py3: use default dict iterator instead of iterkeys

These are the easy cases. Some cases in manifest.py will require more
careful inspection.

Differential Revision: https://phab.mercurial-scm.org/D2315

contrib/synthrepo.py

@@ -1,516 +1,516
 # synthrepo.py - repo synthesis
 #
 # Copyright 2012 Facebook
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 '''synthesize structurally interesting change history
 
 This extension is useful for creating a repository with properties
 that are statistically similar to an existing repository. During
 analysis, a simple probability table is constructed from the history
 of an existing repository.  During synthesis, these properties are
 reconstructed.
 
 Properties that are analyzed and synthesized include the following:
 
 - Lines added or removed when an existing file is modified
 - Number and sizes of files added
 - Number of files removed
 - Line lengths
 - Topological distance to parent changeset(s)
 - Probability of a commit being a merge
 - Probability of a newly added file being added to a new directory
 - Interarrival time, and time zone, of commits
 - Number of files in each directory
 
 A few obvious properties that are not currently handled realistically:
 
 - Merges are treated as regular commits with two parents, which is not
   realistic
 - Modifications are not treated as operations on hunks of lines, but
   as insertions and deletions of randomly chosen single lines
 - Committer ID (always random)
 - Executability of files
 - Symlinks and binary files are ignored
 '''
 
 from __future__ import absolute_import
 import bisect
 import collections
 import itertools
 import json
 import os
 import random
 import sys
 import time
 
 from mercurial.i18n import _
 from mercurial.node import (
     nullid,
     nullrev,
     short,
 )
 from mercurial import (
     context,
     error,
     hg,
     patch,
     registrar,
     scmutil,
     util,
 )
 
 # Note for extension authors: ONLY specify testedwith = 'ships-with-hg-core' for
 # extensions which SHIP WITH MERCURIAL. Non-mainline extensions should
 # be specifying the version(s) of Mercurial they are tested with, or
 # leave the attribute unspecified.
 testedwith = 'ships-with-hg-core'
 
 cmdtable = {}
 command = registrar.command(cmdtable)
 
 newfile = {'new fi', 'rename', 'copy f', 'copy t'}
 
 def zerodict():
     return collections.defaultdict(lambda: 0)
 
 def roundto(x, k):
     if x > k * 2:
         return int(round(x / float(k)) * k)
     return int(round(x))
 
 def parsegitdiff(lines):
     filename, mar, lineadd, lineremove = None, None, zerodict(), 0
     binary = False
     for line in lines:
         start = line[:6]
         if start == 'diff -':
             if filename:
                 yield filename, mar, lineadd, lineremove, binary
             mar, lineadd, lineremove, binary = 'm', zerodict(), 0, False
             filename = patch.gitre.match(line).group(1)
         elif start in newfile:
             mar = 'a'
         elif start == 'GIT bi':
             binary = True
         elif start == 'delete':
             mar = 'r'
         elif start:
             s = start[0]
             if s == '-' and not line.startswith('--- '):
                 lineremove += 1
             elif s == '+' and not line.startswith('+++ '):
                 lineadd[roundto(len(line) - 1, 5)] += 1
     if filename:
         yield filename, mar, lineadd, lineremove, binary
 
 @command('analyze',
          [('o', 'output', '', _('write output to given file'), _('FILE')),
           ('r', 'rev', [], _('analyze specified revisions'), _('REV'))],
          _('hg analyze'), optionalrepo=True)
 def analyze(ui, repo, *revs, **opts):
     '''create a simple model of a repository to use for later synthesis
 
     This command examines every changeset in the given range (or all
     of history if none are specified) and creates a simple statistical
     model of the history of the repository. It also measures the directory
     structure of the repository as checked out.
 
     The model is written out to a JSON file, and can be used by
     :hg:`synthesize` to create or augment a repository with synthetic
     commits that have a structure that is statistically similar to the
     analyzed repository.
     '''
     root = repo.root
     if not root.endswith(os.path.sep):
         root += os.path.sep
 
     revs = list(revs)
     revs.extend(opts['rev'])
     if not revs:
         revs = [':']
 
     output = opts['output']
     if not output:
         output = os.path.basename(root) + '.json'
 
     if output == '-':
         fp = sys.stdout
     else:
         fp = open(output, 'w')
 
     # Always obtain file counts of each directory in the given root directory.
     def onerror(e):
         ui.warn(_('error walking directory structure: %s\n') % e)
 
     dirs = {}
     rootprefixlen = len(root)
     for dirpath, dirnames, filenames in os.walk(root, onerror=onerror):
         dirpathfromroot = dirpath[rootprefixlen:]
         dirs[dirpathfromroot] = len(filenames)
         if '.hg' in dirnames:
             dirnames.remove('.hg')
 
     lineschanged = zerodict()
     children = zerodict()
     p1distance = zerodict()
     p2distance = zerodict()
     linesinfilesadded = zerodict()
     fileschanged = zerodict()
     filesadded = zerodict()
     filesremoved = zerodict()
     linelengths = zerodict()
     interarrival = zerodict()
     parents = zerodict()
     dirsadded = zerodict()
     tzoffset = zerodict()
 
     # If a mercurial repo is available, also model the commit history.
     if repo:
         revs = scmutil.revrange(repo, revs)
         revs.sort()
 
         progress = ui.progress
         _analyzing = _('analyzing')
         _changesets = _('changesets')
         _total = len(revs)
 
         for i, rev in enumerate(revs):
             progress(_analyzing, i, unit=_changesets, total=_total)
             ctx = repo[rev]
             pl = ctx.parents()
             pctx = pl[0]
             prev = pctx.rev()
             children[prev] += 1
             p1distance[rev - prev] += 1
             parents[len(pl)] += 1
             tzoffset[ctx.date()[1]] += 1
             if len(pl) > 1:
                 p2distance[rev - pl[1].rev()] += 1
             if prev == rev - 1:
                 lastctx = pctx
             else:
                 lastctx = repo[rev - 1]
             if lastctx.rev() != nullrev:
                 timedelta = ctx.date()[0] - lastctx.date()[0]
                 interarrival[roundto(timedelta, 300)] += 1
             diff = sum((d.splitlines() for d in ctx.diff(pctx, git=True)), [])
             fileadds, diradds, fileremoves, filechanges = 0, 0, 0, 0
             for filename, mar, lineadd, lineremove, isbin in parsegitdiff(diff):
                 if isbin:
                     continue
                 added = sum(lineadd.itervalues(), 0)
                 if mar == 'm':
                     if added and lineremove:
                         lineschanged[roundto(added, 5),
                                      roundto(lineremove, 5)] += 1
                         filechanges += 1
                 elif mar == 'a':
                     fileadds += 1
                     if '/' in filename:
                         filedir = filename.rsplit('/', 1)[0]
                         if filedir not in pctx.dirs():
                             diradds += 1
                     linesinfilesadded[roundto(added, 5)] += 1
                 elif mar == 'r':
                     fileremoves += 1
                 for length, count in lineadd.iteritems():
                     linelengths[length] += count
             fileschanged[filechanges] += 1
             filesadded[fileadds] += 1
             dirsadded[diradds] += 1
             filesremoved[fileremoves] += 1
 
     invchildren = zerodict()
 
     for rev, count in children.iteritems():
         invchildren[count] += 1
 
     if output != '-':
         ui.status(_('writing output to %s\n') % output)
 
     def pronk(d):
         return sorted(d.iteritems(), key=lambda x: x[1], reverse=True)
 
     json.dump({'revs': len(revs),
                'initdirs': pronk(dirs),
                'lineschanged': pronk(lineschanged),
                'children': pronk(invchildren),
                'fileschanged': pronk(fileschanged),
                'filesadded': pronk(filesadded),
                'linesinfilesadded': pronk(linesinfilesadded),
                'dirsadded': pronk(dirsadded),
                'filesremoved': pronk(filesremoved),
                'linelengths': pronk(linelengths),
                'parents': pronk(parents),
                'p1distance': pronk(p1distance),
                'p2distance': pronk(p2distance),
                'interarrival': pronk(interarrival),
                'tzoffset': pronk(tzoffset),
                },
               fp)
     fp.close()
 
 @command('synthesize',
          [('c', 'count', 0, _('create given number of commits'), _('COUNT')),
           ('', 'dict', '', _('path to a dictionary of words'), _('FILE')),
           ('', 'initfiles', 0, _('initial file count to create'), _('COUNT'))],
          _('hg synthesize [OPTION].. DESCFILE'))
 def synthesize(ui, repo, descpath, **opts):
     '''synthesize commits based on a model of an existing repository
 
     The model must have been generated by :hg:`analyze`. Commits will
     be generated randomly according to the probabilities described in
     the model. If --initfiles is set, the repository will be seeded with
     the given number files following the modeled repository's directory
     structure.
 
     When synthesizing new content, commit descriptions, and user
     names, words will be chosen randomly from a dictionary that is
     presumed to contain one word per line. Use --dict to specify the
     path to an alternate dictionary to use.
     '''
     try:
         fp = hg.openpath(ui, descpath)
     except Exception as err:
         raise error.Abort('%s: %s' % (descpath, err[0].strerror))
     desc = json.load(fp)
     fp.close()
 
     def cdf(l):
         if not l:
             return [], []
         vals, probs = zip(*sorted(l, key=lambda x: x[1], reverse=True))
         t = float(sum(probs, 0))
         s, cdfs = 0, []
         for v in probs:
             s += v
             cdfs.append(s / t)
         return vals, cdfs
 
     lineschanged = cdf(desc['lineschanged'])
     fileschanged = cdf(desc['fileschanged'])
     filesadded = cdf(desc['filesadded'])
     dirsadded = cdf(desc['dirsadded'])
     filesremoved = cdf(desc['filesremoved'])
     linelengths = cdf(desc['linelengths'])
     parents = cdf(desc['parents'])
     p1distance = cdf(desc['p1distance'])
     p2distance = cdf(desc['p2distance'])
     interarrival = cdf(desc['interarrival'])
     linesinfilesadded = cdf(desc['linesinfilesadded'])
     tzoffset = cdf(desc['tzoffset'])
 
     dictfile = opts.get('dict') or '/usr/share/dict/words'
     try:
         fp = open(dictfile, 'rU')
     except IOError as err:
         raise error.Abort('%s: %s' % (dictfile, err.strerror))
     words = fp.read().splitlines()
     fp.close()
 
     initdirs = {}
     if desc['initdirs']:
         for k, v in desc['initdirs']:
             initdirs[k.encode('utf-8').replace('.hg', '_hg')] = v
         initdirs = renamedirs(initdirs, words)
     initdirscdf = cdf(initdirs)
 
     def pick(cdf):
         return cdf[0][bisect.bisect_left(cdf[1], random.random())]
 
     def pickpath():
         return os.path.join(pick(initdirscdf), random.choice(words))
 
     def makeline(minimum=0):
         total = max(minimum, pick(linelengths))
         c, l = 0, []
         while c < total:
             w = random.choice(words)
             c += len(w) + 1
             l.append(w)
         return ' '.join(l)
 
     wlock = repo.wlock()
     lock = repo.lock()
 
     nevertouch = {'.hgsub', '.hgignore', '.hgtags'}
 
     progress = ui.progress
     _synthesizing = _('synthesizing')
     _files = _('initial files')
     _changesets = _('changesets')
 
     # Synthesize a single initial revision adding files to the repo according
     # to the modeled directory structure.
     initcount = int(opts['initfiles'])
     if initcount and initdirs:
         pctx = repo[None].parents()[0]
         dirs = set(pctx.dirs())
         files = {}
 
         def validpath(path):
             # Don't pick filenames which are already directory names.
             if path in dirs:
                 return False
             # Don't pick directories which were used as file names.
             while path:
                 if path in files:
                     return False
                 path = os.path.dirname(path)
             return True
 
         for i in xrange(0, initcount):
             ui.progress(_synthesizing, i, unit=_files, total=initcount)
 
             path = pickpath()
             while not validpath(path):
                 path = pickpath()
             data = '%s contents\n' % path
             files[path] = data
             dir = os.path.dirname(path)
             while dir and dir not in dirs:
                 dirs.add(dir)
                 dir = os.path.dirname(dir)
 
         def filectxfn(repo, memctx, path):
             return context.memfilectx(repo, memctx, path, files[path])
 
         ui.progress(_synthesizing, None)
         message = 'synthesized wide repo with %d files' % (len(files),)
         mc = context.memctx(repo, [pctx.node(), nullid], message,
-                            files.iterkeys(), filectxfn, ui.username(),
+                            files, filectxfn, ui.username(),
                             '%d %d' % util.makedate())
         initnode = mc.commit()
         if ui.debugflag:
             hexfn = hex
         else:
             hexfn = short
         ui.status(_('added commit %s with %d files\n')
                   % (hexfn(initnode), len(files)))
 
     # Synthesize incremental revisions to the repository, adding repo depth.
     count = int(opts['count'])
     heads = set(map(repo.changelog.rev, repo.heads()))
     for i in xrange(count):
         progress(_synthesizing, i, unit=_changesets, total=count)
 
         node = repo.changelog.node
         revs = len(repo)
 
         def pickhead(heads, distance):
             if heads:
                 lheads = sorted(heads)
                 rev = revs - min(pick(distance), revs)
                 if rev < lheads[-1]:
                     rev = lheads[bisect.bisect_left(lheads, rev)]
                 else:
                     rev = lheads[-1]
                 return rev, node(rev)
             return nullrev, nullid
 
         r1 = revs - min(pick(p1distance), revs)
         p1 = node(r1)
 
         # the number of heads will grow without bound if we use a pure
         # model, so artificially constrain their proliferation
         toomanyheads = len(heads) > random.randint(1, 20)
         if p2distance[0] and (pick(parents) == 2 or toomanyheads):
             r2, p2 = pickhead(heads.difference([r1]), p2distance)
         else:
             r2, p2 = nullrev, nullid
 
         pl = [p1, p2]
         pctx = repo[r1]
         mf = pctx.manifest()
         mfk = mf.keys()
         changes = {}
         if mfk:
             for __ in xrange(pick(fileschanged)):
                 for __ in xrange(10):
                     fctx = pctx.filectx(random.choice(mfk))
                     path = fctx.path()
                     if not (path in nevertouch or fctx.isbinary() or
                             'l' in fctx.flags()):
                         break
                 lines = fctx.data().splitlines()
                 add, remove = pick(lineschanged)
                 for __ in xrange(remove):
                     if not lines:
                         break
                     del lines[random.randrange(0, len(lines))]
                 for __ in xrange(add):
                     lines.insert(random.randint(0, len(lines)), makeline())
                 path = fctx.path()
                 changes[path] = '\n'.join(lines) + '\n'
             for __ in xrange(pick(filesremoved)):
                 path = random.choice(mfk)
                 for __ in xrange(10):
                     path = random.choice(mfk)
                     if path not in changes:
                         break
         if filesadded:
             dirs = list(pctx.dirs())
             dirs.insert(0, '')
         for __ in xrange(pick(filesadded)):
             pathstr = ''
             while pathstr in dirs:
                 path = [random.choice(dirs)]
                 if pick(dirsadded):
                     path.append(random.choice(words))
                 path.append(random.choice(words))
                 pathstr = '/'.join(filter(None, path))
             data = '\n'.join(makeline()
                              for __ in xrange(pick(linesinfilesadded))) + '\n'
             changes[pathstr] = data
         def filectxfn(repo, memctx, path):
             if path not in changes:
                 return None
             return context.memfilectx(repo, memctx, path, changes[path])
         if not changes:
             continue
         if revs:
             date = repo['tip'].date()[0] + pick(interarrival)
         else:
             date = time.time() - (86400 * count)
         # dates in mercurial must be positive, fit in 32-bit signed integers.
         date = min(0x7fffffff, max(0, date))
         user = random.choice(words) + '@' + random.choice(words)
         mc = context.memctx(repo, pl, makeline(minimum=2),
                             sorted(changes),
                             filectxfn, user, '%d %d' % (date, pick(tzoffset)))
         newnode = mc.commit()
         heads.add(repo.changelog.rev(newnode))
         heads.discard(r1)
         heads.discard(r2)
 
     lock.release()
     wlock.release()
 
 def renamedirs(dirs, words):
     '''Randomly rename the directory names in the per-dir file count dict.'''
     wordgen = itertools.cycle(words)
     replacements = {'': ''}
     def rename(dirpath):
         '''Recursively rename the directory and all path prefixes.
 
