manifest: get rid of manifest.readshallowfast

This removes manifest.readshallowfast and converts it's one user to use
manifestlog instead.

mercurial/changegroup.py

@@ -1,1043 +1,1044 b''
 # changegroup.py - Mercurial changegroup manipulation functions
 #
 #  Copyright 2006 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 os
 import struct
 import tempfile
 import weakref
 
 from .i18n import _
 from .node import (
     hex,
     nullrev,
     short,
 )
 
 from . import (
     branchmap,
     dagutil,
     discovery,
     error,
     mdiff,
     phases,
     util,
 )
 
 _CHANGEGROUPV1_DELTA_HEADER = "20s20s20s20s"
 _CHANGEGROUPV2_DELTA_HEADER = "20s20s20s20s20s"
 _CHANGEGROUPV3_DELTA_HEADER = ">20s20s20s20s20sH"
 
 def readexactly(stream, n):
     '''read n bytes from stream.read and abort if less was available'''
     s = stream.read(n)
     if len(s) < n:
         raise error.Abort(_("stream ended unexpectedly"
                            " (got %d bytes, expected %d)")
                           % (len(s), n))
     return s
 
 def getchunk(stream):
     """return the next chunk from stream as a string"""
     d = readexactly(stream, 4)
     l = struct.unpack(">l", d)[0]
     if l <= 4:
         if l:
             raise error.Abort(_("invalid chunk length %d") % l)
         return ""
     return readexactly(stream, l - 4)
 
 def chunkheader(length):
     """return a changegroup chunk header (string)"""
     return struct.pack(">l", length + 4)
 
 def closechunk():
     """return a changegroup chunk header (string) for a zero-length chunk"""
     return struct.pack(">l", 0)
 
 def combineresults(results):
     """logic to combine 0 or more addchangegroup results into one"""
     changedheads = 0
     result = 1
     for ret in results:
         # If any changegroup result is 0, return 0
         if ret == 0:
             result = 0
             break
         if ret < -1:
             changedheads += ret + 1
         elif ret > 1:
             changedheads += ret - 1
     if changedheads > 0:
         result = 1 + changedheads
     elif changedheads < 0:
         result = -1 + changedheads
     return result
 
 def writechunks(ui, chunks, filename, vfs=None):
     """Write chunks to a file and return its filename.
 
     The stream is assumed to be a bundle file.
     Existing files will not be overwritten.
     If no filename is specified, a temporary file is created.
     """
     fh = None
     cleanup = None
     try:
         if filename:
             if vfs:
                 fh = vfs.open(filename, "wb")
             else:
                 # Increase default buffer size because default is usually
                 # small (4k is common on Linux).
                 fh = open(filename, "wb", 131072)
         else:
             fd, filename = tempfile.mkstemp(prefix="hg-bundle-", suffix=".hg")
             fh = os.fdopen(fd, "wb")
         cleanup = filename
         for c in chunks:
             fh.write(c)
         cleanup = None
         return filename
     finally:
         if fh is not None:
             fh.close()
         if cleanup is not None:
             if filename and vfs:
                 vfs.unlink(cleanup)
             else:
                 os.unlink(cleanup)
 
 class cg1unpacker(object):
     """Unpacker for cg1 changegroup streams.
 
     A changegroup unpacker handles the framing of the revision data in
     the wire format. Most consumers will want to use the apply()
     method to add the changes from the changegroup to a repository.
 
     If you're forwarding a changegroup unmodified to another consumer,
     use getchunks(), which returns an iterator of changegroup
     chunks. This is mostly useful for cases where you need to know the
     data stream has ended by observing the end of the changegroup.
 
     deltachunk() is useful only if you're applying delta data. Most
     consumers should prefer apply() instead.
 
     A few other public methods exist. Those are used only for
     bundlerepo and some debug commands - their use is discouraged.
     """
     deltaheader = _CHANGEGROUPV1_DELTA_HEADER
     deltaheadersize = struct.calcsize(deltaheader)
     version = '01'
     _grouplistcount = 1 # One list of files after the manifests
 
     def __init__(self, fh, alg, extras=None):
         if alg == 'UN':
             alg = None # get more modern without breaking too much
         if not alg in util.decompressors:
             raise error.Abort(_('unknown stream compression type: %s')
                              % alg)
         if alg == 'BZ':
             alg = '_truncatedBZ'
         self._stream = util.decompressors[alg](fh)
         self._type = alg
         self.extras = extras or {}
         self.callback = None
 
     # These methods (compressed, read, seek, tell) all appear to only
     # be used by bundlerepo, but it's a little hard to tell.
     def compressed(self):
         return self._type is not None
     def read(self, l):
         return self._stream.read(l)
     def seek(self, pos):
         return self._stream.seek(pos)
     def tell(self):
         return self._stream.tell()
     def close(self):
         return self._stream.close()
 
     def _chunklength(self):
         d = readexactly(self._stream, 4)
         l = struct.unpack(">l", d)[0]
         if l <= 4:
             if l:
                 raise error.Abort(_("invalid chunk length %d") % l)
             return 0
         if self.callback:
             self.callback()
         return l - 4
 
     def changelogheader(self):
         """v10 does not have a changelog header chunk"""
         return {}
 
     def manifestheader(self):
         """v10 does not have a manifest header chunk"""
         return {}
 
     def filelogheader(self):
         """return the header of the filelogs chunk, v10 only has the filename"""
         l = self._chunklength()
         if not l:
             return {}
         fname = readexactly(self._stream, l)
         return {'filename': fname}
 
     def _deltaheader(self, headertuple, prevnode):
         node, p1, p2, cs = headertuple
         if prevnode is None:
             deltabase = p1
         else:
             deltabase = prevnode
         flags = 0
         return node, p1, p2, deltabase, cs, flags
 
     def deltachunk(self, prevnode):
         l = self._chunklength()
         if not l:
             return {}
         headerdata = readexactly(self._stream, self.deltaheadersize)
         header = struct.unpack(self.deltaheader, headerdata)
         delta = readexactly(self._stream, l - self.deltaheadersize)
         node, p1, p2, deltabase, cs, flags = self._deltaheader(header, prevnode)
         return {'node': node, 'p1': p1, 'p2': p2, 'cs': cs,
                 'deltabase': deltabase, 'delta': delta, 'flags': flags}
 
     def getchunks(self):
         """returns all the chunks contains in the bundle
 
         Used when you need to forward the binary stream to a file or another
         network API. To do so, it parse the changegroup data, otherwise it will
         block in case of sshrepo because it don't know the end of the stream.
         """
         # an empty chunkgroup is the end of the changegroup
         # a changegroup has at least 2 chunkgroups (changelog and manifest).
         # after that, changegroup versions 1 and 2 have a series of groups
         # with one group per file. changegroup 3 has a series of directory
         # manifests before the files.
         count = 0
         emptycount = 0
         while emptycount < self._grouplistcount:
             empty = True
             count += 1
             while True:
                 chunk = getchunk(self)
                 if not chunk:
                     if empty and count > 2:
                         emptycount += 1
                     break
                 empty = False
                 yield chunkheader(len(chunk))
                 pos = 0
                 while pos < len(chunk):
                     next = pos + 2**20
                     yield chunk[pos:next]
                     pos = next
             yield closechunk()
 
     def _unpackmanifests(self, repo, revmap, trp, prog, numchanges):
         # We know that we'll never have more manifests than we had
         # changesets.
         self.callback = prog(_('manifests'), numchanges)
         # no need to check for empty manifest group here:
         # if the result of the merge of 1 and 2 is the same in 3 and 4,
         # no new manifest will be created and the manifest group will
         # be empty during the pull
         self.manifestheader()
         repo.manifest.addgroup(self, revmap, trp)
         repo.ui.progress(_('manifests'), None)
         self.callback = None
 
     def apply(self, repo, srctype, url, emptyok=False,
               targetphase=phases.draft, expectedtotal=None):
         """Add the changegroup returned by source.read() to this repo.
         srctype is a string like 'push', 'pull', or 'unbundle'.  url is
         the URL of the repo where this changegroup is coming from.
 
         Return an integer summarizing the change to this repo:
         - nothing changed or no source: 0
         - more heads than before: 1+added heads (2..n)
         - fewer heads than before: -1-removed heads (-2..-n)
         - number of heads stays the same: 1
         """
         repo = repo.unfiltered()
         def csmap(x):
             repo.ui.debug("add changeset %s\n" % short(x))
             return len(cl)
 
         def revmap(x):
             return cl.rev(x)
 
         changesets = files = revisions = 0
 
         try:
             with repo.transaction("\n".join([srctype,
                                              util.hidepassword(url)])) as tr:
                 # The transaction could have been created before and already
                 # carries source information. In this case we use the top
                 # level data. We overwrite the argument because we need to use
                 # the top level value (if they exist) in this function.
                 srctype = tr.hookargs.setdefault('source', srctype)
                 url = tr.hookargs.setdefault('url', url)
                 repo.hook('prechangegroup', throw=True, **tr.hookargs)
 
                 # write changelog data to temp files so concurrent readers
                 # will not see an inconsistent view
                 cl = repo.changelog
                 cl.delayupdate(tr)
                 oldheads = cl.heads()
 
                 trp = weakref.proxy(tr)
                 # pull off the changeset group
                 repo.ui.status(_("adding changesets\n"))
                 clstart = len(cl)
                 class prog(object):
                     def __init__(self, step, total):
                         self._step = step
                         self._total = total
                         self._count = 1
                     def __call__(self):
                         repo.ui.progress(self._step, self._count,
                                          unit=_('chunks'), total=self._total)
                         self._count += 1
                 self.callback = prog(_('changesets'), expectedtotal)
 
                 efiles = set()
                 def onchangelog(cl, node):
                     efiles.update(cl.readfiles(node))
 
                 self.changelogheader()
                 srccontent = cl.addgroup(self, csmap, trp,
                                          addrevisioncb=onchangelog)
                 efiles = len(efiles)
 
                 if not (srccontent or emptyok):
                     raise error.Abort(_("received changelog group is empty"))
                 clend = len(cl)
                 changesets = clend - clstart
                 repo.ui.progress(_('changesets'), None)
                 self.callback = None
 
                 # pull off the manifest group
                 repo.ui.status(_("adding manifests\n"))
                 self._unpackmanifests(repo, revmap, trp, prog, changesets)
 
                 needfiles = {}
                 if repo.ui.configbool('server', 'validate', default=False):
                     cl = repo.changelog
                     ml = repo.manifestlog
                     # validate incoming csets have their manifests
                     for cset in xrange(clstart, clend):
                         mfnode = cl.changelogrevision(cset).manifest
                         mfest = ml[mfnode].readdelta()
                         # store file nodes we must see
                         for f, n in mfest.iteritems():
                             needfiles.setdefault(f, set()).add(n)
 
