localrepo: pass root manifest into manifestlog.__init__

Today, localrepository has a method that can be overloaded which
returns an instance of the root manifest storage object. When a
manifestlog is created, it calls this private method and stores
the root manifest object on it.

This "hook" on localrepository isn't part of the documented interface.
It isn't compatible with our desire to make repo storage determined
before the repo object is constructed.

This commit changes manifestlog.__init__ to accept the root
storage object instead of calling into the repo to construct it.
By doing things this way, the repo instance is responsible for
constructing the manifest storage object directly.

This does mean that other derived repo types need to overload
manifestlog(). But they should have been doing this already,
as manifestlog() is typically decorated in a storage-specific way.
e.g. localrepository.manifestlog() is decorated as
@storecache('00manifest.i'). And this assumes that a 00manifest.i
file exists in the store vfs. This condition may not hold for
repository types using non-revlog storage. So it is important
for special repo types to override manifestlog() to remove this
file association.

The code changed in perf is wrong because it isn't compatible with
older Mercurial versions. But I'm pretty sure the code was broken
on older versions before this commit. It only affects `hg perftags`.
I don't care enough to fix that at this time.

.. api::

   ``manifest.manifestlog.__init__()`` now receives the root manifest
   storage instance instead of calling into a private method on
   the repo object to obtain it.

Differential Revision: https://phab.mercurial-scm.org/D4641

contrib/perf.py

@@ -1,2024 +1,2026
 # perf.py - performance test routines
 '''helper extension to measure performance'''
 
 # "historical portability" policy of perf.py:
 #
 # We have to do:
 # - make perf.py "loadable" with as wide Mercurial version as possible
 #   This doesn't mean that perf commands work correctly with that Mercurial.
 #   BTW, perf.py itself has been available since 1.1 (or eb240755386d).
 # - make historical perf command work correctly with as wide Mercurial
 #   version as possible
 #
 # We have to do, if possible with reasonable cost:
 # - make recent perf command for historical feature work correctly
 #   with early Mercurial
 #
 # We don't have to do:
 # - make perf command for recent feature work correctly with early
 #   Mercurial
 
 from __future__ import absolute_import
 import functools
 import gc
 import os
 import random
 import struct
 import sys
 import threading
 import time
 from mercurial import (
     changegroup,
     cmdutil,
     commands,
     copies,
     error,
     extensions,
     mdiff,
     merge,
     revlog,
     util,
 )
 
 # for "historical portability":
 # try to import modules separately (in dict order), and ignore
 # failure, because these aren't available with early Mercurial
 try:
     from mercurial import branchmap # since 2.5 (or bcee63733aad)
 except ImportError:
     pass
 try:
     from mercurial import obsolete # since 2.3 (or ad0d6c2b3279)
 except ImportError:
     pass
 try:
     from mercurial import registrar # since 3.7 (or 37d50250b696)
     dir(registrar) # forcibly load it
 except ImportError:
     registrar = None
 try:
     from mercurial import repoview # since 2.5 (or 3a6ddacb7198)
 except ImportError:
     pass
 try:
     from mercurial import scmutil # since 1.9 (or 8b252e826c68)
 except ImportError:
     pass
 try:
     from mercurial import pycompat
     getargspec = pycompat.getargspec  # added to module after 4.5
 except (ImportError, AttributeError):
     import inspect
     getargspec = inspect.getargspec
 
 try:
     # 4.7+
     queue = pycompat.queue.Queue
 except (AttributeError, ImportError):
     # <4.7.
     try:
         queue = pycompat.queue
     except (AttributeError, ImportError):
         queue = util.queue
 
 try:
     from mercurial import logcmdutil
     makelogtemplater = logcmdutil.maketemplater
 except (AttributeError, ImportError):
     try:
         makelogtemplater = cmdutil.makelogtemplater
     except (AttributeError, ImportError):
         makelogtemplater = None
 
 # for "historical portability":
 # define util.safehasattr forcibly, because util.safehasattr has been
 # available since 1.9.3 (or 94b200a11cf7)
 _undefined = object()
 def safehasattr(thing, attr):
     return getattr(thing, attr, _undefined) is not _undefined
 setattr(util, 'safehasattr', safehasattr)
 
 # for "historical portability":
 # define util.timer forcibly, because util.timer has been available
 # since ae5d60bb70c9
 if safehasattr(time, 'perf_counter'):
     util.timer = time.perf_counter
 elif os.name == b'nt':
     util.timer = time.clock
 else:
     util.timer = time.time
 
 # for "historical portability":
 # use locally defined empty option list, if formatteropts isn't
 # available, because commands.formatteropts has been available since
 # 3.2 (or 7a7eed5176a4), even though formatting itself has been
 # available since 2.2 (or ae5f92e154d3)
 formatteropts = getattr(cmdutil, "formatteropts",
                         getattr(commands, "formatteropts", []))
 
 # for "historical portability":
 # use locally defined option list, if debugrevlogopts isn't available,
 # because commands.debugrevlogopts has been available since 3.7 (or
 # 5606f7d0d063), even though cmdutil.openrevlog() has been available
 # since 1.9 (or a79fea6b3e77).
 revlogopts = getattr(cmdutil, "debugrevlogopts",
                      getattr(commands, "debugrevlogopts", [
         (b'c', b'changelog', False, (b'open changelog')),
         (b'm', b'manifest', False, (b'open manifest')),
         (b'', b'dir', False, (b'open directory manifest')),
         ]))
 
 cmdtable = {}
 
 # for "historical portability":
 # define parsealiases locally, because cmdutil.parsealiases has been
 # available since 1.5 (or 6252852b4332)
 def parsealiases(cmd):
     return cmd.lstrip(b"^").split(b"|")
 
 if safehasattr(registrar, 'command'):
     command = registrar.command(cmdtable)
 elif safehasattr(cmdutil, 'command'):
     command = cmdutil.command(cmdtable)
     if b'norepo' not in getargspec(command).args:
         # for "historical portability":
         # wrap original cmdutil.command, because "norepo" option has
         # been available since 3.1 (or 75a96326cecb)
         _command = command
         def command(name, options=(), synopsis=None, norepo=False):
             if norepo:
                 commands.norepo += b' %s' % b' '.join(parsealiases(name))
             return _command(name, list(options), synopsis)
 else:
     # for "historical portability":
     # define "@command" annotation locally, because cmdutil.command
     # has been available since 1.9 (or 2daa5179e73f)
     def command(name, options=(), synopsis=None, norepo=False):
         def decorator(func):
             if synopsis:
                 cmdtable[name] = func, list(options), synopsis
             else:
                 cmdtable[name] = func, list(options)
             if norepo:
                 commands.norepo += b' %s' % b' '.join(parsealiases(name))
             return func
         return decorator
 
 try:
     import mercurial.registrar
     import mercurial.configitems
     configtable = {}
     configitem = mercurial.registrar.configitem(configtable)
     configitem(b'perf', b'presleep',
         default=mercurial.configitems.dynamicdefault,
     )
     configitem(b'perf', b'stub',
         default=mercurial.configitems.dynamicdefault,
     )
     configitem(b'perf', b'parentscount',
         default=mercurial.configitems.dynamicdefault,
     )
     configitem(b'perf', b'all-timing',
         default=mercurial.configitems.dynamicdefault,
     )
 except (ImportError, AttributeError):
     pass
 
 def getlen(ui):
     if ui.configbool(b"perf", b"stub", False):
         return lambda x: 1
     return len
 
 def gettimer(ui, opts=None):
     """return a timer function and formatter: (timer, formatter)
 
     This function exists to gather the creation of formatter in a single
     place instead of duplicating it in all performance commands."""
 
     # enforce an idle period before execution to counteract power management
     # experimental config: perf.presleep
     time.sleep(getint(ui, b"perf", b"presleep", 1))
 
     if opts is None:
         opts = {}
     # redirect all to stderr unless buffer api is in use
     if not ui._buffers:
         ui = ui.copy()
         uifout = safeattrsetter(ui, b'fout', ignoremissing=True)
         if uifout:
             # for "historical portability":
             # ui.fout/ferr have been available since 1.9 (or 4e1ccd4c2b6d)
             uifout.set(ui.ferr)
 
     # get a formatter
     uiformatter = getattr(ui, 'formatter', None)
     if uiformatter:
         fm = uiformatter(b'perf', opts)
     else:
         # for "historical portability":
         # define formatter locally, because ui.formatter has been
         # available since 2.2 (or ae5f92e154d3)
         from mercurial import node
         class defaultformatter(object):
             """Minimized composition of baseformatter and plainformatter
             """
             def __init__(self, ui, topic, opts):
                 self._ui = ui
                 if ui.debugflag:
                     self.hexfunc = node.hex
                 else:
                     self.hexfunc = node.short
             def __nonzero__(self):
                 return False
             __bool__ = __nonzero__
             def startitem(self):
                 pass
             def data(self, **data):
                 pass
             def write(self, fields, deftext, *fielddata, **opts):
                 self._ui.write(deftext % fielddata, **opts)
             def condwrite(self, cond, fields, deftext, *fielddata, **opts):
                 if cond:
                     self._ui.write(deftext % fielddata, **opts)
             def plain(self, text, **opts):
                 self._ui.write(text, **opts)
             def end(self):
                 pass
         fm = defaultformatter(ui, b'perf', opts)
 
     # stub function, runs code only once instead of in a loop
     # experimental config: perf.stub
     if ui.configbool(b"perf", b"stub", False):
         return functools.partial(stub_timer, fm), fm
 
     # experimental config: perf.all-timing
     displayall = ui.configbool(b"perf", b"all-timing", False)
     return functools.partial(_timer, fm, displayall=displayall), fm
 
 def stub_timer(fm, func, title=None):
     func()
 
 def _timer(fm, func, title=None, displayall=False):
     gc.collect()
     results = []
     begin = util.timer()
     count = 0
     while True:
         ostart = os.times()
         cstart = util.timer()
         r = func()
         cstop = util.timer()
         ostop = os.times()
         count += 1
         a, b = ostart, ostop
         results.append((cstop - cstart, b[0] - a[0], b[1]-a[1]))
         if cstop - begin > 3 and count >= 100:
             break
         if cstop - begin > 10 and count >= 3:
             break
 
     fm.startitem()
 
     if title:
         fm.write(b'title', b'! %s\n', title)
     if r:
         fm.write(b'result', b'! result: %s\n', r)
     def display(role, entry):
         prefix = b''
         if role != b'best':
             prefix = b'%s.' % role
         fm.plain(b'!')
         fm.write(prefix + b'wall', b' wall %f', entry[0])
         fm.write(prefix + b'comb', b' comb %f', entry[1] + entry[2])
         fm.write(prefix + b'user', b' user %f', entry[1])
         fm.write(prefix + b'sys',  b' sys %f', entry[2])
         fm.write(prefix + b'count',  b' (%s of %d)', role, count)
         fm.plain(b'\n')
     results.sort()
     min_val = results[0]
     display(b'best', min_val)
     if displayall:
         max_val = results[-1]
         display(b'max', max_val)
         avg = tuple([sum(x) / count for x in zip(*results)])
         display(b'avg', avg)
         median = results[len(results) // 2]
         display(b'median', median)
 
 # utilities for historical portability
 
 def getint(ui, section, name, default):
     # for "historical portability":
     # ui.configint has been available since 1.9 (or fa2b596db182)
     v = ui.config(section, name, None)
     if v is None:
         return default
     try:
         return int(v)
     except ValueError:
         raise error.ConfigError((b"%s.%s is not an integer ('%s')")
                                 % (section, name, v))
 
 def safeattrsetter(obj, name, ignoremissing=False):
     """Ensure that 'obj' has 'name' attribute before subsequent setattr
 
     This function is aborted, if 'obj' doesn't have 'name' attribute
     at runtime. This avoids overlooking removal of an attribute, which
     breaks assumption of performance measurement, in the future.
 
     This function returns the object to (1) assign a new value, and
     (2) restore an original value to the attribute.
 
     If 'ignoremissing' is true, missing 'name' attribute doesn't cause
     abortion, and this function returns None. This is useful to
     examine an attribute, which isn't ensured in all Mercurial
     versions.
     """
     if not util.safehasattr(obj, name):
         if ignoremissing:
             return None
         raise error.Abort((b"missing attribute %s of %s might break assumption"
                            b" of performance measurement") % (name, obj))
 
     origvalue = getattr(obj, name)
     class attrutil(object):
         def set(self, newvalue):
             setattr(obj, name, newvalue)
         def restore(self):
             setattr(obj, name, origvalue)
 
     return attrutil()
 
 # utilities to examine each internal API changes
 
 def getbranchmapsubsettable():
     # for "historical portability":
     # subsettable is defined in:
     # - branchmap since 2.9 (or 175c6fd8cacc)
     # - repoview since 2.5 (or 59a9f18d4587)
     for mod in (branchmap, repoview):
         subsettable = getattr(mod, 'subsettable', None)
         if subsettable:
             return subsettable
 
     # bisecting in bcee63733aad::59a9f18d4587 can reach here (both
     # branchmap and repoview modules exist, but subsettable attribute
     # doesn't)
     raise error.Abort((b"perfbranchmap not available with this Mercurial"),
                       hint=b"use 2.5 or later")
 
 def getsvfs(repo):
     """Return appropriate object to access files under .hg/store
     """
     # for "historical portability":
     # repo.svfs has been available since 2.3 (or 7034365089bf)
     svfs = getattr(repo, 'svfs', None)
     if svfs:
         return svfs
     else:
         return getattr(repo, 'sopener')
 
 def getvfs(repo):
     """Return appropriate object to access files under .hg
     """
     # for "historical portability":
     # repo.vfs has been available since 2.3 (or 7034365089bf)
     vfs = getattr(repo, 'vfs', None)
     if vfs:
         return vfs
     else:
         return getattr(repo, 'opener')
 
 def repocleartagscachefunc(repo):
     """Return the function to clear tags cache according to repo internal API
     """
     if util.safehasattr(repo, b'_tagscache'): # since 2.0 (or 9dca7653b525)
         # in this case, setattr(repo, '_tagscache', None) or so isn't
         # correct way to clear tags cache, because existing code paths
         # expect _tagscache to be a structured object.
         def clearcache():
             # _tagscache has been filteredpropertycache since 2.5 (or
             # 98c867ac1330), and delattr() can't work in such case
             if b'_tagscache' in vars(repo):
                 del repo.__dict__[b'_tagscache']
         return clearcache
 
     repotags = safeattrsetter(repo, b'_tags', ignoremissing=True)
     if repotags: # since 1.4 (or 5614a628d173)
         return lambda : repotags.set(None)
 
     repotagscache = safeattrsetter(repo, b'tagscache', ignoremissing=True)
     if repotagscache: # since 0.6 (or d7df759d0e97)
         return lambda : repotagscache.set(None)
 
     # Mercurial earlier than 0.6 (or d7df759d0e97) logically reaches
     # this point, but it isn't so problematic, because:
     # - repo.tags of such Mercurial isn't "callable", and repo.tags()
     #   in perftags() causes failure soon
     # - perf.py itself has been available since 1.1 (or eb240755386d)
     raise error.Abort((b"tags API of this hg command is unknown"))
 
 # utilities to clear cache
 
 def clearfilecache(repo, attrname):
     unfi = repo.unfiltered()
     if attrname in vars(unfi):
         delattr(unfi, attrname)
     unfi._filecache.pop(attrname, None)
 
 # perf commands
 
 @command(b'perfwalk', formatteropts)
 def perfwalk(ui, repo, *pats, **opts):
     timer, fm = gettimer(ui, opts)
     m = scmutil.match(repo[None], pats, {})
     timer(lambda: len(list(repo.dirstate.walk(m, subrepos=[], unknown=True,
                                               ignored=False))))
     fm.end()
 
 @command(b'perfannotate', formatteropts)
 def perfannotate(ui, repo, f, **opts):
     timer, fm = gettimer(ui, opts)
     fc = repo[b'.'][f]
     timer(lambda: len(fc.annotate(True)))
     fm.end()
 
 @command(b'perfstatus',
          [(b'u', b'unknown', False,
            b'ask status to look for unknown files')] + formatteropts)
 def perfstatus(ui, repo, **opts):
     #m = match.always(repo.root, repo.getcwd())
     #timer(lambda: sum(map(len, repo.dirstate.status(m, [], False, False,
     #                                                False))))
     timer, fm = gettimer(ui, opts)
     timer(lambda: sum(map(len, repo.status(unknown=opts[b'unknown']))))
     fm.end()
 
 @command(b'perfaddremove', formatteropts)
 def perfaddremove(ui, repo, **opts):
     timer, fm = gettimer(ui, opts)
     try:
         oldquiet = repo.ui.quiet
         repo.ui.quiet = True
         matcher = scmutil.match(repo[None])
         opts[b'dry_run'] = True
         timer(lambda: scmutil.addremove(repo, matcher, b"", opts))
     finally:
         repo.ui.quiet = oldquiet
         fm.end()
 
 def clearcaches(cl):
     # behave somewhat consistently across internal API changes
     if util.safehasattr(cl, b'clearcaches'):
         cl.clearcaches()
     elif util.safehasattr(cl, b'_nodecache'):
         from mercurial.node import nullid, nullrev
         cl._nodecache = {nullid: nullrev}
         cl._nodepos = None
 
 @command(b'perfheads', formatteropts)
 def perfheads(ui, repo, **opts):
     timer, fm = gettimer(ui, opts)
     cl = repo.changelog
     def d():
         len(cl.headrevs())
         clearcaches(cl)
     timer(d)
     fm.end()
 
 @command(b'perftags', formatteropts)
 def perftags(ui, repo, **opts):
     import mercurial.changelog
     import mercurial.manifest
     timer, fm = gettimer(ui, opts)
     svfs = getsvfs(repo)
     repocleartagscache = repocleartagscachefunc(repo)
     def t():
         repo.changelog = mercurial.changelog.changelog(svfs)
-        repo.manifestlog = mercurial.manifest.manifestlog(svfs, repo)
+        rootmanifest = mercurial.manifest.manifestrevlog(svfs)
+        repo.manifestlog = mercurial.manifest.manifestlog(svfs, repo,
+                                                          rootmanifest)
         repocleartagscache()
         return len(repo.tags())
     timer(t)
     fm.end()
 
 @command(b'perfancestors', formatteropts)
 def perfancestors(ui, repo, **opts):
     timer, fm = gettimer(ui, opts)
     heads = repo.changelog.headrevs()
     def d():
         for a in repo.changelog.ancestors(heads):
             pass
     timer(d)
     fm.end()
 
 @command(b'perfancestorset', formatteropts)
 def perfancestorset(ui, repo, revset, **opts):
     timer, fm = gettimer(ui, opts)
     revs = repo.revs(revset)
     heads = repo.changelog.headrevs()
     def d():
         s = repo.changelog.ancestors(heads)
         for rev in revs:
             rev in s
     timer(d)
     fm.end()
 
 @command(b'perfbookmarks', formatteropts)
 def perfbookmarks(ui, repo, **opts):
     """benchmark parsing bookmarks from disk to memory"""
     timer, fm = gettimer(ui, opts)
     def d():
         clearfilecache(repo, b'_bookmarks')
         repo._bookmarks
     timer(d)
     fm.end()
 
 @command(b'perfbundleread', formatteropts, b'BUNDLE')
 def perfbundleread(ui, repo, bundlepath, **opts):
     """Benchmark reading of bundle files.
 
     This command is meant to isolate the I/O part of bundle reading as
     much as possible.
     """
     from mercurial import (
         bundle2,
         exchange,
         streamclone,
     )
 
     def makebench(fn):
         def run():
             with open(bundlepath, b'rb') as fh:
                 bundle = exchange.readbundle(ui, fh, bundlepath)
                 fn(bundle)
 
         return run
 
     def makereadnbytes(size):
         def run():
             with open(bundlepath, b'rb') as fh:
                 bundle = exchange.readbundle(ui, fh, bundlepath)
                 while bundle.read(size):
                     pass
 
         return run
 
     def makestdioread(size):
         def run():
             with open(bundlepath, b'rb') as fh:
                 while fh.read(size):
                     pass
 
         return run
 
     # bundle1
 
     def deltaiter(bundle):
         for delta in bundle.deltaiter():
             pass
 
     def iterchunks(bundle):
         for chunk in bundle.getchunks():
             pass
 
     # bundle2
 
     def forwardchunks(bundle):
         for chunk in bundle._forwardchunks():
             pass
 
     def iterparts(bundle):
         for part in bundle.iterparts():
             pass
 
     def iterpartsseekable(bundle):
         for part in bundle.iterparts(seekable=True):
             pass
 
     def seek(bundle):
         for part in bundle.iterparts(seekable=True):
             part.seek(0, os.SEEK_END)
 
     def makepartreadnbytes(size):
         def run():
             with open(bundlepath, b'rb') as fh:
                 bundle = exchange.readbundle(ui, fh, bundlepath)
                 for part in bundle.iterparts():
                     while part.read(size):
                         pass
 
         return run
 
     benches = [
         (makestdioread(8192), b'read(8k)'),
         (makestdioread(16384), b'read(16k)'),
         (makestdioread(32768), b'read(32k)'),
         (makestdioread(131072), b'read(128k)'),
     ]
 
     with open(bundlepath, b'rb') as fh:
         bundle = exchange.readbundle(ui, fh, bundlepath)
 
         if isinstance(bundle, changegroup.cg1unpacker):
             benches.extend([
                 (makebench(deltaiter), b'cg1 deltaiter()'),
                 (makebench(iterchunks), b'cg1 getchunks()'),
                 (makereadnbytes(8192), b'cg1 read(8k)'),
                 (makereadnbytes(16384), b'cg1 read(16k)'),
                 (makereadnbytes(32768), b'cg1 read(32k)'),
                 (makereadnbytes(131072), b'cg1 read(128k)'),
             ])
         elif isinstance(bundle, bundle2.unbundle20):
             benches.extend([
                 (makebench(forwardchunks), b'bundle2 forwardchunks()'),
                 (makebench(iterparts), b'bundle2 iterparts()'),
                 (makebench(iterpartsseekable), b'bundle2 iterparts() seekable'),
                 (makebench(seek), b'bundle2 part seek()'),
                 (makepartreadnbytes(8192), b'bundle2 part read(8k)'),
                 (makepartreadnbytes(16384), b'bundle2 part read(16k)'),
                 (makepartreadnbytes(32768), b'bundle2 part read(32k)'),
                 (makepartreadnbytes(131072), b'bundle2 part read(128k)'),
             ])
         elif isinstance(bundle, streamclone.streamcloneapplier):
             raise error.Abort(b'stream clone bundles not supported')
         else:
             raise error.Abort(b'unhandled bundle type: %s' % type(bundle))
 
     for fn, title in benches:
         timer, fm = gettimer(ui, opts)
         timer(fn, title=title)
         fm.end()
 
 @command(b'perfchangegroupchangelog', formatteropts +
          [(b'', b'version', b'02', b'changegroup version'),
           (b'r', b'rev', b'', b'revisions to add to changegroup')])
 def perfchangegroupchangelog(ui, repo, version=b'02', rev=None, **opts):
     """Benchmark producing a changelog group for a changegroup.
 
     This measures the time spent processing the changelog during a
     bundle operation. This occurs during `hg bundle` and on a server
     processing a `getbundle` wire protocol request (handles clones
     and pull requests).
 
     By default, all revisions are added to the changegroup.
     """
     cl = repo.changelog
     nodes = [cl.lookup(r) for r in repo.revs(rev or b'all()')]
     bundler = changegroup.getbundler(version, repo)
 
     def d():
         state, chunks = bundler._generatechangelog(cl, nodes)
         for chunk in chunks:
             pass
 
     timer, fm = gettimer(ui, opts)
 
     # Terminal printing can interfere with timing. So disable it.
     with ui.configoverride({(b'progress', b'disable'): True}):
         timer(d)
 
     fm.end()
 
 @command(b'perfdirs', formatteropts)
 def perfdirs(ui, repo, **opts):
     timer, fm = gettimer(ui, opts)
     dirstate = repo.dirstate
     b'a' in dirstate
     def d():
         dirstate.hasdir(b'a')
         del dirstate._map._dirs
     timer(d)
     fm.end()
 
 @command(b'perfdirstate', formatteropts)
 def perfdirstate(ui, repo, **opts):
     timer, fm = gettimer(ui, opts)
     b"a" in repo.dirstate
     def d():
         repo.dirstate.invalidate()
         b"a" in repo.dirstate
     timer(d)
     fm.end()
 
 @command(b'perfdirstatedirs', formatteropts)
 def perfdirstatedirs(ui, repo, **opts):
     timer, fm = gettimer(ui, opts)
     b"a" in repo.dirstate
     def d():
         repo.dirstate.hasdir(b"a")
         del repo.dirstate._map._dirs
     timer(d)
     fm.end()
 
 @command(b'perfdirstatefoldmap', formatteropts)
 def perfdirstatefoldmap(ui, repo, **opts):
     timer, fm = gettimer(ui, opts)
     dirstate = repo.dirstate
     b'a' in dirstate
     def d():
         dirstate._map.filefoldmap.get(b'a')
         del dirstate._map.filefoldmap
     timer(d)
     fm.end()
 
 @command(b'perfdirfoldmap', formatteropts)
 def perfdirfoldmap(ui, repo, **opts):
     timer, fm = gettimer(ui, opts)
     dirstate = repo.dirstate
     b'a' in dirstate
     def d():
         dirstate._map.dirfoldmap.get(b'a')
         del dirstate._map.dirfoldmap
         del dirstate._map._dirs
     timer(d)
     fm.end()
 
 @command(b'perfdirstatewrite', formatteropts)
 def perfdirstatewrite(ui, repo, **opts):
     timer, fm = gettimer(ui, opts)
     ds = repo.dirstate
     b"a" in ds
     def d():
         ds._dirty = True
         ds.write(repo.currenttransaction())
     timer(d)
     fm.end()
 
 @command(b'perfmergecalculate',
          [(b'r', b'rev', b'.', b'rev to merge against')] + formatteropts)
 def perfmergecalculate(ui, repo, rev, **opts):
     timer, fm = gettimer(ui, opts)
     wctx = repo[None]
     rctx = scmutil.revsingle(repo, rev, rev)
     ancestor = wctx.ancestor(rctx)
     # we don't want working dir files to be stat'd in the benchmark, so prime
     # that cache
     wctx.dirty()
     def d():
         # acceptremote is True because we don't want prompts in the middle of
         # our benchmark
         merge.calculateupdates(repo, wctx, rctx, [ancestor], False, False,
                                acceptremote=True, followcopies=True)
     timer(d)
     fm.end()
 
 @command(b'perfpathcopies', [], b"REV REV")
 def perfpathcopies(ui, repo, rev1, rev2, **opts):
     timer, fm = gettimer(ui, opts)
     ctx1 = scmutil.revsingle(repo, rev1, rev1)
     ctx2 = scmutil.revsingle(repo, rev2, rev2)
     def d():
         copies.pathcopies(ctx1, ctx2)
     timer(d)
     fm.end()
 
 @command(b'perfphases',
          [(b'', b'full', False, b'include file reading time too'),
           ], b"")
 def perfphases(ui, repo, **opts):
     """benchmark phasesets computation"""
     timer, fm = gettimer(ui, opts)
     _phases = repo._phasecache
     full = opts.get(b'full')
     def d():
         phases = _phases
         if full:
             clearfilecache(repo, b'_phasecache')
             phases = repo._phasecache
         phases.invalidate()
         phases.loadphaserevs(repo)
     timer(d)
     fm.end()
 
 @command(b'perfphasesremote',
          [], b"[DEST]")
 def perfphasesremote(ui, repo, dest=None, **opts):
     """benchmark time needed to analyse phases of the remote server"""
     from mercurial.node import (
         bin,
     )
     from mercurial import (
         exchange,
         hg,
         phases,
     )
     timer, fm = gettimer(ui, opts)
 
     path = ui.paths.getpath(dest, default=(b'default-push', b'default'))
     if not path:
         raise error.Abort((b'default repository not configured!'),
                           hint=(b"see 'hg help config.paths'"))
     dest = path.pushloc or path.loc
     branches = (path.branch, opts.get(b'branch') or [])
     ui.status((b'analysing phase of %s\n') % util.hidepassword(dest))
     revs, checkout = hg.addbranchrevs(repo, repo, branches, opts.get(b'rev'))
     other = hg.peer(repo, opts, dest)
 
     # easier to perform discovery through the operation
     op = exchange.pushoperation(repo, other)
     exchange._pushdiscoverychangeset(op)
 
     remotesubset = op.fallbackheads
 
     with other.commandexecutor() as e:
         remotephases = e.callcommand(b'listkeys',
                        {b'namespace': b'phases'}).result()
     del other
     publishing = remotephases.get(b'publishing', False)
     if publishing:
         ui.status((b'publishing: yes\n'))
     else:
         ui.status((b'publishing: no\n'))
 
     nodemap = repo.changelog.nodemap
     nonpublishroots = 0
     for nhex, phase in remotephases.iteritems():
         if nhex == b'publishing': # ignore data related to publish option
             continue
         node = bin(nhex)
         if node in nodemap and int(phase):
             nonpublishroots += 1
     ui.status((b'number of roots: %d\n') % len(remotephases))
     ui.status((b'number of known non public roots: %d\n') % nonpublishroots)
     def d():
         phases.remotephasessummary(repo,
                                    remotesubset,
                                    remotephases)
     timer(d)
     fm.end()
 
 @command(b'perfmanifest',[
             (b'm', b'manifest-rev', False, b'Look up a manifest node revision'),
             (b'', b'clear-disk', False, b'clear on-disk caches too'),
          ], b'REV|NODE')
 def perfmanifest(ui, repo, rev, manifest_rev=False, clear_disk=False, **opts):
     """benchmark the time to read a manifest from disk and return a usable
     dict-like object
 