         The mapping from path to renamed path is stored for all path prefixes
         as in dynamic programming, ensuring linear runtime and consistent
         renaming regardless of iteration order through the model.
         '''
         if dirpath in replacements:
             return replacements[dirpath]
         head, _ = os.path.split(dirpath)
         if head:
             head = rename(head)
         else:
             head = ''
         renamed = os.path.join(head, next(wordgen))
         replacements[dirpath] = renamed
         return renamed
     result = []
     for dirpath, count in dirs.iteritems():
         result.append([rename(dirpath.lstrip(os.sep)), count])
     return result

hgext/convert/subversion.py

@@ -1,1356 +1,1356
 # Subversion 1.4/1.5 Python API backend
 #
 # Copyright(C) 2007 Daniel Holth et al
 from __future__ import absolute_import
 
 import os
 import re
 import tempfile
 import xml.dom.minidom
 
 from mercurial.i18n import _
 from mercurial import (
     encoding,
     error,
     pycompat,
     util,
     vfs as vfsmod,
 )
 
 from . import common
 
 pickle = util.pickle
 stringio = util.stringio
 propertycache = util.propertycache
 urlerr = util.urlerr
 urlreq = util.urlreq
 
 commandline = common.commandline
 commit = common.commit
 converter_sink = common.converter_sink
 converter_source = common.converter_source
 decodeargs = common.decodeargs
 encodeargs = common.encodeargs
 makedatetimestamp = common.makedatetimestamp
 mapfile = common.mapfile
 MissingTool = common.MissingTool
 NoRepo = common.NoRepo
 
 # Subversion stuff. Works best with very recent Python SVN bindings
 # e.g. SVN 1.5 or backports. Thanks to the bzr folks for enhancing
 # these bindings.
 
 try:
     import svn
     import svn.client
     import svn.core
     import svn.ra
     import svn.delta
     from . import transport
     import warnings
     warnings.filterwarnings('ignore',
             module='svn.core',
             category=DeprecationWarning)
     svn.core.SubversionException # trigger import to catch error
 
 except ImportError:
     svn = None
 
 class SvnPathNotFound(Exception):
     pass
 
 def revsplit(rev):
     """Parse a revision string and return (uuid, path, revnum).
     >>> revsplit(b'svn:a2147622-4a9f-4db4-a8d3-13562ff547b2'
     ...          b'/proj%20B/mytrunk/mytrunk@1')
     ('a2147622-4a9f-4db4-a8d3-13562ff547b2', '/proj%20B/mytrunk/mytrunk', 1)
     >>> revsplit(b'svn:8af66a51-67f5-4354-b62c-98d67cc7be1d@1')
     ('', '', 1)
     >>> revsplit(b'@7')
     ('', '', 7)
     >>> revsplit(b'7')
     ('', '', 0)
     >>> revsplit(b'bad')
     ('', '', 0)
     """
     parts = rev.rsplit('@', 1)
     revnum = 0
     if len(parts) > 1:
         revnum = int(parts[1])
     parts = parts[0].split('/', 1)
     uuid = ''
     mod = ''
     if len(parts) > 1 and parts[0].startswith('svn:'):
         uuid = parts[0][4:]
         mod = '/' + parts[1]
     return uuid, mod, revnum
 
 def quote(s):
     # As of svn 1.7, many svn calls expect "canonical" paths. In
     # theory, we should call svn.core.*canonicalize() on all paths
     # before passing them to the API.  Instead, we assume the base url
     # is canonical and copy the behaviour of svn URL encoding function
     # so we can extend it safely with new components. The "safe"
     # characters were taken from the "svn_uri__char_validity" table in
     # libsvn_subr/path.c.
     return urlreq.quote(s, "!$&'()*+,-./:=@_~")
 
 def geturl(path):
     try:
         return svn.client.url_from_path(svn.core.svn_path_canonicalize(path))
     except svn.core.SubversionException:
         # svn.client.url_from_path() fails with local repositories
         pass
     if os.path.isdir(path):
         path = os.path.normpath(os.path.abspath(path))
         if pycompat.iswindows:
             path = '/' + util.normpath(path)
         # Module URL is later compared with the repository URL returned
         # by svn API, which is UTF-8.
         path = encoding.tolocal(path)
         path = 'file://%s' % quote(path)
     return svn.core.svn_path_canonicalize(path)
 
 def optrev(number):
     optrev = svn.core.svn_opt_revision_t()
     optrev.kind = svn.core.svn_opt_revision_number
     optrev.value.number = number
     return optrev
 
 class changedpath(object):
     def __init__(self, p):
         self.copyfrom_path = p.copyfrom_path
         self.copyfrom_rev = p.copyfrom_rev
         self.action = p.action
 
 def get_log_child(fp, url, paths, start, end, limit=0,
                   discover_changed_paths=True, strict_node_history=False):
     protocol = -1
     def receiver(orig_paths, revnum, author, date, message, pool):
         paths = {}
         if orig_paths is not None:
             for k, v in orig_paths.iteritems():
                 paths[k] = changedpath(v)
         pickle.dump((paths, revnum, author, date, message),
                     fp, protocol)
 
     try:
         # Use an ra of our own so that our parent can consume
         # our results without confusing the server.
         t = transport.SvnRaTransport(url=url)
         svn.ra.get_log(t.ra, paths, start, end, limit,
                        discover_changed_paths,
                        strict_node_history,
                        receiver)
     except IOError:
         # Caller may interrupt the iteration
         pickle.dump(None, fp, protocol)
     except Exception as inst:
         pickle.dump(str(inst), fp, protocol)
     else:
         pickle.dump(None, fp, protocol)
     fp.close()
     # With large history, cleanup process goes crazy and suddenly
     # consumes *huge* amount of memory. The output file being closed,
     # there is no need for clean termination.
     os._exit(0)
 
 def debugsvnlog(ui, **opts):
     """Fetch SVN log in a subprocess and channel them back to parent to
     avoid memory collection issues.
     """
     if svn is None:
         raise error.Abort(_('debugsvnlog could not load Subversion python '
                            'bindings'))
 
     args = decodeargs(ui.fin.read())
     get_log_child(ui.fout, *args)
 
 class logstream(object):
     """Interruptible revision log iterator."""
     def __init__(self, stdout):
         self._stdout = stdout
 
     def __iter__(self):
         while True:
             try:
                 entry = pickle.load(self._stdout)
             except EOFError:
                 raise error.Abort(_('Mercurial failed to run itself, check'
                                    ' hg executable is in PATH'))
             try:
                 orig_paths, revnum, author, date, message = entry
             except (TypeError, ValueError):
                 if entry is None:
                     break
                 raise error.Abort(_("log stream exception '%s'") % entry)
             yield entry
 
     def close(self):
         if self._stdout:
             self._stdout.close()
             self._stdout = None
 
 class directlogstream(list):
     """Direct revision log iterator.
     This can be used for debugging and development but it will probably leak
     memory and is not suitable for real conversions."""
     def __init__(self, url, paths, start, end, limit=0,
                   discover_changed_paths=True, strict_node_history=False):
 
         def receiver(orig_paths, revnum, author, date, message, pool):
             paths = {}
             if orig_paths is not None:
                 for k, v in orig_paths.iteritems():
                     paths[k] = changedpath(v)
             self.append((paths, revnum, author, date, message))
 
         # Use an ra of our own so that our parent can consume
         # our results without confusing the server.
         t = transport.SvnRaTransport(url=url)
         svn.ra.get_log(t.ra, paths, start, end, limit,
                        discover_changed_paths,
                        strict_node_history,
                        receiver)
 
     def close(self):
         pass
 
 # Check to see if the given path is a local Subversion repo. Verify this by
 # looking for several svn-specific files and directories in the given
 # directory.
 def filecheck(ui, path, proto):
     for x in ('locks', 'hooks', 'format', 'db'):
         if not os.path.exists(os.path.join(path, x)):
             return False
     return True
 
 # Check to see if a given path is the root of an svn repo over http. We verify
 # this by requesting a version-controlled URL we know can't exist and looking
 # for the svn-specific "not found" XML.
 def httpcheck(ui, path, proto):
     try:
         opener = urlreq.buildopener()
         rsp = opener.open('%s://%s/!svn/ver/0/.svn' % (proto, path), 'rb')
         data = rsp.read()
     except urlerr.httperror as inst:
         if inst.code != 404:
             # Except for 404 we cannot know for sure this is not an svn repo
             ui.warn(_('svn: cannot probe remote repository, assume it could '
                       'be a subversion repository. Use --source-type if you '
                       'know better.\n'))
             return True
         data = inst.fp.read()
     except Exception:
         # Could be urlerr.urlerror if the URL is invalid or anything else.
         return False
     return '<m:human-readable errcode="160013">' in data
 
 protomap = {'http': httpcheck,
             'https': httpcheck,
             'file': filecheck,
             }
 def issvnurl(ui, url):
     try:
         proto, path = url.split('://', 1)
         if proto == 'file':
             if (pycompat.iswindows and path[:1] == '/'
                   and path[1:2].isalpha() and path[2:6].lower() == '%3a/'):
                 path = path[:2] + ':/' + path[6:]
             path = urlreq.url2pathname(path)
     except ValueError:
         proto = 'file'
         path = os.path.abspath(url)
     if proto == 'file':
         path = util.pconvert(path)
     check = protomap.get(proto, lambda *args: False)
     while '/' in path:
         if check(ui, path, proto):
             return True
         path = path.rsplit('/', 1)[0]
     return False
 
 # SVN conversion code stolen from bzr-svn and tailor
 #
 # Subversion looks like a versioned filesystem, branches structures
 # are defined by conventions and not enforced by the tool. First,
 # we define the potential branches (modules) as "trunk" and "branches"
 # children directories. Revisions are then identified by their
 # module and revision number (and a repository identifier).
 #
 # The revision graph is really a tree (or a forest). By default, a
 # revision parent is the previous revision in the same module. If the
 # module directory is copied/moved from another module then the
 # revision is the module root and its parent the source revision in
 # the parent module. A revision has at most one parent.
 #
 class svn_source(converter_source):
     def __init__(self, ui, repotype, url, revs=None):
         super(svn_source, self).__init__(ui, repotype, url, revs=revs)
 
         if not (url.startswith('svn://') or url.startswith('svn+ssh://') or
                 (os.path.exists(url) and
                  os.path.exists(os.path.join(url, '.svn'))) or
                 issvnurl(ui, url)):
             raise NoRepo(_("%s does not look like a Subversion repository")
                          % url)
         if svn is None:
             raise MissingTool(_('could not load Subversion python bindings'))
 
         try:
             version = svn.core.SVN_VER_MAJOR, svn.core.SVN_VER_MINOR
             if version < (1, 4):
                 raise MissingTool(_('Subversion python bindings %d.%d found, '
                                     '1.4 or later required') % version)
         except AttributeError:
             raise MissingTool(_('Subversion python bindings are too old, 1.4 '
                                 'or later required'))
 
         self.lastrevs = {}
 
         latest = None
         try:
             # Support file://path@rev syntax. Useful e.g. to convert
             # deleted branches.
             at = url.rfind('@')
             if at >= 0:
                 latest = int(url[at + 1:])
                 url = url[:at]
         except ValueError:
             pass
         self.url = geturl(url)
         self.encoding = 'UTF-8' # Subversion is always nominal UTF-8
         try:
             self.transport = transport.SvnRaTransport(url=self.url)
             self.ra = self.transport.ra
             self.ctx = self.transport.client
             self.baseurl = svn.ra.get_repos_root(self.ra)
             # Module is either empty or a repository path starting with
             # a slash and not ending with a slash.
             self.module = urlreq.unquote(self.url[len(self.baseurl):])
             self.prevmodule = None
             self.rootmodule = self.module
             self.commits = {}
             self.paths = {}
             self.uuid = svn.ra.get_uuid(self.ra)
         except svn.core.SubversionException:
             ui.traceback()
             svnversion = '%d.%d.%d' % (svn.core.SVN_VER_MAJOR,
                                        svn.core.SVN_VER_MINOR,
                                        svn.core.SVN_VER_MICRO)
             raise NoRepo(_("%s does not look like a Subversion repository "
                            "to libsvn version %s")
                          % (self.url, svnversion))
 
         if revs:
             if len(revs) > 1:
                 raise error.Abort(_('subversion source does not support '
                                    'specifying multiple revisions'))
             try:
                 latest = int(revs[0])
             except ValueError:
                 raise error.Abort(_('svn: revision %s is not an integer') %
                                  revs[0])
 
         trunkcfg = self.ui.config('convert', 'svn.trunk')
         if trunkcfg is None:
             trunkcfg = 'trunk'
         self.trunkname = trunkcfg.strip('/')
         self.startrev = self.ui.config('convert', 'svn.startrev')
         try:
             self.startrev = int(self.startrev)
             if self.startrev < 0:
                 self.startrev = 0
         except ValueError:
             raise error.Abort(_('svn: start revision %s is not an integer')
                              % self.startrev)
 
         try:
             self.head = self.latest(self.module, latest)
         except SvnPathNotFound:
             self.head = None
         if not self.head:
             raise error.Abort(_('no revision found in module %s')
                              % self.module)
         self.last_changed = self.revnum(self.head)
 
         self._changescache = (None, None)
 
         if os.path.exists(os.path.join(url, '.svn/entries')):
             self.wc = url
         else:
             self.wc = None
         self.convertfp = None
 
     def setrevmap(self, revmap):
         lastrevs = {}
-        for revid in revmap.iterkeys():
+        for revid in revmap:
             uuid, module, revnum = revsplit(revid)
             lastrevnum = lastrevs.setdefault(module, revnum)
             if revnum > lastrevnum:
                 lastrevs[module] = revnum
         self.lastrevs = lastrevs
 
     def exists(self, path, optrev):
         try:
             svn.client.ls(self.url.rstrip('/') + '/' + quote(path),
                                  optrev, False, self.ctx)
             return True
         except svn.core.SubversionException:
             return False
 
     def getheads(self):
 
         def isdir(path, revnum):
             kind = self._checkpath(path, revnum)
             return kind == svn.core.svn_node_dir
 
         def getcfgpath(name, rev):
             cfgpath = self.ui.config('convert', 'svn.' + name)
             if cfgpath is not None and cfgpath.strip() == '':
                 return None
             path = (cfgpath or name).strip('/')
             if not self.exists(path, rev):
                 if self.module.endswith(path) and name == 'trunk':
                     # we are converting from inside this directory
                     return None
                 if cfgpath:
                     raise error.Abort(_('expected %s to be at %r, but not found'
                                        ) % (name, path))
                 return None
             self.ui.note(_('found %s at %r\n') % (name, path))
             return path
 
         rev = optrev(self.last_changed)
         oldmodule = ''
         trunk = getcfgpath('trunk', rev)
         self.tags = getcfgpath('tags', rev)
         branches = getcfgpath('branches', rev)
 
         # If the project has a trunk or branches, we will extract heads
         # from them. We keep the project root otherwise.
         if trunk:
             oldmodule = self.module or ''
             self.module += '/' + trunk
             self.head = self.latest(self.module, self.last_changed)
             if not self.head:
                 raise error.Abort(_('no revision found in module %s')
                                  % self.module)
 
         # First head in the list is the module's head
         self.heads = [self.head]
         if self.tags is not None:
             self.tags = '%s/%s' % (oldmodule , (self.tags or 'tags'))
 