                 # process the files
                 repo.ui.status(_("adding file changes\n"))
                 newrevs, newfiles = _addchangegroupfiles(
                     repo, self, revmap, trp, efiles, needfiles)
                 revisions += newrevs
                 files += newfiles
 
                 dh = 0
                 if oldheads:
                     heads = cl.heads()
                     dh = len(heads) - len(oldheads)
                     for h in heads:
                         if h not in oldheads and repo[h].closesbranch():
                             dh -= 1
                 htext = ""
                 if dh:
                     htext = _(" (%+d heads)") % dh
 
                 repo.ui.status(_("added %d changesets"
                                  " with %d changes to %d files%s\n")
                                  % (changesets, revisions, files, htext))
                 repo.invalidatevolatilesets()
 
                 if changesets > 0:
                     if 'node' not in tr.hookargs:
                         tr.hookargs['node'] = hex(cl.node(clstart))
                         tr.hookargs['node_last'] = hex(cl.node(clend - 1))
                         hookargs = dict(tr.hookargs)
                     else:
                         hookargs = dict(tr.hookargs)
                         hookargs['node'] = hex(cl.node(clstart))
                         hookargs['node_last'] = hex(cl.node(clend - 1))
                     repo.hook('pretxnchangegroup', throw=True, **hookargs)
 
                 added = [cl.node(r) for r in xrange(clstart, clend)]
                 publishing = repo.publishing()
                 if srctype in ('push', 'serve'):
                     # Old servers can not push the boundary themselves.
                     # New servers won't push the boundary if changeset already
                     # exists locally as secret
                     #
                     # We should not use added here but the list of all change in
                     # the bundle
                     if publishing:
                         phases.advanceboundary(repo, tr, phases.public,
                                                srccontent)
                     else:
                         # Those changesets have been pushed from the
                         # outside, their phases are going to be pushed
                         # alongside. Therefor `targetphase` is
                         # ignored.
                         phases.advanceboundary(repo, tr, phases.draft,
                                                srccontent)
                         phases.retractboundary(repo, tr, phases.draft, added)
                 elif srctype != 'strip':
                     # publishing only alter behavior during push
                     #
                     # strip should not touch boundary at all
                     phases.retractboundary(repo, tr, targetphase, added)
 
                 if changesets > 0:
                     if srctype != 'strip':
                         # During strip, branchcache is invalid but
                         # coming call to `destroyed` will repair it.
                         # In other case we can safely update cache on
                         # disk.
                         repo.ui.debug('updating the branch cache\n')
                         branchmap.updatecache(repo.filtered('served'))
 
                     def runhooks():
                         # These hooks run when the lock releases, not when the
                         # transaction closes. So it's possible for the changelog
                         # to have changed since we last saw it.
                         if clstart >= len(repo):
                             return
 
                         repo.hook("changegroup", **hookargs)
 
                         for n in added:
                             args = hookargs.copy()
                             args['node'] = hex(n)
                             del args['node_last']
                             repo.hook("incoming", **args)
 
                         newheads = [h for h in repo.heads()
                                     if h not in oldheads]
                         repo.ui.log("incoming",
                                     "%s incoming changes - new heads: %s\n",
                                     len(added),
                                     ', '.join([hex(c[:6]) for c in newheads]))
 
                     tr.addpostclose('changegroup-runhooks-%020i' % clstart,
                                     lambda tr: repo._afterlock(runhooks))
         finally:
             repo.ui.flush()
         # never return 0 here:
         if dh < 0:
             return dh - 1
         else:
             return dh + 1
 
 class cg2unpacker(cg1unpacker):
     """Unpacker for cg2 streams.
 
     cg2 streams add support for generaldelta, so the delta header
     format is slightly different. All other features about the data
     remain the same.
     """
     deltaheader = _CHANGEGROUPV2_DELTA_HEADER
     deltaheadersize = struct.calcsize(deltaheader)
     version = '02'
 
     def _deltaheader(self, headertuple, prevnode):
         node, p1, p2, deltabase, cs = headertuple
         flags = 0
         return node, p1, p2, deltabase, cs, flags
 
 class cg3unpacker(cg2unpacker):
     """Unpacker for cg3 streams.
 
     cg3 streams add support for exchanging treemanifests and revlog
     flags. It adds the revlog flags to the delta header and an empty chunk
     separating manifests and files.
     """
     deltaheader = _CHANGEGROUPV3_DELTA_HEADER
     deltaheadersize = struct.calcsize(deltaheader)
     version = '03'
     _grouplistcount = 2 # One list of manifests and one list of files
 
     def _deltaheader(self, headertuple, prevnode):
         node, p1, p2, deltabase, cs, flags = headertuple
         return node, p1, p2, deltabase, cs, flags
 
     def _unpackmanifests(self, repo, revmap, trp, prog, numchanges):
         super(cg3unpacker, self)._unpackmanifests(repo, revmap, trp, prog,
                                                   numchanges)
         for chunkdata in iter(self.filelogheader, {}):
             # If we get here, there are directory manifests in the changegroup
             d = chunkdata["filename"]
             repo.ui.debug("adding %s revisions\n" % d)
             dirlog = repo.manifest.dirlog(d)
             if not dirlog.addgroup(self, revmap, trp):
                 raise error.Abort(_("received dir revlog group is empty"))
 
 class headerlessfixup(object):
     def __init__(self, fh, h):
         self._h = h
         self._fh = fh
     def read(self, n):
         if self._h:
             d, self._h = self._h[:n], self._h[n:]
             if len(d) < n:
                 d += readexactly(self._fh, n - len(d))
             return d
         return readexactly(self._fh, n)
 
 class cg1packer(object):
     deltaheader = _CHANGEGROUPV1_DELTA_HEADER
     version = '01'
     def __init__(self, repo, bundlecaps=None):
         """Given a source repo, construct a bundler.
 
         bundlecaps is optional and can be used to specify the set of
         capabilities which can be used to build the bundle.
         """
         # Set of capabilities we can use to build the bundle.
         if bundlecaps is None:
             bundlecaps = set()
         self._bundlecaps = bundlecaps
         # experimental config: bundle.reorder
         reorder = repo.ui.config('bundle', 'reorder', 'auto')
         if reorder == 'auto':
             reorder = None
         else:
             reorder = util.parsebool(reorder)
         self._repo = repo
         self._reorder = reorder
         self._progress = repo.ui.progress
         if self._repo.ui.verbose and not self._repo.ui.debugflag:
             self._verbosenote = self._repo.ui.note
         else:
             self._verbosenote = lambda s: None
 
     def close(self):
         return closechunk()
 
     def fileheader(self, fname):
         return chunkheader(len(fname)) + fname
 
     # Extracted both for clarity and for overriding in extensions.
     def _sortgroup(self, revlog, nodelist, lookup):
         """Sort nodes for change group and turn them into revnums."""
         # for generaldelta revlogs, we linearize the revs; this will both be
         # much quicker and generate a much smaller bundle
         if (revlog._generaldelta and self._reorder is None) or self._reorder:
             dag = dagutil.revlogdag(revlog)
             return dag.linearize(set(revlog.rev(n) for n in nodelist))
         else:
             return sorted([revlog.rev(n) for n in nodelist])
 
     def group(self, nodelist, revlog, lookup, units=None):
         """Calculate a delta group, yielding a sequence of changegroup chunks
         (strings).
 
         Given a list of changeset revs, return a set of deltas and
         metadata corresponding to nodes. The first delta is
         first parent(nodelist[0]) -> nodelist[0], the receiver is
         guaranteed to have this parent as it has all history before
         these changesets. In the case firstparent is nullrev the
         changegroup starts with a full revision.
 
         If units is not None, progress detail will be generated, units specifies
         the type of revlog that is touched (changelog, manifest, etc.).
         """
         # if we don't have any revisions touched by these changesets, bail
         if len(nodelist) == 0:
             yield self.close()
             return
 
         revs = self._sortgroup(revlog, nodelist, lookup)
 
         # add the parent of the first rev
         p = revlog.parentrevs(revs[0])[0]
         revs.insert(0, p)
 
         # build deltas
         total = len(revs) - 1
         msgbundling = _('bundling')
         for r in xrange(len(revs) - 1):
             if units is not None:
                 self._progress(msgbundling, r + 1, unit=units, total=total)
             prev, curr = revs[r], revs[r + 1]
             linknode = lookup(revlog.node(curr))
             for c in self.revchunk(revlog, curr, prev, linknode):
                 yield c
 
         if units is not None:
             self._progress(msgbundling, None)
         yield self.close()
 
     # filter any nodes that claim to be part of the known set
     def prune(self, revlog, missing, commonrevs):
         rr, rl = revlog.rev, revlog.linkrev
         return [n for n in missing if rl(rr(n)) not in commonrevs]
 
     def _packmanifests(self, dir, mfnodes, lookuplinknode):
         """Pack flat manifests into a changegroup stream."""
         assert not dir
         for chunk in self.group(mfnodes, self._repo.manifest,
                                 lookuplinknode, units=_('manifests')):
             yield chunk
 
     def _manifestsdone(self):
         return ''
 
     def generate(self, commonrevs, clnodes, fastpathlinkrev, source):
         '''yield a sequence of changegroup chunks (strings)'''
         repo = self._repo
         cl = repo.changelog
 
         clrevorder = {}
         mfs = {} # needed manifests
         fnodes = {} # needed file nodes
         changedfiles = set()
 
         # Callback for the changelog, used to collect changed files and manifest
         # nodes.
         # Returns the linkrev node (identity in the changelog case).
         def lookupcl(x):
             c = cl.read(x)
             clrevorder[x] = len(clrevorder)
             n = c[0]
             # record the first changeset introducing this manifest version
             mfs.setdefault(n, x)
             # Record a complete list of potentially-changed files in
             # this manifest.
             changedfiles.update(c[3])
             return x
 
         self._verbosenote(_('uncompressed size of bundle content:\n'))
         size = 0
         for chunk in self.group(clnodes, cl, lookupcl, units=_('changesets')):
             size += len(chunk)
             yield chunk
         self._verbosenote(_('%8.i (changelog)\n') % size)
 