     Manifest caches are cleared before retrieval."""
     timer, fm = gettimer(ui, opts)
     if not manifest_rev:
         ctx = scmutil.revsingle(repo, rev, rev)
         t = ctx.manifestnode()
     else:
         from mercurial.node import bin
 
         if len(rev) == 40:
             t = bin(rev)
         else:
             try:
                 rev = int(rev)
 
                 if util.safehasattr(repo.manifestlog, b'getstorage'):
                     t = repo.manifestlog.getstorage(b'').node(rev)
                 else:
                     t = repo.manifestlog._revlog.lookup(rev)
             except ValueError:
                 raise error.Abort(b'manifest revision must be integer or full '
                                   b'node')
     def d():
         repo.manifestlog.clearcaches(clear_persisted_data=clear_disk)
         repo.manifestlog[t].read()
     timer(d)
     fm.end()
 
 @command(b'perfchangeset', formatteropts)
 def perfchangeset(ui, repo, rev, **opts):
     timer, fm = gettimer(ui, opts)
     n = scmutil.revsingle(repo, rev).node()
     def d():
         repo.changelog.read(n)
         #repo.changelog._cache = None
     timer(d)
     fm.end()
 
 @command(b'perfindex', formatteropts)
 def perfindex(ui, repo, **opts):
     import mercurial.revlog
     timer, fm = gettimer(ui, opts)
     mercurial.revlog._prereadsize = 2**24 # disable lazy parser in old hg
     n = repo[b"tip"].node()
     svfs = getsvfs(repo)
     def d():
         cl = mercurial.revlog.revlog(svfs, b"00changelog.i")
         cl.rev(n)
     timer(d)
     fm.end()
 
 @command(b'perfstartup', formatteropts)
 def perfstartup(ui, repo, **opts):
     timer, fm = gettimer(ui, opts)
     cmd = sys.argv[0]
     def d():
         if os.name != b'nt':
             os.system(b"HGRCPATH= %s version -q > /dev/null" % cmd)
         else:
             os.environ[b'HGRCPATH'] = b' '
             os.system(b"%s version -q > NUL" % cmd)
     timer(d)
     fm.end()
 
 @command(b'perfparents', formatteropts)
 def perfparents(ui, repo, **opts):
     timer, fm = gettimer(ui, opts)
     # control the number of commits perfparents iterates over
     # experimental config: perf.parentscount
     count = getint(ui, b"perf", b"parentscount", 1000)
     if len(repo.changelog) < count:
         raise error.Abort(b"repo needs %d commits for this test" % count)
     repo = repo.unfiltered()
     nl = [repo.changelog.node(i) for i in xrange(count)]
     def d():
         for n in nl:
             repo.changelog.parents(n)
     timer(d)
     fm.end()
 
 @command(b'perfctxfiles', formatteropts)
 def perfctxfiles(ui, repo, x, **opts):
     x = int(x)
     timer, fm = gettimer(ui, opts)
     def d():
         len(repo[x].files())
     timer(d)
     fm.end()
 
 @command(b'perfrawfiles', formatteropts)
 def perfrawfiles(ui, repo, x, **opts):
     x = int(x)
     timer, fm = gettimer(ui, opts)
     cl = repo.changelog
     def d():
         len(cl.read(x)[3])
     timer(d)
     fm.end()
 
 @command(b'perflookup', formatteropts)
 def perflookup(ui, repo, rev, **opts):
     timer, fm = gettimer(ui, opts)
     timer(lambda: len(repo.lookup(rev)))
     fm.end()
 
 @command(b'perflinelogedits',
          [(b'n', b'edits', 10000, b'number of edits'),
           (b'', b'max-hunk-lines', 10, b'max lines in a hunk'),
           ], norepo=True)
 def perflinelogedits(ui, **opts):
     from mercurial import linelog
 
     edits = opts[b'edits']
     maxhunklines = opts[b'max_hunk_lines']
 
     maxb1 = 100000
     random.seed(0)
     randint = random.randint
     currentlines = 0
     arglist = []
     for rev in xrange(edits):
         a1 = randint(0, currentlines)
         a2 = randint(a1, min(currentlines, a1 + maxhunklines))
         b1 = randint(0, maxb1)
         b2 = randint(b1, b1 + maxhunklines)
         currentlines += (b2 - b1) - (a2 - a1)
         arglist.append((rev, a1, a2, b1, b2))
 
     def d():
         ll = linelog.linelog()
         for args in arglist:
             ll.replacelines(*args)
 
     timer, fm = gettimer(ui, opts)
     timer(d)
     fm.end()
 
 @command(b'perfrevrange', formatteropts)
 def perfrevrange(ui, repo, *specs, **opts):
     timer, fm = gettimer(ui, opts)
     revrange = scmutil.revrange
     timer(lambda: len(revrange(repo, specs)))
     fm.end()
 
 @command(b'perfnodelookup', formatteropts)
 def perfnodelookup(ui, repo, rev, **opts):
     timer, fm = gettimer(ui, opts)
     import mercurial.revlog
     mercurial.revlog._prereadsize = 2**24 # disable lazy parser in old hg
     n = scmutil.revsingle(repo, rev).node()
     cl = mercurial.revlog.revlog(getsvfs(repo), b"00changelog.i")
     def d():
         cl.rev(n)
         clearcaches(cl)
     timer(d)
     fm.end()
 
 @command(b'perflog',
          [(b'', b'rename', False, b'ask log to follow renames')
          ] + formatteropts)
 def perflog(ui, repo, rev=None, **opts):
     if rev is None:
         rev=[]
     timer, fm = gettimer(ui, opts)
     ui.pushbuffer()
     timer(lambda: commands.log(ui, repo, rev=rev, date=b'', user=b'',
                                copies=opts.get(b'rename')))
     ui.popbuffer()
     fm.end()
 
 @command(b'perfmoonwalk', formatteropts)
 def perfmoonwalk(ui, repo, **opts):
     """benchmark walking the changelog backwards
 
     This also loads the changelog data for each revision in the changelog.
     """
     timer, fm = gettimer(ui, opts)
     def moonwalk():
         for i in repo.changelog.revs(start=(len(repo) - 1), stop=-1):
             ctx = repo[i]
             ctx.branch() # read changelog data (in addition to the index)
     timer(moonwalk)
     fm.end()
 
 @command(b'perftemplating',
          [(b'r', b'rev', [], b'revisions to run the template on'),
           ] + formatteropts)
 def perftemplating(ui, repo, testedtemplate=None, **opts):
     """test the rendering time of a given template"""
     if makelogtemplater is None:
         raise error.Abort((b"perftemplating not available with this Mercurial"),
                           hint=b"use 4.3 or later")
 
     nullui = ui.copy()
     nullui.fout = open(os.devnull, b'wb')
     nullui.disablepager()
     revs = opts.get(b'rev')
     if not revs:
         revs = [b'all()']
     revs = list(scmutil.revrange(repo, revs))
 
     defaulttemplate = (b'{date|shortdate} [{rev}:{node|short}]'
                        b' {author|person}: {desc|firstline}\n')
     if testedtemplate is None:
         testedtemplate = defaulttemplate
     displayer = makelogtemplater(nullui, repo, testedtemplate)
     def format():
         for r in revs:
             ctx = repo[r]
             displayer.show(ctx)
             displayer.flush(ctx)
 
     timer, fm = gettimer(ui, opts)
     timer(format)
     fm.end()
 
 @command(b'perfcca', formatteropts)
 def perfcca(ui, repo, **opts):
     timer, fm = gettimer(ui, opts)
     timer(lambda: scmutil.casecollisionauditor(ui, False, repo.dirstate))
     fm.end()
 
 @command(b'perffncacheload', formatteropts)
 def perffncacheload(ui, repo, **opts):
     timer, fm = gettimer(ui, opts)
     s = repo.store
     def d():
         s.fncache._load()
     timer(d)
     fm.end()
 
 @command(b'perffncachewrite', formatteropts)
 def perffncachewrite(ui, repo, **opts):
     timer, fm = gettimer(ui, opts)
     s = repo.store
     lock = repo.lock()
     s.fncache._load()
     tr = repo.transaction(b'perffncachewrite')
     tr.addbackup(b'fncache')
     def d():
         s.fncache._dirty = True
         s.fncache.write(tr)
     timer(d)
     tr.close()
     lock.release()
     fm.end()
 
 @command(b'perffncacheencode', formatteropts)
 def perffncacheencode(ui, repo, **opts):
     timer, fm = gettimer(ui, opts)
     s = repo.store
     s.fncache._load()
     def d():
         for p in s.fncache.entries:
             s.encode(p)
     timer(d)
     fm.end()
 
 def _bdiffworker(q, blocks, xdiff, ready, done):
     while not done.is_set():
         pair = q.get()
         while pair is not None:
             if xdiff:
                 mdiff.bdiff.xdiffblocks(*pair)
             elif blocks:
                 mdiff.bdiff.blocks(*pair)
             else:
                 mdiff.textdiff(*pair)
             q.task_done()
             pair = q.get()
         q.task_done() # for the None one
         with ready:
             ready.wait()
 
 def _manifestrevision(repo, mnode):
     ml = repo.manifestlog
 
     if util.safehasattr(ml, b'getstorage'):
         store = ml.getstorage(b'')
     else:
         store = ml._revlog
 
     return store.revision(mnode)
 
 @command(b'perfbdiff', revlogopts + formatteropts + [
     (b'', b'count', 1, b'number of revisions to test (when using --startrev)'),
     (b'', b'alldata', False, b'test bdiffs for all associated revisions'),
     (b'', b'threads', 0, b'number of thread to use (disable with 0)'),
     (b'', b'blocks', False, b'test computing diffs into blocks'),
     (b'', b'xdiff', False, b'use xdiff algorithm'),
     ],
 
     b'-c|-m|FILE REV')
 def perfbdiff(ui, repo, file_, rev=None, count=None, threads=0, **opts):
     """benchmark a bdiff between revisions
 
     By default, benchmark a bdiff between its delta parent and itself.
 
     With ``--count``, benchmark bdiffs between delta parents and self for N
     revisions starting at the specified revision.
 
     With ``--alldata``, assume the requested revision is a changeset and
     measure bdiffs for all changes related to that changeset (manifest
     and filelogs).
     """
     opts = pycompat.byteskwargs(opts)
 
     if opts[b'xdiff'] and not opts[b'blocks']:
         raise error.CommandError(b'perfbdiff', b'--xdiff requires --blocks')
 
     if opts[b'alldata']:
         opts[b'changelog'] = True
 
     if opts.get(b'changelog') or opts.get(b'manifest'):
         file_, rev = None, file_
     elif rev is None:
         raise error.CommandError(b'perfbdiff', b'invalid arguments')
 
     blocks = opts[b'blocks']
     xdiff = opts[b'xdiff']
     textpairs = []
 
     r = cmdutil.openrevlog(repo, b'perfbdiff', file_, opts)
 
     startrev = r.rev(r.lookup(rev))
     for rev in range(startrev, min(startrev + count, len(r) - 1)):
         if opts[b'alldata']:
             # Load revisions associated with changeset.
             ctx = repo[rev]
             mtext = _manifestrevision(repo, ctx.manifestnode())
             for pctx in ctx.parents():
                 pman = _manifestrevision(repo, pctx.manifestnode())
                 textpairs.append((pman, mtext))
 
             # Load filelog revisions by iterating manifest delta.
             man = ctx.manifest()
             pman = ctx.p1().manifest()
             for filename, change in pman.diff(man).items():
                 fctx = repo.file(filename)
                 f1 = fctx.revision(change[0][0] or -1)
                 f2 = fctx.revision(change[1][0] or -1)
                 textpairs.append((f1, f2))
         else:
             dp = r.deltaparent(rev)
             textpairs.append((r.revision(dp), r.revision(rev)))
 
     withthreads = threads > 0
     if not withthreads:
         def d():
             for pair in textpairs:
                 if xdiff:
                     mdiff.bdiff.xdiffblocks(*pair)
                 elif blocks:
                     mdiff.bdiff.blocks(*pair)
                 else:
                     mdiff.textdiff(*pair)
     else:
         q = queue()
         for i in xrange(threads):
             q.put(None)
         ready = threading.Condition()
         done = threading.Event()
         for i in xrange(threads):
             threading.Thread(target=_bdiffworker,
                              args=(q, blocks, xdiff, ready, done)).start()
         q.join()
         def d():
             for pair in textpairs:
                 q.put(pair)
             for i in xrange(threads):
                 q.put(None)
             with ready:
                 ready.notify_all()
             q.join()
     timer, fm = gettimer(ui, opts)
     timer(d)
     fm.end()
 
     if withthreads:
         done.set()
         for i in xrange(threads):
             q.put(None)
         with ready:
             ready.notify_all()
 
 @command(b'perfunidiff', revlogopts + formatteropts + [
     (b'', b'count', 1, b'number of revisions to test (when using --startrev)'),
     (b'', b'alldata', False, b'test unidiffs for all associated revisions'),
     ], b'-c|-m|FILE REV')
 def perfunidiff(ui, repo, file_, rev=None, count=None, **opts):
     """benchmark a unified diff between revisions
 
     This doesn't include any copy tracing - it's just a unified diff
     of the texts.
 
     By default, benchmark a diff between its delta parent and itself.
 
     With ``--count``, benchmark diffs between delta parents and self for N
     revisions starting at the specified revision.
 
     With ``--alldata``, assume the requested revision is a changeset and
     measure diffs for all changes related to that changeset (manifest
     and filelogs).
     """
     if opts[b'alldata']:
         opts[b'changelog'] = True
 
     if opts.get(b'changelog') or opts.get(b'manifest'):
         file_, rev = None, file_
     elif rev is None:
         raise error.CommandError(b'perfunidiff', b'invalid arguments')
 
     textpairs = []
 
     r = cmdutil.openrevlog(repo, b'perfunidiff', file_, opts)
 
     startrev = r.rev(r.lookup(rev))
     for rev in range(startrev, min(startrev + count, len(r) - 1)):
         if opts[b'alldata']:
             # Load revisions associated with changeset.
             ctx = repo[rev]
             mtext = _manifestrevision(repo, ctx.manifestnode())
             for pctx in ctx.parents():
                 pman = _manifestrevision(repo, pctx.manifestnode())
                 textpairs.append((pman, mtext))
 
             # Load filelog revisions by iterating manifest delta.
             man = ctx.manifest()
             pman = ctx.p1().manifest()
             for filename, change in pman.diff(man).items():
                 fctx = repo.file(filename)
                 f1 = fctx.revision(change[0][0] or -1)
                 f2 = fctx.revision(change[1][0] or -1)
                 textpairs.append((f1, f2))
         else:
             dp = r.deltaparent(rev)
             textpairs.append((r.revision(dp), r.revision(rev)))
 
     def d():
         for left, right in textpairs:
             # The date strings don't matter, so we pass empty strings.
             headerlines, hunks = mdiff.unidiff(
                 left, b'', right, b'', b'left', b'right', binary=False)
             # consume iterators in roughly the way patch.py does
             b'\n'.join(headerlines)
             b''.join(sum((list(hlines) for hrange, hlines in hunks), []))
     timer, fm = gettimer(ui, opts)
     timer(d)
     fm.end()
 
 @command(b'perfdiffwd', formatteropts)
 def perfdiffwd(ui, repo, **opts):
     """Profile diff of working directory changes"""
     timer, fm = gettimer(ui, opts)
     options = {
         b'w': b'ignore_all_space',
         b'b': b'ignore_space_change',
         b'B': b'ignore_blank_lines',
         }
 
     for diffopt in (b'', b'w', b'b', b'B', b'wB'):
         opts = dict((options[c], b'1') for c in diffopt)
         def d():
             ui.pushbuffer()
             commands.diff(ui, repo, **opts)
             ui.popbuffer()
         title = b'diffopts: %s' % (diffopt and (b'-' + diffopt) or b'none')
         timer(d, title)
     fm.end()
 
 @command(b'perfrevlogindex', revlogopts + formatteropts,
          b'-c|-m|FILE')
 def perfrevlogindex(ui, repo, file_=None, **opts):
     """Benchmark operations against a revlog index.
 
     This tests constructing a revlog instance, reading index data,
     parsing index data, and performing various operations related to
     index data.
     """
 
     rl = cmdutil.openrevlog(repo, b'perfrevlogindex', file_, opts)
 
     opener = getattr(rl, 'opener')  # trick linter
     indexfile = rl.indexfile
     data = opener.read(indexfile)
 
     header = struct.unpack(b'>I', data[0:4])[0]
     version = header & 0xFFFF
     if version == 1:
         revlogio = revlog.revlogio()
         inline = header & (1 << 16)
     else:
         raise error.Abort((b'unsupported revlog version: %d') % version)
 
     rllen = len(rl)
 
     node0 = rl.node(0)
     node25 = rl.node(rllen // 4)
     node50 = rl.node(rllen // 2)
     node75 = rl.node(rllen // 4 * 3)
     node100 = rl.node(rllen - 1)
 
     allrevs = range(rllen)
     allrevsrev = list(reversed(allrevs))
     allnodes = [rl.node(rev) for rev in range(rllen)]
     allnodesrev = list(reversed(allnodes))
 
     def constructor():
         revlog.revlog(opener, indexfile)
 
     def read():
         with opener(indexfile) as fh:
             fh.read()
 
     def parseindex():
         revlogio.parseindex(data, inline)
 
     def getentry(revornode):
         index = revlogio.parseindex(data, inline)[0]
         index[revornode]
 
     def getentries(revs, count=1):
         index = revlogio.parseindex(data, inline)[0]
 
         for i in range(count):
             for rev in revs:
                 index[rev]
 
     def resolvenode(node):
         nodemap = revlogio.parseindex(data, inline)[1]
         # This only works for the C code.
         if nodemap is None:
             return
 
         try:
             nodemap[node]
         except error.RevlogError:
             pass
 
     def resolvenodes(nodes, count=1):
         nodemap = revlogio.parseindex(data, inline)[1]
         if nodemap is None:
             return
 
         for i in range(count):
             for node in nodes:
                 try:
                     nodemap[node]
                 except error.RevlogError:
                     pass
 
     benches = [
         (constructor, b'revlog constructor'),
         (read, b'read'),
         (parseindex, b'create index object'),
         (lambda: getentry(0), b'retrieve index entry for rev 0'),
         (lambda: resolvenode(b'a' * 20), b'look up missing node'),
         (lambda: resolvenode(node0), b'look up node at rev 0'),
         (lambda: resolvenode(node25), b'look up node at 1/4 len'),
         (lambda: resolvenode(node50), b'look up node at 1/2 len'),
         (lambda: resolvenode(node75), b'look up node at 3/4 len'),
         (lambda: resolvenode(node100), b'look up node at tip'),
         # 2x variation is to measure caching impact.
         (lambda: resolvenodes(allnodes),
          b'look up all nodes (forward)'),
         (lambda: resolvenodes(allnodes, 2),
          b'look up all nodes 2x (forward)'),
         (lambda: resolvenodes(allnodesrev),
          b'look up all nodes (reverse)'),
         (lambda: resolvenodes(allnodesrev, 2),
          b'look up all nodes 2x (reverse)'),
         (lambda: getentries(allrevs),
          b'retrieve all index entries (forward)'),
         (lambda: getentries(allrevs, 2),
          b'retrieve all index entries 2x (forward)'),
         (lambda: getentries(allrevsrev),
          b'retrieve all index entries (reverse)'),
         (lambda: getentries(allrevsrev, 2),
          b'retrieve all index entries 2x (reverse)'),
     ]
 
     for fn, title in benches:
         timer, fm = gettimer(ui, opts)
         timer(fn, title=title)
         fm.end()
 
 @command(b'perfrevlogrevisions', revlogopts + formatteropts +
          [(b'd', b'dist', 100, b'distance between the revisions'),
           (b's', b'startrev', 0, b'revision to start reading at'),
           (b'', b'reverse', False, b'read in reverse')],
          b'-c|-m|FILE')
 def perfrevlogrevisions(ui, repo, file_=None, startrev=0, reverse=False,
                         **opts):
     """Benchmark reading a series of revisions from a revlog.
 
     By default, we read every ``-d/--dist`` revision from 0 to tip of
     the specified revlog.
 
     The start revision can be defined via ``-s/--startrev``.
     """
     rl = cmdutil.openrevlog(repo, b'perfrevlogrevisions', file_, opts)
     rllen = getlen(ui)(rl)
 
     def d():
         rl.clearcaches()
 
         beginrev = startrev
         endrev = rllen
         dist = opts[b'dist']
 
         if reverse:
             beginrev, endrev = endrev, beginrev
             dist = -1 * dist
 
         for x in xrange(beginrev, endrev, dist):
             # Old revisions don't support passing int.
             n = rl.node(x)
             rl.revision(n)
 
     timer, fm = gettimer(ui, opts)
     timer(d)
     fm.end()
 
 @command(b'perfrevlogchunks', revlogopts + formatteropts +
          [(b'e', b'engines', b'', b'compression engines to use'),
           (b's', b'startrev', 0, b'revision to start at')],
          b'-c|-m|FILE')
 def perfrevlogchunks(ui, repo, file_=None, engines=None, startrev=0, **opts):
     """Benchmark operations on revlog chunks.
 
     Logically, each revlog is a collection of fulltext revisions. However,
     stored within each revlog are "chunks" of possibly compressed data. This
     data needs to be read and decompressed or compressed and written.
 
     This command measures the time it takes to read+decompress and recompress
     chunks in a revlog. It effectively isolates I/O and compression performance.
     For measurements of higher-level operations like resolving revisions,
     see ``perfrevlogrevisions`` and ``perfrevlogrevision``.
     """
     rl = cmdutil.openrevlog(repo, b'perfrevlogchunks', file_, opts)
 
     # _chunkraw was renamed to _getsegmentforrevs.
     try:
         segmentforrevs = rl._getsegmentforrevs
     except AttributeError:
         segmentforrevs = rl._chunkraw
 
     # Verify engines argument.
     if engines:
         engines = set(e.strip() for e in engines.split(b','))
         for engine in engines:
             try:
                 util.compressionengines[engine]
             except KeyError:
                 raise error.Abort(b'unknown compression engine: %s' % engine)
     else:
         engines = []
         for e in util.compengines:
             engine = util.compengines[e]
             try:
                 if engine.available():
                     engine.revlogcompressor().compress(b'dummy')
                     engines.append(e)
             except NotImplementedError:
                 pass
 
     revs = list(rl.revs(startrev, len(rl) - 1))
 
     def rlfh(rl):
         if rl._inline:
             return getsvfs(repo)(rl.indexfile)
         else:
             return getsvfs(repo)(rl.datafile)
 
     def doread():
         rl.clearcaches()
         for rev in revs:
             segmentforrevs(rev, rev)
 
     def doreadcachedfh():
         rl.clearcaches()
         fh = rlfh(rl)
         for rev in revs:
             segmentforrevs(rev, rev, df=fh)
 
     def doreadbatch():
         rl.clearcaches()
         segmentforrevs(revs[0], revs[-1])
 
     def doreadbatchcachedfh():
         rl.clearcaches()
         fh = rlfh(rl)
         segmentforrevs(revs[0], revs[-1], df=fh)
 
     def dochunk():
         rl.clearcaches()
         fh = rlfh(rl)
         for rev in revs:
             rl._chunk(rev, df=fh)
 
     chunks = [None]
 
     def dochunkbatch():
         rl.clearcaches()
         fh = rlfh(rl)
         # Save chunks as a side-effect.
         chunks[0] = rl._chunks(revs, df=fh)
 
     def docompress(compressor):
         rl.clearcaches()
 
         try:
             # Swap in the requested compression engine.
             oldcompressor = rl._compressor
             rl._compressor = compressor
             for chunk in chunks[0]:
                 rl.compress(chunk)
         finally:
             rl._compressor = oldcompressor
 
     benches = [
         (lambda: doread(), b'read'),
         (lambda: doreadcachedfh(), b'read w/ reused fd'),
         (lambda: doreadbatch(), b'read batch'),
         (lambda: doreadbatchcachedfh(), b'read batch w/ reused fd'),
         (lambda: dochunk(), b'chunk'),
         (lambda: dochunkbatch(), b'chunk batch'),
     ]
 
     for engine in sorted(engines):
         compressor = util.compengines[engine].revlogcompressor()
         benches.append((functools.partial(docompress, compressor),
                         b'compress w/ %s' % engine))
 
     for fn, title in benches:
         timer, fm = gettimer(ui, opts)
         timer(fn, title=title)
         fm.end()
 
 @command(b'perfrevlogrevision', revlogopts + formatteropts +
          [(b'', b'cache', False, b'use caches instead of clearing')],
          b'-c|-m|FILE REV')
 def perfrevlogrevision(ui, repo, file_, rev=None, cache=None, **opts):
     """Benchmark obtaining a revlog revision.
 
     Obtaining a revlog revision consists of roughly the following steps:
 
     1. Compute the delta chain
     2. Obtain the raw chunks for that delta chain
     3. Decompress each raw chunk
     4. Apply binary patches to obtain fulltext
     5. Verify hash of fulltext
 
     This command measures the time spent in each of these phases.
     """
     if opts.get(b'changelog') or opts.get(b'manifest'):
         file_, rev = None, file_
     elif rev is None:
         raise error.CommandError(b'perfrevlogrevision', b'invalid arguments')
 
     r = cmdutil.openrevlog(repo, b'perfrevlogrevision', file_, opts)
 
     # _chunkraw was renamed to _getsegmentforrevs.
     try:
         segmentforrevs = r._getsegmentforrevs
     except AttributeError:
         segmentforrevs = r._chunkraw
 
     node = r.lookup(rev)
     rev = r.rev(node)
 
     def getrawchunks(data, chain):
         start = r.start
         length = r.length
         inline = r._inline
         iosize = r._io.size
         buffer = util.buffer
         offset = start(chain[0])
 
         chunks = []
         ladd = chunks.append
 
         for rev in chain:
             chunkstart = start(rev)
             if inline:
                 chunkstart += (rev + 1) * iosize
             chunklength = length(rev)
             ladd(buffer(data, chunkstart - offset, chunklength))
 
         return chunks
 
     def dodeltachain(rev):
         if not cache:
             r.clearcaches()
         r._deltachain(rev)
 
     def doread(chain):
         if not cache:
             r.clearcaches()
         segmentforrevs(chain[0], chain[-1])
 
     def dorawchunks(data, chain):
         if not cache:
             r.clearcaches()
         getrawchunks(data, chain)
 
     def dodecompress(chunks):
         decomp = r.decompress
         for chunk in chunks:
             decomp(chunk)
 
     def dopatch(text, bins):
         if not cache:
             r.clearcaches()
         mdiff.patches(text, bins)
 
     def dohash(text):
         if not cache:
             r.clearcaches()
         r.checkhash(text, node, rev=rev)
 
     def dorevision():
         if not cache:
             r.clearcaches()
         r.revision(node)
 
     chain = r._deltachain(rev)[0]
     data = segmentforrevs(chain[0], chain[-1])[1]
     rawchunks = getrawchunks(data, chain)
     bins = r._chunks(chain)
     text = str(bins[0])
     bins = bins[1:]
     text = mdiff.patches(text, bins)
 
     benches = [
         (lambda: dorevision(), b'full'),
         (lambda: dodeltachain(rev), b'deltachain'),
         (lambda: doread(chain), b'read'),
         (lambda: dorawchunks(data, chain), b'rawchunks'),
         (lambda: dodecompress(rawchunks), b'decompress'),
         (lambda: dopatch(text, bins), b'patch'),
         (lambda: dohash(text), b'hash'),
     ]
 
     for fn, title in benches:
         timer, fm = gettimer(ui, opts)
         timer(fn, title=title)
         fm.end()
 
 @command(b'perfrevset',
          [(b'C', b'clear', False, b'clear volatile cache between each call.'),
           (b'', b'contexts', False, b'obtain changectx for each revision')]
          + formatteropts, b"REVSET")
 def perfrevset(ui, repo, expr, clear=False, contexts=False, **opts):
     """benchmark the execution time of a revset
 
     Use the --clean option if need to evaluate the impact of build volatile
     revisions set cache on the revset execution. Volatile cache hold filtered
     and obsolete related cache."""
     timer, fm = gettimer(ui, opts)
     def d():
         if clear:
             repo.invalidatevolatilesets()
         if contexts:
             for ctx in repo.set(expr): pass
         else:
             for r in repo.revs(expr): pass
     timer(d)
     fm.end()
 
 @command(b'perfvolatilesets',
          [(b'', b'clear-obsstore', False, b'drop obsstore between each call.'),
           ] + formatteropts)
 def perfvolatilesets(ui, repo, *names, **opts):
     """benchmark the computation of various volatile set
 
     Volatile set computes element related to filtering and obsolescence."""
     timer, fm = gettimer(ui, opts)
     repo = repo.unfiltered()
 
     def getobs(name):
         def d():
             repo.invalidatevolatilesets()
             if opts[b'clear_obsstore']:
                 clearfilecache(repo, b'obsstore')
             obsolete.getrevs(repo, name)
         return d
 
     allobs = sorted(obsolete.cachefuncs)
     if names:
         allobs = [n for n in allobs if n in names]
 
     for name in allobs:
         timer(getobs(name), title=name)
 
     def getfiltered(name):
         def d():
             repo.invalidatevolatilesets()
             if opts[b'clear_obsstore']:
                 clearfilecache(repo, b'obsstore')
             repoview.filterrevs(repo, name)
         return d
 
     allfilter = sorted(repoview.filtertable)
     if names:
         allfilter = [n for n in allfilter if n in names]
 
     for name in allfilter:
         timer(getfiltered(name), title=name)
     fm.end()
 