         # Check if branches bring a few more heads to the list
         if branches:
             rpath = self.url.strip('/')
             branchnames = svn.client.ls(rpath + '/' + quote(branches),
                                         rev, False, self.ctx)
             for branch in sorted(branchnames):
                 module = '%s/%s/%s' % (oldmodule, branches, branch)
                 if not isdir(module, self.last_changed):
                     continue
                 brevid = self.latest(module, self.last_changed)
                 if not brevid:
                     self.ui.note(_('ignoring empty branch %s\n') % branch)
                     continue
                 self.ui.note(_('found branch %s at %d\n') %
                              (branch, self.revnum(brevid)))
                 self.heads.append(brevid)
 
         if self.startrev and self.heads:
             if len(self.heads) > 1:
                 raise error.Abort(_('svn: start revision is not supported '
                                    'with more than one branch'))
             revnum = self.revnum(self.heads[0])
             if revnum < self.startrev:
                 raise error.Abort(
                     _('svn: no revision found after start revision %d')
                                  % self.startrev)
 
         return self.heads
 
     def _getchanges(self, rev, full):
         (paths, parents) = self.paths[rev]
         copies = {}
         if parents:
             files, self.removed, copies = self.expandpaths(rev, paths, parents)
         if full or not parents:
             # Perform a full checkout on roots
             uuid, module, revnum = revsplit(rev)
             entries = svn.client.ls(self.baseurl + quote(module),
                                     optrev(revnum), True, self.ctx)
             files = [n for n, e in entries.iteritems()
                      if e.kind == svn.core.svn_node_file]
             self.removed = set()
 
         files.sort()
         files = zip(files, [rev] * len(files))
         return (files, copies)
 
     def getchanges(self, rev, full):
         # reuse cache from getchangedfiles
         if self._changescache[0] == rev and not full:
             (files, copies) = self._changescache[1]
         else:
             (files, copies) = self._getchanges(rev, full)
             # caller caches the result, so free it here to release memory
             del self.paths[rev]
         return (files, copies, set())
 
     def getchangedfiles(self, rev, i):
         # called from filemap - cache computed values for reuse in getchanges
         (files, copies) = self._getchanges(rev, False)
         self._changescache = (rev, (files, copies))
         return [f[0] for f in files]
 
     def getcommit(self, rev):
         if rev not in self.commits:
             uuid, module, revnum = revsplit(rev)
             self.module = module
             self.reparent(module)
             # We assume that:
             # - requests for revisions after "stop" come from the
             # revision graph backward traversal. Cache all of them
             # down to stop, they will be used eventually.
             # - requests for revisions before "stop" come to get
             # isolated branches parents. Just fetch what is needed.
             stop = self.lastrevs.get(module, 0)
             if revnum < stop:
                 stop = revnum + 1
             self._fetch_revisions(revnum, stop)
             if rev not in self.commits:
                 raise error.Abort(_('svn: revision %s not found') % revnum)
         revcommit = self.commits[rev]
         # caller caches the result, so free it here to release memory
         del self.commits[rev]
         return revcommit
 
     def checkrevformat(self, revstr, mapname='splicemap'):
         """ fails if revision format does not match the correct format"""
         if not re.match(r'svn:[0-9a-f]{8,8}-[0-9a-f]{4,4}-'
                               r'[0-9a-f]{4,4}-[0-9a-f]{4,4}-[0-9a-f]'
                               r'{12,12}(.*)\@[0-9]+$',revstr):
             raise error.Abort(_('%s entry %s is not a valid revision'
                                ' identifier') % (mapname, revstr))
 
     def numcommits(self):
         return int(self.head.rsplit('@', 1)[1]) - self.startrev
 
     def gettags(self):
         tags = {}
         if self.tags is None:
             return tags
 
         # svn tags are just a convention, project branches left in a
         # 'tags' directory. There is no other relationship than
         # ancestry, which is expensive to discover and makes them hard
         # to update incrementally.  Worse, past revisions may be
         # referenced by tags far away in the future, requiring a deep
         # history traversal on every calculation.  Current code
         # performs a single backward traversal, tracking moves within
         # the tags directory (tag renaming) and recording a new tag
         # everytime a project is copied from outside the tags
         # directory. It also lists deleted tags, this behaviour may
         # change in the future.
         pendings = []
         tagspath = self.tags
         start = svn.ra.get_latest_revnum(self.ra)
         stream = self._getlog([self.tags], start, self.startrev)
         try:
             for entry in stream:
                 origpaths, revnum, author, date, message = entry
                 if not origpaths:
                     origpaths = []
                 copies = [(e.copyfrom_path, e.copyfrom_rev, p) for p, e
                           in origpaths.iteritems() if e.copyfrom_path]
                 # Apply moves/copies from more specific to general
                 copies.sort(reverse=True)
 
                 srctagspath = tagspath
                 if copies and copies[-1][2] == tagspath:
                     # Track tags directory moves
                     srctagspath = copies.pop()[0]
 
                 for source, sourcerev, dest in copies:
                     if not dest.startswith(tagspath + '/'):
                         continue
                     for tag in pendings:
                         if tag[0].startswith(dest):
                             tagpath = source + tag[0][len(dest):]
                             tag[:2] = [tagpath, sourcerev]
                             break
                     else:
                         pendings.append([source, sourcerev, dest])
 
                 # Filter out tags with children coming from different
                 # parts of the repository like:
                 # /tags/tag.1 (from /trunk:10)
                 # /tags/tag.1/foo (from /branches/foo:12)
                 # Here/tags/tag.1 discarded as well as its children.
                 # It happens with tools like cvs2svn. Such tags cannot
                 # be represented in mercurial.
                 addeds = dict((p, e.copyfrom_path) for p, e
                               in origpaths.iteritems()
                               if e.action == 'A' and e.copyfrom_path)
                 badroots = set()
                 for destroot in addeds:
                     for source, sourcerev, dest in pendings:
                         if (not dest.startswith(destroot + '/')
                             or source.startswith(addeds[destroot] + '/')):
                             continue
                         badroots.add(destroot)
                         break
 
                 for badroot in badroots:
                     pendings = [p for p in pendings if p[2] != badroot
                                 and not p[2].startswith(badroot + '/')]
 
                 # Tell tag renamings from tag creations
                 renamings = []
                 for source, sourcerev, dest in pendings:
                     tagname = dest.split('/')[-1]
                     if source.startswith(srctagspath):
                         renamings.append([source, sourcerev, tagname])
                         continue
                     if tagname in tags:
                         # Keep the latest tag value
                         continue
                     # From revision may be fake, get one with changes
                     try:
                         tagid = self.latest(source, sourcerev)
                         if tagid and tagname not in tags:
                             tags[tagname] = tagid
                     except SvnPathNotFound:
                         # It happens when we are following directories
                         # we assumed were copied with their parents
                         # but were really created in the tag
                         # directory.
                         pass
                 pendings = renamings
                 tagspath = srctagspath
         finally:
             stream.close()
         return tags
 
     def converted(self, rev, destrev):
         if not self.wc:
             return
         if self.convertfp is None:
             self.convertfp = open(os.path.join(self.wc, '.svn', 'hg-shamap'),
                                   'ab')
         self.convertfp.write(util.tonativeeol('%s %d\n'
                                               % (destrev, self.revnum(rev))))
         self.convertfp.flush()
 
     def revid(self, revnum, module=None):
         return 'svn:%s%s@%s' % (self.uuid, module or self.module, revnum)
 
     def revnum(self, rev):
         return int(rev.split('@')[-1])
 
     def latest(self, path, stop=None):
         """Find the latest revid affecting path, up to stop revision
         number. If stop is None, default to repository latest
         revision. It may return a revision in a different module,
         since a branch may be moved without a change being
         reported. Return None if computed module does not belong to
         rootmodule subtree.
         """
         def findchanges(path, start, stop=None):
             stream = self._getlog([path], start, stop or 1)
             try:
                 for entry in stream:
                     paths, revnum, author, date, message = entry
                     if stop is None and paths:
                         # We do not know the latest changed revision,
                         # keep the first one with changed paths.
                         break
                     if revnum <= stop:
                         break
 
                     for p in paths:
                         if (not path.startswith(p) or
                             not paths[p].copyfrom_path):
                             continue
                         newpath = paths[p].copyfrom_path + path[len(p):]
                         self.ui.debug("branch renamed from %s to %s at %d\n" %
                                       (path, newpath, revnum))
                         path = newpath
                         break
                 if not paths:
                     revnum = None
                 return revnum, path
             finally:
                 stream.close()
 
         if not path.startswith(self.rootmodule):
             # Requests on foreign branches may be forbidden at server level
             self.ui.debug('ignoring foreign branch %r\n' % path)
             return None
 
         if stop is None:
             stop = svn.ra.get_latest_revnum(self.ra)
         try:
             prevmodule = self.reparent('')
             dirent = svn.ra.stat(self.ra, path.strip('/'), stop)
             self.reparent(prevmodule)
         except svn.core.SubversionException:
             dirent = None
         if not dirent:
             raise SvnPathNotFound(_('%s not found up to revision %d')
                                   % (path, stop))
 
         # stat() gives us the previous revision on this line of
         # development, but it might be in *another module*. Fetch the
         # log and detect renames down to the latest revision.
         revnum, realpath = findchanges(path, stop, dirent.created_rev)
         if revnum is None:
             # Tools like svnsync can create empty revision, when
             # synchronizing only a subtree for instance. These empty
             # revisions created_rev still have their original values
             # despite all changes having disappeared and can be
             # returned by ra.stat(), at least when stating the root
             # module. In that case, do not trust created_rev and scan
             # the whole history.
             revnum, realpath = findchanges(path, stop)
             if revnum is None:
                 self.ui.debug('ignoring empty branch %r\n' % realpath)
                 return None
 
         if not realpath.startswith(self.rootmodule):
             self.ui.debug('ignoring foreign branch %r\n' % realpath)
             return None
         return self.revid(revnum, realpath)
 
     def reparent(self, module):
         """Reparent the svn transport and return the previous parent."""
         if self.prevmodule == module:
             return module
         svnurl = self.baseurl + quote(module)
         prevmodule = self.prevmodule
         if prevmodule is None:
             prevmodule = ''
         self.ui.debug("reparent to %s\n" % svnurl)
         svn.ra.reparent(self.ra, svnurl)
         self.prevmodule = module
         return prevmodule
 
     def expandpaths(self, rev, paths, parents):
         changed, removed = set(), set()
         copies = {}
 
         new_module, revnum = revsplit(rev)[1:]
         if new_module != self.module:
             self.module = new_module
             self.reparent(self.module)
 
         for i, (path, ent) in enumerate(paths):
             self.ui.progress(_('scanning paths'), i, item=path,
                              total=len(paths), unit=_('paths'))
             entrypath = self.getrelpath(path)
 
             kind = self._checkpath(entrypath, revnum)
             if kind == svn.core.svn_node_file:
                 changed.add(self.recode(entrypath))
                 if not ent.copyfrom_path or not parents:
                     continue
                 # Copy sources not in parent revisions cannot be
                 # represented, ignore their origin for now
                 pmodule, prevnum = revsplit(parents[0])[1:]
                 if ent.copyfrom_rev < prevnum:
                     continue
                 copyfrom_path = self.getrelpath(ent.copyfrom_path, pmodule)
                 if not copyfrom_path:
                     continue
                 self.ui.debug("copied to %s from %s@%s\n" %
                               (entrypath, copyfrom_path, ent.copyfrom_rev))
                 copies[self.recode(entrypath)] = self.recode(copyfrom_path)
             elif kind == 0: # gone, but had better be a deleted *file*
                 self.ui.debug("gone from %s\n" % ent.copyfrom_rev)
                 pmodule, prevnum = revsplit(parents[0])[1:]
                 parentpath = pmodule + "/" + entrypath
                 fromkind = self._checkpath(entrypath, prevnum, pmodule)
 
                 if fromkind == svn.core.svn_node_file:
                     removed.add(self.recode(entrypath))
                 elif fromkind == svn.core.svn_node_dir:
                     oroot = parentpath.strip('/')
                     nroot = path.strip('/')
                     children = self._iterfiles(oroot, prevnum)
                     for childpath in children:
                         childpath = childpath.replace(oroot, nroot)
                         childpath = self.getrelpath("/" + childpath, pmodule)
                         if childpath:
                             removed.add(self.recode(childpath))
                 else:
                     self.ui.debug('unknown path in revision %d: %s\n' % \
                                   (revnum, path))
             elif kind == svn.core.svn_node_dir:
                 if ent.action == 'M':
                     # If the directory just had a prop change,
                     # then we shouldn't need to look for its children.
                     continue
                 if ent.action == 'R' and parents:
                     # If a directory is replacing a file, mark the previous
                     # file as deleted
                     pmodule, prevnum = revsplit(parents[0])[1:]
                     pkind = self._checkpath(entrypath, prevnum, pmodule)
                     if pkind == svn.core.svn_node_file:
                         removed.add(self.recode(entrypath))
                     elif pkind == svn.core.svn_node_dir:
                         # We do not know what files were kept or removed,
                         # mark them all as changed.
                         for childpath in self._iterfiles(pmodule, prevnum):
                             childpath = self.getrelpath("/" + childpath)
                             if childpath:
                                 changed.add(self.recode(childpath))
 
                 for childpath in self._iterfiles(path, revnum):
                     childpath = self.getrelpath("/" + childpath)
                     if childpath:
                         changed.add(self.recode(childpath))
 
                 # Handle directory copies
                 if not ent.copyfrom_path or not parents:
                     continue
                 # Copy sources not in parent revisions cannot be
                 # represented, ignore their origin for now
                 pmodule, prevnum = revsplit(parents[0])[1:]
                 if ent.copyfrom_rev < prevnum:
                     continue
                 copyfrompath = self.getrelpath(ent.copyfrom_path, pmodule)
                 if not copyfrompath:
                     continue
                 self.ui.debug("mark %s came from %s:%d\n"
                               % (path, copyfrompath, ent.copyfrom_rev))
                 children = self._iterfiles(ent.copyfrom_path, ent.copyfrom_rev)
                 for childpath in children:
                     childpath = self.getrelpath("/" + childpath, pmodule)
                     if not childpath:
                         continue
                     copytopath = path + childpath[len(copyfrompath):]
                     copytopath = self.getrelpath(copytopath)
                     copies[self.recode(copytopath)] = self.recode(childpath)
 
         self.ui.progress(_('scanning paths'), None)
         changed.update(removed)
         return (list(changed), removed, copies)
 
     def _fetch_revisions(self, from_revnum, to_revnum):
         if from_revnum < to_revnum:
             from_revnum, to_revnum = to_revnum, from_revnum
 
         self.child_cset = None
 
         def parselogentry(orig_paths, revnum, author, date, message):
             """Return the parsed commit object or None, and True if
             the revision is a branch root.
             """
             self.ui.debug("parsing revision %d (%d changes)\n" %
                           (revnum, len(orig_paths)))
 
             branched = False
             rev = self.revid(revnum)
             # branch log might return entries for a parent we already have
 
             if rev in self.commits or revnum < to_revnum:
                 return None, branched
 
             parents = []
             # check whether this revision is the start of a branch or part
             # of a branch renaming
             orig_paths = sorted(orig_paths.iteritems())
             root_paths = [(p, e) for p, e in orig_paths
                           if self.module.startswith(p)]
             if root_paths:
                 path, ent = root_paths[-1]
                 if ent.copyfrom_path:
                     branched = True
                     newpath = ent.copyfrom_path + self.module[len(path):]
                     # ent.copyfrom_rev may not be the actual last revision
                     previd = self.latest(newpath, ent.copyfrom_rev)
                     if previd is not None:
                         prevmodule, prevnum = revsplit(previd)[1:]
                         if prevnum >= self.startrev:
                             parents = [previd]
                             self.ui.note(
                                 _('found parent of branch %s at %d: %s\n') %
                                 (self.module, prevnum, prevmodule))
                 else:
                     self.ui.debug("no copyfrom path, don't know what to do.\n")
 
             paths = []
             # filter out unrelated paths
             for path, ent in orig_paths:
                 if self.getrelpath(path) is None:
                     continue
                 paths.append((path, ent))
 