         # We need to make sure that the linkrev in the changegroup refers to
         # the first changeset that introduced the manifest or file revision.
         # The fastpath is usually safer than the slowpath, because the filelogs
         # are walked in revlog order.
         #
         # When taking the slowpath with reorder=None and the manifest revlog
         # uses generaldelta, the manifest may be walked in the "wrong" order.
         # Without 'clrevorder', we would get an incorrect linkrev (see fix in
         # cc0ff93d0c0c).
         #
         # When taking the fastpath, we are only vulnerable to reordering
         # of the changelog itself. The changelog never uses generaldelta, so
         # it is only reordered when reorder=True. To handle this case, we
         # simply take the slowpath, which already has the 'clrevorder' logic.
         # This was also fixed in cc0ff93d0c0c.
         fastpathlinkrev = fastpathlinkrev and not self._reorder
         # Treemanifests don't work correctly with fastpathlinkrev
         # either, because we don't discover which directory nodes to
         # send along with files. This could probably be fixed.
         fastpathlinkrev = fastpathlinkrev and (
             'treemanifest' not in repo.requirements)
 
         for chunk in self.generatemanifests(commonrevs, clrevorder,
                 fastpathlinkrev, mfs, fnodes):
             yield chunk
         mfs.clear()
         clrevs = set(cl.rev(x) for x in clnodes)
 
         if not fastpathlinkrev:
             def linknodes(unused, fname):
                 return fnodes.get(fname, {})
         else:
             cln = cl.node
             def linknodes(filerevlog, fname):
                 llr = filerevlog.linkrev
                 fln = filerevlog.node
                 revs = ((r, llr(r)) for r in filerevlog)
                 return dict((fln(r), cln(lr)) for r, lr in revs if lr in clrevs)
 
         for chunk in self.generatefiles(changedfiles, linknodes, commonrevs,
                                         source):
             yield chunk
 
         yield self.close()
 
         if clnodes:
             repo.hook('outgoing', node=hex(clnodes[0]), source=source)
 
     def generatemanifests(self, commonrevs, clrevorder, fastpathlinkrev, mfs,
                           fnodes):
         repo = self._repo
-        dirlog = repo.manifest.dirlog
+        mfl = repo.manifestlog
+        dirlog = mfl._revlog.dirlog
         tmfnodes = {'': mfs}
 
         # Callback for the manifest, used to collect linkrevs for filelog
         # revisions.
         # Returns the linkrev node (collected in lookupcl).
         def makelookupmflinknode(dir):
             if fastpathlinkrev:
                 assert not dir
                 return mfs.__getitem__
 
             def lookupmflinknode(x):
                 """Callback for looking up the linknode for manifests.
 
                 Returns the linkrev node for the specified manifest.
 
                 SIDE EFFECT:
 
                 1) fclnodes gets populated with the list of relevant
                    file nodes if we're not using fastpathlinkrev
                 2) When treemanifests are in use, collects treemanifest nodes
                    to send
 
                 Note that this means manifests must be completely sent to
                 the client before you can trust the list of files and
                 treemanifests to send.
                 """
                 clnode = tmfnodes[dir][x]
-                mdata = dirlog(dir).readshallowfast(x)
+                mdata = mfl.get(dir, x).readfast(shallow=True)
                 for p, n, fl in mdata.iterentries():
                     if fl == 't': # subdirectory manifest
                         subdir = dir + p + '/'
                         tmfclnodes = tmfnodes.setdefault(subdir, {})
                         tmfclnode = tmfclnodes.setdefault(n, clnode)
                         if clrevorder[clnode] < clrevorder[tmfclnode]:
                             tmfclnodes[n] = clnode
                     else:
                         f = dir + p
                         fclnodes = fnodes.setdefault(f, {})
                         fclnode = fclnodes.setdefault(n, clnode)
                         if clrevorder[clnode] < clrevorder[fclnode]:
                             fclnodes[n] = clnode
                 return clnode
             return lookupmflinknode
 
         size = 0
         while tmfnodes:
             dir = min(tmfnodes)
             nodes = tmfnodes[dir]
             prunednodes = self.prune(dirlog(dir), nodes, commonrevs)
             if not dir or prunednodes:
                 for x in self._packmanifests(dir, prunednodes,
                                              makelookupmflinknode(dir)):
                     size += len(x)
                     yield x
             del tmfnodes[dir]
         self._verbosenote(_('%8.i (manifests)\n') % size)
         yield self._manifestsdone()
 
     # The 'source' parameter is useful for extensions
     def generatefiles(self, changedfiles, linknodes, commonrevs, source):
         repo = self._repo
         progress = self._progress
         msgbundling = _('bundling')
 
         total = len(changedfiles)
         # for progress output
         msgfiles = _('files')
         for i, fname in enumerate(sorted(changedfiles)):
             filerevlog = repo.file(fname)
             if not filerevlog:
                 raise error.Abort(_("empty or missing revlog for %s") % fname)
 
             linkrevnodes = linknodes(filerevlog, fname)
             # Lookup for filenodes, we collected the linkrev nodes above in the
             # fastpath case and with lookupmf in the slowpath case.
             def lookupfilelog(x):
                 return linkrevnodes[x]
 
             filenodes = self.prune(filerevlog, linkrevnodes, commonrevs)
             if filenodes:
                 progress(msgbundling, i + 1, item=fname, unit=msgfiles,
                          total=total)
                 h = self.fileheader(fname)
                 size = len(h)
                 yield h
                 for chunk in self.group(filenodes, filerevlog, lookupfilelog):
                     size += len(chunk)
                     yield chunk
                 self._verbosenote(_('%8.i  %s\n') % (size, fname))
         progress(msgbundling, None)
 
     def deltaparent(self, revlog, rev, p1, p2, prev):
         return prev
 
     def revchunk(self, revlog, rev, prev, linknode):
         node = revlog.node(rev)
         p1, p2 = revlog.parentrevs(rev)
         base = self.deltaparent(revlog, rev, p1, p2, prev)
 
         prefix = ''
         if revlog.iscensored(base) or revlog.iscensored(rev):
             try:
                 delta = revlog.revision(node)
             except error.CensoredNodeError as e:
                 delta = e.tombstone
             if base == nullrev:
                 prefix = mdiff.trivialdiffheader(len(delta))
             else:
                 baselen = revlog.rawsize(base)
                 prefix = mdiff.replacediffheader(baselen, len(delta))
         elif base == nullrev:
             delta = revlog.revision(node)
             prefix = mdiff.trivialdiffheader(len(delta))
         else:
             delta = revlog.revdiff(base, rev)
         p1n, p2n = revlog.parents(node)
         basenode = revlog.node(base)
         flags = revlog.flags(rev)
         meta = self.builddeltaheader(node, p1n, p2n, basenode, linknode, flags)
         meta += prefix
         l = len(meta) + len(delta)
         yield chunkheader(l)
         yield meta
         yield delta
     def builddeltaheader(self, node, p1n, p2n, basenode, linknode, flags):
         # do nothing with basenode, it is implicitly the previous one in HG10
         # do nothing with flags, it is implicitly 0 for cg1 and cg2
         return struct.pack(self.deltaheader, node, p1n, p2n, linknode)
 
 class cg2packer(cg1packer):
     version = '02'
     deltaheader = _CHANGEGROUPV2_DELTA_HEADER
 
     def __init__(self, repo, bundlecaps=None):
         super(cg2packer, self).__init__(repo, bundlecaps)
         if self._reorder is None:
             # Since generaldelta is directly supported by cg2, reordering
             # generally doesn't help, so we disable it by default (treating
             # bundle.reorder=auto just like bundle.reorder=False).
             self._reorder = False
 
     def deltaparent(self, revlog, rev, p1, p2, prev):
         dp = revlog.deltaparent(rev)
         if dp == nullrev and revlog.storedeltachains:
             # Avoid sending full revisions when delta parent is null. Pick prev
             # in that case. It's tempting to pick p1 in this case, as p1 will
             # be smaller in the common case. However, computing a delta against
             # p1 may require resolving the raw text of p1, which could be
             # expensive. The revlog caches should have prev cached, meaning
             # less CPU for changegroup generation. There is likely room to add
             # a flag and/or config option to control this behavior.
             return prev
         elif dp == nullrev:
             # revlog is configured to use full snapshot for a reason,
             # stick to full snapshot.
             return nullrev
         elif dp not in (p1, p2, prev):
             # Pick prev when we can't be sure remote has the base revision.
             return prev
         else:
             return dp
 
     def builddeltaheader(self, node, p1n, p2n, basenode, linknode, flags):
         # Do nothing with flags, it is implicitly 0 in cg1 and cg2
         return struct.pack(self.deltaheader, node, p1n, p2n, basenode, linknode)
 
 class cg3packer(cg2packer):
     version = '03'
     deltaheader = _CHANGEGROUPV3_DELTA_HEADER
 
     def _packmanifests(self, dir, mfnodes, lookuplinknode):
         if dir:
             yield self.fileheader(dir)
         for chunk in self.group(mfnodes, self._repo.manifest.dirlog(dir),
                                 lookuplinknode, units=_('manifests')):
             yield chunk
 
     def _manifestsdone(self):
         return self.close()
 
     def builddeltaheader(self, node, p1n, p2n, basenode, linknode, flags):
         return struct.pack(
             self.deltaheader, node, p1n, p2n, basenode, linknode, flags)
 
 _packermap = {'01': (cg1packer, cg1unpacker),
              # cg2 adds support for exchanging generaldelta
              '02': (cg2packer, cg2unpacker),
              # cg3 adds support for exchanging revlog flags and treemanifests
              '03': (cg3packer, cg3unpacker),
 }
 
 def allsupportedversions(ui):
     versions = set(_packermap.keys())
     versions.discard('03')
     if (ui.configbool('experimental', 'changegroup3') or
         ui.configbool('experimental', 'treemanifest')):
         versions.add('03')
     return versions
 
 # Changegroup versions that can be applied to the repo
 def supportedincomingversions(repo):
     versions = allsupportedversions(repo.ui)
     if 'treemanifest' in repo.requirements:
         versions.add('03')
     return versions
 
