing=False):
         """wraps unlink for a repo's working directory"""
         self._repo.wvfs.unlinkpath(self._path, ignoremissing=ignoremissing)
 
     def write(self, data, flags, backgroundclose=False, **kwargs):
         """wraps repo.wwrite"""
         self._repo.wwrite(self._path, data, flags,
                           backgroundclose=backgroundclose,
                           **kwargs)
 
     def markcopied(self, src):
         """marks this file a copy of `src`"""
         if self._repo.dirstate[self._path] in "nma":
             self._repo.dirstate.copy(src, self._path)
 
     def clearunknown(self):
         """Removes conflicting items in the working directory so that
         ``write()`` can be called successfully.
         """
         wvfs = self._repo.wvfs
         f = self._path
         wvfs.audit(f)
         if wvfs.isdir(f) and not wvfs.islink(f):
             wvfs.rmtree(f, forcibly=True)
         for p in reversed(list(util.finddirs(f))):
             if wvfs.isfileorlink(p):
                 wvfs.unlink(p)
                 break
 
     def setflags(self, l, x):
         self._repo.wvfs.setflags(self._path, l, x)
 
 class overlayworkingctx(committablectx):
     """Wraps another mutable context with a write-back cache that can be
     converted into a commit context.
 
     self._cache[path] maps to a dict with keys: {
         'exists': bool?
         'date': date?
         'data': str?
         'flags': str?
         'copied': str? (path or None)
     }
     If `exists` is True, `flags` must be non-None and 'date' is non-None. If it
     is `False`, the file was deleted.
     """
 
     def __init__(self, repo):
         super(overlayworkingctx, self).__init__(repo)
         self._repo = repo
         self.clean()
 
     def setbase(self, wrappedctx):
         self._wrappedctx = wrappedctx
         self._parents = [wrappedctx]
         # Drop old manifest cache as it is now out of date.
         # This is necessary when, e.g., rebasing several nodes with one
         # ``overlayworkingctx`` (e.g. with --collapse).
         util.clearcachedproperty(self, '_manifest')
 
     def data(self, path):
         if self.isdirty(path):
             if self._cache[path]['exists']:
                 if self._cache[path]['data']:
                     return self._cache[path]['data']
                 else:
                     # Must fallback here, too, because we only set flags.
                     return self._wrappedctx[path].data()
             else:
                 raise error.ProgrammingError("No such file or directory: %s" %
                                              path)
         else:
             return self._wrappedctx[path].data()
 
     @propertycache
     def _manifest(self):
         parents = self.parents()
         man = parents[0].manifest().copy()
 
         flag = self._flagfunc
         for path in self.added():
             man[path] = addednodeid
             man.setflag(path, flag(path))
         for path in self.modified():
             man[path] = modifiednodeid
             man.setflag(path, flag(path))
         for path in self.removed():
             del man[path]
         return man
 
     @propertycache
     def _flagfunc(self):
         def f(path):
             return self._cache[path]['flags']
         return f
 
     def files(self):
         return sorted(self.added() + self.modified() + self.removed())
 
     def modified(self):
         return [f for f in self._cache.keys() if self._cache[f]['exists'] and
                 self._existsinparent(f)]
 
     def added(self):
         return [f for f in self._cache.keys() if self._cache[f]['exists'] and
                 not self._existsinparent(f)]
 
     def removed(self):
         return [f for f in self._cache.keys() if
                 not self._cache[f]['exists'] and self._existsinparent(f)]
 
     def isinmemory(self):
         return True
 
     def filedate(self, path):
         if self.isdirty(path):
             return self._cache[path]['date']
         else:
             return self._wrappedctx[path].date()
 
     def markcopied(self, path, origin):
         if self.isdirty(path):
             self._cache[path]['copied'] = origin
         else:
             raise error.ProgrammingError('markcopied() called on clean context')
 
     def copydata(self, path):
         if self.isdirty(path):
             return self._cache[path]['copied']
         else:
             raise error.ProgrammingError('copydata() called on clean context')
 
     def flags(self, path):
         if self.isdirty(path):
             if self._cache[path]['exists']:
                 return self._cache[path]['flags']
             else:
                 raise error.ProgrammingError("No such file or directory: %s" %
                                              self._path)
         else:
             return self._wrappedctx[path].flags()
 
     def _existsinparent(self, path):
         try:
             # ``commitctx` raises a ``ManifestLookupError`` if a path does not
             # exist, unlike ``workingctx``, which returns a ``workingfilectx``
             # with an ``exists()`` function.
             self._wrappedctx[path]
             return True
         except error.ManifestLookupError:
             return False
 
     def _auditconflicts(self, path):
         """Replicates conflict checks done by wvfs.write().
 
         Since we never write to the filesystem and never call `applyupdates` in
         IMM, we'll never check that a path is actually writable -- e.g., because
         it adds `a/foo`, but `a` is actually a file in the other commit.
         """
         def fail(path, component):
             # p1() is the base and we're receiving "writes" for p2()'s
             # files.
             if 'l' in self.p1()[component].flags():
                 raise error.Abort("error: %s conflicts with symlink %s "
                                   "in %s." % (path, component,
                                               self.p1().rev()))
             else:
                 raise error.Abort("error: '%s' conflicts with file '%s' in "
                                   "%s." % (path, component,
                                            self.p1().rev()))
 
         # Test that each new directory to be created to write this path from p2
         # is not a file in p1.
         components = path.split('/')
         for i in xrange(len(components)):
             component = "/".join(components[0:i])
             if component in self.p1():
                 fail(path, component)
 
         # Test the other direction -- that this path from p2 isn't a directory
         # in p1 (test that p1 doesn't any paths matching `path/*`).
         match = matchmod.match('/', '', [path + '/'], default=b'relpath')
         matches = self.p1().manifest().matches(match)
         if len(matches) > 0:
             if len(matches) == 1 and matches.keys()[0] == path:
                 return
             raise error.Abort("error: file '%s' cannot be written because "
                               " '%s/' is a folder in %s (containing %d "
                               "entries: %s)"
                               % (path, path, self.p1(), len(matches),
                                  ', '.join(matches.keys())))
 
     def write(self, path, data, flags='', **kwargs):
         if data is None:
             raise error.ProgrammingError("data must be non-None")
         self._auditconflicts(path)
         self._markdirty(path, exists=True, data=data, date=dateutil.makedate(),
                         flags=flags)
 
     def setflags(self, path, l, x):
         self._markdirty(path, exists=True, date=dateutil.makedate(),
                         flags=(l and 'l' or '') + (x and 'x' or ''))
 
     def remove(self, path):
         self._markdirty(path, exists=False)
 
     def exists(self, path):
         """exists behaves like `lexists`, but needs to follow symlinks and
         return False if they are broken.
         """
         if self.isdirty(path):
             # If this path exists and is a symlink, "follow" it by calling
             # exists on the destination path.
             if (self._cache[path]['exists'] and
                         'l' in self._cache[path]['flags']):
                 return self.exists(self._cache[path]['data'].strip())
             else:
                 return self._cache[path]['exists']
 
         return self._existsinparent(path)
 
     def lexists(self, path):
         """lexists returns True if the path exists"""
         if self.isdirty(path):
             return self._cache[path]['exists']
 
         return self._existsinparent(path)
 
     def size(self, path):
         if self.isdirty(path):
             if self._cache[path]['exists']:
                 return len(self._cache[path]['data'])
             else:
                 raise error.ProgrammingError("No such file or directory: %s" %
                                              self._path)
         return self._wrappedctx[path].size()
 
     def tomemctx(self, text, branch=None, extra=None, date=None, parents=None,
                  user=None, editor=None):
         """Converts this ``overlayworkingctx`` into a ``memctx`` ready to be
         committed.
 
         ``text`` is the commit message.
         ``parents`` (optional) are rev numbers.
         """
         # Default parents to the wrapped contexts' if not passed.
         if parents is None:
             parents = self._wrappedctx.parents()
             if len(parents) == 1:
                 parents = (parents[0], None)
 
         # ``parents`` is passed as rev numbers; convert to ``commitctxs``.
         if parents[1] is None:
             parents = (self._repo[parents[0]], None)
         else:
             parents = (self._repo[parents[0]], self._repo[parents[1]])
 
         files = self._cache.keys()
         def getfile(repo, memctx, path):
             if self._cache[path]['exists']:
                 return memfilectx(repo, memctx, path,
                                   self._cache[path]['data'],
                                   'l' in self._cache[path]['flags'],
                                   'x' in self._cache[path]['flags'],
                                   self._cache[path]['copied'])
             else:
                 # Returning None, but including the path in `files`, is
                 # necessary for memctx to register a deletion.
                 return None
         return memctx(self._repo, parents, text, files, getfile, date=date,
                       extra=extra, user=user, branch=branch, editor=editor)
 
     def isdirty(self, path):
         return path in self._cache
 
     def isempty(self):
         # We need to discard any keys that are actually clean before the empty
         # commit check.
         self._compact()
         return len(self._cache) == 0
 
     def clean(self):
         self._cache = {}
 
     def _compact(self):
         """Removes keys from the cache that are actually clean, by comparing
         them with the underlying context.
 
         This can occur during the merge process, e.g. by passing --tool :local
         to resolve a conflict.
         """
         keys = []
         for path in self._cache.keys():
             cache = self._cache[path]
             try:
                 underlying = self._wrappedctx[path]
                 if (underlying.data() == cache['data'] and
                             underlying.flags() == cache['flags']):
                     keys.append(path)
             except error.ManifestLookupError:
                 # Path not in the underlying manifest (created).
                 continue
 
         for path in keys:
             del self._cache[path]
         return keys
 
     def _markdirty(self, path, exists, data=None, date=None, flags=''):
         self._cache[path] = {
             'exists': exists,
             'data': data,
             'date': date,
             'flags': flags,
             'copied': None,
         }
 
     def filectx(self, path, filelog=None):
         return overlayworkingfilectx(self._repo, path, parent=self,
                                      filelog=filelog)
 
 class overlayworkingfilectx(committablefilectx):
     """Wrap a ``workingfilectx`` but intercepts all writes into an in-memory
     cache, which can be flushed through later by calling ``flush()``."""
 
     def __init__(self, repo, path, filelog=None, parent=None):
         super(overlayworkingfilectx, self).__init__(repo, path, filelog,
                                                     parent)
         self._repo = repo
         self._parent = parent
         self._path = path
 
     def cmp(self, fctx):
         return self.data() != fctx.data()
 
     def changectx(self):
         return self._parent
 
     def data(self):
         return self._parent.data(self._path)
 
     def date(self):
         return self._parent.filedate(self._path)
 
     def exists(self):
         return self.lexists()
 
     def lexists(self):
         return self._parent.exists(self._path)
 
     def renamed(self):
         path = self._parent.copydata(self._path)
         if not path:
             return None
         return path, self._changectx._parents[0]._manifest.get(path, nullid)
 
     def size(self):
         return self._parent.size(self._path)
 
     def markcopied(self, origin):
         self._parent.markcopied(self._path, origin)
 
     def audit(self):
         pass
 
     def flags(self):
         return self._parent.flags(self._path)
 
     def setflags(self, islink, isexec):
         return self._parent.setflags(self._path, islink, isexec)
 
     def write(self, data, flags, backgroundclose=False, **kwargs):
         return self._parent.write(self._path, data, flags, **kwargs)
 
     def remove(self, ignoremissing=False):
         return self._parent.remove(self._path)
 
     def clearunknown(self):
         pass
 
 class workingcommitctx(workingctx):
     """A workingcommitctx object makes access to data related to
     the revision being committed convenient.
 
     This hides changes in the working directory, if they aren't
     committed in this context.
     """
     def __init__(self, repo, changes,
                  text="", user=None, date=None, extra=None):
         super(workingctx, self).__init__(repo, text, user, date, extra,
                                          changes)
 
     def _dirstatestatus(self, match, ignored=False, clean=False, unknown=False):
         """Return matched files only in ``self._status``
 
         Uncommitted files appear "clean" via this context, even if
         they aren't actually so in the working directory.
         """
         if clean:
             clean = [f for f in self._manifest if f not in self._changedset]
         else:
             clean = []
         return scmutil.status([f for f in self._status.modified if match(f)],
                               [f for f in self._status.added if match(f)],
                               [f for f in self._status.removed if match(f)],
                               [], [], [], clean)
 
     @propertycache
     def _changedset(self):
         """Return the set of files changed in this context
         """
         changed = set(self._status.modified)
         changed.update(self._status.added)
         changed.update(self._status.removed)
         return changed
 
 def makecachingfilectxfn(func):
     """Create a filectxfn that caches based on the path.
 
     We can't use util.cachefunc because it uses all arguments as the cache
     key and this creates a cycle since the arguments include the repo and
     memctx.
     """
     cache = {}
 
     def getfilectx(repo, memctx, path):
         if path not in cache:
             cache[path] = func(repo, memctx, path)
         return cache[path]
 
     return getfilectx
 
 def memfilefromctx(ctx):
     """Given a context return a memfilectx for ctx[path]
 
     This is a convenience method for building a memctx based on another
     context.
     """
     def getfilectx(repo, memctx, path):
         fctx = ctx[path]
         # this is weird but apparently we only keep track of one parent
         # (why not only store that instead of a tuple?)
         copied = fctx.renamed()
         if copied:
             copied = copied[0]
         return memfilectx(repo, memctx, path, fctx.data(),
                           islink=fctx.islink(), isexec=fctx.isexec(),
                           copied=copied)
 
     return getfilectx
 
 def memfilefrompatch(patchstore):
     """Given a patch (e.g. patchstore object) return a memfilectx
 
     This is a convenience method for building a memctx based on a patchstore.
     """
     def getfilectx(repo, memctx, path):
         data, mode, copied = patchstore.getfile(path)
         if data is None:
             return None
         islink, isexec = mode
         return memfilectx(repo, memctx, path, data, islink=islink,
                           isexec=isexec, copied=copied)
 
     return getfilectx
 
 class memctx(committablectx):
     """Use memctx to perform in-memory commits via localrepo.commitctx().
 
     Revision information is supplied at initialization time while
     related files data and is made available through a callback
     mechanism.  'repo' is the current localrepo, 'parents' is a
     sequence of two parent revisions identifiers (pass None for every
     missing parent), 'text' is the commit message and 'files' lists
     names of files touched by the revision (normalized and relative to
     repository root).
 
     filectxfn(repo, memctx, path) is a callable receiving the
     repository, the current memctx object and the normalized path of
     requested file, relative to repository root. It is fired by the
     commit function for every file in 'files', but calls order is
     undefined. If the file is available in the revision being
     committed (updated or added), filectxfn returns a memfilectx
     object. If the file was removed, filectxfn return None for recent
     Mercurial. Moved files are represented by marking the source file
     removed and the new file added with copy information (see
     memfilectx).
 
     user receives the committer name and defaults to current
     repository username, date is the commit date in any format
     supported by dateutil.parsedate() and defaults to current date, extra
     is a dictionary of metadata or is left empty.
     """
 
     # Mercurial <= 3.1 expects the filectxfn to raise IOError for missing files.
     # Extensions that need to retain compatibility across Mercurial 3.1 can use
     # this field to determine what to do in filectxfn.
     _returnnoneformissingfiles = True
 
     def __init__(self, repo, parents, text, files, filectxfn, user=None,
                  date=None, extra=None, branch=None, editor=False):
         super(memctx, self).__init__(repo, text, user, date, extra)
         self._rev = None
         self._node = None
         parents = [(p or nullid) for p in parents]
         p1, p2 = parents
         self._parents = [changectx(self._repo, p) for p in (p1, p2)]
         files = sorted(set(files))
         self._files = files
         if branch is not None:
             self._extra['branch'] = encoding.fromlocal(branch)
         self.substate = {}
 
         if isinstance(filectxfn, patch.filestore):
             filectxfn = memfilefrompatch(filectxfn)
         elif not callable(filectxfn):
             # if store is not callable, wrap it in a function
             filectxfn = memfilefromctx(filectxfn)
 
         # memoizing increases performance for e.g. vcs convert scenarios.
         self._filectxfn = makecachingfilectxfn(filectxfn)
 
         if editor:
             self._text = editor(self._repo, self, [])
             self._repo.savecommitmessage(self._text)
 
     def filectx(self, path, filelog=None):
         """get a file context from the working directory
 
         Returns None if file doesn't exist and should be removed."""
         return self._filectxfn(self._repo, self, path)
 
     def commit(self):
         """commit context to the repo"""
         return self._repo.commitctx(self)
 
     @propertycache
     def _manifest(self):
         """generate a manifest based on the return values of filectxfn"""
 
         # keep this simple for now; just worry about p1
         pctx = self._parents[0]
         man = pctx.manifest().copy()
 
         for f in self._status.modified:
             p1node = nullid
             p2node = nullid
             p = pctx[f].parents() # if file isn't in pctx, check p2?
             if len(p) > 0:
                 p1node = p[0].filenode()
                 if len(p) > 1:
                     p2node = p[1].filenode()
             man[f] = revlog.hash(self[f].data(), p1node, p2node)
 
         for f in self._status.added:
             man[f] = revlog.hash(self[f].data(), nullid, nullid)
 
         for f in self._status.removed:
             if f in man:
                 del man[f]
 
         return man
 
     @propertycache
     def _status(self):
         """Calculate exact status from ``files`` specified at construction
         """
         man1 = self.p1().manifest()
         p2 = self._parents[1]
         # "1 < len(self._parents)" can't be used for checking
         # existence of the 2nd parent, because "memctx._parents" is
         # explicitly initialized by the list, of which length is 2.
         if p2.node() != nullid:
             man2 = p2.manifest()
             managing = lambda f: f in man1 or f in man2
         else:
             managing = lambda f: f in man1
 
         modified, added, removed = [], [], []
         for f in self._files:
             if not managing(f):
                 added.append(f)
             elif self[f]:
                 modified.append(f)
             else:
                 removed.append(f)
 
         return scmutil.status(modified, added, removed, [], [], [], [])
 
 class memfilectx(committablefilectx):
     """memfilectx represents an in-memory file to commit.
 