             # Example SVN datetime. Includes microseconds.
             # ISO-8601 conformant
             # '2007-01-04T17:35:00.902377Z'
             date = util.parsedate(date[:19] + " UTC", ["%Y-%m-%dT%H:%M:%S"])
             if self.ui.configbool('convert', 'localtimezone'):
                 date = makedatetimestamp(date[0])
 
             if message:
                 log = self.recode(message)
             else:
                 log = ''
 
             if author:
                 author = self.recode(author)
             else:
                 author = ''
 
             try:
                 branch = self.module.split("/")[-1]
                 if branch == self.trunkname:
                     branch = None
             except IndexError:
                 branch = None
 
             cset = commit(author=author,
                           date=util.datestr(date, '%Y-%m-%d %H:%M:%S %1%2'),
                           desc=log,
                           parents=parents,
                           branch=branch,
                           rev=rev)
 
             self.commits[rev] = cset
             # The parents list is *shared* among self.paths and the
             # commit object. Both will be updated below.
             self.paths[rev] = (paths, cset.parents)
             if self.child_cset and not self.child_cset.parents:
                 self.child_cset.parents[:] = [rev]
             self.child_cset = cset
             return cset, branched
 
         self.ui.note(_('fetching revision log for "%s" from %d to %d\n') %
                      (self.module, from_revnum, to_revnum))
 
         try:
             firstcset = None
             lastonbranch = False
             stream = self._getlog([self.module], from_revnum, to_revnum)
             try:
                 for entry in stream:
                     paths, revnum, author, date, message = entry
                     if revnum < self.startrev:
                         lastonbranch = True
                         break
                     if not paths:
                         self.ui.debug('revision %d has no entries\n' % revnum)
                         # If we ever leave the loop on an empty
                         # revision, do not try to get a parent branch
                         lastonbranch = lastonbranch or revnum == 0
                         continue
                     cset, lastonbranch = parselogentry(paths, revnum, author,
                                                        date, message)
                     if cset:
                         firstcset = cset
                     if lastonbranch:
                         break
             finally:
                 stream.close()
 
             if not lastonbranch and firstcset and not firstcset.parents:
                 # The first revision of the sequence (the last fetched one)
                 # has invalid parents if not a branch root. Find the parent
                 # revision now, if any.
                 try:
                     firstrevnum = self.revnum(firstcset.rev)
                     if firstrevnum > 1:
                         latest = self.latest(self.module, firstrevnum - 1)
                         if latest:
                             firstcset.parents.append(latest)
                 except SvnPathNotFound:
                     pass
         except svn.core.SubversionException as xxx_todo_changeme:
             (inst, num) = xxx_todo_changeme.args
             if num == svn.core.SVN_ERR_FS_NO_SUCH_REVISION:
                 raise error.Abort(_('svn: branch has no revision %s')
                                  % to_revnum)
             raise
 
     def getfile(self, file, rev):
         # TODO: ra.get_file transmits the whole file instead of diffs.
         if file in self.removed:
             return None, None
         mode = ''
         try:
             new_module, revnum = revsplit(rev)[1:]
             if self.module != new_module:
                 self.module = new_module
                 self.reparent(self.module)
             io = stringio()
             info = svn.ra.get_file(self.ra, file, revnum, io)
             data = io.getvalue()
             # ra.get_file() seems to keep a reference on the input buffer
             # preventing collection. Release it explicitly.
             io.close()
             if isinstance(info, list):
                 info = info[-1]
             mode = ("svn:executable" in info) and 'x' or ''
             mode = ("svn:special" in info) and 'l' or mode
         except svn.core.SubversionException as e:
             notfound = (svn.core.SVN_ERR_FS_NOT_FOUND,
                 svn.core.SVN_ERR_RA_DAV_PATH_NOT_FOUND)
             if e.apr_err in notfound: # File not found
                 return None, None
             raise
         if mode == 'l':
             link_prefix = "link "
             if data.startswith(link_prefix):
                 data = data[len(link_prefix):]
         return data, mode
 
     def _iterfiles(self, path, revnum):
         """Enumerate all files in path at revnum, recursively."""
         path = path.strip('/')
         pool = svn.core.Pool()
         rpath = '/'.join([self.baseurl, quote(path)]).strip('/')
         entries = svn.client.ls(rpath, optrev(revnum), True, self.ctx, pool)
         if path:
             path += '/'
         return ((path + p) for p, e in entries.iteritems()
                 if e.kind == svn.core.svn_node_file)
 
     def getrelpath(self, path, module=None):
         if module is None:
             module = self.module
         # Given the repository url of this wc, say
         #   "http://server/plone/CMFPlone/branches/Plone-2_0-branch"
         # extract the "entry" portion (a relative path) from what
         # svn log --xml says, i.e.
         #   "/CMFPlone/branches/Plone-2_0-branch/tests/PloneTestCase.py"
         # that is to say "tests/PloneTestCase.py"
         if path.startswith(module):
             relative = path.rstrip('/')[len(module):]
             if relative.startswith('/'):
                 return relative[1:]
             elif relative == '':
                 return relative
 
         # The path is outside our tracked tree...
         self.ui.debug('%r is not under %r, ignoring\n' % (path, module))
         return None
 
     def _checkpath(self, path, revnum, module=None):
         if module is not None:
             prevmodule = self.reparent('')
             path = module + '/' + path
         try:
             # ra.check_path does not like leading slashes very much, it leads
             # to PROPFIND subversion errors
             return svn.ra.check_path(self.ra, path.strip('/'), revnum)
         finally:
             if module is not None:
                 self.reparent(prevmodule)
 
     def _getlog(self, paths, start, end, limit=0, discover_changed_paths=True,
                 strict_node_history=False):
         # Normalize path names, svn >= 1.5 only wants paths relative to
         # supplied URL
         relpaths = []
         for p in paths:
             if not p.startswith('/'):
                 p = self.module + '/' + p
             relpaths.append(p.strip('/'))
         args = [self.baseurl, relpaths, start, end, limit,
                 discover_changed_paths, strict_node_history]
         # developer config: convert.svn.debugsvnlog
         if not self.ui.configbool('convert', 'svn.debugsvnlog'):
             return directlogstream(*args)
         arg = encodeargs(args)
         hgexe = util.hgexecutable()
         cmd = '%s debugsvnlog' % util.shellquote(hgexe)
         stdin, stdout = util.popen2(util.quotecommand(cmd))
         stdin.write(arg)
         try:
             stdin.close()
         except IOError:
             raise error.Abort(_('Mercurial failed to run itself, check'
                                ' hg executable is in PATH'))
         return logstream(stdout)
 
 pre_revprop_change = '''#!/bin/sh
 
 REPOS="$1"
 REV="$2"
 USER="$3"
 PROPNAME="$4"
 ACTION="$5"
 
 if [ "$ACTION" = "M" -a "$PROPNAME" = "svn:log" ]; then exit 0; fi
 if [ "$ACTION" = "A" -a "$PROPNAME" = "hg:convert-branch" ]; then exit 0; fi
 if [ "$ACTION" = "A" -a "$PROPNAME" = "hg:convert-rev" ]; then exit 0; fi
 
 echo "Changing prohibited revision property" >&2
 exit 1
 '''
 
 class svn_sink(converter_sink, commandline):
     commit_re = re.compile(r'Committed revision (\d+).', re.M)
     uuid_re = re.compile(r'Repository UUID:\s*(\S+)', re.M)
 
     def prerun(self):
         if self.wc:
             os.chdir(self.wc)
 
     def postrun(self):
         if self.wc:
             os.chdir(self.cwd)
 
     def join(self, name):
         return os.path.join(self.wc, '.svn', name)
 
     def revmapfile(self):
         return self.join('hg-shamap')
 
     def authorfile(self):
         return self.join('hg-authormap')
 
     def __init__(self, ui, repotype, path):
 
         converter_sink.__init__(self, ui, repotype, path)
         commandline.__init__(self, ui, 'svn')
         self.delete = []
         self.setexec = []
         self.delexec = []
         self.copies = []
         self.wc = None
         self.cwd = pycompat.getcwd()
 
         created = False
         if os.path.isfile(os.path.join(path, '.svn', 'entries')):
             self.wc = os.path.realpath(path)
             self.run0('update')
         else:
             if not re.search(r'^(file|http|https|svn|svn\+ssh)\://', path):
                 path = os.path.realpath(path)
                 if os.path.isdir(os.path.dirname(path)):
                     if not os.path.exists(os.path.join(path, 'db', 'fs-type')):
                         ui.status(_('initializing svn repository %r\n') %
                                   os.path.basename(path))
                         commandline(ui, 'svnadmin').run0('create', path)
                         created = path
                     path = util.normpath(path)
                     if not path.startswith('/'):
                         path = '/' + path
                     path = 'file://' + path
 
             wcpath = os.path.join(pycompat.getcwd(), os.path.basename(path) +
                                 '-wc')
             ui.status(_('initializing svn working copy %r\n')
                       % os.path.basename(wcpath))
             self.run0('checkout', path, wcpath)
 
             self.wc = wcpath
         self.opener = vfsmod.vfs(self.wc)
         self.wopener = vfsmod.vfs(self.wc)
         self.childmap = mapfile(ui, self.join('hg-childmap'))
         if util.checkexec(self.wc):
             self.is_exec = util.isexec
         else:
             self.is_exec = None
 
         if created:
             hook = os.path.join(created, 'hooks', 'pre-revprop-change')
             fp = open(hook, 'wb')
             fp.write(pre_revprop_change)
             fp.close()
             util.setflags(hook, False, True)
 
         output = self.run0('info')
         self.uuid = self.uuid_re.search(output).group(1).strip()
 
     def wjoin(self, *names):
         return os.path.join(self.wc, *names)
 
     @propertycache
     def manifest(self):
         # As of svn 1.7, the "add" command fails when receiving
         # already tracked entries, so we have to track and filter them
         # ourselves.
         m = set()
         output = self.run0('ls', recursive=True, xml=True)
         doc = xml.dom.minidom.parseString(output)
         for e in doc.getElementsByTagName('entry'):
             for n in e.childNodes:
                 if n.nodeType != n.ELEMENT_NODE or n.tagName != 'name':
                     continue
                 name = ''.join(c.data for c in n.childNodes
                                if c.nodeType == c.TEXT_NODE)
                 # Entries are compared with names coming from
                 # mercurial, so bytes with undefined encoding. Our
                 # best bet is to assume they are in local
                 # encoding. They will be passed to command line calls
                 # later anyway, so they better be.
                 m.add(encoding.unitolocal(name))
                 break
         return m
 
     def putfile(self, filename, flags, data):
         if 'l' in flags:
             self.wopener.symlink(data, filename)
         else:
             try:
                 if os.path.islink(self.wjoin(filename)):
                     os.unlink(filename)
             except OSError:
                 pass
             self.wopener.write(filename, data)
 
             if self.is_exec:
                 if self.is_exec(self.wjoin(filename)):
                     if 'x' not in flags:
                         self.delexec.append(filename)
                 else:
                     if 'x' in flags:
                         self.setexec.append(filename)
                 util.setflags(self.wjoin(filename), False, 'x' in flags)
 
     def _copyfile(self, source, dest):
         # SVN's copy command pukes if the destination file exists, but
         # our copyfile method expects to record a copy that has
         # already occurred.  Cross the semantic gap.
         wdest = self.wjoin(dest)
         exists = os.path.lexists(wdest)
         if exists:
             fd, tempname = tempfile.mkstemp(
                 prefix='hg-copy-', dir=os.path.dirname(wdest))
             os.close(fd)
             os.unlink(tempname)
             os.rename(wdest, tempname)
         try:
             self.run0('copy', source, dest)
         finally:
             self.manifest.add(dest)
             if exists:
                 try:
                     os.unlink(wdest)
                 except OSError:
                     pass
                 os.rename(tempname, wdest)
 
     def dirs_of(self, files):
         dirs = set()
         for f in files:
             if os.path.isdir(self.wjoin(f)):
                 dirs.add(f)
             i = len(f)
             for i in iter(lambda: f.rfind('/', 0, i), -1):
                 dirs.add(f[:i])
         return dirs
 
     def add_dirs(self, files):
         add_dirs = [d for d in sorted(self.dirs_of(files))
                     if d not in self.manifest]
         if add_dirs:
             self.manifest.update(add_dirs)
             self.xargs(add_dirs, 'add', non_recursive=True, quiet=True)
         return add_dirs
 
     def add_files(self, files):
         files = [f for f in files if f not in self.manifest]
         if files:
             self.manifest.update(files)
             self.xargs(files, 'add', quiet=True)
         return files
 
     def addchild(self, parent, child):
         self.childmap[parent] = child
 
     def revid(self, rev):
         return u"svn:%s@%s" % (self.uuid, rev)
 
     def putcommit(self, files, copies, parents, commit, source, revmap, full,
                   cleanp2):
         for parent in parents:
             try:
                 return self.revid(self.childmap[parent])
             except KeyError:
                 pass
 
         # Apply changes to working copy
         for f, v in files:
             data, mode = source.getfile(f, v)
             if data is None:
                 self.delete.append(f)
             else:
                 self.putfile(f, mode, data)
                 if f in copies:
                     self.copies.append([copies[f], f])
         if full:
             self.delete.extend(sorted(self.manifest.difference(files)))
         files = [f[0] for f in files]
 
         entries = set(self.delete)
         files = frozenset(files)
         entries.update(self.add_dirs(files.difference(entries)))
         if self.copies:
             for s, d in self.copies:
                 self._copyfile(s, d)
             self.copies = []
         if self.delete:
             self.xargs(self.delete, 'delete')
             for f in self.delete:
                 self.manifest.remove(f)
             self.delete = []
         entries.update(self.add_files(files.difference(entries)))
         if self.delexec:
             self.xargs(self.delexec, 'propdel', 'svn:executable')
             self.delexec = []
         if self.setexec:
             self.xargs(self.setexec, 'propset', 'svn:executable', '*')
             self.setexec = []
 
         fd, messagefile = tempfile.mkstemp(prefix='hg-convert-')
         fp = os.fdopen(fd, pycompat.sysstr('wb'))
         fp.write(util.tonativeeol(commit.desc))
         fp.close()
         try:
             output = self.run0('commit',
                                username=util.shortuser(commit.author),
                                file=messagefile,
                                encoding='utf-8')
             try:
                 rev = self.commit_re.search(output).group(1)
             except AttributeError:
                 if parents and not files:
                     return parents[0]
                 self.ui.warn(_('unexpected svn output:\n'))
                 self.ui.warn(output)
                 raise error.Abort(_('unable to cope with svn output'))
             if commit.rev:
                 self.run('propset', 'hg:convert-rev', commit.rev,
                          revprop=True, revision=rev)
             if commit.branch and commit.branch != 'default':
                 self.run('propset', 'hg:convert-branch', commit.branch,
                          revprop=True, revision=rev)
             for parent in parents:
                 self.addchild(parent, rev)
             return self.revid(rev)
         finally:
             os.unlink(messagefile)
 
     def puttags(self, tags):
         self.ui.warn(_('writing Subversion tags is not yet implemented\n'))
         return None, None
 
     def hascommitfrommap(self, rev):
         # We trust that revisions referenced in a map still is present
         # TODO: implement something better if necessary and feasible
         return True
 
     def hascommitforsplicemap(self, rev):
         # This is not correct as one can convert to an existing subversion
         # repository and childmap would not list all revisions. Too bad.
         if rev in self.childmap:
             return True
         raise error.Abort(_('splice map revision %s not found in subversion '
                            'child map (revision lookups are not implemented)')
                          % rev)

hgext/rebase.py

@@ -1,1870 +1,1870
 # rebase.py - rebasing feature for mercurial
 #
 # Copyright 2008 Stefano Tortarolo <stefano.tortarolo at gmail dot com>
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 '''command to move sets of revisions to a different ancestor
 
 This extension lets you rebase changesets in an existing Mercurial
 repository.
 
 For more information:
 https://mercurial-scm.org/wiki/RebaseExtension
 '''
 
 from __future__ import absolute_import
 
 import errno
 import os
 
 from mercurial.i18n import _
 from mercurial.node import (
     nullid,
     nullrev,
     short,
 )
 from mercurial import (
     bookmarks,
     cmdutil,
     commands,
     copies,
     destutil,
     dirstateguard,
     error,
     extensions,
     hg,
     lock,
     merge as mergemod,
     mergeutil,
     obsolete,
     obsutil,
     patch,
     phases,
     pycompat,
     registrar,
     repair,
     revset,
     revsetlang,
     scmutil,
     smartset,
     util,
 )
 
 release = lock.release
 
 # The following constants are used throughout the rebase module. The ordering of
 # their values must be maintained.
 
 # Indicates that a revision needs to be rebased
 revtodo = -1
 revtodostr = '-1'
 
 # legacy revstates no longer needed in current code
 # -2: nullmerge, -3: revignored, -4: revprecursor, -5: revpruned
 legacystates = {'-2', '-3', '-4', '-5'}
 
 cmdtable = {}
 command = registrar.command(cmdtable)
 # Note for extension authors: ONLY specify testedwith = 'ships-with-hg-core' for
 # extensions which SHIP WITH MERCURIAL. Non-mainline extensions should
 # be specifying the version(s) of Mercurial they are tested with, or
 # leave the attribute unspecified.
 testedwith = 'ships-with-hg-core'
 
 def _nothingtorebase():
     return 1
 
 def _savegraft(ctx, extra):
     s = ctx.extra().get('source', None)
     if s is not None:
         extra['source'] = s
     s = ctx.extra().get('intermediate-source', None)
     if s is not None:
         extra['intermediate-source'] = s
 
 def _savebranch(ctx, extra):
     extra['branch'] = ctx.branch()
 
 def _makeextrafn(copiers):
     """make an extrafn out of the given copy-functions.
 