 # Changegroup versions that can be created from the repo
 def supportedoutgoingversions(repo):
     versions = allsupportedversions(repo.ui)
     if 'treemanifest' in repo.requirements:
         # Versions 01 and 02 support only flat manifests and it's just too
         # expensive to convert between the flat manifest and tree manifest on
         # the fly. Since tree manifests are hashed differently, all of history
         # would have to be converted. Instead, we simply don't even pretend to
         # support versions 01 and 02.
         versions.discard('01')
         versions.discard('02')
         versions.add('03')
     return versions
 
 def safeversion(repo):
     # Finds the smallest version that it's safe to assume clients of the repo
     # will support. For example, all hg versions that support generaldelta also
     # support changegroup 02.
     versions = supportedoutgoingversions(repo)
     if 'generaldelta' in repo.requirements:
         versions.discard('01')
     assert versions
     return min(versions)
 
 def getbundler(version, repo, bundlecaps=None):
     assert version in supportedoutgoingversions(repo)
     return _packermap[version][0](repo, bundlecaps)
 
 def getunbundler(version, fh, alg, extras=None):
     return _packermap[version][1](fh, alg, extras=extras)
 
 def _changegroupinfo(repo, nodes, source):
     if repo.ui.verbose or source == 'bundle':
         repo.ui.status(_("%d changesets found\n") % len(nodes))
     if repo.ui.debugflag:
         repo.ui.debug("list of changesets:\n")
         for node in nodes:
             repo.ui.debug("%s\n" % hex(node))
 
 def getsubsetraw(repo, outgoing, bundler, source, fastpath=False):
     repo = repo.unfiltered()
     commonrevs = outgoing.common
     csets = outgoing.missing
     heads = outgoing.missingheads
     # We go through the fast path if we get told to, or if all (unfiltered
     # heads have been requested (since we then know there all linkrevs will
     # be pulled by the client).
     heads.sort()
     fastpathlinkrev = fastpath or (
             repo.filtername is None and heads == sorted(repo.heads()))
 
     repo.hook('preoutgoing', throw=True, source=source)
     _changegroupinfo(repo, csets, source)
     return bundler.generate(commonrevs, csets, fastpathlinkrev, source)
 
 def getsubset(repo, outgoing, bundler, source, fastpath=False):
     gengroup = getsubsetraw(repo, outgoing, bundler, source, fastpath)
     return getunbundler(bundler.version, util.chunkbuffer(gengroup), None,
                         {'clcount': len(outgoing.missing)})
 
 def changegroupsubset(repo, roots, heads, source, version='01'):
     """Compute a changegroup consisting of all the nodes that are
     descendants of any of the roots and ancestors of any of the heads.
     Return a chunkbuffer object whose read() method will return
     successive changegroup chunks.
 
     It is fairly complex as determining which filenodes and which
     manifest nodes need to be included for the changeset to be complete
     is non-trivial.
 
     Another wrinkle is doing the reverse, figuring out which changeset in
     the changegroup a particular filenode or manifestnode belongs to.
     """
     outgoing = discovery.outgoing(repo, missingroots=roots, missingheads=heads)
     bundler = getbundler(version, repo)
     return getsubset(repo, outgoing, bundler, source)
 
 def getlocalchangegroupraw(repo, source, outgoing, bundlecaps=None,
                            version='01'):
     """Like getbundle, but taking a discovery.outgoing as an argument.
 
     This is only implemented for local repos and reuses potentially
     precomputed sets in outgoing. Returns a raw changegroup generator."""
     if not outgoing.missing:
         return None
     bundler = getbundler(version, repo, bundlecaps)
     return getsubsetraw(repo, outgoing, bundler, source)
 
 def getlocalchangegroup(repo, source, outgoing, bundlecaps=None,
                         version='01'):
     """Like getbundle, but taking a discovery.outgoing as an argument.
 
     This is only implemented for local repos and reuses potentially
     precomputed sets in outgoing."""
     if not outgoing.missing:
         return None
     bundler = getbundler(version, repo, bundlecaps)
     return getsubset(repo, outgoing, bundler, source)
 
 def getchangegroup(repo, source, outgoing, bundlecaps=None,
                    version='01'):
     """Like changegroupsubset, but returns the set difference between the
     ancestors of heads and the ancestors common.
 
     If heads is None, use the local heads. If common is None, use [nullid].
 
     The nodes in common might not all be known locally due to the way the
     current discovery protocol works.
     """
     return getlocalchangegroup(repo, source, outgoing, bundlecaps=bundlecaps,
                                version=version)
 
 def changegroup(repo, basenodes, source):
     # to avoid a race we use changegroupsubset() (issue1320)
     return changegroupsubset(repo, basenodes, repo.heads(), source)
 
 def _addchangegroupfiles(repo, source, revmap, trp, expectedfiles, needfiles):
     revisions = 0
     files = 0
     for chunkdata in iter(source.filelogheader, {}):
         files += 1
         f = chunkdata["filename"]
         repo.ui.debug("adding %s revisions\n" % f)
         repo.ui.progress(_('files'), files, unit=_('files'),
                          total=expectedfiles)
         fl = repo.file(f)
         o = len(fl)
         try:
             if not fl.addgroup(source, revmap, trp):
                 raise error.Abort(_("received file revlog group is empty"))
         except error.CensoredBaseError as e:
             raise error.Abort(_("received delta base is censored: %s") % e)
         revisions += len(fl) - o
         if f in needfiles:
             needs = needfiles[f]
             for new in xrange(o, len(fl)):
                 n = fl.node(new)
                 if n in needs:
                     needs.remove(n)
                 else:
                     raise error.Abort(
                         _("received spurious file revlog entry"))
             if not needs:
                 del needfiles[f]
     repo.ui.progress(_('files'), None)
 
     for f, needs in needfiles.iteritems():
         fl = repo.file(f)
         for n in needs:
             try:
                 fl.rev(n)
             except error.LookupError:
                 raise error.Abort(
                     _('missing file data for %s:%s - run hg verify') %
                     (f, hex(n)))
 
     return revisions, files

mercurial/manifest.py

@@ -1,1579 +1,1576 b''
 # manifest.py - manifest revision class for mercurial
 #
 # 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 array
 import heapq
 import os
 import struct
 
 from .i18n import _
 from . import (
     error,
     mdiff,
     parsers,
     revlog,
     util,
 )
 
 propertycache = util.propertycache
 
 def _parsev1(data):
     # This method does a little bit of excessive-looking
     # precondition checking. This is so that the behavior of this
     # class exactly matches its C counterpart to try and help
     # prevent surprise breakage for anyone that develops against
     # the pure version.
     if data and data[-1] != '\n':
         raise ValueError('Manifest did not end in a newline.')
     prev = None
     for l in data.splitlines():
         if prev is not None and prev > l:
             raise ValueError('Manifest lines not in sorted order.')
         prev = l
         f, n = l.split('\0')
         if len(n) > 40:
             yield f, revlog.bin(n[:40]), n[40:]
         else:
             yield f, revlog.bin(n), ''
 
 def _parsev2(data):
     metadataend = data.find('\n')
     # Just ignore metadata for now
     pos = metadataend + 1
     prevf = ''
     while pos < len(data):
         end = data.find('\n', pos + 1) # +1 to skip stem length byte
         if end == -1:
             raise ValueError('Manifest ended with incomplete file entry.')
         stemlen = ord(data[pos])
         items = data[pos + 1:end].split('\0')
         f = prevf[:stemlen] + items[0]
         if prevf > f:
             raise ValueError('Manifest entries not in sorted order.')
         fl = items[1]
         # Just ignore metadata (items[2:] for now)
         n = data[end + 1:end + 21]
         yield f, n, fl
         pos = end + 22
         prevf = f
 
 def _parse(data):
     """Generates (path, node, flags) tuples from a manifest text"""
     if data.startswith('\0'):
         return iter(_parsev2(data))
     else:
         return iter(_parsev1(data))
 
 def _text(it, usemanifestv2):
     """Given an iterator over (path, node, flags) tuples, returns a manifest
     text"""
     if usemanifestv2:
         return _textv2(it)
     else:
         return _textv1(it)
 
 def _textv1(it):
     files = []
     lines = []
     _hex = revlog.hex
     for f, n, fl in it:
         files.append(f)
         # if this is changed to support newlines in filenames,
         # be sure to check the templates/ dir again (especially *-raw.tmpl)
         lines.append("%s\0%s%s\n" % (f, _hex(n), fl))
 
     _checkforbidden(files)
     return ''.join(lines)
 
 def _textv2(it):
     files = []
     lines = ['\0\n']
     prevf = ''
     for f, n, fl in it:
         files.append(f)
         stem = os.path.commonprefix([prevf, f])
         stemlen = min(len(stem), 255)
         lines.append("%c%s\0%s\n%s\n" % (stemlen, f[stemlen:], fl, n))
         prevf = f
     _checkforbidden(files)
     return ''.join(lines)
 
 class lazymanifestiter(object):
     def __init__(self, lm):
         self.pos = 0
         self.lm = lm
 
     def __iter__(self):
         return self
 
     def next(self):
         try:
             data, pos = self.lm._get(self.pos)
         except IndexError:
             raise StopIteration
         if pos == -1:
             self.pos += 1
             return data[0]
         self.pos += 1
         zeropos = data.find('\x00', pos)
         return data[pos:zeropos]
 
 class lazymanifestiterentries(object):
     def __init__(self, lm):
         self.lm = lm
         self.pos = 0
 
     def __iter__(self):
         return self
 
     def next(self):
         try:
             data, pos = self.lm._get(self.pos)
         except IndexError:
             raise StopIteration
         if pos == -1:
             self.pos += 1
             return data
         zeropos = data.find('\x00', pos)
         hashval = unhexlify(data, self.lm.extrainfo[self.pos],
                             zeropos + 1, 40)
         flags = self.lm._getflags(data, self.pos, zeropos)
         self.pos += 1
         return (data[pos:zeropos], hashval, flags)
 
 def unhexlify(data, extra, pos, length):
     s = data[pos:pos + length].decode('hex')
     if extra:
         s += chr(extra & 0xff)
     return s
 
 def _cmp(a, b):
     return (a > b) - (a < b)
 
 class _lazymanifest(object):
     def __init__(self, data, positions=None, extrainfo=None, extradata=None):
         if positions is None:
             self.positions = self.findlines(data)
             self.extrainfo = [0] * len(self.positions)
             self.data = data
             self.extradata = []
         else:
             self.positions = positions[:]
             self.extrainfo = extrainfo[:]
             self.extradata = extradata[:]
             self.data = data
 
     def findlines(self, data):
         if not data:
             return []
         pos = data.find("\n")
         if pos == -1 or data[-1] != '\n':
             raise ValueError("Manifest did not end in a newline.")
         positions = [0]
         prev = data[:data.find('\x00')]
         while pos < len(data) - 1 and pos != -1:
             positions.append(pos + 1)
             nexts = data[pos + 1:data.find('\x00', pos + 1)]
             if nexts < prev:
                 raise ValueError("Manifest lines not in sorted order.")
             prev = nexts
             pos = data.find("\n", pos + 1)
         return positions
 
     def _get(self, index):
         # get the position encoded in pos:
         #   positive number is an index in 'data'
         #   negative number is in extrapieces
         pos = self.positions[index]
         if pos >= 0:
             return self.data, pos
         return self.extradata[-pos - 1], -1
 
     def _getkey(self, pos):
         if pos >= 0:
             return self.data[pos:self.data.find('\x00', pos + 1)]
         return self.extradata[-pos - 1][0]
 
     def bsearch(self, key):
         first = 0
         last = len(self.positions) - 1
 
