patch: accept second matcher that applies only to copy sources (API)

See previous patch for motivation.

Differential Revision: https://phab.mercurial-scm.org/D5893

mercurial/context.py

@@ -1,2485 +1,2486 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 stat
 
 from .i18n import _
 from .node import (
     addednodeid,
     hex,
     modifiednodeid,
     nullid,
     nullrev,
     short,
     wdirfilenodeids,
     wdirid,
 )
 from . import (
     dagop,
     encoding,
     error,
     fileset,
     match as matchmod,
     obsolete as obsmod,
     patch,
     pathutil,
     phases,
     pycompat,
     repoview,
     scmutil,
     sparse,
     subrepo,
     subrepoutil,
     util,
 )
 from .utils import (
     dateutil,
     stringutil,
 )
 
 propertycache = util.propertycache
 
 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 __init__(self, repo):
         self._repo = repo
 
     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 wdirfilenodeids:
                 # 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 matchfileset(self, expr, badfn=None):
         return fileset.match(self, expr, badfn=badfn)
 
     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 its ancestor is"""
         return self.rev() in obsmod.getrevs(self._repo, 'orphan')
 
     def phasedivergent(self):
         """True if the changeset tries to be a successor of a public changeset
 
         Only non-public and non-obsolete changesets may be phase-divergent.
         """
         return self.rev() in obsmod.getrevs(self._repo, 'phasedivergent')
 
     def contentdivergent(self):
         """Is a successor of a changeset with multiple possible successor sets
 
         Only non-public and non-obsolete changesets may be content-divergent.
         """
         return self.rev() in obsmod.getrevs(self._repo, 'contentdivergent')
 
     def isunstable(self):
         """True if the changeset is either orphan, phase-divergent or
         content-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 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, changes=None, opts=None,
              losedatafn=None, prefix='', relroot='', copy=None,
-             hunksfilterfn=None):
+             copysourcematch=None, hunksfilterfn=None):
         """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]
         return patch.diff(self._repo, ctx2, self, match=match, changes=changes,
                           opts=opts, losedatafn=losedatafn, prefix=prefix,
                           relroot=relroot, copy=copy,
+                          copysourcematch=copysourcematch,
                           hunksfilterfn=hunksfilterfn)
 
     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 = self._repo.narrowmatch(match)
         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
 
 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, rev, node):
         super(changectx, self).__init__(repo)
         self._rev = rev
         self._node = node
 
     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 [repo[p1]]
         return [repo[p1], 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 [self._repo[x] for x in c]
 
     def ancestors(self):
         for a in self._repo.changelog.ancestors([self._rev]):
             yield 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 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 = scmutil.revsymbol(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 self._repo[anc]
 
     def isancestorof(self, other):
         """True if this changeset is an ancestor of other"""
         return self._repo.changelog.isancestorrev(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(self._repo.narrowmatch(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,
     """
     @propertycache
     def _filelog(self):
         return self._repo.file(self._path)
 
     @propertycache
     def _changeid(self):
         if 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 r"<%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 stringutil.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 self._filenode is None:
             raise error.ProgrammingError(
                 'filectx.cmp() must be reimplemented if not backed by revlog')
 
         if fctx._filenode is None:
             if self._repo._encodefilterpats:
                 # can't rely on size() because wdir content may be decoded
                 return self._filelog.cmp(self._filenode, fctx.data())
             if self.size() - 4 == fctx.size():
                 # size() can match:
                 # if file data starts with '\1\n', empty metadata block is
                 # prepended, which adds 4 bytes to filelog.size().
                 return self._filelog.cmp(self._filenode, fctx.data())
         if self.size() == fctx.size():
             # size() matches: need to compare content
             return self._filelog.cmp(self._filenode, fctx.data())
 
         # size() differs
         return True
 
     def _adjustlinkrev(self, srcrev, inclusive=False, stoprev=None):
         """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
         :stoprev: an optional revision to stop the walk at. If no introduction
                   of this file content could be found before this floor
                   revision, the function will returns "None" and stops its
                   iteration.
         """
         repo = self._repo
         cl = repo.unfiltered().changelog
         mfl = repo.manifestlog
         # fetch the linkrev
         lkr = self.linkrev()
         if srcrev == lkr:
             return lkr
         # 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:
                 if stoprev is not None and a < stoprev:
                     return None
                 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 isintroducedafter(self, changelogrev):
         """True if a filectx has been introduced after a given floor revision
         """
         if self.linkrev() >= changelogrev:
             return True
         introrev = self._introrev(stoprev=changelogrev)
         if introrev is None:
             return False
         return introrev >= changelogrev
 
     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.
         """
         return self._introrev()
 
     def _introrev(self, stoprev=None):
         """
         Same as `introrev` but, with an extra argument to limit changelog
         iteration range in some internal usecase.
 
         If `stoprev` is set, the `introrev` will not be searched past that
         `stoprev` revision and "None" might be returned. This is useful to
         limit the iteration range.
         """
         toprev = None
         attrs = vars(self)
         if r'_changeid' in attrs:
             # We have a cached value already
             toprev = self._changeid
         elif r'_changectx' in attrs:
             # We know which changelog entry we are coming from
             toprev = self._changectx.rev()
 
         if toprev is not None:
             return self._adjustlinkrev(toprev, inclusive=True, stoprev=stoprev)
         elif r'_descendantrev' in attrs:
             introrev = self._adjustlinkrev(self._descendantrev, stoprev=stoprev)
             # be nice and cache the result of the computation
             if introrev is not None:
                 self._changeid = introrev
             return introrev
         else:
             return self.linkrev()
 
     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, skiprevs=None, diffopts=None):
         """Returns a list of annotateline objects for each line in the file
 
         - line.fctx is the filectx of the node where that line was last changed
         - line.lineno is the line number at the first appearance in the managed
           file
         - line.text is the data on that line (including newline character)
         """
         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
 
         return dagop.annotate(base, parents, skiprevs=skiprevs,
                               diffopts=diffopts)
 
     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())
 
 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 must be a revision number, if specified.
            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 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 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 None
 
         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):
         super(committablectx, self).__init__(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(self._repo.narrowmatch(match),
                                                subrepos=sorted(self.substate),
                                                unknown=True, ignored=False))
 
     def matches(self, match):
         match = self._repo.narrowmatch(match)
         ds = self._repo.dirstate
         return sorted(f for f in ds.matches(match) if ds[f] != 'r')
 
     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 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]
         # use unfiltered repo to delay/avoid loading obsmarkers
         unfi = self._repo.unfiltered()
         return [changectx(self._repo, unfi.changelog.rev(n), n) for n in p]
 
     def _fileinfo(self, path):
         # populate __dict__['_manifest'] as workingctx has no _manifestdelta
         self._manifest
         return super(workingctx, self)._fileinfo(path)
 
     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
                 limit = ui.configbytes('ui', 'large-file-limit')
                 if limit != 0 and st.st_size > limit:
                     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 ds:
                     self._repo.ui.warn(_("%s not tracked!\n") % uipath(f))
                     rejected.append(f)
                 elif ds[f] != 'a':
                     ds.remove(f)
                 else:
                     ds.drop(f)
             return rejected
 
     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():
                 ds = self._repo.dirstate
                 if ds[dest] in '?':
                     ds.add(dest)
                 elif ds[dest] in 'r':
                     ds.normallookup(dest)
                 ds.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 stringutil.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)[stat.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"""
         rmdir = self._repo.ui.configbool('experimental', 'removeemptydirs')
         self._repo.wvfs.unlinkpath(self._path, ignoremissing=ignoremissing,
                                    rmdir=rmdir)
 
     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 self._repo.ui.configbool('experimental', 'merge.checkpathconflicts'):
             # remove files under the directory as they should already be
             # warned and backed up
             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
         else:
             # don't remove files if path conflicts are not processed
             if wvfs.isdir(f) and not wvfs.islink(f):
                 wvfs.removedirs(f)
 
     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.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 __contains__(self, key):
         if key in self._cache:
             return self._cache[key]['exists']
         return key in self.p1()
 
     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 %d." % (path, component,
                                               self.p1().rev()))
             else:
                 raise error.Abort("error: '%s' conflicts with file '%s' in "
                                   "%d." % (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 pycompat.xrange(len(components)):
             component = "/".join(components[0:i])
             if component in self:
                 fail(path, component)
 
         # Test the other direction -- that this path from p2 isn't a directory
         # in p1 (test that p1 doesn't have any paths matching `path/*`).
         match = self.match(pats=[path + '/'], default=b'path')
         matches = self.p1().manifest().matches(match)
         mfiles = matches.keys()
         if len(mfiles) > 0:
             if len(mfiles) == 1 and mfiles[0] == path:
                 return
             # omit the files which are deleted in current IMM wctx
             mfiles = [m for m in mfiles if m in self]
             if not mfiles:
                 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(mfiles),
                                  ', '.join(mfiles)))
 
     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):
         flag = ''
         if l:
             flag = 'l'
         elif x:
             flag = 'x'
         self._markdirty(path, exists=True, date=dateutil.makedate(),
                         flags=flag)
 
     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 = []
         # This won't be perfect, but can help performance significantly when
         # using things like remotefilelog.
         scmutil.prefetchfiles(
             self.repo(), [self.p1().rev()],
             scmutil.matchfiles(self.repo(), self._cache.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=''):
         # data not provided, let's see if we already have some; if not, let's
         # grab it from our underlying context, so that we always have data if
         # the file is marked as existing.
         if exists and data is None:
             oldentry = self._cache.get(path) or {}
             data = oldentry.get('data') or self._wrappedctx[path].data()
 
         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(workingcommitctx, 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]
         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 = [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:
             man[f] = modifiednodeid
 
         for f in self._status.added:
             man[f] = addednodeid
 
         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
         if islink:
             self._flags = 'l'
         elif isexec:
             self._flags = 'x'
         else:
             self._flags = ''
         self._copied = None
         if copied:
             self._copied = (copied, nullid)
 
     def cmp(self, fctx):
         return self.data() != fctx.data()
 
     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 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 __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(r"can't reuse the manifest: its p1 "
                                r"doesn't match the new ctx p1")
         if p2 != nullid and p2.manifestnode() != mp2:
             raise RuntimeError(r"can't reuse the manifest: "
                                r"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, "wb") as f:
             f.write(data)

mercurial/logcmdutil.py

@@ -1,916 +1,919 b''
 # logcmdutil.py - utility for log-like commands
 #
 # Copyright 2005-2007 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 itertools
 import os
 
 from .i18n import _
 from .node import (
     nullid,
     wdirid,
     wdirrev,
 )
 
 from . import (
     dagop,
     error,
     formatter,
     graphmod,
     match as matchmod,
     mdiff,
     patch,
     pathutil,
     pycompat,
     revset,
     revsetlang,
     scmutil,
     smartset,
     templatekw,
     templater,
     util,
 )
 from .utils import (
     dateutil,
     stringutil,
 )
 
 def getlimit(opts):
     """get the log limit according to option -l/--limit"""
     limit = opts.get('limit')
     if limit:
         try:
             limit = int(limit)
         except ValueError:
             raise error.Abort(_('limit must be a positive integer'))
         if limit <= 0:
             raise error.Abort(_('limit must be positive'))
     else:
         limit = None
     return limit
 
 def diffordiffstat(ui, repo, diffopts, node1, node2, match,
                    changes=None, stat=False, fp=None, graphwidth=0,
                    prefix='', root='', listsubrepos=False, hunksfilterfn=None):
     '''show diff or diffstat.'''
     ctx1 = repo[node1]
     ctx2 = repo[node2]
     if root:
         relroot = pathutil.canonpath(repo.root, repo.getcwd(), root)
     else:
         relroot = ''
+    copysourcematch = None
     if relroot != '':
         # XXX relative roots currently don't work if the root is within a
         # subrepo
         uirelroot = match.uipath(relroot)
         relroot += '/'
         for matchroot in match.files():
             if not matchroot.startswith(relroot):
                 ui.warn(_('warning: %s not inside relative root %s\n') % (
                     match.uipath(matchroot), uirelroot))
 
         relrootmatch = scmutil.match(ctx2, pats=[relroot], default='path')
         match = matchmod.intersectmatchers(match, relrootmatch)
+        copysourcematch = relrootmatch
 
     if stat:
         diffopts = diffopts.copy(context=0, noprefix=False)
         width = 80
         if not ui.plain():
             width = ui.termwidth() - graphwidth
 
     chunks = ctx2.diff(ctx1, match, changes, opts=diffopts, prefix=prefix,
-                       relroot=relroot, hunksfilterfn=hunksfilterfn)
+                       relroot=relroot, copysourcematch=copysourcematch,
+                       hunksfilterfn=hunksfilterfn)
 
     if fp is not None or ui.canwritewithoutlabels():
         out = fp or ui
         if stat:
             chunks = [patch.diffstat(util.iterlines(chunks), width=width)]
         for chunk in util.filechunkiter(util.chunkbuffer(chunks)):
             out.write(chunk)
     else:
         if stat:
             chunks = patch.diffstatui(util.iterlines(chunks), width=width)
         else:
             chunks = patch.difflabel(lambda chunks, **kwargs: chunks, chunks,
                                      opts=diffopts)
         if ui.canbatchlabeledwrites():
             def gen():
                 for chunk, label in chunks:
                     yield ui.label(chunk, label=label)
             for chunk in util.filechunkiter(util.chunkbuffer(gen())):
                 ui.write(chunk)
         else:
             for chunk, label in chunks:
                 ui.write(chunk, label=label)
 
     if listsubrepos:
         for subpath, sub in scmutil.itersubrepos(ctx1, ctx2):
             tempnode2 = node2
             try:
                 if node2 is not None:
                     tempnode2 = ctx2.substate[subpath][1]
             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 is to ignore it.
                 tempnode2 = None
             submatch = matchmod.subdirmatcher(subpath, match)
             sub.diff(ui, diffopts, tempnode2, submatch, changes=changes,
                      stat=stat, fp=fp, prefix=prefix)
 
 class changesetdiffer(object):
     """Generate diff of changeset with pre-configured filtering functions"""
 
     def _makefilematcher(self, ctx):
         return scmutil.matchall(ctx.repo())
 
     def _makehunksfilter(self, ctx):
         return None
 
     def showdiff(self, ui, ctx, diffopts, graphwidth=0, stat=False):
         repo = ctx.repo()
         node = ctx.node()
         prev = ctx.p1().node()
         diffordiffstat(ui, repo, diffopts, prev, node,
                        match=self._makefilematcher(ctx), stat=stat,
                        graphwidth=graphwidth,
                        hunksfilterfn=self._makehunksfilter(ctx))
 
 def changesetlabels(ctx):
     labels = ['log.changeset', 'changeset.%s' % ctx.phasestr()]
     if ctx.obsolete():
         labels.append('changeset.obsolete')
     if ctx.isunstable():
         labels.append('changeset.unstable')
         for instability in ctx.instabilities():
             labels.append('instability.%s' % instability)
     return ' '.join(labels)
 
 class changesetprinter(object):
     '''show changeset information when templating not requested.'''
 
     def __init__(self, ui, repo, differ=None, diffopts=None, buffered=False):
         self.ui = ui
         self.repo = repo
         self.buffered = buffered
         self._differ = differ or changesetdiffer()
         self._diffopts = patch.diffallopts(ui, diffopts)
         self._includestat = diffopts and diffopts.get('stat')
         self._includediff = diffopts and diffopts.get('patch')
         self.header = {}
         self.hunk = {}
         self.lastheader = None
         self.footer = None
         self._columns = templatekw.getlogcolumns()
 
     def flush(self, ctx):
         rev = ctx.rev()
         if rev in self.header:
             h = self.header[rev]
             if h != self.lastheader:
                 self.lastheader = h
                 self.ui.write(h)
             del self.header[rev]
         if rev in self.hunk:
             self.ui.write(self.hunk[rev])
             del self.hunk[rev]
 
     def close(self):
         if self.footer:
             self.ui.write(self.footer)
 
     def show(self, ctx, copies=None, **props):
         props = pycompat.byteskwargs(props)
         if self.buffered:
             self.ui.pushbuffer(labeled=True)
             self._show(ctx, copies, props)
             self.hunk[ctx.rev()] = self.ui.popbuffer()
         else:
             self._show(ctx, copies, props)
 
     def _show(self, ctx, copies, props):
         '''show a single changeset or file revision'''
         changenode = ctx.node()
         graphwidth = props.get('graphwidth', 0)
 
         if self.ui.quiet:
             self.ui.write("%s\n" % scmutil.formatchangeid(ctx),
                           label='log.node')
             return
 
         columns = self._columns
         self.ui.write(columns['changeset'] % scmutil.formatchangeid(ctx),
                       label=changesetlabels(ctx))
 