     A copy function takes a context and an extra dict, and mutates the
     extra dict as needed based on the given context.
     """
     def extrafn(ctx, extra):
         for c in copiers:
             c(ctx, extra)
     return extrafn
 
 def _destrebase(repo, sourceset, destspace=None):
     """small wrapper around destmerge to pass the right extra args
 
     Please wrap destutil.destmerge instead."""
     return destutil.destmerge(repo, action='rebase', sourceset=sourceset,
                               onheadcheck=False, destspace=destspace)
 
 revsetpredicate = registrar.revsetpredicate()
 
 @revsetpredicate('_destrebase')
 def _revsetdestrebase(repo, subset, x):
     # ``_rebasedefaultdest()``
 
     # default destination for rebase.
     # # XXX: Currently private because I expect the signature to change.
     # # XXX: - bailing out in case of ambiguity vs returning all data.
     # i18n: "_rebasedefaultdest" is a keyword
     sourceset = None
     if x is not None:
         sourceset = revset.getset(repo, smartset.fullreposet(repo), x)
     return subset & smartset.baseset([_destrebase(repo, sourceset)])
 
 def _ctxdesc(ctx):
     """short description for a context"""
     desc = '%d:%s "%s"' % (ctx.rev(), ctx,
                            ctx.description().split('\n', 1)[0])
     repo = ctx.repo()
     names = []
     for nsname, ns in repo.names.iteritems():
         if nsname == 'branches':
             continue
         names.extend(ns.names(repo, ctx.node()))
     if names:
         desc += ' (%s)' % ' '.join(names)
     return desc
 
 class rebaseruntime(object):
     """This class is a container for rebase runtime state"""
     def __init__(self, repo, ui, inmemory=False, opts=None):
         if opts is None:
             opts = {}
 
         # prepared: whether we have rebasestate prepared or not. Currently it
         # decides whether "self.repo" is unfiltered or not.
         # The rebasestate has explicit hash to hash instructions not depending
         # on visibility. If rebasestate exists (in-memory or on-disk), use
         # unfiltered repo to avoid visibility issues.
         # Before knowing rebasestate (i.e. when starting a new rebase (not
         # --continue or --abort)), the original repo should be used so
         # visibility-dependent revsets are correct.
         self.prepared = False
         self._repo = repo
 
         self.ui = ui
         self.opts = opts
         self.originalwd = None
         self.external = nullrev
         # Mapping between the old revision id and either what is the new rebased
         # revision or what needs to be done with the old revision. The state
         # dict will be what contains most of the rebase progress state.
         self.state = {}
         self.activebookmark = None
         self.destmap = {}
         self.skipped = set()
 
         self.collapsef = opts.get('collapse', False)
         self.collapsemsg = cmdutil.logmessage(ui, opts)
         self.date = opts.get('date', None)
 
         e = opts.get('extrafn') # internal, used by e.g. hgsubversion
         self.extrafns = [_savegraft]
         if e:
             self.extrafns = [e]
 
         self.keepf = opts.get('keep', False)
         self.keepbranchesf = opts.get('keepbranches', False)
         # keepopen is not meant for use on the command line, but by
         # other extensions
         self.keepopen = opts.get('keepopen', False)
         self.obsoletenotrebased = {}
         self.obsoletewithoutsuccessorindestination = set()
         self.inmemory = inmemory
 
     @property
     def repo(self):
         if self.prepared:
             return self._repo.unfiltered()
         else:
             return self._repo
 
     def storestatus(self, tr=None):
         """Store the current status to allow recovery"""
         if tr:
             tr.addfilegenerator('rebasestate', ('rebasestate',),
                                 self._writestatus, location='plain')
         else:
             with self.repo.vfs("rebasestate", "w") as f:
                 self._writestatus(f)
 
     def _writestatus(self, f):
         repo = self.repo
         assert repo.filtername is None
         f.write(repo[self.originalwd].hex() + '\n')
         # was "dest". we now write dest per src root below.
         f.write('\n')
         f.write(repo[self.external].hex() + '\n')
         f.write('%d\n' % int(self.collapsef))
         f.write('%d\n' % int(self.keepf))
         f.write('%d\n' % int(self.keepbranchesf))
         f.write('%s\n' % (self.activebookmark or ''))
         destmap = self.destmap
         for d, v in self.state.iteritems():
             oldrev = repo[d].hex()
             if v >= 0:
                 newrev = repo[v].hex()
             else:
                 newrev = "%d" % v
             destnode = repo[destmap[d]].hex()
             f.write("%s:%s:%s\n" % (oldrev, newrev, destnode))
         repo.ui.debug('rebase status stored\n')
 
     def restorestatus(self):
         """Restore a previously stored status"""
         self.prepared = True
         repo = self.repo
         assert repo.filtername is None
         keepbranches = None
         legacydest = None
         collapse = False
         external = nullrev
         activebookmark = None
         state = {}
         destmap = {}
 
         try:
             f = repo.vfs("rebasestate")
             for i, l in enumerate(f.read().splitlines()):
                 if i == 0:
                     originalwd = repo[l].rev()
                 elif i == 1:
                     # this line should be empty in newer version. but legacy
                     # clients may still use it
                     if l:
                         legacydest = repo[l].rev()
                 elif i == 2:
                     external = repo[l].rev()
                 elif i == 3:
                     collapse = bool(int(l))
                 elif i == 4:
                     keep = bool(int(l))
                 elif i == 5:
                     keepbranches = bool(int(l))
                 elif i == 6 and not (len(l) == 81 and ':' in l):
                     # line 6 is a recent addition, so for backwards
                     # compatibility check that the line doesn't look like the
                     # oldrev:newrev lines
                     activebookmark = l
                 else:
                     args = l.split(':')
                     oldrev = args[0]
                     newrev = args[1]
                     if newrev in legacystates:
                         continue
                     if len(args) > 2:
                         destnode = args[2]
                     else:
                         destnode = legacydest
                     destmap[repo[oldrev].rev()] = repo[destnode].rev()
                     if newrev in (nullid, revtodostr):
                         state[repo[oldrev].rev()] = revtodo
                         # Legacy compat special case
                     else:
                         state[repo[oldrev].rev()] = repo[newrev].rev()
 
         except IOError as err:
             if err.errno != errno.ENOENT:
                 raise
             cmdutil.wrongtooltocontinue(repo, _('rebase'))
 
         if keepbranches is None:
             raise error.Abort(_('.hg/rebasestate is incomplete'))
 
         skipped = set()
         # recompute the set of skipped revs
         if not collapse:
             seen = set(destmap.values())
             for old, new in sorted(state.items()):
                 if new != revtodo and new in seen:
                     skipped.add(old)
                 seen.add(new)
         repo.ui.debug('computed skipped revs: %s\n' %
                         (' '.join('%d' % r for r in sorted(skipped)) or ''))
         repo.ui.debug('rebase status resumed\n')
 
         self.originalwd = originalwd
         self.destmap = destmap
         self.state = state
         self.skipped = skipped
         self.collapsef = collapse
         self.keepf = keep
         self.keepbranchesf = keepbranches
         self.external = external
         self.activebookmark = activebookmark
 
     def _handleskippingobsolete(self, obsoleterevs, destmap):
         """Compute structures necessary for skipping obsolete revisions
 
         obsoleterevs:   iterable of all obsolete revisions in rebaseset
         destmap:        {srcrev: destrev} destination revisions
         """
         self.obsoletenotrebased = {}
         if not self.ui.configbool('experimental', 'rebaseskipobsolete'):
             return
         obsoleteset = set(obsoleterevs)
         (self.obsoletenotrebased,
          self.obsoletewithoutsuccessorindestination,
          obsoleteextinctsuccessors) = _computeobsoletenotrebased(
              self.repo, obsoleteset, destmap)
         skippedset = set(self.obsoletenotrebased)
         skippedset.update(self.obsoletewithoutsuccessorindestination)
         skippedset.update(obsoleteextinctsuccessors)
         _checkobsrebase(self.repo, self.ui, obsoleteset, skippedset)
 
     def _prepareabortorcontinue(self, isabort):
         try:
             self.restorestatus()
             self.collapsemsg = restorecollapsemsg(self.repo, isabort)
         except error.RepoLookupError:
             if isabort:
                 clearstatus(self.repo)
                 clearcollapsemsg(self.repo)
                 self.repo.ui.warn(_('rebase aborted (no revision is removed,'
                                     ' only broken state is cleared)\n'))
                 return 0
             else:
                 msg = _('cannot continue inconsistent rebase')
                 hint = _('use "hg rebase --abort" to clear broken state')
                 raise error.Abort(msg, hint=hint)
         if isabort:
             return abort(self.repo, self.originalwd, self.destmap,
                          self.state, activebookmark=self.activebookmark)
 
     def _preparenewrebase(self, destmap):
         if not destmap:
             return _nothingtorebase()
 
         rebaseset = destmap.keys()
         allowunstable = obsolete.isenabled(self.repo, obsolete.allowunstableopt)
         if (not (self.keepf or allowunstable)
               and self.repo.revs('first(children(%ld) - %ld)',
                                  rebaseset, rebaseset)):
             raise error.Abort(
                 _("can't remove original changesets with"
                   " unrebased descendants"),
                 hint=_('use --keep to keep original changesets'))
 
         result = buildstate(self.repo, destmap, self.collapsef)
 
         if not result:
             # Empty state built, nothing to rebase
             self.ui.status(_('nothing to rebase\n'))
             return _nothingtorebase()
 
         for root in self.repo.set('roots(%ld)', rebaseset):
             if not self.keepf and not root.mutable():
                 raise error.Abort(_("can't rebase public changeset %s")
                                   % root,
                                   hint=_("see 'hg help phases' for details"))
 
         (self.originalwd, self.destmap, self.state) = result
         if self.collapsef:
             dests = set(self.destmap.values())
             if len(dests) != 1:
                 raise error.Abort(
                     _('--collapse does not work with multiple destinations'))
             destrev = next(iter(dests))
             destancestors = self.repo.changelog.ancestors([destrev],
                                                           inclusive=True)
             self.external = externalparent(self.repo, self.state, destancestors)
 
         for destrev in sorted(set(destmap.values())):
             dest = self.repo[destrev]
             if dest.closesbranch() and not self.keepbranchesf:
                 self.ui.status(_('reopening closed branch head %s\n') % dest)
 
         self.prepared = True
 
     def _assignworkingcopy(self):
         if self.inmemory:
             from mercurial.context import overlayworkingctx
             self.wctx = overlayworkingctx(self.repo)
             self.repo.ui.debug("rebasing in-memory\n")
         else:
             self.wctx = self.repo[None]
             self.repo.ui.debug("rebasing on disk\n")
         self.repo.ui.log("rebase", "", rebase_imm_used=self.wctx.isinmemory())
 
     def _performrebase(self, tr):
         self._assignworkingcopy()
         repo, ui = self.repo, self.ui
         if self.keepbranchesf:
             # insert _savebranch at the start of extrafns so if
             # there's a user-provided extrafn it can clobber branch if
             # desired
             self.extrafns.insert(0, _savebranch)
             if self.collapsef:
                 branches = set()
                 for rev in self.state:
                     branches.add(repo[rev].branch())
                     if len(branches) > 1:
                         raise error.Abort(_('cannot collapse multiple named '
                             'branches'))
 
         # Calculate self.obsoletenotrebased
         obsrevs = _filterobsoleterevs(self.repo, self.state)
         self._handleskippingobsolete(obsrevs, self.destmap)
 
         # Keep track of the active bookmarks in order to reset them later
         self.activebookmark = self.activebookmark or repo._activebookmark
         if self.activebookmark:
             bookmarks.deactivate(repo)
 
         # Store the state before we begin so users can run 'hg rebase --abort'
         # if we fail before the transaction closes.
         self.storestatus()
 
         cands = [k for k, v in self.state.iteritems() if v == revtodo]
         total = len(cands)
         pos = 0
         for subset in sortsource(self.destmap):
             pos = self._performrebasesubset(tr, subset, pos, total)
         ui.progress(_('rebasing'), None)
         ui.note(_('rebase merging completed\n'))
 
     def _performrebasesubset(self, tr, subset, pos, total):
         repo, ui, opts = self.repo, self.ui, self.opts
         sortedrevs = repo.revs('sort(%ld, -topo)', subset)
         allowdivergence = self.ui.configbool(
             'experimental', 'evolution.allowdivergence')
         if not allowdivergence:
             sortedrevs -= repo.revs(
                 'descendants(%ld) and not %ld',
                 self.obsoletewithoutsuccessorindestination,
                 self.obsoletewithoutsuccessorindestination,
             )
         for rev in sortedrevs:
             dest = self.destmap[rev]
             ctx = repo[rev]
             desc = _ctxdesc(ctx)
             if self.state[rev] == rev:
                 ui.status(_('already rebased %s\n') % desc)
             elif (not allowdivergence
                   and rev in self.obsoletewithoutsuccessorindestination):
                 msg = _('note: not rebasing %s and its descendants as '
                         'this would cause divergence\n') % desc
                 repo.ui.status(msg)
                 self.skipped.add(rev)
             elif rev in self.obsoletenotrebased:
                 succ = self.obsoletenotrebased[rev]
                 if succ is None:
                     msg = _('note: not rebasing %s, it has no '
                             'successor\n') % desc
                 else:
                     succdesc = _ctxdesc(repo[succ])
                     msg = (_('note: not rebasing %s, already in '
                              'destination as %s\n') % (desc, succdesc))
                 repo.ui.status(msg)
                 # Make clearrebased aware state[rev] is not a true successor
                 self.skipped.add(rev)
                 # Record rev as moved to its desired destination in self.state.
                 # This helps bookmark and working parent movement.
                 dest = max(adjustdest(repo, rev, self.destmap, self.state,
                                       self.skipped))
                 self.state[rev] = dest
             elif self.state[rev] == revtodo:
                 pos += 1
                 ui.status(_('rebasing %s\n') % desc)
                 ui.progress(_("rebasing"), pos, ("%d:%s" % (rev, ctx)),
                             _('changesets'), total)
                 p1, p2, base = defineparents(repo, rev, self.destmap,
                                              self.state, self.skipped,
                                              self.obsoletenotrebased)
                 self.storestatus(tr=tr)
                 storecollapsemsg(repo, self.collapsemsg)
                 if len(repo[None].parents()) == 2:
                     repo.ui.debug('resuming interrupted rebase\n')
                 else:
                     try:
                         ui.setconfig('ui', 'forcemerge', opts.get('tool', ''),
                                      'rebase')
                         stats = rebasenode(repo, rev, p1, base, self.state,
                                            self.collapsef, dest, wctx=self.wctx)
                         if stats and stats[3] > 0:
                             if self.wctx.isinmemory():
                                 raise error.InMemoryMergeConflictsError()
                             else:
                                 raise error.InterventionRequired(
                                     _('unresolved conflicts (see hg '
                                       'resolve, then hg rebase --continue)'))
                     finally:
                         ui.setconfig('ui', 'forcemerge', '', 'rebase')
                 if not self.collapsef:
                     merging = p2 != nullrev
                     editform = cmdutil.mergeeditform(merging, 'rebase')
                     editor = cmdutil.getcommiteditor(editform=editform,
                                                      **pycompat.strkwargs(opts))
                     if self.wctx.isinmemory():
                         newnode = concludememorynode(repo, rev, p1, p2,
                             wctx=self.wctx,
                             extrafn=_makeextrafn(self.extrafns),
                             editor=editor,
                             keepbranches=self.keepbranchesf,
                             date=self.date)
                         mergemod.mergestate.clean(repo)
                     else:
                         newnode = concludenode(repo, rev, p1, p2,
                             extrafn=_makeextrafn(self.extrafns),
                             editor=editor,
                             keepbranches=self.keepbranchesf,
                             date=self.date)
 
                     if newnode is None:
                         # If it ended up being a no-op commit, then the normal
                         # merge state clean-up path doesn't happen, so do it
                         # here. Fix issue5494
                         mergemod.mergestate.clean(repo)
                 else:
                     # Skip commit if we are collapsing
                     if self.wctx.isinmemory():
                         self.wctx.setbase(repo[p1])
                     else:
                         repo.setparents(repo[p1].node())
                     newnode = None
                 # Update the state
                 if newnode is not None:
                     self.state[rev] = repo[newnode].rev()
                     ui.debug('rebased as %s\n' % short(newnode))
                 else:
                     if not self.collapsef:
                         ui.warn(_('note: rebase of %d:%s created no changes '
                                   'to commit\n') % (rev, ctx))
                         self.skipped.add(rev)
                     self.state[rev] = p1
                     ui.debug('next revision set to %d\n' % p1)
             else:
                 ui.status(_('already rebased %s as %s\n') %
                           (desc, repo[self.state[rev]]))
         return pos
 
     def _finishrebase(self):
         repo, ui, opts = self.repo, self.ui, self.opts
         fm = ui.formatter('rebase', opts)
         fm.startitem()
         if self.collapsef and not self.keepopen:
             p1, p2, _base = defineparents(repo, min(self.state), self.destmap,
                                           self.state, self.skipped,
                                           self.obsoletenotrebased)
             editopt = opts.get('edit')
             editform = 'rebase.collapse'
             if self.collapsemsg:
                 commitmsg = self.collapsemsg
             else:
                 commitmsg = 'Collapsed revision'
                 for rebased in sorted(self.state):
                     if rebased not in self.skipped:
                         commitmsg += '\n* %s' % repo[rebased].description()
                 editopt = True
             editor = cmdutil.getcommiteditor(edit=editopt, editform=editform)
             revtoreuse = max(self.state)
 
             dsguard = None
             if self.inmemory:
                 newnode = concludememorynode(repo, revtoreuse, p1,
                     self.external,
                     commitmsg=commitmsg,
                     extrafn=_makeextrafn(self.extrafns),
                     editor=editor,
                     keepbranches=self.keepbranchesf,
                     date=self.date, wctx=self.wctx)
             else:
                 if ui.configbool('rebase', 'singletransaction'):
                     dsguard = dirstateguard.dirstateguard(repo, 'rebase')
                 with util.acceptintervention(dsguard):
                     newnode = concludenode(repo, revtoreuse, p1, self.external,
                         commitmsg=commitmsg,
                         extrafn=_makeextrafn(self.extrafns),
                         editor=editor,
                         keepbranches=self.keepbranchesf,
                         date=self.date)
             if newnode is not None:
                 newrev = repo[newnode].rev()
-                for oldrev in self.state.iterkeys():
+                for oldrev in self.state:
                     self.state[oldrev] = newrev
 
         if 'qtip' in repo.tags():
             updatemq(repo, self.state, self.skipped, **opts)
 