         while first <= last:
             midpoint = (first + last)//2
             nextpos = self.positions[midpoint]
             candidate = self._getkey(nextpos)
             r = _cmp(key, candidate)
             if r == 0:
                 return midpoint
             else:
                 if r < 0:
                     last = midpoint - 1
                 else:
                     first = midpoint + 1
         return -1
 
     def bsearch2(self, key):
         # same as the above, but will always return the position
         # done for performance reasons
         first = 0
         last = len(self.positions) - 1
 
         while first <= last:
             midpoint = (first + last)//2
             nextpos = self.positions[midpoint]
             candidate = self._getkey(nextpos)
             r = _cmp(key, candidate)
             if r == 0:
                 return (midpoint, True)
             else:
                 if r < 0:
                     last = midpoint - 1
                 else:
                     first = midpoint + 1
         return (first, False)
 
     def __contains__(self, key):
         return self.bsearch(key) != -1
 
     def _getflags(self, data, needle, pos):
         start = pos + 41
         end = data.find("\n", start)
         if end == -1:
             end = len(data) - 1
         if start == end:
             return ''
         return self.data[start:end]
 
     def __getitem__(self, key):
         if not isinstance(key, str):
             raise TypeError("getitem: manifest keys must be a string.")
         needle = self.bsearch(key)
         if needle == -1:
             raise KeyError
         data, pos = self._get(needle)
         if pos == -1:
             return (data[1], data[2])
         zeropos = data.find('\x00', pos)
         assert 0 <= needle <= len(self.positions)
         assert len(self.extrainfo) == len(self.positions)
         hashval = unhexlify(data, self.extrainfo[needle], zeropos + 1, 40)
         flags = self._getflags(data, needle, zeropos)
         return (hashval, flags)
 
     def __delitem__(self, key):
         needle, found = self.bsearch2(key)
         if not found:
             raise KeyError
         cur = self.positions[needle]
         self.positions = self.positions[:needle] + self.positions[needle + 1:]
         self.extrainfo = self.extrainfo[:needle] + self.extrainfo[needle + 1:]
         if cur >= 0:
             self.data = self.data[:cur] + '\x00' + self.data[cur + 1:]
 
     def __setitem__(self, key, value):
         if not isinstance(key, str):
             raise TypeError("setitem: manifest keys must be a string.")
         if not isinstance(value, tuple) or len(value) != 2:
             raise TypeError("Manifest values must be a tuple of (node, flags).")
         hashval = value[0]
         if not isinstance(hashval, str) or not 20 <= len(hashval) <= 22:
             raise TypeError("node must be a 20-byte string")
         flags = value[1]
         if len(hashval) == 22:
             hashval = hashval[:-1]
         if not isinstance(flags, str) or len(flags) > 1:
             raise TypeError("flags must a 0 or 1 byte string, got %r", flags)
         needle, found = self.bsearch2(key)
         if found:
             # put the item
             pos = self.positions[needle]
             if pos < 0:
                 self.extradata[-pos - 1] = (key, hashval, value[1])
             else:
                 # just don't bother
                 self.extradata.append((key, hashval, value[1]))
                 self.positions[needle] = -len(self.extradata)
         else:
             # not found, put it in with extra positions
             self.extradata.append((key, hashval, value[1]))
             self.positions = (self.positions[:needle] + [-len(self.extradata)]
                               + self.positions[needle:])
             self.extrainfo = (self.extrainfo[:needle] + [0] +
                               self.extrainfo[needle:])
 
     def copy(self):
         # XXX call _compact like in C?
         return _lazymanifest(self.data, self.positions, self.extrainfo,
             self.extradata)
 
     def _compact(self):
         # hopefully not called TOO often
         if len(self.extradata) == 0:
             return
         l = []
         last_cut = 0
         i = 0
         offset = 0
         self.extrainfo = [0] * len(self.positions)
         while i < len(self.positions):
             if self.positions[i] >= 0:
                 cur = self.positions[i]
                 last_cut = cur
                 while True:
                     self.positions[i] = offset
                     i += 1
                     if i == len(self.positions) or self.positions[i] < 0:
                         break
                     offset += self.positions[i] - cur
                     cur = self.positions[i]
                 end_cut = self.data.find('\n', cur)
                 if end_cut != -1:
                     end_cut += 1
                 offset += end_cut - cur
                 l.append(self.data[last_cut:end_cut])
             else:
                 while i < len(self.positions) and self.positions[i] < 0:
                     cur = self.positions[i]
                     t = self.extradata[-cur - 1]
                     l.append(self._pack(t))
                     self.positions[i] = offset
                     if len(t[1]) > 20:
                         self.extrainfo[i] = ord(t[1][21])
                     offset += len(l[-1])
                     i += 1
         self.data = ''.join(l)
         self.extradata = []
 
     def _pack(self, d):
         return d[0] + '\x00' + d[1][:20].encode('hex') + d[2] + '\n'
 
     def text(self):
         self._compact()
         return self.data
 
     def diff(self, m2, clean=False):
         '''Finds changes between the current manifest and m2.'''
         # XXX think whether efficiency matters here
         diff = {}
 
         for fn, e1, flags in self.iterentries():
             if fn not in m2:
                 diff[fn] = (e1, flags), (None, '')
             else:
                 e2 = m2[fn]
                 if (e1, flags) != e2:
                     diff[fn] = (e1, flags), e2
                 elif clean:
                     diff[fn] = None
 
         for fn, e2, flags in m2.iterentries():
             if fn not in self:
                 diff[fn] = (None, ''), (e2, flags)
 
         return diff
 
     def iterentries(self):
         return lazymanifestiterentries(self)
 
     def iterkeys(self):
         return lazymanifestiter(self)
 
     def __iter__(self):
         return lazymanifestiter(self)
 
     def __len__(self):
         return len(self.positions)
 
     def filtercopy(self, filterfn):
         # XXX should be optimized
         c = _lazymanifest('')
         for f, n, fl in self.iterentries():
             if filterfn(f):
                 c[f] = n, fl
         return c
 
 try:
     _lazymanifest = parsers.lazymanifest
 except AttributeError:
     pass
 
 class manifestdict(object):
     def __init__(self, data=''):
         if data.startswith('\0'):
             #_lazymanifest can not parse v2
             self._lm = _lazymanifest('')
             for f, n, fl in _parsev2(data):
                 self._lm[f] = n, fl
         else:
             self._lm = _lazymanifest(data)
 
     def __getitem__(self, key):
         return self._lm[key][0]
 
     def find(self, key):
         return self._lm[key]
 
     def __len__(self):
         return len(self._lm)
 
     def __setitem__(self, key, node):
         self._lm[key] = node, self.flags(key, '')
 
     def __contains__(self, key):
         return key in self._lm
 
     def __delitem__(self, key):
         del self._lm[key]
 
     def __iter__(self):
         return self._lm.__iter__()
 
     def iterkeys(self):
         return self._lm.iterkeys()
 
     def keys(self):
         return list(self.iterkeys())
 
     def filesnotin(self, m2):
         '''Set of files in this manifest that are not in the other'''
         diff = self.diff(m2)
         files = set(filepath
                     for filepath, hashflags in diff.iteritems()
                     if hashflags[1][0] is None)
         return files
 
     @propertycache
     def _dirs(self):
         return util.dirs(self)
 
     def dirs(self):
         return self._dirs
 
     def hasdir(self, dir):
         return dir in self._dirs
 
     def _filesfastpath(self, match):
         '''Checks whether we can correctly and quickly iterate over matcher
         files instead of over manifest files.'''
         files = match.files()
         return (len(files) < 100 and (match.isexact() or
             (match.prefix() and all(fn in self for fn in files))))
 
     def walk(self, match):
         '''Generates matching file names.
 
         Equivalent to manifest.matches(match).iterkeys(), but without creating
         an entirely new manifest.
 
         It also reports nonexistent files by marking them bad with match.bad().
         '''
         if match.always():
             for f in iter(self):
                 yield f
             return
 
         fset = set(match.files())
 
         # avoid the entire walk if we're only looking for specific files
         if self._filesfastpath(match):
             for fn in sorted(fset):
                 yield fn
             return
 
         for fn in self:
             if fn in fset:
                 # specified pattern is the exact name
                 fset.remove(fn)
             if match(fn):
                 yield fn
 
         # for dirstate.walk, files=['.'] means "walk the whole tree".
         # follow that here, too
         fset.discard('.')
 
         for fn in sorted(fset):
             if not self.hasdir(fn):
                 match.bad(fn, None)
 
     def matches(self, match):
         '''generate a new manifest filtered by the match argument'''
         if match.always():
             return self.copy()
 
         if self._filesfastpath(match):
             m = manifestdict()
             lm = self._lm
             for fn in match.files():
                 if fn in lm:
                     m._lm[fn] = lm[fn]
             return m
 
         m = manifestdict()
         m._lm = self._lm.filtercopy(match)
         return m
 
     def diff(self, m2, clean=False):
         '''Finds changes between the current manifest and m2.
 
         Args:
           m2: the manifest to which this manifest should be compared.
           clean: if true, include files unchanged between these manifests
                  with a None value in the returned dictionary.
 
         The result is returned as a dict with filename as key and
         values of the form ((n1,fl1),(n2,fl2)), where n1/n2 is the
         nodeid in the current/other manifest and fl1/fl2 is the flag
         in the current/other manifest. Where the file does not exist,
         the nodeid will be None and the flags will be the empty
         string.
         '''
         return self._lm.diff(m2._lm, clean)
 
     def setflag(self, key, flag):
         self._lm[key] = self[key], flag
 
     def get(self, key, default=None):
         try:
             return self._lm[key][0]
         except KeyError:
             return default
 
     def flags(self, key, default=''):
         try:
             return self._lm[key][1]
         except KeyError:
             return default
 
     def copy(self):
         c = manifestdict()
         c._lm = self._lm.copy()
         return c
 
     def iteritems(self):
         return (x[:2] for x in self._lm.iterentries())
 
     def iterentries(self):
         return self._lm.iterentries()
 
     def text(self, usemanifestv2=False):
         if usemanifestv2:
             return _textv2(self._lm.iterentries())
         else:
             # use (probably) native version for v1
             return self._lm.text()
 
     def fastdelta(self, base, changes):
         """Given a base manifest text as an array.array and a list of changes
         relative to that text, compute a delta that can be used by revlog.
         """
         delta = []
         dstart = None
         dend = None
         dline = [""]
         start = 0
         # zero copy representation of base as a buffer
         addbuf = util.buffer(base)
 
         changes = list(changes)
         if len(changes) < 1000:
             # start with a readonly loop that finds the offset of
             # each line and creates the deltas
             for f, todelete in changes:
                 # bs will either be the index of the item or the insert point
                 start, end = _msearch(addbuf, f, start)
                 if not todelete:
                     h, fl = self._lm[f]
                     l = "%s\0%s%s\n" % (f, revlog.hex(h), fl)
                 else:
                     if start == end:
                         # item we want to delete was not found, error out
                         raise AssertionError(
                                 _("failed to remove %s from manifest") % f)
                     l = ""
                 if dstart is not None and dstart <= start and dend >= start:
                     if dend < end:
                         dend = end
                     if l:
                         dline.append(l)
                 else:
                     if dstart is not None:
                         delta.append([dstart, dend, "".join(dline)])
                     dstart = start
                     dend = end
                     dline = [l]
 
             if dstart is not None:
                 delta.append([dstart, dend, "".join(dline)])
             # apply the delta to the base, and get a delta for addrevision
             deltatext, arraytext = _addlistdelta(base, delta)
         else:
             # For large changes, it's much cheaper to just build the text and
             # diff it.
             arraytext = array.array('c', self.text())
             deltatext = mdiff.textdiff(base, arraytext)
 
         return arraytext, deltatext
 
 def _msearch(m, s, lo=0, hi=None):
     '''return a tuple (start, end) that says where to find s within m.
 