     See memctx and committablefilectx for more details.
     """
     def __init__(self, repo, changectx, path, data, islink=False,
                  isexec=False, copied=None):
         """
         path is the normalized file path relative to repository root.
         data is the file content as a string.
         islink is True if the file is a symbolic link.
         isexec is True if the file is executable.
         copied is the source file path if current file was copied in the
         revision being committed, or None."""
         super(memfilectx, self).__init__(repo, path, None, changectx)
         self._data = data
         self._flags = (islink and 'l' or '') + (isexec and 'x' or '')
         self._copied = None
         if copied:
             self._copied = (copied, nullid)
 
     def data(self):
         return self._data
 
     def remove(self, ignoremissing=False):
         """wraps unlink for a repo's working directory"""
         # need to figure out what to do here
         del self._changectx[self._path]
 
     def write(self, data, flags, **kwargs):
         """wraps repo.wwrite"""
         self._data = data
 
 class overlayfilectx(committablefilectx):
     """Like memfilectx but take an original filectx and optional parameters to
     override parts of it. This is useful when fctx.data() is expensive (i.e.
     flag processor is expensive) and raw data, flags, and filenode could be
     reused (ex. rebase or mode-only amend a REVIDX_EXTSTORED file).
     """
 
     def __init__(self, originalfctx, datafunc=None, path=None, flags=None,
                  copied=None, ctx=None):
         """originalfctx: filecontext to duplicate
 
         datafunc: None or a function to override data (file content). It is a
         function to be lazy. path, flags, copied, ctx: None or overridden value
 
         copied could be (path, rev), or False. copied could also be just path,
         and will be converted to (path, nullid). This simplifies some callers.
         """
 
         if path is None:
             path = originalfctx.path()
         if ctx is None:
             ctx = originalfctx.changectx()
             ctxmatch = lambda: True
         else:
             ctxmatch = lambda: ctx == originalfctx.changectx()
 
         repo = originalfctx.repo()
         flog = originalfctx.filelog()
         super(overlayfilectx, self).__init__(repo, path, flog, ctx)
 
         if copied is None:
             copied = originalfctx.renamed()
             copiedmatch = lambda: True
         else:
             if copied and not isinstance(copied, tuple):
                 # repo._filecommit will recalculate copyrev so nullid is okay
                 copied = (copied, nullid)
             copiedmatch = lambda: copied == originalfctx.renamed()
 
         # When data, copied (could affect data), ctx (could affect filelog
         # parents) are not overridden, rawdata, rawflags, and filenode may be
         # reused (repo._filecommit should double check filelog parents).
         #
         # path, flags are not hashed in filelog (but in manifestlog) so they do
         # not affect reusable here.
         #
         # If ctx or copied is overridden to a same value with originalfctx,
         # still consider it's reusable. originalfctx.renamed() may be a bit
         # expensive so it's not called unless necessary. Assuming datafunc is
         # always expensive, do not call it for this "reusable" test.
         reusable = datafunc is None and ctxmatch() and copiedmatch()
 
         if datafunc is None:
             datafunc = originalfctx.data
         if flags is None:
             flags = originalfctx.flags()
 
         self._datafunc = datafunc
         self._flags = flags
         self._copied = copied
 
         if reusable:
             # copy extra fields from originalfctx
             attrs = ['rawdata', 'rawflags', '_filenode', '_filerev']
             for attr_ in attrs:
                 if util.safehasattr(originalfctx, attr_):
                     setattr(self, attr_, getattr(originalfctx, attr_))
 
     def data(self):
         return self._datafunc()
 
 class metadataonlyctx(committablectx):
     """Like memctx but it's reusing the manifest of different commit.
     Intended to be used by lightweight operations that are creating
     metadata-only changes.
 
     Revision information is supplied at initialization time.  'repo' is the
     current localrepo, 'ctx' is original revision which manifest we're reuisng
     'parents' is a sequence of two parent revisions identifiers (pass None for
     every missing parent), 'text' is the commit.
 
     user receives the committer name and defaults to current repository
     username, date is the commit date in any format supported by
     dateutil.parsedate() and defaults to current date, extra is a dictionary of
     metadata or is left empty.
     """
     def __new__(cls, repo, originalctx, *args, **kwargs):
         return super(metadataonlyctx, cls).__new__(cls, repo)
 
     def __init__(self, repo, originalctx, parents=None, text=None, user=None,
                  date=None, extra=None, editor=False):
         if text is None:
             text = originalctx.description()
         super(metadataonlyctx, self).__init__(repo, text, user, date, extra)
         self._rev = None
         self._node = None
         self._originalctx = originalctx
         self._manifestnode = originalctx.manifestnode()
         if parents is None:
             parents = originalctx.parents()
         else:
             parents = [repo[p] for p in parents if p is not None]
         parents = parents[:]
         while len(parents) < 2:
             parents.append(repo[nullid])
         p1, p2 = self._parents = parents
 
         # sanity check to ensure that the reused manifest parents are
         # manifests of our commit parents
         mp1, mp2 = self.manifestctx().parents
         if p1 != nullid and p1.manifestnode() != mp1:
             raise RuntimeError('can\'t reuse the manifest: '
                                'its p1 doesn\'t match the new ctx p1')
         if p2 != nullid and p2.manifestnode() != mp2:
             raise RuntimeError('can\'t reuse the manifest: '
                                'its p2 doesn\'t match the new ctx p2')
 
         self._files = originalctx.files()
         self.substate = {}
 
         if editor:
             self._text = editor(self._repo, self, [])
             self._repo.savecommitmessage(self._text)
 
     def manifestnode(self):
         return self._manifestnode
 
     @property
     def _manifestctx(self):
         return self._repo.manifestlog[self._manifestnode]
 
     def filectx(self, path, filelog=None):
         return self._originalctx.filectx(path, filelog=filelog)
 
     def commit(self):
         """commit context to the repo"""
         return self._repo.commitctx(self)
 
     @property
     def _manifest(self):
         return self._originalctx.manifest()
 
     @propertycache
     def _status(self):
         """Calculate exact status from ``files`` specified in the ``origctx``
         and parents manifests.
         """
         man1 = self.p1().manifest()
         p2 = self._parents[1]
         # "1 < len(self._parents)" can't be used for checking
         # existence of the 2nd parent, because "metadataonlyctx._parents" is
         # explicitly initialized by the list, of which length is 2.
         if p2.node() != nullid:
             man2 = p2.manifest()
             managing = lambda f: f in man1 or f in man2
         else:
             managing = lambda f: f in man1
 
         modified, added, removed = [], [], []
         for f in self._files:
             if not managing(f):
                 added.append(f)
             elif f in self:
                 modified.append(f)
             else:
                 removed.append(f)
 
         return scmutil.status(modified, added, removed, [], [], [], [])
 
 class arbitraryfilectx(object):
     """Allows you to use filectx-like functions on a file in an arbitrary
     location on disk, possibly not in the working directory.
     """
     def __init__(self, path, repo=None):
         # Repo is optional because contrib/simplemerge uses this class.
         self._repo = repo
         self._path = path
 
     def cmp(self, fctx):
         # filecmp follows symlinks whereas `cmp` should not, so skip the fast
         # path if either side is a symlink.
         symlinks = ('l' in self.flags() or 'l' in fctx.flags())
         if not symlinks and isinstance(fctx, workingfilectx) and self._repo:
             # Add a fast-path for merge if both sides are disk-backed.
             # Note that filecmp uses the opposite return values (True if same)
             # from our cmp functions (True if different).
             return not filecmp.cmp(self.path(), self._repo.wjoin(fctx.path()))
         return self.data() != fctx.data()
 
     def path(self):
         return self._path
 
     def flags(self):
         return ''
 
     def data(self):
         return util.readfile(self._path)
 
     def decodeddata(self):
         with open(self._path, "rb") as f:
             return f.read()
 
     def remove(self):
         util.unlink(self._path)
 
     def write(self, data, flags, **kwargs):
         assert not flags
         with open(self._path, "w") as f:
             f.write(data)

mercurial/subrepo.py

@@ -1,1788 +1,1808 b''
 # subrepo.py - sub-repository classes and factory
 #
 # Copyright 2009-2010 Matt Mackall <[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 copy
 import errno
 import hashlib
 import os
 import posixpath
 import re
 import stat
 import subprocess
 import sys
 import tarfile
 import xml.dom.minidom
 
 from .i18n import _
 from . import (
     cmdutil,
     encoding,
     error,
     exchange,
     logcmdutil,
     match as matchmod,
     node,
     pathutil,
     phases,
     pycompat,
     scmutil,
     subrepoutil,
     util,
     vfs as vfsmod,
 )
 from .utils import dateutil
 
 hg = None
 reporelpath = subrepoutil.reporelpath
 subrelpath = subrepoutil.subrelpath
 _abssource = subrepoutil._abssource
 propertycache = util.propertycache
 
 def _expandedabspath(path):
     '''
     get a path or url and if it is a path expand it and return an absolute path
     '''
     expandedpath = util.urllocalpath(util.expandpath(path))
     u = util.url(expandedpath)
     if not u.scheme:
         path = util.normpath(os.path.abspath(u.path))
     return path
 
 def _getstorehashcachename(remotepath):
     '''get a unique filename for the store hash cache of a remote repository'''
     return node.hex(hashlib.sha1(_expandedabspath(remotepath)).digest())[0:12]
 
 class SubrepoAbort(error.Abort):
     """Exception class used to avoid handling a subrepo error more than once"""
     def __init__(self, *args, **kw):
         self.subrepo = kw.pop(r'subrepo', None)
         self.cause = kw.pop(r'cause', None)
         error.Abort.__init__(self, *args, **kw)
 
 def annotatesubrepoerror(func):
     def decoratedmethod(self, *args, **kargs):
         try:
             res = func(self, *args, **kargs)
         except SubrepoAbort as ex:
             # This exception has already been handled
             raise ex
         except error.Abort as ex:
             subrepo = subrelpath(self)
             errormsg = (util.forcebytestr(ex) + ' '
                         + _('(in subrepository "%s")') % subrepo)
             # avoid handling this exception by raising a SubrepoAbort exception
             raise SubrepoAbort(errormsg, hint=ex.hint, subrepo=subrepo,
                                cause=sys.exc_info())
         return res
     return decoratedmethod
 
 def _updateprompt(ui, sub, dirty, local, remote):
     if dirty:
         msg = (_(' subrepository sources for %s differ\n'
                  'use (l)ocal source (%s) or (r)emote source (%s)?'
                  '$$ &Local $$ &Remote')
                % (subrelpath(sub), local, remote))
     else:
         msg = (_(' subrepository sources for %s differ (in checked out '
                  'version)\n'
                  'use (l)ocal source (%s) or (r)emote source (%s)?'
                  '$$ &Local $$ &Remote')
                % (subrelpath(sub), local, remote))
     return ui.promptchoice(msg, 0)
 
 def _sanitize(ui, vfs, ignore):
     for dirname, dirs, names in vfs.walk():
         for i, d in enumerate(dirs):
             if d.lower() == ignore:
                 del dirs[i]
                 break
         if vfs.basename(dirname).lower() != '.hg':
             continue
         for f in names:
             if f.lower() == 'hgrc':
                 ui.warn(_("warning: removing potentially hostile 'hgrc' "
                           "in '%s'\n") % vfs.join(dirname))
                 vfs.unlink(vfs.reljoin(dirname, f))
 
 def _auditsubrepopath(repo, path):
     # auditor doesn't check if the path itself is a symlink
     pathutil.pathauditor(repo.root)(path)
     if repo.wvfs.islink(path):
         raise error.Abort(_("subrepo '%s' traverses symbolic link") % path)
 
 SUBREPO_ALLOWED_DEFAULTS = {
     'hg': True,
     'git': False,
     'svn': False,
 }
 
 def _checktype(ui, kind):
     # subrepos.allowed is a master kill switch. If disabled, subrepos are
     # disabled period.
     if not ui.configbool('subrepos', 'allowed', True):
         raise error.Abort(_('subrepos not enabled'),
                           hint=_("see 'hg help config.subrepos' for details"))
 
     default = SUBREPO_ALLOWED_DEFAULTS.get(kind, False)
     if not ui.configbool('subrepos', '%s:allowed' % kind, default):
         raise error.Abort(_('%s subrepos not allowed') % kind,
                           hint=_("see 'hg help config.subrepos' for details"))
 
     if kind not in types:
         raise error.Abort(_('unknown subrepo type %s') % kind)
 
 def subrepo(ctx, path, allowwdir=False, allowcreate=True):
     """return instance of the right subrepo class for subrepo in path"""
     # subrepo inherently violates our import layering rules
     # because it wants to make repo objects from deep inside the stack
     # so we manually delay the circular imports to not break
     # scripts that don't use our demand-loading
     global hg
     from . import hg as h
     hg = h
 
     repo = ctx.repo()
     _auditsubrepopath(repo, path)
     state = ctx.substate[path]
     _checktype(repo.ui, state[2])
     if allowwdir:
         state = (state[0], ctx.subrev(path), state[2])
     return types[state[2]](ctx, path, state[:2], allowcreate)
 
 def nullsubrepo(ctx, path, pctx):
     """return an empty subrepo in pctx for the extant subrepo in ctx"""
     # subrepo inherently violates our import layering rules
     # because it wants to make repo objects from deep inside the stack
     # so we manually delay the circular imports to not break
     # scripts that don't use our demand-loading
     global hg
     from . import hg as h
     hg = h
 
     repo = ctx.repo()
     _auditsubrepopath(repo, path)
     state = ctx.substate[path]
     _checktype(repo.ui, state[2])
     subrev = ''
     if state[2] == 'hg':
         subrev = "0" * 40
     return types[state[2]](pctx, path, (state[0], subrev), True)
 
 # subrepo classes need to implement the following abstract class:
 
 class abstractsubrepo(object):
 
     def __init__(self, ctx, path):
         """Initialize abstractsubrepo part
 
         ``ctx`` is the context referring this subrepository in the
         parent repository.
 
         ``path`` is the path to this subrepository as seen from
         innermost repository.
         """
         self.ui = ctx.repo().ui
         self._ctx = ctx
         self._path = path
 
     def addwebdirpath(self, serverpath, webconf):
         """Add the hgwebdir entries for this subrepo, and any of its subrepos.
 
         ``serverpath`` is the path component of the URL for this repo.
 
         ``webconf`` is the dictionary of hgwebdir entries.
         """
         pass
 
     def storeclean(self, path):
         """
         returns true if the repository has not changed since it was last
         cloned from or pushed to a given repository.
         """
         return False
 
     def dirty(self, ignoreupdate=False, missing=False):
         """returns true if the dirstate of the subrepo is dirty or does not
         match current stored state. If ignoreupdate is true, only check
         whether the subrepo has uncommitted changes in its dirstate.  If missing
         is true, check for deleted files.
         """
         raise NotImplementedError
 
     def dirtyreason(self, ignoreupdate=False, missing=False):
         """return reason string if it is ``dirty()``
 
         Returned string should have enough information for the message
         of exception.
 
         This returns None, otherwise.
         """
         if self.dirty(ignoreupdate=ignoreupdate, missing=missing):
             return _('uncommitted changes in subrepository "%s"'
                      ) % subrelpath(self)
 
     def bailifchanged(self, ignoreupdate=False, hint=None):
         """raise Abort if subrepository is ``dirty()``
         """
         dirtyreason = self.dirtyreason(ignoreupdate=ignoreupdate,
                                        missing=True)
         if dirtyreason:
             raise error.Abort(dirtyreason, hint=hint)
 
     def basestate(self):
         """current working directory base state, disregarding .hgsubstate
         state and working directory modifications"""
         raise NotImplementedError
 
     def checknested(self, path):
         """check if path is a subrepository within this repository"""
         return False
 
     def commit(self, text, user, date):
         """commit the current changes to the subrepo with the given
         log message. Use given user and date if possible. Return the
         new state of the subrepo.
         """
         raise NotImplementedError
 
     def phase(self, state):
         """returns phase of specified state in the subrepository.
         """
         return phases.public
 
     def remove(self):
         """remove the subrepo
 
         (should verify the dirstate is not dirty first)
         """
         raise NotImplementedError
 
     def get(self, state, overwrite=False):
         """run whatever commands are needed to put the subrepo into
         this state
         """
         raise NotImplementedError
 
     def merge(self, state):
         """merge currently-saved state with the new state."""
         raise NotImplementedError
 
     def push(self, opts):
         """perform whatever action is analogous to 'hg push'
 
         This may be a no-op on some systems.
         """
         raise NotImplementedError
 
     def add(self, ui, match, prefix, explicitonly, **opts):
         return []
 
     def addremove(self, matcher, prefix, opts, dry_run, similarity):
         self.ui.warn("%s: %s" % (prefix, _("addremove is not supported")))
         return 1
 
     def cat(self, match, fm, fntemplate, prefix, **opts):
         return 1
 
     def status(self, rev2, **opts):
         return scmutil.status([], [], [], [], [], [], [])
 
     def diff(self, ui, diffopts, node2, match, prefix, **opts):
         pass
 
     def outgoing(self, ui, dest, opts):
         return 1
 
     def incoming(self, ui, source, opts):
         return 1
 
     def files(self):
         """return filename iterator"""
         raise NotImplementedError
 
     def filedata(self, name, decode):
         """return file data, optionally passed through repo decoders"""
         raise NotImplementedError
 
     def fileflags(self, name):
         """return file flags"""
         return ''
 
     def getfileset(self, expr):
         """Resolve the fileset expression for this repo"""
         return set()
 
     def printfiles(self, ui, m, fm, fmt, subrepos):
         """handle the files command for this subrepo"""
         return 1
 
     def archive(self, archiver, prefix, match=None, decode=True):
         if match is not None:
             files = [f for f in self.files() if match(f)]
         else:
             files = self.files()
         total = len(files)
         relpath = subrelpath(self)
         self.ui.progress(_('archiving (%s)') % relpath, 0,
                          unit=_('files'), total=total)
         for i, name in enumerate(files):
             flags = self.fileflags(name)
             mode = 'x' in flags and 0o755 or 0o644
             symlink = 'l' in flags
             archiver.addfile(prefix + self._path + '/' + name,
                              mode, symlink, self.filedata(name, decode))
             self.ui.progress(_('archiving (%s)') % relpath, i + 1,
                              unit=_('files'), total=total)
         self.ui.progress(_('archiving (%s)') % relpath, None)
         return total
 
     def walk(self, match):
         '''
         walk recursively through the directory tree, finding all files
         matched by the match function
         '''
 
     def forget(self, match, prefix):
         return ([], [])
 
     def removefiles(self, matcher, prefix, after, force, subrepos, warnings):
         """remove the matched files from the subrepository and the filesystem,
         possibly by force and/or after the file has been removed from the
         filesystem.  Return 0 on success, 1 on any warning.
         """
         warnings.append(_("warning: removefiles not implemented (%s)")
                         % self._path)
         return 1
 
     def revert(self, substate, *pats, **opts):
         self.ui.warn(_('%s: reverting %s subrepos is unsupported\n') \
             % (substate[0], substate[2]))
         return []
 
     def shortid(self, revid):
         return revid
 
     def unshare(self):
         '''
         convert this repository from shared to normal storage.
         '''
 
     def verify(self):
         '''verify the integrity of the repository.  Return 0 on success or
         warning, 1 on any error.
         '''
         return 0
 
     @propertycache
     def wvfs(self):
         """return vfs to access the working directory of this subrepository
         """
         return vfsmod.vfs(self._ctx.repo().wvfs.join(self._path))
 
     @propertycache
     def _relpath(self):
         """return path to this subrepository as seen from outermost repository
         """
         return self.wvfs.reljoin(reporelpath(self._ctx.repo()), self._path)
 
 class hgsubrepo(abstractsubrepo):
     def __init__(self, ctx, path, state, allowcreate):
         super(hgsubrepo, self).__init__(ctx, path)
         self._state = state
         r = ctx.repo()
         root = r.wjoin(path)
         create = allowcreate and not r.wvfs.exists('%s/.hg' % path)
         self._repo = hg.repository(r.baseui, root, create=create)
 