 @command(b'perfbranchmap',
          [(b'f', b'full', False,
            b'Includes build time of subset'),
           (b'', b'clear-revbranch', False,
            b'purge the revbranch cache between computation'),
           ] + formatteropts)
 def perfbranchmap(ui, repo, *filternames, **opts):
     """benchmark the update of a branchmap
 
     This benchmarks the full repo.branchmap() call with read and write disabled
     """
     full = opts.get(b"full", False)
     clear_revbranch = opts.get(b"clear_revbranch", False)
     timer, fm = gettimer(ui, opts)
     def getbranchmap(filtername):
         """generate a benchmark function for the filtername"""
         if filtername is None:
             view = repo
         else:
             view = repo.filtered(filtername)
         def d():
             if clear_revbranch:
                 repo.revbranchcache()._clear()
             if full:
                 view._branchcaches.clear()
             else:
                 view._branchcaches.pop(filtername, None)
             view.branchmap()
         return d
     # add filter in smaller subset to bigger subset
     possiblefilters = set(repoview.filtertable)
     if filternames:
         possiblefilters &= set(filternames)
     subsettable = getbranchmapsubsettable()
     allfilters = []
     while possiblefilters:
         for name in possiblefilters:
             subset = subsettable.get(name)
             if subset not in possiblefilters:
                 break
         else:
             assert False, b'subset cycle %s!' % possiblefilters
         allfilters.append(name)
         possiblefilters.remove(name)
 
     # warm the cache
     if not full:
         for name in allfilters:
             repo.filtered(name).branchmap()
     if not filternames or b'unfiltered' in filternames:
         # add unfiltered
         allfilters.append(None)
 
     branchcacheread = safeattrsetter(branchmap, b'read')
     branchcachewrite = safeattrsetter(branchmap.branchcache, b'write')
     branchcacheread.set(lambda repo: None)
     branchcachewrite.set(lambda bc, repo: None)
     try:
         for name in allfilters:
             printname = name
             if name is None:
                 printname = b'unfiltered'
             timer(getbranchmap(name), title=str(printname))
     finally:
         branchcacheread.restore()
         branchcachewrite.restore()
     fm.end()
 
 @command(b'perfbranchmapload', [
      (b'f', b'filter', b'', b'Specify repoview filter'),
      (b'', b'list', False, b'List brachmap filter caches'),
     ] + formatteropts)
 def perfbranchmapread(ui, repo, filter=b'', list=False, **opts):
     """benchmark reading the branchmap"""
     if list:
         for name, kind, st in repo.cachevfs.readdir(stat=True):
             if name.startswith(b'branch2'):
                 filtername = name.partition(b'-')[2] or b'unfiltered'
                 ui.status(b'%s - %s\n'
                           % (filtername, util.bytecount(st.st_size)))
         return
     if filter:
         repo = repoview.repoview(repo, filter)
     else:
         repo = repo.unfiltered()
     # try once without timer, the filter may not be cached
     if branchmap.read(repo) is None:
         raise error.Abort(b'No brachmap cached for %s repo'
                           % (filter or b'unfiltered'))
     timer, fm = gettimer(ui, opts)
     timer(lambda: branchmap.read(repo) and None)
     fm.end()
 
 @command(b'perfloadmarkers')
 def perfloadmarkers(ui, repo):
     """benchmark the time to parse the on-disk markers for a repo
 
     Result is the number of markers in the repo."""
     timer, fm = gettimer(ui)
     svfs = getsvfs(repo)
     timer(lambda: len(obsolete.obsstore(svfs)))
     fm.end()
 
 @command(b'perflrucachedict', formatteropts +
     [(b'', b'costlimit', 0, b'maximum total cost of items in cache'),
      (b'', b'mincost', 0, b'smallest cost of items in cache'),
      (b'', b'maxcost', 100, b'maximum cost of items in cache'),
      (b'', b'size', 4, b'size of cache'),
      (b'', b'gets', 10000, b'number of key lookups'),
      (b'', b'sets', 10000, b'number of key sets'),
      (b'', b'mixed', 10000, b'number of mixed mode operations'),
      (b'', b'mixedgetfreq', 50, b'frequency of get vs set ops in mixed mode')],
     norepo=True)
 def perflrucache(ui, mincost=0, maxcost=100, costlimit=0, size=4,
                  gets=10000, sets=10000, mixed=10000, mixedgetfreq=50, **opts):
     def doinit():
         for i in xrange(10000):
             util.lrucachedict(size)
 
     costrange = list(range(mincost, maxcost + 1))
 
     values = []
     for i in xrange(size):
         values.append(random.randint(0, sys.maxint))
 
     # Get mode fills the cache and tests raw lookup performance with no
     # eviction.
     getseq = []
     for i in xrange(gets):
         getseq.append(random.choice(values))
 
     def dogets():
         d = util.lrucachedict(size)
         for v in values:
             d[v] = v
         for key in getseq:
             value = d[key]
             value # silence pyflakes warning
 
     def dogetscost():
         d = util.lrucachedict(size, maxcost=costlimit)
         for i, v in enumerate(values):
             d.insert(v, v, cost=costs[i])
         for key in getseq:
             try:
                 value = d[key]
                 value # silence pyflakes warning
             except KeyError:
                 pass
 
     # Set mode tests insertion speed with cache eviction.
     setseq = []
     costs = []
     for i in xrange(sets):
         setseq.append(random.randint(0, sys.maxint))
         costs.append(random.choice(costrange))
 
     def doinserts():
         d = util.lrucachedict(size)
         for v in setseq:
             d.insert(v, v)
 
     def doinsertscost():
         d = util.lrucachedict(size, maxcost=costlimit)
         for i, v in enumerate(setseq):
             d.insert(v, v, cost=costs[i])
 
     def dosets():
         d = util.lrucachedict(size)
         for v in setseq:
             d[v] = v
 
     # Mixed mode randomly performs gets and sets with eviction.
     mixedops = []
     for i in xrange(mixed):
         r = random.randint(0, 100)
         if r < mixedgetfreq:
             op = 0
         else:
             op = 1
 
         mixedops.append((op,
                          random.randint(0, size * 2),
                          random.choice(costrange)))
 
     def domixed():
         d = util.lrucachedict(size)
 
         for op, v, cost in mixedops:
             if op == 0:
                 try:
                     d[v]
                 except KeyError:
                     pass
             else:
                 d[v] = v
 
     def domixedcost():
         d = util.lrucachedict(size, maxcost=costlimit)
 
         for op, v, cost in mixedops:
             if op == 0:
                 try:
                     d[v]
                 except KeyError:
                     pass
             else:
                 d.insert(v, v, cost=cost)
 
     benches = [
         (doinit, b'init'),
     ]
 
     if costlimit:
         benches.extend([
             (dogetscost, b'gets w/ cost limit'),
             (doinsertscost, b'inserts w/ cost limit'),
             (domixedcost, b'mixed w/ cost limit'),
         ])
     else:
         benches.extend([
             (dogets, b'gets'),
             (doinserts, b'inserts'),
             (dosets, b'sets'),
             (domixed, b'mixed')
         ])
 
     for fn, title in benches:
         timer, fm = gettimer(ui, opts)
         timer(fn, title=title)
         fm.end()
 
 @command(b'perfwrite', formatteropts)
 def perfwrite(ui, repo, **opts):
     """microbenchmark ui.write
     """
     timer, fm = gettimer(ui, opts)
     def write():
         for i in range(100000):
             ui.write((b'Testing write performance\n'))
     timer(write)
     fm.end()
 
 def uisetup(ui):
     if (util.safehasattr(cmdutil, b'openrevlog') and
         not util.safehasattr(commands, b'debugrevlogopts')):
         # for "historical portability":
         # In this case, Mercurial should be 1.9 (or a79fea6b3e77) -
         # 3.7 (or 5606f7d0d063). Therefore, '--dir' option for
         # openrevlog() should cause failure, because it has been
         # available since 3.5 (or 49c583ca48c4).
         def openrevlog(orig, repo, cmd, file_, opts):
             if opts.get(b'dir') and not util.safehasattr(repo, b'dirlog'):
                 raise error.Abort(b"This version doesn't support --dir option",
                                   hint=b"use 3.5 or later")
             return orig(repo, cmd, file_, opts)
         extensions.wrapfunction(cmdutil, b'openrevlog', openrevlog)

mercurial/bundlerepo.py

@@ -1,657 +1,659
 # bundlerepo.py - repository class for viewing uncompressed bundles
 #
 # Copyright 2006, 2007 Benoit Boissinot <[e-mail]>
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 """Repository class for viewing uncompressed bundles.
 
 This provides a read-only repository interface to bundles as if they
 were part of the actual repository.
 """
 
 from __future__ import absolute_import
 
 import os
 import shutil
 
 from .i18n import _
 from .node import nullid
 
 from . import (
     bundle2,
     changegroup,
     changelog,
     cmdutil,
     discovery,
     error,
     exchange,
     filelog,
     localrepo,
     manifest,
     mdiff,
     node as nodemod,
     pathutil,
     phases,
     pycompat,
     revlog,
     util,
     vfs as vfsmod,
 )
 
 class bundlerevlog(revlog.revlog):
     def __init__(self, opener, indexfile, cgunpacker, linkmapper):
         # How it works:
         # To retrieve a revision, we need to know the offset of the revision in
         # the bundle (an unbundle object). We store this offset in the index
         # (start). The base of the delta is stored in the base field.
         #
         # To differentiate a rev in the bundle from a rev in the revlog, we
         # check revision against repotiprev.
         opener = vfsmod.readonlyvfs(opener)
         revlog.revlog.__init__(self, opener, indexfile)
         self.bundle = cgunpacker
         n = len(self)
         self.repotiprev = n - 1
         self.bundlerevs = set() # used by 'bundle()' revset expression
         for deltadata in cgunpacker.deltaiter():
             node, p1, p2, cs, deltabase, delta, flags = deltadata
 
             size = len(delta)
             start = cgunpacker.tell() - size
 
             link = linkmapper(cs)
             if node in self.nodemap:
                 # this can happen if two branches make the same change
                 self.bundlerevs.add(self.nodemap[node])
                 continue
 
             for p in (p1, p2):
                 if p not in self.nodemap:
                     raise error.LookupError(p, self.indexfile,
                                             _("unknown parent"))
 
             if deltabase not in self.nodemap:
                 raise LookupError(deltabase, self.indexfile,
                                   _('unknown delta base'))
 
             baserev = self.rev(deltabase)
             # start, size, full unc. size, base (unused), link, p1, p2, node
             e = (revlog.offset_type(start, flags), size, -1, baserev, link,
                  self.rev(p1), self.rev(p2), node)
             self.index.append(e)
             self.nodemap[node] = n
             self.bundlerevs.add(n)
             n += 1
 
     def _chunk(self, rev, df=None):
         # Warning: in case of bundle, the diff is against what we stored as
         # delta base, not against rev - 1
         # XXX: could use some caching
         if rev <= self.repotiprev:
             return revlog.revlog._chunk(self, rev)
         self.bundle.seek(self.start(rev))
         return self.bundle.read(self.length(rev))
 
     def revdiff(self, rev1, rev2):
         """return or calculate a delta between two revisions"""
         if rev1 > self.repotiprev and rev2 > self.repotiprev:
             # hot path for bundle
             revb = self.index[rev2][3]
             if revb == rev1:
                 return self._chunk(rev2)
         elif rev1 <= self.repotiprev and rev2 <= self.repotiprev:
             return revlog.revlog.revdiff(self, rev1, rev2)
 
         return mdiff.textdiff(self.revision(rev1, raw=True),
                               self.revision(rev2, raw=True))
 
     def revision(self, nodeorrev, _df=None, raw=False):
         """return an uncompressed revision of a given node or revision
         number.
         """
         if isinstance(nodeorrev, int):
             rev = nodeorrev
             node = self.node(rev)
         else:
             node = nodeorrev
             rev = self.rev(node)
 
         if node == nullid:
             return ""
 
         rawtext = None
         chain = []
         iterrev = rev
         # reconstruct the revision if it is from a changegroup
         while iterrev > self.repotiprev:
             if self._cache and self._cache[1] == iterrev:
                 rawtext = self._cache[2]
                 break
             chain.append(iterrev)
             iterrev = self.index[iterrev][3]
         if rawtext is None:
             rawtext = self.baserevision(iterrev)
 
         while chain:
             delta = self._chunk(chain.pop())
             rawtext = mdiff.patches(rawtext, [delta])
 
         text, validatehash = self._processflags(rawtext, self.flags(rev),
                                                 'read', raw=raw)
         if validatehash:
             self.checkhash(text, node, rev=rev)
         self._cache = (node, rev, rawtext)
         return text
 
     def baserevision(self, nodeorrev):
         # Revlog subclasses may override 'revision' method to modify format of
         # content retrieved from revlog. To use bundlerevlog with such class one
         # needs to override 'baserevision' and make more specific call here.
         return revlog.revlog.revision(self, nodeorrev, raw=True)
 
     def addrevision(self, *args, **kwargs):
         raise NotImplementedError
 
     def addgroup(self, *args, **kwargs):
         raise NotImplementedError
 
     def strip(self, *args, **kwargs):
         raise NotImplementedError
 
     def checksize(self):
         raise NotImplementedError
 
 class bundlechangelog(bundlerevlog, changelog.changelog):
     def __init__(self, opener, cgunpacker):
         changelog.changelog.__init__(self, opener)
         linkmapper = lambda x: x
         bundlerevlog.__init__(self, opener, self.indexfile, cgunpacker,
                               linkmapper)
 
     def baserevision(self, nodeorrev):
         # Although changelog doesn't override 'revision' method, some extensions
         # may replace this class with another that does. Same story with
         # manifest and filelog classes.
 
         # This bypasses filtering on changelog.node() and rev() because we need
         # revision text of the bundle base even if it is hidden.
         oldfilter = self.filteredrevs
         try:
             self.filteredrevs = ()
             return changelog.changelog.revision(self, nodeorrev, raw=True)
         finally:
             self.filteredrevs = oldfilter
 
 class bundlemanifest(bundlerevlog, manifest.manifestrevlog):
     def __init__(self, opener, cgunpacker, linkmapper, dirlogstarts=None,
                  dir=''):
         manifest.manifestrevlog.__init__(self, opener, tree=dir)
         bundlerevlog.__init__(self, opener, self.indexfile, cgunpacker,
                               linkmapper)
         if dirlogstarts is None:
             dirlogstarts = {}
             if self.bundle.version == "03":
                 dirlogstarts = _getfilestarts(self.bundle)
         self._dirlogstarts = dirlogstarts
         self._linkmapper = linkmapper
 
     def baserevision(self, nodeorrev):
         node = nodeorrev
         if isinstance(node, int):
             node = self.node(node)
 
         if node in self.fulltextcache:
             result = '%s' % self.fulltextcache[node]
         else:
             result = manifest.manifestrevlog.revision(self, nodeorrev, raw=True)
         return result
 
     def dirlog(self, d):
         if d in self._dirlogstarts:
             self.bundle.seek(self._dirlogstarts[d])
             return bundlemanifest(
                 self.opener, self.bundle, self._linkmapper,
                 self._dirlogstarts, dir=d)
         return super(bundlemanifest, self).dirlog(d)
 
 class bundlefilelog(filelog.filelog):
     def __init__(self, opener, path, cgunpacker, linkmapper):
         filelog.filelog.__init__(self, opener, path)
         self._revlog = bundlerevlog(opener, self.indexfile,
                                     cgunpacker, linkmapper)
 
     def baserevision(self, nodeorrev):
         return filelog.filelog.revision(self, nodeorrev, raw=True)
 
 class bundlepeer(localrepo.localpeer):
     def canpush(self):
         return False
 
 class bundlephasecache(phases.phasecache):
     def __init__(self, *args, **kwargs):
         super(bundlephasecache, self).__init__(*args, **kwargs)
         if util.safehasattr(self, 'opener'):
             self.opener = vfsmod.readonlyvfs(self.opener)
 
     def write(self):
         raise NotImplementedError
 
     def _write(self, fp):
         raise NotImplementedError
 
     def _updateroots(self, phase, newroots, tr):
         self.phaseroots[phase] = newroots
         self.invalidate()
         self.dirty = True
 
 def _getfilestarts(cgunpacker):
     filespos = {}
     for chunkdata in iter(cgunpacker.filelogheader, {}):
         fname = chunkdata['filename']
         filespos[fname] = cgunpacker.tell()
         for chunk in iter(lambda: cgunpacker.deltachunk(None), {}):
             pass
     return filespos
 
 class bundlerepository(object):
     """A repository instance that is a union of a local repo and a bundle.
 
     Instances represent a read-only repository composed of a local repository
     with the contents of a bundle file applied. The repository instance is
     conceptually similar to the state of a repository after an
     ``hg unbundle`` operation. However, the contents of the bundle are never
     applied to the actual base repository.
 
     Instances constructed directly are not usable as repository objects.
     Use instance() or makebundlerepository() to create instances.
     """
     def __init__(self, bundlepath, url, tempparent):
         self._tempparent = tempparent
         self._url = url
 
         self.ui.setconfig('phases', 'publish', False, 'bundlerepo')
 
         self.tempfile = None
         f = util.posixfile(bundlepath, "rb")
         bundle = exchange.readbundle(self.ui, f, bundlepath)
 
         if isinstance(bundle, bundle2.unbundle20):
             self._bundlefile = bundle
             self._cgunpacker = None
 
             cgpart = None
             for part in bundle.iterparts(seekable=True):
                 if part.type == 'changegroup':
                     if cgpart:
                         raise NotImplementedError("can't process "
                                                   "multiple changegroups")
                     cgpart = part
 
                 self._handlebundle2part(bundle, part)
 
             if not cgpart:
                 raise error.Abort(_("No changegroups found"))
 
             # This is required to placate a later consumer, which expects
             # the payload offset to be at the beginning of the changegroup.
             # We need to do this after the iterparts() generator advances
             # because iterparts() will seek to end of payload after the
             # generator returns control to iterparts().
             cgpart.seek(0, os.SEEK_SET)
 
         elif isinstance(bundle, changegroup.cg1unpacker):
             if bundle.compressed():
                 f = self._writetempbundle(bundle.read, '.hg10un',
                                           header='HG10UN')
                 bundle = exchange.readbundle(self.ui, f, bundlepath, self.vfs)
 
             self._bundlefile = bundle
             self._cgunpacker = bundle
         else:
             raise error.Abort(_('bundle type %s cannot be read') %
                               type(bundle))
 
         # dict with the mapping 'filename' -> position in the changegroup.
         self._cgfilespos = {}
 
         self.firstnewrev = self.changelog.repotiprev + 1
         phases.retractboundary(self, None, phases.draft,
                                [ctx.node() for ctx in self[self.firstnewrev:]])
 
     def _handlebundle2part(self, bundle, part):
         if part.type != 'changegroup':
             return
 
         cgstream = part
         version = part.params.get('version', '01')
         legalcgvers = changegroup.supportedincomingversions(self)
         if version not in legalcgvers:
             msg = _('Unsupported changegroup version: %s')
             raise error.Abort(msg % version)
         if bundle.compressed():
             cgstream = self._writetempbundle(part.read, '.cg%sun' % version)
 
         self._cgunpacker = changegroup.getunbundler(version, cgstream, 'UN')
 
     def _writetempbundle(self, readfn, suffix, header=''):
         """Write a temporary file to disk
         """
         fdtemp, temp = self.vfs.mkstemp(prefix="hg-bundle-",
                                         suffix=suffix)
         self.tempfile = temp
 
         with os.fdopen(fdtemp, r'wb') as fptemp:
             fptemp.write(header)
             while True:
                 chunk = readfn(2**18)
                 if not chunk:
                     break
                 fptemp.write(chunk)
 
         return self.vfs.open(self.tempfile, mode="rb")
 
     @localrepo.unfilteredpropertycache
     def _phasecache(self):
         return bundlephasecache(self, self._phasedefaults)
 
     @localrepo.unfilteredpropertycache
     def changelog(self):
         # consume the header if it exists
         self._cgunpacker.changelogheader()
         c = bundlechangelog(self.svfs, self._cgunpacker)
         self.manstart = self._cgunpacker.tell()
         return c
 
-    def _constructmanifest(self):
+    @localrepo.unfilteredpropertycache
+    def manifestlog(self):
         self._cgunpacker.seek(self.manstart)
         # consume the header if it exists
         self._cgunpacker.manifestheader()
         linkmapper = self.unfiltered().changelog.rev
-        m = bundlemanifest(self.svfs, self._cgunpacker, linkmapper)
+        rootstore = bundlemanifest(self.svfs, self._cgunpacker, linkmapper)
         self.filestart = self._cgunpacker.tell()
-        return m
+
+        return manifest.manifestlog(self.svfs, self, rootstore)
 
     def _consumemanifest(self):
         """Consumes the manifest portion of the bundle, setting filestart so the
         file portion can be read."""
         self._cgunpacker.seek(self.manstart)
         self._cgunpacker.manifestheader()
         for delta in self._cgunpacker.deltaiter():
             pass
         self.filestart = self._cgunpacker.tell()
 
     @localrepo.unfilteredpropertycache
     def manstart(self):
         self.changelog
         return self.manstart
 
     @localrepo.unfilteredpropertycache
     def filestart(self):
         self.manifestlog
 
         # If filestart was not set by self.manifestlog, that means the
         # manifestlog implementation did not consume the manifests from the
         # changegroup (ex: it might be consuming trees from a separate bundle2
         # part instead). So we need to manually consume it.
         if r'filestart' not in self.__dict__:
             self._consumemanifest()
 
         return self.filestart
 
     def url(self):
         return self._url
 
     def file(self, f):
         if not self._cgfilespos:
             self._cgunpacker.seek(self.filestart)
             self._cgfilespos = _getfilestarts(self._cgunpacker)
 
         if f in self._cgfilespos:
             self._cgunpacker.seek(self._cgfilespos[f])
             linkmapper = self.unfiltered().changelog.rev
             return bundlefilelog(self.svfs, f, self._cgunpacker, linkmapper)
         else:
             return super(bundlerepository, self).file(f)
 
     def close(self):
         """Close assigned bundle file immediately."""
         self._bundlefile.close()
         if self.tempfile is not None:
             self.vfs.unlink(self.tempfile)
         if self._tempparent:
             shutil.rmtree(self._tempparent, True)
 
     def cancopy(self):
         return False
 
     def peer(self):
         return bundlepeer(self)
 
     def getcwd(self):
         return pycompat.getcwd() # always outside the repo
 
     # Check if parents exist in localrepo before setting
     def setparents(self, p1, p2=nullid):
         p1rev = self.changelog.rev(p1)
         p2rev = self.changelog.rev(p2)
         msg = _("setting parent to node %s that only exists in the bundle\n")
         if self.changelog.repotiprev < p1rev:
             self.ui.warn(msg % nodemod.hex(p1))
         if self.changelog.repotiprev < p2rev:
             self.ui.warn(msg % nodemod.hex(p2))
         return super(bundlerepository, self).setparents(p1, p2)
 
 def instance(ui, path, create, intents=None, createopts=None):
     if create:
         raise error.Abort(_('cannot create new bundle repository'))
     # internal config: bundle.mainreporoot
     parentpath = ui.config("bundle", "mainreporoot")
     if not parentpath:
         # try to find the correct path to the working directory repo
         parentpath = cmdutil.findrepo(pycompat.getcwd())
         if parentpath is None:
             parentpath = ''
     if parentpath:
         # Try to make the full path relative so we get a nice, short URL.
         # In particular, we don't want temp dir names in test outputs.
         cwd = pycompat.getcwd()
         if parentpath == cwd:
             parentpath = ''
         else:
             cwd = pathutil.normasprefix(cwd)
             if parentpath.startswith(cwd):
                 parentpath = parentpath[len(cwd):]
     u = util.url(path)
     path = u.localpath()
     if u.scheme == 'bundle':
         s = path.split("+", 1)
         if len(s) == 1:
             repopath, bundlename = parentpath, s[0]
         else:
             repopath, bundlename = s
     else:
         repopath, bundlename = parentpath, path
 
     return makebundlerepository(ui, repopath, bundlename)
 
 def makebundlerepository(ui, repopath, bundlepath):
     """Make a bundle repository object based on repo and bundle paths."""
     if repopath:
         url = 'bundle:%s+%s' % (util.expandpath(repopath), bundlepath)
     else:
         url = 'bundle:%s' % bundlepath
 
     # Because we can't make any guarantees about the type of the base
     # repository, we can't have a static class representing the bundle
     # repository. We also can't make any guarantees about how to even
     # call the base repository's constructor!
     #
     # So, our strategy is to go through ``localrepo.instance()`` to construct
     # a repo instance. Then, we dynamically create a new type derived from
     # both it and our ``bundlerepository`` class which overrides some
     # functionality. We then change the type of the constructed repository
     # to this new type and initialize the bundle-specific bits of it.
 
     try:
         parentrepo = localrepo.instance(ui, repopath, create=False)
         tempparent = None
     except error.RepoError:
         tempparent = pycompat.mkdtemp()
         try:
             parentrepo = localrepo.instance(ui, tempparent, create=True)
         except Exception:
             shutil.rmtree(tempparent)
             raise
 
     class derivedbundlerepository(bundlerepository, parentrepo.__class__):
         pass
 
     repo = parentrepo
     repo.__class__ = derivedbundlerepository
     bundlerepository.__init__(repo, bundlepath, url, tempparent)
 
     return repo
 
 class bundletransactionmanager(object):
     def transaction(self):
         return None
 
     def close(self):
         raise NotImplementedError
 
     def release(self):
         raise NotImplementedError
 
 def getremotechanges(ui, repo, peer, onlyheads=None, bundlename=None,
                      force=False):
     '''obtains a bundle of changes incoming from peer
 
     "onlyheads" restricts the returned changes to those reachable from the
       specified heads.
     "bundlename", if given, stores the bundle to this file path permanently;
       otherwise it's stored to a temp file and gets deleted again when you call
       the returned "cleanupfn".
     "force" indicates whether to proceed on unrelated repos.
 