     If the string is found m[start:end] are the line containing
     that string.  If start == end the string was not found and
     they indicate the proper sorted insertion point.
 
     m should be a buffer or a string
     s is a string'''
     def advance(i, c):
         while i < lenm and m[i] != c:
             i += 1
         return i
     if not s:
         return (lo, lo)
     lenm = len(m)
     if not hi:
         hi = lenm
     while lo < hi:
         mid = (lo + hi) // 2
         start = mid
         while start > 0 and m[start - 1] != '\n':
             start -= 1
         end = advance(start, '\0')
         if m[start:end] < s:
             # we know that after the null there are 40 bytes of sha1
             # this translates to the bisect lo = mid + 1
             lo = advance(end + 40, '\n') + 1
         else:
             # this translates to the bisect hi = mid
             hi = start
     end = advance(lo, '\0')
     found = m[lo:end]
     if s == found:
         # we know that after the null there are 40 bytes of sha1
         end = advance(end + 40, '\n')
         return (lo, end + 1)
     else:
         return (lo, lo)
 
 def _checkforbidden(l):
     """Check filenames for illegal characters."""
     for f in l:
         if '\n' in f or '\r' in f:
             raise error.RevlogError(
                 _("'\\n' and '\\r' disallowed in filenames: %r") % f)
 
 
 # apply the changes collected during the bisect loop to our addlist
 # return a delta suitable for addrevision
 def _addlistdelta(addlist, x):
     # for large addlist arrays, building a new array is cheaper
     # than repeatedly modifying the existing one
     currentposition = 0
     newaddlist = array.array('c')
 
     for start, end, content in x:
         newaddlist += addlist[currentposition:start]
         if content:
             newaddlist += array.array('c', content)
 
         currentposition = end
 
     newaddlist += addlist[currentposition:]
 
     deltatext = "".join(struct.pack(">lll", start, end, len(content))
                    + content for start, end, content in x)
     return deltatext, newaddlist
 
 def _splittopdir(f):
     if '/' in f:
         dir, subpath = f.split('/', 1)
         return dir + '/', subpath
     else:
         return '', f
 
 _noop = lambda s: None
 
 class treemanifest(object):
     def __init__(self, dir='', text=''):
         self._dir = dir
         self._node = revlog.nullid
         self._loadfunc = _noop
         self._copyfunc = _noop
         self._dirty = False
         self._dirs = {}
         # Using _lazymanifest here is a little slower than plain old dicts
         self._files = {}
         self._flags = {}
         if text:
             def readsubtree(subdir, subm):
                 raise AssertionError('treemanifest constructor only accepts '
                                      'flat manifests')
             self.parse(text, readsubtree)
             self._dirty = True # Mark flat manifest dirty after parsing
 
     def _subpath(self, path):
         return self._dir + path
 
     def __len__(self):
         self._load()
         size = len(self._files)
         for m in self._dirs.values():
             size += m.__len__()
         return size
 
     def _isempty(self):
         self._load() # for consistency; already loaded by all callers
         return (not self._files and (not self._dirs or
                 all(m._isempty() for m in self._dirs.values())))
 
     def __repr__(self):
         return ('<treemanifest dir=%s, node=%s, loaded=%s, dirty=%s at 0x%x>' %
                 (self._dir, revlog.hex(self._node),
                  bool(self._loadfunc is _noop),
                  self._dirty, id(self)))
 
     def dir(self):
         '''The directory that this tree manifest represents, including a
         trailing '/'. Empty string for the repo root directory.'''
         return self._dir
 
     def node(self):
         '''This node of this instance. nullid for unsaved instances. Should
         be updated when the instance is read or written from a revlog.
         '''
         assert not self._dirty
         return self._node
 
     def setnode(self, node):
         self._node = node
         self._dirty = False
 
     def iterentries(self):
         self._load()
         for p, n in sorted(self._dirs.items() + self._files.items()):
             if p in self._files:
                 yield self._subpath(p), n, self._flags.get(p, '')
             else:
                 for x in n.iterentries():
                     yield x
 
     def iteritems(self):
         self._load()
         for p, n in sorted(self._dirs.items() + self._files.items()):
             if p in self._files:
                 yield self._subpath(p), n
             else:
                 for f, sn in n.iteritems():
                     yield f, sn
 
     def iterkeys(self):
         self._load()
         for p in sorted(self._dirs.keys() + self._files.keys()):
             if p in self._files:
                 yield self._subpath(p)
             else:
                 for f in self._dirs[p].iterkeys():
                     yield f
 
     def keys(self):
         return list(self.iterkeys())
 
     def __iter__(self):
         return self.iterkeys()
 
     def __contains__(self, f):
         if f is None:
             return False
         self._load()
         dir, subpath = _splittopdir(f)
         if dir:
             if dir not in self._dirs:
                 return False
             return self._dirs[dir].__contains__(subpath)
         else:
             return f in self._files
 
     def get(self, f, default=None):
         self._load()
         dir, subpath = _splittopdir(f)
         if dir:
             if dir not in self._dirs:
                 return default
             return self._dirs[dir].get(subpath, default)
         else:
             return self._files.get(f, default)
 
     def __getitem__(self, f):
         self._load()
         dir, subpath = _splittopdir(f)
         if dir:
             return self._dirs[dir].__getitem__(subpath)
         else:
             return self._files[f]
 
     def flags(self, f):
         self._load()
         dir, subpath = _splittopdir(f)
         if dir:
             if dir not in self._dirs:
                 return ''
             return self._dirs[dir].flags(subpath)
         else:
             if f in self._dirs:
                 return ''
             return self._flags.get(f, '')
 
     def find(self, f):
         self._load()
         dir, subpath = _splittopdir(f)
         if dir:
             return self._dirs[dir].find(subpath)
         else:
             return self._files[f], self._flags.get(f, '')
 
     def __delitem__(self, f):
         self._load()
         dir, subpath = _splittopdir(f)
         if dir:
             self._dirs[dir].__delitem__(subpath)
             # If the directory is now empty, remove it
             if self._dirs[dir]._isempty():
                 del self._dirs[dir]
         else:
             del self._files[f]
             if f in self._flags:
                 del self._flags[f]
         self._dirty = True
 
     def __setitem__(self, f, n):
         assert n is not None
         self._load()
         dir, subpath = _splittopdir(f)
         if dir:
             if dir not in self._dirs:
                 self._dirs[dir] = treemanifest(self._subpath(dir))
             self._dirs[dir].__setitem__(subpath, n)
         else:
             self._files[f] = n[:21] # to match manifestdict's behavior
         self._dirty = True
 
     def _load(self):
         if self._loadfunc is not _noop:
             lf, self._loadfunc = self._loadfunc, _noop
             lf(self)
         elif self._copyfunc is not _noop:
             cf, self._copyfunc = self._copyfunc, _noop
             cf(self)
 
     def setflag(self, f, flags):
         """Set the flags (symlink, executable) for path f."""
         self._load()
         dir, subpath = _splittopdir(f)
         if dir:
             if dir not in self._dirs:
                 self._dirs[dir] = treemanifest(self._subpath(dir))
             self._dirs[dir].setflag(subpath, flags)
         else:
             self._flags[f] = flags
         self._dirty = True
 
     def copy(self):
         copy = treemanifest(self._dir)
         copy._node = self._node
         copy._dirty = self._dirty
         if self._copyfunc is _noop:
             def _copyfunc(s):
                 self._load()
                 for d in self._dirs:
                     s._dirs[d] = self._dirs[d].copy()
                 s._files = dict.copy(self._files)
                 s._flags = dict.copy(self._flags)
             if self._loadfunc is _noop:
                 _copyfunc(copy)
             else:
                 copy._copyfunc = _copyfunc
         else:
             copy._copyfunc = self._copyfunc
         return copy
 
     def filesnotin(self, m2):
         '''Set of files in this manifest that are not in the other'''
         files = set()
         def _filesnotin(t1, t2):
             if t1._node == t2._node and not t1._dirty and not t2._dirty:
                 return
             t1._load()
             t2._load()
             for d, m1 in t1._dirs.iteritems():
                 if d in t2._dirs:
                     m2 = t2._dirs[d]
                     _filesnotin(m1, m2)
                 else:
                     files.update(m1.iterkeys())
 
             for fn in t1._files.iterkeys():
                 if fn not in t2._files:
                     files.add(t1._subpath(fn))
 
         _filesnotin(self, m2)
         return files
 
     @propertycache
     def _alldirs(self):
         return util.dirs(self)
 
     def dirs(self):
         return self._alldirs
 
     def hasdir(self, dir):
         self._load()
         topdir, subdir = _splittopdir(dir)
         if topdir:
             if topdir in self._dirs:
                 return self._dirs[topdir].hasdir(subdir)
             return False
         return (dir + '/') in self._dirs
 
     def walk(self, match):
         '''Generates matching file names.
 
         Equivalent to manifest.matches(match).iterkeys(), but without creating
         an entirely new manifest.
 