         # Propagate the parent's --hidden option
         if r is r.unfiltered():
             self._repo = self._repo.unfiltered()
 
         self.ui = self._repo.ui
         for s, k in [('ui', 'commitsubrepos')]:
             v = r.ui.config(s, k)
             if v:
                 self.ui.setconfig(s, k, v, 'subrepo')
         # internal config: ui._usedassubrepo
         self.ui.setconfig('ui', '_usedassubrepo', 'True', 'subrepo')
         self._initrepo(r, state[0], create)
 
     @annotatesubrepoerror
     def addwebdirpath(self, serverpath, webconf):
         cmdutil.addwebdirpath(self._repo, subrelpath(self), webconf)
 
     def storeclean(self, path):
         with self._repo.lock():
             return self._storeclean(path)
 
     def _storeclean(self, path):
         clean = True
         itercache = self._calcstorehash(path)
         for filehash in self._readstorehashcache(path):
             if filehash != next(itercache, None):
                 clean = False
                 break
         if clean:
             # if not empty:
             # the cached and current pull states have a different size
             clean = next(itercache, None) is None
         return clean
 
     def _calcstorehash(self, remotepath):
         '''calculate a unique "store hash"
 
         This method is used to to detect when there are changes that may
         require a push to a given remote path.'''
         # sort the files that will be hashed in increasing (likely) file size
         filelist = ('bookmarks', 'store/phaseroots', 'store/00changelog.i')
         yield '# %s\n' % _expandedabspath(remotepath)
         vfs = self._repo.vfs
         for relname in filelist:
             filehash = node.hex(hashlib.sha1(vfs.tryread(relname)).digest())
             yield '%s = %s\n' % (relname, filehash)
 
     @propertycache
     def _cachestorehashvfs(self):
         return vfsmod.vfs(self._repo.vfs.join('cache/storehash'))
 
     def _readstorehashcache(self, remotepath):
         '''read the store hash cache for a given remote repository'''
         cachefile = _getstorehashcachename(remotepath)
         return self._cachestorehashvfs.tryreadlines(cachefile, 'r')
 
     def _cachestorehash(self, remotepath):
         '''cache the current store hash
 
         Each remote repo requires its own store hash cache, because a subrepo
         store may be "clean" versus a given remote repo, but not versus another
         '''
         cachefile = _getstorehashcachename(remotepath)
         with self._repo.lock():
             storehash = list(self._calcstorehash(remotepath))
             vfs = self._cachestorehashvfs
             vfs.writelines(cachefile, storehash, mode='wb', notindexed=True)
 
     def _getctx(self):
         '''fetch the context for this subrepo revision, possibly a workingctx
         '''
         if self._ctx.rev() is None:
             return self._repo[None]  # workingctx if parent is workingctx
         else:
             rev = self._state[1]
             return self._repo[rev]
 
     @annotatesubrepoerror
     def _initrepo(self, parentrepo, source, create):
         self._repo._subparent = parentrepo
         self._repo._subsource = source
 
         if create:
             lines = ['[paths]\n']
 
             def addpathconfig(key, value):
                 if value:
                     lines.append('%s = %s\n' % (key, value))
                     self.ui.setconfig('paths', key, value, 'subrepo')
 
             defpath = _abssource(self._repo, abort=False)
             defpushpath = _abssource(self._repo, True, abort=False)
             addpathconfig('default', defpath)
             if defpath != defpushpath:
                 addpathconfig('default-push', defpushpath)
 
             self._repo.vfs.write('hgrc', util.tonativeeol(''.join(lines)))
 
     @annotatesubrepoerror
     def add(self, ui, match, prefix, explicitonly, **opts):
         return cmdutil.add(ui, self._repo, match,
                            self.wvfs.reljoin(prefix, self._path),
                            explicitonly, **opts)
 
     @annotatesubrepoerror
     def addremove(self, m, prefix, opts, dry_run, similarity):
         # In the same way as sub directories are processed, once in a subrepo,
         # always entry any of its subrepos.  Don't corrupt the options that will
         # be used to process sibling subrepos however.
         opts = copy.copy(opts)
         opts['subrepos'] = True
         return scmutil.addremove(self._repo, m,
                                  self.wvfs.reljoin(prefix, self._path), opts,
                                  dry_run, similarity)
 
     @annotatesubrepoerror
     def cat(self, match, fm, fntemplate, prefix, **opts):
         rev = self._state[1]
         ctx = self._repo[rev]
         return cmdutil.cat(self.ui, self._repo, ctx, match, fm, fntemplate,
                            prefix, **opts)
 
     @annotatesubrepoerror
     def status(self, rev2, **opts):
         try:
             rev1 = self._state[1]
             ctx1 = self._repo[rev1]
             ctx2 = self._repo[rev2]
             return self._repo.status(ctx1, ctx2, **opts)
         except error.RepoLookupError as inst:
             self.ui.warn(_('warning: error "%s" in subrepository "%s"\n')
                          % (inst, subrelpath(self)))
             return scmutil.status([], [], [], [], [], [], [])
 
     @annotatesubrepoerror
     def diff(self, ui, diffopts, node2, match, prefix, **opts):
         try:
             node1 = node.bin(self._state[1])
             # We currently expect node2 to come from substate and be
             # in hex format
             if node2 is not None:
                 node2 = node.bin(node2)
             logcmdutil.diffordiffstat(ui, self._repo, diffopts,
                                       node1, node2, match,
                                       prefix=posixpath.join(prefix, self._path),
                                       listsubrepos=True, **opts)
         except error.RepoLookupError as inst:
             self.ui.warn(_('warning: error "%s" in subrepository "%s"\n')
                           % (inst, subrelpath(self)))
 
     @annotatesubrepoerror
     def archive(self, archiver, prefix, match=None, decode=True):
         self._get(self._state + ('hg',))
         files = self.files()
         if match:
             files = [f for f in files if match(f)]
         rev = self._state[1]
         ctx = self._repo[rev]
         scmutil.fileprefetchhooks(self._repo, ctx, files)
         total = abstractsubrepo.archive(self, archiver, prefix, match)
         for subpath in ctx.substate:
             s = subrepo(ctx, subpath, True)
             submatch = matchmod.subdirmatcher(subpath, match)
             total += s.archive(archiver, prefix + self._path + '/', submatch,
                                decode)
         return total
 
     @annotatesubrepoerror
     def dirty(self, ignoreupdate=False, missing=False):
         r = self._state[1]
         if r == '' and not ignoreupdate: # no state recorded
             return True
         w = self._repo[None]
         if r != w.p1().hex() and not ignoreupdate:
             # different version checked out
             return True
         return w.dirty(missing=missing) # working directory changed
 
     def basestate(self):
         return self._repo['.'].hex()
 
     def checknested(self, path):
         return self._repo._checknested(self._repo.wjoin(path))
 
     @annotatesubrepoerror
     def commit(self, text, user, date):
         # don't bother committing in the subrepo if it's only been
         # updated
         if not self.dirty(True):
             return self._repo['.'].hex()
         self.ui.debug("committing subrepo %s\n" % subrelpath(self))
         n = self._repo.commit(text, user, date)
         if not n:
             return self._repo['.'].hex() # different version checked out
         return node.hex(n)
 
     @annotatesubrepoerror
     def phase(self, state):
         return self._repo[state].phase()
 
     @annotatesubrepoerror
     def remove(self):
         # we can't fully delete the repository as it may contain
         # local-only history
         self.ui.note(_('removing subrepo %s\n') % subrelpath(self))
         hg.clean(self._repo, node.nullid, False)
 
     def _get(self, state):
         source, revision, kind = state
         parentrepo = self._repo._subparent
 
         if revision in self._repo.unfiltered():
             # Allow shared subrepos tracked at null to setup the sharedpath
             if len(self._repo) != 0 or not parentrepo.shared():
                 return True
         self._repo._subsource = source
         srcurl = _abssource(self._repo)
         other = hg.peer(self._repo, {}, srcurl)
         if len(self._repo) == 0:
             # use self._repo.vfs instead of self.wvfs to remove .hg only
             self._repo.vfs.rmtree()
-            if parentrepo.shared():
+
+            # A remote subrepo could be shared if there is a local copy
+            # relative to the parent's share source.  But clone pooling doesn't
+            # assemble the repos in a tree, so that can't be consistently done.
+            # A simpler option is for the user to configure clone pooling, and
+            # work with that.
+            if parentrepo.shared() and hg.islocal(srcurl):
                 self.ui.status(_('sharing subrepo %s from %s\n')
                                % (subrelpath(self), srcurl))
                 shared = hg.share(self._repo._subparent.baseui,
                                   other, self._repo.root,
                                   update=False, bookmarks=False)
                 self._repo = shared.local()
             else:
+                # TODO: find a common place for this and this code in the
+                # share.py wrap of the clone command.
+                if parentrepo.shared():
+                    pool = self.ui.config('share', 'pool')
+                    if pool:
+                        pool = util.expandpath(pool)
+
+                    shareopts = {
+                        'pool': pool,
+                        'mode': self.ui.config('share', 'poolnaming'),
+                    }
+                else:
+                    shareopts = {}
+
                 self.ui.status(_('cloning subrepo %s from %s\n')
                                % (subrelpath(self), srcurl))
                 other, cloned = hg.clone(self._repo._subparent.baseui, {},
                                          other, self._repo.root,
-                                         update=False)
+                                         update=False, shareopts=shareopts)
                 self._repo = cloned.local()
             self._initrepo(parentrepo, source, create=True)
             self._cachestorehash(srcurl)
         else:
             self.ui.status(_('pulling subrepo %s from %s\n')
                            % (subrelpath(self), srcurl))
             cleansub = self.storeclean(srcurl)
             exchange.pull(self._repo, other)
             if cleansub:
                 # keep the repo clean after pull
                 self._cachestorehash(srcurl)
         return False
 
     @annotatesubrepoerror
     def get(self, state, overwrite=False):
         inrepo = self._get(state)
         source, revision, kind = state
         repo = self._repo
         repo.ui.debug("getting subrepo %s\n" % self._path)
         if inrepo:
             urepo = repo.unfiltered()
             ctx = urepo[revision]
             if ctx.hidden():
                 urepo.ui.warn(
                     _('revision %s in subrepository "%s" is hidden\n') \
                     % (revision[0:12], self._path))
                 repo = urepo
         hg.updaterepo(repo, revision, overwrite)
 
     @annotatesubrepoerror
     def merge(self, state):
         self._get(state)
         cur = self._repo['.']
         dst = self._repo[state[1]]
         anc = dst.ancestor(cur)
 
         def mergefunc():
             if anc == cur and dst.branch() == cur.branch():
                 self.ui.debug('updating subrepository "%s"\n'
                               % subrelpath(self))
                 hg.update(self._repo, state[1])
             elif anc == dst:
                 self.ui.debug('skipping subrepository "%s"\n'
                               % subrelpath(self))
             else:
                 self.ui.debug('merging subrepository "%s"\n' % subrelpath(self))
                 hg.merge(self._repo, state[1], remind=False)
 
         wctx = self._repo[None]
         if self.dirty():
             if anc != dst:
                 if _updateprompt(self.ui, self, wctx.dirty(), cur, dst):
                     mergefunc()
             else:
                 mergefunc()
         else:
             mergefunc()
 
     @annotatesubrepoerror
     def push(self, opts):
         force = opts.get('force')
         newbranch = opts.get('new_branch')
         ssh = opts.get('ssh')
 
         # push subrepos depth-first for coherent ordering
         c = self._repo['']
         subs = c.substate # only repos that are committed
         for s in sorted(subs):
             if c.sub(s).push(opts) == 0:
                 return False
 
         dsturl = _abssource(self._repo, True)
         if not force:
             if self.storeclean(dsturl):
                 self.ui.status(
                     _('no changes made to subrepo %s since last push to %s\n')
                     % (subrelpath(self), dsturl))
                 return None
         self.ui.status(_('pushing subrepo %s to %s\n') %
             (subrelpath(self), dsturl))
         other = hg.peer(self._repo, {'ssh': ssh}, dsturl)
         res = exchange.push(self._repo, other, force, newbranch=newbranch)
 
         # the repo is now clean
         self._cachestorehash(dsturl)
         return res.cgresult
 
     @annotatesubrepoerror
     def outgoing(self, ui, dest, opts):
         if 'rev' in opts or 'branch' in opts:
             opts = copy.copy(opts)
             opts.pop('rev', None)
             opts.pop('branch', None)
         return hg.outgoing(ui, self._repo, _abssource(self._repo, True), opts)
 
     @annotatesubrepoerror
     def incoming(self, ui, source, opts):
         if 'rev' in opts or 'branch' in opts:
             opts = copy.copy(opts)
             opts.pop('rev', None)
             opts.pop('branch', None)
         return hg.incoming(ui, self._repo, _abssource(self._repo, False), opts)
 
     @annotatesubrepoerror
     def files(self):
         rev = self._state[1]
         ctx = self._repo[rev]
         return ctx.manifest().keys()
 
     def filedata(self, name, decode):
         rev = self._state[1]
         data = self._repo[rev][name].data()
         if decode:
             data = self._repo.wwritedata(name, data)
         return data
 
     def fileflags(self, name):
         rev = self._state[1]
         ctx = self._repo[rev]
         return ctx.flags(name)
 
     @annotatesubrepoerror
     def printfiles(self, ui, m, fm, fmt, subrepos):
         # If the parent context is a workingctx, use the workingctx here for
         # consistency.
         if self._ctx.rev() is None:
             ctx = self._repo[None]
         else:
             rev = self._state[1]
             ctx = self._repo[rev]
         return cmdutil.files(ui, ctx, m, fm, fmt, subrepos)
 
     @annotatesubrepoerror
     def getfileset(self, expr):
         if self._ctx.rev() is None:
             ctx = self._repo[None]
         else:
             rev = self._state[1]
             ctx = self._repo[rev]
 
         files = ctx.getfileset(expr)
 
         for subpath in ctx.substate:
             sub = ctx.sub(subpath)
 
             try:
                 files.extend(subpath + '/' + f for f in sub.getfileset(expr))
             except error.LookupError:
                 self.ui.status(_("skipping missing subrepository: %s\n")
                                % self.wvfs.reljoin(reporelpath(self), subpath))
         return files
 
     def walk(self, match):
         ctx = self._repo[None]
         return ctx.walk(match)
 
     @annotatesubrepoerror
     def forget(self, match, prefix):
         return cmdutil.forget(self.ui, self._repo, match,
                               self.wvfs.reljoin(prefix, self._path), True)
 
     @annotatesubrepoerror
     def removefiles(self, matcher, prefix, after, force, subrepos, warnings):
         return cmdutil.remove(self.ui, self._repo, matcher,
                               self.wvfs.reljoin(prefix, self._path),
                               after, force, subrepos)
 
     @annotatesubrepoerror
     def revert(self, substate, *pats, **opts):
         # reverting a subrepo is a 2 step process:
         # 1. if the no_backup is not set, revert all modified
         #    files inside the subrepo
         # 2. update the subrepo to the revision specified in
         #    the corresponding substate dictionary
         self.ui.status(_('reverting subrepo %s\n') % substate[0])
         if not opts.get(r'no_backup'):
             # Revert all files on the subrepo, creating backups
             # Note that this will not recursively revert subrepos
             # We could do it if there was a set:subrepos() predicate
             opts = opts.copy()
             opts[r'date'] = None
             opts[r'rev'] = substate[1]
 
             self.filerevert(*pats, **opts)
 
         # Update the repo to the revision specified in the given substate
         if not opts.get(r'dry_run'):
             self.get(substate, overwrite=True)
 
     def filerevert(self, *pats, **opts):
         ctx = self._repo[opts[r'rev']]
         parents = self._repo.dirstate.parents()
         if opts.get(r'all'):
             pats = ['set:modified()']
         else:
             pats = []
         cmdutil.revert(self.ui, self._repo, ctx, parents, *pats, **opts)
 
     def shortid(self, revid):
         return revid[:12]
 
     @annotatesubrepoerror
     def unshare(self):
         # subrepo inherently violates our import layering rules
         # because it wants to make repo objects from deep inside the stack
         # so we manually delay the circular imports to not break
         # scripts that don't use our demand-loading
         global hg
         from . import hg as h
         hg = h
 
         # Nothing prevents a user from sharing in a repo, and then making that a
         # subrepo.  Alternately, the previous unshare attempt may have failed
         # part way through.  So recurse whether or not this layer is shared.
         if self._repo.shared():
             self.ui.status(_("unsharing subrepo '%s'\n") % self._relpath)
 
         hg.unshare(self.ui, self._repo)
 
     def verify(self):
         try:
             rev = self._state[1]
             ctx = self._repo.unfiltered()[rev]
             if ctx.hidden():
                 # Since hidden revisions aren't pushed/pulled, it seems worth an
                 # explicit warning.
                 ui = self._repo.ui
                 ui.warn(_("subrepo '%s' is hidden in revision %s\n") %
                         (self._relpath, node.short(self._ctx.node())))
             return 0
         except error.RepoLookupError:
             # A missing subrepo revision may be a case of needing to pull it, so
             # don't treat this as an error.
             self._repo.ui.warn(_("subrepo '%s' not found in revision %s\n") %
                                (self._relpath, node.short(self._ctx.node())))
             return 0
 
     @propertycache
     def wvfs(self):
         """return own wvfs for efficiency and consistency
         """
         return self._repo.wvfs
 
     @propertycache
     def _relpath(self):
         """return path to this subrepository as seen from outermost repository
         """
         # Keep consistent dir separators by avoiding vfs.join(self._path)
         return reporelpath(self._repo)
 
 class svnsubrepo(abstractsubrepo):
     def __init__(self, ctx, path, state, allowcreate):
         super(svnsubrepo, self).__init__(ctx, path)
         self._state = state
         self._exe = util.findexe('svn')
         if not self._exe:
             raise error.Abort(_("'svn' executable not found for subrepo '%s'")
                              % self._path)
 
     def _svncommand(self, commands, filename='', failok=False):
         cmd = [self._exe]
         extrakw = {}
         if not self.ui.interactive():
             # Making stdin be a pipe should prevent svn from behaving
             # interactively even if we can't pass --non-interactive.
             extrakw[r'stdin'] = subprocess.PIPE
             # Starting in svn 1.5 --non-interactive is a global flag
             # instead of being per-command, but we need to support 1.4 so
             # we have to be intelligent about what commands take
             # --non-interactive.
             if commands[0] in ('update', 'checkout', 'commit'):
                 cmd.append('--non-interactive')
         cmd.extend(commands)
         if filename is not None:
             path = self.wvfs.reljoin(self._ctx.repo().origroot,
                                      self._path, filename)
             cmd.append(path)
         env = dict(encoding.environ)
         # Avoid localized output, preserve current locale for everything else.
         lc_all = env.get('LC_ALL')
         if lc_all:
             env['LANG'] = lc_all
             del env['LC_ALL']
         env['LC_MESSAGES'] = 'C'
         p = subprocess.Popen(cmd, bufsize=-1, close_fds=util.closefds,
                              stdout=subprocess.PIPE, stderr=subprocess.PIPE,
                               universal_newlines=True, env=env, **extrakw)
         stdout, stderr = p.communicate()
         stderr = stderr.strip()
         if not failok:
             if p.returncode:
                 raise error.Abort(stderr or 'exited with code %d'
                                   % p.returncode)
             if stderr:
                 self.ui.warn(stderr + '\n')
         return stdout, stderr
 