         # branches are shown first before any other names due to backwards
         # compatibility
         branch = ctx.branch()
         # don't show the default branch name
         if branch != 'default':
             self.ui.write(columns['branch'] % branch, label='log.branch')
 
         for nsname, ns in self.repo.names.iteritems():
             # branches has special logic already handled above, so here we just
             # skip it
             if nsname == 'branches':
                 continue
             # we will use the templatename as the color name since those two
             # should be the same
             for name in ns.names(self.repo, changenode):
                 self.ui.write(ns.logfmt % name,
                               label='log.%s' % ns.colorname)
         if self.ui.debugflag:
             self.ui.write(columns['phase'] % ctx.phasestr(), label='log.phase')
         for pctx in scmutil.meaningfulparents(self.repo, ctx):
             label = 'log.parent changeset.%s' % pctx.phasestr()
             self.ui.write(columns['parent'] % scmutil.formatchangeid(pctx),
                           label=label)
 
         if self.ui.debugflag:
             mnode = ctx.manifestnode()
             if mnode is None:
                 mnode = wdirid
                 mrev = wdirrev
             else:
                 mrev = self.repo.manifestlog.rev(mnode)
             self.ui.write(columns['manifest']
                           % scmutil.formatrevnode(self.ui, mrev, mnode),
                           label='ui.debug log.manifest')
         self.ui.write(columns['user'] % ctx.user(), label='log.user')
         self.ui.write(columns['date'] % dateutil.datestr(ctx.date()),
                       label='log.date')
 
         if ctx.isunstable():
             instabilities = ctx.instabilities()
             self.ui.write(columns['instability'] % ', '.join(instabilities),
                           label='log.instability')
 
         elif ctx.obsolete():
             self._showobsfate(ctx)
 
         self._exthook(ctx)
 
         if self.ui.debugflag:
             files = ctx.p1().status(ctx)[:3]
             for key, value in zip(['files', 'files+', 'files-'], files):
                 if value:
                     self.ui.write(columns[key] % " ".join(value),
                                   label='ui.debug log.files')
         elif ctx.files() and self.ui.verbose:
             self.ui.write(columns['files'] % " ".join(ctx.files()),
                           label='ui.note log.files')
         if copies and self.ui.verbose:
             copies = ['%s (%s)' % c for c in copies]
             self.ui.write(columns['copies'] % ' '.join(copies),
                           label='ui.note log.copies')
 
         extra = ctx.extra()
         if extra and self.ui.debugflag:
             for key, value in sorted(extra.items()):
                 self.ui.write(columns['extra']
                               % (key, stringutil.escapestr(value)),
                               label='ui.debug log.extra')
 
         description = ctx.description().strip()
         if description:
             if self.ui.verbose:
                 self.ui.write(_("description:\n"),
                               label='ui.note log.description')
                 self.ui.write(description,
                               label='ui.note log.description')
                 self.ui.write("\n\n")
             else:
                 self.ui.write(columns['summary'] % description.splitlines()[0],
                               label='log.summary')
         self.ui.write("\n")
 
         self._showpatch(ctx, graphwidth)
 
     def _showobsfate(self, ctx):
         # TODO: do not depend on templater
         tres = formatter.templateresources(self.repo.ui, self.repo)
         t = formatter.maketemplater(self.repo.ui, '{join(obsfate, "\n")}',
                                     defaults=templatekw.keywords,
                                     resources=tres)
         obsfate = t.renderdefault({'ctx': ctx}).splitlines()
 
         if obsfate:
             for obsfateline in obsfate:
                 self.ui.write(self._columns['obsolete'] % obsfateline,
                               label='log.obsfate')
 
     def _exthook(self, ctx):
         '''empty method used by extension as a hook point
         '''
 
     def _showpatch(self, ctx, graphwidth=0):
         if self._includestat:
             self._differ.showdiff(self.ui, ctx, self._diffopts,
                                   graphwidth, stat=True)
         if self._includestat and self._includediff:
             self.ui.write("\n")
         if self._includediff:
             self._differ.showdiff(self.ui, ctx, self._diffopts,
                                   graphwidth, stat=False)
         if self._includestat or self._includediff:
             self.ui.write("\n")
 
 class changesetformatter(changesetprinter):
     """Format changeset information by generic formatter"""
 
     def __init__(self, ui, repo, fm, differ=None, diffopts=None,
                  buffered=False):
         changesetprinter.__init__(self, ui, repo, differ, diffopts, buffered)
         self._diffopts = patch.difffeatureopts(ui, diffopts, git=True)
         self._fm = fm
 
     def close(self):
         self._fm.end()
 
     def _show(self, ctx, copies, props):
         '''show a single changeset or file revision'''
         fm = self._fm
         fm.startitem()
         fm.context(ctx=ctx)
         fm.data(rev=scmutil.intrev(ctx),
                 node=fm.hexfunc(scmutil.binnode(ctx)))
 
         if self.ui.quiet:
             return
 
         fm.data(branch=ctx.branch(),
                 phase=ctx.phasestr(),
                 user=ctx.user(),
                 date=fm.formatdate(ctx.date()),
                 desc=ctx.description(),
                 bookmarks=fm.formatlist(ctx.bookmarks(), name='bookmark'),
                 tags=fm.formatlist(ctx.tags(), name='tag'),
                 parents=fm.formatlist([fm.hexfunc(c.node())
                                        for c in ctx.parents()], name='node'))
 
         if self.ui.debugflag:
             fm.data(manifest=fm.hexfunc(ctx.manifestnode() or wdirid),
                     extra=fm.formatdict(ctx.extra()))
 
             files = ctx.p1().status(ctx)
             fm.data(modified=fm.formatlist(files[0], name='file'),
                     added=fm.formatlist(files[1], name='file'),
                     removed=fm.formatlist(files[2], name='file'))
 
         elif self.ui.verbose:
             fm.data(files=fm.formatlist(ctx.files(), name='file'))
             if copies:
                 fm.data(copies=fm.formatdict(copies,
                                              key='name', value='source'))
 
         if self._includestat:
             self.ui.pushbuffer()
             self._differ.showdiff(self.ui, ctx, self._diffopts, stat=True)
             fm.data(diffstat=self.ui.popbuffer())
         if self._includediff:
             self.ui.pushbuffer()
             self._differ.showdiff(self.ui, ctx, self._diffopts, stat=False)
             fm.data(diff=self.ui.popbuffer())
 
 class changesettemplater(changesetprinter):
     '''format changeset information.
 
     Note: there are a variety of convenience functions to build a
     changesettemplater for common cases. See functions such as:
     maketemplater, changesetdisplayer, buildcommittemplate, or other
     functions that use changesest_templater.
     '''
 
     # Arguments before "buffered" used to be positional. Consider not
     # adding/removing arguments before "buffered" to not break callers.
     def __init__(self, ui, repo, tmplspec, differ=None, diffopts=None,
                  buffered=False):
         changesetprinter.__init__(self, ui, repo, differ, diffopts, buffered)
         # tres is shared with _graphnodeformatter()
         self._tresources = tres = formatter.templateresources(ui, repo)
         self.t = formatter.loadtemplater(ui, tmplspec,
                                          defaults=templatekw.keywords,
                                          resources=tres,
                                          cache=templatekw.defaulttempl)
         self._counter = itertools.count()
 
         self._tref = tmplspec.ref
         self._parts = {'header': '', 'footer': '',
                        tmplspec.ref: tmplspec.ref,
                        'docheader': '', 'docfooter': '',
                        'separator': ''}
         if tmplspec.mapfile:
             # find correct templates for current mode, for backward
             # compatibility with 'log -v/-q/--debug' using a mapfile
             tmplmodes = [
                 (True, ''),
                 (self.ui.verbose, '_verbose'),
                 (self.ui.quiet, '_quiet'),
                 (self.ui.debugflag, '_debug'),
             ]
             for mode, postfix in tmplmodes:
                 for t in self._parts:
                     cur = t + postfix
                     if mode and cur in self.t:
                         self._parts[t] = cur
         else:
             partnames = [p for p in self._parts.keys() if p != tmplspec.ref]
             m = formatter.templatepartsmap(tmplspec, self.t, partnames)
             self._parts.update(m)
 
         if self._parts['docheader']:
             self.ui.write(self.t.render(self._parts['docheader'], {}))
 
     def close(self):
         if self._parts['docfooter']:
             if not self.footer:
                 self.footer = ""
             self.footer += self.t.render(self._parts['docfooter'], {})
         return super(changesettemplater, self).close()
 
     def _show(self, ctx, copies, props):
         '''show a single changeset or file revision'''
         props = props.copy()
         props['ctx'] = ctx
         props['index'] = index = next(self._counter)
         props['revcache'] = {'copies': copies}
         graphwidth = props.get('graphwidth', 0)
 
         # write separator, which wouldn't work well with the header part below
         # since there's inherently a conflict between header (across items) and
         # separator (per item)
         if self._parts['separator'] and index > 0:
             self.ui.write(self.t.render(self._parts['separator'], {}))
 
         # write header
         if self._parts['header']:
             h = self.t.render(self._parts['header'], props)
             if self.buffered:
                 self.header[ctx.rev()] = h
             else:
                 if self.lastheader != h:
                     self.lastheader = h
                     self.ui.write(h)
 
         # write changeset metadata, then patch if requested
         key = self._parts[self._tref]
         self.ui.write(self.t.render(key, props))
         self._showpatch(ctx, graphwidth)
 
         if self._parts['footer']:
             if not self.footer:
                 self.footer = self.t.render(self._parts['footer'], props)
 
 def templatespec(tmpl, mapfile):
     if pycompat.ispy3:
         assert not isinstance(tmpl, str), 'tmpl must not be a str'
     if mapfile:
         return formatter.templatespec('changeset', tmpl, mapfile)
     else:
         return formatter.templatespec('', tmpl, None)
 
 def _lookuptemplate(ui, tmpl, style):
     """Find the template matching the given template spec or style
 
     See formatter.lookuptemplate() for details.
     """
 
     # ui settings
     if not tmpl and not style: # template are stronger than style
         tmpl = ui.config('ui', 'logtemplate')
         if tmpl:
             return templatespec(templater.unquotestring(tmpl), None)
         else:
             style = util.expandpath(ui.config('ui', 'style'))
 
     if not tmpl and style:
         mapfile = style
         if not os.path.split(mapfile)[0]:
             mapname = (templater.templatepath('map-cmdline.' + mapfile)
                        or templater.templatepath(mapfile))
             if mapname:
                 mapfile = mapname
         return templatespec(None, mapfile)
 
     if not tmpl:
         return templatespec(None, None)
 
     return formatter.lookuptemplate(ui, 'changeset', tmpl)
 
 def maketemplater(ui, repo, tmpl, buffered=False):
     """Create a changesettemplater from a literal template 'tmpl'
     byte-string."""
     spec = templatespec(tmpl, None)
     return changesettemplater(ui, repo, spec, buffered=buffered)
 
 def changesetdisplayer(ui, repo, opts, differ=None, buffered=False):
     """show one changeset using template or regular display.
 
     Display format will be the first non-empty hit of:
     1. option 'template'
     2. option 'style'
     3. [ui] setting 'logtemplate'
     4. [ui] setting 'style'
     If all of these values are either the unset or the empty string,
     regular display via changesetprinter() is done.
     """
     postargs = (differ, opts, buffered)
     if opts.get('template') == 'json':
         fm = ui.formatter('log', opts)
         return changesetformatter(ui, repo, fm, *postargs)
 
     spec = _lookuptemplate(ui, opts.get('template'), opts.get('style'))
 
     if not spec.ref and not spec.tmpl and not spec.mapfile:
         return changesetprinter(ui, repo, *postargs)
 
     return changesettemplater(ui, repo, spec, *postargs)
 
 def _makematcher(repo, revs, pats, opts):
     """Build matcher and expanded patterns from log options
 
     If --follow, revs are the revisions to follow from.
 
     Returns (match, pats, slowpath) where
     - match: a matcher built from the given pats and -I/-X opts
     - pats: patterns used (globs are expanded on Windows)
     - slowpath: True if patterns aren't as simple as scanning filelogs
     """
     # pats/include/exclude are passed to match.match() directly in
     # _matchfiles() revset but walkchangerevs() builds its matcher with
     # scmutil.match(). The difference is input pats are globbed on
     # platforms without shell expansion (windows).
     wctx = repo[None]
     match, pats = scmutil.matchandpats(wctx, pats, opts)
     slowpath = match.anypats() or (not match.always() and opts.get('removed'))
     if not slowpath:
         follow = opts.get('follow') or opts.get('follow_first')
         startctxs = []
         if follow and opts.get('rev'):
             startctxs = [repo[r] for r in revs]
         for f in match.files():
             if follow and startctxs:
                 # No idea if the path was a directory at that revision, so
                 # take the slow path.
                 if any(f not in c for c in startctxs):
                     slowpath = True
                     continue
             elif follow and f not in wctx:
                 # If the file exists, it may be a directory, so let it
                 # take the slow path.
                 if os.path.exists(repo.wjoin(f)):
                     slowpath = True
                     continue
                 else:
                     raise error.Abort(_('cannot follow file not in parent '
                                         'revision: "%s"') % f)
             filelog = repo.file(f)
             if not filelog:
                 # A zero count may be a directory or deleted file, so
                 # try to find matching entries on the slow path.
                 if follow:
                     raise error.Abort(
                         _('cannot follow nonexistent file: "%s"') % f)
                 slowpath = True
 
         # We decided to fall back to the slowpath because at least one
         # of the paths was not a file. Check to see if at least one of them
         # existed in history - in that case, we'll continue down the
         # slowpath; otherwise, we can turn off the slowpath
         if slowpath:
             for path in match.files():
                 if path == '.' or path in repo.store:
                     break
             else:
                 slowpath = False
 
     return match, pats, slowpath
 
 def _fileancestors(repo, revs, match, followfirst):
     fctxs = []
     for r in revs:
         ctx = repo[r]
         fctxs.extend(ctx[f].introfilectx() for f in ctx.walk(match))
 
     # When displaying a revision with --patch --follow FILE, we have
     # to know which file of the revision must be diffed. With
     # --follow, we want the names of the ancestors of FILE in the
     # revision, stored in "fcache". "fcache" is populated as a side effect
     # of the graph traversal.
     fcache = {}
     def filematcher(ctx):
         return scmutil.matchfiles(repo, fcache.get(ctx.rev(), []))
 
     def revgen():
         for rev, cs in dagop.filectxancestors(fctxs, followfirst=followfirst):
             fcache[rev] = [c.path() for c in cs]
             yield rev
     return smartset.generatorset(revgen(), iterasc=False), filematcher
 
 def _makenofollowfilematcher(repo, pats, opts):
     '''hook for extensions to override the filematcher for non-follow cases'''
     return None
 
 _opt2logrevset = {
     'no_merges':        ('not merge()', None),
     'only_merges':      ('merge()', None),
     '_matchfiles':      (None, '_matchfiles(%ps)'),
     'date':             ('date(%s)', None),
     'branch':           ('branch(%s)', '%lr'),
     '_patslog':         ('filelog(%s)', '%lr'),
     'keyword':          ('keyword(%s)', '%lr'),
     'prune':            ('ancestors(%s)', 'not %lr'),
     'user':             ('user(%s)', '%lr'),
 }
 
 def _makerevset(repo, match, pats, slowpath, opts):
     """Return a revset string built from log options and file patterns"""
     opts = dict(opts)
     # follow or not follow?
     follow = opts.get('follow') or opts.get('follow_first')
 