         # restore original working directory
         # (we do this before stripping)
         newwd = self.state.get(self.originalwd, self.originalwd)
         if newwd < 0:
             # original directory is a parent of rebase set root or ignored
             newwd = self.originalwd
         if (newwd not in [c.rev() for c in repo[None].parents()] and
                 not self.inmemory):
             ui.note(_("update back to initial working directory parent\n"))
             hg.updaterepo(repo, newwd, False)
 
         collapsedas = None
         if not self.keepf:
             if self.collapsef:
                 collapsedas = newnode
         clearrebased(ui, repo, self.destmap, self.state, self.skipped,
                      collapsedas, self.keepf, fm=fm)
 
         clearstatus(repo)
         clearcollapsemsg(repo)
 
         ui.note(_("rebase completed\n"))
         util.unlinkpath(repo.sjoin('undo'), ignoremissing=True)
         if self.skipped:
             skippedlen = len(self.skipped)
             ui.note(_("%d revisions have been skipped\n") % skippedlen)
         fm.end()
 
         if (self.activebookmark and self.activebookmark in repo._bookmarks and
             repo['.'].node() == repo._bookmarks[self.activebookmark]):
                 bookmarks.activate(repo, self.activebookmark)
 
 @command('rebase',
     [('s', 'source', '',
      _('rebase the specified changeset and descendants'), _('REV')),
     ('b', 'base', '',
      _('rebase everything from branching point of specified changeset'),
      _('REV')),
     ('r', 'rev', [],
      _('rebase these revisions'),
      _('REV')),
     ('d', 'dest', '',
      _('rebase onto the specified changeset'), _('REV')),
     ('', 'collapse', False, _('collapse the rebased changesets')),
     ('m', 'message', '',
      _('use text as collapse commit message'), _('TEXT')),
     ('e', 'edit', False, _('invoke editor on commit messages')),
     ('l', 'logfile', '',
      _('read collapse commit message from file'), _('FILE')),
     ('k', 'keep', False, _('keep original changesets')),
     ('', 'keepbranches', False, _('keep original branch names')),
     ('D', 'detach', False, _('(DEPRECATED)')),
     ('i', 'interactive', False, _('(DEPRECATED)')),
     ('t', 'tool', '', _('specify merge tool')),
     ('c', 'continue', False, _('continue an interrupted rebase')),
     ('a', 'abort', False, _('abort an interrupted rebase'))] +
     cmdutil.formatteropts,
     _('[-s REV | -b REV] [-d REV] [OPTION]'))
 def rebase(ui, repo, **opts):
     """move changeset (and descendants) to a different branch
 
     Rebase uses repeated merging to graft changesets from one part of
     history (the source) onto another (the destination). This can be
     useful for linearizing *local* changes relative to a master
     development tree.
 
     Published commits cannot be rebased (see :hg:`help phases`).
     To copy commits, see :hg:`help graft`.
 
     If you don't specify a destination changeset (``-d/--dest``), rebase
     will use the same logic as :hg:`merge` to pick a destination.  if
     the current branch contains exactly one other head, the other head
     is merged with by default.  Otherwise, an explicit revision with
     which to merge with must be provided.  (destination changeset is not
     modified by rebasing, but new changesets are added as its
     descendants.)
 
     Here are the ways to select changesets:
 
       1. Explicitly select them using ``--rev``.
 
       2. Use ``--source`` to select a root changeset and include all of its
          descendants.
 
       3. Use ``--base`` to select a changeset; rebase will find ancestors
          and their descendants which are not also ancestors of the destination.
 
       4. If you do not specify any of ``--rev``, ``source``, or ``--base``,
          rebase will use ``--base .`` as above.
 
     If ``--source`` or ``--rev`` is used, special names ``SRC`` and ``ALLSRC``
     can be used in ``--dest``. Destination would be calculated per source
     revision with ``SRC`` substituted by that single source revision and
     ``ALLSRC`` substituted by all source revisions.
 
     Rebase will destroy original changesets unless you use ``--keep``.
     It will also move your bookmarks (even if you do).
 
     Some changesets may be dropped if they do not contribute changes
     (e.g. merges from the destination branch).
 
     Unlike ``merge``, rebase will do nothing if you are at the branch tip of
     a named branch with two heads. You will need to explicitly specify source
     and/or destination.
 
     If you need to use a tool to automate merge/conflict decisions, you
     can specify one with ``--tool``, see :hg:`help merge-tools`.
     As a caveat: the tool will not be used to mediate when a file was
     deleted, there is no hook presently available for this.
 
     If a rebase is interrupted to manually resolve a conflict, it can be
     continued with --continue/-c or aborted with --abort/-a.
 
     .. container:: verbose
 
       Examples:
 
       - move "local changes" (current commit back to branching point)
         to the current branch tip after a pull::
 
           hg rebase
 
       - move a single changeset to the stable branch::
 
           hg rebase -r 5f493448 -d stable
 
       - splice a commit and all its descendants onto another part of history::
 
           hg rebase --source c0c3 --dest 4cf9
 
       - rebase everything on a branch marked by a bookmark onto the
         default branch::
 
           hg rebase --base myfeature --dest default
 
       - collapse a sequence of changes into a single commit::
 
           hg rebase --collapse -r 1520:1525 -d .
 
       - move a named branch while preserving its name::
 
           hg rebase -r "branch(featureX)" -d 1.3 --keepbranches
 
       - stabilize orphaned changesets so history looks linear::
 
           hg rebase -r 'orphan()-obsolete()'\
  -d 'first(max((successors(max(roots(ALLSRC) & ::SRC)^)-obsolete())::) +\
  max(::((roots(ALLSRC) & ::SRC)^)-obsolete()))'
 
     Configuration Options:
 
     You can make rebase require a destination if you set the following config
     option::
 
       [commands]
       rebase.requiredest = True
 
     By default, rebase will close the transaction after each commit. For
     performance purposes, you can configure rebase to use a single transaction
     across the entire rebase. WARNING: This setting introduces a significant
     risk of losing the work you've done in a rebase if the rebase aborts
     unexpectedly::
 
       [rebase]
       singletransaction = True
 
     By default, rebase writes to the working copy, but you can configure it to
     run in-memory for for better performance, and to allow it to run if the
     working copy is dirty::
 
       [rebase]
       experimental.inmemory = True
 
     Return Values:
 
     Returns 0 on success, 1 if nothing to rebase or there are
     unresolved conflicts.
 
     """
     inmemory = ui.configbool('rebase', 'experimental.inmemory')
     if (opts.get('continue') or opts.get('abort') or
         repo.currenttransaction() is not None):
         # in-memory rebase is not compatible with resuming rebases.
         # (Or if it is run within a transaction, since the restart logic can
         # fail the entire transaction.)
         inmemory = False
 
     if inmemory:
         try:
             # in-memory merge doesn't support conflicts, so if we hit any, abort
             # and re-run as an on-disk merge.
             return _origrebase(ui, repo, inmemory=inmemory, **opts)
         except error.InMemoryMergeConflictsError:
             ui.warn(_('hit merge conflicts; re-running rebase without in-memory'
                       ' merge\n'))
             _origrebase(ui, repo, **{'abort': True})
             return _origrebase(ui, repo, inmemory=False, **opts)
     else:
         return _origrebase(ui, repo, **opts)
 
 def _origrebase(ui, repo, inmemory=False, **opts):
     opts = pycompat.byteskwargs(opts)
     rbsrt = rebaseruntime(repo, ui, inmemory, opts)
 
     with repo.wlock(), repo.lock():
         # Validate input and define rebasing points
         destf = opts.get('dest', None)
         srcf = opts.get('source', None)
         basef = opts.get('base', None)
         revf = opts.get('rev', [])
         # search default destination in this space
         # used in the 'hg pull --rebase' case, see issue 5214.
         destspace = opts.get('_destspace')
         contf = opts.get('continue')
         abortf = opts.get('abort')
         if opts.get('interactive'):
             try:
                 if extensions.find('histedit'):
                     enablehistedit = ''
             except KeyError:
                 enablehistedit = " --config extensions.histedit="
             help = "hg%s help -e histedit" % enablehistedit
             msg = _("interactive history editing is supported by the "
                     "'histedit' extension (see \"%s\")") % help
             raise error.Abort(msg)
 
         if rbsrt.collapsemsg and not rbsrt.collapsef:
             raise error.Abort(
                 _('message can only be specified with collapse'))
 
         if contf or abortf:
             if contf and abortf:
                 raise error.Abort(_('cannot use both abort and continue'))
             if rbsrt.collapsef:
                 raise error.Abort(
                     _('cannot use collapse with continue or abort'))
             if srcf or basef or destf:
                 raise error.Abort(
                     _('abort and continue do not allow specifying revisions'))
             if abortf and opts.get('tool', False):
                 ui.warn(_('tool option will be ignored\n'))
             if contf:
                 ms = mergemod.mergestate.read(repo)
                 mergeutil.checkunresolved(ms)
 
             retcode = rbsrt._prepareabortorcontinue(abortf)
             if retcode is not None:
                 return retcode
         else:
             destmap = _definedestmap(ui, repo, rbsrt, destf, srcf, basef, revf,
                                      destspace=destspace)
             retcode = rbsrt._preparenewrebase(destmap)
             if retcode is not None:
                 return retcode
 
         tr = None
         dsguard = None
 
         singletr = ui.configbool('rebase', 'singletransaction')
         if singletr:
             tr = repo.transaction('rebase')
 
         # If `rebase.singletransaction` is enabled, wrap the entire operation in
         # one transaction here. Otherwise, transactions are obtained when
         # committing each node, which is slower but allows partial success.
         with util.acceptintervention(tr):
             # Same logic for the dirstate guard, except we don't create one when
             # rebasing in-memory (it's not needed).
             if singletr and not inmemory:
                 dsguard = dirstateguard.dirstateguard(repo, 'rebase')
             with util.acceptintervention(dsguard):
                 rbsrt._performrebase(tr)
 
         rbsrt._finishrebase()
 
 def _definedestmap(ui, repo, rbsrt, destf=None, srcf=None, basef=None,
                    revf=None, destspace=None):
     """use revisions argument to define destmap {srcrev: destrev}"""
     if revf is None:
         revf = []
 
     # destspace is here to work around issues with `hg pull --rebase` see
     # issue5214 for details
     if srcf and basef:
         raise error.Abort(_('cannot specify both a source and a base'))
     if revf and basef:
         raise error.Abort(_('cannot specify both a revision and a base'))
     if revf and srcf:
         raise error.Abort(_('cannot specify both a revision and a source'))
 
     if not rbsrt.inmemory:
         cmdutil.checkunfinished(repo)
         cmdutil.bailifchanged(repo)
 
     if ui.configbool('commands', 'rebase.requiredest') and not destf:
         raise error.Abort(_('you must specify a destination'),
                           hint=_('use: hg rebase -d REV'))
 
     dest = None
 
     if revf:
         rebaseset = scmutil.revrange(repo, revf)
         if not rebaseset:
             ui.status(_('empty "rev" revision set - nothing to rebase\n'))
             return None
     elif srcf:
         src = scmutil.revrange(repo, [srcf])
         if not src:
             ui.status(_('empty "source" revision set - nothing to rebase\n'))
             return None
         rebaseset = repo.revs('(%ld)::', src)
         assert rebaseset
     else:
         base = scmutil.revrange(repo, [basef or '.'])
         if not base:
             ui.status(_('empty "base" revision set - '
                         "can't compute rebase set\n"))
             return None
         if destf:
             # --base does not support multiple destinations
             dest = scmutil.revsingle(repo, destf)
         else:
             dest = repo[_destrebase(repo, base, destspace=destspace)]
             destf = str(dest)
 
         roots = [] # selected children of branching points
         bpbase = {} # {branchingpoint: [origbase]}
         for b in base: # group bases by branching points
             bp = repo.revs('ancestor(%d, %d)', b, dest).first()
             bpbase[bp] = bpbase.get(bp, []) + [b]
         if None in bpbase:
             # emulate the old behavior, showing "nothing to rebase" (a better
             # behavior may be abort with "cannot find branching point" error)
             bpbase.clear()
         for bp, bs in bpbase.iteritems(): # calculate roots
             roots += list(repo.revs('children(%d) & ancestors(%ld)', bp, bs))
 
         rebaseset = repo.revs('%ld::', roots)
 
         if not rebaseset:
             # transform to list because smartsets are not comparable to
             # lists. This should be improved to honor laziness of
             # smartset.
             if list(base) == [dest.rev()]:
                 if basef:
                     ui.status(_('nothing to rebase - %s is both "base"'
                                 ' and destination\n') % dest)
                 else:
                     ui.status(_('nothing to rebase - working directory '
                                 'parent is also destination\n'))
             elif not repo.revs('%ld - ::%d', base, dest):
                 if basef:
                     ui.status(_('nothing to rebase - "base" %s is '
                                 'already an ancestor of destination '
                                 '%s\n') %
                               ('+'.join(str(repo[r]) for r in base),
                                dest))
                 else:
                     ui.status(_('nothing to rebase - working '
                                 'directory parent is already an '
                                 'ancestor of destination %s\n') % dest)
             else: # can it happen?
                 ui.status(_('nothing to rebase from %s to %s\n') %
                           ('+'.join(str(repo[r]) for r in base), dest))
             return None
     # If rebasing the working copy parent, force in-memory merge to be off.
     #
     # This is because the extra work of checking out the newly rebased commit
     # outweights the benefits of rebasing in-memory, and executing an extra
     # update command adds a bit of overhead, so better to just do it on disk. In
     # all other cases leave it on.
     #
     # Note that there are cases where this isn't true -- e.g., rebasing large
     # stacks that include the WCP. However, I'm not yet sure where the cutoff
     # is.
     rebasingwcp = repo['.'].rev() in rebaseset
     ui.log("rebase", "", rebase_rebasing_wcp=rebasingwcp)
     if rbsrt.inmemory and rebasingwcp:
         rbsrt.inmemory = False
         # Check these since we did not before.
         cmdutil.checkunfinished(repo)
         cmdutil.bailifchanged(repo)
 
     if not destf:
         dest = repo[_destrebase(repo, rebaseset, destspace=destspace)]
         destf = str(dest)
 