     @propertycache
     def _svnversion(self):
         output, err = self._svncommand(['--version', '--quiet'], filename=None)
         m = re.search(br'^(\d+)\.(\d+)', output)
         if not m:
             raise error.Abort(_('cannot retrieve svn tool version'))
         return (int(m.group(1)), int(m.group(2)))
 
     def _svnmissing(self):
         return not self.wvfs.exists('.svn')
 
     def _wcrevs(self):
         # Get the working directory revision as well as the last
         # commit revision so we can compare the subrepo state with
         # both. We used to store the working directory one.
         output, err = self._svncommand(['info', '--xml'])
         doc = xml.dom.minidom.parseString(output)
         entries = doc.getElementsByTagName('entry')
         lastrev, rev = '0', '0'
         if entries:
             rev = str(entries[0].getAttribute('revision')) or '0'
             commits = entries[0].getElementsByTagName('commit')
             if commits:
                 lastrev = str(commits[0].getAttribute('revision')) or '0'
         return (lastrev, rev)
 
     def _wcrev(self):
         return self._wcrevs()[0]
 
     def _wcchanged(self):
         """Return (changes, extchanges, missing) where changes is True
         if the working directory was changed, extchanges is
         True if any of these changes concern an external entry and missing
         is True if any change is a missing entry.
         """
         output, err = self._svncommand(['status', '--xml'])
         externals, changes, missing = [], [], []
         doc = xml.dom.minidom.parseString(output)
         for e in doc.getElementsByTagName('entry'):
             s = e.getElementsByTagName('wc-status')
             if not s:
                 continue
             item = s[0].getAttribute('item')
             props = s[0].getAttribute('props')
             path = e.getAttribute('path')
             if item == 'external':
                 externals.append(path)
             elif item == 'missing':
                 missing.append(path)
             if (item not in ('', 'normal', 'unversioned', 'external')
                 or props not in ('', 'none', 'normal')):
                 changes.append(path)
         for path in changes:
             for ext in externals:
                 if path == ext or path.startswith(ext + pycompat.ossep):
                     return True, True, bool(missing)
         return bool(changes), False, bool(missing)
 
     @annotatesubrepoerror
     def dirty(self, ignoreupdate=False, missing=False):
         if self._svnmissing():
             return self._state[1] != ''
         wcchanged = self._wcchanged()
         changed = wcchanged[0] or (missing and wcchanged[2])
         if not changed:
             if self._state[1] in self._wcrevs() or ignoreupdate:
                 return False
         return True
 
     def basestate(self):
         lastrev, rev = self._wcrevs()
         if lastrev != rev:
             # Last committed rev is not the same than rev. We would
             # like to take lastrev but we do not know if the subrepo
             # URL exists at lastrev.  Test it and fallback to rev it
             # is not there.
             try:
                 self._svncommand(['list', '%s@%s' % (self._state[0], lastrev)])
                 return lastrev
             except error.Abort:
                 pass
         return rev
 
     @annotatesubrepoerror
     def commit(self, text, user, date):
         # user and date are out of our hands since svn is centralized
         changed, extchanged, missing = self._wcchanged()
         if not changed:
             return self.basestate()
         if extchanged:
             # Do not try to commit externals
             raise error.Abort(_('cannot commit svn externals'))
         if missing:
             # svn can commit with missing entries but aborting like hg
             # seems a better approach.
             raise error.Abort(_('cannot commit missing svn entries'))
         commitinfo, err = self._svncommand(['commit', '-m', text])
         self.ui.status(commitinfo)
         newrev = re.search('Committed revision ([0-9]+).', commitinfo)
         if not newrev:
             if not commitinfo.strip():
                 # Sometimes, our definition of "changed" differs from
                 # svn one. For instance, svn ignores missing files
                 # when committing. If there are only missing files, no
                 # commit is made, no output and no error code.
                 raise error.Abort(_('failed to commit svn changes'))
             raise error.Abort(commitinfo.splitlines()[-1])
         newrev = newrev.groups()[0]
         self.ui.status(self._svncommand(['update', '-r', newrev])[0])
         return newrev
 
     @annotatesubrepoerror
     def remove(self):
         if self.dirty():
             self.ui.warn(_('not removing repo %s because '
                            'it has changes.\n') % self._path)
             return
         self.ui.note(_('removing subrepo %s\n') % self._path)
 
         self.wvfs.rmtree(forcibly=True)
         try:
             pwvfs = self._ctx.repo().wvfs
             pwvfs.removedirs(pwvfs.dirname(self._path))
         except OSError:
             pass
 
     @annotatesubrepoerror
     def get(self, state, overwrite=False):
         if overwrite:
             self._svncommand(['revert', '--recursive'])
         args = ['checkout']
         if self._svnversion >= (1, 5):
             args.append('--force')
         # The revision must be specified at the end of the URL to properly
         # update to a directory which has since been deleted and recreated.
         args.append('%s@%s' % (state[0], state[1]))
 
         # SEC: check that the ssh url is safe
         util.checksafessh(state[0])
 
         status, err = self._svncommand(args, failok=True)
         _sanitize(self.ui, self.wvfs, '.svn')
         if not re.search('Checked out revision [0-9]+.', status):
             if ('is already a working copy for a different URL' in err
                 and (self._wcchanged()[:2] == (False, False))):
                 # obstructed but clean working copy, so just blow it away.
                 self.remove()
                 self.get(state, overwrite=False)
                 return
             raise error.Abort((status or err).splitlines()[-1])
         self.ui.status(status)
 
     @annotatesubrepoerror
     def merge(self, state):
         old = self._state[1]
         new = state[1]
         wcrev = self._wcrev()
         if new != wcrev:
             dirty = old == wcrev or self._wcchanged()[0]
             if _updateprompt(self.ui, self, dirty, wcrev, new):
                 self.get(state, False)
 
     def push(self, opts):
         # push is a no-op for SVN
         return True
 
     @annotatesubrepoerror
     def files(self):
         output = self._svncommand(['list', '--recursive', '--xml'])[0]
         doc = xml.dom.minidom.parseString(output)
         paths = []
         for e in doc.getElementsByTagName('entry'):
             kind = str(e.getAttribute('kind'))
             if kind != 'file':
                 continue
             name = ''.join(c.data for c
                            in e.getElementsByTagName('name')[0].childNodes
                            if c.nodeType == c.TEXT_NODE)
             paths.append(name.encode('utf-8'))
         return paths
 
     def filedata(self, name, decode):
         return self._svncommand(['cat'], name)[0]
 
 
 class gitsubrepo(abstractsubrepo):
     def __init__(self, ctx, path, state, allowcreate):
         super(gitsubrepo, self).__init__(ctx, path)
         self._state = state
         self._abspath = ctx.repo().wjoin(path)
         self._subparent = ctx.repo()
         self._ensuregit()
 
     def _ensuregit(self):
         try:
             self._gitexecutable = 'git'
             out, err = self._gitnodir(['--version'])
         except OSError as e:
             genericerror = _("error executing git for subrepo '%s': %s")
             notfoundhint = _("check git is installed and in your PATH")
             if e.errno != errno.ENOENT:
                 raise error.Abort(genericerror % (
                     self._path, encoding.strtolocal(e.strerror)))
             elif pycompat.iswindows:
                 try:
                     self._gitexecutable = 'git.cmd'
                     out, err = self._gitnodir(['--version'])
                 except OSError as e2:
                     if e2.errno == errno.ENOENT:
                         raise error.Abort(_("couldn't find 'git' or 'git.cmd'"
                             " for subrepo '%s'") % self._path,
                             hint=notfoundhint)
                     else:
                         raise error.Abort(genericerror % (self._path,
                             encoding.strtolocal(e2.strerror)))
             else:
                 raise error.Abort(_("couldn't find git for subrepo '%s'")
                     % self._path, hint=notfoundhint)
         versionstatus = self._checkversion(out)
         if versionstatus == 'unknown':
             self.ui.warn(_('cannot retrieve git version\n'))
         elif versionstatus == 'abort':
             raise error.Abort(_('git subrepo requires at least 1.6.0 or later'))
         elif versionstatus == 'warning':
             self.ui.warn(_('git subrepo requires at least 1.6.0 or later\n'))
 
     @staticmethod
     def _gitversion(out):
         m = re.search(br'^git version (\d+)\.(\d+)\.(\d+)', out)
         if m:
             return (int(m.group(1)), int(m.group(2)), int(m.group(3)))
 
         m = re.search(br'^git version (\d+)\.(\d+)', out)
         if m:
             return (int(m.group(1)), int(m.group(2)), 0)
 
         return -1
 
     @staticmethod
     def _checkversion(out):
         '''ensure git version is new enough
 
         >>> _checkversion = gitsubrepo._checkversion
         >>> _checkversion(b'git version 1.6.0')
         'ok'
         >>> _checkversion(b'git version 1.8.5')
         'ok'
         >>> _checkversion(b'git version 1.4.0')
         'abort'
         >>> _checkversion(b'git version 1.5.0')
         'warning'
         >>> _checkversion(b'git version 1.9-rc0')
         'ok'
         >>> _checkversion(b'git version 1.9.0.265.g81cdec2')
         'ok'
         >>> _checkversion(b'git version 1.9.0.GIT')
         'ok'
         >>> _checkversion(b'git version 12345')
         'unknown'
         >>> _checkversion(b'no')
         'unknown'
         '''
         version = gitsubrepo._gitversion(out)
         # git 1.4.0 can't work at all, but 1.5.X can in at least some cases,
         # despite the docstring comment.  For now, error on 1.4.0, warn on
         # 1.5.0 but attempt to continue.
         if version == -1:
             return 'unknown'
         if version < (1, 5, 0):
             return 'abort'
         elif version < (1, 6, 0):
             return 'warning'
         return 'ok'
 
     def _gitcommand(self, commands, env=None, stream=False):
         return self._gitdir(commands, env=env, stream=stream)[0]
 
     def _gitdir(self, commands, env=None, stream=False):
         return self._gitnodir(commands, env=env, stream=stream,
                               cwd=self._abspath)
 
     def _gitnodir(self, commands, env=None, stream=False, cwd=None):
         """Calls the git command
 
         The methods tries to call the git command. versions prior to 1.6.0
         are not supported and very probably fail.
         """
         self.ui.debug('%s: git %s\n' % (self._relpath, ' '.join(commands)))
         if env is None:
             env = encoding.environ.copy()
         # disable localization for Git output (issue5176)
         env['LC_ALL'] = 'C'
         # fix for Git CVE-2015-7545
         if 'GIT_ALLOW_PROTOCOL' not in env:
             env['GIT_ALLOW_PROTOCOL'] = 'file:git:http:https:ssh'
         # unless ui.quiet is set, print git's stderr,
         # which is mostly progress and useful info
         errpipe = None
         if self.ui.quiet:
             errpipe = open(os.devnull, 'w')
         if self.ui._colormode and len(commands) and commands[0] == "diff":
             # insert the argument in the front,
             # the end of git diff arguments is used for paths
             commands.insert(1, '--color')
         p = subprocess.Popen([self._gitexecutable] + commands, bufsize=-1,
                              cwd=cwd, env=env, close_fds=util.closefds,
                              stdout=subprocess.PIPE, stderr=errpipe)
         if stream:
             return p.stdout, None
 
         retdata = p.stdout.read().strip()
         # wait for the child to exit to avoid race condition.
         p.wait()
 
         if p.returncode != 0 and p.returncode != 1:
             # there are certain error codes that are ok
             command = commands[0]
             if command in ('cat-file', 'symbolic-ref'):
                 return retdata, p.returncode
             # for all others, abort
             raise error.Abort(_('git %s error %d in %s') %
                              (command, p.returncode, self._relpath))
 
         return retdata, p.returncode
 
     def _gitmissing(self):
         return not self.wvfs.exists('.git')
 
     def _gitstate(self):
         return self._gitcommand(['rev-parse', 'HEAD'])
 
     def _gitcurrentbranch(self):
         current, err = self._gitdir(['symbolic-ref', 'HEAD', '--quiet'])
         if err:
             current = None
         return current
 
     def _gitremote(self, remote):
         out = self._gitcommand(['remote', 'show', '-n', remote])
         line = out.split('\n')[1]
         i = line.index('URL: ') + len('URL: ')
         return line[i:]
 
     def _githavelocally(self, revision):
         out, code = self._gitdir(['cat-file', '-e', revision])
         return code == 0
 
     def _gitisancestor(self, r1, r2):
         base = self._gitcommand(['merge-base', r1, r2])
         return base == r1
 
     def _gitisbare(self):
         return self._gitcommand(['config', '--bool', 'core.bare']) == 'true'
 
     def _gitupdatestat(self):
         """This must be run before git diff-index.
         diff-index only looks at changes to file stat;
         this command looks at file contents and updates the stat."""
         self._gitcommand(['update-index', '-q', '--refresh'])
 
     def _gitbranchmap(self):
         '''returns 2 things:
         a map from git branch to revision
         a map from revision to branches'''
         branch2rev = {}
         rev2branch = {}
 
         out = self._gitcommand(['for-each-ref', '--format',
                                 '%(objectname) %(refname)'])
         for line in out.split('\n'):
             revision, ref = line.split(' ')
             if (not ref.startswith('refs/heads/') and
                 not ref.startswith('refs/remotes/')):
                 continue
             if ref.startswith('refs/remotes/') and ref.endswith('/HEAD'):
                 continue # ignore remote/HEAD redirects
             branch2rev[ref] = revision
             rev2branch.setdefault(revision, []).append(ref)
         return branch2rev, rev2branch
 
     def _gittracking(self, branches):
         'return map of remote branch to local tracking branch'
         # assumes no more than one local tracking branch for each remote
         tracking = {}
         for b in branches:
             if b.startswith('refs/remotes/'):
                 continue
             bname = b.split('/', 2)[2]
             remote = self._gitcommand(['config', 'branch.%s.remote' % bname])
             if remote:
                 ref = self._gitcommand(['config', 'branch.%s.merge' % bname])
                 tracking['refs/remotes/%s/%s' %
                          (remote, ref.split('/', 2)[2])] = b
         return tracking
 
     def _abssource(self, source):
         if '://' not in source:
             # recognize the scp syntax as an absolute source
             colon = source.find(':')
             if colon != -1 and '/' not in source[:colon]:
                 return source
         self._subsource = source
         return _abssource(self)
 
     def _fetch(self, source, revision):
         if self._gitmissing():
             # SEC: check for safe ssh url
             util.checksafessh(source)
 
             source = self._abssource(source)
             self.ui.status(_('cloning subrepo %s from %s\n') %
                             (self._relpath, source))
             self._gitnodir(['clone', source, self._abspath])
         if self._githavelocally(revision):
             return
         self.ui.status(_('pulling subrepo %s from %s\n') %
                         (self._relpath, self._gitremote('origin')))
         # try only origin: the originally cloned repo
         self._gitcommand(['fetch'])
         if not self._githavelocally(revision):
             raise error.Abort(_('revision %s does not exist in subrepository '
                                 '"%s"\n') % (revision, self._relpath))
 
     @annotatesubrepoerror
     def dirty(self, ignoreupdate=False, missing=False):
         if self._gitmissing():
             return self._state[1] != ''
         if self._gitisbare():
             return True
         if not ignoreupdate and self._state[1] != self._gitstate():
             # different version checked out
             return True
         # check for staged changes or modified files; ignore untracked files
         self._gitupdatestat()
         out, code = self._gitdir(['diff-index', '--quiet', 'HEAD'])
         return code == 1
 
     def basestate(self):
         return self._gitstate()
 
     @annotatesubrepoerror
     def get(self, state, overwrite=False):
         source, revision, kind = state
         if not revision:
             self.remove()
             return
         self._fetch(source, revision)
         # if the repo was set to be bare, unbare it
         if self._gitisbare():
             self._gitcommand(['config', 'core.bare', 'false'])
             if self._gitstate() == revision:
                 self._gitcommand(['reset', '--hard', 'HEAD'])
                 return
         elif self._gitstate() == revision:
             if overwrite:
                 # first reset the index to unmark new files for commit, because
                 # reset --hard will otherwise throw away files added for commit,
                 # not just unmark them.
                 self._gitcommand(['reset', 'HEAD'])
                 self._gitcommand(['reset', '--hard', 'HEAD'])
             return
         branch2rev, rev2branch = self._gitbranchmap()
 
         def checkout(args):
             cmd = ['checkout']
             if overwrite:
                 # first reset the index to unmark new files for commit, because
                 # the -f option will otherwise throw away files added for
                 # commit, not just unmark them.
                 self._gitcommand(['reset', 'HEAD'])
                 cmd.append('-f')
             self._gitcommand(cmd + args)
             _sanitize(self.ui, self.wvfs, '.git')
 
         def rawcheckout():
             # no branch to checkout, check it out with no branch
             self.ui.warn(_('checking out detached HEAD in '
                            'subrepository "%s"\n') % self._relpath)
             self.ui.warn(_('check out a git branch if you intend '
                             'to make changes\n'))
             checkout(['-q', revision])
 
         if revision not in rev2branch:
             rawcheckout()
             return
         branches = rev2branch[revision]
         firstlocalbranch = None
         for b in branches:
             if b == 'refs/heads/master':
                 # master trumps all other branches
                 checkout(['refs/heads/master'])
                 return
             if not firstlocalbranch and not b.startswith('refs/remotes/'):
                 firstlocalbranch = b
         if firstlocalbranch:
             checkout([firstlocalbranch])
             return
 
         tracking = self._gittracking(branch2rev.keys())
         # choose a remote branch already tracked if possible
         remote = branches[0]
         if remote not in tracking:
             for b in branches:
                 if b in tracking:
                     remote = b
                     break
 
         if remote not in tracking:
             # create a new local tracking branch
             local = remote.split('/', 3)[3]
             checkout(['-b', local, remote])
         elif self._gitisancestor(branch2rev[tracking[remote]], remote):
             # When updating to a tracked remote branch,
             # if the local tracking branch is downstream of it,
             # a normal `git pull` would have performed a "fast-forward merge"
             # which is equivalent to updating the local branch to the remote.
             # Since we are only looking at branching at update, we need to
             # detect this situation and perform this action lazily.
             if tracking[remote] != self._gitcurrentbranch():
                 checkout([tracking[remote]])
             self._gitcommand(['merge', '--ff', remote])
             _sanitize(self.ui, self.wvfs, '.git')
         else:
             # a real merge would be required, just checkout the revision
             rawcheckout()
 