     # branch and only_branch are really aliases and must be handled at
     # the same time
     opts['branch'] = opts.get('branch', []) + opts.get('only_branch', [])
     opts['branch'] = [repo.lookupbranch(b) for b in opts['branch']]
 
     if slowpath:
         # See walkchangerevs() slow path.
         #
         # pats/include/exclude cannot be represented as separate
         # revset expressions as their filtering logic applies at file
         # level. For instance "-I a -X b" matches a revision touching
         # "a" and "b" while "file(a) and not file(b)" does
         # not. Besides, filesets are evaluated against the working
         # directory.
         matchargs = ['r:', 'd:relpath']
         for p in pats:
             matchargs.append('p:' + p)
         for p in opts.get('include', []):
             matchargs.append('i:' + p)
         for p in opts.get('exclude', []):
             matchargs.append('x:' + p)
         opts['_matchfiles'] = matchargs
     elif not follow:
         opts['_patslog'] = list(pats)
 
     expr = []
     for op, val in sorted(opts.iteritems()):
         if not val:
             continue
         if op not in _opt2logrevset:
             continue
         revop, listop = _opt2logrevset[op]
         if revop and '%' not in revop:
             expr.append(revop)
         elif not listop:
             expr.append(revsetlang.formatspec(revop, val))
         else:
             if revop:
                 val = [revsetlang.formatspec(revop, v) for v in val]
             expr.append(revsetlang.formatspec(listop, val))
 
     if expr:
         expr = '(' + ' and '.join(expr) + ')'
     else:
         expr = None
     return expr
 
 def _initialrevs(repo, opts):
     """Return the initial set of revisions to be filtered or followed"""
     follow = opts.get('follow') or opts.get('follow_first')
     if opts.get('rev'):
         revs = scmutil.revrange(repo, opts['rev'])
     elif follow and repo.dirstate.p1() == nullid:
         revs = smartset.baseset()
     elif follow:
         revs = repo.revs('.')
     else:
         revs = smartset.spanset(repo)
         revs.reverse()
     return revs
 
 def getrevs(repo, pats, opts):
     """Return (revs, differ) where revs is a smartset
 
     differ is a changesetdiffer with pre-configured file matcher.
     """
     follow = opts.get('follow') or opts.get('follow_first')
     followfirst = opts.get('follow_first')
     limit = getlimit(opts)
     revs = _initialrevs(repo, opts)
     if not revs:
         return smartset.baseset(), None
     match, pats, slowpath = _makematcher(repo, revs, pats, opts)
     filematcher = None
     if follow:
         if slowpath or match.always():
             revs = dagop.revancestors(repo, revs, followfirst=followfirst)
         else:
             revs, filematcher = _fileancestors(repo, revs, match, followfirst)
         revs.reverse()
     if filematcher is None:
         filematcher = _makenofollowfilematcher(repo, pats, opts)
     if filematcher is None:
         def filematcher(ctx):
             return match
 
     expr = _makerevset(repo, match, pats, slowpath, opts)
     if opts.get('graph') and opts.get('rev'):
         # User-specified revs might be unsorted, but don't sort before
         # _makerevset because it might depend on the order of revs
         if not (revs.isdescending() or revs.istopo()):
             revs.sort(reverse=True)
     if expr:
         matcher = revset.match(None, expr)
         revs = matcher(repo, revs)
     if limit is not None:
         revs = revs.slice(0, limit)
 
     differ = changesetdiffer()
     differ._makefilematcher = filematcher
     return revs, differ
 
 def _parselinerangeopt(repo, opts):
     """Parse --line-range log option and return a list of tuples (filename,
     (fromline, toline)).
     """
     linerangebyfname = []
     for pat in opts.get('line_range', []):
         try:
             pat, linerange = pat.rsplit(',', 1)
         except ValueError:
             raise error.Abort(_('malformatted line-range pattern %s') % pat)
         try:
             fromline, toline = map(int, linerange.split(':'))
         except ValueError:
             raise error.Abort(_("invalid line range for %s") % pat)
         msg = _("line range pattern '%s' must match exactly one file") % pat
         fname = scmutil.parsefollowlinespattern(repo, None, pat, msg)
         linerangebyfname.append(
             (fname, util.processlinerange(fromline, toline)))
     return linerangebyfname
 
 def getlinerangerevs(repo, userrevs, opts):
     """Return (revs, differ).
 
     "revs" are revisions obtained by processing "line-range" log options and
     walking block ancestors of each specified file/line-range.
 
     "differ" is a changesetdiffer with pre-configured file matcher and hunks
     filter.
     """
     wctx = repo[None]
 
     # Two-levels map of "rev -> file ctx -> [line range]".
     linerangesbyrev = {}
     for fname, (fromline, toline) in _parselinerangeopt(repo, opts):
         if fname not in wctx:
             raise error.Abort(_('cannot follow file not in parent '
                                 'revision: "%s"') % fname)
         fctx = wctx.filectx(fname)
         for fctx, linerange in dagop.blockancestors(fctx, fromline, toline):
             rev = fctx.introrev()
             if rev not in userrevs:
                 continue
             linerangesbyrev.setdefault(
                 rev, {}).setdefault(
                     fctx.path(), []).append(linerange)
 
     def nofilterhunksfn(fctx, hunks):
         return hunks
 
     def hunksfilter(ctx):
         fctxlineranges = linerangesbyrev.get(ctx.rev())
         if fctxlineranges is None:
             return nofilterhunksfn
 
         def filterfn(fctx, hunks):
             lineranges = fctxlineranges.get(fctx.path())
             if lineranges is not None:
                 for hr, lines in hunks:
                     if hr is None: # binary
                         yield hr, lines
                         continue
                     if any(mdiff.hunkinrange(hr[2:], lr)
                            for lr in lineranges):
                         yield hr, lines
             else:
                 for hunk in hunks:
                     yield hunk
 
         return filterfn
 
     def filematcher(ctx):
         files = list(linerangesbyrev.get(ctx.rev(), []))
         return scmutil.matchfiles(repo, files)
 
     revs = sorted(linerangesbyrev, reverse=True)
 
     differ = changesetdiffer()
     differ._makefilematcher = filematcher
     differ._makehunksfilter = hunksfilter
     return revs, differ
 
 def _graphnodeformatter(ui, displayer):
     spec = ui.config('ui', 'graphnodetemplate')
     if not spec:
         return templatekw.getgraphnode  # fast path for "{graphnode}"
 
     spec = templater.unquotestring(spec)
     if isinstance(displayer, changesettemplater):
         # reuse cache of slow templates
         tres = displayer._tresources
     else:
         tres = formatter.templateresources(ui)
     templ = formatter.maketemplater(ui, spec, defaults=templatekw.keywords,
                                     resources=tres)
     def formatnode(repo, ctx):
         props = {'ctx': ctx, 'repo': repo}
         return templ.renderdefault(props)
     return formatnode
 
 def displaygraph(ui, repo, dag, displayer, edgefn, getrenamed=None, props=None):
     props = props or {}
     formatnode = _graphnodeformatter(ui, displayer)
     state = graphmod.asciistate()
     styles = state['styles']
 
     # only set graph styling if HGPLAIN is not set.
     if ui.plain('graph'):
         # set all edge styles to |, the default pre-3.8 behaviour
         styles.update(dict.fromkeys(styles, '|'))
     else:
         edgetypes = {
             'parent': graphmod.PARENT,
             'grandparent': graphmod.GRANDPARENT,
             'missing': graphmod.MISSINGPARENT
         }
         for name, key in edgetypes.items():
             # experimental config: experimental.graphstyle.*
             styles[key] = ui.config('experimental', 'graphstyle.%s' % name,
                                     styles[key])
             if not styles[key]:
                 styles[key] = None
 
         # experimental config: experimental.graphshorten
         state['graphshorten'] = ui.configbool('experimental', 'graphshorten')
 
     for rev, type, ctx, parents in dag:
         char = formatnode(repo, ctx)
         copies = None
         if getrenamed and ctx.rev():
             copies = []
             for fn in ctx.files():
                 rename = getrenamed(fn, ctx.rev())
                 if rename:
                     copies.append((fn, rename))
         edges = edgefn(type, char, state, rev, parents)
         firstedge = next(edges)
         width = firstedge[2]
         displayer.show(ctx, copies=copies,
                        graphwidth=width, **pycompat.strkwargs(props))
         lines = displayer.hunk.pop(rev).split('\n')
         if not lines[-1]:
             del lines[-1]
         displayer.flush(ctx)
         for type, char, width, coldata in itertools.chain([firstedge], edges):
             graphmod.ascii(ui, state, type, char, lines, coldata)
             lines = []
     displayer.close()
 
 def displaygraphrevs(ui, repo, revs, displayer, getrenamed):
     revdag = graphmod.dagwalker(repo, revs)
     displaygraph(ui, repo, revdag, displayer, graphmod.asciiedges, getrenamed)
 
 def displayrevs(ui, repo, revs, displayer, getrenamed):
     for rev in revs:
         ctx = repo[rev]
         copies = None
         if getrenamed is not None and rev:
             copies = []
             for fn in ctx.files():
                 rename = getrenamed(fn, rev)
                 if rename:
                     copies.append((fn, rename))
         displayer.show(ctx, copies=copies)
         displayer.flush(ctx)
     displayer.close()
 
 def checkunsupportedgraphflags(pats, opts):
     for op in ["newest_first"]:
         if op in opts and opts[op]:
             raise error.Abort(_("-G/--graph option is incompatible with --%s")
                              % op.replace("_", "-"))
 
 def graphrevs(repo, nodes, opts):
     limit = getlimit(opts)
     nodes.reverse()
     if limit is not None:
         nodes = nodes[:limit]
     return graphmod.nodes(repo, nodes)

mercurial/patch.py

@@ -1,2857 +1,2861 b''
 # patch.py - patch file parsing routines
 #
 # Copyright 2006 Brendan Cully <[e-mail]>
 # Copyright 2007 Chris Mason <[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, print_function
 
 import collections
 import contextlib
 import copy
 import email
 import errno
 import hashlib
 import os
 import posixpath
 import re
 import shutil
 import zlib
 
 from .i18n import _
 from .node import (
     hex,
     short,
 )
 from . import (
     copies,
     diffhelper,
     diffutil,
     encoding,
     error,
     mail,
     mdiff,
     pathutil,
     pycompat,
     scmutil,
     similar,
     util,
     vfs as vfsmod,
 )
 from .utils import (
     dateutil,
     procutil,
     stringutil,
 )
 
 stringio = util.stringio
 
 gitre = re.compile(br'diff --git a/(.*) b/(.*)')
 tabsplitter = re.compile(br'(\t+|[^\t]+)')
 wordsplitter = re.compile(br'(\t+| +|[a-zA-Z0-9_\x80-\xff]+|'
                           b'[^ \ta-zA-Z0-9_\x80-\xff])')
 
 PatchError = error.PatchError
 
 # public functions
 
 def split(stream):
     '''return an iterator of individual patches from a stream'''
     def isheader(line, inheader):
         if inheader and line.startswith((' ', '\t')):
             # continuation
             return True
         if line.startswith((' ', '-', '+')):
             # diff line - don't check for header pattern in there
             return False
         l = line.split(': ', 1)
         return len(l) == 2 and ' ' not in l[0]
 
     def chunk(lines):
         return stringio(''.join(lines))
 
     def hgsplit(stream, cur):
         inheader = True
 
         for line in stream:
             if not line.strip():
                 inheader = False
             if not inheader and line.startswith('# HG changeset patch'):
                 yield chunk(cur)
                 cur = []
                 inheader = True
 
             cur.append(line)
 
         if cur:
             yield chunk(cur)
 
     def mboxsplit(stream, cur):
         for line in stream:
             if line.startswith('From '):
                 for c in split(chunk(cur[1:])):
                     yield c
                 cur = []
 
             cur.append(line)
 
         if cur:
             for c in split(chunk(cur[1:])):
                 yield c
 
     def mimesplit(stream, cur):
         def msgfp(m):
             fp = stringio()
             g = email.Generator.Generator(fp, mangle_from_=False)
             g.flatten(m)
             fp.seek(0)
             return fp
 
         for line in stream:
             cur.append(line)
         c = chunk(cur)
 
         m = mail.parse(c)
         if not m.is_multipart():
             yield msgfp(m)
         else:
             ok_types = ('text/plain', 'text/x-diff', 'text/x-patch')
             for part in m.walk():
                 ct = part.get_content_type()
                 if ct not in ok_types:
                     continue
                 yield msgfp(part)
 
     def headersplit(stream, cur):
         inheader = False
 
         for line in stream:
             if not inheader and isheader(line, inheader):
                 yield chunk(cur)
                 cur = []
                 inheader = True
             if inheader and not isheader(line, inheader):
                 inheader = False
 
             cur.append(line)
 
         if cur:
             yield chunk(cur)
 
     def remainder(cur):
         yield chunk(cur)
 
     class fiter(object):
         def __init__(self, fp):
             self.fp = fp
 
         def __iter__(self):
             return self
 
         def next(self):
             l = self.fp.readline()
             if not l:
                 raise StopIteration
             return l
 
         __next__ = next
 
     inheader = False
     cur = []
 
     mimeheaders = ['content-type']
 
     if not util.safehasattr(stream, 'next'):
         # http responses, for example, have readline but not next
         stream = fiter(stream)
 
     for line in stream:
         cur.append(line)
         if line.startswith('# HG changeset patch'):
             return hgsplit(stream, cur)
         elif line.startswith('From '):
             return mboxsplit(stream, cur)
         elif isheader(line, inheader):
             inheader = True
             if line.split(':', 1)[0].lower() in mimeheaders:
                 # let email parser handle this
                 return mimesplit(stream, cur)
         elif line.startswith('--- ') and inheader:
             # No evil headers seen by diff start, split by hand
             return headersplit(stream, cur)
         # Not enough info, keep reading
 
     # if we are here, we have a very plain patch
     return remainder(cur)
 
 ## Some facility for extensible patch parsing:
 # list of pairs ("header to match", "data key")
 patchheadermap = [('Date', 'date'),
                   ('Branch', 'branch'),
                   ('Node ID', 'nodeid'),
                  ]
 
 @contextlib.contextmanager
 def extract(ui, fileobj):
     '''extract patch from data read from fileobj.
 
     patch can be a normal patch or contained in an email message.
 
     return a dictionary. Standard keys are:
       - filename,
       - message,
       - user,
       - date,
       - branch,
       - node,
       - p1,
       - p2.
     Any item can be missing from the dictionary. If filename is missing,
     fileobj did not contain a patch. Caller must unlink filename when done.'''
 
     fd, tmpname = pycompat.mkstemp(prefix='hg-patch-')
     tmpfp = os.fdopen(fd, r'wb')
     try:
         yield _extract(ui, fileobj, tmpname, tmpfp)
     finally:
         tmpfp.close()
         os.unlink(tmpname)
 
 def _extract(ui, fileobj, tmpname, tmpfp):
 
     # attempt to detect the start of a patch
     # (this heuristic is borrowed from quilt)
     diffre = re.compile(br'^(?:Index:[ \t]|diff[ \t]-|RCS file: |'
                         br'retrieving revision [0-9]+(\.[0-9]+)*$|'
                         br'---[ \t].*?^\+\+\+[ \t]|'
                         br'\*\*\*[ \t].*?^---[ \t])',
                         re.MULTILINE | re.DOTALL)
 
     data = {}
 
     msg = mail.parse(fileobj)
 
     subject = msg[r'Subject'] and mail.headdecode(msg[r'Subject'])
     data['user'] = msg[r'From'] and mail.headdecode(msg[r'From'])
     if not subject and not data['user']:
         # Not an email, restore parsed headers if any
         subject = '\n'.join(': '.join(map(encoding.strtolocal, h))
                             for h in msg.items()) + '\n'
 