     Returns a tuple (local, csets, cleanupfn):
 
     "local" is a local repo from which to obtain the actual incoming
       changesets; it is a bundlerepo for the obtained bundle when the
       original "peer" is remote.
     "csets" lists the incoming changeset node ids.
     "cleanupfn" must be called without arguments when you're done processing
       the changes; it closes both the original "peer" and the one returned
       here.
     '''
     tmp = discovery.findcommonincoming(repo, peer, heads=onlyheads,
                                        force=force)
     common, incoming, rheads = tmp
     if not incoming:
         try:
             if bundlename:
                 os.unlink(bundlename)
         except OSError:
             pass
         return repo, [], peer.close
 
     commonset = set(common)
     rheads = [x for x in rheads if x not in commonset]
 
     bundle = None
     bundlerepo = None
     localrepo = peer.local()
     if bundlename or not localrepo:
         # create a bundle (uncompressed if peer repo is not local)
 
         # developer config: devel.legacy.exchange
         legexc = ui.configlist('devel', 'legacy.exchange')
         forcebundle1 = 'bundle2' not in legexc and 'bundle1' in legexc
         canbundle2 = (not forcebundle1
                       and peer.capable('getbundle')
                       and peer.capable('bundle2'))
         if canbundle2:
             with peer.commandexecutor() as e:
                 b2 = e.callcommand('getbundle', {
                     'source': 'incoming',
                     'common': common,
                     'heads': rheads,
                     'bundlecaps': exchange.caps20to10(repo, role='client'),
                     'cg': True,
                 }).result()
 
                 fname = bundle = changegroup.writechunks(ui,
                                                          b2._forwardchunks(),
                                                          bundlename)
         else:
             if peer.capable('getbundle'):
                 with peer.commandexecutor() as e:
                     cg = e.callcommand('getbundle', {
                         'source': 'incoming',
                         'common': common,
                         'heads': rheads,
                     }).result()
             elif onlyheads is None and not peer.capable('changegroupsubset'):
                 # compat with older servers when pulling all remote heads
 
                 with peer.commandexecutor() as e:
                     cg = e.callcommand('changegroup', {
                         'nodes': incoming,
                         'source': 'incoming',
                     }).result()
 
                 rheads = None
             else:
                 with peer.commandexecutor() as e:
                     cg = e.callcommand('changegroupsubset', {
                         'bases': incoming,
                         'heads': rheads,
                         'source': 'incoming',
                     }).result()
 
             if localrepo:
                 bundletype = "HG10BZ"
             else:
                 bundletype = "HG10UN"
             fname = bundle = bundle2.writebundle(ui, cg, bundlename,
                                                      bundletype)
         # keep written bundle?
         if bundlename:
             bundle = None
         if not localrepo:
             # use the created uncompressed bundlerepo
             localrepo = bundlerepo = makebundlerepository(repo. baseui,
                                                           repo.root,
                                                           fname)
 
             # this repo contains local and peer now, so filter out local again
             common = repo.heads()
     if localrepo:
         # Part of common may be remotely filtered
         # So use an unfiltered version
         # The discovery process probably need cleanup to avoid that
         localrepo = localrepo.unfiltered()
 
     csets = localrepo.changelog.findmissing(common, rheads)
 
     if bundlerepo:
         reponodes = [ctx.node() for ctx in bundlerepo[bundlerepo.firstnewrev:]]
 
         with peer.commandexecutor() as e:
             remotephases = e.callcommand('listkeys', {
                 'namespace': 'phases',
             }).result()
 
         pullop = exchange.pulloperation(bundlerepo, peer, heads=reponodes)
         pullop.trmanager = bundletransactionmanager()
         exchange._pullapplyphases(pullop, remotephases)
 
     def cleanup():
         if bundlerepo:
             bundlerepo.close()
         if bundle:
             os.unlink(bundle)
         peer.close()
 
     return (localrepo, csets, cleanup)

mercurial/localrepo.py

@@ -1,2742 +1,2737
 # localrepo.py - read/write repository class for mercurial
 #
 # Copyright 2005-2007 Matt Mackall <[e-mail]>
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 from __future__ import absolute_import
 
 import errno
 import hashlib
 import os
 import random
 import sys
 import time
 import weakref
 
 from .i18n import _
 from .node import (
     hex,
     nullid,
     short,
 )
 from . import (
     bookmarks,
     branchmap,
     bundle2,
     changegroup,
     changelog,
     color,
     context,
     dirstate,
     dirstateguard,
     discovery,
     encoding,
     error,
     exchange,
     extensions,
     filelog,
     hook,
     lock as lockmod,
     manifest,
     match as matchmod,
     merge as mergemod,
     mergeutil,
     namespaces,
     narrowspec,
     obsolete,
     pathutil,
     phases,
     pushkey,
     pycompat,
     repository,
     repoview,
     revset,
     revsetlang,
     scmutil,
     sparse,
     store as storemod,
     subrepoutil,
     tags as tagsmod,
     transaction,
     txnutil,
     util,
     vfs as vfsmod,
 )
 from .utils import (
     interfaceutil,
     procutil,
     stringutil,
 )
 
 from .revlogutils import (
     constants as revlogconst,
 )
 
 release = lockmod.release
 urlerr = util.urlerr
 urlreq = util.urlreq
 
 # set of (path, vfs-location) tuples. vfs-location is:
 # - 'plain for vfs relative paths
 # - '' for svfs relative paths
 _cachedfiles = set()
 
 class _basefilecache(scmutil.filecache):
     """All filecache usage on repo are done for logic that should be unfiltered
     """
     def __get__(self, repo, type=None):
         if repo is None:
             return self
         return super(_basefilecache, self).__get__(repo.unfiltered(), type)
     def __set__(self, repo, value):
         return super(_basefilecache, self).__set__(repo.unfiltered(), value)
     def __delete__(self, repo):
         return super(_basefilecache, self).__delete__(repo.unfiltered())
 
 class repofilecache(_basefilecache):
     """filecache for files in .hg but outside of .hg/store"""
     def __init__(self, *paths):
         super(repofilecache, self).__init__(*paths)
         for path in paths:
             _cachedfiles.add((path, 'plain'))
 
     def join(self, obj, fname):
         return obj.vfs.join(fname)
 
 class storecache(_basefilecache):
     """filecache for files in the store"""
     def __init__(self, *paths):
         super(storecache, self).__init__(*paths)
         for path in paths:
             _cachedfiles.add((path, ''))
 
     def join(self, obj, fname):
         return obj.sjoin(fname)
 
 def isfilecached(repo, name):
     """check if a repo has already cached "name" filecache-ed property
 
     This returns (cachedobj-or-None, iscached) tuple.
     """
     cacheentry = repo.unfiltered()._filecache.get(name, None)
     if not cacheentry:
         return None, False
     return cacheentry.obj, True
 
 class unfilteredpropertycache(util.propertycache):
     """propertycache that apply to unfiltered repo only"""
 
     def __get__(self, repo, type=None):
         unfi = repo.unfiltered()
         if unfi is repo:
             return super(unfilteredpropertycache, self).__get__(unfi)
         return getattr(unfi, self.name)
 
 class filteredpropertycache(util.propertycache):
     """propertycache that must take filtering in account"""
 
     def cachevalue(self, obj, value):
         object.__setattr__(obj, self.name, value)
 
 
 def hasunfilteredcache(repo, name):
     """check if a repo has an unfilteredpropertycache value for <name>"""
     return name in vars(repo.unfiltered())
 
 def unfilteredmethod(orig):
     """decorate method that always need to be run on unfiltered version"""
     def wrapper(repo, *args, **kwargs):
         return orig(repo.unfiltered(), *args, **kwargs)
     return wrapper
 
 moderncaps = {'lookup', 'branchmap', 'pushkey', 'known', 'getbundle',
               'unbundle'}
 legacycaps = moderncaps.union({'changegroupsubset'})
 
 @interfaceutil.implementer(repository.ipeercommandexecutor)
 class localcommandexecutor(object):
     def __init__(self, peer):
         self._peer = peer
         self._sent = False
         self._closed = False
 
     def __enter__(self):
         return self
 
     def __exit__(self, exctype, excvalue, exctb):
         self.close()
 
     def callcommand(self, command, args):
         if self._sent:
             raise error.ProgrammingError('callcommand() cannot be used after '
                                          'sendcommands()')
 
         if self._closed:
             raise error.ProgrammingError('callcommand() cannot be used after '
                                          'close()')
 
         # We don't need to support anything fancy. Just call the named
         # method on the peer and return a resolved future.
         fn = getattr(self._peer, pycompat.sysstr(command))
 
         f = pycompat.futures.Future()
 
         try:
             result = fn(**pycompat.strkwargs(args))
         except Exception:
             pycompat.future_set_exception_info(f, sys.exc_info()[1:])
         else:
             f.set_result(result)
 
         return f
 
     def sendcommands(self):
         self._sent = True
 
     def close(self):
         self._closed = True
 
 @interfaceutil.implementer(repository.ipeercommands)
 class localpeer(repository.peer):
     '''peer for a local repo; reflects only the most recent API'''
 
     def __init__(self, repo, caps=None):
         super(localpeer, self).__init__()
 
         if caps is None:
             caps = moderncaps.copy()
         self._repo = repo.filtered('served')
         self.ui = repo.ui
         self._caps = repo._restrictcapabilities(caps)
 
     # Begin of _basepeer interface.
 
     def url(self):
         return self._repo.url()
 
     def local(self):
         return self._repo
 
     def peer(self):
         return self
 
     def canpush(self):
         return True
 
     def close(self):
         self._repo.close()
 
     # End of _basepeer interface.
 
     # Begin of _basewirecommands interface.
 
     def branchmap(self):
         return self._repo.branchmap()
 
     def capabilities(self):
         return self._caps
 
     def clonebundles(self):
         return self._repo.tryread('clonebundles.manifest')
 
     def debugwireargs(self, one, two, three=None, four=None, five=None):
         """Used to test argument passing over the wire"""
         return "%s %s %s %s %s" % (one, two, pycompat.bytestr(three),
                                    pycompat.bytestr(four),
                                    pycompat.bytestr(five))
 
     def getbundle(self, source, heads=None, common=None, bundlecaps=None,
                   **kwargs):
         chunks = exchange.getbundlechunks(self._repo, source, heads=heads,
                                           common=common, bundlecaps=bundlecaps,
                                           **kwargs)[1]
         cb = util.chunkbuffer(chunks)
 
         if exchange.bundle2requested(bundlecaps):
             # When requesting a bundle2, getbundle returns a stream to make the
             # wire level function happier. We need to build a proper object
             # from it in local peer.
             return bundle2.getunbundler(self.ui, cb)
         else:
             return changegroup.getunbundler('01', cb, None)
 
     def heads(self):
         return self._repo.heads()
 
     def known(self, nodes):
         return self._repo.known(nodes)
 
     def listkeys(self, namespace):
         return self._repo.listkeys(namespace)
 
     def lookup(self, key):
         return self._repo.lookup(key)
 
     def pushkey(self, namespace, key, old, new):
         return self._repo.pushkey(namespace, key, old, new)
 
     def stream_out(self):
         raise error.Abort(_('cannot perform stream clone against local '
                             'peer'))
 
     def unbundle(self, bundle, heads, url):
         """apply a bundle on a repo
 
         This function handles the repo locking itself."""
         try:
             try:
                 bundle = exchange.readbundle(self.ui, bundle, None)
                 ret = exchange.unbundle(self._repo, bundle, heads, 'push', url)
                 if util.safehasattr(ret, 'getchunks'):
                     # This is a bundle20 object, turn it into an unbundler.
                     # This little dance should be dropped eventually when the
                     # API is finally improved.
                     stream = util.chunkbuffer(ret.getchunks())
                     ret = bundle2.getunbundler(self.ui, stream)
                 return ret
             except Exception as exc:
                 # If the exception contains output salvaged from a bundle2
                 # reply, we need to make sure it is printed before continuing
                 # to fail. So we build a bundle2 with such output and consume
                 # it directly.
                 #
                 # This is not very elegant but allows a "simple" solution for
                 # issue4594
                 output = getattr(exc, '_bundle2salvagedoutput', ())
                 if output:
                     bundler = bundle2.bundle20(self._repo.ui)
                     for out in output:
                         bundler.addpart(out)
                     stream = util.chunkbuffer(bundler.getchunks())
                     b = bundle2.getunbundler(self.ui, stream)
                     bundle2.processbundle(self._repo, b)
                 raise
         except error.PushRaced as exc:
             raise error.ResponseError(_('push failed:'),
                                       stringutil.forcebytestr(exc))
 
     # End of _basewirecommands interface.
 
     # Begin of peer interface.
 
     def commandexecutor(self):
         return localcommandexecutor(self)
 
     # End of peer interface.
 
 @interfaceutil.implementer(repository.ipeerlegacycommands)
 class locallegacypeer(localpeer):
     '''peer extension which implements legacy methods too; used for tests with
     restricted capabilities'''
 
     def __init__(self, repo):
         super(locallegacypeer, self).__init__(repo, caps=legacycaps)
 
     # Begin of baselegacywirecommands interface.
 
     def between(self, pairs):
         return self._repo.between(pairs)
 
     def branches(self, nodes):
         return self._repo.branches(nodes)
 
     def changegroup(self, nodes, source):
         outgoing = discovery.outgoing(self._repo, missingroots=nodes,
                                       missingheads=self._repo.heads())
         return changegroup.makechangegroup(self._repo, outgoing, '01', source)
 
     def changegroupsubset(self, bases, heads, source):
         outgoing = discovery.outgoing(self._repo, missingroots=bases,
                                       missingheads=heads)
         return changegroup.makechangegroup(self._repo, outgoing, '01', source)
 
     # End of baselegacywirecommands interface.
 
 # Increment the sub-version when the revlog v2 format changes to lock out old
 # clients.
 REVLOGV2_REQUIREMENT = 'exp-revlogv2.0'
 
 # A repository with the sparserevlog feature will have delta chains that
 # can spread over a larger span. Sparse reading cuts these large spans into
 # pieces, so that each piece isn't too big.
 # Without the sparserevlog capability, reading from the repository could use
 # huge amounts of memory, because the whole span would be read at once,
 # including all the intermediate revisions that aren't pertinent for the chain.
 # This is why once a repository has enabled sparse-read, it becomes required.
 SPARSEREVLOG_REQUIREMENT = 'sparserevlog'
 
 # Functions receiving (ui, features) that extensions can register to impact
 # the ability to load repositories with custom requirements. Only
 # functions defined in loaded extensions are called.
 #
 # The function receives a set of requirement strings that the repository
 # is capable of opening. Functions will typically add elements to the
 # set to reflect that the extension knows how to handle that requirements.
 featuresetupfuncs = set()
 
 def makelocalrepository(baseui, path, intents=None):
     """Create a local repository object.
 
     Given arguments needed to construct a local repository, this function
     derives a type suitable for representing that repository and returns an
     instance of it.
 
     The returned object conforms to the ``repository.completelocalrepository``
     interface.
     """
     ui = baseui.copy()
     # Prevent copying repo configuration.
     ui.copy = baseui.copy
 
     # Working directory VFS rooted at repository root.
     wdirvfs = vfsmod.vfs(path, expandpath=True, realpath=True)
 
     # Main VFS for .hg/ directory.
     hgpath = wdirvfs.join(b'.hg')
     hgvfs = vfsmod.vfs(hgpath, cacheaudited=True)
 
     # The .hg/ path should exist and should be a directory. All other
     # cases are errors.
     if not hgvfs.isdir():
         try:
             hgvfs.stat()
         except OSError as e:
             if e.errno != errno.ENOENT:
                 raise
 
         raise error.RepoError(_(b'repository %s not found') % path)
 
     # .hg/requires file contains a newline-delimited list of
     # features/capabilities the opener (us) must have in order to use
     # the repository. This file was introduced in Mercurial 0.9.2,
     # which means very old repositories may not have one. We assume
     # a missing file translates to no requirements.
     try:
         requirements = set(hgvfs.read(b'requires').splitlines())
     except IOError as e:
         if e.errno != errno.ENOENT:
             raise
         requirements = set()
 
     # The .hg/hgrc file may load extensions or contain config options
     # that influence repository construction. Attempt to load it and
     # process any new extensions that it may have pulled in.
     try:
         ui.readconfig(hgvfs.join(b'hgrc'), root=wdirvfs.base)
     except IOError:
         pass
     else:
         extensions.loadall(ui)
 
     supportedrequirements = gathersupportedrequirements(ui)
 
     # We first validate the requirements are known.
     ensurerequirementsrecognized(requirements, supportedrequirements)
 
     # Then we validate that the known set is reasonable to use together.
     ensurerequirementscompatible(ui, requirements)
 
     # TODO there are unhandled edge cases related to opening repositories with
     # shared storage. If storage is shared, we should also test for requirements
     # compatibility in the pointed-to repo. This entails loading the .hg/hgrc in
     # that repo, as that repo may load extensions needed to open it. This is a
     # bit complicated because we don't want the other hgrc to overwrite settings
     # in this hgrc.
     #
     # This bug is somewhat mitigated by the fact that we copy the .hg/requires
     # file when sharing repos. But if a requirement is added after the share is
     # performed, thereby introducing a new requirement for the opener, we may
     # will not see that and could encounter a run-time error interacting with
     # that shared store since it has an unknown-to-us requirement.
 
     # At this point, we know we should be capable of opening the repository.
     # Now get on with doing that.
 
     # The "store" part of the repository holds versioned data. How it is
     # accessed is determined by various requirements. The ``shared`` or
     # ``relshared`` requirements indicate the store lives in the path contained
     # in the ``.hg/sharedpath`` file. This is an absolute path for
     # ``shared`` and relative to ``.hg/`` for ``relshared``.
     if b'shared' in requirements or b'relshared' in requirements:
         sharedpath = hgvfs.read(b'sharedpath').rstrip(b'\n')
         if b'relshared' in requirements:
             sharedpath = hgvfs.join(sharedpath)
 
         sharedvfs = vfsmod.vfs(sharedpath, realpath=True)
 
         if not sharedvfs.exists():
             raise error.RepoError(_(b'.hg/sharedpath points to nonexistent '
                                     b'directory %s') % sharedvfs.base)
 
         storebasepath = sharedvfs.base
         cachepath = sharedvfs.join(b'cache')
     else:
         storebasepath = hgvfs.base
         cachepath = hgvfs.join(b'cache')
 
     # The store has changed over time and the exact layout is dictated by
     # requirements. The store interface abstracts differences across all
     # of them.
     store = makestore(requirements, storebasepath,
                       lambda base: vfsmod.vfs(base, cacheaudited=True))
     hgvfs.createmode = store.createmode
 
     storevfs = store.vfs
     storevfs.options = resolvestorevfsoptions(ui, requirements)
 
     # The cache vfs is used to manage cache files.
     cachevfs = vfsmod.vfs(cachepath, cacheaudited=True)
     cachevfs.createmode = store.createmode
 
     return localrepository(
         baseui=baseui,
         ui=ui,
         origroot=path,
         wdirvfs=wdirvfs,
         hgvfs=hgvfs,
         requirements=requirements,
         supportedrequirements=supportedrequirements,
         sharedpath=storebasepath,
         store=store,
         cachevfs=cachevfs,
         intents=intents)
 
 def gathersupportedrequirements(ui):
     """Determine the complete set of recognized requirements."""
     # Start with all requirements supported by this file.
     supported = set(localrepository._basesupported)
 
     # Execute ``featuresetupfuncs`` entries if they belong to an extension
     # relevant to this ui instance.
     modules = {m.__name__ for n, m in extensions.extensions(ui)}
 
     for fn in featuresetupfuncs:
         if fn.__module__ in modules:
             fn(ui, supported)
 
     # Add derived requirements from registered compression engines.
     for name in util.compengines:
         engine = util.compengines[name]
         if engine.revlogheader():
             supported.add(b'exp-compression-%s' % name)
 
     return supported
 
 def ensurerequirementsrecognized(requirements, supported):
     """Validate that a set of local requirements is recognized.
 
     Receives a set of requirements. Raises an ``error.RepoError`` if there
     exists any requirement in that set that currently loaded code doesn't
     recognize.
 
     Returns a set of supported requirements.
     """
     missing = set()
 
     for requirement in requirements:
         if requirement in supported:
             continue
 
         if not requirement or not requirement[0:1].isalnum():
             raise error.RequirementError(_(b'.hg/requires file is corrupt'))
 
         missing.add(requirement)
 
     if missing:
         raise error.RequirementError(
             _(b'repository requires features unknown to this Mercurial: %s') %
             b' '.join(sorted(missing)),
             hint=_(b'see https://mercurial-scm.org/wiki/MissingRequirement '
                    b'for more information'))
 
 def ensurerequirementscompatible(ui, requirements):
     """Validates that a set of recognized requirements is mutually compatible.
 
     Some requirements may not be compatible with others or require
     config options that aren't enabled. This function is called during
     repository opening to ensure that the set of requirements needed
     to open a repository is sane and compatible with config options.
 
     Extensions can monkeypatch this function to perform additional
     checking.
 
     ``error.RepoError`` should be raised on failure.
     """
     if b'exp-sparse' in requirements and not sparse.enabled:
         raise error.RepoError(_(b'repository is using sparse feature but '
                                 b'sparse is not enabled; enable the '
                                 b'"sparse" extensions to access'))
 
 def makestore(requirements, path, vfstype):
     """Construct a storage object for a repository."""
     if b'store' in requirements:
         if b'fncache' in requirements:
             return storemod.fncachestore(path, vfstype,
                                          b'dotencode' in requirements)
 
         return storemod.encodedstore(path, vfstype)
 
     return storemod.basicstore(path, vfstype)
 
 def resolvestorevfsoptions(ui, requirements):
     """Resolve the options to pass to the store vfs opener.
 
     The returned dict is used to influence behavior of the storage layer.
     """
     options = {}
 
     if b'treemanifest' in requirements:
         options[b'treemanifest'] = True
 
     # experimental config: format.manifestcachesize
     manifestcachesize = ui.configint(b'format', b'manifestcachesize')
     if manifestcachesize is not None:
         options[b'manifestcachesize'] = manifestcachesize
 
     # In the absence of another requirement superseding a revlog-related
     # requirement, we have to assume the repo is using revlog version 0.
     # This revlog format is super old and we don't bother trying to parse
     # opener options for it because those options wouldn't do anything
     # meaningful on such old repos.
     if b'revlogv1' in requirements or REVLOGV2_REQUIREMENT in requirements:
         options.update(resolverevlogstorevfsoptions(ui, requirements))
 
     return options
 
 def resolverevlogstorevfsoptions(ui, requirements):
     """Resolve opener options specific to revlogs."""
 
     options = {}
 
     if b'revlogv1' in requirements:
         options[b'revlogv1'] = True
     if REVLOGV2_REQUIREMENT in requirements:
         options[b'revlogv2'] = True
 
     if b'generaldelta' in requirements:
         options[b'generaldelta'] = True
 
     # experimental config: format.chunkcachesize
     chunkcachesize = ui.configint(b'format', b'chunkcachesize')
     if chunkcachesize is not None:
         options[b'chunkcachesize'] = chunkcachesize
 
     deltabothparents = ui.configbool(b'storage',
                                      b'revlog.optimize-delta-parent-choice')
     options[b'deltabothparents'] = deltabothparents
 
     options[b'lazydeltabase'] = not scmutil.gddeltaconfig(ui)
 
     chainspan = ui.configbytes(b'experimental', b'maxdeltachainspan')
     if 0 <= chainspan:
         options[b'maxdeltachainspan'] = chainspan
 
     mmapindexthreshold = ui.configbytes(b'experimental',
                                         b'mmapindexthreshold')
     if mmapindexthreshold is not None:
         options[b'mmapindexthreshold'] = mmapindexthreshold
 
     withsparseread = ui.configbool(b'experimental', b'sparse-read')
     srdensitythres = float(ui.config(b'experimental',
                                      b'sparse-read.density-threshold'))
     srmingapsize = ui.configbytes(b'experimental',
                                   b'sparse-read.min-gap-size')
     options[b'with-sparse-read'] = withsparseread
     options[b'sparse-read-density-threshold'] = srdensitythres
     options[b'sparse-read-min-gap-size'] = srmingapsize
 
     sparserevlog = SPARSEREVLOG_REQUIREMENT in requirements
     options[b'sparse-revlog'] = sparserevlog
     if sparserevlog:
         options[b'generaldelta'] = True
 
     maxchainlen = None
     if sparserevlog:
         maxchainlen = revlogconst.SPARSE_REVLOG_MAX_CHAIN_LENGTH
     # experimental config: format.maxchainlen
     maxchainlen = ui.configint(b'format', b'maxchainlen', maxchainlen)
     if maxchainlen is not None:
         options[b'maxchainlen'] = maxchainlen
 
     for r in requirements:
         if r.startswith(b'exp-compression-'):
             options[b'compengine'] = r[len(b'exp-compression-'):]
 
     return options
 
 @interfaceutil.implementer(repository.completelocalrepository)
 class localrepository(object):
 
     # obsolete experimental requirements:
     #  - manifestv2: An experimental new manifest format that allowed
     #    for stem compression of long paths. Experiment ended up not
     #    being successful (repository sizes went up due to worse delta
     #    chains), and the code was deleted in 4.6.
     supportedformats = {
         'revlogv1',
         'generaldelta',
         'treemanifest',
         REVLOGV2_REQUIREMENT,
         SPARSEREVLOG_REQUIREMENT,
     }
     _basesupported = supportedformats | {
         'store',
         'fncache',
         'shared',
         'relshared',
         'dotencode',
         'exp-sparse',
         'internal-phase'
     }
 
     # list of prefix for file which can be written without 'wlock'
     # Extensions should extend this list when needed
     _wlockfreeprefix = {
         # We migh consider requiring 'wlock' for the next
         # two, but pretty much all the existing code assume
         # wlock is not needed so we keep them excluded for
         # now.
         'hgrc',
         'requires',
         # XXX cache is a complicatged business someone
         # should investigate this in depth at some point
         'cache/',
         # XXX shouldn't be dirstate covered by the wlock?
         'dirstate',
         # XXX bisect was still a bit too messy at the time
         # this changeset was introduced. Someone should fix
         # the remainig bit and drop this line
         'bisect.state',
     }
 
     def __init__(self, baseui, ui, origroot, wdirvfs, hgvfs, requirements,
                  supportedrequirements, sharedpath, store, cachevfs,
                  intents=None):
         """Create a new local repository instance.
 
         Most callers should use ``hg.repository()``, ``localrepo.instance()``,
         or ``localrepo.makelocalrepository()`` for obtaining a new repository
         object.
 
         Arguments:
 
         baseui
            ``ui.ui`` instance that ``ui`` argument was based off of.
 
         ui
            ``ui.ui`` instance for use by the repository.
 
         origroot
            ``bytes`` path to working directory root of this repository.
 
         wdirvfs
            ``vfs.vfs`` rooted at the working directory.
 
         hgvfs
            ``vfs.vfs`` rooted at .hg/
 
         requirements
            ``set`` of bytestrings representing repository opening requirements.
 
         supportedrequirements
            ``set`` of bytestrings representing repository requirements that we
            know how to open. May be a supetset of ``requirements``.
 
         sharedpath
            ``bytes`` Defining path to storage base directory. Points to a
            ``.hg/`` directory somewhere.
 
         store
            ``store.basicstore`` (or derived) instance providing access to
            versioned storage.
 
         cachevfs
            ``vfs.vfs`` used for cache files.
 
         intents
            ``set`` of system strings indicating what this repo will be used
            for.
         """
         self.baseui = baseui
         self.ui = ui
         self.origroot = origroot
         # vfs rooted at working directory.
         self.wvfs = wdirvfs
         self.root = wdirvfs.base
         # vfs rooted at .hg/. Used to access most non-store paths.
         self.vfs = hgvfs
         self.path = hgvfs.base
         self.requirements = requirements
         self.supported = supportedrequirements
         self.sharedpath = sharedpath
         self.store = store
         self.cachevfs = cachevfs
 
         self.filtername = None
 
         if (self.ui.configbool('devel', 'all-warnings') or
             self.ui.configbool('devel', 'check-locks')):
             self.vfs.audit = self._getvfsward(self.vfs.audit)
         # A list of callback to shape the phase if no data were found.
         # Callback are in the form: func(repo, roots) --> processed root.
         # This list it to be filled by extension during repo setup
         self._phasedefaults = []
 
         color.setup(self.ui)
 
         self.spath = self.store.path
         self.svfs = self.store.vfs
         self.sjoin = self.store.join
         if (self.ui.configbool('devel', 'all-warnings') or
             self.ui.configbool('devel', 'check-locks')):
             if util.safehasattr(self.svfs, 'vfs'): # this is filtervfs
                 self.svfs.vfs.audit = self._getsvfsward(self.svfs.vfs.audit)
             else: # standard vfs
                 self.svfs.audit = self._getsvfsward(self.svfs.audit)
 
         self._dirstatevalidatewarned = False
 
         self._branchcaches = {}
         self._revbranchcache = None
         self._filterpats = {}
         self._datafilters = {}
         self._transref = self._lockref = self._wlockref = None
 
         # A cache for various files under .hg/ that tracks file changes,
         # (used by the filecache decorator)
         #
         # Maps a property name to its util.filecacheentry
         self._filecache = {}
 
         # hold sets of revision to be filtered
         # should be cleared when something might have changed the filter value:
         # - new changesets,
         # - phase change,
         # - new obsolescence marker,
         # - working directory parent change,
         # - bookmark changes
         self.filteredrevcache = {}
 
         # post-dirstate-status hooks
         self._postdsstatus = []
 
         # generic mapping between names and nodes
         self.names = namespaces.namespaces()
 
         # Key to signature value.
         self._sparsesignaturecache = {}
         # Signature to cached matcher instance.
         self._sparsematchercache = {}
 
     def _getvfsward(self, origfunc):
         """build a ward for self.vfs"""
         rref = weakref.ref(self)
         def checkvfs(path, mode=None):
             ret = origfunc(path, mode=mode)
             repo = rref()
             if (repo is None
                 or not util.safehasattr(repo, '_wlockref')
                 or not util.safehasattr(repo, '_lockref')):
                 return
             if mode in (None, 'r', 'rb'):
                 return
             if path.startswith(repo.path):
                 # truncate name relative to the repository (.hg)
                 path = path[len(repo.path) + 1:]
             if path.startswith('cache/'):
                 msg = 'accessing cache with vfs instead of cachevfs: "%s"'
                 repo.ui.develwarn(msg % path, stacklevel=2, config="cache-vfs")
             if path.startswith('journal.'):
                 # journal is covered by 'lock'
                 if repo._currentlock(repo._lockref) is None:
                     repo.ui.develwarn('write with no lock: "%s"' % path,
                                       stacklevel=2, config='check-locks')
             elif repo._currentlock(repo._wlockref) is None:
                 # rest of vfs files are covered by 'wlock'
                 #
                 # exclude special files
                 for prefix in self._wlockfreeprefix:
                     if path.startswith(prefix):
                         return
                 repo.ui.develwarn('write with no wlock: "%s"' % path,
                                   stacklevel=2, config='check-locks')
             return ret
         return checkvfs
 
     def _getsvfsward(self, origfunc):
         """build a ward for self.svfs"""
         rref = weakref.ref(self)
         def checksvfs(path, mode=None):
             ret = origfunc(path, mode=mode)
             repo = rref()
             if repo is None or not util.safehasattr(repo, '_lockref'):
                 return
             if mode in (None, 'r', 'rb'):
                 return
             if path.startswith(repo.sharedpath):
                 # truncate name relative to the repository (.hg)
                 path = path[len(repo.sharedpath) + 1:]
             if repo._currentlock(repo._lockref) is None:
                 repo.ui.develwarn('write with no lock: "%s"' % path,
                                   stacklevel=3)
             return ret
         return checksvfs
 
     def close(self):
         self._writecaches()
 
     def _writecaches(self):
         if self._revbranchcache:
             self._revbranchcache.write()
 
     def _restrictcapabilities(self, caps):
         if self.ui.configbool('experimental', 'bundle2-advertise'):
             caps = set(caps)
             capsblob = bundle2.encodecaps(bundle2.getrepocaps(self,
                                                               role='client'))
             caps.add('bundle2=' + urlreq.quote(capsblob))
         return caps
 
     def _writerequirements(self):
         scmutil.writerequires(self.vfs, self.requirements)
 
     # Don't cache auditor/nofsauditor, or you'll end up with reference cycle:
     # self -> auditor -> self._checknested -> self
 
     @property
     def auditor(self):
         # This is only used by context.workingctx.match in order to
         # detect files in subrepos.
         return pathutil.pathauditor(self.root, callback=self._checknested)
 
     @property
     def nofsauditor(self):
         # This is only used by context.basectx.match in order to detect
         # files in subrepos.
         return pathutil.pathauditor(self.root, callback=self._checknested,
                                     realfs=False, cached=True)
 
     def _checknested(self, path):
         """Determine if path is a legal nested repository."""
         if not path.startswith(self.root):
             return False
         subpath = path[len(self.root) + 1:]
         normsubpath = util.pconvert(subpath)
 