     @annotatesubrepoerror
     def commit(self, text, user, date):
         if self._gitmissing():
             raise error.Abort(_("subrepo %s is missing") % self._relpath)
         cmd = ['commit', '-a', '-m', text]
         env = encoding.environ.copy()
         if user:
             cmd += ['--author', user]
         if date:
             # git's date parser silently ignores when seconds < 1e9
             # convert to ISO8601
             env['GIT_AUTHOR_DATE'] = dateutil.datestr(date,
                                                   '%Y-%m-%dT%H:%M:%S %1%2')
         self._gitcommand(cmd, env=env)
         # make sure commit works otherwise HEAD might not exist under certain
         # circumstances
         return self._gitstate()
 
     @annotatesubrepoerror
     def merge(self, state):
         source, revision, kind = state
         self._fetch(source, revision)
         base = self._gitcommand(['merge-base', revision, self._state[1]])
         self._gitupdatestat()
         out, code = self._gitdir(['diff-index', '--quiet', 'HEAD'])
 
         def mergefunc():
             if base == revision:
                 self.get(state) # fast forward merge
             elif base != self._state[1]:
                 self._gitcommand(['merge', '--no-commit', revision])
             _sanitize(self.ui, self.wvfs, '.git')
 
         if self.dirty():
             if self._gitstate() != revision:
                 dirty = self._gitstate() == self._state[1] or code != 0
                 if _updateprompt(self.ui, self, dirty,
                                  self._state[1][:7], revision[:7]):
                     mergefunc()
         else:
             mergefunc()
 
     @annotatesubrepoerror
     def push(self, opts):
         force = opts.get('force')
 
         if not self._state[1]:
             return True
         if self._gitmissing():
             raise error.Abort(_("subrepo %s is missing") % self._relpath)
         # if a branch in origin contains the revision, nothing to do
         branch2rev, rev2branch = self._gitbranchmap()
         if self._state[1] in rev2branch:
             for b in rev2branch[self._state[1]]:
                 if b.startswith('refs/remotes/origin/'):
                     return True
         for b, revision in branch2rev.iteritems():
             if b.startswith('refs/remotes/origin/'):
                 if self._gitisancestor(self._state[1], revision):
                     return True
         # otherwise, try to push the currently checked out branch
         cmd = ['push']
         if force:
             cmd.append('--force')
 
         current = self._gitcurrentbranch()
         if current:
             # determine if the current branch is even useful
             if not self._gitisancestor(self._state[1], current):
                 self.ui.warn(_('unrelated git branch checked out '
                                'in subrepository "%s"\n') % self._relpath)
                 return False
             self.ui.status(_('pushing branch %s of subrepository "%s"\n') %
                            (current.split('/', 2)[2], self._relpath))
             ret = self._gitdir(cmd + ['origin', current])
             return ret[1] == 0
         else:
             self.ui.warn(_('no branch checked out in subrepository "%s"\n'
                            'cannot push revision %s\n') %
                           (self._relpath, self._state[1]))
             return False
 
     @annotatesubrepoerror
     def add(self, ui, match, prefix, explicitonly, **opts):
         if self._gitmissing():
             return []
 
         (modified, added, removed,
          deleted, unknown, ignored, clean) = self.status(None, unknown=True,
                                                          clean=True)
 
         tracked = set()
         # dirstates 'amn' warn, 'r' is added again
         for l in (modified, added, deleted, clean):
             tracked.update(l)
 
         # Unknown files not of interest will be rejected by the matcher
         files = unknown
         files.extend(match.files())
 
         rejected = []
 
         files = [f for f in sorted(set(files)) if match(f)]
         for f in files:
             exact = match.exact(f)
             command = ["add"]
             if exact:
                 command.append("-f") #should be added, even if ignored
             if ui.verbose or not exact:
                 ui.status(_('adding %s\n') % match.rel(f))
 
             if f in tracked:  # hg prints 'adding' even if already tracked
                 if exact:
                     rejected.append(f)
                 continue
             if not opts.get(r'dry_run'):
                 self._gitcommand(command + [f])
 
         for f in rejected:
             ui.warn(_("%s already tracked!\n") % match.abs(f))
 
         return rejected
 
     @annotatesubrepoerror
     def remove(self):
         if self._gitmissing():
             return
         if self.dirty():
             self.ui.warn(_('not removing repo %s because '
                            'it has changes.\n') % self._relpath)
             return
         # we can't fully delete the repository as it may contain
         # local-only history
         self.ui.note(_('removing subrepo %s\n') % self._relpath)
         self._gitcommand(['config', 'core.bare', 'true'])
         for f, kind in self.wvfs.readdir():
             if f == '.git':
                 continue
             if kind == stat.S_IFDIR:
                 self.wvfs.rmtree(f)
             else:
                 self.wvfs.unlink(f)
 
     def archive(self, archiver, prefix, match=None, decode=True):
         total = 0
         source, revision = self._state
         if not revision:
             return total
         self._fetch(source, revision)
 
         # Parse git's native archive command.
         # This should be much faster than manually traversing the trees
         # and objects with many subprocess calls.
         tarstream = self._gitcommand(['archive', revision], stream=True)
         tar = tarfile.open(fileobj=tarstream, mode='r|')
         relpath = subrelpath(self)
         self.ui.progress(_('archiving (%s)') % relpath, 0, unit=_('files'))
         for i, info in enumerate(tar):
             if info.isdir():
                 continue
             if match and not match(info.name):
                 continue
             if info.issym():
                 data = info.linkname
             else:
                 data = tar.extractfile(info).read()
             archiver.addfile(prefix + self._path + '/' + info.name,
                              info.mode, info.issym(), data)
             total += 1
             self.ui.progress(_('archiving (%s)') % relpath, i + 1,
                              unit=_('files'))
         self.ui.progress(_('archiving (%s)') % relpath, None)
         return total
 
 
     @annotatesubrepoerror
     def cat(self, match, fm, fntemplate, prefix, **opts):
         rev = self._state[1]
         if match.anypats():
             return 1 #No support for include/exclude yet
 
         if not match.files():
             return 1
 
         # TODO: add support for non-plain formatter (see cmdutil.cat())
         for f in match.files():
             output = self._gitcommand(["show", "%s:%s" % (rev, f)])
             fp = cmdutil.makefileobj(self._ctx, fntemplate,
                                      pathname=self.wvfs.reljoin(prefix, f))
             fp.write(output)
             fp.close()
         return 0
 
 
     @annotatesubrepoerror
     def status(self, rev2, **opts):
         rev1 = self._state[1]
         if self._gitmissing() or not rev1:
             # if the repo is missing, return no results
             return scmutil.status([], [], [], [], [], [], [])
         modified, added, removed = [], [], []
         self._gitupdatestat()
         if rev2:
             command = ['diff-tree', '--no-renames', '-r', rev1, rev2]
         else:
             command = ['diff-index', '--no-renames', rev1]
         out = self._gitcommand(command)
         for line in out.split('\n'):
             tab = line.find('\t')
             if tab == -1:
                 continue
             status, f = line[tab - 1], line[tab + 1:]
             if status == 'M':
                 modified.append(f)
             elif status == 'A':
                 added.append(f)
             elif status == 'D':
                 removed.append(f)
 
         deleted, unknown, ignored, clean = [], [], [], []
 
         command = ['status', '--porcelain', '-z']
         if opts.get(r'unknown'):
             command += ['--untracked-files=all']
         if opts.get(r'ignored'):
             command += ['--ignored']
         out = self._gitcommand(command)
 
         changedfiles = set()
         changedfiles.update(modified)
         changedfiles.update(added)
         changedfiles.update(removed)
         for line in out.split('\0'):
             if not line:
                 continue
             st = line[0:2]
             #moves and copies show 2 files on one line
             if line.find('\0') >= 0:
                 filename1, filename2 = line[3:].split('\0')
             else:
                 filename1 = line[3:]
                 filename2 = None
 
             changedfiles.add(filename1)
             if filename2:
                 changedfiles.add(filename2)
 
             if st == '??':
                 unknown.append(filename1)
             elif st == '!!':
                 ignored.append(filename1)
 
         if opts.get(r'clean'):
             out = self._gitcommand(['ls-files'])
             for f in out.split('\n'):
                 if not f in changedfiles:
                     clean.append(f)
 
         return scmutil.status(modified, added, removed, deleted,
                               unknown, ignored, clean)
 
     @annotatesubrepoerror
     def diff(self, ui, diffopts, node2, match, prefix, **opts):
         node1 = self._state[1]
         cmd = ['diff', '--no-renames']
         if opts[r'stat']:
             cmd.append('--stat')
         else:
             # for Git, this also implies '-p'
             cmd.append('-U%d' % diffopts.context)
 
         gitprefix = self.wvfs.reljoin(prefix, self._path)
 
         if diffopts.noprefix:
             cmd.extend(['--src-prefix=%s/' % gitprefix,
                         '--dst-prefix=%s/' % gitprefix])
         else:
             cmd.extend(['--src-prefix=a/%s/' % gitprefix,
                         '--dst-prefix=b/%s/' % gitprefix])
 
         if diffopts.ignorews:
             cmd.append('--ignore-all-space')
         if diffopts.ignorewsamount:
             cmd.append('--ignore-space-change')
         if self._gitversion(self._gitcommand(['--version'])) >= (1, 8, 4) \
                 and diffopts.ignoreblanklines:
             cmd.append('--ignore-blank-lines')
 
         cmd.append(node1)
         if node2:
             cmd.append(node2)
 
         output = ""
         if match.always():
             output += self._gitcommand(cmd) + '\n'
         else:
             st = self.status(node2)[:3]
             files = [f for sublist in st for f in sublist]
             for f in files:
                 if match(f):
                     output += self._gitcommand(cmd + ['--', f]) + '\n'
 
         if output.strip():
             ui.write(output)
 
     @annotatesubrepoerror
     def revert(self, substate, *pats, **opts):
         self.ui.status(_('reverting subrepo %s\n') % substate[0])
         if not opts.get(r'no_backup'):
             status = self.status(None)
             names = status.modified
             for name in names:
                 bakname = scmutil.origpath(self.ui, self._subparent, name)
                 self.ui.note(_('saving current version of %s as %s\n') %
                         (name, bakname))
                 self.wvfs.rename(name, bakname)
 
         if not opts.get(r'dry_run'):
             self.get(substate, overwrite=True)
         return []
 
     def shortid(self, revid):
         return revid[:7]
 
 types = {
     'hg': hgsubrepo,
     'svn': svnsubrepo,
     'git': gitsubrepo,
     }

tests/test-annotate.t

@@ -1,1003 +1,1035 b''
   $ HGMERGE=true; export HGMERGE
 
 init
 
   $ hg init repo
   $ cd repo
 
 commit
 
   $ echo 'a' > a
   $ hg ci -A -m test -u nobody -d '1 0'
   adding a
 
 annotate -c
 
   $ hg annotate -c a
   8435f90966e4: a
 
 annotate -cl
 
   $ hg annotate -cl a
   8435f90966e4:1: a
 
 annotate -d
 
   $ hg annotate -d a
   Thu Jan 01 00:00:01 1970 +0000: a
 
 annotate -n
 
   $ hg annotate -n a
   0: a
 
 annotate -nl
 
   $ hg annotate -nl a
   0:1: a
 
 annotate -u
 
   $ hg annotate -u a
   nobody: a
 
 annotate -cdnu
 
   $ hg annotate -cdnu a
   nobody 0 8435f90966e4 Thu Jan 01 00:00:01 1970 +0000: a
 
 annotate -cdnul
 
   $ hg annotate -cdnul a
   nobody 0 8435f90966e4 Thu Jan 01 00:00:01 1970 +0000:1: a
 
 annotate (JSON)
 
   $ hg annotate -Tjson a
   [
    {
     "abspath": "a",
     "lines": [{"line": "a\n", "rev": 0}],
     "path": "a"
    }
   ]
 
   $ hg annotate -Tjson -cdfnul a
   [
    {
     "abspath": "a",
     "lines": [{"date": [1.0, 0], "file": "a", "line": "a\n", "line_number": 1, "node": "8435f90966e442695d2ded29fdade2bac5ad8065", "rev": 0, "user": "nobody"}],
     "path": "a"
    }
   ]
 
   $ cat <<EOF >>a
   > a
   > a
   > EOF
   $ hg ci -ma1 -d '1 0'
   $ hg cp a b
   $ hg ci -mb -d '1 0'
   $ cat <<EOF >> b
   > b4
   > b5
   > b6
   > EOF
   $ hg ci -mb2 -d '2 0'
 
 annotate multiple files (JSON)
 
   $ hg annotate -Tjson a b
   [
    {
     "abspath": "a",
     "lines": [{"line": "a\n", "rev": 0}, {"line": "a\n", "rev": 1}, {"line": "a\n", "rev": 1}],
     "path": "a"
    },
    {
     "abspath": "b",
     "lines": [{"line": "a\n", "rev": 0}, {"line": "a\n", "rev": 1}, {"line": "a\n", "rev": 1}, {"line": "b4\n", "rev": 3}, {"line": "b5\n", "rev": 3}, {"line": "b6\n", "rev": 3}],
     "path": "b"
    }
   ]
 
 annotate multiple files (template)
 
   $ hg annotate -T'== {abspath} ==\n{lines % "{rev}: {line}"}' a b
   == a ==
   0: a
   1: a
   1: a
   == b ==
   0: a
   1: a
   1: a
   3: b4
   3: b5
   3: b6
 
 annotate -n b
 
   $ hg annotate -n b
   0: a
   1: a
   1: a
   3: b4
   3: b5
   3: b6
 
 annotate --no-follow b
 
   $ hg annotate --no-follow b
   2: a
   2: a
   2: a
   3: b4
   3: b5
   3: b6
 
 annotate -nl b
 
   $ hg annotate -nl b
   0:1: a
   1:2: a
   1:3: a
   3:4: b4
   3:5: b5
   3:6: b6
 
 annotate -nf b
 
   $ hg annotate -nf b
   0 a: a
   1 a: a
   1 a: a
   3 b: b4
   3 b: b5
   3 b: b6
 
 annotate -nlf b
 
   $ hg annotate -nlf b
   0 a:1: a
   1 a:2: a
   1 a:3: a
   3 b:4: b4
   3 b:5: b5
   3 b:6: b6
 
   $ hg up -C 2
   1 files updated, 0 files merged, 0 files removed, 0 files unresolved
   $ cat <<EOF >> b
   > b4
   > c
   > b5
   > EOF
   $ hg ci -mb2.1 -d '2 0'
   created new head
   $ hg merge
   merging b
   0 files updated, 1 files merged, 0 files removed, 0 files unresolved
   (branch merge, don't forget to commit)
   $ hg ci -mmergeb -d '3 0'
 
 annotate after merge
 
   $ hg annotate -nf b
   0 a: a
   1 a: a
   1 a: a
   3 b: b4
   4 b: c
   3 b: b5
 
 annotate after merge with -l
 
   $ hg annotate -nlf b
   0 a:1: a
   1 a:2: a
   1 a:3: a
   3 b:4: b4
   4 b:5: c
   3 b:5: b5
 
   $ hg up -C 1
   0 files updated, 0 files merged, 1 files removed, 0 files unresolved
   $ hg cp a b
   $ cat <<EOF > b
   > a
   > z
   > a
   > EOF
   $ hg ci -mc -d '3 0'
   created new head
   $ hg merge
   merging b
   0 files updated, 1 files merged, 0 files removed, 0 files unresolved
   (branch merge, don't forget to commit)
   $ cat <<EOF >> b
   > b4
   > c
   > b5
   > EOF
   $ echo d >> b
   $ hg ci -mmerge2 -d '4 0'
 
 annotate after rename merge
 
   $ hg annotate -nf b
   0 a: a
   6 b: z
   1 a: a
   3 b: b4
   4 b: c
   3 b: b5
   7 b: d
 
 annotate after rename merge with -l
 
   $ hg annotate -nlf b
   0 a:1: a
   6 b:2: z
   1 a:3: a
   3 b:4: b4
   4 b:5: c
   3 b:5: b5
   7 b:7: d
 
 --skip nothing (should be the same as no --skip at all)
 
   $ hg annotate -nlf b --skip '1::0'
   0 a:1: a
   6 b:2: z
   1 a:3: a
   3 b:4: b4
   4 b:5: c
   3 b:5: b5
   7 b:7: d
 
 --skip a modified line. Note a slight behavior difference in pure - this is
 because the pure code comes up with slightly different deltas internally.
 
   $ hg annotate -nlf b --skip 6
   0 a:1: a
   1 a:2* z (no-pure !)
   0 a:1* z (pure !)
   1 a:3: a
   3 b:4: b4
   4 b:5: c
   3 b:5: b5
   7 b:7: d
 
 --skip added lines (and test multiple skip)
 
   $ hg annotate -nlf b --skip 3
   0 a:1: a
   6 b:2: z
   1 a:3: a
   1 a:3* b4
   4 b:5: c
   1 a:3* b5
   7 b:7: d
 
   $ hg annotate -nlf b --skip 4
   0 a:1: a
   6 b:2: z
   1 a:3: a
   3 b:4: b4
   1 a:3* c
   3 b:5: b5
   7 b:7: d
 
   $ hg annotate -nlf b --skip 3 --skip 4
   0 a:1: a
   6 b:2: z
   1 a:3: a
   1 a:3* b4
   1 a:3* c
   1 a:3* b5
   7 b:7: d
 
   $ hg annotate -nlf b --skip 'merge()'
   0 a:1: a
   6 b:2: z
   1 a:3: a
   3 b:4: b4
   4 b:5: c
   3 b:5: b5
   3 b:5* d
 
 --skip everything -- use the revision the file was introduced in
 
   $ hg annotate -nlf b --skip 'all()'
   0 a:1: a
   0 a:1* z
   0 a:1* a
   0 a:1* b4
   0 a:1* c
   0 a:1* b5
   0 a:1* d
 
 Issue2807: alignment of line numbers with -l
 
   $ echo more >> b
   $ hg ci -mmore -d '5 0'
   $ echo more >> b
   $ hg ci -mmore -d '6 0'
   $ echo more >> b
   $ hg ci -mmore -d '7 0'
   $ hg annotate -nlf b
    0 a: 1: a
    6 b: 2: z
    1 a: 3: a
    3 b: 4: b4
    4 b: 5: c
    3 b: 5: b5
    7 b: 7: d
    8 b: 8: more
    9 b: 9: more
   10 b:10: more
 
 linkrev vs rev
 
   $ hg annotate -r tip -n a
   0: a
   1: a
   1: a
 
 linkrev vs rev with -l
 
   $ hg annotate -r tip -nl a
   0:1: a
   1:2: a
   1:3: a
 
 Issue589: "undelete" sequence leads to crash
 
 annotate was crashing when trying to --follow something
 
 like A -> B -> A
 
 generate ABA rename configuration
 
   $ echo foo > foo
   $ hg add foo
   $ hg ci -m addfoo
   $ hg rename foo bar
   $ hg ci -m renamefoo
   $ hg rename bar foo
   $ hg ci -m renamebar
 
 annotate after ABA with follow
 
   $ hg annotate --follow foo
   foo: foo
 
 missing file
 
   $ hg ann nosuchfile
   abort: nosuchfile: no such file in rev e9e6b4fa872f
   [255]
 
 annotate file without '\n' on last line
 
   $ printf "" > c
   $ hg ci -A -m test -u nobody -d '1 0'
   adding c
   $ hg annotate c
   $ printf "a\nb" > c
   $ hg ci -m test
   $ hg annotate c
   [0-9]+: a (re)
   [0-9]+: b (re)
 
 Issue3841: check annotation of the file of which filelog includes
 merging between the revision and its ancestor
 
 to reproduce the situation with recent Mercurial, this script uses (1)
 "hg debugsetparents" to merge without ancestor check by "hg merge",
 and (2) the extension to allow filelog merging between the revision
 and its ancestor by overriding "repo._filecommit".
 