     # should try to parse msg['Date']
     parents = []
 
     if subject:
         if subject.startswith('[PATCH'):
             pend = subject.find(']')
             if pend >= 0:
                 subject = subject[pend + 1:].lstrip()
         subject = re.sub(br'\n[ \t]+', ' ', subject)
         ui.debug('Subject: %s\n' % subject)
     if data['user']:
         ui.debug('From: %s\n' % data['user'])
     diffs_seen = 0
     ok_types = ('text/plain', 'text/x-diff', 'text/x-patch')
     message = ''
     for part in msg.walk():
         content_type = pycompat.bytestr(part.get_content_type())
         ui.debug('Content-Type: %s\n' % content_type)
         if content_type not in ok_types:
             continue
         payload = part.get_payload(decode=True)
         m = diffre.search(payload)
         if m:
             hgpatch = False
             hgpatchheader = False
             ignoretext = False
 
             ui.debug('found patch at byte %d\n' % m.start(0))
             diffs_seen += 1
             cfp = stringio()
             for line in payload[:m.start(0)].splitlines():
                 if line.startswith('# HG changeset patch') and not hgpatch:
                     ui.debug('patch generated by hg export\n')
                     hgpatch = True
                     hgpatchheader = True
                     # drop earlier commit message content
                     cfp.seek(0)
                     cfp.truncate()
                     subject = None
                 elif hgpatchheader:
                     if line.startswith('# User '):
                         data['user'] = line[7:]
                         ui.debug('From: %s\n' % data['user'])
                     elif line.startswith("# Parent "):
                         parents.append(line[9:].lstrip())
                     elif line.startswith("# "):
                         for header, key in patchheadermap:
                             prefix = '# %s ' % header
                             if line.startswith(prefix):
                                 data[key] = line[len(prefix):]
                     else:
                         hgpatchheader = False
                 elif line == '---':
                     ignoretext = True
                 if not hgpatchheader and not ignoretext:
                     cfp.write(line)
                     cfp.write('\n')
             message = cfp.getvalue()
             if tmpfp:
                 tmpfp.write(payload)
                 if not payload.endswith('\n'):
                     tmpfp.write('\n')
         elif not diffs_seen and message and content_type == 'text/plain':
             message += '\n' + payload
 
     if subject and not message.startswith(subject):
         message = '%s\n%s' % (subject, message)
     data['message'] = message
     tmpfp.close()
     if parents:
         data['p1'] = parents.pop(0)
         if parents:
             data['p2'] = parents.pop(0)
 
     if diffs_seen:
         data['filename'] = tmpname
 
     return data
 
 class patchmeta(object):
     """Patched file metadata
 
     'op' is the performed operation within ADD, DELETE, RENAME, MODIFY
     or COPY.  'path' is patched file path. 'oldpath' is set to the
     origin file when 'op' is either COPY or RENAME, None otherwise. If
     file mode is changed, 'mode' is a tuple (islink, isexec) where
     'islink' is True if the file is a symlink and 'isexec' is True if
     the file is executable. Otherwise, 'mode' is None.
     """
     def __init__(self, path):
         self.path = path
         self.oldpath = None
         self.mode = None
         self.op = 'MODIFY'
         self.binary = False
 
     def setmode(self, mode):
         islink = mode & 0o20000
         isexec = mode & 0o100
         self.mode = (islink, isexec)
 
     def copy(self):
         other = patchmeta(self.path)
         other.oldpath = self.oldpath
         other.mode = self.mode
         other.op = self.op
         other.binary = self.binary
         return other
 
     def _ispatchinga(self, afile):
         if afile == '/dev/null':
             return self.op == 'ADD'
         return afile == 'a/' + (self.oldpath or self.path)
 
     def _ispatchingb(self, bfile):
         if bfile == '/dev/null':
             return self.op == 'DELETE'
         return bfile == 'b/' + self.path
 
     def ispatching(self, afile, bfile):
         return self._ispatchinga(afile) and self._ispatchingb(bfile)
 
     def __repr__(self):
         return "<patchmeta %s %r>" % (self.op, self.path)
 
 def readgitpatch(lr):
     """extract git-style metadata about patches from <patchname>"""
 
     # Filter patch for git information
     gp = None
     gitpatches = []
     for line in lr:
         line = line.rstrip(' \r\n')
         if line.startswith('diff --git a/'):
             m = gitre.match(line)
             if m:
                 if gp:
                     gitpatches.append(gp)
                 dst = m.group(2)
                 gp = patchmeta(dst)
         elif gp:
             if line.startswith('--- '):
                 gitpatches.append(gp)
                 gp = None
                 continue
             if line.startswith('rename from '):
                 gp.op = 'RENAME'
                 gp.oldpath = line[12:]
             elif line.startswith('rename to '):
                 gp.path = line[10:]
             elif line.startswith('copy from '):
                 gp.op = 'COPY'
                 gp.oldpath = line[10:]
             elif line.startswith('copy to '):
                 gp.path = line[8:]
             elif line.startswith('deleted file'):
                 gp.op = 'DELETE'
             elif line.startswith('new file mode '):
                 gp.op = 'ADD'
                 gp.setmode(int(line[-6:], 8))
             elif line.startswith('new mode '):
                 gp.setmode(int(line[-6:], 8))
             elif line.startswith('GIT binary patch'):
                 gp.binary = True
     if gp:
         gitpatches.append(gp)
 
     return gitpatches
 
 class linereader(object):
     # simple class to allow pushing lines back into the input stream
     def __init__(self, fp):
         self.fp = fp
         self.buf = []
 
     def push(self, line):
         if line is not None:
             self.buf.append(line)
 
     def readline(self):
         if self.buf:
             l = self.buf[0]
             del self.buf[0]
             return l
         return self.fp.readline()
 
     def __iter__(self):
         return iter(self.readline, '')
 
 class abstractbackend(object):
     def __init__(self, ui):
         self.ui = ui
 
     def getfile(self, fname):
         """Return target file data and flags as a (data, (islink,
         isexec)) tuple. Data is None if file is missing/deleted.
         """
         raise NotImplementedError
 
     def setfile(self, fname, data, mode, copysource):
         """Write data to target file fname and set its mode. mode is a
         (islink, isexec) tuple. If data is None, the file content should
         be left unchanged. If the file is modified after being copied,
         copysource is set to the original file name.
         """
         raise NotImplementedError
 
     def unlink(self, fname):
         """Unlink target file."""
         raise NotImplementedError
 
     def writerej(self, fname, failed, total, lines):
         """Write rejected lines for fname. total is the number of hunks
         which failed to apply and total the total number of hunks for this
         files.
         """
 
     def exists(self, fname):
         raise NotImplementedError
 
     def close(self):
         raise NotImplementedError
 
 class fsbackend(abstractbackend):
     def __init__(self, ui, basedir):
         super(fsbackend, self).__init__(ui)
         self.opener = vfsmod.vfs(basedir)
 
     def getfile(self, fname):
         if self.opener.islink(fname):
             return (self.opener.readlink(fname), (True, False))
 
         isexec = False
         try:
             isexec = self.opener.lstat(fname).st_mode & 0o100 != 0
         except OSError as e:
             if e.errno != errno.ENOENT:
                 raise
         try:
             return (self.opener.read(fname), (False, isexec))
         except IOError as e:
             if e.errno != errno.ENOENT:
                 raise
             return None, None
 
     def setfile(self, fname, data, mode, copysource):
         islink, isexec = mode
         if data is None:
             self.opener.setflags(fname, islink, isexec)
             return
         if islink:
             self.opener.symlink(data, fname)
         else:
             self.opener.write(fname, data)
             if isexec:
                 self.opener.setflags(fname, False, True)
 
     def unlink(self, fname):
         rmdir = self.ui.configbool('experimental', 'removeemptydirs')
         self.opener.unlinkpath(fname, ignoremissing=True, rmdir=rmdir)
 
     def writerej(self, fname, failed, total, lines):
         fname = fname + ".rej"
         self.ui.warn(
             _("%d out of %d hunks FAILED -- saving rejects to file %s\n") %
             (failed, total, fname))
         fp = self.opener(fname, 'w')
         fp.writelines(lines)
         fp.close()
 
     def exists(self, fname):
         return self.opener.lexists(fname)
 
 class workingbackend(fsbackend):
     def __init__(self, ui, repo, similarity):
         super(workingbackend, self).__init__(ui, repo.root)
         self.repo = repo
         self.similarity = similarity
         self.removed = set()
         self.changed = set()
         self.copied = []
 
     def _checkknown(self, fname):
         if self.repo.dirstate[fname] == '?' and self.exists(fname):
             raise PatchError(_('cannot patch %s: file is not tracked') % fname)
 
     def setfile(self, fname, data, mode, copysource):
         self._checkknown(fname)
         super(workingbackend, self).setfile(fname, data, mode, copysource)
         if copysource is not None:
             self.copied.append((copysource, fname))
         self.changed.add(fname)
 
     def unlink(self, fname):
         self._checkknown(fname)
         super(workingbackend, self).unlink(fname)
         self.removed.add(fname)
         self.changed.add(fname)
 
     def close(self):
         wctx = self.repo[None]
         changed = set(self.changed)
         for src, dst in self.copied:
             scmutil.dirstatecopy(self.ui, self.repo, wctx, src, dst)
         if self.removed:
             wctx.forget(sorted(self.removed))
             for f in self.removed:
                 if f not in self.repo.dirstate:
                     # File was deleted and no longer belongs to the
                     # dirstate, it was probably marked added then
                     # deleted, and should not be considered by
                     # marktouched().
                     changed.discard(f)
         if changed:
             scmutil.marktouched(self.repo, changed, self.similarity)
         return sorted(self.changed)
 
 class filestore(object):
     def __init__(self, maxsize=None):
         self.opener = None
         self.files = {}
         self.created = 0
         self.maxsize = maxsize
         if self.maxsize is None:
             self.maxsize = 4*(2**20)
         self.size = 0
         self.data = {}
 
     def setfile(self, fname, data, mode, copied=None):
         if self.maxsize < 0 or (len(data) + self.size) <= self.maxsize:
             self.data[fname] = (data, mode, copied)
             self.size += len(data)
         else:
             if self.opener is None:
                 root = pycompat.mkdtemp(prefix='hg-patch-')
                 self.opener = vfsmod.vfs(root)
             # Avoid filename issues with these simple names
             fn = '%d' % self.created
             self.opener.write(fn, data)
             self.created += 1
             self.files[fname] = (fn, mode, copied)
 
     def getfile(self, fname):
         if fname in self.data:
             return self.data[fname]
         if not self.opener or fname not in self.files:
             return None, None, None
         fn, mode, copied = self.files[fname]
         return self.opener.read(fn), mode, copied
 
     def close(self):
         if self.opener:
             shutil.rmtree(self.opener.base)
 
 class repobackend(abstractbackend):
     def __init__(self, ui, repo, ctx, store):
         super(repobackend, self).__init__(ui)
         self.repo = repo
         self.ctx = ctx
         self.store = store
         self.changed = set()
         self.removed = set()
         self.copied = {}
 
     def _checkknown(self, fname):
         if fname not in self.ctx:
             raise PatchError(_('cannot patch %s: file is not tracked') % fname)
 
     def getfile(self, fname):
         try:
             fctx = self.ctx[fname]
         except error.LookupError:
             return None, None
         flags = fctx.flags()
         return fctx.data(), ('l' in flags, 'x' in flags)
 
     def setfile(self, fname, data, mode, copysource):
         if copysource:
             self._checkknown(copysource)
         if data is None:
             data = self.ctx[fname].data()
         self.store.setfile(fname, data, mode, copysource)
         self.changed.add(fname)
         if copysource:
             self.copied[fname] = copysource
 
     def unlink(self, fname):
         self._checkknown(fname)
         self.removed.add(fname)
 
     def exists(self, fname):
         return fname in self.ctx
 
     def close(self):
         return self.changed | self.removed
 
 # @@ -start,len +start,len @@ or @@ -start +start @@ if len is 1
 unidesc = re.compile(br'@@ -(\d+)(?:,(\d+))? \+(\d+)(?:,(\d+))? @@')
 contextdesc = re.compile(br'(?:---|\*\*\*) (\d+)(?:,(\d+))? (?:---|\*\*\*)')
 eolmodes = ['strict', 'crlf', 'lf', 'auto']
 
 class patchfile(object):
     def __init__(self, ui, gp, backend, store, eolmode='strict'):
         self.fname = gp.path
         self.eolmode = eolmode
         self.eol = None
         self.backend = backend
         self.ui = ui
         self.lines = []
         self.exists = False
         self.missing = True
         self.mode = gp.mode
         self.copysource = gp.oldpath
         self.create = gp.op in ('ADD', 'COPY', 'RENAME')
         self.remove = gp.op == 'DELETE'
         if self.copysource is None:
             data, mode = backend.getfile(self.fname)
         else:
             data, mode = store.getfile(self.copysource)[:2]
         if data is not None:
             self.exists = self.copysource is None or backend.exists(self.fname)
             self.missing = False
             if data:
                 self.lines = mdiff.splitnewlines(data)
             if self.mode is None:
                 self.mode = mode
             if self.lines:
                 # Normalize line endings
                 if self.lines[0].endswith('\r\n'):
                     self.eol = '\r\n'
                 elif self.lines[0].endswith('\n'):
                     self.eol = '\n'
                 if eolmode != 'strict':
                     nlines = []
                     for l in self.lines:
                         if l.endswith('\r\n'):
                             l = l[:-2] + '\n'
                         nlines.append(l)
                     self.lines = nlines
         else:
             if self.create:
                 self.missing = False
             if self.mode is None:
                 self.mode = (False, False)
         if self.missing:
             self.ui.warn(_("unable to find '%s' for patching\n") % self.fname)
             self.ui.warn(_("(use '--prefix' to apply patch relative to the "
                            "current directory)\n"))
 
         self.hash = {}
         self.dirty = 0
         self.offset = 0
         self.skew = 0
         self.rej = []
         self.fileprinted = False
         self.printfile(False)
         self.hunks = 0
 
     def writelines(self, fname, lines, mode):
         if self.eolmode == 'auto':
             eol = self.eol
         elif self.eolmode == 'crlf':
             eol = '\r\n'
         else:
             eol = '\n'
 
         if self.eolmode != 'strict' and eol and eol != '\n':
             rawlines = []
             for l in lines:
                 if l and l.endswith('\n'):
                     l = l[:-1] + eol
                 rawlines.append(l)
             lines = rawlines
 
         self.backend.setfile(fname, ''.join(lines), mode, self.copysource)
 
     def printfile(self, warn):
         if self.fileprinted:
             return
         if warn or self.ui.verbose:
             self.fileprinted = True
         s = _("patching file %s\n") % self.fname
         if warn:
             self.ui.warn(s)
         else:
             self.ui.note(s)
 
 
     def findlines(self, l, linenum):
         # looks through the hash and finds candidate lines.  The
         # result is a list of line numbers sorted based on distance
         # from linenum
 
         cand = self.hash.get(l, [])
         if len(cand) > 1:
             # resort our list of potentials forward then back.
             cand.sort(key=lambda x: abs(x - linenum))
         return cand
 
     def write_rej(self):
         # our rejects are a little different from patch(1).  This always
         # creates rejects in the same form as the original patch.  A file
         # header is inserted so that you can run the reject through patch again
         # without having to type the filename.
         if not self.rej:
             return
         base = os.path.basename(self.fname)
         lines = ["--- %s\n+++ %s\n" % (base, base)]
         for x in self.rej:
             for l in x.hunk:
                 lines.append(l)
                 if l[-1:] != '\n':
                     lines.append("\n\\ No newline at end of file\n")
         self.backend.writerej(self.fname, len(self.rej), self.hunks, lines)
 
     def apply(self, h):
         if not h.complete():
             raise PatchError(_("bad hunk #%d %s (%d %d %d %d)") %
                             (h.number, h.desc, len(h.a), h.lena, len(h.b),
                             h.lenb))
 
         self.hunks += 1
 
         if self.missing:
             self.rej.append(h)
             return -1
 
         if self.exists and self.create:
             if self.copysource:
                 self.ui.warn(_("cannot create %s: destination already "
                                "exists\n") % self.fname)
             else:
                 self.ui.warn(_("file %s already exists\n") % self.fname)
             self.rej.append(h)
             return -1
 
         if isinstance(h, binhunk):
             if self.remove:
                 self.backend.unlink(self.fname)
             else:
                 l = h.new(self.lines)
                 self.lines[:] = l
                 self.offset += len(l)
                 self.dirty = True
             return 0
 
         horig = h
         if (self.eolmode in ('crlf', 'lf')
             or self.eolmode == 'auto' and self.eol):
             # If new eols are going to be normalized, then normalize
             # hunk data before patching. Otherwise, preserve input
             # line-endings.
             h = h.getnormalized()
 
         # fast case first, no offsets, no fuzz
         old, oldstart, new, newstart = h.fuzzit(0, False)
         oldstart += self.offset
         orig_start = oldstart
         # if there's skew we want to emit the "(offset %d lines)" even
         # when the hunk cleanly applies at start + skew, so skip the
         # fast case code
         if self.skew == 0 and diffhelper.testhunk(old, self.lines, oldstart):
             if self.remove:
                 self.backend.unlink(self.fname)
             else:
                 self.lines[oldstart:oldstart + len(old)] = new
                 self.offset += len(new) - len(old)
                 self.dirty = True
             return 0
 