         It also reports nonexistent files by marking them bad with match.bad().
         '''
         if match.always():
             for f in iter(self):
                 yield f
             return
 
         fset = set(match.files())
 
         for fn in self._walk(match):
             if fn in fset:
                 # specified pattern is the exact name
                 fset.remove(fn)
             yield fn
 
         # for dirstate.walk, files=['.'] means "walk the whole tree".
         # follow that here, too
         fset.discard('.')
 
         for fn in sorted(fset):
             if not self.hasdir(fn):
                 match.bad(fn, None)
 
     def _walk(self, match):
         '''Recursively generates matching file names for walk().'''
         if not match.visitdir(self._dir[:-1] or '.'):
             return
 
         # yield this dir's files and walk its submanifests
         self._load()
         for p in sorted(self._dirs.keys() + self._files.keys()):
             if p in self._files:
                 fullp = self._subpath(p)
                 if match(fullp):
                     yield fullp
             else:
                 for f in self._dirs[p]._walk(match):
                     yield f
 
     def matches(self, match):
         '''generate a new manifest filtered by the match argument'''
         if match.always():
             return self.copy()
 
         return self._matches(match)
 
     def _matches(self, match):
         '''recursively generate a new manifest filtered by the match argument.
         '''
 
         visit = match.visitdir(self._dir[:-1] or '.')
         if visit == 'all':
             return self.copy()
         ret = treemanifest(self._dir)
         if not visit:
             return ret
 
         self._load()
         for fn in self._files:
             fullp = self._subpath(fn)
             if not match(fullp):
                 continue
             ret._files[fn] = self._files[fn]
             if fn in self._flags:
                 ret._flags[fn] = self._flags[fn]
 
         for dir, subm in self._dirs.iteritems():
             m = subm._matches(match)
             if not m._isempty():
                 ret._dirs[dir] = m
 
         if not ret._isempty():
             ret._dirty = True
         return ret
 
     def diff(self, m2, clean=False):
         '''Finds changes between the current manifest and m2.
 
         Args:
           m2: the manifest to which this manifest should be compared.
           clean: if true, include files unchanged between these manifests
                  with a None value in the returned dictionary.
 
         The result is returned as a dict with filename as key and
         values of the form ((n1,fl1),(n2,fl2)), where n1/n2 is the
         nodeid in the current/other manifest and fl1/fl2 is the flag
         in the current/other manifest. Where the file does not exist,
         the nodeid will be None and the flags will be the empty
         string.
         '''
         result = {}
         emptytree = treemanifest()
         def _diff(t1, t2):
             if t1._node == t2._node and not t1._dirty and not t2._dirty:
                 return
             t1._load()
             t2._load()
             for d, m1 in t1._dirs.iteritems():
                 m2 = t2._dirs.get(d, emptytree)
                 _diff(m1, m2)
 
             for d, m2 in t2._dirs.iteritems():
                 if d not in t1._dirs:
                     _diff(emptytree, m2)
 
             for fn, n1 in t1._files.iteritems():
                 fl1 = t1._flags.get(fn, '')
                 n2 = t2._files.get(fn, None)
                 fl2 = t2._flags.get(fn, '')
                 if n1 != n2 or fl1 != fl2:
                     result[t1._subpath(fn)] = ((n1, fl1), (n2, fl2))
                 elif clean:
                     result[t1._subpath(fn)] = None
 
             for fn, n2 in t2._files.iteritems():
                 if fn not in t1._files:
                     fl2 = t2._flags.get(fn, '')
                     result[t2._subpath(fn)] = ((None, ''), (n2, fl2))
 
         _diff(self, m2)
         return result
 
     def unmodifiedsince(self, m2):
         return not self._dirty and not m2._dirty and self._node == m2._node
 
     def parse(self, text, readsubtree):
         for f, n, fl in _parse(text):
             if fl == 't':
                 f = f + '/'
                 self._dirs[f] = readsubtree(self._subpath(f), n)
             elif '/' in f:
                 # This is a flat manifest, so use __setitem__ and setflag rather
                 # than assigning directly to _files and _flags, so we can
                 # assign a path in a subdirectory, and to mark dirty (compared
                 # to nullid).
                 self[f] = n
                 if fl:
                     self.setflag(f, fl)
             else:
                 # Assigning to _files and _flags avoids marking as dirty,
                 # and should be a little faster.
                 self._files[f] = n
                 if fl:
                     self._flags[f] = fl
 
     def text(self, usemanifestv2=False):
         """Get the full data of this manifest as a bytestring."""
         self._load()
         return _text(self.iterentries(), usemanifestv2)
 
     def dirtext(self, usemanifestv2=False):
         """Get the full data of this directory as a bytestring. Make sure that
         any submanifests have been written first, so their nodeids are correct.
         """
         self._load()
         flags = self.flags
         dirs = [(d[:-1], self._dirs[d]._node, 't') for d in self._dirs]
         files = [(f, self._files[f], flags(f)) for f in self._files]
         return _text(sorted(dirs + files), usemanifestv2)
 
     def read(self, gettext, readsubtree):
         def _load_for_read(s):
             s.parse(gettext(), readsubtree)
             s._dirty = False
         self._loadfunc = _load_for_read
 
     def writesubtrees(self, m1, m2, writesubtree):
         self._load() # for consistency; should never have any effect here
         m1._load()
         m2._load()
         emptytree = treemanifest()
         for d, subm in self._dirs.iteritems():
             subp1 = m1._dirs.get(d, emptytree)._node
             subp2 = m2._dirs.get(d, emptytree)._node
             if subp1 == revlog.nullid:
                 subp1, subp2 = subp2, subp1
             writesubtree(subm, subp1, subp2)
 
 class manifestrevlog(revlog.revlog):
     '''A revlog that stores manifest texts. This is responsible for caching the
     full-text manifest contents.
     '''
     def __init__(self, opener, dir='', dirlogcache=None):
         # During normal operations, we expect to deal with not more than four
         # revs at a time (such as during commit --amend). When rebasing large
         # stacks of commits, the number can go up, hence the config knob below.
         cachesize = 4
         usetreemanifest = False
         usemanifestv2 = False
         opts = getattr(opener, 'options', None)
         if opts is not None:
             cachesize = opts.get('manifestcachesize', cachesize)
             usetreemanifest = opts.get('treemanifest', usetreemanifest)
             usemanifestv2 = opts.get('manifestv2', usemanifestv2)
 
         self._treeondisk = usetreemanifest
         self._usemanifestv2 = usemanifestv2
 
         self._fulltextcache = util.lrucachedict(cachesize)
 
         indexfile = "00manifest.i"
         if dir:
             assert self._treeondisk, 'opts is %r' % opts
             if not dir.endswith('/'):
                 dir = dir + '/'
             indexfile = "meta/" + dir + "00manifest.i"
         self._dir = dir
         # The dirlogcache is kept on the root manifest log
         if dir:
             self._dirlogcache = dirlogcache
         else:
             self._dirlogcache = {'': self}
 
         super(manifestrevlog, self).__init__(opener, indexfile,
                                              checkambig=bool(dir))
 
     @property
     def fulltextcache(self):
         return self._fulltextcache
 
     def clearcaches(self):
         super(manifestrevlog, self).clearcaches()
         self._fulltextcache.clear()
         self._dirlogcache = {'': self}
 
     def dirlog(self, dir):
         if dir:
             assert self._treeondisk
         if dir not in self._dirlogcache:
             self._dirlogcache[dir] = manifestrevlog(self.opener, dir,
                                                     self._dirlogcache)
         return self._dirlogcache[dir]
 
     def add(self, m, transaction, link, p1, p2, added, removed):
         if (p1 in self.fulltextcache and util.safehasattr(m, 'fastdelta')
             and not self._usemanifestv2):
             # If our first parent is in the manifest cache, we can
             # compute a delta here using properties we know about the
             # manifest up-front, which may save time later for the
             # revlog layer.
 
             _checkforbidden(added)
             # combine the changed lists into one sorted iterator
             work = heapq.merge([(x, False) for x in added],
                                [(x, True) for x in removed])
 
             arraytext, deltatext = m.fastdelta(self.fulltextcache[p1], work)
             cachedelta = self.rev(p1), deltatext
             text = util.buffer(arraytext)
             n = self.addrevision(text, transaction, link, p1, p2, cachedelta)
         else:
             # The first parent manifest isn't already loaded, so we'll
             # just encode a fulltext of the manifest and pass that
             # through to the revlog layer, and let it handle the delta
             # process.
             if self._treeondisk:
                 m1 = self.read(p1)
                 m2 = self.read(p2)
                 n = self._addtree(m, transaction, link, m1, m2)
                 arraytext = None
             else:
                 text = m.text(self._usemanifestv2)
                 n = self.addrevision(text, transaction, link, p1, p2)
                 arraytext = array.array('c', text)
 
         if arraytext is not None:
             self.fulltextcache[n] = arraytext
 
         return n
 
     def _addtree(self, m, transaction, link, m1, m2):
         # If the manifest is unchanged compared to one parent,
         # don't write a new revision
         if m.unmodifiedsince(m1) or m.unmodifiedsince(m2):
             return m.node()
         def writesubtree(subm, subp1, subp2):
             sublog = self.dirlog(subm.dir())
             sublog.add(subm, transaction, link, subp1, subp2, None, None)
         m.writesubtrees(m1, m2, writesubtree)
         text = m.dirtext(self._usemanifestv2)
         # Double-check whether contents are unchanged to one parent
         if text == m1.dirtext(self._usemanifestv2):
             n = m1.node()
         elif text == m2.dirtext(self._usemanifestv2):
             n = m2.node()
         else:
             n = self.addrevision(text, transaction, link, m1.node(), m2.node())
         # Save nodeid so parent manifest can calculate its nodeid
         m.setnode(n)
         return n
 
 class manifestlog(object):
     """A collection class representing the collection of manifest snapshots
     referenced by commits in the repository.
 