     allsrc = revsetlang.formatspec('%ld', rebaseset)
     alias = {'ALLSRC': allsrc}
 
     if dest is None:
         try:
             # fast path: try to resolve dest without SRC alias
             dest = scmutil.revsingle(repo, destf, localalias=alias)
         except error.RepoLookupError:
             # multi-dest path: resolve dest for each SRC separately
             destmap = {}
             for r in rebaseset:
                 alias['SRC'] = revsetlang.formatspec('%d', r)
                 # use repo.anyrevs instead of scmutil.revsingle because we
                 # don't want to abort if destset is empty.
                 destset = repo.anyrevs([destf], user=True, localalias=alias)
                 size = len(destset)
                 if size == 1:
                     destmap[r] = destset.first()
                 elif size == 0:
                     ui.note(_('skipping %s - empty destination\n') % repo[r])
                 else:
                     raise error.Abort(_('rebase destination for %s is not '
                                         'unique') % repo[r])
 
     if dest is not None:
         # single-dest case: assign dest to each rev in rebaseset
         destrev = dest.rev()
         destmap = {r: destrev for r in rebaseset} # {srcrev: destrev}
 
     if not destmap:
         ui.status(_('nothing to rebase - empty destination\n'))
         return None
 
     return destmap
 
 def externalparent(repo, state, destancestors):
     """Return the revision that should be used as the second parent
     when the revisions in state is collapsed on top of destancestors.
     Abort if there is more than one parent.
     """
     parents = set()
     source = min(state)
     for rev in state:
         if rev == source:
             continue
         for p in repo[rev].parents():
             if (p.rev() not in state
                         and p.rev() not in destancestors):
                 parents.add(p.rev())
     if not parents:
         return nullrev
     if len(parents) == 1:
         return parents.pop()
     raise error.Abort(_('unable to collapse on top of %s, there is more '
                        'than one external parent: %s') %
                      (max(destancestors),
                       ', '.join(str(p) for p in sorted(parents))))
 
 def concludememorynode(repo, rev, p1, p2, wctx=None,
                        commitmsg=None, editor=None, extrafn=None,
                        keepbranches=False, date=None):
     '''Commit the memory changes with parents p1 and p2. Reuse commit info from
     rev but also store useful information in extra.
     Return node of committed revision.'''
     ctx = repo[rev]
     if commitmsg is None:
         commitmsg = ctx.description()
     keepbranch = keepbranches and repo[p1].branch() != ctx.branch()
     extra = {'rebase_source': ctx.hex()}
     if extrafn:
         extrafn(ctx, extra)
 
     destphase = max(ctx.phase(), phases.draft)
     overrides = {('phases', 'new-commit'): destphase}
     with repo.ui.configoverride(overrides, 'rebase'):
         if keepbranch:
             repo.ui.setconfig('ui', 'allowemptycommit', True)
         # Replicates the empty check in ``repo.commit``.
         if wctx.isempty() and not repo.ui.configbool('ui', 'allowemptycommit'):
             return None
 
         if date is None:
             date = ctx.date()
 
         # By convention, ``extra['branch']`` (set by extrafn) clobbers
         # ``branch`` (used when passing ``--keepbranches``).
         branch = repo[p1].branch()
         if 'branch' in extra:
             branch = extra['branch']
 
         memctx = wctx.tomemctx(commitmsg, parents=(p1, p2), date=date,
             extra=extra, user=ctx.user(), branch=branch, editor=editor)
         commitres = repo.commitctx(memctx)
         wctx.clean() # Might be reused
         return commitres
 
 def concludenode(repo, rev, p1, p2, commitmsg=None, editor=None, extrafn=None,
                  keepbranches=False, date=None):
     '''Commit the wd changes with parents p1 and p2. Reuse commit info from rev
     but also store useful information in extra.
     Return node of committed revision.'''
     dsguard = util.nullcontextmanager()
     if not repo.ui.configbool('rebase', 'singletransaction'):
         dsguard = dirstateguard.dirstateguard(repo, 'rebase')
     with dsguard:
         repo.setparents(repo[p1].node(), repo[p2].node())
         ctx = repo[rev]
         if commitmsg is None:
             commitmsg = ctx.description()
         keepbranch = keepbranches and repo[p1].branch() != ctx.branch()
         extra = {'rebase_source': ctx.hex()}
         if extrafn:
             extrafn(ctx, extra)
 
         destphase = max(ctx.phase(), phases.draft)
         overrides = {('phases', 'new-commit'): destphase}
         with repo.ui.configoverride(overrides, 'rebase'):
             if keepbranch:
                 repo.ui.setconfig('ui', 'allowemptycommit', True)
             # Commit might fail if unresolved files exist
             if date is None:
                 date = ctx.date()
             newnode = repo.commit(text=commitmsg, user=ctx.user(),
                                   date=date, extra=extra, editor=editor)
 
         repo.dirstate.setbranch(repo[newnode].branch())
         return newnode
 
 def rebasenode(repo, rev, p1, base, state, collapse, dest, wctx):
     'Rebase a single revision rev on top of p1 using base as merge ancestor'
     # Merge phase
     # Update to destination and merge it with local
     if wctx.isinmemory():
         wctx.setbase(repo[p1])
     else:
         if repo['.'].rev() != p1:
             repo.ui.debug(" update to %d:%s\n" % (p1, repo[p1]))
             mergemod.update(repo, p1, False, True)
         else:
             repo.ui.debug(" already in destination\n")
         # This is, alas, necessary to invalidate workingctx's manifest cache,
         # as well as other data we litter on it in other places.
         wctx = repo[None]
         repo.dirstate.write(repo.currenttransaction())
     repo.ui.debug(" merge against %d:%s\n" % (rev, repo[rev]))
     if base is not None:
         repo.ui.debug("   detach base %d:%s\n" % (base, repo[base]))
     # When collapsing in-place, the parent is the common ancestor, we
     # have to allow merging with it.
     stats = mergemod.update(repo, rev, True, True, base, collapse,
                             labels=['dest', 'source'], wc=wctx)
     if collapse:
         copies.duplicatecopies(repo, wctx, rev, dest)
     else:
         # If we're not using --collapse, we need to
         # duplicate copies between the revision we're
         # rebasing and its first parent, but *not*
         # duplicate any copies that have already been
         # performed in the destination.
         p1rev = repo[rev].p1().rev()
         copies.duplicatecopies(repo, wctx, rev, p1rev, skiprev=dest)
     return stats
 
 def adjustdest(repo, rev, destmap, state, skipped):
     """adjust rebase destination given the current rebase state
 
     rev is what is being rebased. Return a list of two revs, which are the
     adjusted destinations for rev's p1 and p2, respectively. If a parent is
     nullrev, return dest without adjustment for it.
 
     For example, when doing rebasing B+E to F, C to G, rebase will first move B
     to B1, and E's destination will be adjusted from F to B1.
 
         B1 <- written during rebasing B
         |
         F <- original destination of B, E
         |
         | E <- rev, which is being rebased
         | |
         | D <- prev, one parent of rev being checked
         | |
         | x <- skipped, ex. no successor or successor in (::dest)
         | |
         | C <- rebased as C', different destination
         | |
         | B <- rebased as B1     C'
         |/                       |
         A                        G <- destination of C, different
 
     Another example about merge changeset, rebase -r C+G+H -d K, rebase will
     first move C to C1, G to G1, and when it's checking H, the adjusted
     destinations will be [C1, G1].
 
             H       C1 G1
            /|       | /
           F G       |/
         K | |  ->   K
         | C D       |
         | |/        |
         | B         | ...
         |/          |/
         A           A
 
     Besides, adjust dest according to existing rebase information. For example,
 
       B C D    B needs to be rebased on top of C, C needs to be rebased on top
        \|/     of D. We will rebase C first.
         A
 
           C'   After rebasing C, when considering B's destination, use C'
           |    instead of the original C.
       B   D
        \ /
         A
     """
     # pick already rebased revs with same dest from state as interesting source
     dest = destmap[rev]
     source = [s for s, d in state.items()
               if d > 0 and destmap[s] == dest and s not in skipped]
 
     result = []
     for prev in repo.changelog.parentrevs(rev):
         adjusted = dest
         if prev != nullrev:
             candidate = repo.revs('max(%ld and (::%d))', source, prev).first()
             if candidate is not None:
                 adjusted = state[candidate]
         if adjusted == dest and dest in state:
             adjusted = state[dest]
             if adjusted == revtodo:
                 # sortsource should produce an order that makes this impossible
                 raise error.ProgrammingError(
                     'rev %d should be rebased already at this time' % dest)
         result.append(adjusted)
     return result
 
 def _checkobsrebase(repo, ui, rebaseobsrevs, rebaseobsskipped):
     """
     Abort if rebase will create divergence or rebase is noop because of markers
 
     `rebaseobsrevs`: set of obsolete revision in source
     `rebaseobsskipped`: set of revisions from source skipped because they have
     successors in destination or no non-obsolete successor.
     """
     # Obsolete node with successors not in dest leads to divergence
     divergenceok = ui.configbool('experimental',
                                  'evolution.allowdivergence')
     divergencebasecandidates = rebaseobsrevs - rebaseobsskipped
 
     if divergencebasecandidates and not divergenceok:
         divhashes = (str(repo[r])
                      for r in divergencebasecandidates)
         msg = _("this rebase will cause "
                 "divergences from: %s")
         h = _("to force the rebase please set "
               "experimental.evolution.allowdivergence=True")
         raise error.Abort(msg % (",".join(divhashes),), hint=h)
 
 def successorrevs(unfi, rev):
     """yield revision numbers for successors of rev"""
     assert unfi.filtername is None
     nodemap = unfi.changelog.nodemap
     for s in obsutil.allsuccessors(unfi.obsstore, [unfi[rev].node()]):
         if s in nodemap:
             yield nodemap[s]
 
 def defineparents(repo, rev, destmap, state, skipped, obsskipped):
     """Return new parents and optionally a merge base for rev being rebased
 
     The destination specified by "dest" cannot always be used directly because
     previously rebase result could affect destination. For example,
 
           D E    rebase -r C+D+E -d B
           |/     C will be rebased to C'
         B C      D's new destination will be C' instead of B
         |/       E's new destination will be C' instead of B
         A
 
     The new parents of a merge is slightly more complicated. See the comment
     block below.
     """
     # use unfiltered changelog since successorrevs may return filtered nodes
     assert repo.filtername is None
     cl = repo.changelog
     def isancestor(a, b):
         # take revision numbers instead of nodes
         if a == b:
             return True
         elif a > b:
             return False
         return cl.isancestor(cl.node(a), cl.node(b))
 
     dest = destmap[rev]
     oldps = repo.changelog.parentrevs(rev) # old parents
     newps = [nullrev, nullrev] # new parents
     dests = adjustdest(repo, rev, destmap, state, skipped)
     bases = list(oldps) # merge base candidates, initially just old parents
 
     if all(r == nullrev for r in oldps[1:]):
         # For non-merge changeset, just move p to adjusted dest as requested.
         newps[0] = dests[0]
     else:
         # For merge changeset, if we move p to dests[i] unconditionally, both
         # parents may change and the end result looks like "the merge loses a
         # parent", which is a surprise. This is a limit because "--dest" only
         # accepts one dest per src.
         #
         # Therefore, only move p with reasonable conditions (in this order):
         #   1. use dest, if dest is a descendent of (p or one of p's successors)
         #   2. use p's rebased result, if p is rebased (state[p] > 0)
         #
         # Comparing with adjustdest, the logic here does some additional work:
         #   1. decide which parents will not be moved towards dest
         #   2. if the above decision is "no", should a parent still be moved
         #      because it was rebased?
         #
         # For example:
         #
         #     C    # "rebase -r C -d D" is an error since none of the parents
         #    /|    # can be moved. "rebase -r B+C -d D" will move C's parent
         #   A B D  # B (using rule "2."), since B will be rebased.
         #
         # The loop tries to be not rely on the fact that a Mercurial node has
         # at most 2 parents.
         for i, p in enumerate(oldps):
             np = p # new parent
             if any(isancestor(x, dests[i]) for x in successorrevs(repo, p)):
                 np = dests[i]
             elif p in state and state[p] > 0:
                 np = state[p]
 
             # "bases" only record "special" merge bases that cannot be
             # calculated from changelog DAG (i.e. isancestor(p, np) is False).
             # For example:
             #
             #   B'   # rebase -s B -d D, when B was rebased to B'. dest for C
             #   | C  # is B', but merge base for C is B, instead of
             #   D |  # changelog.ancestor(C, B') == A. If changelog DAG and
             #   | B  # "state" edges are merged (so there will be an edge from
             #   |/   # B to B'), the merge base is still ancestor(C, B') in
             #   A    # the merged graph.
             #
             # Also see https://bz.mercurial-scm.org/show_bug.cgi?id=1950#c8
             # which uses "virtual null merge" to explain this situation.
             if isancestor(p, np):
                 bases[i] = nullrev
 
             # If one parent becomes an ancestor of the other, drop the ancestor
             for j, x in enumerate(newps[:i]):
                 if x == nullrev:
                     continue
                 if isancestor(np, x): # CASE-1
                     np = nullrev
                 elif isancestor(x, np): # CASE-2
                     newps[j] = np
                     np = nullrev
                     # New parents forming an ancestor relationship does not
                     # mean the old parents have a similar relationship. Do not
                     # set bases[x] to nullrev.
                     bases[j], bases[i] = bases[i], bases[j]
 
             newps[i] = np
 
         # "rebasenode" updates to new p1, and the old p1 will be used as merge
         # base. If only p2 changes, merging using unchanged p1 as merge base is
         # suboptimal. Therefore swap parents to make the merge sane.
         if newps[1] != nullrev and oldps[0] == newps[0]:
             assert len(newps) == 2 and len(oldps) == 2
             newps.reverse()
             bases.reverse()
 
         # No parent change might be an error because we fail to make rev a
         # descendent of requested dest. This can happen, for example:
         #
         #     C    # rebase -r C -d D
         #    /|    # None of A and B will be changed to D and rebase fails.
         #   A B D
         if set(newps) == set(oldps) and dest not in newps:
             raise error.Abort(_('cannot rebase %d:%s without '
                                 'moving at least one of its parents')
                               % (rev, repo[rev]))
 
     # Source should not be ancestor of dest. The check here guarantees it's
     # impossible. With multi-dest, the initial check does not cover complex
     # cases since we don't have abstractions to dry-run rebase cheaply.
     if any(p != nullrev and isancestor(rev, p) for p in newps):
         raise error.Abort(_('source is ancestor of destination'))
 
     # "rebasenode" updates to new p1, use the corresponding merge base.
     if bases[0] != nullrev:
         base = bases[0]
     else:
         base = None
 
     # Check if the merge will contain unwanted changes. That may happen if
     # there are multiple special (non-changelog ancestor) merge bases, which
     # cannot be handled well by the 3-way merge algorithm. For example:
     #
     #     F
     #    /|
     #   D E  # "rebase -r D+E+F -d Z", when rebasing F, if "D" was chosen
     #   | |  # as merge base, the difference between D and F will include
     #   B C  # C, so the rebased F will contain C surprisingly. If "E" was
     #   |/   #  chosen, the rebased F will contain B.
     #   A Z
     #
     # But our merge base candidates (D and E in above case) could still be
     # better than the default (ancestor(F, Z) == null). Therefore still
     # pick one (so choose p1 above).
     if sum(1 for b in bases if b != nullrev) > 1:
         unwanted = [None, None] # unwanted[i]: unwanted revs if choose bases[i]
         for i, base in enumerate(bases):
             if base == nullrev:
                 continue
             # Revisions in the side (not chosen as merge base) branch that
             # might contain "surprising" contents
             siderevs = list(repo.revs('((%ld-%d) %% (%d+%d))',
                                       bases, base, base, dest))
 
             # If those revisions are covered by rebaseset, the result is good.
             # A merge in rebaseset would be considered to cover its ancestors.
             if siderevs:
                 rebaseset = [r for r, d in state.items()
                              if d > 0 and r not in obsskipped]
                 merges = [r for r in rebaseset
                           if cl.parentrevs(r)[1] != nullrev]
                 unwanted[i] = list(repo.revs('%ld - (::%ld) - %ld',
                                              siderevs, merges, rebaseset))
 
         # Choose a merge base that has a minimal number of unwanted revs.
         l, i = min((len(revs), i)
                    for i, revs in enumerate(unwanted) if revs is not None)
         base = bases[i]
 