         # ok, we couldn't match the hunk. Lets look for offsets and fuzz it
         self.hash = {}
         for x, s in enumerate(self.lines):
             self.hash.setdefault(s, []).append(x)
 
         for fuzzlen in pycompat.xrange(self.ui.configint("patch", "fuzz") + 1):
             for toponly in [True, False]:
                 old, oldstart, new, newstart = h.fuzzit(fuzzlen, toponly)
                 oldstart = oldstart + self.offset + self.skew
                 oldstart = min(oldstart, len(self.lines))
                 if old:
                     cand = self.findlines(old[0][1:], oldstart)
                 else:
                     # Only adding lines with no or fuzzed context, just
                     # take the skew in account
                     cand = [oldstart]
 
                 for l in cand:
                     if not old or diffhelper.testhunk(old, self.lines, l):
                         self.lines[l : l + len(old)] = new
                         self.offset += len(new) - len(old)
                         self.skew = l - orig_start
                         self.dirty = True
                         offset = l - orig_start - fuzzlen
                         if fuzzlen:
                             msg = _("Hunk #%d succeeded at %d "
                                     "with fuzz %d "
                                     "(offset %d lines).\n")
                             self.printfile(True)
                             self.ui.warn(msg %
                                 (h.number, l + 1, fuzzlen, offset))
                         else:
                             msg = _("Hunk #%d succeeded at %d "
                                     "(offset %d lines).\n")
                             self.ui.note(msg % (h.number, l + 1, offset))
                         return fuzzlen
         self.printfile(True)
         self.ui.warn(_("Hunk #%d FAILED at %d\n") % (h.number, orig_start))
         self.rej.append(horig)
         return -1
 
     def close(self):
         if self.dirty:
             self.writelines(self.fname, self.lines, self.mode)
         self.write_rej()
         return len(self.rej)
 
 class header(object):
     """patch header
     """
     diffgit_re = re.compile('diff --git a/(.*) b/(.*)$')
     diff_re = re.compile('diff -r .* (.*)$')
     allhunks_re = re.compile('(?:index|deleted file) ')
     pretty_re = re.compile('(?:new file|deleted file) ')
     special_re = re.compile('(?:index|deleted|copy|rename) ')
     newfile_re = re.compile('(?:new file)')
 
     def __init__(self, header):
         self.header = header
         self.hunks = []
 
     def binary(self):
         return any(h.startswith('index ') for h in self.header)
 
     def pretty(self, fp):
         for h in self.header:
             if h.startswith('index '):
                 fp.write(_('this modifies a binary file (all or nothing)\n'))
                 break
             if self.pretty_re.match(h):
                 fp.write(h)
                 if self.binary():
                     fp.write(_('this is a binary file\n'))
                 break
             if h.startswith('---'):
                 fp.write(_('%d hunks, %d lines changed\n') %
                          (len(self.hunks),
                           sum([max(h.added, h.removed) for h in self.hunks])))
                 break
             fp.write(h)
 
     def write(self, fp):
         fp.write(''.join(self.header))
 
     def allhunks(self):
         return any(self.allhunks_re.match(h) for h in self.header)
 
     def files(self):
         match = self.diffgit_re.match(self.header[0])
         if match:
             fromfile, tofile = match.groups()
             if fromfile == tofile:
                 return [fromfile]
             return [fromfile, tofile]
         else:
             return self.diff_re.match(self.header[0]).groups()
 
     def filename(self):
         return self.files()[-1]
 
     def __repr__(self):
         return '<header %s>' % (' '.join(map(repr, self.files())))
 
     def isnewfile(self):
         return any(self.newfile_re.match(h) for h in self.header)
 
     def special(self):
         # Special files are shown only at the header level and not at the hunk
         # level for example a file that has been deleted is a special file.
         # The user cannot change the content of the operation, in the case of
         # the deleted file he has to take the deletion or not take it, he
         # cannot take some of it.
         # Newly added files are special if they are empty, they are not special
         # if they have some content as we want to be able to change it
         nocontent = len(self.header) == 2
         emptynewfile = self.isnewfile() and nocontent
         return emptynewfile or \
                 any(self.special_re.match(h) for h in self.header)
 
 class recordhunk(object):
     """patch hunk
 
     XXX shouldn't we merge this with the other hunk class?
     """
 
     def __init__(self, header, fromline, toline, proc, before, hunk, after,
                  maxcontext=None):
         def trimcontext(lines, reverse=False):
             if maxcontext is not None:
                 delta = len(lines) - maxcontext
                 if delta > 0:
                     if reverse:
                         return delta, lines[delta:]
                     else:
                         return delta, lines[:maxcontext]
             return 0, lines
 
         self.header = header
         trimedbefore, self.before = trimcontext(before, True)
         self.fromline = fromline + trimedbefore
         self.toline = toline + trimedbefore
         _trimedafter, self.after = trimcontext(after, False)
         self.proc = proc
         self.hunk = hunk
         self.added, self.removed = self.countchanges(self.hunk)
 
     def __eq__(self, v):
         if not isinstance(v, recordhunk):
             return False
 
         return ((v.hunk == self.hunk) and
                 (v.proc == self.proc) and
                 (self.fromline == v.fromline) and
                 (self.header.files() == v.header.files()))
 
     def __hash__(self):
         return hash((tuple(self.hunk),
             tuple(self.header.files()),
             self.fromline,
             self.proc))
 
     def countchanges(self, hunk):
         """hunk -> (n+,n-)"""
         add = len([h for h in hunk if h.startswith('+')])
         rem = len([h for h in hunk if h.startswith('-')])
         return add, rem
 
     def reversehunk(self):
         """return another recordhunk which is the reverse of the hunk
 
         If this hunk is diff(A, B), the returned hunk is diff(B, A). To do
         that, swap fromline/toline and +/- signs while keep other things
         unchanged.
         """
         m = {'+': '-', '-': '+', '\\': '\\'}
         hunk = ['%s%s' % (m[l[0:1]], l[1:]) for l in self.hunk]
         return recordhunk(self.header, self.toline, self.fromline, self.proc,
                           self.before, hunk, self.after)
 
     def write(self, fp):
         delta = len(self.before) + len(self.after)
         if self.after and self.after[-1] == '\\ No newline at end of file\n':
             delta -= 1
         fromlen = delta + self.removed
         tolen = delta + self.added
         fp.write('@@ -%d,%d +%d,%d @@%s\n' %
                  (self.fromline, fromlen, self.toline, tolen,
                   self.proc and (' ' + self.proc)))
         fp.write(''.join(self.before + self.hunk + self.after))
 
     pretty = write
 
     def filename(self):
         return self.header.filename()
 
     def __repr__(self):
         return '<hunk %r@%d>' % (self.filename(), self.fromline)
 
 def getmessages():
     return {
         'multiple': {
             'apply': _("apply change %d/%d to '%s'?"),
             'discard': _("discard change %d/%d to '%s'?"),
             'record': _("record change %d/%d to '%s'?"),
         },
         'single': {
             'apply': _("apply this change to '%s'?"),
             'discard': _("discard this change to '%s'?"),
             'record': _("record this change to '%s'?"),
         },
         'help': {
             'apply': _('[Ynesfdaq?]'
                          '$$ &Yes, apply this change'
                          '$$ &No, skip this change'
                          '$$ &Edit this change manually'
                          '$$ &Skip remaining changes to this file'
                          '$$ Apply remaining changes to this &file'
                          '$$ &Done, skip remaining changes and files'
                          '$$ Apply &all changes to all remaining files'
                          '$$ &Quit, applying no changes'
                          '$$ &? (display help)'),
             'discard': _('[Ynesfdaq?]'
                          '$$ &Yes, discard this change'
                          '$$ &No, skip this change'
                          '$$ &Edit this change manually'
                          '$$ &Skip remaining changes to this file'
                          '$$ Discard remaining changes to this &file'
                          '$$ &Done, skip remaining changes and files'
                          '$$ Discard &all changes to all remaining files'
                          '$$ &Quit, discarding no changes'
                          '$$ &? (display help)'),
             'record': _('[Ynesfdaq?]'
                         '$$ &Yes, record this change'
                         '$$ &No, skip this change'
                         '$$ &Edit this change manually'
                         '$$ &Skip remaining changes to this file'
                         '$$ Record remaining changes to this &file'
                         '$$ &Done, skip remaining changes and files'
                         '$$ Record &all changes to all remaining files'
                         '$$ &Quit, recording no changes'
                         '$$ &? (display help)'),
         }
     }
 
 def filterpatch(ui, headers, operation=None):
     """Interactively filter patch chunks into applied-only chunks"""
     messages = getmessages()
 
     if operation is None:
         operation = 'record'
 
     def prompt(skipfile, skipall, query, chunk):
         """prompt query, and process base inputs
 
         - y/n for the rest of file
         - y/n for the rest
         - ? (help)
         - q (quit)
 
         Return True/False and possibly updated skipfile and skipall.
         """
         newpatches = None
         if skipall is not None:
             return skipall, skipfile, skipall, newpatches
         if skipfile is not None:
             return skipfile, skipfile, skipall, newpatches
         while True:
             resps = messages['help'][operation]
             r = ui.promptchoice("%s %s" % (query, resps))
             ui.write("\n")
             if r == 8: # ?
                 for c, t in ui.extractchoices(resps)[1]:
                     ui.write('%s - %s\n' % (c, encoding.lower(t)))
                 continue
             elif r == 0: # yes
                 ret = True
             elif r == 1: # no
                 ret = False
             elif r == 2: # Edit patch
                 if chunk is None:
                     ui.write(_('cannot edit patch for whole file'))
                     ui.write("\n")
                     continue
                 if chunk.header.binary():
                     ui.write(_('cannot edit patch for binary file'))
                     ui.write("\n")
                     continue
                 # Patch comment based on the Git one (based on comment at end of
                 # https://mercurial-scm.org/wiki/RecordExtension)
                 phelp = '---' + _("""
 To remove '-' lines, make them ' ' lines (context).
 To remove '+' lines, delete them.
 Lines starting with # will be removed from the patch.
 
 If the patch applies cleanly, the edited hunk will immediately be
 added to the record list. If it does not apply cleanly, a rejects
 file will be generated: you can use that when you try again. If
 all lines of the hunk are removed, then the edit is aborted and
 the hunk is left unchanged.
 """)
                 (patchfd, patchfn) = pycompat.mkstemp(prefix="hg-editor-",
                                                       suffix=".diff")
                 ncpatchfp = None
                 try:
                     # Write the initial patch
                     f = util.nativeeolwriter(os.fdopen(patchfd, r'wb'))
                     chunk.header.write(f)
                     chunk.write(f)
                     f.write('\n'.join(['# ' + i for i in phelp.splitlines()]))
                     f.close()
                     # Start the editor and wait for it to complete
                     editor = ui.geteditor()
                     ret = ui.system("%s \"%s\"" % (editor, patchfn),
                                     environ={'HGUSER': ui.username()},
                                     blockedtag='filterpatch')
                     if ret != 0:
                         ui.warn(_("editor exited with exit code %d\n") % ret)
                         continue
                     # Remove comment lines
                     patchfp = open(patchfn, r'rb')
                     ncpatchfp = stringio()
                     for line in util.iterfile(patchfp):
                         line = util.fromnativeeol(line)
                         if not line.startswith('#'):
                             ncpatchfp.write(line)
                     patchfp.close()
                     ncpatchfp.seek(0)
                     newpatches = parsepatch(ncpatchfp)
                 finally:
                     os.unlink(patchfn)
                     del ncpatchfp
                 # Signal that the chunk shouldn't be applied as-is, but
                 # provide the new patch to be used instead.
                 ret = False
             elif r == 3: # Skip
                 ret = skipfile = False
             elif r == 4: # file (Record remaining)
                 ret = skipfile = True
             elif r == 5: # done, skip remaining
                 ret = skipall = False
             elif r == 6: # all
                 ret = skipall = True
             elif r == 7: # quit
                 raise error.Abort(_('user quit'))
             return ret, skipfile, skipall, newpatches
 
     seen = set()
     applied = {}        # 'filename' -> [] of chunks
     skipfile, skipall = None, None
     pos, total = 1, sum(len(h.hunks) for h in headers)
     for h in headers:
         pos += len(h.hunks)
         skipfile = None
         fixoffset = 0
         hdr = ''.join(h.header)
         if hdr in seen:
             continue
         seen.add(hdr)
         if skipall is None:
             h.pretty(ui)
         msg = (_('examine changes to %s?') %
                _(' and ').join("'%s'" % f for f in h.files()))
         r, skipfile, skipall, np = prompt(skipfile, skipall, msg, None)
         if not r:
             continue
         applied[h.filename()] = [h]
         if h.allhunks():
             applied[h.filename()] += h.hunks
             continue
         for i, chunk in enumerate(h.hunks):
             if skipfile is None and skipall is None:
                 chunk.pretty(ui)
             if total == 1:
                 msg = messages['single'][operation] % chunk.filename()
             else:
                 idx = pos - len(h.hunks) + i
                 msg = messages['multiple'][operation] % (idx, total,
                                                          chunk.filename())
             r, skipfile, skipall, newpatches = prompt(skipfile,
                     skipall, msg, chunk)
             if r:
                 if fixoffset:
                     chunk = copy.copy(chunk)
                     chunk.toline += fixoffset
                 applied[chunk.filename()].append(chunk)
             elif newpatches is not None:
                 for newpatch in newpatches:
                     for newhunk in newpatch.hunks:
                         if fixoffset:
                             newhunk.toline += fixoffset
                         applied[newhunk.filename()].append(newhunk)
             else:
                 fixoffset += chunk.removed - chunk.added
     return (sum([h for h in applied.itervalues()
                if h[0].special() or len(h) > 1], []), {})
 class hunk(object):
     def __init__(self, desc, num, lr, context):
         self.number = num
         self.desc = desc
         self.hunk = [desc]
         self.a = []
         self.b = []
         self.starta = self.lena = None
         self.startb = self.lenb = None
         if lr is not None:
             if context:
                 self.read_context_hunk(lr)
             else:
                 self.read_unified_hunk(lr)
 
     def getnormalized(self):
         """Return a copy with line endings normalized to LF."""
 
         def normalize(lines):
             nlines = []
             for line in lines:
                 if line.endswith('\r\n'):
                     line = line[:-2] + '\n'
                 nlines.append(line)
             return nlines
 
         # Dummy object, it is rebuilt manually
         nh = hunk(self.desc, self.number, None, None)
         nh.number = self.number
         nh.desc = self.desc
         nh.hunk = self.hunk
         nh.a = normalize(self.a)
         nh.b = normalize(self.b)
         nh.starta = self.starta
         nh.startb = self.startb
         nh.lena = self.lena
         nh.lenb = self.lenb
         return nh
 
     def read_unified_hunk(self, lr):
         m = unidesc.match(self.desc)
         if not m:
             raise PatchError(_("bad hunk #%d") % self.number)
         self.starta, self.lena, self.startb, self.lenb = m.groups()
         if self.lena is None:
             self.lena = 1
         else:
             self.lena = int(self.lena)
         if self.lenb is None:
             self.lenb = 1
         else:
             self.lenb = int(self.lenb)
         self.starta = int(self.starta)
         self.startb = int(self.startb)
         try:
             diffhelper.addlines(lr, self.hunk, self.lena, self.lenb,
                                 self.a, self.b)
         except error.ParseError as e:
             raise PatchError(_("bad hunk #%d: %s") % (self.number, e))
         # if we hit eof before finishing out the hunk, the last line will
         # be zero length.  Lets try to fix it up.
         while len(self.hunk[-1]) == 0:
             del self.hunk[-1]
             del self.a[-1]
             del self.b[-1]
             self.lena -= 1
             self.lenb -= 1
         self._fixnewline(lr)
 
     def read_context_hunk(self, lr):
         self.desc = lr.readline()
         m = contextdesc.match(self.desc)
         if not m:
             raise PatchError(_("bad hunk #%d") % self.number)
         self.starta, aend = m.groups()
         self.starta = int(self.starta)
         if aend is None:
             aend = self.starta
         self.lena = int(aend) - self.starta
         if self.starta:
             self.lena += 1
         for x in pycompat.xrange(self.lena):
             l = lr.readline()
             if l.startswith('---'):
                 # lines addition, old block is empty
                 lr.push(l)
                 break
             s = l[2:]
             if l.startswith('- ') or l.startswith('! '):
                 u = '-' + s
             elif l.startswith('  '):
                 u = ' ' + s
             else:
                 raise PatchError(_("bad hunk #%d old text line %d") %
                                  (self.number, x))
             self.a.append(u)
             self.hunk.append(u)
 
         l = lr.readline()
         if l.startswith(br'\ '):
             s = self.a[-1][:-1]
             self.a[-1] = s
             self.hunk[-1] = s
             l = lr.readline()
         m = contextdesc.match(l)
         if not m:
             raise PatchError(_("bad hunk #%d") % self.number)
         self.startb, bend = m.groups()
         self.startb = int(self.startb)
         if bend is None:
             bend = self.startb
         self.lenb = int(bend) - self.startb
         if self.startb:
             self.lenb += 1
         hunki = 1
         for x in pycompat.xrange(self.lenb):
             l = lr.readline()
             if l.startswith(br'\ '):
                 # XXX: the only way to hit this is with an invalid line range.
                 # The no-eol marker is not counted in the line range, but I
                 # guess there are diff(1) out there which behave differently.
                 s = self.b[-1][:-1]
                 self.b[-1] = s
                 self.hunk[hunki - 1] = s
                 continue
             if not l:
                 # line deletions, new block is empty and we hit EOF
                 lr.push(l)
                 break
             s = l[2:]
             if l.startswith('+ ') or l.startswith('! '):
                 u = '+' + s
             elif l.startswith('  '):
                 u = ' ' + s
             elif len(self.b) == 0:
                 # line deletions, new block is empty
                 lr.push(l)
                 break
             else:
                 raise PatchError(_("bad hunk #%d old text line %d") %
                                  (self.number, x))
             self.b.append(s)
             while True:
                 if hunki >= len(self.hunk):
                     h = ""
                 else:
                     h = self.hunk[hunki]
                 hunki += 1
                 if h == u:
                     break
                 elif h.startswith('-'):
                     continue
                 else:
                     self.hunk.insert(hunki - 1, u)
                     break
 