   $ cat > ../legacyrepo.py <<EOF
   > from __future__ import absolute_import
   > from mercurial import error, node
   > def reposetup(ui, repo):
   >     class legacyrepo(repo.__class__):
   >         def _filecommit(self, fctx, manifest1, manifest2,
   >                         linkrev, tr, changelist):
   >             fname = fctx.path()
   >             text = fctx.data()
   >             flog = self.file(fname)
   >             fparent1 = manifest1.get(fname, node.nullid)
   >             fparent2 = manifest2.get(fname, node.nullid)
   >             meta = {}
   >             copy = fctx.renamed()
   >             if copy and copy[0] != fname:
   >                 raise error.Abort('copying is not supported')
   >             if fparent2 != node.nullid:
   >                 changelist.append(fname)
   >                 return flog.add(text, meta, tr, linkrev,
   >                                 fparent1, fparent2)
   >             raise error.Abort('only merging is supported')
   >     repo.__class__ = legacyrepo
   > EOF
 
   $ cat > baz <<EOF
   > 1
   > 2
   > 3
   > 4
   > 5
   > EOF
   $ hg add baz
   $ hg commit -m "baz:0"
 
   $ cat > baz <<EOF
   > 1 baz:1
   > 2
   > 3
   > 4
   > 5
   > EOF
   $ hg commit -m "baz:1"
 
   $ cat > baz <<EOF
   > 1 baz:1
   > 2 baz:2
   > 3
   > 4
   > 5
   > EOF
   $ hg debugsetparents 17 17
   $ hg --config extensions.legacyrepo=../legacyrepo.py  commit -m "baz:2"
   $ hg debugindexdot .hg/store/data/baz.i
   digraph G {
   	-1 -> 0
   	0 -> 1
   	1 -> 2
   	1 -> 2
   }
   $ hg annotate baz
   17: 1 baz:1
   18: 2 baz:2
   16: 3
   16: 4
   16: 5
 
   $ cat > baz <<EOF
   > 1 baz:1
   > 2 baz:2
   > 3 baz:3
   > 4
   > 5
   > EOF
   $ hg commit -m "baz:3"
 
   $ cat > baz <<EOF
   > 1 baz:1
   > 2 baz:2
   > 3 baz:3
   > 4 baz:4
   > 5
   > EOF
   $ hg debugsetparents 19 18
   $ hg --config extensions.legacyrepo=../legacyrepo.py  commit -m "baz:4"
   $ hg debugindexdot .hg/store/data/baz.i
   digraph G {
   	-1 -> 0
   	0 -> 1
   	1 -> 2
   	1 -> 2
   	2 -> 3
   	3 -> 4
   	2 -> 4
   }
   $ hg annotate baz
   17: 1 baz:1
   18: 2 baz:2
   19: 3 baz:3
   20: 4 baz:4
   16: 5
 
 annotate clean file
 
   $ hg annotate -ncr "wdir()" foo
   11 472b18db256d : foo
 
 annotate modified file
 
   $ echo foofoo >> foo
   $ hg annotate -r "wdir()" foo
   11 : foo
   20+: foofoo
 
   $ hg annotate -cr "wdir()" foo
   472b18db256d : foo
   b6bedd5477e7+: foofoo
 
   $ hg annotate -ncr "wdir()" foo
   11 472b18db256d : foo
   20 b6bedd5477e7+: foofoo
 
   $ hg annotate --debug -ncr "wdir()" foo
   11 472b18db256d1e8282064eab4bfdaf48cbfe83cd : foo
   20 b6bedd5477e797f25e568a6402d4697f3f895a72+: foofoo
 
   $ hg annotate -udr "wdir()" foo
   test Thu Jan 01 00:00:00 1970 +0000: foo
   test [A-Za-z0-9:+ ]+: foofoo (re)
 
   $ hg annotate -ncr "wdir()" -Tjson foo
   [
    {
     "abspath": "foo",
     "lines": [{"line": "foo\n", "node": "472b18db256d1e8282064eab4bfdaf48cbfe83cd", "rev": 11}, {"line": "foofoo\n", "node": null, "rev": null}],
     "path": "foo"
    }
   ]
 
 annotate added file
 
   $ echo bar > bar
   $ hg add bar
   $ hg annotate -ncr "wdir()" bar
   20 b6bedd5477e7+: bar
 
 annotate renamed file
 
   $ hg rename foo renamefoo2
   $ hg annotate -ncr "wdir()" renamefoo2
   11 472b18db256d : foo
   20 b6bedd5477e7+: foofoo
 
 annotate missing file
 
   $ rm baz
 
   $ hg annotate -ncr "wdir()" baz
   abort: $TESTTMP\repo\baz: $ENOENT$ (windows !)
   abort: $ENOENT$: $TESTTMP/repo/baz (no-windows !)
   [255]
 
 annotate removed file
 
   $ hg rm baz
 
   $ hg annotate -ncr "wdir()" baz
   abort: $TESTTMP\repo\baz: $ENOENT$ (windows !)
   abort: $ENOENT$: $TESTTMP/repo/baz (no-windows !)
   [255]
 
   $ hg revert --all --no-backup --quiet
   $ hg id -n
   20
 
 Test followlines() revset; we usually check both followlines(pat, range) and
 followlines(pat, range, descend=True) to make sure both give the same result
 when they should.
 
   $ echo a >> foo
   $ hg ci -m 'foo: add a'
   $ hg log -T '{rev}: {desc}\n' -r 'followlines(baz, 3:5)'
   16: baz:0
   19: baz:3
   20: baz:4
   $ hg log -T '{rev}: {desc}\n' -r 'followlines(baz, 3:5, startrev=20)'
   16: baz:0
   19: baz:3
   20: baz:4
   $ hg log -T '{rev}: {desc}\n' -r 'followlines(baz, 3:5, startrev=19)'
   16: baz:0
   19: baz:3
   $ hg log -T '{rev}: {desc}\n' -r 'followlines(baz, 3:5, startrev=19, descend=True)'
   19: baz:3
   20: baz:4
   $ printf "0\n0\n" | cat - baz > baz1
   $ mv baz1 baz
   $ hg ci -m 'added two lines with 0'
   $ hg log -T '{rev}: {desc}\n' -r 'followlines(baz, 5:7)'
   16: baz:0
   19: baz:3
   20: baz:4
   $ hg log -T '{rev}: {desc}\n' -r 'followlines(baz, 3:5, descend=true, startrev=19)'
   19: baz:3
   20: baz:4
   $ echo 6 >> baz
   $ hg ci -m 'added line 8'
   $ hg log -T '{rev}: {desc}\n' -r 'followlines(baz, 5:7)'
   16: baz:0
   19: baz:3
   20: baz:4
   $ hg log -T '{rev}: {desc}\n' -r 'followlines(baz, 3:5, startrev=19, descend=1)'
   19: baz:3
   20: baz:4
   $ sed 's/3/3+/' baz > baz.new
   $ mv baz.new baz
   $ hg ci -m 'baz:3->3+'
   $ hg log -T '{rev}: {desc}\n' -r 'followlines(baz, 5:7, descend=0)'
   16: baz:0
   19: baz:3
   20: baz:4
   24: baz:3->3+
   $ hg log -T '{rev}: {desc}\n' -r 'followlines(baz, 3:5, startrev=17, descend=True)'
   19: baz:3
   20: baz:4
   24: baz:3->3+
   $ hg log -T '{rev}: {desc}\n' -r 'followlines(baz, 1:2, descend=false)'
   22: added two lines with 0
 
 file patterns are okay
   $ hg log -T '{rev}: {desc}\n' -r 'followlines("path:baz", 1:2)'
   22: added two lines with 0
 
 renames are followed
   $ hg mv baz qux
   $ sed 's/4/4+/' qux > qux.new
   $ mv qux.new qux
   $ hg ci -m 'qux:4->4+'
   $ hg log -T '{rev}: {desc}\n' -r 'followlines(qux, 5:7)'
   16: baz:0
   19: baz:3
   20: baz:4
   24: baz:3->3+
   25: qux:4->4+
 
 but are missed when following children
   $ hg log -T '{rev}: {desc}\n' -r 'followlines(baz, 5:7, startrev=22, descend=True)'
   24: baz:3->3+
 
 merge
   $ hg up 24 --quiet
   $ echo 7 >> baz
   $ hg ci -m 'one more line, out of line range'
   created new head
   $ sed 's/3+/3-/' baz > baz.new
   $ mv baz.new baz
   $ hg ci -m 'baz:3+->3-'
   $ hg log -T '{rev}: {desc}\n' -r 'followlines(baz, 5:7)'
   16: baz:0
   19: baz:3
   20: baz:4
   24: baz:3->3+
   27: baz:3+->3-
   $ hg merge 25
   merging baz and qux to qux
   0 files updated, 1 files merged, 0 files removed, 0 files unresolved
   (branch merge, don't forget to commit)
   $ hg ci -m merge
   $ hg log -T '{rev}: {desc}\n' -r 'followlines(qux, 5:7)'
   16: baz:0
   19: baz:3
   20: baz:4
   24: baz:3->3+
   25: qux:4->4+
   27: baz:3+->3-
   28: merge
   $ hg up 25 --quiet
   $ hg merge 27
   merging qux and baz to qux
   0 files updated, 1 files merged, 0 files removed, 0 files unresolved
   (branch merge, don't forget to commit)
   $ hg ci -m 'merge from other side'
   created new head
   $ hg log -T '{rev}: {desc}\n' -r 'followlines(qux, 5:7)'
   16: baz:0
   19: baz:3
   20: baz:4
   24: baz:3->3+
   25: qux:4->4+
   27: baz:3+->3-
   29: merge from other side
   $ hg up 24 --quiet
 
 we are missing the branch with rename when following children
   $ hg log -T '{rev}: {desc}\n' -r 'followlines(baz, 5:7, startrev=26, descend=True)'
   27: baz:3+->3-
 
 we follow all branches in descending direction
   $ hg up 23 --quiet
   $ sed 's/3/+3/' baz > baz.new
   $ mv baz.new baz
   $ hg ci -m 'baz:3->+3'
   created new head
   $ hg log -T '{rev}: {desc}\n' -r 'followlines(baz, 2:5, startrev=16, descend=True)' --graph
   @  30: baz:3->+3
   :
   : o  27: baz:3+->3-
   : :
   : o  24: baz:3->3+
   :/
   o    20: baz:4
   |\
   | o  19: baz:3
   |/
   o  18: baz:2
   :
   o  16: baz:0
   |
   ~
 
 Issue5595: on a merge changeset with different line ranges depending on
 parent, be conservative and use the surrounding interval to avoid loosing
 track of possible further descendants in specified range.
 
   $ hg up 23 --quiet
   $ hg cat baz -r 24
   0
   0
   1 baz:1
   2 baz:2
   3+ baz:3
   4 baz:4
   5
   6
   $ cat > baz << EOF
   > 0
   > 0
   > a
   > b
   > 3+ baz:3
   > 4 baz:4
   > y
   > z
   > EOF
   $ hg ci -m 'baz: mostly rewrite with some content from 24'
   created new head
   $ hg merge --tool :merge-other 24
   merging baz
   1 files updated, 0 files merged, 0 files removed, 0 files unresolved
   (branch merge, don't forget to commit)
   $ hg ci -m 'merge forgetting about baz rewrite'
   $ cat > baz << EOF
   > 0
   > 0
   > 1 baz:1
   > 2+ baz:2
   > 3+ baz:3
   > 4 baz:4
   > 5
   > 6
   > EOF
   $ hg ci -m 'baz: narrow change (2->2+)'
   $ hg log -T '{rev}: {desc}\n' -r 'followlines(baz, 3:4, startrev=20, descend=True)' --graph
   @  33: baz: narrow change (2->2+)
   |
   o    32: merge forgetting about baz rewrite
   |\
   | o  31: baz: mostly rewrite with some content from 24
   | :
   | : o  30: baz:3->+3
   | :/
   +---o  27: baz:3+->3-
   | :
   o :  24: baz:3->3+
   :/
   o    20: baz:4
   |\
   ~ ~
 
 check error cases
   $ hg up 24 --quiet
   $ hg log -r 'followlines()'
   hg: parse error: followlines takes at least 1 positional arguments
   [255]
   $ hg log -r 'followlines(baz)'
   hg: parse error: followlines requires a line range
   [255]
   $ hg log -r 'followlines(baz, 1)'
   hg: parse error: followlines expects a line range
   [255]
   $ hg log -r 'followlines(baz, 1:2, startrev=desc("b"))'
   hg: parse error: followlines expects exactly one revision
   [255]
   $ hg log -r 'followlines("glob:*", 1:2)'
   hg: parse error: followlines expects exactly one file
   [255]
   $ hg log -r 'followlines(baz, 1:)'
   hg: parse error: line range bounds must be integers
   [255]
   $ hg log -r 'followlines(baz, :1)'
   hg: parse error: line range bounds must be integers
   [255]
   $ hg log -r 'followlines(baz, x:4)'
   hg: parse error: line range bounds must be integers
   [255]
   $ hg log -r 'followlines(baz, 5:4)'
   hg: parse error: line range must be positive
   [255]
   $ hg log -r 'followlines(baz, 0:4)'
   hg: parse error: fromline must be strictly positive
   [255]
   $ hg log -r 'followlines(baz, 2:40)'
   abort: line range exceeds file size
   [255]
   $ hg log -r 'followlines(baz, 2:4, startrev=20, descend=[1])'
   hg: parse error at 43: not a prefix: [
   (followlines(baz, 2:4, startrev=20, descend=[1])
                                               ^ here)
   [255]
   $ hg log -r 'followlines(baz, 2:4, startrev=20, descend=a)'
   hg: parse error: descend argument must be a boolean
   [255]
 
 Test empty annotate output
 
   $ printf '\0' > binary
   $ touch empty
   $ hg ci -qAm 'add binary and empty files'
 
   $ hg annotate binary empty
   binary: binary file
 
   $ hg annotate -Tjson binary empty
   [
    {
     "abspath": "binary",
     "path": "binary"
    },
    {
     "abspath": "empty",
     "lines": [],
     "path": "empty"
    }
   ]
 
 Test annotate with whitespace options
 
   $ cd ..
   $ hg init repo-ws
   $ cd repo-ws
   $ cat > a <<EOF
   > aa
   > 
   > b b
   > EOF
   $ hg ci -Am "adda"
   adding a
   $ sed 's/EOL$//g' > a <<EOF
   > a  a
   > 
   >  EOL
   > b  b
   > EOF
   $ hg ci -m "changea"
 
 Annotate with no option
 
   $ hg annotate a
   1: a  a
   0: 
   1:  
   1: b  b
 
 Annotate with --ignore-space-change
 
   $ hg annotate --ignore-space-change a
   1: a  a
   1: 
   0:  
   0: b  b
 
 Annotate with --ignore-all-space
 
   $ hg annotate --ignore-all-space a
   0: a  a
   0: 
   1:  
   0: b  b
 
 Annotate with --ignore-blank-lines (similar to no options case)
 
   $ hg annotate --ignore-blank-lines a
   1: a  a
   0: 
   1:  
   1: b  b
 
   $ cd ..
 
+Annotate with orphaned CR (issue5798)
+-------------------------------------
+
+  $ hg init repo-cr
+  $ cd repo-cr
+
+  $ substcr() {
+  > sed 's/\r/[CR]/g'
+  > }
+
+  >>> with open('a', 'wb') as f:
+  ...     f.write(b'0a\r0b\r\n0c\r0d\r\n0e\n0f\n0g')
+  $ hg ci -qAm0
+  >>> with open('a', 'wb') as f:
+  ...     f.write(b'0a\r0b\r\n1c\r1d\r\n0e\n1f\n0g')
+  $ hg ci -m1
+
+  $ hg annotate -r0 a | substcr
+  0: 0a[CR]0b[CR]
+  0: 0c[CR]0d[CR]
+  0: 0e
+  0: 0f
+  0: 0g
+  $ hg annotate -r1 a | substcr
+  0: 0a[CR]0b[CR]
+  1: 1c[CR]1d[CR]
+  0: 0e
+  1: 1f
+  0: 0g
+
+  $ cd ..
+
 Annotate with linkrev pointing to another branch
 ------------------------------------------------
 
 create history with a filerev whose linkrev points to another branch
 
   $ hg init branchedlinkrev
   $ cd branchedlinkrev
   $ echo A > a
   $ hg commit -Am 'contentA'
   adding a
   $ echo B >> a
   $ hg commit -m 'contentB'
   $ hg up --rev 'desc(contentA)'
   1 files updated, 0 files merged, 0 files removed, 0 files unresolved
   $ echo unrelated > unrelated
   $ hg commit -Am 'unrelated'
   adding unrelated
   created new head
   $ hg graft -r 'desc(contentB)'
   grafting 1:fd27c222e3e6 "contentB"
   $ echo C >> a
   $ hg commit -m 'contentC'
   $ echo W >> a
   $ hg log -G
   @  changeset:   4:072f1e8df249
   |  tag:         tip
   |  user:        test
   |  date:        Thu Jan 01 00:00:00 1970 +0000
   |  summary:     contentC
   |
   o  changeset:   3:ff38df03cc4b
   |  user:        test
   |  date:        Thu Jan 01 00:00:00 1970 +0000
   |  summary:     contentB
   |
   o  changeset:   2:62aaf3f6fc06
   |  parent:      0:f0932f74827e
   |  user:        test
   |  date:        Thu Jan 01 00:00:00 1970 +0000
   |  summary:     unrelated
   |
   | o  changeset:   1:fd27c222e3e6
   |/   user:        test
   |    date:        Thu Jan 01 00:00:00 1970 +0000
   |    summary:     contentB
   |
   o  changeset:   0:f0932f74827e
      user:        test
      date:        Thu Jan 01 00:00:00 1970 +0000
      summary:     contentA
   
 
 Annotate should list ancestor of starting revision only
 
   $ hg annotate a
   0: A
   3: B
   4: C
 
   $ hg annotate a -r 'wdir()'
   0 : A
   3 : B
   4 : C
   4+: W
 
 Even when the starting revision is the linkrev-shadowed one:
 
   $ hg annotate a -r 3
   0: A
   3: B
 
   $ cd ..
 