         # newps[0] should match merge base if possible. Currently, if newps[i]
         # is nullrev, the only case is newps[i] and newps[j] (j < i), one is
         # the other's ancestor. In that case, it's fine to not swap newps here.
         # (see CASE-1 and CASE-2 above)
         if i != 0 and newps[i] != nullrev:
             newps[0], newps[i] = newps[i], newps[0]
 
         # The merge will include unwanted revisions. Abort now. Revisit this if
         # we have a more advanced merge algorithm that handles multiple bases.
         if l > 0:
             unwanteddesc = _(' or ').join(
                 (', '.join('%d:%s' % (r, repo[r]) for r in revs)
                  for revs in unwanted if revs is not None))
             raise error.Abort(
                 _('rebasing %d:%s will include unwanted changes from %s')
                 % (rev, repo[rev], unwanteddesc))
 
     repo.ui.debug(" future parents are %d and %d\n" % tuple(newps))
 
     return newps[0], newps[1], base
 
 def isagitpatch(repo, patchname):
     'Return true if the given patch is in git format'
     mqpatch = os.path.join(repo.mq.path, patchname)
     for line in patch.linereader(file(mqpatch, 'rb')):
         if line.startswith('diff --git'):
             return True
     return False
 
 def updatemq(repo, state, skipped, **opts):
     'Update rebased mq patches - finalize and then import them'
     mqrebase = {}
     mq = repo.mq
     original_series = mq.fullseries[:]
     skippedpatches = set()
 
     for p in mq.applied:
         rev = repo[p.node].rev()
         if rev in state:
             repo.ui.debug('revision %d is an mq patch (%s), finalize it.\n' %
                                         (rev, p.name))
             mqrebase[rev] = (p.name, isagitpatch(repo, p.name))
         else:
             # Applied but not rebased, not sure this should happen
             skippedpatches.add(p.name)
 
     if mqrebase:
         mq.finish(repo, mqrebase.keys())
 
         # We must start import from the newest revision
         for rev in sorted(mqrebase, reverse=True):
             if rev not in skipped:
                 name, isgit = mqrebase[rev]
                 repo.ui.note(_('updating mq patch %s to %s:%s\n') %
                              (name, state[rev], repo[state[rev]]))
                 mq.qimport(repo, (), patchname=name, git=isgit,
                                 rev=[str(state[rev])])
             else:
                 # Rebased and skipped
                 skippedpatches.add(mqrebase[rev][0])
 
         # Patches were either applied and rebased and imported in
         # order, applied and removed or unapplied. Discard the removed
         # ones while preserving the original series order and guards.
         newseries = [s for s in original_series
                      if mq.guard_re.split(s, 1)[0] not in skippedpatches]
         mq.fullseries[:] = newseries
         mq.seriesdirty = True
         mq.savedirty()
 
 def storecollapsemsg(repo, collapsemsg):
     'Store the collapse message to allow recovery'
     collapsemsg = collapsemsg or ''
     f = repo.vfs("last-message.txt", "w")
     f.write("%s\n" % collapsemsg)
     f.close()
 
 def clearcollapsemsg(repo):
     'Remove collapse message file'
     repo.vfs.unlinkpath("last-message.txt", ignoremissing=True)
 
 def restorecollapsemsg(repo, isabort):
     'Restore previously stored collapse message'
     try:
         f = repo.vfs("last-message.txt")
         collapsemsg = f.readline().strip()
         f.close()
     except IOError as err:
         if err.errno != errno.ENOENT:
             raise
         if isabort:
             # Oh well, just abort like normal
             collapsemsg = ''
         else:
             raise error.Abort(_('missing .hg/last-message.txt for rebase'))
     return collapsemsg
 
 def clearstatus(repo):
     'Remove the status files'
     # Make sure the active transaction won't write the state file
     tr = repo.currenttransaction()
     if tr:
         tr.removefilegenerator('rebasestate')
     repo.vfs.unlinkpath("rebasestate", ignoremissing=True)
 
 def needupdate(repo, state):
     '''check whether we should `update --clean` away from a merge, or if
     somehow the working dir got forcibly updated, e.g. by older hg'''
     parents = [p.rev() for p in repo[None].parents()]
 
     # Are we in a merge state at all?
     if len(parents) < 2:
         return False
 
     # We should be standing on the first as-of-yet unrebased commit.
     firstunrebased = min([old for old, new in state.iteritems()
                           if new == nullrev])
     if firstunrebased in parents:
         return True
 
     return False
 
 def abort(repo, originalwd, destmap, state, activebookmark=None):
     '''Restore the repository to its original state.  Additional args:
 
     activebookmark: the name of the bookmark that should be active after the
         restore'''
 
     try:
         # If the first commits in the rebased set get skipped during the rebase,
         # their values within the state mapping will be the dest rev id. The
         # dstates list must must not contain the dest rev (issue4896)
         dstates = [s for r, s in state.items() if s >= 0 and s != destmap[r]]
         immutable = [d for d in dstates if not repo[d].mutable()]
         cleanup = True
         if immutable:
             repo.ui.warn(_("warning: can't clean up public changesets %s\n")
                         % ', '.join(str(repo[r]) for r in immutable),
                         hint=_("see 'hg help phases' for details"))
             cleanup = False
 
         descendants = set()
         if dstates:
             descendants = set(repo.changelog.descendants(dstates))
         if descendants - set(dstates):
             repo.ui.warn(_("warning: new changesets detected on destination "
                            "branch, can't strip\n"))
             cleanup = False
 
         if cleanup:
             shouldupdate = False
             rebased = [s for r, s in state.items()
                        if s >= 0 and s != destmap[r]]
             if rebased:
                 strippoints = [
                         c.node() for c in repo.set('roots(%ld)', rebased)]
 
             updateifonnodes = set(rebased)
             updateifonnodes.update(destmap.values())
             updateifonnodes.add(originalwd)
             shouldupdate = repo['.'].rev() in updateifonnodes
 
             # Update away from the rebase if necessary
             if shouldupdate or needupdate(repo, state):
                 mergemod.update(repo, originalwd, False, True)
 
             # Strip from the first rebased revision
             if rebased:
                 # no backup of rebased cset versions needed
                 repair.strip(repo.ui, repo, strippoints)
 
         if activebookmark and activebookmark in repo._bookmarks:
             bookmarks.activate(repo, activebookmark)
 
     finally:
         clearstatus(repo)
         clearcollapsemsg(repo)
         repo.ui.warn(_('rebase aborted\n'))
     return 0
 
 def sortsource(destmap):
     """yield source revisions in an order that we only rebase things once
 
     If source and destination overlaps, we should filter out revisions
     depending on other revisions which hasn't been rebased yet.
 
     Yield a sorted list of revisions each time.
 
     For example, when rebasing A to B, B to C. This function yields [B], then
     [A], indicating B needs to be rebased first.
 
     Raise if there is a cycle so the rebase is impossible.
     """
     srcset = set(destmap)
     while srcset:
         srclist = sorted(srcset)
         result = []
         for r in srclist:
             if destmap[r] not in srcset:
                 result.append(r)
         if not result:
             raise error.Abort(_('source and destination form a cycle'))
         srcset -= set(result)
         yield result
 
 def buildstate(repo, destmap, collapse):
     '''Define which revisions are going to be rebased and where
 
     repo: repo
     destmap: {srcrev: destrev}
     '''
     rebaseset = destmap.keys()
     originalwd = repo['.'].rev()
 
     # This check isn't strictly necessary, since mq detects commits over an
     # applied patch. But it prevents messing up the working directory when
     # a partially completed rebase is blocked by mq.
     if 'qtip' in repo.tags():
         mqapplied = set(repo[s.node].rev() for s in repo.mq.applied)
         if set(destmap.values()) & mqapplied:
             raise error.Abort(_('cannot rebase onto an applied mq patch'))
 
     # Get "cycle" error early by exhausting the generator.
     sortedsrc = list(sortsource(destmap)) # a list of sorted revs
     if not sortedsrc:
         raise error.Abort(_('no matching revisions'))
 
     # Only check the first batch of revisions to rebase not depending on other
     # rebaseset. This means "source is ancestor of destination" for the second
     # (and following) batches of revisions are not checked here. We rely on
     # "defineparents" to do that check.
     roots = list(repo.set('roots(%ld)', sortedsrc[0]))
     if not roots:
         raise error.Abort(_('no matching revisions'))
     def revof(r):
         return r.rev()
     roots = sorted(roots, key=revof)
     state = dict.fromkeys(rebaseset, revtodo)
     emptyrebase = (len(sortedsrc) == 1)
     for root in roots:
         dest = repo[destmap[root.rev()]]
         commonbase = root.ancestor(dest)
         if commonbase == root:
             raise error.Abort(_('source is ancestor of destination'))
         if commonbase == dest:
             wctx = repo[None]
             if dest == wctx.p1():
                 # when rebasing to '.', it will use the current wd branch name
                 samebranch = root.branch() == wctx.branch()
             else:
                 samebranch = root.branch() == dest.branch()
             if not collapse and samebranch and dest in root.parents():
                 # mark the revision as done by setting its new revision
                 # equal to its old (current) revisions
                 state[root.rev()] = root.rev()
                 repo.ui.debug('source is a child of destination\n')
                 continue
 
         emptyrebase = False
         repo.ui.debug('rebase onto %s starting from %s\n' % (dest, root))
     if emptyrebase:
         return None
     for rev in sorted(state):
         parents = [p for p in repo.changelog.parentrevs(rev) if p != nullrev]
         # if all parents of this revision are done, then so is this revision
         if parents and all((state.get(p) == p for p in parents)):
             state[rev] = rev
     return originalwd, destmap, state
 
 def clearrebased(ui, repo, destmap, state, skipped, collapsedas=None,
                  keepf=False, fm=None):
     """dispose of rebased revision at the end of the rebase
 
     If `collapsedas` is not None, the rebase was a collapse whose result if the
     `collapsedas` node.
 
     If `keepf` is not True, the rebase has --keep set and no nodes should be
     removed (but bookmarks still need to be moved).
     """
     tonode = repo.changelog.node
     replacements = {}
     moves = {}
     for rev, newrev in sorted(state.items()):
         if newrev >= 0 and newrev != rev:
             oldnode = tonode(rev)
             newnode = collapsedas or tonode(newrev)
             moves[oldnode] = newnode
             if not keepf:
                 if rev in skipped:
                     succs = ()
                 else:
                     succs = (newnode,)
                 replacements[oldnode] = succs
     scmutil.cleanupnodes(repo, replacements, 'rebase', moves)
     if fm:
         hf = fm.hexfunc
         fl = fm.formatlist
         fd = fm.formatdict
         nodechanges = fd({hf(oldn): fl([hf(n) for n in newn], name='node')
                           for oldn, newn in replacements.iteritems()},
                          key="oldnode", value="newnodes")
         fm.data(nodechanges=nodechanges)
 
 def pullrebase(orig, ui, repo, *args, **opts):
     'Call rebase after pull if the latter has been invoked with --rebase'
     ret = None
     if opts.get(r'rebase'):
         if ui.configbool('commands', 'rebase.requiredest'):
             msg = _('rebase destination required by configuration')
             hint = _('use hg pull followed by hg rebase -d DEST')
             raise error.Abort(msg, hint=hint)
 
         with repo.wlock(), repo.lock():
             if opts.get(r'update'):
                 del opts[r'update']
                 ui.debug('--update and --rebase are not compatible, ignoring '
                          'the update flag\n')
 
             cmdutil.checkunfinished(repo)
             cmdutil.bailifchanged(repo, hint=_('cannot pull with rebase: '
                 'please commit or shelve your changes first'))
 
             revsprepull = len(repo)
             origpostincoming = commands.postincoming
             def _dummy(*args, **kwargs):
                 pass
             commands.postincoming = _dummy
             try:
                 ret = orig(ui, repo, *args, **opts)
             finally:
                 commands.postincoming = origpostincoming
             revspostpull = len(repo)
             if revspostpull > revsprepull:
                 # --rev option from pull conflict with rebase own --rev
                 # dropping it
                 if r'rev' in opts:
                     del opts[r'rev']
                 # positional argument from pull conflicts with rebase's own
                 # --source.
                 if r'source' in opts:
                     del opts[r'source']
                 # revsprepull is the len of the repo, not revnum of tip.
                 destspace = list(repo.changelog.revs(start=revsprepull))
                 opts[r'_destspace'] = destspace
                 try:
                     rebase(ui, repo, **opts)
                 except error.NoMergeDestAbort:
                     # we can maybe update instead
                     rev, _a, _b = destutil.destupdate(repo)
                     if rev == repo['.'].rev():
                         ui.status(_('nothing to rebase\n'))
                     else:
                         ui.status(_('nothing to rebase - updating instead\n'))
                         # not passing argument to get the bare update behavior
                         # with warning and trumpets
                         commands.update(ui, repo)
     else:
         if opts.get(r'tool'):
             raise error.Abort(_('--tool can only be used with --rebase'))
         ret = orig(ui, repo, *args, **opts)
 
     return ret
 
 def _filterobsoleterevs(repo, revs):
     """returns a set of the obsolete revisions in revs"""
     return set(r for r in revs if repo[r].obsolete())
 
 def _computeobsoletenotrebased(repo, rebaseobsrevs, destmap):
     """Return (obsoletenotrebased, obsoletewithoutsuccessorindestination).
 
     `obsoletenotrebased` is a mapping mapping obsolete => successor for all
     obsolete nodes to be rebased given in `rebaseobsrevs`.
 
     `obsoletewithoutsuccessorindestination` is a set with obsolete revisions
     without a successor in destination.
 
     `obsoleteextinctsuccessors` is a set of obsolete revisions with only
     obsolete successors.
     """
     obsoletenotrebased = {}
     obsoletewithoutsuccessorindestination = set([])
     obsoleteextinctsuccessors = set([])
 
     assert repo.filtername is None
     cl = repo.changelog
     nodemap = cl.nodemap
     extinctnodes = set(cl.node(r) for r in repo.revs('extinct()'))
     for srcrev in rebaseobsrevs:
         srcnode = cl.node(srcrev)
         destnode = cl.node(destmap[srcrev])
         # XXX: more advanced APIs are required to handle split correctly
         successors = set(obsutil.allsuccessors(repo.obsstore, [srcnode]))
         # obsutil.allsuccessors includes node itself
         successors.remove(srcnode)
         if successors.issubset(extinctnodes):
             # all successors are extinct
             obsoleteextinctsuccessors.add(srcrev)
         if not successors:
             # no successor
             obsoletenotrebased[srcrev] = None
         else:
             for succnode in successors:
                 if succnode not in nodemap:
                     continue
                 if cl.isancestor(succnode, destnode):
                     obsoletenotrebased[srcrev] = nodemap[succnode]
                     break
             else:
                 # If 'srcrev' has a successor in rebase set but none in
                 # destination (which would be catched above), we shall skip it
                 # and its descendants to avoid divergence.
                 if any(nodemap[s] in destmap for s in successors):
                     obsoletewithoutsuccessorindestination.add(srcrev)
 
     return (
         obsoletenotrebased,
         obsoletewithoutsuccessorindestination,
         obsoleteextinctsuccessors,
     )
 
 def summaryhook(ui, repo):
     if not repo.vfs.exists('rebasestate'):
         return
     try:
         rbsrt = rebaseruntime(repo, ui, {})
         rbsrt.restorestatus()
         state = rbsrt.state
     except error.RepoLookupError:
         # i18n: column positioning for "hg summary"
         msg = _('rebase: (use "hg rebase --abort" to clear broken state)\n')
         ui.write(msg)
         return
     numrebased = len([i for i in state.itervalues() if i >= 0])
     # i18n: column positioning for "hg summary"
     ui.write(_('rebase: %s, %s (rebase --continue)\n') %
              (ui.label(_('%d rebased'), 'rebase.rebased') % numrebased,
               ui.label(_('%d remaining'), 'rebase.remaining') %
               (len(state) - numrebased)))
 
 def uisetup(ui):
     #Replace pull with a decorator to provide --rebase option
     entry = extensions.wrapcommand(commands.table, 'pull', pullrebase)
     entry[1].append(('', 'rebase', None,
                      _("rebase working directory to branch head")))
     entry[1].append(('t', 'tool', '',
                      _("specify merge tool for rebase")))
     cmdutil.summaryhooks.add('rebase', summaryhook)
     cmdutil.unfinishedstates.append(
         ['rebasestate', False, False, _('rebase in progress'),
          _("use 'hg rebase --continue' or 'hg rebase --abort'")])
     cmdutil.afterresolvedstates.append(
         ['rebasestate', _('hg rebase --continue')])