         if not self.a:
             # this happens when lines were only added to the hunk
             for x in self.hunk:
                 if x.startswith('-') or x.startswith(' '):
                     self.a.append(x)
         if not self.b:
             # this happens when lines were only deleted from the hunk
             for x in self.hunk:
                 if x.startswith('+') or x.startswith(' '):
                     self.b.append(x[1:])
         # @@ -start,len +start,len @@
         self.desc = "@@ -%d,%d +%d,%d @@\n" % (self.starta, self.lena,
                                              self.startb, self.lenb)
         self.hunk[0] = self.desc
         self._fixnewline(lr)
 
     def _fixnewline(self, lr):
         l = lr.readline()
         if l.startswith(br'\ '):
             diffhelper.fixnewline(self.hunk, self.a, self.b)
         else:
             lr.push(l)
 
     def complete(self):
         return len(self.a) == self.lena and len(self.b) == self.lenb
 
     def _fuzzit(self, old, new, fuzz, toponly):
         # this removes context lines from the top and bottom of list 'l'.  It
         # checks the hunk to make sure only context lines are removed, and then
         # returns a new shortened list of lines.
         fuzz = min(fuzz, len(old))
         if fuzz:
             top = 0
             bot = 0
             hlen = len(self.hunk)
             for x in pycompat.xrange(hlen - 1):
                 # the hunk starts with the @@ line, so use x+1
                 if self.hunk[x + 1].startswith(' '):
                     top += 1
                 else:
                     break
             if not toponly:
                 for x in pycompat.xrange(hlen - 1):
                     if self.hunk[hlen - bot - 1].startswith(' '):
                         bot += 1
                     else:
                         break
 
             bot = min(fuzz, bot)
             top = min(fuzz, top)
             return old[top:len(old) - bot], new[top:len(new) - bot], top
         return old, new, 0
 
     def fuzzit(self, fuzz, toponly):
         old, new, top = self._fuzzit(self.a, self.b, fuzz, toponly)
         oldstart = self.starta + top
         newstart = self.startb + top
         # zero length hunk ranges already have their start decremented
         if self.lena and oldstart > 0:
             oldstart -= 1
         if self.lenb and newstart > 0:
             newstart -= 1
         return old, oldstart, new, newstart
 
 class binhunk(object):
     'A binary patch file.'
     def __init__(self, lr, fname):
         self.text = None
         self.delta = False
         self.hunk = ['GIT binary patch\n']
         self._fname = fname
         self._read(lr)
 
     def complete(self):
         return self.text is not None
 
     def new(self, lines):
         if self.delta:
             return [applybindelta(self.text, ''.join(lines))]
         return [self.text]
 
     def _read(self, lr):
         def getline(lr, hunk):
             l = lr.readline()
             hunk.append(l)
             return l.rstrip('\r\n')
 
         while True:
             line = getline(lr, self.hunk)
             if not line:
                 raise PatchError(_('could not extract "%s" binary data')
                                  % self._fname)
             if line.startswith('literal '):
                 size = int(line[8:].rstrip())
                 break
             if line.startswith('delta '):
                 size = int(line[6:].rstrip())
                 self.delta = True
                 break
         dec = []
         line = getline(lr, self.hunk)
         while len(line) > 1:
             l = line[0:1]
             if l <= 'Z' and l >= 'A':
                 l = ord(l) - ord('A') + 1
             else:
                 l = ord(l) - ord('a') + 27
             try:
                 dec.append(util.b85decode(line[1:])[:l])
             except ValueError as e:
                 raise PatchError(_('could not decode "%s" binary patch: %s')
                                  % (self._fname, stringutil.forcebytestr(e)))
             line = getline(lr, self.hunk)
         text = zlib.decompress(''.join(dec))
         if len(text) != size:
             raise PatchError(_('"%s" length is %d bytes, should be %d')
                              % (self._fname, len(text), size))
         self.text = text
 
 def parsefilename(str):
     # --- filename \t|space stuff
     s = str[4:].rstrip('\r\n')
     i = s.find('\t')
     if i < 0:
         i = s.find(' ')
         if i < 0:
             return s
     return s[:i]
 
 def reversehunks(hunks):
     '''reverse the signs in the hunks given as argument
 
     This function operates on hunks coming out of patch.filterpatch, that is
     a list of the form: [header1, hunk1, hunk2, header2...]. Example usage:
 
     >>> rawpatch = b"""diff --git a/folder1/g b/folder1/g
     ... --- a/folder1/g
     ... +++ b/folder1/g
     ... @@ -1,7 +1,7 @@
     ... +firstline
     ...  c
     ...  1
     ...  2
     ... + 3
     ... -4
     ...  5
     ...  d
     ... +lastline"""
     >>> hunks = parsepatch([rawpatch])
     >>> hunkscomingfromfilterpatch = []
     >>> for h in hunks:
     ...     hunkscomingfromfilterpatch.append(h)
     ...     hunkscomingfromfilterpatch.extend(h.hunks)
 
     >>> reversedhunks = reversehunks(hunkscomingfromfilterpatch)
     >>> from . import util
     >>> fp = util.stringio()
     >>> for c in reversedhunks:
     ...      c.write(fp)
     >>> fp.seek(0) or None
     >>> reversedpatch = fp.read()
     >>> print(pycompat.sysstr(reversedpatch))
     diff --git a/folder1/g b/folder1/g
     --- a/folder1/g
     +++ b/folder1/g
     @@ -1,4 +1,3 @@
     -firstline
      c
      1
      2
     @@ -2,6 +1,6 @@
      c
      1
      2
     - 3
     +4
      5
      d
     @@ -6,3 +5,2 @@
      5
      d
     -lastline
 
     '''
 
     newhunks = []
     for c in hunks:
         if util.safehasattr(c, 'reversehunk'):
             c = c.reversehunk()
         newhunks.append(c)
     return newhunks
 
 def parsepatch(originalchunks, maxcontext=None):
     """patch -> [] of headers -> [] of hunks
 
     If maxcontext is not None, trim context lines if necessary.
 
     >>> rawpatch = b'''diff --git a/folder1/g b/folder1/g
     ... --- a/folder1/g
     ... +++ b/folder1/g
     ... @@ -1,8 +1,10 @@
     ...  1
     ...  2
     ... -3
     ...  4
     ...  5
     ...  6
     ... +6.1
     ... +6.2
     ...  7
     ...  8
     ... +9'''
     >>> out = util.stringio()
     >>> headers = parsepatch([rawpatch], maxcontext=1)
     >>> for header in headers:
     ...     header.write(out)
     ...     for hunk in header.hunks:
     ...         hunk.write(out)
     >>> print(pycompat.sysstr(out.getvalue()))
     diff --git a/folder1/g b/folder1/g
     --- a/folder1/g
     +++ b/folder1/g
     @@ -2,3 +2,2 @@
      2
     -3
      4
     @@ -6,2 +5,4 @@
      6
     +6.1
     +6.2
      7
     @@ -8,1 +9,2 @@
      8
     +9
     """
     class parser(object):
         """patch parsing state machine"""
         def __init__(self):
             self.fromline = 0
             self.toline = 0
             self.proc = ''
             self.header = None
             self.context = []
             self.before = []
             self.hunk = []
             self.headers = []
 
         def addrange(self, limits):
             self.addcontext([])
             fromstart, fromend, tostart, toend, proc = limits
             self.fromline = int(fromstart)
             self.toline = int(tostart)
             self.proc = proc
 
         def addcontext(self, context):
             if self.hunk:
                 h = recordhunk(self.header, self.fromline, self.toline,
                         self.proc, self.before, self.hunk, context, maxcontext)
                 self.header.hunks.append(h)
                 self.fromline += len(self.before) + h.removed
                 self.toline += len(self.before) + h.added
                 self.before = []
                 self.hunk = []
             self.context = context
 
         def addhunk(self, hunk):
             if self.context:
                 self.before = self.context
                 self.context = []
             if self.hunk:
                 self.addcontext([])
             self.hunk = hunk
 
         def newfile(self, hdr):
             self.addcontext([])
             h = header(hdr)
             self.headers.append(h)
             self.header = h
 
         def addother(self, line):
             pass # 'other' lines are ignored
 
         def finished(self):
             self.addcontext([])
             return self.headers
 
         transitions = {
             'file': {'context': addcontext,
                      'file': newfile,
                      'hunk': addhunk,
                      'range': addrange},
             'context': {'file': newfile,
                         'hunk': addhunk,
                         'range': addrange,
                         'other': addother},
             'hunk': {'context': addcontext,
                      'file': newfile,
                      'range': addrange},
             'range': {'context': addcontext,
                       'hunk': addhunk},
             'other': {'other': addother},
             }
 
     p = parser()
     fp = stringio()
     fp.write(''.join(originalchunks))
     fp.seek(0)
 
     state = 'context'
     for newstate, data in scanpatch(fp):
         try:
             p.transitions[state][newstate](p, data)
         except KeyError:
             raise PatchError('unhandled transition: %s -> %s' %
                                    (state, newstate))
         state = newstate
     del fp
     return p.finished()
 
 def pathtransform(path, strip, prefix):
     '''turn a path from a patch into a path suitable for the repository
 
     prefix, if not empty, is expected to be normalized with a / at the end.
 
     Returns (stripped components, path in repository).
 
     >>> pathtransform(b'a/b/c', 0, b'')
     ('', 'a/b/c')
     >>> pathtransform(b'   a/b/c   ', 0, b'')
     ('', '   a/b/c')
     >>> pathtransform(b'   a/b/c   ', 2, b'')
     ('a/b/', 'c')
     >>> pathtransform(b'a/b/c', 0, b'd/e/')
     ('', 'd/e/a/b/c')
     >>> pathtransform(b'   a//b/c   ', 2, b'd/e/')
     ('a//b/', 'd/e/c')
     >>> pathtransform(b'a/b/c', 3, b'')
     Traceback (most recent call last):
     PatchError: unable to strip away 1 of 3 dirs from a/b/c
     '''
     pathlen = len(path)
     i = 0
     if strip == 0:
         return '', prefix + path.rstrip()
     count = strip
     while count > 0:
         i = path.find('/', i)
         if i == -1:
             raise PatchError(_("unable to strip away %d of %d dirs from %s") %
                              (count, strip, path))
         i += 1
         # consume '//' in the path
         while i < pathlen - 1 and path[i:i + 1] == '/':
             i += 1
         count -= 1
     return path[:i].lstrip(), prefix + path[i:].rstrip()
 
 def makepatchmeta(backend, afile_orig, bfile_orig, hunk, strip, prefix):
     nulla = afile_orig == "/dev/null"
     nullb = bfile_orig == "/dev/null"
     create = nulla and hunk.starta == 0 and hunk.lena == 0
     remove = nullb and hunk.startb == 0 and hunk.lenb == 0
     abase, afile = pathtransform(afile_orig, strip, prefix)
     gooda = not nulla and backend.exists(afile)
     bbase, bfile = pathtransform(bfile_orig, strip, prefix)
     if afile == bfile:
         goodb = gooda
     else:
         goodb = not nullb and backend.exists(bfile)
     missing = not goodb and not gooda and not create
 
     # some diff programs apparently produce patches where the afile is
     # not /dev/null, but afile starts with bfile
     abasedir = afile[:afile.rfind('/') + 1]
     bbasedir = bfile[:bfile.rfind('/') + 1]
     if (missing and abasedir == bbasedir and afile.startswith(bfile)
         and hunk.starta == 0 and hunk.lena == 0):
         create = True
         missing = False
 
     # If afile is "a/b/foo" and bfile is "a/b/foo.orig" we assume the
     # diff is between a file and its backup. In this case, the original
     # file should be patched (see original mpatch code).
     isbackup = (abase == bbase and bfile.startswith(afile))
     fname = None
     if not missing:
         if gooda and goodb:
             if isbackup:
                 fname = afile
             else:
                 fname = bfile
         elif gooda:
             fname = afile
 
     if not fname:
         if not nullb:
             if isbackup:
                 fname = afile
             else:
                 fname = bfile
         elif not nulla:
             fname = afile
         else:
             raise PatchError(_("undefined source and destination files"))
 
     gp = patchmeta(fname)
     if create:
         gp.op = 'ADD'
     elif remove:
         gp.op = 'DELETE'
     return gp
 
 def scanpatch(fp):
     """like patch.iterhunks, but yield different events
 
     - ('file',    [header_lines + fromfile + tofile])
     - ('context', [context_lines])
     - ('hunk',    [hunk_lines])
     - ('range',   (-start,len, +start,len, proc))
     """
     lines_re = re.compile(br'@@ -(\d+),(\d+) \+(\d+),(\d+) @@\s*(.*)')
     lr = linereader(fp)
 
     def scanwhile(first, p):
         """scan lr while predicate holds"""
         lines = [first]
         for line in iter(lr.readline, ''):
             if p(line):
                 lines.append(line)
             else:
                 lr.push(line)
                 break
         return lines
 
     for line in iter(lr.readline, ''):
         if line.startswith('diff --git a/') or line.startswith('diff -r '):
             def notheader(line):
                 s = line.split(None, 1)
                 return not s or s[0] not in ('---', 'diff')
             header = scanwhile(line, notheader)
             fromfile = lr.readline()
             if fromfile.startswith('---'):
                 tofile = lr.readline()
                 header += [fromfile, tofile]
             else:
                 lr.push(fromfile)
             yield 'file', header
         elif line.startswith(' '):
             cs = (' ', '\\')
             yield 'context', scanwhile(line, lambda l: l.startswith(cs))
         elif line.startswith(('-', '+')):
             cs = ('-', '+', '\\')
             yield 'hunk', scanwhile(line, lambda l: l.startswith(cs))
         else:
             m = lines_re.match(line)
             if m:
                 yield 'range', m.groups()
             else:
                 yield 'other', line
 
 def scangitpatch(lr, firstline):
     """
     Git patches can emit:
     - rename a to b
     - change b
     - copy a to c
     - change c
 
     We cannot apply this sequence as-is, the renamed 'a' could not be
     found for it would have been renamed already. And we cannot copy
     from 'b' instead because 'b' would have been changed already. So
     we scan the git patch for copy and rename commands so we can
     perform the copies ahead of time.
     """
     pos = 0
     try:
         pos = lr.fp.tell()
         fp = lr.fp
     except IOError:
         fp = stringio(lr.fp.read())
     gitlr = linereader(fp)
     gitlr.push(firstline)
     gitpatches = readgitpatch(gitlr)
     fp.seek(pos)
     return gitpatches
 
 def iterhunks(fp):
     """Read a patch and yield the following events:
     - ("file", afile, bfile, firsthunk): select a new target file.
     - ("hunk", hunk): a new hunk is ready to be applied, follows a
     "file" event.
     - ("git", gitchanges): current diff is in git format, gitchanges
     maps filenames to gitpatch records. Unique event.
     """
     afile = ""
     bfile = ""
     state = None
     hunknum = 0
     emitfile = newfile = False
     gitpatches = None
 