         # XXX: Checking against the current working copy is wrong in
         # the sense that it can reject things like
         #
         #   $ hg cat -r 10 sub/x.txt
         #
         # if sub/ is no longer a subrepository in the working copy
         # parent revision.
         #
         # However, it can of course also allow things that would have
         # been rejected before, such as the above cat command if sub/
         # is a subrepository now, but was a normal directory before.
         # The old path auditor would have rejected by mistake since it
         # panics when it sees sub/.hg/.
         #
         # All in all, checking against the working copy seems sensible
         # since we want to prevent access to nested repositories on
         # the filesystem *now*.
         ctx = self[None]
         parts = util.splitpath(subpath)
         while parts:
             prefix = '/'.join(parts)
             if prefix in ctx.substate:
                 if prefix == normsubpath:
                     return True
                 else:
                     sub = ctx.sub(prefix)
                     return sub.checknested(subpath[len(prefix) + 1:])
             else:
                 parts.pop()
         return False
 
     def peer(self):
         return localpeer(self) # not cached to avoid reference cycle
 
     def unfiltered(self):
         """Return unfiltered version of the repository
 
         Intended to be overwritten by filtered repo."""
         return self
 
     def filtered(self, name, visibilityexceptions=None):
         """Return a filtered version of a repository"""
         cls = repoview.newtype(self.unfiltered().__class__)
         return cls(self, name, visibilityexceptions)
 
     @repofilecache('bookmarks', 'bookmarks.current')
     def _bookmarks(self):
         return bookmarks.bmstore(self)
 
     @property
     def _activebookmark(self):
         return self._bookmarks.active
 
     # _phasesets depend on changelog. what we need is to call
     # _phasecache.invalidate() if '00changelog.i' was changed, but it
     # can't be easily expressed in filecache mechanism.
     @storecache('phaseroots', '00changelog.i')
     def _phasecache(self):
         return phases.phasecache(self, self._phasedefaults)
 
     @storecache('obsstore')
     def obsstore(self):
         return obsolete.makestore(self.ui, self)
 
     @storecache('00changelog.i')
     def changelog(self):
         return changelog.changelog(self.svfs,
                                    trypending=txnutil.mayhavepending(self.root))
 
-    def _constructmanifest(self):
-        # This is a temporary function while we migrate from manifest to
-        # manifestlog. It allows bundlerepo and unionrepo to intercept the
-        # manifest creation.
-        return manifest.manifestrevlog(self.svfs)
-
     @storecache('00manifest.i')
     def manifestlog(self):
-        return manifest.manifestlog(self.svfs, self)
+        rootstore = manifest.manifestrevlog(self.svfs)
+        return manifest.manifestlog(self.svfs, self, rootstore)
 
     @repofilecache('dirstate')
     def dirstate(self):
         return self._makedirstate()
 
     def _makedirstate(self):
         """Extension point for wrapping the dirstate per-repo."""
         sparsematchfn = lambda: sparse.matcher(self)
 
         return dirstate.dirstate(self.vfs, self.ui, self.root,
                                  self._dirstatevalidate, sparsematchfn)
 
     def _dirstatevalidate(self, node):
         try:
             self.changelog.rev(node)
             return node
         except error.LookupError:
             if not self._dirstatevalidatewarned:
                 self._dirstatevalidatewarned = True
                 self.ui.warn(_("warning: ignoring unknown"
                                " working parent %s!\n") % short(node))
             return nullid
 
     @storecache(narrowspec.FILENAME)
     def narrowpats(self):
         """matcher patterns for this repository's narrowspec
 
         A tuple of (includes, excludes).
         """
         return narrowspec.load(self)
 
     @storecache(narrowspec.FILENAME)
     def _narrowmatch(self):
         if repository.NARROW_REQUIREMENT not in self.requirements:
             return matchmod.always(self.root, '')
         include, exclude = self.narrowpats
         return narrowspec.match(self.root, include=include, exclude=exclude)
 
     # TODO(martinvonz): make this property-like instead?
     def narrowmatch(self):
         return self._narrowmatch
 
     def setnarrowpats(self, newincludes, newexcludes):
         narrowspec.save(self, newincludes, newexcludes)
         self.invalidate(clearfilecache=True)
 
     def __getitem__(self, changeid):
         if changeid is None:
             return context.workingctx(self)
         if isinstance(changeid, context.basectx):
             return changeid
         if isinstance(changeid, slice):
             # wdirrev isn't contiguous so the slice shouldn't include it
             return [context.changectx(self, i)
                     for i in pycompat.xrange(*changeid.indices(len(self)))
                     if i not in self.changelog.filteredrevs]
         try:
             return context.changectx(self, changeid)
         except error.WdirUnsupported:
             return context.workingctx(self)
 
     def __contains__(self, changeid):
         """True if the given changeid exists
 
         error.AmbiguousPrefixLookupError is raised if an ambiguous node
         specified.
         """
         try:
             self[changeid]
             return True
         except error.RepoLookupError:
             return False
 
     def __nonzero__(self):
         return True
 
     __bool__ = __nonzero__
 
     def __len__(self):
         # no need to pay the cost of repoview.changelog
         unfi = self.unfiltered()
         return len(unfi.changelog)
 
     def __iter__(self):
         return iter(self.changelog)
 
     def revs(self, expr, *args):
         '''Find revisions matching a revset.
 
         The revset is specified as a string ``expr`` that may contain
         %-formatting to escape certain types. See ``revsetlang.formatspec``.
 
         Revset aliases from the configuration are not expanded. To expand
         user aliases, consider calling ``scmutil.revrange()`` or
         ``repo.anyrevs([expr], user=True)``.
 
         Returns a revset.abstractsmartset, which is a list-like interface
         that contains integer revisions.
         '''
         expr = revsetlang.formatspec(expr, *args)
         m = revset.match(None, expr)
         return m(self)
 
     def set(self, expr, *args):
         '''Find revisions matching a revset and emit changectx instances.
 
         This is a convenience wrapper around ``revs()`` that iterates the
         result and is a generator of changectx instances.
 
         Revset aliases from the configuration are not expanded. To expand
         user aliases, consider calling ``scmutil.revrange()``.
         '''
         for r in self.revs(expr, *args):
             yield self[r]
 
     def anyrevs(self, specs, user=False, localalias=None):
         '''Find revisions matching one of the given revsets.
 
         Revset aliases from the configuration are not expanded by default. To
         expand user aliases, specify ``user=True``. To provide some local
         definitions overriding user aliases, set ``localalias`` to
         ``{name: definitionstring}``.
         '''
         if user:
             m = revset.matchany(self.ui, specs,
                                 lookup=revset.lookupfn(self),
                                 localalias=localalias)
         else:
             m = revset.matchany(None, specs, localalias=localalias)
         return m(self)
 
     def url(self):
         return 'file:' + self.root
 
     def hook(self, name, throw=False, **args):
         """Call a hook, passing this repo instance.
 
         This a convenience method to aid invoking hooks. Extensions likely
         won't call this unless they have registered a custom hook or are
         replacing code that is expected to call a hook.
         """
         return hook.hook(self.ui, self, name, throw, **args)
 
     @filteredpropertycache
     def _tagscache(self):
         '''Returns a tagscache object that contains various tags related
         caches.'''
 
         # This simplifies its cache management by having one decorated
         # function (this one) and the rest simply fetch things from it.
         class tagscache(object):
             def __init__(self):
                 # These two define the set of tags for this repository. tags
                 # maps tag name to node; tagtypes maps tag name to 'global' or
                 # 'local'. (Global tags are defined by .hgtags across all
                 # heads, and local tags are defined in .hg/localtags.)
                 # They constitute the in-memory cache of tags.
                 self.tags = self.tagtypes = None
 
                 self.nodetagscache = self.tagslist = None
 
         cache = tagscache()
         cache.tags, cache.tagtypes = self._findtags()
 
         return cache
 
     def tags(self):
         '''return a mapping of tag to node'''
         t = {}
         if self.changelog.filteredrevs:
             tags, tt = self._findtags()
         else:
             tags = self._tagscache.tags
         for k, v in tags.iteritems():
             try:
                 # ignore tags to unknown nodes
                 self.changelog.rev(v)
                 t[k] = v
             except (error.LookupError, ValueError):
                 pass
         return t
 
     def _findtags(self):
         '''Do the hard work of finding tags.  Return a pair of dicts
         (tags, tagtypes) where tags maps tag name to node, and tagtypes
         maps tag name to a string like \'global\' or \'local\'.
         Subclasses or extensions are free to add their own tags, but
         should be aware that the returned dicts will be retained for the
         duration of the localrepo object.'''
 
         # XXX what tagtype should subclasses/extensions use?  Currently
         # mq and bookmarks add tags, but do not set the tagtype at all.
         # Should each extension invent its own tag type?  Should there
         # be one tagtype for all such "virtual" tags?  Or is the status
         # quo fine?
 
 
         # map tag name to (node, hist)
         alltags = tagsmod.findglobaltags(self.ui, self)
         # map tag name to tag type
         tagtypes = dict((tag, 'global') for tag in alltags)
 
         tagsmod.readlocaltags(self.ui, self, alltags, tagtypes)
 
         # Build the return dicts.  Have to re-encode tag names because
         # the tags module always uses UTF-8 (in order not to lose info
         # writing to the cache), but the rest of Mercurial wants them in
         # local encoding.
         tags = {}
         for (name, (node, hist)) in alltags.iteritems():
             if node != nullid:
                 tags[encoding.tolocal(name)] = node
         tags['tip'] = self.changelog.tip()
         tagtypes = dict([(encoding.tolocal(name), value)
                          for (name, value) in tagtypes.iteritems()])
         return (tags, tagtypes)
 
     def tagtype(self, tagname):
         '''
         return the type of the given tag. result can be:
 
         'local'  : a local tag
         'global' : a global tag
         None     : tag does not exist
         '''
 
         return self._tagscache.tagtypes.get(tagname)
 
     def tagslist(self):
         '''return a list of tags ordered by revision'''
         if not self._tagscache.tagslist:
             l = []
             for t, n in self.tags().iteritems():
                 l.append((self.changelog.rev(n), t, n))
             self._tagscache.tagslist = [(t, n) for r, t, n in sorted(l)]
 
         return self._tagscache.tagslist
 
     def nodetags(self, node):
         '''return the tags associated with a node'''
         if not self._tagscache.nodetagscache:
             nodetagscache = {}
             for t, n in self._tagscache.tags.iteritems():
                 nodetagscache.setdefault(n, []).append(t)
             for tags in nodetagscache.itervalues():
                 tags.sort()
             self._tagscache.nodetagscache = nodetagscache
         return self._tagscache.nodetagscache.get(node, [])
 
     def nodebookmarks(self, node):
         """return the list of bookmarks pointing to the specified node"""
         return self._bookmarks.names(node)
 
     def branchmap(self):
         '''returns a dictionary {branch: [branchheads]} with branchheads
         ordered by increasing revision number'''
         branchmap.updatecache(self)
         return self._branchcaches[self.filtername]
 
     @unfilteredmethod
     def revbranchcache(self):
         if not self._revbranchcache:
             self._revbranchcache = branchmap.revbranchcache(self.unfiltered())
         return self._revbranchcache
 
     def branchtip(self, branch, ignoremissing=False):
         '''return the tip node for a given branch
 
         If ignoremissing is True, then this method will not raise an error.
         This is helpful for callers that only expect None for a missing branch
         (e.g. namespace).
 
         '''
         try:
             return self.branchmap().branchtip(branch)
         except KeyError:
             if not ignoremissing:
                 raise error.RepoLookupError(_("unknown branch '%s'") % branch)
             else:
                 pass
 
     def lookup(self, key):
         return scmutil.revsymbol(self, key).node()
 
     def lookupbranch(self, key):
         if key in self.branchmap():
             return key
 
         return scmutil.revsymbol(self, key).branch()
 
     def known(self, nodes):
         cl = self.changelog
         nm = cl.nodemap
         filtered = cl.filteredrevs
         result = []
         for n in nodes:
             r = nm.get(n)
             resp = not (r is None or r in filtered)
             result.append(resp)
         return result
 
     def local(self):
         return self
 
     def publishing(self):
         # it's safe (and desirable) to trust the publish flag unconditionally
         # so that we don't finalize changes shared between users via ssh or nfs
         return self.ui.configbool('phases', 'publish', untrusted=True)
 
     def cancopy(self):
         # so statichttprepo's override of local() works
         if not self.local():
             return False
         if not self.publishing():
             return True
         # if publishing we can't copy if there is filtered content
         return not self.filtered('visible').changelog.filteredrevs
 
     def shared(self):
         '''the type of shared repository (None if not shared)'''
         if self.sharedpath != self.path:
             return 'store'
         return None
 
     def wjoin(self, f, *insidef):
         return self.vfs.reljoin(self.root, f, *insidef)
 
     def file(self, f):
         if f[0] == '/':
             f = f[1:]
         return filelog.filelog(self.svfs, f)
 
     def setparents(self, p1, p2=nullid):
         with self.dirstate.parentchange():
             copies = self.dirstate.setparents(p1, p2)
             pctx = self[p1]
             if copies:
                 # Adjust copy records, the dirstate cannot do it, it
                 # requires access to parents manifests. Preserve them
                 # only for entries added to first parent.
                 for f in copies:
                     if f not in pctx and copies[f] in pctx:
                         self.dirstate.copy(copies[f], f)
             if p2 == nullid:
                 for f, s in sorted(self.dirstate.copies().items()):
                     if f not in pctx and s not in pctx:
                         self.dirstate.copy(None, f)
 
     def filectx(self, path, changeid=None, fileid=None, changectx=None):
         """changeid can be a changeset revision, node, or tag.
            fileid can be a file revision or node."""
         return context.filectx(self, path, changeid, fileid,
                                changectx=changectx)
 
     def getcwd(self):
         return self.dirstate.getcwd()
 
     def pathto(self, f, cwd=None):
         return self.dirstate.pathto(f, cwd)
 
     def _loadfilter(self, filter):
         if filter not in self._filterpats:
             l = []
             for pat, cmd in self.ui.configitems(filter):
                 if cmd == '!':
                     continue
                 mf = matchmod.match(self.root, '', [pat])
                 fn = None
                 params = cmd
                 for name, filterfn in self._datafilters.iteritems():
                     if cmd.startswith(name):
                         fn = filterfn
                         params = cmd[len(name):].lstrip()
                         break
                 if not fn:
                     fn = lambda s, c, **kwargs: procutil.filter(s, c)
                 # Wrap old filters not supporting keyword arguments
                 if not pycompat.getargspec(fn)[2]:
                     oldfn = fn
                     fn = lambda s, c, **kwargs: oldfn(s, c)
                 l.append((mf, fn, params))
             self._filterpats[filter] = l
         return self._filterpats[filter]
 
     def _filter(self, filterpats, filename, data):
         for mf, fn, cmd in filterpats:
             if mf(filename):
                 self.ui.debug("filtering %s through %s\n" % (filename, cmd))
                 data = fn(data, cmd, ui=self.ui, repo=self, filename=filename)
                 break
 
         return data
 
     @unfilteredpropertycache
     def _encodefilterpats(self):
         return self._loadfilter('encode')
 
     @unfilteredpropertycache
     def _decodefilterpats(self):
         return self._loadfilter('decode')
 
     def adddatafilter(self, name, filter):
         self._datafilters[name] = filter
 
     def wread(self, filename):
         if self.wvfs.islink(filename):
             data = self.wvfs.readlink(filename)
         else:
             data = self.wvfs.read(filename)
         return self._filter(self._encodefilterpats, filename, data)
 
     def wwrite(self, filename, data, flags, backgroundclose=False, **kwargs):
         """write ``data`` into ``filename`` in the working directory
 
         This returns length of written (maybe decoded) data.
         """
         data = self._filter(self._decodefilterpats, filename, data)
         if 'l' in flags:
             self.wvfs.symlink(data, filename)
         else:
             self.wvfs.write(filename, data, backgroundclose=backgroundclose,
                             **kwargs)
             if 'x' in flags:
                 self.wvfs.setflags(filename, False, True)
             else:
                 self.wvfs.setflags(filename, False, False)
         return len(data)
 
     def wwritedata(self, filename, data):
         return self._filter(self._decodefilterpats, filename, data)
 
     def currenttransaction(self):
         """return the current transaction or None if non exists"""
         if self._transref:
             tr = self._transref()
         else:
             tr = None
 
         if tr and tr.running():
             return tr
         return None
 
     def transaction(self, desc, report=None):
         if (self.ui.configbool('devel', 'all-warnings')
                 or self.ui.configbool('devel', 'check-locks')):
             if self._currentlock(self._lockref) is None:
                 raise error.ProgrammingError('transaction requires locking')
         tr = self.currenttransaction()
         if tr is not None:
             return tr.nest(name=desc)
 
         # abort here if the journal already exists
         if self.svfs.exists("journal"):
             raise error.RepoError(
                 _("abandoned transaction found"),
                 hint=_("run 'hg recover' to clean up transaction"))
 
         idbase = "%.40f#%f" % (random.random(), time.time())
         ha = hex(hashlib.sha1(idbase).digest())
         txnid = 'TXN:' + ha
         self.hook('pretxnopen', throw=True, txnname=desc, txnid=txnid)
 
         self._writejournal(desc)
         renames = [(vfs, x, undoname(x)) for vfs, x in self._journalfiles()]
         if report:
             rp = report
         else:
             rp = self.ui.warn
         vfsmap = {'plain': self.vfs} # root of .hg/
         # we must avoid cyclic reference between repo and transaction.
         reporef = weakref.ref(self)
         # Code to track tag movement
         #
         # Since tags are all handled as file content, it is actually quite hard
         # to track these movement from a code perspective. So we fallback to a
         # tracking at the repository level. One could envision to track changes
         # to the '.hgtags' file through changegroup apply but that fails to
         # cope with case where transaction expose new heads without changegroup
         # being involved (eg: phase movement).
         #
         # For now, We gate the feature behind a flag since this likely comes
         # with performance impacts. The current code run more often than needed
         # and do not use caches as much as it could.  The current focus is on
         # the behavior of the feature so we disable it by default. The flag
         # will be removed when we are happy with the performance impact.
         #
         # Once this feature is no longer experimental move the following
         # documentation to the appropriate help section:
         #
         # The ``HG_TAG_MOVED`` variable will be set if the transaction touched
         # tags (new or changed or deleted tags). In addition the details of
         # these changes are made available in a file at:
         #     ``REPOROOT/.hg/changes/tags.changes``.
         # Make sure you check for HG_TAG_MOVED before reading that file as it
         # might exist from a previous transaction even if no tag were touched
         # in this one. Changes are recorded in a line base format::
         #
         #     <action> <hex-node> <tag-name>\n
         #
         # Actions are defined as follow:
         #   "-R": tag is removed,
         #   "+A": tag is added,
         #   "-M": tag is moved (old value),
         #   "+M": tag is moved (new value),
         tracktags = lambda x: None
         # experimental config: experimental.hook-track-tags
         shouldtracktags = self.ui.configbool('experimental', 'hook-track-tags')
         if desc != 'strip' and shouldtracktags:
             oldheads = self.changelog.headrevs()
             def tracktags(tr2):
                 repo = reporef()
                 oldfnodes = tagsmod.fnoderevs(repo.ui, repo, oldheads)
                 newheads = repo.changelog.headrevs()
                 newfnodes = tagsmod.fnoderevs(repo.ui, repo, newheads)
                 # notes: we compare lists here.
                 # As we do it only once buiding set would not be cheaper
                 changes = tagsmod.difftags(repo.ui, repo, oldfnodes, newfnodes)
                 if changes:
                     tr2.hookargs['tag_moved'] = '1'
                     with repo.vfs('changes/tags.changes', 'w',
                                   atomictemp=True) as changesfile:
                         # note: we do not register the file to the transaction
                         # because we needs it to still exist on the transaction
                         # is close (for txnclose hooks)
                         tagsmod.writediff(changesfile, changes)
         def validate(tr2):
             """will run pre-closing hooks"""
             # XXX the transaction API is a bit lacking here so we take a hacky
             # path for now
             #
             # We cannot add this as a "pending" hooks since the 'tr.hookargs'
             # dict is copied before these run. In addition we needs the data
             # available to in memory hooks too.
             #
             # Moreover, we also need to make sure this runs before txnclose
             # hooks and there is no "pending" mechanism that would execute
             # logic only if hooks are about to run.
             #
             # Fixing this limitation of the transaction is also needed to track
             # other families of changes (bookmarks, phases, obsolescence).
             #
             # This will have to be fixed before we remove the experimental
             # gating.
             tracktags(tr2)
             repo = reporef()
             if repo.ui.configbool('experimental', 'single-head-per-branch'):
                 scmutil.enforcesinglehead(repo, tr2, desc)
             if hook.hashook(repo.ui, 'pretxnclose-bookmark'):
                 for name, (old, new) in sorted(tr.changes['bookmarks'].items()):
                     args = tr.hookargs.copy()
                     args.update(bookmarks.preparehookargs(name, old, new))
                     repo.hook('pretxnclose-bookmark', throw=True,
                               txnname=desc,
                               **pycompat.strkwargs(args))
             if hook.hashook(repo.ui, 'pretxnclose-phase'):
                 cl = repo.unfiltered().changelog
                 for rev, (old, new) in tr.changes['phases'].items():
                     args = tr.hookargs.copy()
                     node = hex(cl.node(rev))
                     args.update(phases.preparehookargs(node, old, new))
                     repo.hook('pretxnclose-phase', throw=True, txnname=desc,
                               **pycompat.strkwargs(args))
 
             repo.hook('pretxnclose', throw=True,
                       txnname=desc, **pycompat.strkwargs(tr.hookargs))
         def releasefn(tr, success):
             repo = reporef()
             if success:
                 # this should be explicitly invoked here, because
                 # in-memory changes aren't written out at closing
                 # transaction, if tr.addfilegenerator (via
                 # dirstate.write or so) isn't invoked while
                 # transaction running
                 repo.dirstate.write(None)
             else:
                 # discard all changes (including ones already written
                 # out) in this transaction
                 narrowspec.restorebackup(self, 'journal.narrowspec')
                 repo.dirstate.restorebackup(None, 'journal.dirstate')
 
                 repo.invalidate(clearfilecache=True)
 
         tr = transaction.transaction(rp, self.svfs, vfsmap,
                                      "journal",
                                      "undo",
                                      aftertrans(renames),
                                      self.store.createmode,
                                      validator=validate,
                                      releasefn=releasefn,
                                      checkambigfiles=_cachedfiles,
                                      name=desc)
         tr.changes['origrepolen'] = len(self)
         tr.changes['obsmarkers'] = set()
         tr.changes['phases'] = {}
         tr.changes['bookmarks'] = {}
 
         tr.hookargs['txnid'] = txnid
         # note: writing the fncache only during finalize mean that the file is
         # outdated when running hooks. As fncache is used for streaming clone,
         # this is not expected to break anything that happen during the hooks.
         tr.addfinalize('flush-fncache', self.store.write)
         def txnclosehook(tr2):
             """To be run if transaction is successful, will schedule a hook run
             """
             # Don't reference tr2 in hook() so we don't hold a reference.
             # This reduces memory consumption when there are multiple
             # transactions per lock. This can likely go away if issue5045
             # fixes the function accumulation.
             hookargs = tr2.hookargs
 
             def hookfunc():
                 repo = reporef()
                 if hook.hashook(repo.ui, 'txnclose-bookmark'):
                     bmchanges = sorted(tr.changes['bookmarks'].items())
                     for name, (old, new) in bmchanges:
                         args = tr.hookargs.copy()
                         args.update(bookmarks.preparehookargs(name, old, new))
                         repo.hook('txnclose-bookmark', throw=False,
                                   txnname=desc, **pycompat.strkwargs(args))
 
                 if hook.hashook(repo.ui, 'txnclose-phase'):
                     cl = repo.unfiltered().changelog
                     phasemv = sorted(tr.changes['phases'].items())
                     for rev, (old, new) in phasemv:
                         args = tr.hookargs.copy()
                         node = hex(cl.node(rev))
                         args.update(phases.preparehookargs(node, old, new))
                         repo.hook('txnclose-phase', throw=False, txnname=desc,
                                   **pycompat.strkwargs(args))
 
                 repo.hook('txnclose', throw=False, txnname=desc,
                           **pycompat.strkwargs(hookargs))
             reporef()._afterlock(hookfunc)
         tr.addfinalize('txnclose-hook', txnclosehook)
         # Include a leading "-" to make it happen before the transaction summary
         # reports registered via scmutil.registersummarycallback() whose names
         # are 00-txnreport etc. That way, the caches will be warm when the
         # callbacks run.
         tr.addpostclose('-warm-cache', self._buildcacheupdater(tr))
         def txnaborthook(tr2):
             """To be run if transaction is aborted
             """
             reporef().hook('txnabort', throw=False, txnname=desc,
                            **pycompat.strkwargs(tr2.hookargs))
         tr.addabort('txnabort-hook', txnaborthook)
         # avoid eager cache invalidation. in-memory data should be identical
         # to stored data if transaction has no error.
         tr.addpostclose('refresh-filecachestats', self._refreshfilecachestats)
         self._transref = weakref.ref(tr)
         scmutil.registersummarycallback(self, tr, desc)
         return tr
 
     def _journalfiles(self):
         return ((self.svfs, 'journal'),
                 (self.vfs, 'journal.dirstate'),
                 (self.vfs, 'journal.branch'),
                 (self.vfs, 'journal.desc'),
                 (self.vfs, 'journal.bookmarks'),
                 (self.svfs, 'journal.phaseroots'))
 
     def undofiles(self):
         return [(vfs, undoname(x)) for vfs, x in self._journalfiles()]
 
     @unfilteredmethod
     def _writejournal(self, desc):
         self.dirstate.savebackup(None, 'journal.dirstate')
         narrowspec.savebackup(self, 'journal.narrowspec')
         self.vfs.write("journal.branch",
                           encoding.fromlocal(self.dirstate.branch()))
         self.vfs.write("journal.desc",
                           "%d\n%s\n" % (len(self), desc))
         self.vfs.write("journal.bookmarks",
                           self.vfs.tryread("bookmarks"))
         self.svfs.write("journal.phaseroots",
                            self.svfs.tryread("phaseroots"))
 
     def recover(self):
         with self.lock():
             if self.svfs.exists("journal"):
                 self.ui.status(_("rolling back interrupted transaction\n"))
                 vfsmap = {'': self.svfs,
                           'plain': self.vfs,}
                 transaction.rollback(self.svfs, vfsmap, "journal",
                                      self.ui.warn,
                                      checkambigfiles=_cachedfiles)
                 self.invalidate()
                 return True
             else:
                 self.ui.warn(_("no interrupted transaction available\n"))
                 return False
 
     def rollback(self, dryrun=False, force=False):
         wlock = lock = dsguard = None
         try:
             wlock = self.wlock()
             lock = self.lock()
             if self.svfs.exists("undo"):
                 dsguard = dirstateguard.dirstateguard(self, 'rollback')
 
                 return self._rollback(dryrun, force, dsguard)
             else:
                 self.ui.warn(_("no rollback information available\n"))
                 return 1
         finally:
             release(dsguard, lock, wlock)
 
     @unfilteredmethod # Until we get smarter cache management
     def _rollback(self, dryrun, force, dsguard):
         ui = self.ui
         try:
             args = self.vfs.read('undo.desc').splitlines()
             (oldlen, desc, detail) = (int(args[0]), args[1], None)
             if len(args) >= 3:
                 detail = args[2]
             oldtip = oldlen - 1
 
             if detail and ui.verbose:
                 msg = (_('repository tip rolled back to revision %d'
                          ' (undo %s: %s)\n')
                        % (oldtip, desc, detail))
             else:
                 msg = (_('repository tip rolled back to revision %d'
                          ' (undo %s)\n')
                        % (oldtip, desc))
         except IOError:
             msg = _('rolling back unknown transaction\n')
             desc = None
 
         if not force and self['.'] != self['tip'] and desc == 'commit':
             raise error.Abort(
                 _('rollback of last commit while not checked out '
                   'may lose data'), hint=_('use -f to force'))
 
         ui.status(msg)
         if dryrun:
             return 0
 
         parents = self.dirstate.parents()
         self.destroying()
         vfsmap = {'plain': self.vfs, '': self.svfs}
         transaction.rollback(self.svfs, vfsmap, 'undo', ui.warn,
                              checkambigfiles=_cachedfiles)
         if self.vfs.exists('undo.bookmarks'):
             self.vfs.rename('undo.bookmarks', 'bookmarks', checkambig=True)
         if self.svfs.exists('undo.phaseroots'):
             self.svfs.rename('undo.phaseroots', 'phaseroots', checkambig=True)
         self.invalidate()
 
         parentgone = (parents[0] not in self.changelog.nodemap or
                       parents[1] not in self.changelog.nodemap)
         if parentgone:
             # prevent dirstateguard from overwriting already restored one
             dsguard.close()
 
             narrowspec.restorebackup(self, 'undo.narrowspec')
             self.dirstate.restorebackup(None, 'undo.dirstate')
             try:
                 branch = self.vfs.read('undo.branch')
                 self.dirstate.setbranch(encoding.tolocal(branch))
             except IOError:
                 ui.warn(_('named branch could not be reset: '
                           'current branch is still \'%s\'\n')
                         % self.dirstate.branch())
 
             parents = tuple([p.rev() for p in self[None].parents()])
             if len(parents) > 1:
                 ui.status(_('working directory now based on '
                             'revisions %d and %d\n') % parents)
             else:
                 ui.status(_('working directory now based on '
                             'revision %d\n') % parents)
             mergemod.mergestate.clean(self, self['.'].node())
 
         # TODO: if we know which new heads may result from this rollback, pass
         # them to destroy(), which will prevent the branchhead cache from being
         # invalidated.
         self.destroyed()
         return 0
 
     def _buildcacheupdater(self, newtransaction):
         """called during transaction to build the callback updating cache
 
         Lives on the repository to help extension who might want to augment
         this logic. For this purpose, the created transaction is passed to the
         method.
         """
         # we must avoid cyclic reference between repo and transaction.
         reporef = weakref.ref(self)
         def updater(tr):
             repo = reporef()
             repo.updatecaches(tr)
         return updater
 
     @unfilteredmethod
     def updatecaches(self, tr=None, full=False):
         """warm appropriate caches
 
         If this function is called after a transaction closed. The transaction
         will be available in the 'tr' argument. This can be used to selectively
         update caches relevant to the changes in that transaction.
 