     In this situation, 'manifest' refers to the abstract concept of a snapshot
     of the list of files in the given commit. Consumers of the output of this
     class do not care about the implementation details of the actual manifests
     they receive (i.e. tree or flat or lazily loaded, etc)."""
     def __init__(self, opener, repo):
         self._repo = repo
 
         usetreemanifest = False
 
         opts = getattr(opener, 'options', None)
         if opts is not None:
             usetreemanifest = opts.get('treemanifest', usetreemanifest)
         self._treeinmem = usetreemanifest
 
         self._oldmanifest = repo._constructmanifest()
         self._revlog = self._oldmanifest
 
         # A cache of the manifestctx or treemanifestctx for each directory
         self._dirmancache = {}
 
         # We'll separate this into it's own cache once oldmanifest is no longer
         # used
         self._mancache = self._oldmanifest._mancache
         self._dirmancache[''] = self._mancache
 
         # A future patch makes this use the same config value as the existing
         # mancache
         self.cachesize = 4
 
     def __getitem__(self, node):
         """Retrieves the manifest instance for the given node. Throws a
         LookupError if not found.
         """
         return self.get('', node)
 
     def get(self, dir, node):
         """Retrieves the manifest instance for the given node. Throws a
         LookupError if not found.
         """
         if node in self._dirmancache.get(dir, ()):
             cachemf = self._dirmancache[dir][node]
             # The old manifest may put non-ctx manifests in the cache, so
             # skip those since they don't implement the full api.
             if (isinstance(cachemf, manifestctx) or
                 isinstance(cachemf, treemanifestctx)):
                 return cachemf
 
         if dir:
             if self._revlog._treeondisk:
                 dirlog = self._revlog.dirlog(dir)
                 if node not in dirlog.nodemap:
                     raise LookupError(node, dirlog.indexfile,
                                       _('no node'))
                 m = treemanifestctx(self._repo, dir, node)
             else:
                 raise error.Abort(
                         _("cannot ask for manifest directory '%s' in a flat "
                           "manifest") % dir)
         else:
             if node not in self._revlog.nodemap:
                 raise LookupError(node, self._revlog.indexfile,
                                   _('no node'))
             if self._treeinmem:
                 m = treemanifestctx(self._repo, '', node)
             else:
                 m = manifestctx(self._repo, node)
 
         if node != revlog.nullid:
             mancache = self._dirmancache.get(dir)
             if not mancache:
                 mancache = util.lrucachedict(self.cachesize)
                 self._dirmancache[dir] = mancache
             mancache[node] = m
         return m
 
     def add(self, m, transaction, link, p1, p2, added, removed):
         return self._revlog.add(m, transaction, link, p1, p2, added, removed)
 
 class manifestctx(object):
     """A class representing a single revision of a manifest, including its
     contents, its parent revs, and its linkrev.
     """
     def __init__(self, repo, node):
         self._repo = repo
         self._data = None
 
         self._node = node
 
         # TODO: We eventually want p1, p2, and linkrev exposed on this class,
         # but let's add it later when something needs it and we can load it
         # lazily.
         #self.p1, self.p2 = revlog.parents(node)
         #rev = revlog.rev(node)
         #self.linkrev = revlog.linkrev(rev)
 
     def node(self):
         return self._node
 
     def read(self):
         if not self._data:
             if self._node == revlog.nullid:
                 self._data = manifestdict()
             else:
                 rl = self._repo.manifestlog._revlog
                 text = rl.revision(self._node)
                 arraytext = array.array('c', text)
                 rl._fulltextcache[self._node] = arraytext
                 self._data = manifestdict(text)
         return self._data
 
     def readfast(self, shallow=False):
+        '''Calls either readdelta or read, based on which would be less work.
+        readdelta is called if the delta is against the p1, and therefore can be
+        read quickly.
+
+        If `shallow` is True, nothing changes since this is a flat manifest.
+        '''
         rl = self._repo.manifestlog._revlog
         r = rl.rev(self._node)
         deltaparent = rl.deltaparent(r)
         if deltaparent != revlog.nullrev and deltaparent in rl.parentrevs(r):
             return self.readdelta()
         return self.read()
 
     def readdelta(self, shallow=False):
         revlog = self._repo.manifestlog._revlog
         if revlog._usemanifestv2:
             # Need to perform a slow delta
             r0 = revlog.deltaparent(revlog.rev(self._node))
             m0 = manifestctx(self._repo, revlog.node(r0)).read()
             m1 = self.read()
             md = manifestdict()
             for f, ((n0, fl0), (n1, fl1)) in m0.diff(m1).iteritems():
                 if n1:
                     md[f] = n1
                     if fl1:
                         md.setflag(f, fl1)
             return md
 
         r = revlog.rev(self._node)
         d = mdiff.patchtext(revlog.revdiff(revlog.deltaparent(r), r))
         return manifestdict(d)
 
 class treemanifestctx(object):
     def __init__(self, repo, dir, node):
         self._repo = repo
         self._dir = dir
         self._data = None
 
         self._node = node
 
         # TODO: Load p1/p2/linkrev lazily. They need to be lazily loaded so that
         # we can instantiate treemanifestctx objects for directories we don't
         # have on disk.
         #self.p1, self.p2 = revlog.parents(node)
         #rev = revlog.rev(node)
         #self.linkrev = revlog.linkrev(rev)
 
     def _revlog(self):
         return self._repo.manifestlog._revlog.dirlog(self._dir)
 
     def read(self):
         if not self._data:
             rl = self._revlog()
             if self._node == revlog.nullid:
                 self._data = treemanifest()
             elif rl._treeondisk:
                 m = treemanifest(dir=self._dir)
                 def gettext():
                     return rl.revision(self._node)
                 def readsubtree(dir, subm):
                     return treemanifestctx(self._repo, dir, subm).read()
                 m.read(gettext, readsubtree)
                 m.setnode(self._node)
                 self._data = m
             else:
                 text = revlog.revision(self._node)
                 arraytext = array.array('c', text)
                 rl.fulltextcache[self._node] = arraytext
                 self._data = treemanifest(dir=self._dir, text=text)
 
         return self._data
 
     def node(self):
         return self._node
 
     def readdelta(self, shallow=False):
         revlog = self._revlog()
         if shallow and not revlog._usemanifestv2:
             r = revlog.rev(self._node)
             d = mdiff.patchtext(revlog.revdiff(revlog.deltaparent(r), r))
             return manifestdict(d)
         else:
             # Need to perform a slow delta
             r0 = revlog.deltaparent(revlog.rev(self._node))
             m0 = treemanifestctx(self._repo, self._dir, revlog.node(r0)).read()
             m1 = self.read()
             md = treemanifest(dir=self._dir)
             for f, ((n0, fl0), (n1, fl1)) in m0.diff(m1).iteritems():
                 if n1:
                     md[f] = n1
                     if fl1:
                         md.setflag(f, fl1)
             return md
 
     def readfast(self, shallow=False):
+        '''Calls either readdelta or read, based on which would be less work.
+        readdelta is called if the delta is against the p1, and therefore can be
+        read quickly.
+
+        If `shallow` is True, it only returns the entries from this manifest,
+        and not any submanifests.
+        '''
         rl = self._revlog()
         r = rl.rev(self._node)
         deltaparent = rl.deltaparent(r)
         if (deltaparent != revlog.nullrev and
             deltaparent in rl.parentrevs(r)):
             return self.readdelta(shallow=shallow)
 
         if shallow:
             return manifestdict(rl.revision(self._node))
         else:
             return self.read()
 
 class manifest(manifestrevlog):
     def __init__(self, opener, dir='', dirlogcache=None):
         '''The 'dir' and 'dirlogcache' arguments are for internal use by
         manifest.manifest only. External users should create a root manifest
         log with manifest.manifest(opener) and call dirlog() on it.
         '''
         # During normal operations, we expect to deal with not more than four
         # revs at a time (such as during commit --amend). When rebasing large
         # stacks of commits, the number can go up, hence the config knob below.
         cachesize = 4
         usetreemanifest = False
         opts = getattr(opener, 'options', None)
         if opts is not None:
             cachesize = opts.get('manifestcachesize', cachesize)
             usetreemanifest = opts.get('treemanifest', usetreemanifest)
         self._mancache = util.lrucachedict(cachesize)
         self._treeinmem = usetreemanifest
         super(manifest, self).__init__(opener, dir=dir, dirlogcache=dirlogcache)
 
     def _newmanifest(self, data=''):
         if self._treeinmem:
             return treemanifest(self._dir, data)
         return manifestdict(data)
 
     def dirlog(self, dir):
         """This overrides the base revlog implementation to allow construction
         'manifest' types instead of manifestrevlog types. This is only needed
         until we migrate off the 'manifest' type."""
         if dir:
             assert self._treeondisk
         if dir not in self._dirlogcache:
             self._dirlogcache[dir] = manifest(self.opener, dir,
                                               self._dirlogcache)
         return self._dirlogcache[dir]
 
     def _slowreaddelta(self, node):
         r0 = self.deltaparent(self.rev(node))
         m0 = self.read(self.node(r0))
         m1 = self.read(node)
         md = self._newmanifest()
         for f, ((n0, fl0), (n1, fl1)) in m0.diff(m1).iteritems():
             if n1:
                 md[f] = n1
                 if fl1:
                     md.setflag(f, fl1)
         return md
 
     def readdelta(self, node):
         if self._usemanifestv2 or self._treeondisk:
             return self._slowreaddelta(node)
         r = self.rev(node)
         d = mdiff.patchtext(self.revdiff(self.deltaparent(r), r))
         return self._newmanifest(d)
 
     def readshallowdelta(self, node):
         '''For flat manifests, this is the same as readdelta(). For
         treemanifests, this will read the delta for this revlog's directory,
         without recursively reading subdirectory manifests. Instead, any
         subdirectory entry will be reported as it appears in the manifests, i.e.
         the subdirectory will be reported among files and distinguished only by
         its 't' flag.'''
         if not self._treeondisk:
             return self.readdelta(node)
         if self._usemanifestv2:
             raise error.Abort(
                 _("readshallowdelta() not implemented for manifestv2"))
         r = self.rev(node)
         d = mdiff.patchtext(self.revdiff(self.deltaparent(r), r))
         return manifestdict(d)
 
-    def readshallowfast(self, node):
-        '''like readfast(), but calls readshallowdelta() instead of readdelta()
-        '''
-        r = self.rev(node)
-        deltaparent = self.deltaparent(r)
-        if deltaparent != revlog.nullrev and deltaparent in self.parentrevs(r):
-            return self.readshallowdelta(node)
-        return self.readshallow(node)
-
     def read(self, node):
         if node == revlog.nullid:
             return self._newmanifest() # don't upset local cache
         if node in self._mancache:
             cached = self._mancache[node]
             if (isinstance(cached, manifestctx) or
                 isinstance(cached, treemanifestctx)):
                 cached = cached.read()
             return cached
         if self._treeondisk:
             def gettext():
                 return self.revision(node)
             def readsubtree(dir, subm):
                 return self.dirlog(dir).read(subm)
             m = self._newmanifest()
             m.read(gettext, readsubtree)
             m.setnode(node)
             arraytext = None
         else:
             text = self.revision(node)
             m = self._newmanifest(text)
             arraytext = array.array('c', text)
         self._mancache[node] = m
         if arraytext is not None:
             self.fulltextcache[node] = arraytext
         return m
 
-    def readshallow(self, node):
-        '''Reads the manifest in this directory. When using flat manifests,
-        this manifest will generally have files in subdirectories in it. Does
-        not cache the manifest as the callers generally do not read the same
-        version twice.'''
-        return manifestdict(self.revision(node))
-
     def find(self, node, f):
         '''look up entry for a single file efficiently.
         return (node, flags) pair if found, (None, None) if not.'''
         m = self.read(node)
         try:
             return m.find(f)
         except KeyError:
             return None, None
 
     def clearcaches(self):
         super(manifest, self).clearcaches()
         self._mancache.clear()