     # our states
     BFILE = 1
     context = None
     lr = linereader(fp)
 
     for x in iter(lr.readline, ''):
         if state == BFILE and (
             (not context and x.startswith('@'))
             or (context is not False and x.startswith('***************'))
             or x.startswith('GIT binary patch')):
             gp = None
             if (gitpatches and
                 gitpatches[-1].ispatching(afile, bfile)):
                 gp = gitpatches.pop()
             if x.startswith('GIT binary patch'):
                 h = binhunk(lr, gp.path)
             else:
                 if context is None and x.startswith('***************'):
                     context = True
                 h = hunk(x, hunknum + 1, lr, context)
             hunknum += 1
             if emitfile:
                 emitfile = False
                 yield 'file', (afile, bfile, h, gp and gp.copy() or None)
             yield 'hunk', h
         elif x.startswith('diff --git a/'):
             m = gitre.match(x.rstrip(' \r\n'))
             if not m:
                 continue
             if gitpatches is None:
                 # scan whole input for git metadata
                 gitpatches = scangitpatch(lr, x)
                 yield 'git', [g.copy() for g in gitpatches
                               if g.op in ('COPY', 'RENAME')]
                 gitpatches.reverse()
             afile = 'a/' + m.group(1)
             bfile = 'b/' + m.group(2)
             while gitpatches and not gitpatches[-1].ispatching(afile, bfile):
                 gp = gitpatches.pop()
                 yield 'file', ('a/' + gp.path, 'b/' + gp.path, None, gp.copy())
             if not gitpatches:
                 raise PatchError(_('failed to synchronize metadata for "%s"')
                                  % afile[2:])
             newfile = True
         elif x.startswith('---'):
             # check for a unified diff
             l2 = lr.readline()
             if not l2.startswith('+++'):
                 lr.push(l2)
                 continue
             newfile = True
             context = False
             afile = parsefilename(x)
             bfile = parsefilename(l2)
         elif x.startswith('***'):
             # check for a context diff
             l2 = lr.readline()
             if not l2.startswith('---'):
                 lr.push(l2)
                 continue
             l3 = lr.readline()
             lr.push(l3)
             if not l3.startswith("***************"):
                 lr.push(l2)
                 continue
             newfile = True
             context = True
             afile = parsefilename(x)
             bfile = parsefilename(l2)
 
         if newfile:
             newfile = False
             emitfile = True
             state = BFILE
             hunknum = 0
 
     while gitpatches:
         gp = gitpatches.pop()
         yield 'file', ('a/' + gp.path, 'b/' + gp.path, None, gp.copy())
 
 def applybindelta(binchunk, data):
     """Apply a binary delta hunk
     The algorithm used is the algorithm from git's patch-delta.c
     """
     def deltahead(binchunk):
         i = 0
         for c in pycompat.bytestr(binchunk):
             i += 1
             if not (ord(c) & 0x80):
                 return i
         return i
     out = ""
     s = deltahead(binchunk)
     binchunk = binchunk[s:]
     s = deltahead(binchunk)
     binchunk = binchunk[s:]
     i = 0
     while i < len(binchunk):
         cmd = ord(binchunk[i:i + 1])
         i += 1
         if (cmd & 0x80):
             offset = 0
             size = 0
             if (cmd & 0x01):
                 offset = ord(binchunk[i:i + 1])
                 i += 1
             if (cmd & 0x02):
                 offset |= ord(binchunk[i:i + 1]) << 8
                 i += 1
             if (cmd & 0x04):
                 offset |= ord(binchunk[i:i + 1]) << 16
                 i += 1
             if (cmd & 0x08):
                 offset |= ord(binchunk[i:i + 1]) << 24
                 i += 1
             if (cmd & 0x10):
                 size = ord(binchunk[i:i + 1])
                 i += 1
             if (cmd & 0x20):
                 size |= ord(binchunk[i:i + 1]) << 8
                 i += 1
             if (cmd & 0x40):
                 size |= ord(binchunk[i:i + 1]) << 16
                 i += 1
             if size == 0:
                 size = 0x10000
             offset_end = offset + size
             out += data[offset:offset_end]
         elif cmd != 0:
             offset_end = i + cmd
             out += binchunk[i:offset_end]
             i += cmd
         else:
             raise PatchError(_('unexpected delta opcode 0'))
     return out
 
 def applydiff(ui, fp, backend, store, strip=1, prefix='', eolmode='strict'):
     """Reads a patch from fp and tries to apply it.
 
     Returns 0 for a clean patch, -1 if any rejects were found and 1 if
     there was any fuzz.
 
     If 'eolmode' is 'strict', the patch content and patched file are
     read in binary mode. Otherwise, line endings are ignored when
     patching then normalized according to 'eolmode'.
     """
     return _applydiff(ui, fp, patchfile, backend, store, strip=strip,
                       prefix=prefix, eolmode=eolmode)
 
 def _canonprefix(repo, prefix):
     if prefix:
         prefix = pathutil.canonpath(repo.root, repo.getcwd(), prefix)
         if prefix != '':
             prefix += '/'
     return prefix
 
 def _applydiff(ui, fp, patcher, backend, store, strip=1, prefix='',
                eolmode='strict'):
     prefix = _canonprefix(backend.repo, prefix)
     def pstrip(p):
         return pathtransform(p, strip - 1, prefix)[1]
 
     rejects = 0
     err = 0
     current_file = None
 
     for state, values in iterhunks(fp):
         if state == 'hunk':
             if not current_file:
                 continue
             ret = current_file.apply(values)
             if ret > 0:
                 err = 1
         elif state == 'file':
             if current_file:
                 rejects += current_file.close()
                 current_file = None
             afile, bfile, first_hunk, gp = values
             if gp:
                 gp.path = pstrip(gp.path)
                 if gp.oldpath:
                     gp.oldpath = pstrip(gp.oldpath)
             else:
                 gp = makepatchmeta(backend, afile, bfile, first_hunk, strip,
                                    prefix)
             if gp.op == 'RENAME':
                 backend.unlink(gp.oldpath)
             if not first_hunk:
                 if gp.op == 'DELETE':
                     backend.unlink(gp.path)
                     continue
                 data, mode = None, None
                 if gp.op in ('RENAME', 'COPY'):
                     data, mode = store.getfile(gp.oldpath)[:2]
                     if data is None:
                         # This means that the old path does not exist
                         raise PatchError(_("source file '%s' does not exist")
                                            % gp.oldpath)
                 if gp.mode:
                     mode = gp.mode
                     if gp.op == 'ADD':
                         # Added files without content have no hunk and
                         # must be created
                         data = ''
                 if data or mode:
                     if (gp.op in ('ADD', 'RENAME', 'COPY')
                         and backend.exists(gp.path)):
                         raise PatchError(_("cannot create %s: destination "
                                            "already exists") % gp.path)
                     backend.setfile(gp.path, data, mode, gp.oldpath)
                 continue
             try:
                 current_file = patcher(ui, gp, backend, store,
                                        eolmode=eolmode)
             except PatchError as inst:
                 ui.warn(str(inst) + '\n')
                 current_file = None
                 rejects += 1
                 continue
         elif state == 'git':
             for gp in values:
                 path = pstrip(gp.oldpath)
                 data, mode = backend.getfile(path)
                 if data is None:
                     # The error ignored here will trigger a getfile()
                     # error in a place more appropriate for error
                     # handling, and will not interrupt the patching
                     # process.
                     pass
                 else:
                     store.setfile(path, data, mode)
         else:
             raise error.Abort(_('unsupported parser state: %s') % state)
 
     if current_file:
         rejects += current_file.close()
 
     if rejects:
         return -1
     return err
 
 def _externalpatch(ui, repo, patcher, patchname, strip, files,
                    similarity):
     """use <patcher> to apply <patchname> to the working directory.
     returns whether patch was applied with fuzz factor."""
 
     fuzz = False
     args = []
     cwd = repo.root
     if cwd:
         args.append('-d %s' % procutil.shellquote(cwd))
     cmd = ('%s %s -p%d < %s'
            % (patcher, ' '.join(args), strip, procutil.shellquote(patchname)))
     ui.debug('Using external patch tool: %s\n' % cmd)
     fp = procutil.popen(cmd, 'rb')
     try:
         for line in util.iterfile(fp):
             line = line.rstrip()
             ui.note(line + '\n')
             if line.startswith('patching file '):
                 pf = util.parsepatchoutput(line)
                 printed_file = False
                 files.add(pf)
             elif line.find('with fuzz') >= 0:
                 fuzz = True
                 if not printed_file:
                     ui.warn(pf + '\n')
                     printed_file = True
                 ui.warn(line + '\n')
             elif line.find('saving rejects to file') >= 0:
                 ui.warn(line + '\n')
             elif line.find('FAILED') >= 0:
                 if not printed_file:
                     ui.warn(pf + '\n')
                     printed_file = True
                 ui.warn(line + '\n')
     finally:
         if files:
             scmutil.marktouched(repo, files, similarity)
     code = fp.close()
     if code:
         raise PatchError(_("patch command failed: %s") %
                          procutil.explainexit(code))
     return fuzz
 
 def patchbackend(ui, backend, patchobj, strip, prefix, files=None,
                  eolmode='strict'):
     if files is None:
         files = set()
     if eolmode is None:
         eolmode = ui.config('patch', 'eol')
     if eolmode.lower() not in eolmodes:
         raise error.Abort(_('unsupported line endings type: %s') % eolmode)
     eolmode = eolmode.lower()
 
     store = filestore()
     try:
         fp = open(patchobj, 'rb')
     except TypeError:
         fp = patchobj
     try:
         ret = applydiff(ui, fp, backend, store, strip=strip, prefix=prefix,
                         eolmode=eolmode)
     finally:
         if fp != patchobj:
             fp.close()
         files.update(backend.close())
         store.close()
     if ret < 0:
         raise PatchError(_('patch failed to apply'))
     return ret > 0
 
 def internalpatch(ui, repo, patchobj, strip, prefix='', files=None,
                   eolmode='strict', similarity=0):
     """use builtin patch to apply <patchobj> to the working directory.
     returns whether patch was applied with fuzz factor."""
     backend = workingbackend(ui, repo, similarity)
     return patchbackend(ui, backend, patchobj, strip, prefix, files, eolmode)
 
 def patchrepo(ui, repo, ctx, store, patchobj, strip, prefix, files=None,
               eolmode='strict'):
     backend = repobackend(ui, repo, ctx, store)
     return patchbackend(ui, backend, patchobj, strip, prefix, files, eolmode)
 
 def patch(ui, repo, patchname, strip=1, prefix='', files=None, eolmode='strict',
           similarity=0):
     """Apply <patchname> to the working directory.
 
     'eolmode' specifies how end of lines should be handled. It can be:
     - 'strict': inputs are read in binary mode, EOLs are preserved
     - 'crlf': EOLs are ignored when patching and reset to CRLF
     - 'lf': EOLs are ignored when patching and reset to LF
     - None: get it from user settings, default to 'strict'
     'eolmode' is ignored when using an external patcher program.
 
     Returns whether patch was applied with fuzz factor.
     """
     patcher = ui.config('ui', 'patch')
     if files is None:
         files = set()
     if patcher:
         return _externalpatch(ui, repo, patcher, patchname, strip,
                               files, similarity)
     return internalpatch(ui, repo, patchname, strip, prefix, files, eolmode,
                          similarity)
 
 def changedfiles(ui, repo, patchpath, strip=1, prefix=''):
     backend = fsbackend(ui, repo.root)
     prefix = _canonprefix(repo, prefix)
     with open(patchpath, 'rb') as fp:
         changed = set()
         for state, values in iterhunks(fp):
             if state == 'file':
                 afile, bfile, first_hunk, gp = values
                 if gp:
                     gp.path = pathtransform(gp.path, strip - 1, prefix)[1]
                     if gp.oldpath:
                         gp.oldpath = pathtransform(gp.oldpath, strip - 1,
                                                    prefix)[1]
                 else:
                     gp = makepatchmeta(backend, afile, bfile, first_hunk, strip,
                                        prefix)
                 changed.add(gp.path)
                 if gp.op == 'RENAME':
                     changed.add(gp.oldpath)
             elif state not in ('hunk', 'git'):
                 raise error.Abort(_('unsupported parser state: %s') % state)
         return changed
 
 class GitDiffRequired(Exception):
     pass
 
 diffopts = diffutil.diffallopts
 diffallopts = diffutil.diffallopts
 difffeatureopts = diffutil.difffeatureopts
 
 def diff(repo, node1=None, node2=None, match=None, changes=None,
          opts=None, losedatafn=None, prefix='', relroot='', copy=None,
-         hunksfilterfn=None):
+         copysourcematch=None, hunksfilterfn=None):
     '''yields diff of changes to files between two nodes, or node and
     working directory.
 
     if node1 is None, use first dirstate parent instead.
     if node2 is None, compare node1 with working directory.
 
     losedatafn(**kwarg) is a callable run when opts.upgrade=True and
     every time some change cannot be represented with the current
     patch format. Return False to upgrade to git patch format, True to
     accept the loss or raise an exception to abort the diff. It is
     called with the name of current file being diffed as 'fn'. If set
     to None, patches will always be upgraded to git format when
     necessary.
 
     prefix is a filename prefix that is prepended to all filenames on
     display (used for subrepos).
 
     relroot, if not empty, must be normalized with a trailing /. Any match
     patterns that fall outside it will be ignored.
 
     copy, if not empty, should contain mappings {dst@y: src@x} of copy
     information.
 
+    if copysourcematch is not None, then copy sources will be filtered by this
+    matcher
+
     hunksfilterfn, if not None, should be a function taking a filectx and
     hunks generator that may yield filtered hunks.
     '''
     if not node1 and not node2:
         node1 = repo.dirstate.p1()
 
     ctx1 = repo[node1]
     ctx2 = repo[node2]
 
     for fctx1, fctx2, hdr, hunks in diffhunks(
             repo, ctx1=ctx1, ctx2=ctx2,
             match=match, changes=changes, opts=opts,
             losedatafn=losedatafn, prefix=prefix, relroot=relroot, copy=copy,
-    ):
+            copysourcematch=copysourcematch):
         if hunksfilterfn is not None:
             # If the file has been removed, fctx2 is None; but this should
             # not occur here since we catch removed files early in
             # logcmdutil.getlinerangerevs() for 'hg log -L'.
             assert fctx2 is not None, \
                 'fctx2 unexpectly None in diff hunks filtering'
             hunks = hunksfilterfn(fctx2, hunks)
         text = ''.join(sum((list(hlines) for hrange, hlines in hunks), []))
         if hdr and (text or len(hdr) > 1):
             yield '\n'.join(hdr) + '\n'
         if text:
             yield text
 
 def diffhunks(repo, ctx1, ctx2, match=None, changes=None,
-              opts=None, losedatafn=None, prefix='', relroot='', copy=None):
+              opts=None, losedatafn=None, prefix='', relroot='', copy=None,
+              copysourcematch=None):
     """Yield diff of changes to files in the form of (`header`, `hunks`) tuples
     where `header` is a list of diff headers and `hunks` is an iterable of
     (`hunkrange`, `hunklines`) tuples.
 