         If 'full' is set, make sure all caches the function knows about have
         up-to-date data. Even the ones usually loaded more lazily.
         """
         if tr is not None and tr.hookargs.get('source') == 'strip':
             # During strip, many caches are invalid but
             # later call to `destroyed` will refresh them.
             return
 
         if tr is None or tr.changes['origrepolen'] < len(self):
             # updating the unfiltered branchmap should refresh all the others,
             self.ui.debug('updating the branch cache\n')
             branchmap.updatecache(self.filtered('served'))
 
         if full:
             rbc = self.revbranchcache()
             for r in self.changelog:
                 rbc.branchinfo(r)
             rbc.write()
 
             # ensure the working copy parents are in the manifestfulltextcache
             for ctx in self['.'].parents():
                 ctx.manifest()  # accessing the manifest is enough
 
     def invalidatecaches(self):
 
         if '_tagscache' in vars(self):
             # can't use delattr on proxy
             del self.__dict__['_tagscache']
 
         self.unfiltered()._branchcaches.clear()
         self.invalidatevolatilesets()
         self._sparsesignaturecache.clear()
 
     def invalidatevolatilesets(self):
         self.filteredrevcache.clear()
         obsolete.clearobscaches(self)
 
     def invalidatedirstate(self):
         '''Invalidates the dirstate, causing the next call to dirstate
         to check if it was modified since the last time it was read,
         rereading it if it has.
 
         This is different to dirstate.invalidate() that it doesn't always
         rereads the dirstate. Use dirstate.invalidate() if you want to
         explicitly read the dirstate again (i.e. restoring it to a previous
         known good state).'''
         if hasunfilteredcache(self, 'dirstate'):
             for k in self.dirstate._filecache:
                 try:
                     delattr(self.dirstate, k)
                 except AttributeError:
                     pass
             delattr(self.unfiltered(), 'dirstate')
 
     def invalidate(self, clearfilecache=False):
         '''Invalidates both store and non-store parts other than dirstate
 
         If a transaction is running, invalidation of store is omitted,
         because discarding in-memory changes might cause inconsistency
         (e.g. incomplete fncache causes unintentional failure, but
         redundant one doesn't).
         '''
         unfiltered = self.unfiltered() # all file caches are stored unfiltered
         for k in list(self._filecache.keys()):
             # dirstate is invalidated separately in invalidatedirstate()
             if k == 'dirstate':
                 continue
             if (k == 'changelog' and
                 self.currenttransaction() and
                 self.changelog._delayed):
                 # The changelog object may store unwritten revisions. We don't
                 # want to lose them.
                 # TODO: Solve the problem instead of working around it.
                 continue
 
             if clearfilecache:
                 del self._filecache[k]
             try:
                 delattr(unfiltered, k)
             except AttributeError:
                 pass
         self.invalidatecaches()
         if not self.currenttransaction():
             # TODO: Changing contents of store outside transaction
             # causes inconsistency. We should make in-memory store
             # changes detectable, and abort if changed.
             self.store.invalidatecaches()
 
     def invalidateall(self):
         '''Fully invalidates both store and non-store parts, causing the
         subsequent operation to reread any outside changes.'''
         # extension should hook this to invalidate its caches
         self.invalidate()
         self.invalidatedirstate()
 
     @unfilteredmethod
     def _refreshfilecachestats(self, tr):
         """Reload stats of cached files so that they are flagged as valid"""
         for k, ce in self._filecache.items():
             k = pycompat.sysstr(k)
             if k == r'dirstate' or k not in self.__dict__:
                 continue
             ce.refresh()
 
     def _lock(self, vfs, lockname, wait, releasefn, acquirefn, desc,
               inheritchecker=None, parentenvvar=None):
         parentlock = None
         # the contents of parentenvvar are used by the underlying lock to
         # determine whether it can be inherited
         if parentenvvar is not None:
             parentlock = encoding.environ.get(parentenvvar)
 
         timeout = 0
         warntimeout = 0
         if wait:
             timeout = self.ui.configint("ui", "timeout")
             warntimeout = self.ui.configint("ui", "timeout.warn")
         # internal config: ui.signal-safe-lock
         signalsafe = self.ui.configbool('ui', 'signal-safe-lock')
 
         l = lockmod.trylock(self.ui, vfs, lockname, timeout, warntimeout,
                             releasefn=releasefn,
                             acquirefn=acquirefn, desc=desc,
                             inheritchecker=inheritchecker,
                             parentlock=parentlock,
                             signalsafe=signalsafe)
         return l
 
     def _afterlock(self, callback):
         """add a callback to be run when the repository is fully unlocked
 
         The callback will be executed when the outermost lock is released
         (with wlock being higher level than 'lock')."""
         for ref in (self._wlockref, self._lockref):
             l = ref and ref()
             if l and l.held:
                 l.postrelease.append(callback)
                 break
         else: # no lock have been found.
             callback()
 
     def lock(self, wait=True):
         '''Lock the repository store (.hg/store) and return a weak reference
         to the lock. Use this before modifying the store (e.g. committing or
         stripping). If you are opening a transaction, get a lock as well.)
 
         If both 'lock' and 'wlock' must be acquired, ensure you always acquires
         'wlock' first to avoid a dead-lock hazard.'''
         l = self._currentlock(self._lockref)
         if l is not None:
             l.lock()
             return l
 
         l = self._lock(self.svfs, "lock", wait, None,
                        self.invalidate, _('repository %s') % self.origroot)
         self._lockref = weakref.ref(l)
         return l
 
     def _wlockchecktransaction(self):
         if self.currenttransaction() is not None:
             raise error.LockInheritanceContractViolation(
                 'wlock cannot be inherited in the middle of a transaction')
 
     def wlock(self, wait=True):
         '''Lock the non-store parts of the repository (everything under
         .hg except .hg/store) and return a weak reference to the lock.
 
         Use this before modifying files in .hg.
 
         If both 'lock' and 'wlock' must be acquired, ensure you always acquires
         'wlock' first to avoid a dead-lock hazard.'''
         l = self._wlockref and self._wlockref()
         if l is not None and l.held:
             l.lock()
             return l
 
         # We do not need to check for non-waiting lock acquisition.  Such
         # acquisition would not cause dead-lock as they would just fail.
         if wait and (self.ui.configbool('devel', 'all-warnings')
                      or self.ui.configbool('devel', 'check-locks')):
             if self._currentlock(self._lockref) is not None:
                 self.ui.develwarn('"wlock" acquired after "lock"')
 
         def unlock():
             if self.dirstate.pendingparentchange():
                 self.dirstate.invalidate()
             else:
                 self.dirstate.write(None)
 
             self._filecache['dirstate'].refresh()
 
         l = self._lock(self.vfs, "wlock", wait, unlock,
                        self.invalidatedirstate, _('working directory of %s') %
                        self.origroot,
                        inheritchecker=self._wlockchecktransaction,
                        parentenvvar='HG_WLOCK_LOCKER')
         self._wlockref = weakref.ref(l)
         return l
 
     def _currentlock(self, lockref):
         """Returns the lock if it's held, or None if it's not."""
         if lockref is None:
             return None
         l = lockref()
         if l is None or not l.held:
             return None
         return l
 
     def currentwlock(self):
         """Returns the wlock if it's held, or None if it's not."""
         return self._currentlock(self._wlockref)
 
     def _filecommit(self, fctx, manifest1, manifest2, linkrev, tr, changelist):
         """
         commit an individual file as part of a larger transaction
         """
 
         fname = fctx.path()
         fparent1 = manifest1.get(fname, nullid)
         fparent2 = manifest2.get(fname, nullid)
         if isinstance(fctx, context.filectx):
             node = fctx.filenode()
             if node in [fparent1, fparent2]:
                 self.ui.debug('reusing %s filelog entry\n' % fname)
                 if manifest1.flags(fname) != fctx.flags():
                     changelist.append(fname)
                 return node
 
         flog = self.file(fname)
         meta = {}
         copy = fctx.renamed()
         if copy and copy[0] != fname:
             # Mark the new revision of this file as a copy of another
             # file.  This copy data will effectively act as a parent
             # of this new revision.  If this is a merge, the first
             # parent will be the nullid (meaning "look up the copy data")
             # and the second one will be the other parent.  For example:
             #
             # 0 --- 1 --- 3   rev1 changes file foo
             #   \       /     rev2 renames foo to bar and changes it
             #    \- 2 -/      rev3 should have bar with all changes and
             #                      should record that bar descends from
             #                      bar in rev2 and foo in rev1
             #
             # this allows this merge to succeed:
             #
             # 0 --- 1 --- 3   rev4 reverts the content change from rev2
             #   \       /     merging rev3 and rev4 should use bar@rev2
             #    \- 2 --- 4        as the merge base
             #
 
             cfname = copy[0]
             crev = manifest1.get(cfname)
             newfparent = fparent2
 
             if manifest2: # branch merge
                 if fparent2 == nullid or crev is None: # copied on remote side
                     if cfname in manifest2:
                         crev = manifest2[cfname]
                         newfparent = fparent1
 
             # Here, we used to search backwards through history to try to find
             # where the file copy came from if the source of a copy was not in
             # the parent directory. However, this doesn't actually make sense to
             # do (what does a copy from something not in your working copy even
             # mean?) and it causes bugs (eg, issue4476). Instead, we will warn
             # the user that copy information was dropped, so if they didn't
             # expect this outcome it can be fixed, but this is the correct
             # behavior in this circumstance.
 
             if crev:
                 self.ui.debug(" %s: copy %s:%s\n" % (fname, cfname, hex(crev)))
                 meta["copy"] = cfname
                 meta["copyrev"] = hex(crev)
                 fparent1, fparent2 = nullid, newfparent
             else:
                 self.ui.warn(_("warning: can't find ancestor for '%s' "
                                "copied from '%s'!\n") % (fname, cfname))
 
         elif fparent1 == nullid:
             fparent1, fparent2 = fparent2, nullid
         elif fparent2 != nullid:
             # is one parent an ancestor of the other?
             fparentancestors = flog.commonancestorsheads(fparent1, fparent2)
             if fparent1 in fparentancestors:
                 fparent1, fparent2 = fparent2, nullid
             elif fparent2 in fparentancestors:
                 fparent2 = nullid
 
         # is the file changed?
         text = fctx.data()
         if fparent2 != nullid or flog.cmp(fparent1, text) or meta:
             changelist.append(fname)
             return flog.add(text, meta, tr, linkrev, fparent1, fparent2)
         # are just the flags changed during merge?
         elif fname in manifest1 and manifest1.flags(fname) != fctx.flags():
             changelist.append(fname)
 
         return fparent1
 
     def checkcommitpatterns(self, wctx, vdirs, match, status, fail):
         """check for commit arguments that aren't committable"""
         if match.isexact() or match.prefix():
             matched = set(status.modified + status.added + status.removed)
 
             for f in match.files():
                 f = self.dirstate.normalize(f)
                 if f == '.' or f in matched or f in wctx.substate:
                     continue
                 if f in status.deleted:
                     fail(f, _('file not found!'))
                 if f in vdirs: # visited directory
                     d = f + '/'
                     for mf in matched:
                         if mf.startswith(d):
                             break
                     else:
                         fail(f, _("no match under directory!"))
                 elif f not in self.dirstate:
                     fail(f, _("file not tracked!"))
 
     @unfilteredmethod
     def commit(self, text="", user=None, date=None, match=None, force=False,
                editor=False, extra=None):
         """Add a new revision to current repository.
 
         Revision information is gathered from the working directory,
         match can be used to filter the committed files. If editor is
         supplied, it is called to get a commit message.
         """
         if extra is None:
             extra = {}
 
         def fail(f, msg):
             raise error.Abort('%s: %s' % (f, msg))
 
         if not match:
             match = matchmod.always(self.root, '')
 
         if not force:
             vdirs = []
             match.explicitdir = vdirs.append
             match.bad = fail
 
         wlock = lock = tr = None
         try:
             wlock = self.wlock()
             lock = self.lock() # for recent changelog (see issue4368)
 
             wctx = self[None]
             merge = len(wctx.parents()) > 1
 
             if not force and merge and not match.always():
                 raise error.Abort(_('cannot partially commit a merge '
                                    '(do not specify files or patterns)'))
 
             status = self.status(match=match, clean=force)
             if force:
                 status.modified.extend(status.clean) # mq may commit clean files
 
             # check subrepos
             subs, commitsubs, newstate = subrepoutil.precommit(
                 self.ui, wctx, status, match, force=force)
 
             # make sure all explicit patterns are matched
             if not force:
                 self.checkcommitpatterns(wctx, vdirs, match, status, fail)
 
             cctx = context.workingcommitctx(self, status,
                                             text, user, date, extra)
 
             # internal config: ui.allowemptycommit
             allowemptycommit = (wctx.branch() != wctx.p1().branch()
                                 or extra.get('close') or merge or cctx.files()
                                 or self.ui.configbool('ui', 'allowemptycommit'))
             if not allowemptycommit:
                 return None
 
             if merge and cctx.deleted():
                 raise error.Abort(_("cannot commit merge with missing files"))
 
             ms = mergemod.mergestate.read(self)
             mergeutil.checkunresolved(ms)
 
             if editor:
                 cctx._text = editor(self, cctx, subs)
             edited = (text != cctx._text)
 
             # Save commit message in case this transaction gets rolled back
             # (e.g. by a pretxncommit hook).  Leave the content alone on
             # the assumption that the user will use the same editor again.
             msgfn = self.savecommitmessage(cctx._text)
 
             # commit subs and write new state
             if subs:
                 for s in sorted(commitsubs):
                     sub = wctx.sub(s)
                     self.ui.status(_('committing subrepository %s\n') %
                                    subrepoutil.subrelpath(sub))
                     sr = sub.commit(cctx._text, user, date)
                     newstate[s] = (newstate[s][0], sr)
                 subrepoutil.writestate(self, newstate)
 
             p1, p2 = self.dirstate.parents()
             hookp1, hookp2 = hex(p1), (p2 != nullid and hex(p2) or '')
             try:
                 self.hook("precommit", throw=True, parent1=hookp1,
                           parent2=hookp2)
                 tr = self.transaction('commit')
                 ret = self.commitctx(cctx, True)
             except: # re-raises
                 if edited:
                     self.ui.write(
                         _('note: commit message saved in %s\n') % msgfn)
                 raise
             # update bookmarks, dirstate and mergestate
             bookmarks.update(self, [p1, p2], ret)
             cctx.markcommitted(ret)
             ms.reset()
             tr.close()
 
         finally:
             lockmod.release(tr, lock, wlock)
 
         def commithook(node=hex(ret), parent1=hookp1, parent2=hookp2):
             # hack for command that use a temporary commit (eg: histedit)
             # temporary commit got stripped before hook release
             if self.changelog.hasnode(ret):
                 self.hook("commit", node=node, parent1=parent1,
                           parent2=parent2)
         self._afterlock(commithook)
         return ret
 
     @unfilteredmethod
     def commitctx(self, ctx, error=False):
         """Add a new revision to current repository.
         Revision information is passed via the context argument.
 
         ctx.files() should list all files involved in this commit, i.e.
         modified/added/removed files. On merge, it may be wider than the
         ctx.files() to be committed, since any file nodes derived directly
         from p1 or p2 are excluded from the committed ctx.files().
         """
 
         tr = None
         p1, p2 = ctx.p1(), ctx.p2()
         user = ctx.user()
 
         lock = self.lock()
         try:
             tr = self.transaction("commit")
             trp = weakref.proxy(tr)
 
             if ctx.manifestnode():
                 # reuse an existing manifest revision
                 self.ui.debug('reusing known manifest\n')
                 mn = ctx.manifestnode()
                 files = ctx.files()
             elif ctx.files():
                 m1ctx = p1.manifestctx()
                 m2ctx = p2.manifestctx()
                 mctx = m1ctx.copy()
 
                 m = mctx.read()
                 m1 = m1ctx.read()
                 m2 = m2ctx.read()
 
                 # check in files
                 added = []
                 changed = []
                 removed = list(ctx.removed())
                 linkrev = len(self)
                 self.ui.note(_("committing files:\n"))
                 for f in sorted(ctx.modified() + ctx.added()):
                     self.ui.note(f + "\n")
                     try:
                         fctx = ctx[f]
                         if fctx is None:
                             removed.append(f)
                         else:
                             added.append(f)
                             m[f] = self._filecommit(fctx, m1, m2, linkrev,
                                                     trp, changed)
                             m.setflag(f, fctx.flags())
                     except OSError as inst:
                         self.ui.warn(_("trouble committing %s!\n") % f)
                         raise
                     except IOError as inst:
                         errcode = getattr(inst, 'errno', errno.ENOENT)
                         if error or errcode and errcode != errno.ENOENT:
                             self.ui.warn(_("trouble committing %s!\n") % f)
                         raise
 
                 # update manifest
                 removed = [f for f in sorted(removed) if f in m1 or f in m2]
                 drop = [f for f in removed if f in m]
                 for f in drop:
                     del m[f]
                 files = changed + removed
                 md = None
                 if not files:
                     # if no "files" actually changed in terms of the changelog,
                     # try hard to detect unmodified manifest entry so that the
                     # exact same commit can be reproduced later on convert.
                     md = m1.diff(m, scmutil.matchfiles(self, ctx.files()))
                 if not files and md:
                     self.ui.debug('not reusing manifest (no file change in '
                                   'changelog, but manifest differs)\n')
                 if files or md:
                     self.ui.note(_("committing manifest\n"))
                     # we're using narrowmatch here since it's already applied at
                     # other stages (such as dirstate.walk), so we're already
                     # ignoring things outside of narrowspec in most cases. The
                     # one case where we might have files outside the narrowspec
                     # at this point is merges, and we already error out in the
                     # case where the merge has files outside of the narrowspec,
                     # so this is safe.
                     mn = mctx.write(trp, linkrev,
                                     p1.manifestnode(), p2.manifestnode(),
                                     added, drop, match=self.narrowmatch())
                 else:
                     self.ui.debug('reusing manifest form p1 (listed files '
                                   'actually unchanged)\n')
                     mn = p1.manifestnode()
             else:
                 self.ui.debug('reusing manifest from p1 (no file change)\n')
                 mn = p1.manifestnode()
                 files = []
 
             # update changelog
             self.ui.note(_("committing changelog\n"))
             self.changelog.delayupdate(tr)
             n = self.changelog.add(mn, files, ctx.description(),
                                    trp, p1.node(), p2.node(),
                                    user, ctx.date(), ctx.extra().copy())
             xp1, xp2 = p1.hex(), p2 and p2.hex() or ''
             self.hook('pretxncommit', throw=True, node=hex(n), parent1=xp1,
                       parent2=xp2)
             # set the new commit is proper phase
             targetphase = subrepoutil.newcommitphase(self.ui, ctx)
             if targetphase:
                 # retract boundary do not alter parent changeset.
                 # if a parent have higher the resulting phase will
                 # be compliant anyway
                 #
                 # if minimal phase was 0 we don't need to retract anything
                 phases.registernew(self, tr, targetphase, [n])
             tr.close()
             return n
         finally:
             if tr:
                 tr.release()
             lock.release()
 
     @unfilteredmethod
     def destroying(self):
         '''Inform the repository that nodes are about to be destroyed.
         Intended for use by strip and rollback, so there's a common
         place for anything that has to be done before destroying history.
 
         This is mostly useful for saving state that is in memory and waiting
         to be flushed when the current lock is released. Because a call to
         destroyed is imminent, the repo will be invalidated causing those
         changes to stay in memory (waiting for the next unlock), or vanish
         completely.
         '''
         # When using the same lock to commit and strip, the phasecache is left
         # dirty after committing. Then when we strip, the repo is invalidated,
         # causing those changes to disappear.
         if '_phasecache' in vars(self):
             self._phasecache.write()
 
     @unfilteredmethod
     def destroyed(self):
         '''Inform the repository that nodes have been destroyed.
         Intended for use by strip and rollback, so there's a common
         place for anything that has to be done after destroying history.
         '''
         # When one tries to:
         # 1) destroy nodes thus calling this method (e.g. strip)
         # 2) use phasecache somewhere (e.g. commit)
         #
         # then 2) will fail because the phasecache contains nodes that were
         # removed. We can either remove phasecache from the filecache,
         # causing it to reload next time it is accessed, or simply filter
         # the removed nodes now and write the updated cache.
         self._phasecache.filterunknown(self)
         self._phasecache.write()
 
         # refresh all repository caches
         self.updatecaches()
 
         # Ensure the persistent tag cache is updated.  Doing it now
         # means that the tag cache only has to worry about destroyed
         # heads immediately after a strip/rollback.  That in turn
         # guarantees that "cachetip == currenttip" (comparing both rev
         # and node) always means no nodes have been added or destroyed.
 
         # XXX this is suboptimal when qrefresh'ing: we strip the current
         # head, refresh the tag cache, then immediately add a new head.
         # But I think doing it this way is necessary for the "instant
         # tag cache retrieval" case to work.
         self.invalidate()
 
     def status(self, node1='.', node2=None, match=None,
                ignored=False, clean=False, unknown=False,
                listsubrepos=False):
         '''a convenience method that calls node1.status(node2)'''
         return self[node1].status(node2, match, ignored, clean, unknown,
                                   listsubrepos)
 
     def addpostdsstatus(self, ps):
         """Add a callback to run within the wlock, at the point at which status
         fixups happen.
 
         On status completion, callback(wctx, status) will be called with the
         wlock held, unless the dirstate has changed from underneath or the wlock
         couldn't be grabbed.
 
         Callbacks should not capture and use a cached copy of the dirstate --
         it might change in the meanwhile. Instead, they should access the
         dirstate via wctx.repo().dirstate.
 
         This list is emptied out after each status run -- extensions should
         make sure it adds to this list each time dirstate.status is called.
         Extensions should also make sure they don't call this for statuses
         that don't involve the dirstate.
         """
 
         # The list is located here for uniqueness reasons -- it is actually
         # managed by the workingctx, but that isn't unique per-repo.
         self._postdsstatus.append(ps)
 
     def postdsstatus(self):
         """Used by workingctx to get the list of post-dirstate-status hooks."""
         return self._postdsstatus
 
     def clearpostdsstatus(self):
         """Used by workingctx to clear post-dirstate-status hooks."""
         del self._postdsstatus[:]
 
     def heads(self, start=None):
         if start is None:
             cl = self.changelog
             headrevs = reversed(cl.headrevs())
             return [cl.node(rev) for rev in headrevs]
 
         heads = self.changelog.heads(start)
         # sort the output in rev descending order
         return sorted(heads, key=self.changelog.rev, reverse=True)
 
     def branchheads(self, branch=None, start=None, closed=False):
         '''return a (possibly filtered) list of heads for the given branch
 
         Heads are returned in topological order, from newest to oldest.
         If branch is None, use the dirstate branch.
         If start is not None, return only heads reachable from start.
         If closed is True, return heads that are marked as closed as well.
         '''
         if branch is None:
             branch = self[None].branch()
         branches = self.branchmap()
         if branch not in branches:
             return []
         # the cache returns heads ordered lowest to highest
         bheads = list(reversed(branches.branchheads(branch, closed=closed)))
         if start is not None:
             # filter out the heads that cannot be reached from startrev
             fbheads = set(self.changelog.nodesbetween([start], bheads)[2])
             bheads = [h for h in bheads if h in fbheads]
         return bheads
 
     def branches(self, nodes):
         if not nodes:
             nodes = [self.changelog.tip()]
         b = []
         for n in nodes:
             t = n
             while True:
                 p = self.changelog.parents(n)
                 if p[1] != nullid or p[0] == nullid:
                     b.append((t, n, p[0], p[1]))
                     break
                 n = p[0]
         return b
 
     def between(self, pairs):
         r = []
 
         for top, bottom in pairs:
             n, l, i = top, [], 0
             f = 1
 
             while n != bottom and n != nullid:
                 p = self.changelog.parents(n)[0]
                 if i == f:
                     l.append(n)
                     f = f * 2
                 n = p
                 i += 1
 
             r.append(l)
 
         return r
 
     def checkpush(self, pushop):
         """Extensions can override this function if additional checks have
         to be performed before pushing, or call it if they override push
         command.
         """
 
     @unfilteredpropertycache
     def prepushoutgoinghooks(self):
         """Return util.hooks consists of a pushop with repo, remote, outgoing
         methods, which are called before pushing changesets.
         """
         return util.hooks()
 
     def pushkey(self, namespace, key, old, new):
         try:
             tr = self.currenttransaction()
             hookargs = {}
             if tr is not None:
                 hookargs.update(tr.hookargs)
             hookargs = pycompat.strkwargs(hookargs)
             hookargs[r'namespace'] = namespace
             hookargs[r'key'] = key
             hookargs[r'old'] = old
             hookargs[r'new'] = new
             self.hook('prepushkey', throw=True, **hookargs)
         except error.HookAbort as exc:
             self.ui.write_err(_("pushkey-abort: %s\n") % exc)
             if exc.hint:
                 self.ui.write_err(_("(%s)\n") % exc.hint)
             return False
         self.ui.debug('pushing key for "%s:%s"\n' % (namespace, key))
         ret = pushkey.push(self, namespace, key, old, new)
         def runhook():
             self.hook('pushkey', namespace=namespace, key=key, old=old, new=new,
                       ret=ret)
         self._afterlock(runhook)
         return ret
 
     def listkeys(self, namespace):
         self.hook('prelistkeys', throw=True, namespace=namespace)
         self.ui.debug('listing keys for "%s"\n' % namespace)
         values = pushkey.list(self, namespace)
         self.hook('listkeys', namespace=namespace, values=values)
         return values
 
     def debugwireargs(self, one, two, three=None, four=None, five=None):
         '''used to test argument passing over the wire'''
         return "%s %s %s %s %s" % (one, two, pycompat.bytestr(three),
                                    pycompat.bytestr(four),
                                    pycompat.bytestr(five))
 
     def savecommitmessage(self, text):
         fp = self.vfs('last-message.txt', 'wb')
         try:
             fp.write(text)
         finally:
             fp.close()
         return self.pathto(fp.name[len(self.root) + 1:])
 
 # used to avoid circular references so destructors work
 def aftertrans(files):
     renamefiles = [tuple(t) for t in files]
     def a():
         for vfs, src, dest in renamefiles:
             # if src and dest refer to a same file, vfs.rename is a no-op,
             # leaving both src and dest on disk. delete dest to make sure
             # the rename couldn't be such a no-op.
             vfs.tryunlink(dest)
             try:
                 vfs.rename(src, dest)
             except OSError: # journal file does not yet exist
                 pass
     return a
 
 def undoname(fn):
     base, name = os.path.split(fn)
     assert name.startswith('journal')
     return os.path.join(base, name.replace('journal', 'undo', 1))
 
 def instance(ui, path, create, intents=None, createopts=None):
     localpath = util.urllocalpath(path)
     if create:
         createrepository(ui, localpath, createopts=createopts)
 
     return makelocalrepository(ui, localpath, intents=intents)
 
 def islocal(path):
     return True
 
 def newreporequirements(ui, createopts=None):
     """Determine the set of requirements for a new local repository.
 
     Extensions can wrap this function to specify custom requirements for
     new repositories.
     """
     createopts = createopts or {}
 
     requirements = {'revlogv1'}
     if ui.configbool('format', 'usestore'):
         requirements.add('store')
         if ui.configbool('format', 'usefncache'):
             requirements.add('fncache')
             if ui.configbool('format', 'dotencode'):
                 requirements.add('dotencode')
 
     compengine = ui.config('experimental', 'format.compression')
     if compengine not in util.compengines:
         raise error.Abort(_('compression engine %s defined by '
                             'experimental.format.compression not available') %
                           compengine,
                           hint=_('run "hg debuginstall" to list available '
                                  'compression engines'))
 
     # zlib is the historical default and doesn't need an explicit requirement.
     if compengine != 'zlib':
         requirements.add('exp-compression-%s' % compengine)
 
     if scmutil.gdinitconfig(ui):
         requirements.add('generaldelta')
     if ui.configbool('experimental', 'treemanifest'):
         requirements.add('treemanifest')
     # experimental config: format.sparse-revlog
     if ui.configbool('format', 'sparse-revlog'):
         requirements.add(SPARSEREVLOG_REQUIREMENT)
 
     revlogv2 = ui.config('experimental', 'revlogv2')
     if revlogv2 == 'enable-unstable-format-and-corrupt-my-data':
         requirements.remove('revlogv1')
         # generaldelta is implied by revlogv2.
         requirements.discard('generaldelta')
         requirements.add(REVLOGV2_REQUIREMENT)
     # experimental config: format.internal-phase
     if ui.configbool('format', 'internal-phase'):
         requirements.add('internal-phase')
 
     if createopts.get('narrowfiles'):
         requirements.add(repository.NARROW_REQUIREMENT)
 
     return requirements
 
 def filterknowncreateopts(ui, createopts):
     """Filters a dict of repo creation options against options that are known.
 
     Receives a dict of repo creation options and returns a dict of those
     options that we don't know how to handle.
 
     This function is called as part of repository creation. If the
     returned dict contains any items, repository creation will not
     be allowed, as it means there was a request to create a repository
     with options not recognized by loaded code.
 
     Extensions can wrap this function to filter out creation options
     they know how to handle.
     """
     known = {'narrowfiles'}
 
     return {k: v for k, v in createopts.items() if k not in known}
 
 def createrepository(ui, path, createopts=None):
     """Create a new repository in a vfs.
 