 Issue5360: Deleted chunk in p1 of a merge changeset
 
   $ hg init repo-5360
   $ cd repo-5360
   $ echo 1 > a
   $ hg commit -A a -m 1
   $ echo 2 >> a
   $ hg commit -m 2
   $ echo a > a
   $ hg commit -m a
   $ hg update '.^' -q
   $ echo 3 >> a
   $ hg commit -m 3 -q
   $ hg merge 2 -q
   $ cat > a << EOF
   > b
   > 1
   > 2
   > 3
   > a
   > EOF
   $ hg resolve --mark -q
   $ hg commit -m m
   $ hg annotate a
   4: b
   0: 1
   1: 2
   3: 3
   2: a
 
   $ cd ..

tests/test-subrepo-recursion.t

@@ -1,622 +1,696 b''
 Create test repository:
 
   $ hg init repo
   $ cd repo
   $ echo x1 > x.txt
 
   $ hg init foo
   $ cd foo
   $ echo y1 > y.txt
 
   $ hg init bar
   $ cd bar
   $ echo z1 > z.txt
 
   $ cd ..
   $ echo 'bar = bar' > .hgsub
 
   $ cd ..
   $ echo 'foo = foo' > .hgsub
 
 Add files --- .hgsub files must go first to trigger subrepos:
 
   $ hg add -S .hgsub
   $ hg add -S foo/.hgsub
   $ hg add -S foo/bar
   adding foo/bar/z.txt
   $ hg add -S
   adding x.txt
   adding foo/y.txt
 
 Test recursive status without committing anything:
 
   $ hg status -S
   A .hgsub
   A foo/.hgsub
   A foo/bar/z.txt
   A foo/y.txt
   A x.txt
 
 Test recursive diff without committing anything:
 
   $ hg diff --nodates -S foo
   diff -r 000000000000 foo/.hgsub
   --- /dev/null
   +++ b/foo/.hgsub
   @@ -0,0 +1,1 @@
   +bar = bar
   diff -r 000000000000 foo/y.txt
   --- /dev/null
   +++ b/foo/y.txt
   @@ -0,0 +1,1 @@
   +y1
   diff -r 000000000000 foo/bar/z.txt
   --- /dev/null
   +++ b/foo/bar/z.txt
   @@ -0,0 +1,1 @@
   +z1
 
 Commits:
 
   $ hg commit -m fails
   abort: uncommitted changes in subrepository "foo"
   (use --subrepos for recursive commit)
   [255]
 
 The --subrepos flag overwrite the config setting:
 
   $ hg commit -m 0-0-0 --config ui.commitsubrepos=No --subrepos
   committing subrepository foo
   committing subrepository foo/bar
 
   $ cd foo
   $ echo y2 >> y.txt
   $ hg commit -m 0-1-0
 
   $ cd bar
   $ echo z2 >> z.txt
   $ hg commit -m 0-1-1
 
   $ cd ..
   $ hg commit -m 0-2-1
 
   $ cd ..
   $ hg commit -m 1-2-1
 
 Change working directory:
 
   $ echo y3 >> foo/y.txt
   $ echo z3 >> foo/bar/z.txt
   $ hg status -S
   M foo/bar/z.txt
   M foo/y.txt
   $ hg diff --nodates -S
   diff -r d254738c5f5e foo/y.txt
   --- a/foo/y.txt
   +++ b/foo/y.txt
   @@ -1,2 +1,3 @@
    y1
    y2
   +y3
   diff -r 9647f22de499 foo/bar/z.txt
   --- a/foo/bar/z.txt
   +++ b/foo/bar/z.txt
   @@ -1,2 +1,3 @@
    z1
    z2
   +z3
 
 Status call crossing repository boundaries:
 
   $ hg status -S foo/bar/z.txt
   M foo/bar/z.txt
   $ hg status -S -I 'foo/?.txt'
   M foo/y.txt
   $ hg status -S -I '**/?.txt'
   M foo/bar/z.txt
   M foo/y.txt
   $ hg diff --nodates -S -I '**/?.txt'
   diff -r d254738c5f5e foo/y.txt
   --- a/foo/y.txt
   +++ b/foo/y.txt
   @@ -1,2 +1,3 @@
    y1
    y2
   +y3
   diff -r 9647f22de499 foo/bar/z.txt
   --- a/foo/bar/z.txt
   +++ b/foo/bar/z.txt
   @@ -1,2 +1,3 @@
    z1
    z2
   +z3
 
 Status from within a subdirectory:
 
   $ mkdir dir
   $ cd dir
   $ echo a1 > a.txt
   $ hg status -S
   M foo/bar/z.txt
   M foo/y.txt
   ? dir/a.txt
   $ hg diff --nodates -S
   diff -r d254738c5f5e foo/y.txt
   --- a/foo/y.txt
   +++ b/foo/y.txt
   @@ -1,2 +1,3 @@
    y1
    y2
   +y3
   diff -r 9647f22de499 foo/bar/z.txt
   --- a/foo/bar/z.txt
   +++ b/foo/bar/z.txt
   @@ -1,2 +1,3 @@
    z1
    z2
   +z3
 
 Status with relative path:
 
   $ hg status -S ..
   M ../foo/bar/z.txt
   M ../foo/y.txt
   ? a.txt
 
 XXX: filtering lfilesrepo.status() in 3.3-rc causes these files to be listed as
 added instead of modified.
   $ hg status -S .. --config extensions.largefiles=
   M ../foo/bar/z.txt
   M ../foo/y.txt
   ? a.txt
 
   $ hg diff --nodates -S ..
   diff -r d254738c5f5e foo/y.txt
   --- a/foo/y.txt
   +++ b/foo/y.txt
   @@ -1,2 +1,3 @@
    y1
    y2
   +y3
   diff -r 9647f22de499 foo/bar/z.txt
   --- a/foo/bar/z.txt
   +++ b/foo/bar/z.txt
   @@ -1,2 +1,3 @@
    z1
    z2
   +z3
   $ cd ..
 
 Cleanup and final commit:
 
   $ rm -r dir
   $ hg commit --subrepos -m 2-3-2
   committing subrepository foo
   committing subrepository foo/bar
 
 Test explicit path commands within subrepos: add/forget
   $ echo z1 > foo/bar/z2.txt
   $ hg status -S
   ? foo/bar/z2.txt
   $ hg add foo/bar/z2.txt
   $ hg status -S
   A foo/bar/z2.txt
   $ hg forget foo/bar/z2.txt
   $ hg status -S
   ? foo/bar/z2.txt
   $ hg forget foo/bar/z2.txt
   not removing foo/bar/z2.txt: file is already untracked
   [1]
   $ hg status -S
   ? foo/bar/z2.txt
   $ rm foo/bar/z2.txt
 
 Log with the relationships between repo and its subrepo:
 
   $ hg log --template '{rev}:{node|short} {desc}\n'
   2:1326fa26d0c0 2-3-2
   1:4b3c9ff4f66b 1-2-1
   0:23376cbba0d8 0-0-0
 
   $ hg -R foo log --template '{rev}:{node|short} {desc}\n'
   3:65903cebad86 2-3-2
   2:d254738c5f5e 0-2-1
   1:8629ce7dcc39 0-1-0
   0:af048e97ade2 0-0-0
 
   $ hg -R foo/bar log --template '{rev}:{node|short} {desc}\n'
   2:31ecbdafd357 2-3-2
   1:9647f22de499 0-1-1
   0:4904098473f9 0-0-0
 
 Status between revisions:
 
   $ hg status -S
   $ hg status -S --rev 0:1
   M .hgsubstate
   M foo/.hgsubstate
   M foo/bar/z.txt
   M foo/y.txt
   $ hg diff --nodates -S -I '**/?.txt' --rev 0:1
   diff -r af048e97ade2 -r d254738c5f5e foo/y.txt
   --- a/foo/y.txt
   +++ b/foo/y.txt
   @@ -1,1 +1,2 @@
    y1
   +y2
   diff -r 4904098473f9 -r 9647f22de499 foo/bar/z.txt
   --- a/foo/bar/z.txt
   +++ b/foo/bar/z.txt
   @@ -1,1 +1,2 @@
    z1
   +z2
 
 #if serve
   $ cd ..
   $ hg serve -R repo --debug -S -p $HGPORT -d --pid-file=hg1.pid -E error.log -A access.log
   adding  = $TESTTMP/repo
   adding foo = $TESTTMP/repo/foo
   adding foo/bar = $TESTTMP/repo/foo/bar
   listening at http://*:$HGPORT/ (bound to *:$HGPORT) (glob) (?)
   adding  = $TESTTMP/repo (?)
   adding foo = $TESTTMP/repo/foo (?)
   adding foo/bar = $TESTTMP/repo/foo/bar (?)
   $ cat hg1.pid >> $DAEMON_PIDS
 
   $ hg clone http://localhost:$HGPORT clone  --config progress.disable=True
   requesting all changes
   adding changesets
   adding manifests
   adding file changes
   added 3 changesets with 5 changes to 3 files
   new changesets 23376cbba0d8:1326fa26d0c0
   updating to branch default
   cloning subrepo foo from http://localhost:$HGPORT/foo
   requesting all changes
   adding changesets
   adding manifests
   adding file changes
   added 4 changesets with 7 changes to 3 files
   new changesets af048e97ade2:65903cebad86
   cloning subrepo foo/bar from http://localhost:$HGPORT/foo/bar
   requesting all changes
   adding changesets
   adding manifests
   adding file changes
   added 3 changesets with 3 changes to 1 files
   new changesets 4904098473f9:31ecbdafd357
   3 files updated, 0 files merged, 0 files removed, 0 files unresolved
 
   $ cat clone/foo/bar/z.txt
   z1
   z2
   z3
 
+Clone pooling from a remote URL will share the top level repo and the subrepos,
+even if they are referenced by remote URL.
+
+  $ hg --config extensions.share= --config share.pool=$TESTTMP/pool \
+  >    clone http://localhost:$HGPORT shared
+  (sharing from new pooled repository 23376cbba0d87c15906bb3652584927c140907bf)
+  requesting all changes
+  adding changesets
+  adding manifests
+  adding file changes
+  added 3 changesets with 5 changes to 3 files
+  new changesets 23376cbba0d8:1326fa26d0c0
+  searching for changes
+  no changes found
+  updating working directory
+  cloning subrepo foo from http://localhost:$HGPORT/foo
+  (sharing from new pooled repository af048e97ade2e236f754f05d07013e586af0f8bf)
+  requesting all changes
+  adding changesets
+  adding manifests
+  adding file changes
+  added 4 changesets with 7 changes to 3 files
+  new changesets af048e97ade2:65903cebad86
+  searching for changes
+  no changes found
+  cloning subrepo foo/bar from http://localhost:$HGPORT/foo/bar
+  (sharing from new pooled repository 4904098473f96c900fec436dad267edd4da59fad)
+  requesting all changes
+  adding changesets
+  adding manifests
+  adding file changes
+  added 3 changesets with 3 changes to 1 files
+  new changesets 4904098473f9:31ecbdafd357
+  searching for changes
+  no changes found
+  3 files updated, 0 files merged, 0 files removed, 0 files unresolved
+
   $ cat access.log
   * "GET /?cmd=capabilities HTTP/1.1" 200 - (glob)
   * "GET /?cmd=batch HTTP/1.1" 200 - * (glob)
   * "GET /?cmd=getbundle HTTP/1.1" 200 - * (glob)
   * "GET /foo?cmd=capabilities HTTP/1.1" 200 - (glob)
   * "GET /foo?cmd=batch HTTP/1.1" 200 - * (glob)
   * "GET /foo?cmd=getbundle HTTP/1.1" 200 - * (glob)
   * "GET /foo/bar?cmd=capabilities HTTP/1.1" 200 - (glob)
   * "GET /foo/bar?cmd=batch HTTP/1.1" 200 - * (glob)
   * "GET /foo/bar?cmd=getbundle HTTP/1.1" 200 - * (glob)
+  $LOCALIP - - [$LOGDATE$] "GET /?cmd=capabilities HTTP/1.1" 200 - (glob)
+  $LOCALIP - - [$LOGDATE$] "GET /?cmd=lookup HTTP/1.1" 200 - x-hgarg-1:key=0 x-hgproto-1:0.1 0.2 comp=$USUAL_COMPRESSIONS$ (glob)
+  $LOCALIP - - [$LOGDATE$] "GET /?cmd=capabilities HTTP/1.1" 200 - (glob)
+  $LOCALIP - - [$LOGDATE$] "GET /?cmd=batch HTTP/1.1" 200 - x-hgarg-1:cmds=heads+%3Bknown+nodes%3D x-hgproto-1:0.1 0.2 comp=$USUAL_COMPRESSIONS$ (glob)
+  $LOCALIP - - [$LOGDATE$] "GET /?cmd=getbundle HTTP/1.1" 200 - x-hgarg-1:bookmarks=1&$USUAL_BUNDLE_CAPS$&cg=1&common=0000000000000000000000000000000000000000&heads=1326fa26d0c00d2146c63b56bb6a45149d7325ac&listkeys=bookmarks&phases=1 x-hgproto-1:0.1 0.2 comp=$USUAL_COMPRESSIONS$ (glob)
+  $LOCALIP - - [$LOGDATE$] "GET /?cmd=batch HTTP/1.1" 200 - x-hgarg-1:cmds=heads+%3Bknown+nodes%3D1326fa26d0c00d2146c63b56bb6a45149d7325ac x-hgproto-1:0.1 0.2 comp=$USUAL_COMPRESSIONS$ (glob)
+  $LOCALIP - - [$LOGDATE$] "GET /?cmd=getbundle HTTP/1.1" 200 - x-hgarg-1:bookmarks=1&$USUAL_BUNDLE_CAPS$&cg=0&common=1326fa26d0c00d2146c63b56bb6a45149d7325ac&heads=1326fa26d0c00d2146c63b56bb6a45149d7325ac&listkeys=bookmarks&phases=1 x-hgproto-1:0.1 0.2 comp=$USUAL_COMPRESSIONS$ (glob)
+  $LOCALIP - - [$LOGDATE$] "GET /foo?cmd=capabilities HTTP/1.1" 200 - (glob)
+  $LOCALIP - - [$LOGDATE$] "GET /foo?cmd=lookup HTTP/1.1" 200 - x-hgarg-1:key=0 x-hgproto-1:0.1 0.2 comp=$USUAL_COMPRESSIONS$ (glob)
+  $LOCALIP - - [$LOGDATE$] "GET /foo?cmd=capabilities HTTP/1.1" 200 - (glob)
+  $LOCALIP - - [$LOGDATE$] "GET /foo?cmd=batch HTTP/1.1" 200 - x-hgarg-1:cmds=heads+%3Bknown+nodes%3D x-hgproto-1:0.1 0.2 comp=$USUAL_COMPRESSIONS$ (glob)
+  $LOCALIP - - [$LOGDATE$] "GET /foo?cmd=getbundle HTTP/1.1" 200 - x-hgarg-1:bookmarks=1&$USUAL_BUNDLE_CAPS$&cg=1&common=0000000000000000000000000000000000000000&heads=65903cebad86f1a84bd4f1134f62fa7dcb7a1c98&listkeys=bookmarks&phases=1 x-hgproto-1:0.1 0.2 comp=$USUAL_COMPRESSIONS$ (glob)
+  $LOCALIP - - [$LOGDATE$] "GET /foo?cmd=batch HTTP/1.1" 200 - x-hgarg-1:cmds=heads+%3Bknown+nodes%3D65903cebad86f1a84bd4f1134f62fa7dcb7a1c98 x-hgproto-1:0.1 0.2 comp=$USUAL_COMPRESSIONS$ (glob)
+  $LOCALIP - - [$LOGDATE$] "GET /foo?cmd=getbundle HTTP/1.1" 200 - x-hgarg-1:bookmarks=1&$USUAL_BUNDLE_CAPS$&cg=0&common=65903cebad86f1a84bd4f1134f62fa7dcb7a1c98&heads=65903cebad86f1a84bd4f1134f62fa7dcb7a1c98&listkeys=bookmarks&phases=1 x-hgproto-1:0.1 0.2 comp=$USUAL_COMPRESSIONS$ (glob)
+  $LOCALIP - - [$LOGDATE$] "GET /foo/bar?cmd=capabilities HTTP/1.1" 200 - (glob)
+  $LOCALIP - - [$LOGDATE$] "GET /foo/bar?cmd=lookup HTTP/1.1" 200 - x-hgarg-1:key=0 x-hgproto-1:0.1 0.2 comp=$USUAL_COMPRESSIONS$ (glob)
+  $LOCALIP - - [$LOGDATE$] "GET /foo/bar?cmd=capabilities HTTP/1.1" 200 - (glob)
+  $LOCALIP - - [$LOGDATE$] "GET /foo/bar?cmd=batch HTTP/1.1" 200 - x-hgarg-1:cmds=heads+%3Bknown+nodes%3D x-hgproto-1:0.1 0.2 comp=$USUAL_COMPRESSIONS$ (glob)
+  $LOCALIP - - [$LOGDATE$] "GET /foo/bar?cmd=getbundle HTTP/1.1" 200 - x-hgarg-1:bookmarks=1&$USUAL_BUNDLE_CAPS$&cg=1&common=0000000000000000000000000000000000000000&heads=31ecbdafd357f54b281c9bd1d681bb90de219e22&listkeys=bookmarks&phases=1 x-hgproto-1:0.1 0.2 comp=$USUAL_COMPRESSIONS$ (glob)
+  $LOCALIP - - [$LOGDATE$] "GET /foo/bar?cmd=batch HTTP/1.1" 200 - x-hgarg-1:cmds=heads+%3Bknown+nodes%3D31ecbdafd357f54b281c9bd1d681bb90de219e22 x-hgproto-1:0.1 0.2 comp=$USUAL_COMPRESSIONS$ (glob)
+  $LOCALIP - - [$LOGDATE$] "GET /foo/bar?cmd=getbundle HTTP/1.1" 200 - x-hgarg-1:bookmarks=1&$USUAL_BUNDLE_CAPS$&cg=0&common=31ecbdafd357f54b281c9bd1d681bb90de219e22&heads=31ecbdafd357f54b281c9bd1d681bb90de219e22&listkeys=bookmarks&phases=1 x-hgproto-1:0.1 0.2 comp=$USUAL_COMPRESSIONS$ (glob)
 
   $ killdaemons.py
   $ rm hg1.pid error.log access.log
   $ cd repo
 #endif
 
 Enable progress extension for archive tests:
 
   $ cp $HGRCPATH $HGRCPATH.no-progress
   $ cat >> $HGRCPATH <<EOF
   > [progress]
   > disable=False
   > assume-tty = 1
   > delay = 0
   > # set changedelay really large so we don't see nested topics
   > changedelay = 30000
   > format = topic bar number
   > refresh = 0
   > width = 60
   > EOF
 
 Test archiving to a directory tree (the doubled lines in the output
 only show up in the test output, not in real usage):
 