     See diff() for the meaning of parameters.
     """
 
     if opts is None:
         opts = mdiff.defaultopts
 
     def lrugetfilectx():
         cache = {}
         order = collections.deque()
         def getfilectx(f, ctx):
             fctx = ctx.filectx(f, filelog=cache.get(f))
             if f not in cache:
                 if len(cache) > 20:
                     del cache[order.popleft()]
                 cache[f] = fctx.filelog()
             else:
                 order.remove(f)
             order.append(f)
             return fctx
         return getfilectx
     getfilectx = lrugetfilectx()
 
     if not changes:
         changes = ctx1.status(ctx2, match=match)
     modified, added, removed = changes[:3]
 
     if not modified and not added and not removed:
         return []
 
     if repo.ui.debugflag:
         hexfunc = hex
     else:
         hexfunc = short
     revs = [hexfunc(node) for node in [ctx1.node(), ctx2.node()] if node]
 
     if copy is None:
         copy = {}
         if opts.git or opts.upgrade:
             copy = copies.pathcopies(ctx1, ctx2, match=match)
 
-    if relroot:
-        # filter out copies where source side isn't inside the relative root
+    if copysourcematch:
+        # filter out copies where source side isn't inside the matcher
         # (copies.pathcopies() already filtered out the destination)
         copy = {dst: src for dst, src in copy.iteritems()
-                if src.startswith(relroot)}
+                if copysourcematch(src)}
 
     modifiedset = set(modified)
     addedset = set(added)
     removedset = set(removed)
     for f in modified:
         if f not in ctx1:
             # Fix up added, since merged-in additions appear as
             # modifications during merges
             modifiedset.remove(f)
             addedset.add(f)
     for f in removed:
         if f not in ctx1:
             # Merged-in additions that are then removed are reported as removed.
             # They are not in ctx1, so We don't want to show them in the diff.
             removedset.remove(f)
     modified = sorted(modifiedset)
     added = sorted(addedset)
     removed = sorted(removedset)
     for dst, src in list(copy.items()):
         if src not in ctx1:
             # Files merged in during a merge and then copied/renamed are
             # reported as copies. We want to show them in the diff as additions.
             del copy[dst]
 
     prefetchmatch = scmutil.matchfiles(
         repo, list(modifiedset | addedset | removedset))
     scmutil.prefetchfiles(repo, [ctx1.rev(), ctx2.rev()], prefetchmatch)
 
     def difffn(opts, losedata):
         return trydiff(repo, revs, ctx1, ctx2, modified, added, removed,
                        copy, getfilectx, opts, losedata, prefix, relroot)
     if opts.upgrade and not opts.git:
         try:
             def losedata(fn):
                 if not losedatafn or not losedatafn(fn=fn):
                     raise GitDiffRequired
             # Buffer the whole output until we are sure it can be generated
             return list(difffn(opts.copy(git=False), losedata))
         except GitDiffRequired:
             return difffn(opts.copy(git=True), None)
     else:
         return difffn(opts, None)
 
 def diffsinglehunk(hunklines):
     """yield tokens for a list of lines in a single hunk"""
     for line in hunklines:
         # chomp
         chompline = line.rstrip('\r\n')
         # highlight tabs and trailing whitespace
         stripline = chompline.rstrip()
         if line.startswith('-'):
             label = 'diff.deleted'
         elif line.startswith('+'):
             label = 'diff.inserted'
         else:
             raise error.ProgrammingError('unexpected hunk line: %s' % line)
         for token in tabsplitter.findall(stripline):
             if token.startswith('\t'):
                 yield (token, 'diff.tab')
             else:
                 yield (token, label)
 
         if chompline != stripline:
             yield (chompline[len(stripline):], 'diff.trailingwhitespace')
         if chompline != line:
             yield (line[len(chompline):], '')
 
 def diffsinglehunkinline(hunklines):
     """yield tokens for a list of lines in a single hunk, with inline colors"""
     # prepare deleted, and inserted content
     a = ''
     b = ''
     for line in hunklines:
         if line[0:1] == '-':
             a += line[1:]
         elif line[0:1] == '+':
             b += line[1:]
         else:
             raise error.ProgrammingError('unexpected hunk line: %s' % line)
     # fast path: if either side is empty, use diffsinglehunk
     if not a or not b:
         for t in diffsinglehunk(hunklines):
             yield t
         return
     # re-split the content into words
     al = wordsplitter.findall(a)
     bl = wordsplitter.findall(b)
     # re-arrange the words to lines since the diff algorithm is line-based
     aln = [s if s == '\n' else s + '\n' for s in al]
     bln = [s if s == '\n' else s + '\n' for s in bl]
     an = ''.join(aln)
     bn = ''.join(bln)
     # run the diff algorithm, prepare atokens and btokens
     atokens = []
     btokens = []
     blocks = mdiff.allblocks(an, bn, lines1=aln, lines2=bln)
     for (a1, a2, b1, b2), btype in blocks:
         changed = btype == '!'
         for token in mdiff.splitnewlines(''.join(al[a1:a2])):
             atokens.append((changed, token))
         for token in mdiff.splitnewlines(''.join(bl[b1:b2])):
             btokens.append((changed, token))
 
     # yield deleted tokens, then inserted ones
     for prefix, label, tokens in [('-', 'diff.deleted', atokens),
                                   ('+', 'diff.inserted', btokens)]:
         nextisnewline = True
         for changed, token in tokens:
             if nextisnewline:
                 yield (prefix, label)
                 nextisnewline = False
             # special handling line end
             isendofline = token.endswith('\n')
             if isendofline:
                 chomp = token[:-1] # chomp
                 if chomp.endswith('\r'):
                     chomp = chomp[:-1]
                 endofline = token[len(chomp):]
                 token = chomp.rstrip() # detect spaces at the end
                 endspaces = chomp[len(token):]
             # scan tabs
             for maybetab in tabsplitter.findall(token):
                 if b'\t' == maybetab[0:1]:
                     currentlabel = 'diff.tab'
                 else:
                     if changed:
                         currentlabel = label + '.changed'
                     else:
                         currentlabel = label + '.unchanged'
                 yield (maybetab, currentlabel)
             if isendofline:
                 if endspaces:
                     yield (endspaces, 'diff.trailingwhitespace')
                 yield (endofline, '')
                 nextisnewline = True
 
 def difflabel(func, *args, **kw):
     '''yields 2-tuples of (output, label) based on the output of func()'''
     if kw.get(r'opts') and kw[r'opts'].worddiff:
         dodiffhunk = diffsinglehunkinline
     else:
         dodiffhunk = diffsinglehunk
     headprefixes = [('diff', 'diff.diffline'),
                     ('copy', 'diff.extended'),
                     ('rename', 'diff.extended'),
                     ('old', 'diff.extended'),
                     ('new', 'diff.extended'),
                     ('deleted', 'diff.extended'),
                     ('index', 'diff.extended'),
                     ('similarity', 'diff.extended'),
                     ('---', 'diff.file_a'),
                     ('+++', 'diff.file_b')]
     textprefixes = [('@', 'diff.hunk'),
                     # - and + are handled by diffsinglehunk
                    ]
     head = False
 
     # buffers a hunk, i.e. adjacent "-", "+" lines without other changes.
     hunkbuffer = []
     def consumehunkbuffer():
         if hunkbuffer:
             for token in dodiffhunk(hunkbuffer):
                 yield token
             hunkbuffer[:] = []
 
     for chunk in func(*args, **kw):
         lines = chunk.split('\n')
         linecount = len(lines)
         for i, line in enumerate(lines):
             if head:
                 if line.startswith('@'):
                     head = False
             else:
                 if line and not line.startswith((' ', '+', '-', '@', '\\')):
                     head = True
             diffline = False
             if not head and line and line.startswith(('+', '-')):
                 diffline = True
 
             prefixes = textprefixes
             if head:
                 prefixes = headprefixes
             if diffline:
                 # buffered
                 bufferedline = line
                 if i + 1 < linecount:
                     bufferedline += "\n"
                 hunkbuffer.append(bufferedline)
             else:
                 # unbuffered
                 for token in consumehunkbuffer():
                     yield token
                 stripline = line.rstrip()
                 for prefix, label in prefixes:
                     if stripline.startswith(prefix):
                         yield (stripline, label)
                         if line != stripline:
                             yield (line[len(stripline):],
                                    'diff.trailingwhitespace')
                         break
                 else:
                     yield (line, '')
                 if i + 1 < linecount:
                     yield ('\n', '')
         for token in consumehunkbuffer():
             yield token
 
 def diffui(*args, **kw):
     '''like diff(), but yields 2-tuples of (output, label) for ui.write()'''
     return difflabel(diff, *args, **kw)
 
 def _filepairs(modified, added, removed, copy, opts):
     '''generates tuples (f1, f2, copyop), where f1 is the name of the file
     before and f2 is the the name after. For added files, f1 will be None,
     and for removed files, f2 will be None. copyop may be set to None, 'copy'
     or 'rename' (the latter two only if opts.git is set).'''
     gone = set()
 
     copyto = dict([(v, k) for k, v in copy.items()])
 
     addedset, removedset = set(added), set(removed)
 
     for f in sorted(modified + added + removed):
         copyop = None
         f1, f2 = f, f
         if f in addedset:
             f1 = None
             if f in copy:
                 if opts.git:
                     f1 = copy[f]
                     if f1 in removedset and f1 not in gone:
                         copyop = 'rename'
                         gone.add(f1)
                     else:
                         copyop = 'copy'
         elif f in removedset:
             f2 = None
             if opts.git:
                 # have we already reported a copy above?
                 if (f in copyto and copyto[f] in addedset
                     and copy[copyto[f]] == f):
                     continue
         yield f1, f2, copyop
 
 def trydiff(repo, revs, ctx1, ctx2, modified, added, removed,
             copy, getfilectx, opts, losedatafn, prefix, relroot):
     '''given input data, generate a diff and yield it in blocks
 
     If generating a diff would lose data like flags or binary data and
     losedatafn is not None, it will be called.
 
     relroot is removed and prefix is added to every path in the diff output.
 
     If relroot is not empty, this function expects every path in modified,
     added, removed and copy to start with it.'''
 
     def gitindex(text):
         if not text:
             text = ""
         l = len(text)
         s = hashlib.sha1('blob %d\0' % l)
         s.update(text)
         return hex(s.digest())
 
     if opts.noprefix:
         aprefix = bprefix = ''
     else:
         aprefix = 'a/'
         bprefix = 'b/'
 
     def diffline(f, revs):
         revinfo = ' '.join(["-r %s" % rev for rev in revs])
         return 'diff %s %s' % (revinfo, f)
 
     def isempty(fctx):
         return fctx is None or fctx.size() == 0
 
     date1 = dateutil.datestr(ctx1.date())
     date2 = dateutil.datestr(ctx2.date())
 
     gitmode = {'l': '120000', 'x': '100755', '': '100644'}
 
     if relroot != '' and (repo.ui.configbool('devel', 'all-warnings')
                           or repo.ui.configbool('devel', 'check-relroot')):
         for f in modified + added + removed + list(copy) + list(copy.values()):
             if f is not None and not f.startswith(relroot):
                 raise AssertionError(
                     "file %s doesn't start with relroot %s" % (f, relroot))
 
     for f1, f2, copyop in _filepairs(modified, added, removed, copy, opts):
         content1 = None
         content2 = None
         fctx1 = None
         fctx2 = None
         flag1 = None
         flag2 = None
         if f1:
             fctx1 = getfilectx(f1, ctx1)
             if opts.git or losedatafn:
                 flag1 = ctx1.flags(f1)
         if f2:
             fctx2 = getfilectx(f2, ctx2)
             if opts.git or losedatafn:
                 flag2 = ctx2.flags(f2)
         # if binary is True, output "summary" or "base85", but not "text diff"
         if opts.text:
             binary = False
         else:
             binary = any(f.isbinary() for f in [fctx1, fctx2] if f is not None)
 
         if losedatafn and not opts.git:
             if (binary or
                 # copy/rename
                 f2 in copy or
                 # empty file creation
                 (not f1 and isempty(fctx2)) or
                 # empty file deletion
                 (isempty(fctx1) and not f2) or
                 # create with flags
                 (not f1 and flag2) or
                 # change flags
                 (f1 and f2 and flag1 != flag2)):
                 losedatafn(f2 or f1)
 
         path1 = f1 or f2
         path2 = f2 or f1
         path1 = posixpath.join(prefix, path1[len(relroot):])
         path2 = posixpath.join(prefix, path2[len(relroot):])
         header = []
         if opts.git:
             header.append('diff --git %s%s %s%s' %
                           (aprefix, path1, bprefix, path2))
             if not f1: # added
                 header.append('new file mode %s' % gitmode[flag2])
             elif not f2: # removed
                 header.append('deleted file mode %s' % gitmode[flag1])
             else:  # modified/copied/renamed
                 mode1, mode2 = gitmode[flag1], gitmode[flag2]
                 if mode1 != mode2:
                     header.append('old mode %s' % mode1)
                     header.append('new mode %s' % mode2)
                 if copyop is not None:
                     if opts.showsimilarity:
                         sim = similar.score(ctx1[path1], ctx2[path2]) * 100
                         header.append('similarity index %d%%' % sim)
                     header.append('%s from %s' % (copyop, path1))
                     header.append('%s to %s' % (copyop, path2))
         elif revs and not repo.ui.quiet:
             header.append(diffline(path1, revs))
 
         #  fctx.is  | diffopts                | what to   | is fctx.data()
         #  binary() | text nobinary git index | output?   | outputted?
         # ------------------------------------|----------------------------
         #  yes      | no   no       no  *     | summary   | no
         #  yes      | no   no       yes *     | base85    | yes
         #  yes      | no   yes      no  *     | summary   | no
         #  yes      | no   yes      yes 0     | summary   | no
         #  yes      | no   yes      yes >0    | summary   | semi [1]
         #  yes      | yes  *        *   *     | text diff | yes
         #  no       | *    *        *   *     | text diff | yes
         # [1]: hash(fctx.data()) is outputted. so fctx.data() cannot be faked
         if binary and (not opts.git or (opts.git and opts.nobinary and not
                                         opts.index)):
             # fast path: no binary content will be displayed, content1 and
             # content2 are only used for equivalent test. cmp() could have a
             # fast path.
             if fctx1 is not None:
                 content1 = b'\0'
             if fctx2 is not None:
                 if fctx1 is not None and not fctx1.cmp(fctx2):
                     content2 = b'\0' # not different
                 else:
                     content2 = b'\0\0'
         else:
             # normal path: load contents
             if fctx1 is not None:
                 content1 = fctx1.data()
             if fctx2 is not None:
                 content2 = fctx2.data()
 
         if binary and opts.git and not opts.nobinary:
             text = mdiff.b85diff(content1, content2)
             if text:
                 header.append('index %s..%s' %
                               (gitindex(content1), gitindex(content2)))
             hunks = (None, [text]),
         else:
             if opts.git and opts.index > 0:
                 flag = flag1
                 if flag is None:
                     flag = flag2
                 header.append('index %s..%s %s' %
                               (gitindex(content1)[0:opts.index],
                                gitindex(content2)[0:opts.index],
                                gitmode[flag]))
 
             uheaders, hunks = mdiff.unidiff(content1, date1,
                                             content2, date2,
                                             path1, path2,
                                             binary=binary, opts=opts)
             header.extend(uheaders)
         yield fctx1, fctx2, header, hunks
 
 def diffstatsum(stats):
     maxfile, maxtotal, addtotal, removetotal, binary = 0, 0, 0, 0, False
     for f, a, r, b in stats:
         maxfile = max(maxfile, encoding.colwidth(f))
         maxtotal = max(maxtotal, a + r)
         addtotal += a
         removetotal += r
         binary = binary or b
 
     return maxfile, maxtotal, addtotal, removetotal, binary
 
 def diffstatdata(lines):
     diffre = re.compile(br'^diff .*-r [a-z0-9]+\s(.*)$')
 
     results = []
     filename, adds, removes, isbinary = None, 0, 0, False
 
     def addresult():
         if filename:
             results.append((filename, adds, removes, isbinary))
 
     # inheader is used to track if a line is in the
     # header portion of the diff.  This helps properly account
     # for lines that start with '--' or '++'
     inheader = False
 
     for line in lines:
         if line.startswith('diff'):
             addresult()
             # starting a new file diff
             # set numbers to 0 and reset inheader
             inheader = True
             adds, removes, isbinary = 0, 0, False
             if line.startswith('diff --git a/'):
                 filename = gitre.search(line).group(2)
             elif line.startswith('diff -r'):
                 # format: "diff -r ... -r ... filename"
                 filename = diffre.search(line).group(1)
         elif line.startswith('@@'):
             inheader = False
         elif line.startswith('+') and not inheader:
             adds += 1
         elif line.startswith('-') and not inheader:
             removes += 1
         elif (line.startswith('GIT binary patch') or
               line.startswith('Binary file')):
             isbinary = True
         elif line.startswith('rename from'):
             filename = line[12:]
         elif line.startswith('rename to'):
             filename += ' => %s' % line[10:]
     addresult()
     return results
 
 def diffstat(lines, width=80):
     output = []
     stats = diffstatdata(lines)
     maxname, maxtotal, totaladds, totalremoves, hasbinary = diffstatsum(stats)
 
     countwidth = len(str(maxtotal))
     if hasbinary and countwidth < 3:
         countwidth = 3
     graphwidth = width - countwidth - maxname - 6
     if graphwidth < 10:
         graphwidth = 10
 
     def scale(i):
         if maxtotal <= graphwidth:
             return i
         # If diffstat runs out of room it doesn't print anything,
         # which isn't very useful, so always print at least one + or -
         # if there were at least some changes.
         return max(i * graphwidth // maxtotal, int(bool(i)))
 
     for filename, adds, removes, isbinary in stats:
         if isbinary:
             count = 'Bin'
         else:
             count = '%d' % (adds + removes)
         pluses = '+' * scale(adds)
         minuses = '-' * scale(removes)
         output.append(' %s%s |  %*s %s%s\n' %
                       (filename, ' ' * (maxname - encoding.colwidth(filename)),
                        countwidth, count, pluses, minuses))
 
     if stats:
         output.append(_(' %d files changed, %d insertions(+), '
                         '%d deletions(-)\n')
                       % (len(stats), totaladds, totalremoves))
 
     return ''.join(output)
 
 def diffstatui(*args, **kw):
     '''like diffstat(), but yields 2-tuples of (output, label) for
     ui.write()
     '''
 
     for line in diffstat(*args, **kw).splitlines():
         if line and line[-1] in '+-':
             name, graph = line.rsplit(' ', 1)
             yield (name + ' ', '')
             m = re.search(br'\++', graph)
             if m:
                 yield (m.group(0), 'diffstat.inserted')
             m = re.search(br'-+', graph)
             if m:
                 yield (m.group(0), 'diffstat.deleted')
         else:
             yield (line, '')
         yield ('\n', '')