     ``path`` path to the new repo's working directory.
     ``createopts`` options for the new repository.
     """
     createopts = createopts or {}
 
     unknownopts = filterknowncreateopts(ui, createopts)
 
     if not isinstance(unknownopts, dict):
         raise error.ProgrammingError('filterknowncreateopts() did not return '
                                      'a dict')
 
     if unknownopts:
         raise error.Abort(_('unable to create repository because of unknown '
                             'creation option: %s') %
                           ', '.sorted(unknownopts),
                           hint=_('is a required extension not loaded?'))
 
     requirements = newreporequirements(ui, createopts=createopts)
 
     wdirvfs = vfsmod.vfs(path, expandpath=True, realpath=True)
     if not wdirvfs.exists():
         wdirvfs.makedirs()
 
     hgvfs = vfsmod.vfs(wdirvfs.join(b'.hg'))
     if hgvfs.exists():
         raise error.RepoError(_('repository %s already exists') % path)
 
     hgvfs.makedir(notindexed=True)
 
     if b'store' in requirements:
         hgvfs.mkdir(b'store')
 
         # We create an invalid changelog outside the store so very old
         # Mercurial versions (which didn't know about the requirements
         # file) encounter an error on reading the changelog. This
         # effectively locks out old clients and prevents them from
         # mucking with a repo in an unknown format.
         #
         # The revlog header has version 2, which won't be recognized by
         # such old clients.
         hgvfs.append(b'00changelog.i',
                      b'\0\0\0\2 dummy changelog to prevent using the old repo '
                      b'layout')
 
     scmutil.writerequires(hgvfs, requirements)
 
 def poisonrepository(repo):
     """Poison a repository instance so it can no longer be used."""
     # Perform any cleanup on the instance.
     repo.close()
 
     # Our strategy is to replace the type of the object with one that
     # has all attribute lookups result in error.
     #
     # But we have to allow the close() method because some constructors
     # of repos call close() on repo references.
     class poisonedrepository(object):
         def __getattribute__(self, item):
             if item == r'close':
                 return object.__getattribute__(self, item)
 
             raise error.ProgrammingError('repo instances should not be used '
                                          'after unshare')
 
         def close(self):
             pass
 
     # We may have a repoview, which intercepts __setattr__. So be sure
     # we operate at the lowest level possible.
     object.__setattr__(repo, r'__class__', poisonedrepository)

mercurial/manifest.py

@@ -1,2017 +1,2017
 # manifest.py - manifest revision class for mercurial
 #
 # Copyright 2005-2007 Matt Mackall <[e-mail]>
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 from __future__ import absolute_import
 
 import heapq
 import itertools
 import struct
 import weakref
 
 from .i18n import _
 from .node import (
     bin,
     hex,
     nullid,
     nullrev,
 )
 from . import (
     error,
     mdiff,
     policy,
     pycompat,
     repository,
     revlog,
     util,
 )
 from .utils import (
     interfaceutil,
 )
 
 parsers = policy.importmod(r'parsers')
 propertycache = util.propertycache
 
 def _parse(data):
     # This method does a little bit of excessive-looking
     # precondition checking. This is so that the behavior of this
     # class exactly matches its C counterpart to try and help
     # prevent surprise breakage for anyone that develops against
     # the pure version.
     if data and data[-1:] != '\n':
         raise ValueError('Manifest did not end in a newline.')
     prev = None
     for l in data.splitlines():
         if prev is not None and prev > l:
             raise ValueError('Manifest lines not in sorted order.')
         prev = l
         f, n = l.split('\0')
         if len(n) > 40:
             yield f, bin(n[:40]), n[40:]
         else:
             yield f, bin(n), ''
 
 def _text(it):
     files = []
     lines = []
     for f, n, fl in it:
         files.append(f)
         # if this is changed to support newlines in filenames,
         # be sure to check the templates/ dir again (especially *-raw.tmpl)
         lines.append("%s\0%s%s\n" % (f, hex(n), fl))
 
     _checkforbidden(files)
     return ''.join(lines)
 
 class lazymanifestiter(object):
     def __init__(self, lm):
         self.pos = 0
         self.lm = lm
 
     def __iter__(self):
         return self
 
     def next(self):
         try:
             data, pos = self.lm._get(self.pos)
         except IndexError:
             raise StopIteration
         if pos == -1:
             self.pos += 1
             return data[0]
         self.pos += 1
         zeropos = data.find('\x00', pos)
         return data[pos:zeropos]
 
     __next__ = next
 
 class lazymanifestiterentries(object):
     def __init__(self, lm):
         self.lm = lm
         self.pos = 0
 
     def __iter__(self):
         return self
 
     def next(self):
         try:
             data, pos = self.lm._get(self.pos)
         except IndexError:
             raise StopIteration
         if pos == -1:
             self.pos += 1
             return data
         zeropos = data.find('\x00', pos)
         hashval = unhexlify(data, self.lm.extrainfo[self.pos],
                             zeropos + 1, 40)
         flags = self.lm._getflags(data, self.pos, zeropos)
         self.pos += 1
         return (data[pos:zeropos], hashval, flags)
 
     __next__ = next
 
 def unhexlify(data, extra, pos, length):
     s = bin(data[pos:pos + length])
     if extra:
         s += chr(extra & 0xff)
     return s
 
 def _cmp(a, b):
     return (a > b) - (a < b)
 
 class _lazymanifest(object):
     def __init__(self, data, positions=None, extrainfo=None, extradata=None):
         if positions is None:
             self.positions = self.findlines(data)
             self.extrainfo = [0] * len(self.positions)
             self.data = data
             self.extradata = []
         else:
             self.positions = positions[:]
             self.extrainfo = extrainfo[:]
             self.extradata = extradata[:]
             self.data = data
 
     def findlines(self, data):
         if not data:
             return []
         pos = data.find("\n")
         if pos == -1 or data[-1:] != '\n':
             raise ValueError("Manifest did not end in a newline.")
         positions = [0]
         prev = data[:data.find('\x00')]
         while pos < len(data) - 1 and pos != -1:
             positions.append(pos + 1)
             nexts = data[pos + 1:data.find('\x00', pos + 1)]
             if nexts < prev:
                 raise ValueError("Manifest lines not in sorted order.")
             prev = nexts
             pos = data.find("\n", pos + 1)
         return positions
 
     def _get(self, index):
         # get the position encoded in pos:
         #   positive number is an index in 'data'
         #   negative number is in extrapieces
         pos = self.positions[index]
         if pos >= 0:
             return self.data, pos
         return self.extradata[-pos - 1], -1
 
     def _getkey(self, pos):
         if pos >= 0:
             return self.data[pos:self.data.find('\x00', pos + 1)]
         return self.extradata[-pos - 1][0]
 
     def bsearch(self, key):
         first = 0
         last = len(self.positions) - 1
 
         while first <= last:
             midpoint = (first + last)//2
             nextpos = self.positions[midpoint]
             candidate = self._getkey(nextpos)
             r = _cmp(key, candidate)
             if r == 0:
                 return midpoint
             else:
                 if r < 0:
                     last = midpoint - 1
                 else:
                     first = midpoint + 1
         return -1
 
     def bsearch2(self, key):
         # same as the above, but will always return the position
         # done for performance reasons
         first = 0
         last = len(self.positions) - 1
 
         while first <= last:
             midpoint = (first + last)//2
             nextpos = self.positions[midpoint]
             candidate = self._getkey(nextpos)
             r = _cmp(key, candidate)
             if r == 0:
                 return (midpoint, True)
             else:
                 if r < 0:
                     last = midpoint - 1
                 else:
                     first = midpoint + 1
         return (first, False)
 
     def __contains__(self, key):
         return self.bsearch(key) != -1
 
     def _getflags(self, data, needle, pos):
         start = pos + 41
         end = data.find("\n", start)
         if end == -1:
             end = len(data) - 1
         if start == end:
             return ''
         return self.data[start:end]
 
     def __getitem__(self, key):
         if not isinstance(key, bytes):
             raise TypeError("getitem: manifest keys must be a bytes.")
         needle = self.bsearch(key)
         if needle == -1:
             raise KeyError
         data, pos = self._get(needle)
         if pos == -1:
             return (data[1], data[2])
         zeropos = data.find('\x00', pos)
         assert 0 <= needle <= len(self.positions)
         assert len(self.extrainfo) == len(self.positions)
         hashval = unhexlify(data, self.extrainfo[needle], zeropos + 1, 40)
         flags = self._getflags(data, needle, zeropos)
         return (hashval, flags)
 
     def __delitem__(self, key):
         needle, found = self.bsearch2(key)
         if not found:
             raise KeyError
         cur = self.positions[needle]
         self.positions = self.positions[:needle] + self.positions[needle + 1:]
         self.extrainfo = self.extrainfo[:needle] + self.extrainfo[needle + 1:]
         if cur >= 0:
             self.data = self.data[:cur] + '\x00' + self.data[cur + 1:]
 
     def __setitem__(self, key, value):
         if not isinstance(key, bytes):
             raise TypeError("setitem: manifest keys must be a byte string.")
         if not isinstance(value, tuple) or len(value) != 2:
             raise TypeError("Manifest values must be a tuple of (node, flags).")
         hashval = value[0]
         if not isinstance(hashval, bytes) or not 20 <= len(hashval) <= 22:
             raise TypeError("node must be a 20-byte byte string")
         flags = value[1]
         if len(hashval) == 22:
             hashval = hashval[:-1]
         if not isinstance(flags, bytes) or len(flags) > 1:
             raise TypeError("flags must a 0 or 1 byte string, got %r", flags)
         needle, found = self.bsearch2(key)
         if found:
             # put the item
             pos = self.positions[needle]
             if pos < 0:
                 self.extradata[-pos - 1] = (key, hashval, value[1])
             else:
                 # just don't bother
                 self.extradata.append((key, hashval, value[1]))
                 self.positions[needle] = -len(self.extradata)
         else:
             # not found, put it in with extra positions
             self.extradata.append((key, hashval, value[1]))
             self.positions = (self.positions[:needle] + [-len(self.extradata)]
                               + self.positions[needle:])
             self.extrainfo = (self.extrainfo[:needle] + [0] +
                               self.extrainfo[needle:])
 
     def copy(self):
         # XXX call _compact like in C?
         return _lazymanifest(self.data, self.positions, self.extrainfo,
             self.extradata)
 
     def _compact(self):
         # hopefully not called TOO often
         if len(self.extradata) == 0:
             return
         l = []
         last_cut = 0
         i = 0
         offset = 0
         self.extrainfo = [0] * len(self.positions)
         while i < len(self.positions):
             if self.positions[i] >= 0:
                 cur = self.positions[i]
                 last_cut = cur
                 while True:
                     self.positions[i] = offset
                     i += 1
                     if i == len(self.positions) or self.positions[i] < 0:
                         break
                     offset += self.positions[i] - cur
                     cur = self.positions[i]
                 end_cut = self.data.find('\n', cur)
                 if end_cut != -1:
                     end_cut += 1
                 offset += end_cut - cur
                 l.append(self.data[last_cut:end_cut])
             else:
                 while i < len(self.positions) and self.positions[i] < 0:
                     cur = self.positions[i]
                     t = self.extradata[-cur - 1]
                     l.append(self._pack(t))
                     self.positions[i] = offset
                     if len(t[1]) > 20:
                         self.extrainfo[i] = ord(t[1][21])
                     offset += len(l[-1])
                     i += 1
         self.data = ''.join(l)
         self.extradata = []
 
     def _pack(self, d):
         return d[0] + '\x00' + hex(d[1][:20]) + d[2] + '\n'
 
     def text(self):
         self._compact()
         return self.data
 
     def diff(self, m2, clean=False):
         '''Finds changes between the current manifest and m2.'''
         # XXX think whether efficiency matters here
         diff = {}
 
         for fn, e1, flags in self.iterentries():
             if fn not in m2:
                 diff[fn] = (e1, flags), (None, '')
             else:
                 e2 = m2[fn]
                 if (e1, flags) != e2:
                     diff[fn] = (e1, flags), e2
                 elif clean:
                     diff[fn] = None
 
         for fn, e2, flags in m2.iterentries():
             if fn not in self:
                 diff[fn] = (None, ''), (e2, flags)
 
         return diff
 
     def iterentries(self):
         return lazymanifestiterentries(self)
 
     def iterkeys(self):
         return lazymanifestiter(self)
 
     def __iter__(self):
         return lazymanifestiter(self)
 
     def __len__(self):
         return len(self.positions)
 
     def filtercopy(self, filterfn):
         # XXX should be optimized
         c = _lazymanifest('')
         for f, n, fl in self.iterentries():
             if filterfn(f):
                 c[f] = n, fl
         return c
 
 try:
     _lazymanifest = parsers.lazymanifest
 except AttributeError:
     pass
 
 @interfaceutil.implementer(repository.imanifestdict)
 class manifestdict(object):
     def __init__(self, data=''):
         self._lm = _lazymanifest(data)
 
     def __getitem__(self, key):
         return self._lm[key][0]
 
     def find(self, key):
         return self._lm[key]
 
     def __len__(self):
         return len(self._lm)
 
     def __nonzero__(self):
         # nonzero is covered by the __len__ function, but implementing it here
         # makes it easier for extensions to override.
         return len(self._lm) != 0
 
     __bool__ = __nonzero__
 
     def __setitem__(self, key, node):
         self._lm[key] = node, self.flags(key, '')
 
     def __contains__(self, key):
         if key is None:
             return False
         return key in self._lm
 
     def __delitem__(self, key):
         del self._lm[key]
 
     def __iter__(self):
         return self._lm.__iter__()
 
     def iterkeys(self):
         return self._lm.iterkeys()
 
     def keys(self):
         return list(self.iterkeys())
 
     def filesnotin(self, m2, match=None):
         '''Set of files in this manifest that are not in the other'''
         if match:
             m1 = self.matches(match)
             m2 = m2.matches(match)
             return m1.filesnotin(m2)
         diff = self.diff(m2)
         files = set(filepath
                     for filepath, hashflags in diff.iteritems()
                     if hashflags[1][0] is None)
         return files
 
     @propertycache
     def _dirs(self):
         return util.dirs(self)
 
     def dirs(self):
         return self._dirs
 
     def hasdir(self, dir):
         return dir in self._dirs
 
     def _filesfastpath(self, match):
         '''Checks whether we can correctly and quickly iterate over matcher
         files instead of over manifest files.'''
         files = match.files()
         return (len(files) < 100 and (match.isexact() or
             (match.prefix() and all(fn in self for fn in files))))
 
     def walk(self, match):
         '''Generates matching file names.
 
         Equivalent to manifest.matches(match).iterkeys(), but without creating
         an entirely new manifest.
 
         It also reports nonexistent files by marking them bad with match.bad().
         '''
         if match.always():
             for f in iter(self):
                 yield f
             return
 
         fset = set(match.files())
 
         # avoid the entire walk if we're only looking for specific files
         if self._filesfastpath(match):
             for fn in sorted(fset):
                 yield fn
             return
 
         for fn in self:
             if fn in fset:
                 # specified pattern is the exact name
                 fset.remove(fn)
             if match(fn):
                 yield fn
 
         # for dirstate.walk, files=['.'] means "walk the whole tree".
         # follow that here, too
         fset.discard('.')
 
         for fn in sorted(fset):
             if not self.hasdir(fn):
                 match.bad(fn, None)
 
     def matches(self, match):
         '''generate a new manifest filtered by the match argument'''
         if match.always():
             return self.copy()
 
         if self._filesfastpath(match):
             m = manifestdict()
             lm = self._lm
             for fn in match.files():
                 if fn in lm:
                     m._lm[fn] = lm[fn]
             return m
 
         m = manifestdict()
         m._lm = self._lm.filtercopy(match)
         return m
 
     def diff(self, m2, match=None, clean=False):
         '''Finds changes between the current manifest and m2.
 
         Args:
           m2: the manifest to which this manifest should be compared.
           clean: if true, include files unchanged between these manifests
                  with a None value in the returned dictionary.
 
         The result is returned as a dict with filename as key and
         values of the form ((n1,fl1),(n2,fl2)), where n1/n2 is the
         nodeid in the current/other manifest and fl1/fl2 is the flag
         in the current/other manifest. Where the file does not exist,
         the nodeid will be None and the flags will be the empty
         string.
         '''
         if match:
             m1 = self.matches(match)
             m2 = m2.matches(match)
             return m1.diff(m2, clean=clean)
         return self._lm.diff(m2._lm, clean)
 
     def setflag(self, key, flag):
         self._lm[key] = self[key], flag
 
     def get(self, key, default=None):
         try:
             return self._lm[key][0]
         except KeyError:
             return default
 
     def flags(self, key, default=''):
         try:
             return self._lm[key][1]
         except KeyError:
             return default
 
     def copy(self):
         c = manifestdict()
         c._lm = self._lm.copy()
         return c
 
     def items(self):
         return (x[:2] for x in self._lm.iterentries())
 
     def iteritems(self):
         return (x[:2] for x in self._lm.iterentries())
 
     def iterentries(self):
         return self._lm.iterentries()
 
     def text(self):
         # most likely uses native version
         return self._lm.text()
 
     def fastdelta(self, base, changes):
         """Given a base manifest text as a bytearray and a list of changes
         relative to that text, compute a delta that can be used by revlog.
         """
         delta = []
         dstart = None
         dend = None
         dline = [""]
         start = 0
         # zero copy representation of base as a buffer
         addbuf = util.buffer(base)
 
         changes = list(changes)
         if len(changes) < 1000:
             # start with a readonly loop that finds the offset of
             # each line and creates the deltas
             for f, todelete in changes:
                 # bs will either be the index of the item or the insert point
                 start, end = _msearch(addbuf, f, start)
                 if not todelete:
                     h, fl = self._lm[f]
                     l = "%s\0%s%s\n" % (f, hex(h), fl)
                 else:
                     if start == end:
                         # item we want to delete was not found, error out
                         raise AssertionError(
                                 _("failed to remove %s from manifest") % f)
                     l = ""
                 if dstart is not None and dstart <= start and dend >= start:
                     if dend < end:
                         dend = end
                     if l:
                         dline.append(l)
                 else:
                     if dstart is not None:
                         delta.append([dstart, dend, "".join(dline)])
                     dstart = start
                     dend = end
                     dline = [l]
 
             if dstart is not None:
                 delta.append([dstart, dend, "".join(dline)])
             # apply the delta to the base, and get a delta for addrevision
             deltatext, arraytext = _addlistdelta(base, delta)
         else:
             # For large changes, it's much cheaper to just build the text and
             # diff it.
             arraytext = bytearray(self.text())
             deltatext = mdiff.textdiff(
                 util.buffer(base), util.buffer(arraytext))
 
         return arraytext, deltatext
 
 def _msearch(m, s, lo=0, hi=None):
     '''return a tuple (start, end) that says where to find s within m.
 
     If the string is found m[start:end] are the line containing
     that string.  If start == end the string was not found and
     they indicate the proper sorted insertion point.
 
     m should be a buffer, a memoryview or a byte string.
     s is a byte string'''
     def advance(i, c):
         while i < lenm and m[i:i + 1] != c:
             i += 1
         return i
     if not s:
         return (lo, lo)
     lenm = len(m)
     if not hi:
         hi = lenm
     while lo < hi:
         mid = (lo + hi) // 2
         start = mid
         while start > 0 and m[start - 1:start] != '\n':
             start -= 1
         end = advance(start, '\0')
         if bytes(m[start:end]) < s:
             # we know that after the null there are 40 bytes of sha1
             # this translates to the bisect lo = mid + 1
             lo = advance(end + 40, '\n') + 1
         else:
             # this translates to the bisect hi = mid
             hi = start
     end = advance(lo, '\0')
     found = m[lo:end]
     if s == found:
         # we know that after the null there are 40 bytes of sha1
         end = advance(end + 40, '\n')
         return (lo, end + 1)
     else:
         return (lo, lo)
 
 def _checkforbidden(l):
     """Check filenames for illegal characters."""
     for f in l:
         if '\n' in f or '\r' in f:
             raise error.RevlogError(
                 _("'\\n' and '\\r' disallowed in filenames: %r")
                 % pycompat.bytestr(f))
 
 
 # apply the changes collected during the bisect loop to our addlist
 # return a delta suitable for addrevision
 def _addlistdelta(addlist, x):
     # for large addlist arrays, building a new array is cheaper
     # than repeatedly modifying the existing one
     currentposition = 0
     newaddlist = bytearray()
 
     for start, end, content in x:
         newaddlist += addlist[currentposition:start]
         if content:
             newaddlist += bytearray(content)
 
         currentposition = end
 
     newaddlist += addlist[currentposition:]
 
     deltatext = "".join(struct.pack(">lll", start, end, len(content))
                    + content for start, end, content in x)
     return deltatext, newaddlist
 
 def _splittopdir(f):
     if '/' in f:
         dir, subpath = f.split('/', 1)
         return dir + '/', subpath
     else:
         return '', f
 
 _noop = lambda s: None
 
 class treemanifest(object):
     def __init__(self, dir='', text=''):
         self._dir = dir
         self._node = nullid
         self._loadfunc = _noop
         self._copyfunc = _noop
         self._dirty = False
         self._dirs = {}
         self._lazydirs = {}
         # Using _lazymanifest here is a little slower than plain old dicts
         self._files = {}
         self._flags = {}
         if text:
             def readsubtree(subdir, subm):
                 raise AssertionError('treemanifest constructor only accepts '
                                      'flat manifests')
             self.parse(text, readsubtree)
             self._dirty = True # Mark flat manifest dirty after parsing
 
     def _subpath(self, path):
         return self._dir + path
 
     def _loadalllazy(self):
         for k, (path, node, readsubtree) in self._lazydirs.iteritems():
             self._dirs[k] = readsubtree(path, node)
         self._lazydirs = {}
 
     def _loadlazy(self, d):
         path, node, readsubtree = self._lazydirs[d]
         self._dirs[d] = readsubtree(path, node)
         del self._lazydirs[d]
 
     def _loadchildrensetlazy(self, visit):
         if not visit:
             return None
         if visit == 'all' or visit == 'this':
             self._loadalllazy()
             return None
 
         todel = []
         for k in visit:
             kslash = k + '/'
             ld = self._lazydirs.get(kslash)
             if ld:
                 path, node, readsubtree = ld
                 self._dirs[kslash] = readsubtree(path, node)
                 todel.append(kslash)
         for kslash in todel:
             del self._lazydirs[kslash]
         return visit
 
     def __len__(self):
         self._load()
         size = len(self._files)
         self._loadalllazy()
         for m in self._dirs.values():
             size += m.__len__()
         return size
 
     def __nonzero__(self):
         # Faster than "__len() != 0" since it avoids loading sub-manifests
         return not self._isempty()
 
     __bool__ = __nonzero__
 
     def _isempty(self):
         self._load() # for consistency; already loaded by all callers
         # See if we can skip loading everything.
         if self._files or (self._dirs and
                            any(not m._isempty() for m in self._dirs.values())):
             return False
         self._loadalllazy()
         return (not self._dirs or
                 all(m._isempty() for m in self._dirs.values()))
 
     def __repr__(self):
         return ('<treemanifest dir=%s, node=%s, loaded=%s, dirty=%s at 0x%x>' %
                 (self._dir, hex(self._node),
                  bool(self._loadfunc is _noop),
                  self._dirty, id(self)))
 
     def dir(self):
         '''The directory that this tree manifest represents, including a
         trailing '/'. Empty string for the repo root directory.'''
         return self._dir
 
     def node(self):
         '''This node of this instance. nullid for unsaved instances. Should
         be updated when the instance is read or written from a revlog.
         '''
         assert not self._dirty
         return self._node
 
     def setnode(self, node):
         self._node = node
         self._dirty = False
 
     def iterentries(self):
         self._load()
         self._loadalllazy()
         for p, n in sorted(itertools.chain(self._dirs.items(),
                                            self._files.items())):
             if p in self._files:
                 yield self._subpath(p), n, self._flags.get(p, '')
             else:
                 for x in n.iterentries():
                     yield x
 
     def items(self):
         self._load()
         self._loadalllazy()
         for p, n in sorted(itertools.chain(self._dirs.items(),
                                            self._files.items())):
             if p in self._files:
                 yield self._subpath(p), n
             else:
                 for f, sn in n.iteritems():
                     yield f, sn
 
     iteritems = items
 
     def iterkeys(self):
         self._load()
         self._loadalllazy()
         for p in sorted(itertools.chain(self._dirs, self._files)):
             if p in self._files:
                 yield self._subpath(p)
             else:
                 for f in self._dirs[p]:
                     yield f
 
     def keys(self):
         return list(self.iterkeys())
 
     def __iter__(self):
         return self.iterkeys()
 
     def __contains__(self, f):
         if f is None:
             return False
         self._load()
         dir, subpath = _splittopdir(f)
         if dir:
             if dir in self._lazydirs:
                 self._loadlazy(dir)
 
             if dir not in self._dirs:
                 return False
 
             return self._dirs[dir].__contains__(subpath)
         else:
             return f in self._files
 
     def get(self, f, default=None):
         self._load()
         dir, subpath = _splittopdir(f)
         if dir:
             if dir in self._lazydirs:
                 self._loadlazy(dir)
 
             if dir not in self._dirs:
                 return default
             return self._dirs[dir].get(subpath, default)
         else:
             return self._files.get(f, default)
 
     def __getitem__(self, f):
         self._load()
         dir, subpath = _splittopdir(f)
         if dir:
             if dir in self._lazydirs:
                 self._loadlazy(dir)
 
             return self._dirs[dir].__getitem__(subpath)
         else:
             return self._files[f]
 
     def flags(self, f):
         self._load()
         dir, subpath = _splittopdir(f)
         if dir:
             if dir in self._lazydirs:
                 self._loadlazy(dir)
 
             if dir not in self._dirs:
                 return ''
             return self._dirs[dir].flags(subpath)
         else:
             if f in self._lazydirs or f in self._dirs:
                 return ''
             return self._flags.get(f, '')
 
     def find(self, f):
         self._load()
         dir, subpath = _splittopdir(f)
         if dir:
             if dir in self._lazydirs:
                 self._loadlazy(dir)
 
             return self._dirs[dir].find(subpath)
         else:
             return self._files[f], self._flags.get(f, '')
 
     def __delitem__(self, f):
         self._load()
         dir, subpath = _splittopdir(f)
         if dir:
             if dir in self._lazydirs:
                 self._loadlazy(dir)
 
             self._dirs[dir].__delitem__(subpath)
             # If the directory is now empty, remove it
             if self._dirs[dir]._isempty():
                 del self._dirs[dir]
         else:
             del self._files[f]
             if f in self._flags:
                 del self._flags[f]
         self._dirty = True
 
     def __setitem__(self, f, n):
         assert n is not None
         self._load()
         dir, subpath = _splittopdir(f)
         if dir:
             if dir in self._lazydirs:
                 self._loadlazy(dir)
             if dir not in self._dirs:
                 self._dirs[dir] = treemanifest(self._subpath(dir))
             self._dirs[dir].__setitem__(subpath, n)
         else:
             self._files[f] = n[:21] # to match manifestdict's behavior
         self._dirty = True
 
     def _load(self):
         if self._loadfunc is not _noop:
             lf, self._loadfunc = self._loadfunc, _noop
             lf(self)
         elif self._copyfunc is not _noop:
             cf, self._copyfunc = self._copyfunc, _noop
             cf(self)
 
     def setflag(self, f, flags):
         """Set the flags (symlink, executable) for path f."""
         self._load()
         dir, subpath = _splittopdir(f)
         if dir:
             if dir in self._lazydirs:
                 self._loadlazy(dir)
             if dir not in self._dirs:
                 self._dirs[dir] = treemanifest(self._subpath(dir))
             self._dirs[dir].setflag(subpath, flags)
         else:
             self._flags[f] = flags
         self._dirty = True
 
     def copy(self):
         copy = treemanifest(self._dir)
         copy._node = self._node
         copy._dirty = self._dirty
         if self._copyfunc is _noop:
             def _copyfunc(s):
                 self._load()
                 # OPT: it'd be nice to not load everything here. Unfortunately
                 # this makes a mess of the "dirty" state tracking if we don't.
                 self._loadalllazy()
                 sdirs = s._dirs
                 for d, v in self._dirs.iteritems():
                     sdirs[d] = v.copy()
                 s._files = dict.copy(self._files)
                 s._flags = dict.copy(self._flags)
             if self._loadfunc is _noop:
                 _copyfunc(copy)
             else:
                 copy._copyfunc = _copyfunc
         else:
             copy._copyfunc = self._copyfunc
         return copy
 
     def filesnotin(self, m2, match=None):
         '''Set of files in this manifest that are not in the other'''
         if match and not match.always():
             m1 = self.matches(match)
             m2 = m2.matches(match)
             return m1.filesnotin(m2)
 
         files = set()
         def _filesnotin(t1, t2):
             if t1._node == t2._node and not t1._dirty and not t2._dirty:
                 return
             t1._load()
             t2._load()
             t1._loadalllazy()
             t2._loadalllazy()
             for d, m1 in t1._dirs.iteritems():
                 if d in t2._dirs:
                     m2 = t2._dirs[d]
                     _filesnotin(m1, m2)
                 else:
                     files.update(m1.iterkeys())
 
             for fn in t1._files:
                 if fn not in t2._files:
                     files.add(t1._subpath(fn))
 
         _filesnotin(self, m2)
         return files
 
     @propertycache
     def _alldirs(self):
         return util.dirs(self)
 
     def dirs(self):
         return self._alldirs
 
     def hasdir(self, dir):
         self._load()
         topdir, subdir = _splittopdir(dir)
         if topdir:
             if topdir in self._lazydirs:
                 self._loadlazy(topdir)
             if topdir in self._dirs:
                 return self._dirs[topdir].hasdir(subdir)
             return False
         dirslash = dir + '/'
         return dirslash in self._dirs or dirslash in self._lazydirs
 
     def walk(self, match):
         '''Generates matching file names.
 
         Equivalent to manifest.matches(match).iterkeys(), but without creating
         an entirely new manifest.
 