   $ hg archive --subrepos ../archive
   \r (no-eol) (esc)
   archiving [                                           ] 0/3\r (no-eol) (esc)
   archiving [=============>                             ] 1/3\r (no-eol) (esc)
   archiving [===========================>               ] 2/3\r (no-eol) (esc)
   archiving [==========================================>] 3/3\r (no-eol) (esc)
                                                               \r (no-eol) (esc)
   \r (no-eol) (esc)
   archiving (foo) [                                     ] 0/3\r (no-eol) (esc)
   archiving (foo) [===========>                         ] 1/3\r (no-eol) (esc)
   archiving (foo) [=======================>             ] 2/3\r (no-eol) (esc)
   archiving (foo) [====================================>] 3/3\r (no-eol) (esc)
                                                               \r (no-eol) (esc)
   \r (no-eol) (esc)
   archiving (foo/bar) [                                 ] 0/1\r (no-eol) (esc)
   archiving (foo/bar) [================================>] 1/1\r (no-eol) (esc)
                                                               \r (no-eol) (esc)
   $ find ../archive | sort
   ../archive
   ../archive/.hg_archival.txt
   ../archive/.hgsub
   ../archive/.hgsubstate
   ../archive/foo
   ../archive/foo/.hgsub
   ../archive/foo/.hgsubstate
   ../archive/foo/bar
   ../archive/foo/bar/z.txt
   ../archive/foo/y.txt
   ../archive/x.txt
 
 Test archiving to zip file (unzip output is unstable):
 
   $ hg archive --subrepos --prefix '.' ../archive.zip
   \r (no-eol) (esc)
   archiving [                                           ] 0/3\r (no-eol) (esc)
   archiving [=============>                             ] 1/3\r (no-eol) (esc)
   archiving [===========================>               ] 2/3\r (no-eol) (esc)
   archiving [==========================================>] 3/3\r (no-eol) (esc)
                                                               \r (no-eol) (esc)
   \r (no-eol) (esc)
   archiving (foo) [                                     ] 0/3\r (no-eol) (esc)
   archiving (foo) [===========>                         ] 1/3\r (no-eol) (esc)
   archiving (foo) [=======================>             ] 2/3\r (no-eol) (esc)
   archiving (foo) [====================================>] 3/3\r (no-eol) (esc)
                                                               \r (no-eol) (esc)
   \r (no-eol) (esc)
   archiving (foo/bar) [                                 ] 0/1\r (no-eol) (esc)
   archiving (foo/bar) [================================>] 1/1\r (no-eol) (esc)
                                                               \r (no-eol) (esc)
 
 (unzip date formating is unstable, we do not care about it and glob it out)
 
   $ unzip -l ../archive.zip | grep -v -- ----- | egrep -v files$
   Archive:  ../archive.zip
     Length [ ]* Date [ ]* Time [ ]* Name (re)
         172  [0-9:\- ]*  .hg_archival.txt (re)
          10  [0-9:\- ]*  .hgsub (re)
          45  [0-9:\- ]*  .hgsubstate (re)
           3  [0-9:\- ]*  x.txt (re)
          10  [0-9:\- ]*  foo/.hgsub (re)
          45  [0-9:\- ]*  foo/.hgsubstate (re)
           9  [0-9:\- ]*  foo/y.txt (re)
           9  [0-9:\- ]*  foo/bar/z.txt (re)
 
 Test archiving a revision that references a subrepo that is not yet
 cloned:
 
 #if hardlink
   $ hg clone -U . ../empty
   \r (no-eol) (esc)
   linking [ <=>                                           ] 1\r (no-eol) (esc)
   linking [  <=>                                          ] 2\r (no-eol) (esc)
   linking [   <=>                                         ] 3\r (no-eol) (esc)
   linking [    <=>                                        ] 4\r (no-eol) (esc)
   linking [     <=>                                       ] 5\r (no-eol) (esc)
   linking [      <=>                                      ] 6\r (no-eol) (esc)
   linking [       <=>                                     ] 7\r (no-eol) (esc)
   linking [        <=>                                    ] 8\r (no-eol) (esc)
                                                               \r (no-eol) (esc)
 #else
   $ hg clone -U . ../empty
   \r (no-eol) (esc)
   linking [ <=>                                           ] 1 (no-eol)
 #endif
 
   $ cd ../empty
 #if hardlink
   $ hg archive --subrepos -r tip --prefix './' ../archive.tar.gz
   \r (no-eol) (esc)
   archiving [                                           ] 0/3\r (no-eol) (esc)
   archiving [=============>                             ] 1/3\r (no-eol) (esc)
   archiving [===========================>               ] 2/3\r (no-eol) (esc)
   archiving [==========================================>] 3/3\r (no-eol) (esc)
                                                               \r (no-eol) (esc)
   \r (no-eol) (esc)
   linking [ <=>                                           ] 1\r (no-eol) (esc)
   linking [  <=>                                          ] 2\r (no-eol) (esc)
   linking [   <=>                                         ] 3\r (no-eol) (esc)
   linking [    <=>                                        ] 4\r (no-eol) (esc)
   linking [     <=>                                       ] 5\r (no-eol) (esc)
   linking [      <=>                                      ] 6\r (no-eol) (esc)
   linking [       <=>                                     ] 7\r (no-eol) (esc)
   linking [        <=>                                    ] 8\r (no-eol) (esc)
                                                               \r (no-eol) (esc)
   \r (no-eol) (esc)
   archiving (foo) [                                     ] 0/3\r (no-eol) (esc)
   archiving (foo) [===========>                         ] 1/3\r (no-eol) (esc)
   archiving (foo) [=======================>             ] 2/3\r (no-eol) (esc)
   archiving (foo) [====================================>] 3/3\r (no-eol) (esc)
                                                               \r (no-eol) (esc)
   \r (no-eol) (esc)
   linking [ <=>                                           ] 1\r (no-eol) (esc)
   linking [  <=>                                          ] 2\r (no-eol) (esc)
   linking [   <=>                                         ] 3\r (no-eol) (esc)
   linking [    <=>                                        ] 4\r (no-eol) (esc)
   linking [     <=>                                       ] 5\r (no-eol) (esc)
   linking [      <=>                                      ] 6\r (no-eol) (esc)
                                                               \r (no-eol) (esc)
   \r (no-eol) (esc)
   archiving (foo/bar) [                                 ] 0/1\r (no-eol) (esc)
   archiving (foo/bar) [================================>] 1/1\r (no-eol) (esc)
                                                               \r (no-eol) (esc)
   cloning subrepo foo from $TESTTMP/repo/foo
   cloning subrepo foo/bar from $TESTTMP/repo/foo/bar
 #else
 Note there's a slight output glitch on non-hardlink systems: the last
 "linking" progress topic never gets closed, leading to slight output corruption on that platform.
   $ hg archive --subrepos -r tip --prefix './' ../archive.tar.gz
   \r (no-eol) (esc)
   archiving [                                           ] 0/3\r (no-eol) (esc)
   archiving [=============>                             ] 1/3\r (no-eol) (esc)
   archiving [===========================>               ] 2/3\r (no-eol) (esc)
   archiving [==========================================>] 3/3\r (no-eol) (esc)
                                                               \r (no-eol) (esc)
   \r (no-eol) (esc)
   linking [ <=>                                           ] 1\r (no-eol) (esc)
   cloning subrepo foo/bar from $TESTTMP/repo/foo/bar
 #endif
 
 Archive + subrepos uses '/' for all component separators
 
   $ tar -tzf ../archive.tar.gz | sort
   .hg_archival.txt
   .hgsub
   .hgsubstate
   foo/.hgsub
   foo/.hgsubstate
   foo/bar/z.txt
   foo/y.txt
   x.txt
 
 The newly cloned subrepos contain no working copy:
 
   $ hg -R foo summary
   parent: -1:000000000000  (no revision checked out)
   branch: default
   commit: (clean)
   update: 4 new changesets (update)
 
+Sharing a local repo without the locally referenced subrepo (i.e. it was never
+updated from null), fails the same as a clone operation.
+
+  $ hg --config progress.disable=True clone -U ../empty ../empty2
+
+  $ hg --config extensions.share= --config progress.disable=True \
+  >    share ../empty2 ../empty_share
+  updating working directory
+  abort: repository $TESTTMP/empty2/foo not found!
+  [255]
+
+  $ hg --config progress.disable=True clone ../empty2 ../empty_clone
+  updating to branch default
+  abort: repository $TESTTMP/empty2/foo not found!
+  [255]
+
 Disable progress extension and cleanup:
 
   $ mv $HGRCPATH.no-progress $HGRCPATH
 
 Test archiving when there is a directory in the way for a subrepo
 created by archive:
 
   $ hg clone -U . ../almost-empty
   $ cd ../almost-empty
   $ mkdir foo
   $ echo f > foo/f
   $ hg archive --subrepos -r tip archive
   cloning subrepo foo from $TESTTMP/empty/foo
   abort: destination '$TESTTMP/almost-empty/foo' is not empty (in subrepository "foo")
   [255]
 
 Clone and test outgoing:
 
   $ cd ..
   $ hg clone repo repo2
   updating to branch default
   cloning subrepo foo from $TESTTMP/repo/foo
   cloning subrepo foo/bar from $TESTTMP/repo/foo/bar
   3 files updated, 0 files merged, 0 files removed, 0 files unresolved
   $ cd repo2
   $ hg outgoing -S
   comparing with $TESTTMP/repo
   searching for changes
   no changes found
   comparing with $TESTTMP/repo/foo
   searching for changes
   no changes found
   comparing with $TESTTMP/repo/foo/bar
   searching for changes
   no changes found
   [1]
 
 Make nested change:
 
   $ echo y4 >> foo/y.txt
   $ hg diff --nodates -S
   diff -r 65903cebad86 foo/y.txt
   --- a/foo/y.txt
   +++ b/foo/y.txt
   @@ -1,3 +1,4 @@
    y1
    y2
    y3
   +y4
   $ hg commit --subrepos -m 3-4-2
   committing subrepository foo
   $ hg outgoing -S
   comparing with $TESTTMP/repo
   searching for changes
   changeset:   3:2655b8ecc4ee
   tag:         tip
   user:        test
   date:        Thu Jan 01 00:00:00 1970 +0000
   summary:     3-4-2
   
   comparing with $TESTTMP/repo/foo
   searching for changes
   changeset:   4:e96193d6cb36
   tag:         tip
   user:        test
   date:        Thu Jan 01 00:00:00 1970 +0000
   summary:     3-4-2
   
   comparing with $TESTTMP/repo/foo/bar
   searching for changes
   no changes found
 
 
 Switch to original repo and setup default path:
 
   $ cd ../repo
   $ echo '[paths]' >> .hg/hgrc
   $ echo 'default = ../repo2' >> .hg/hgrc
 
 Test incoming:
 
   $ hg incoming -S
   comparing with $TESTTMP/repo2
   searching for changes
   changeset:   3:2655b8ecc4ee
   tag:         tip
   user:        test
   date:        Thu Jan 01 00:00:00 1970 +0000
   summary:     3-4-2
   
   comparing with $TESTTMP/repo2/foo
   searching for changes
   changeset:   4:e96193d6cb36
   tag:         tip
   user:        test
   date:        Thu Jan 01 00:00:00 1970 +0000
   summary:     3-4-2
   
   comparing with $TESTTMP/repo2/foo/bar
   searching for changes
   no changes found
 
   $ hg incoming -S --bundle incoming.hg
   abort: cannot combine --bundle and --subrepos
   [255]
 
 Test missing subrepo:
 
   $ rm -r foo
   $ hg status -S
   warning: error "unknown revision '65903cebad86f1a84bd4f1134f62fa7dcb7a1c98'" in subrepository "foo"
 
 Issue2619: IndexError: list index out of range on hg add with subrepos
 The subrepo must sorts after the explicit filename.
 
   $ cd ..
   $ hg init test
   $ cd test
   $ hg init x
   $ echo abc > abc.txt
   $ hg ci -Am "abc"
   adding abc.txt
   $ echo "x = x" >> .hgsub
   $ hg add .hgsub
   $ touch a x/a
   $ hg add a x/a
 
   $ hg ci -Sm "added x"
   committing subrepository x
   $ echo abc > x/a
   $ hg revert --rev '.^' "set:subrepo('glob:x*')"
   abort: subrepository 'x' does not exist in 25ac2c9b3180!
   [255]
 
   $ cd ..

tests/test-subrepo-relative-path.t

@@ -1,109 +1,190 b''
 #require killdaemons
 
 Preparing the subrepository 'sub'
 
   $ hg init sub
   $ echo sub > sub/sub
   $ hg add -R sub
   adding sub/sub
   $ hg commit -R sub -m "sub import"
 
 Preparing the 'main' repo which depends on the subrepo 'sub'
 
   $ hg init main
   $ echo main > main/main
   $ echo "sub = ../sub" > main/.hgsub
   $ hg clone sub main/sub
   updating to branch default
   1 files updated, 0 files merged, 0 files removed, 0 files unresolved
   $ hg add -R main
   adding main/.hgsub
   adding main/main
   $ hg commit -R main -m "main import"
 
 Cleaning both repositories, just as a clone -U
 
   $ hg up -C -R sub null
   0 files updated, 0 files merged, 1 files removed, 0 files unresolved
   $ hg up -C -R main null
   0 files updated, 0 files merged, 3 files removed, 0 files unresolved
   $ rm -rf main/sub
 
 hide outer repo
   $ hg init
 
 Serving them both using hgweb
 
   $ printf '[paths]\n/main = main\nsub = sub\n' > webdir.conf
   $ hg serve --webdir-conf webdir.conf -a localhost -p $HGPORT \
   >    -A /dev/null -E /dev/null --pid-file hg.pid -d
   $ cat hg.pid >> $DAEMON_PIDS
 
 Clone main from hgweb
 
   $ hg clone "http://localhost:$HGPORT/main" cloned
   requesting all changes
   adding changesets
   adding manifests
   adding file changes
   added 1 changesets with 3 changes to 3 files
   new changesets fdfeeb3e979e
   updating to branch default
   cloning subrepo sub from http://localhost:$HGPORT/sub
   requesting all changes
   adding changesets
   adding manifests
   adding file changes
   added 1 changesets with 1 changes to 1 files
   new changesets 863c1745b441
   3 files updated, 0 files merged, 0 files removed, 0 files unresolved
 
 Checking cloned repo ids
 
   $ hg id -R cloned
   fdfeeb3e979e tip
   $ hg id -R cloned/sub
   863c1745b441 tip
 
 subrepo debug for 'main' clone
 
   $ hg debugsub -R cloned
   path sub
    source   ../sub
    revision 863c1745b441bd97a8c4a096e87793073f4fb215
 
+Test sharing with a remote URL reference
+
+  $ hg init absolute_subrepo
+  $ cd absolute_subrepo
+  $ echo foo > foo.txt
+  $ hg ci -Am 'initial commit'
+  adding foo.txt
+  $ echo "sub = http://localhost:$HGPORT/sub" > .hgsub
+  $ hg ci -Am 'add absolute subrepo'
+  adding .hgsub
+  $ cd ..
+
+Clone pooling works for local clones with a remote subrepo reference.  The
+subrepo is cloned to the pool and shared from there, so that all clones will
+share the same subrepo.
+
+  $ hg --config extensions.share= --config share.pool=$TESTTMP/pool \
+  >    clone absolute_subrepo cloned_from_abs
+  (sharing from new pooled repository 8d6a2f1e993b34b6557de0042cfe825ae12a8dae)
+  requesting all changes
+  adding changesets
+  adding manifests
+  adding file changes
+  added 2 changesets with 3 changes to 3 files
+  new changesets 8d6a2f1e993b:* (glob)
+  searching for changes
+  no changes found
+  updating working directory
+  cloning subrepo sub from http://localhost:$HGPORT/sub
+  (sharing from new pooled repository 863c1745b441bd97a8c4a096e87793073f4fb215)
+  requesting all changes
+  adding changesets
+  adding manifests
+  adding file changes
+  added 1 changesets with 1 changes to 1 files
+  new changesets 863c1745b441
+  searching for changes
+  no changes found
+  3 files updated, 0 files merged, 0 files removed, 0 files unresolved
+
+Vanilla sharing with a subrepo remote path reference will clone the subrepo.
+Each share of these top level repos will end up with independent subrepo copies
+(potentially leaving the shared parent with dangling cset references).
+
+  $ hg --config extensions.share= share absolute_subrepo shared_from_abs
+  updating working directory
+  cloning subrepo sub from http://localhost:$HGPORT/sub
+  requesting all changes
+  adding changesets
+  adding manifests
+  adding file changes
+  added 1 changesets with 1 changes to 1 files
+  new changesets 863c1745b441
+  3 files updated, 0 files merged, 0 files removed, 0 files unresolved
+
+  $ hg --config extensions.share= share -U absolute_subrepo shared_from_abs2
+  $ hg -R shared_from_abs2 update -r tip
+  cloning subrepo sub from http://localhost:$HGPORT/sub
+  requesting all changes
+  adding changesets
+  adding manifests
+  adding file changes
+  added 1 changesets with 1 changes to 1 files
+  new changesets 863c1745b441
+  3 files updated, 0 files merged, 0 files removed, 0 files unresolved
+
+A parent repo without its subrepo available locally can be shared if the
+subrepo is referenced by absolute path.
+
+  $ hg clone -U absolute_subrepo cloned_null_from_abs
+  $ hg --config extensions.share= share cloned_null_from_abs shared_from_null_abs
+  updating working directory
+  cloning subrepo sub from http://localhost:$HGPORT/sub
+  requesting all changes
+  adding changesets
+  adding manifests
+  adding file changes
+  added 1 changesets with 1 changes to 1 files
+  new changesets 863c1745b441
+  3 files updated, 0 files merged, 0 files removed, 0 files unresolved
+
   $ killdaemons.py
 
 subrepo paths with ssh urls
 
   $ hg clone -e "\"$PYTHON\" \"$TESTDIR/dummyssh\"" ssh://user@dummy/cloned sshclone
   requesting all changes
   adding changesets
   adding manifests
   adding file changes
   added 1 changesets with 3 changes to 3 files
   new changesets fdfeeb3e979e
   updating to branch default
   cloning subrepo sub from ssh://user@dummy/sub
   requesting all changes
   adding changesets
   adding manifests
   adding file changes
   added 1 changesets with 1 changes to 1 files
   new changesets 863c1745b441
   3 files updated, 0 files merged, 0 files removed, 0 files unresolved
 
   $ hg -R sshclone push -e "\"$PYTHON\" \"$TESTDIR/dummyssh\"" ssh://user@dummy/`pwd`/cloned
   pushing to ssh://user@dummy/$TESTTMP/cloned
   pushing subrepo sub to ssh://user@dummy/$TESTTMP/sub
   searching for changes
   no changes found
   searching for changes
   no changes found
   [1]
 
   $ cat dummylog
   Got arguments 1:user@dummy 2:hg -R cloned serve --stdio
   Got arguments 1:user@dummy 2:hg -R sub serve --stdio
   Got arguments 1:user@dummy 2:hg -R $TESTTMP/cloned serve --stdio
   Got arguments 1:user@dummy 2:hg -R $TESTTMP/sub serve --stdio