         It also reports nonexistent files by marking them bad with match.bad().
         '''
         if match.always():
             for f in iter(self):
                 yield f
             return
 
         fset = set(match.files())
 
         for fn in self._walk(match):
             if fn in fset:
                 # specified pattern is the exact name
                 fset.remove(fn)
             yield fn
 
         # for dirstate.walk, files=['.'] means "walk the whole tree".
         # follow that here, too
         fset.discard('.')
 
         for fn in sorted(fset):
             if not self.hasdir(fn):
                 match.bad(fn, None)
 
     def _walk(self, match):
         '''Recursively generates matching file names for walk().'''
         visit = match.visitchildrenset(self._dir[:-1] or '.')
         if not visit:
             return
 
         # yield this dir's files and walk its submanifests
         self._load()
         visit = self._loadchildrensetlazy(visit)
         for p in sorted(list(self._dirs) + list(self._files)):
             if p in self._files:
                 fullp = self._subpath(p)
                 if match(fullp):
                     yield fullp
             else:
                 if not visit or p[:-1] in visit:
                     for f in self._dirs[p]._walk(match):
                         yield f
 
     def matches(self, match):
         '''generate a new manifest filtered by the match argument'''
         if match.always():
             return self.copy()
 
         return self._matches(match)
 
     def _matches(self, match):
         '''recursively generate a new manifest filtered by the match argument.
         '''
 
         visit = match.visitchildrenset(self._dir[:-1] or '.')
         if visit == 'all':
             return self.copy()
         ret = treemanifest(self._dir)
         if not visit:
             return ret
 
         self._load()
         for fn in self._files:
             # While visitchildrenset *usually* lists only subdirs, this is
             # actually up to the matcher and may have some files in the set().
             # If visit == 'this', we should obviously look at the files in this
             # directory; if visit is a set, and fn is in it, we should inspect
             # fn (but no need to inspect things not in the set).
             if visit != 'this' and fn not in visit:
                 continue
             fullp = self._subpath(fn)
             # visitchildrenset isn't perfect, we still need to call the regular
             # matcher code to further filter results.
             if not match(fullp):
                 continue
             ret._files[fn] = self._files[fn]
             if fn in self._flags:
                 ret._flags[fn] = self._flags[fn]
 
         visit = self._loadchildrensetlazy(visit)
         for dir, subm in self._dirs.iteritems():
             if visit and dir[:-1] not in visit:
                 continue
             m = subm._matches(match)
             if not m._isempty():
                 ret._dirs[dir] = m
 
         if not ret._isempty():
             ret._dirty = True
         return ret
 
     def diff(self, m2, match=None, clean=False):
         '''Finds changes between the current manifest and m2.
 
         Args:
           m2: the manifest to which this manifest should be compared.
           clean: if true, include files unchanged between these manifests
                  with a None value in the returned dictionary.
 
         The result is returned as a dict with filename as key and
         values of the form ((n1,fl1),(n2,fl2)), where n1/n2 is the
         nodeid in the current/other manifest and fl1/fl2 is the flag
         in the current/other manifest. Where the file does not exist,
         the nodeid will be None and the flags will be the empty
         string.
         '''
         if match and not match.always():
             m1 = self.matches(match)
             m2 = m2.matches(match)
             return m1.diff(m2, clean=clean)
         result = {}
         emptytree = treemanifest()
         def _diff(t1, t2):
             if t1._node == t2._node and not t1._dirty and not t2._dirty:
                 return
             t1._load()
             t2._load()
             # OPT: do we need to load everything?
             t1._loadalllazy()
             t2._loadalllazy()
             for d, m1 in t1._dirs.iteritems():
                 m2 = t2._dirs.get(d, emptytree)
                 _diff(m1, m2)
 
             for d, m2 in t2._dirs.iteritems():
                 if d not in t1._dirs:
                     _diff(emptytree, m2)
 
             for fn, n1 in t1._files.iteritems():
                 fl1 = t1._flags.get(fn, '')
                 n2 = t2._files.get(fn, None)
                 fl2 = t2._flags.get(fn, '')
                 if n1 != n2 or fl1 != fl2:
                     result[t1._subpath(fn)] = ((n1, fl1), (n2, fl2))
                 elif clean:
                     result[t1._subpath(fn)] = None
 
             for fn, n2 in t2._files.iteritems():
                 if fn not in t1._files:
                     fl2 = t2._flags.get(fn, '')
                     result[t2._subpath(fn)] = ((None, ''), (n2, fl2))
 
         _diff(self, m2)
         return result
 
     def unmodifiedsince(self, m2):
         return not self._dirty and not m2._dirty and self._node == m2._node
 
     def parse(self, text, readsubtree):
         selflazy = self._lazydirs
         subpath = self._subpath
         for f, n, fl in _parse(text):
             if fl == 't':
                 f = f + '/'
                 selflazy[f] = (subpath(f), n, readsubtree)
             elif '/' in f:
                 # This is a flat manifest, so use __setitem__ and setflag rather
                 # than assigning directly to _files and _flags, so we can
                 # assign a path in a subdirectory, and to mark dirty (compared
                 # to nullid).
                 self[f] = n
                 if fl:
                     self.setflag(f, fl)
             else:
                 # Assigning to _files and _flags avoids marking as dirty,
                 # and should be a little faster.
                 self._files[f] = n
                 if fl:
                     self._flags[f] = fl
 
     def text(self):
         """Get the full data of this manifest as a bytestring."""
         self._load()
         return _text(self.iterentries())
 
     def dirtext(self):
         """Get the full data of this directory as a bytestring. Make sure that
         any submanifests have been written first, so their nodeids are correct.
         """
         self._load()
         flags = self.flags
         lazydirs = [(d[:-1], node, 't') for
                     d, (path, node, readsubtree) in self._lazydirs.iteritems()]
         dirs = [(d[:-1], self._dirs[d]._node, 't') for d in self._dirs]
         files = [(f, self._files[f], flags(f)) for f in self._files]
         return _text(sorted(dirs + files + lazydirs))
 
     def read(self, gettext, readsubtree):
         def _load_for_read(s):
             s.parse(gettext(), readsubtree)
             s._dirty = False
         self._loadfunc = _load_for_read
 
     def writesubtrees(self, m1, m2, writesubtree, match):
         self._load() # for consistency; should never have any effect here
         m1._load()
         m2._load()
         emptytree = treemanifest()
         def getnode(m, d):
             ld = m._lazydirs.get(d)
             if ld:
                 return ld[1]
             return m._dirs.get(d, emptytree)._node
 
         # we should have always loaded everything by the time we get here for
         # `self`, but possibly not in `m1` or `m2`.
         assert not self._lazydirs
         # let's skip investigating things that `match` says we do not need.
         visit = match.visitchildrenset(self._dir[:-1] or '.')
         if visit == 'this' or visit == 'all':
             visit = None
         for d, subm in self._dirs.iteritems():
             if visit and d[:-1] not in visit:
                 continue
             subp1 = getnode(m1, d)
             subp2 = getnode(m2, d)
             if subp1 == nullid:
                 subp1, subp2 = subp2, subp1
             writesubtree(subm, subp1, subp2, match)
 
     def walksubtrees(self, matcher=None):
         """Returns an iterator of the subtrees of this manifest, including this
         manifest itself.
 
         If `matcher` is provided, it only returns subtrees that match.
         """
         if matcher and not matcher.visitdir(self._dir[:-1] or '.'):
             return
         if not matcher or matcher(self._dir[:-1]):
             yield self
 
         self._load()
         # OPT: use visitchildrenset to avoid loading everything.
         self._loadalllazy()
         for d, subm in self._dirs.iteritems():
             for subtree in subm.walksubtrees(matcher=matcher):
                 yield subtree
 
 class manifestfulltextcache(util.lrucachedict):
     """File-backed LRU cache for the manifest cache
 
     File consists of entries, up to EOF:
 
     - 20 bytes node, 4 bytes length, <length> manifest data
 
     These are written in reverse cache order (oldest to newest).
 
     """
     def __init__(self, max):
         super(manifestfulltextcache, self).__init__(max)
         self._dirty = False
         self._read = False
         self._opener = None
 
     def read(self):
         if self._read or self._opener is None:
             return
 
         try:
             with self._opener('manifestfulltextcache') as fp:
                 set = super(manifestfulltextcache, self).__setitem__
                 # ignore trailing data, this is a cache, corruption is skipped
                 while True:
                     node = fp.read(20)
                     if len(node) < 20:
                         break
                     try:
                         size = struct.unpack('>L', fp.read(4))[0]
                     except struct.error:
                         break
                     value = bytearray(fp.read(size))
                     if len(value) != size:
                         break
                     set(node, value)
         except IOError:
             # the file is allowed to be missing
             pass
 
         self._read = True
         self._dirty = False
 
     def write(self):
         if not self._dirty or self._opener is None:
             return
         # rotate backwards to the first used node
         with self._opener(
                 'manifestfulltextcache', 'w', atomictemp=True, checkambig=True
             ) as fp:
             node = self._head.prev
             while True:
                 if node.key in self._cache:
                     fp.write(node.key)
                     fp.write(struct.pack('>L', len(node.value)))
                     fp.write(node.value)
                 if node is self._head:
                     break
                 node = node.prev
 
     def __len__(self):
         if not self._read:
             self.read()
         return super(manifestfulltextcache, self).__len__()
 
     def __contains__(self, k):
         if not self._read:
             self.read()
         return super(manifestfulltextcache, self).__contains__(k)
 
     def __iter__(self):
         if not self._read:
             self.read()
         return super(manifestfulltextcache, self).__iter__()
 
     def __getitem__(self, k):
         if not self._read:
             self.read()
         # the cache lru order can change on read
         setdirty = self._cache.get(k) is not self._head
         value = super(manifestfulltextcache, self).__getitem__(k)
         if setdirty:
             self._dirty = True
         return value
 
     def __setitem__(self, k, v):
         if not self._read:
             self.read()
         super(manifestfulltextcache, self).__setitem__(k, v)
         self._dirty = True
 
     def __delitem__(self, k):
         if not self._read:
             self.read()
         super(manifestfulltextcache, self).__delitem__(k)
         self._dirty = True
 
     def get(self, k, default=None):
         if not self._read:
             self.read()
         return super(manifestfulltextcache, self).get(k, default=default)
 
     def clear(self, clear_persisted_data=False):
         super(manifestfulltextcache, self).clear()
         if clear_persisted_data:
             self._dirty = True
             self.write()
         self._read = False
 
 @interfaceutil.implementer(repository.imanifeststorage)
 class manifestrevlog(object):
     '''A revlog that stores manifest texts. This is responsible for caching the
     full-text manifest contents.
     '''
     def __init__(self, opener, tree='', dirlogcache=None, indexfile=None,
                  treemanifest=False):
         """Constructs a new manifest revlog
 
         `indexfile` - used by extensions to have two manifests at once, like
         when transitioning between flatmanifeset and treemanifests.
 
         `treemanifest` - used to indicate this is a tree manifest revlog. Opener
         options can also be used to make this a tree manifest revlog. The opener
         option takes precedence, so if it is set to True, we ignore whatever
         value is passed in to the constructor.
         """
         # During normal operations, we expect to deal with not more than four
         # revs at a time (such as during commit --amend). When rebasing large
         # stacks of commits, the number can go up, hence the config knob below.
         cachesize = 4
         optiontreemanifest = False
         opts = getattr(opener, 'options', None)
         if opts is not None:
             cachesize = opts.get('manifestcachesize', cachesize)
             optiontreemanifest = opts.get('treemanifest', False)
 
         self._treeondisk = optiontreemanifest or treemanifest
 
         self._fulltextcache = manifestfulltextcache(cachesize)
 
         if tree:
             assert self._treeondisk, 'opts is %r' % opts
 
         if indexfile is None:
             indexfile = '00manifest.i'
             if tree:
                 indexfile = "meta/" + tree + indexfile
 
         self.tree = tree
 
         # The dirlogcache is kept on the root manifest log
         if tree:
             self._dirlogcache = dirlogcache
         else:
             self._dirlogcache = {'': self}
 
         self._revlog = revlog.revlog(opener, indexfile,
                                      # only root indexfile is cached
                                      checkambig=not bool(tree),
                                      mmaplargeindex=True)
 
         self.index = self._revlog.index
         self.version = self._revlog.version
         self._generaldelta = self._revlog._generaldelta
 
     def _setupmanifestcachehooks(self, repo):
         """Persist the manifestfulltextcache on lock release"""
         if not util.safehasattr(repo, '_lockref'):
             return
 
         self._fulltextcache._opener = repo.cachevfs
         reporef = weakref.ref(repo)
         manifestrevlogref = weakref.ref(self)
 
         def persistmanifestcache():
             repo = reporef()
             self = manifestrevlogref()
             if repo is None or self is None:
                 return
             if repo.manifestlog.getstorage(b'') is not self:
                 # there's a different manifest in play now, abort
                 return
             self._fulltextcache.write()
 
         if repo._currentlock(repo._lockref) is not None:
             repo._afterlock(persistmanifestcache)
 
     @property
     def fulltextcache(self):
         return self._fulltextcache
 
     def clearcaches(self, clear_persisted_data=False):
         self._revlog.clearcaches()
         self._fulltextcache.clear(clear_persisted_data=clear_persisted_data)
         self._dirlogcache = {self.tree: self}
 
     def dirlog(self, d):
         if d:
             assert self._treeondisk
         if d not in self._dirlogcache:
             mfrevlog = manifestrevlog(self.opener, d,
                                       self._dirlogcache,
                                       treemanifest=self._treeondisk)
             self._dirlogcache[d] = mfrevlog
         return self._dirlogcache[d]
 
     def add(self, m, transaction, link, p1, p2, added, removed, readtree=None,
             match=None):
         if p1 in self.fulltextcache and util.safehasattr(m, 'fastdelta'):
             # If our first parent is in the manifest cache, we can
             # compute a delta here using properties we know about the
             # manifest up-front, which may save time later for the
             # revlog layer.
 
             _checkforbidden(added)
             # combine the changed lists into one sorted iterator
             work = heapq.merge([(x, False) for x in added],
                                [(x, True) for x in removed])
 
             arraytext, deltatext = m.fastdelta(self.fulltextcache[p1], work)
             cachedelta = self._revlog.rev(p1), deltatext
             text = util.buffer(arraytext)
             n = self._revlog.addrevision(text, transaction, link, p1, p2,
                                          cachedelta)
         else:
             # The first parent manifest isn't already loaded, so we'll
             # just encode a fulltext of the manifest and pass that
             # through to the revlog layer, and let it handle the delta
             # process.
             if self._treeondisk:
                 assert readtree, "readtree must be set for treemanifest writes"
                 assert match, "match must be specified for treemanifest writes"
                 m1 = readtree(self.tree, p1)
                 m2 = readtree(self.tree, p2)
                 n = self._addtree(m, transaction, link, m1, m2, readtree,
                                   match=match)
                 arraytext = None
             else:
                 text = m.text()
                 n = self._revlog.addrevision(text, transaction, link, p1, p2)
                 arraytext = bytearray(text)
 
         if arraytext is not None:
             self.fulltextcache[n] = arraytext
 
         return n
 
     def _addtree(self, m, transaction, link, m1, m2, readtree, match):
         # If the manifest is unchanged compared to one parent,
         # don't write a new revision
         if self.tree != '' and (m.unmodifiedsince(m1) or m.unmodifiedsince(
             m2)):
             return m.node()
         def writesubtree(subm, subp1, subp2, match):
             sublog = self.dirlog(subm.dir())
             sublog.add(subm, transaction, link, subp1, subp2, None, None,
                        readtree=readtree, match=match)
         m.writesubtrees(m1, m2, writesubtree, match)
         text = m.dirtext()
         n = None
         if self.tree != '':
             # Double-check whether contents are unchanged to one parent
             if text == m1.dirtext():
                 n = m1.node()
             elif text == m2.dirtext():
                 n = m2.node()
 
         if not n:
             n = self._revlog.addrevision(text, transaction, link, m1.node(),
                                          m2.node())
 
         # Save nodeid so parent manifest can calculate its nodeid
         m.setnode(n)
         return n
 
     def __len__(self):
         return len(self._revlog)
 
     def __iter__(self):
         return self._revlog.__iter__()
 
     def rev(self, node):
         return self._revlog.rev(node)
 
     def node(self, rev):
         return self._revlog.node(rev)
 
     def lookup(self, value):
         return self._revlog.lookup(value)
 
     def parentrevs(self, rev):
         return self._revlog.parentrevs(rev)
 
     def parents(self, node):
         return self._revlog.parents(node)
 
     def linkrev(self, rev):
         return self._revlog.linkrev(rev)
 
     def checksize(self):
         return self._revlog.checksize()
 
     def revision(self, node, _df=None, raw=False):
         return self._revlog.revision(node, _df=_df, raw=raw)
 
     def revdiff(self, rev1, rev2):
         return self._revlog.revdiff(rev1, rev2)
 
     def cmp(self, node, text):
         return self._revlog.cmp(node, text)
 
     def deltaparent(self, rev):
         return self._revlog.deltaparent(rev)
 
     def emitrevisiondeltas(self, requests):
         return self._revlog.emitrevisiondeltas(requests)
 
     def addgroup(self, deltas, linkmapper, transaction, addrevisioncb=None):
         return self._revlog.addgroup(deltas, linkmapper, transaction,
                                      addrevisioncb=addrevisioncb)
 
     def getstrippoint(self, minlink):
         return self._revlog.getstrippoint(minlink)
 
     def strip(self, minlink, transaction):
         return self._revlog.strip(minlink, transaction)
 
     def files(self):
         return self._revlog.files()
 
     def clone(self, tr, destrevlog, **kwargs):
         if not isinstance(destrevlog, manifestrevlog):
             raise error.ProgrammingError('expected manifestrevlog to clone()')
 
         return self._revlog.clone(tr, destrevlog._revlog, **kwargs)
 
     @property
     def indexfile(self):
         return self._revlog.indexfile
 
     @indexfile.setter
     def indexfile(self, value):
         self._revlog.indexfile = value
 
     @property
     def opener(self):
         return self._revlog.opener
 
     @opener.setter
     def opener(self, value):
         self._revlog.opener = value
 
 @interfaceutil.implementer(repository.imanifestlog)
 class manifestlog(object):
     """A collection class representing the collection of manifest snapshots
     referenced by commits in the repository.
 
     In this situation, 'manifest' refers to the abstract concept of a snapshot
     of the list of files in the given commit. Consumers of the output of this
     class do not care about the implementation details of the actual manifests
     they receive (i.e. tree or flat or lazily loaded, etc)."""
-    def __init__(self, opener, repo):
+    def __init__(self, opener, repo, rootstore):
         usetreemanifest = False
         cachesize = 4
 
         opts = getattr(opener, 'options', None)
         if opts is not None:
             usetreemanifest = opts.get('treemanifest', usetreemanifest)
             cachesize = opts.get('manifestcachesize', cachesize)
 
         self._treemanifests = usetreemanifest
 
-        self._rootstore = repo._constructmanifest()
+        self._rootstore = rootstore
         self._rootstore._setupmanifestcachehooks(repo)
         self._narrowmatch = repo.narrowmatch()
 
         # A cache of the manifestctx or treemanifestctx for each directory
         self._dirmancache = {}
         self._dirmancache[''] = util.lrucachedict(cachesize)
 
         self._cachesize = cachesize
 
     def __getitem__(self, node):
         """Retrieves the manifest instance for the given node. Throws a
         LookupError if not found.
         """
         return self.get('', node)
 
     def get(self, tree, node, verify=True):
         """Retrieves the manifest instance for the given node. Throws a
         LookupError if not found.
 
         `verify` - if True an exception will be thrown if the node is not in
                    the revlog
         """
         if node in self._dirmancache.get(tree, ()):
             return self._dirmancache[tree][node]
 
         if not self._narrowmatch.always():
             if not self._narrowmatch.visitdir(tree[:-1] or '.'):
                 return excludeddirmanifestctx(tree, node)
         if tree:
             if self._rootstore._treeondisk:
                 if verify:
                     # Side-effect is LookupError is raised if node doesn't
                     # exist.
                     self.getstorage(tree).rev(node)
 
                 m = treemanifestctx(self, tree, node)
             else:
                 raise error.Abort(
                         _("cannot ask for manifest directory '%s' in a flat "
                           "manifest") % tree)
         else:
             if verify:
                 # Side-effect is LookupError is raised if node doesn't exist.
                 self._rootstore.rev(node)
 
             if self._treemanifests:
                 m = treemanifestctx(self, '', node)
             else:
                 m = manifestctx(self, node)
 
         if node != nullid:
             mancache = self._dirmancache.get(tree)
             if not mancache:
                 mancache = util.lrucachedict(self._cachesize)
                 self._dirmancache[tree] = mancache
             mancache[node] = m
         return m
 
     def getstorage(self, tree):
         return self._rootstore.dirlog(tree)
 
     def clearcaches(self, clear_persisted_data=False):
         self._dirmancache.clear()
         self._rootstore.clearcaches(clear_persisted_data=clear_persisted_data)
 
     def rev(self, node):
         return self._rootstore.rev(node)
 
 @interfaceutil.implementer(repository.imanifestrevisionwritable)
 class memmanifestctx(object):
     def __init__(self, manifestlog):
         self._manifestlog = manifestlog
         self._manifestdict = manifestdict()
 
     def _storage(self):
         return self._manifestlog.getstorage(b'')
 
     def new(self):
         return memmanifestctx(self._manifestlog)
 
     def copy(self):
         memmf = memmanifestctx(self._manifestlog)
         memmf._manifestdict = self.read().copy()
         return memmf
 
     def read(self):
         return self._manifestdict
 
     def write(self, transaction, link, p1, p2, added, removed, match=None):
         return self._storage().add(self._manifestdict, transaction, link,
                                    p1, p2, added, removed, match=match)
 
 @interfaceutil.implementer(repository.imanifestrevisionstored)
 class manifestctx(object):
     """A class representing a single revision of a manifest, including its
     contents, its parent revs, and its linkrev.
     """
     def __init__(self, manifestlog, node):
         self._manifestlog = manifestlog
         self._data = None
 
         self._node = node
 
         # TODO: We eventually want p1, p2, and linkrev exposed on this class,
         # but let's add it later when something needs it and we can load it
         # lazily.
         #self.p1, self.p2 = store.parents(node)
         #rev = store.rev(node)
         #self.linkrev = store.linkrev(rev)
 
     def _storage(self):
         return self._manifestlog.getstorage(b'')
 
     def node(self):
         return self._node
 
     def new(self):
         return memmanifestctx(self._manifestlog)
 
     def copy(self):
         memmf = memmanifestctx(self._manifestlog)
         memmf._manifestdict = self.read().copy()
         return memmf
 
     @propertycache
     def parents(self):
         return self._storage().parents(self._node)
 
     def read(self):
         if self._data is None:
             if self._node == nullid:
                 self._data = manifestdict()
             else:
                 store = self._storage()
                 if self._node in store.fulltextcache:
                     text = pycompat.bytestr(store.fulltextcache[self._node])
                 else:
                     text = store.revision(self._node)
                     arraytext = bytearray(text)
                     store.fulltextcache[self._node] = arraytext
                 self._data = manifestdict(text)
         return self._data
 
     def readfast(self, shallow=False):
         '''Calls either readdelta or read, based on which would be less work.
         readdelta is called if the delta is against the p1, and therefore can be
         read quickly.
 
         If `shallow` is True, nothing changes since this is a flat manifest.
         '''
         store = self._storage()
         r = store.rev(self._node)
         deltaparent = store.deltaparent(r)
         if deltaparent != nullrev and deltaparent in store.parentrevs(r):
             return self.readdelta()
         return self.read()
 
     def readdelta(self, shallow=False):
         '''Returns a manifest containing just the entries that are present
         in this manifest, but not in its p1 manifest. This is efficient to read
         if the revlog delta is already p1.
 
         Changing the value of `shallow` has no effect on flat manifests.
         '''
         store = self._storage()
         r = store.rev(self._node)
         d = mdiff.patchtext(store.revdiff(store.deltaparent(r), r))
         return manifestdict(d)
 
     def find(self, key):
         return self.read().find(key)
 
 @interfaceutil.implementer(repository.imanifestrevisionwritable)
 class memtreemanifestctx(object):
     def __init__(self, manifestlog, dir=''):
         self._manifestlog = manifestlog
         self._dir = dir
         self._treemanifest = treemanifest()
 
     def _storage(self):
         return self._manifestlog.getstorage(b'')
 
     def new(self, dir=''):
         return memtreemanifestctx(self._manifestlog, dir=dir)
 
     def copy(self):
         memmf = memtreemanifestctx(self._manifestlog, dir=self._dir)
         memmf._treemanifest = self._treemanifest.copy()
         return memmf
 
     def read(self):
         return self._treemanifest
 
     def write(self, transaction, link, p1, p2, added, removed, match=None):
         def readtree(dir, node):
             return self._manifestlog.get(dir, node).read()
         return self._storage().add(self._treemanifest, transaction, link,
                                    p1, p2, added, removed, readtree=readtree,
                                    match=match)
 
 @interfaceutil.implementer(repository.imanifestrevisionstored)
 class treemanifestctx(object):
     def __init__(self, manifestlog, dir, node):
         self._manifestlog = manifestlog
         self._dir = dir
         self._data = None
 
         self._node = node
 
         # TODO: Load p1/p2/linkrev lazily. They need to be lazily loaded so that
         # we can instantiate treemanifestctx objects for directories we don't
         # have on disk.
         #self.p1, self.p2 = store.parents(node)
         #rev = store.rev(node)
         #self.linkrev = store.linkrev(rev)
 
     def _storage(self):
         narrowmatch = self._manifestlog._narrowmatch
         if not narrowmatch.always():
             if not narrowmatch.visitdir(self._dir[:-1] or '.'):
                 return excludedmanifestrevlog(self._dir)
         return self._manifestlog.getstorage(self._dir)
 
     def read(self):
         if self._data is None:
             store = self._storage()
             if self._node == nullid:
                 self._data = treemanifest()
             # TODO accessing non-public API
             elif store._treeondisk:
                 m = treemanifest(dir=self._dir)
                 def gettext():
                     return store.revision(self._node)
                 def readsubtree(dir, subm):
                     # Set verify to False since we need to be able to create
                     # subtrees for trees that don't exist on disk.
                     return self._manifestlog.get(dir, subm, verify=False).read()
                 m.read(gettext, readsubtree)
                 m.setnode(self._node)
                 self._data = m
             else:
                 if self._node in store.fulltextcache:
                     text = pycompat.bytestr(store.fulltextcache[self._node])
                 else:
                     text = store.revision(self._node)
                     arraytext = bytearray(text)
                     store.fulltextcache[self._node] = arraytext
                 self._data = treemanifest(dir=self._dir, text=text)
 
         return self._data
 
     def node(self):
         return self._node
 
     def new(self, dir=''):
         return memtreemanifestctx(self._manifestlog, dir=dir)
 
     def copy(self):
         memmf = memtreemanifestctx(self._manifestlog, dir=self._dir)
         memmf._treemanifest = self.read().copy()
         return memmf
 
     @propertycache
     def parents(self):
         return self._storage().parents(self._node)
 
     def readdelta(self, shallow=False):
         '''Returns a manifest containing just the entries that are present
         in this manifest, but not in its p1 manifest. This is efficient to read
         if the revlog delta is already p1.
 
         If `shallow` is True, this will read the delta for this directory,
         without recursively reading subdirectory manifests. Instead, any
         subdirectory entry will be reported as it appears in the manifest, i.e.
         the subdirectory will be reported among files and distinguished only by
         its 't' flag.
         '''
         store = self._storage()
         if shallow:
             r = store.rev(self._node)
             d = mdiff.patchtext(store.revdiff(store.deltaparent(r), r))
             return manifestdict(d)
         else:
             # Need to perform a slow delta
             r0 = store.deltaparent(store.rev(self._node))
             m0 = self._manifestlog.get(self._dir, store.node(r0)).read()
             m1 = self.read()
             md = treemanifest(dir=self._dir)
             for f, ((n0, fl0), (n1, fl1)) in m0.diff(m1).iteritems():
                 if n1:
                     md[f] = n1
                     if fl1:
                         md.setflag(f, fl1)
             return md
 
     def readfast(self, shallow=False):
         '''Calls either readdelta or read, based on which would be less work.
         readdelta is called if the delta is against the p1, and therefore can be
         read quickly.
 
         If `shallow` is True, it only returns the entries from this manifest,
         and not any submanifests.
         '''
         store = self._storage()
         r = store.rev(self._node)
         deltaparent = store.deltaparent(r)
         if (deltaparent != nullrev and
             deltaparent in store.parentrevs(r)):
             return self.readdelta(shallow=shallow)
 
         if shallow:
             return manifestdict(store.revision(self._node))
         else:
             return self.read()
 
     def find(self, key):
         return self.read().find(key)
 
 class excludeddir(treemanifest):
     """Stand-in for a directory that is excluded from the repository.
 
     With narrowing active on a repository that uses treemanifests,
     some of the directory revlogs will be excluded from the resulting
     clone. This is a huge storage win for clients, but means we need
     some sort of pseudo-manifest to surface to internals so we can
     detect a merge conflict outside the narrowspec. That's what this
     class is: it stands in for a directory whose node is known, but
     whose contents are unknown.
     """
     def __init__(self, dir, node):
         super(excludeddir, self).__init__(dir)
         self._node = node
         # Add an empty file, which will be included by iterators and such,
         # appearing as the directory itself (i.e. something like "dir/")
         self._files[''] = node
         self._flags[''] = 't'
 
     # Manifests outside the narrowspec should never be modified, so avoid
     # copying. This makes a noticeable difference when there are very many
     # directories outside the narrowspec. Also, it makes sense for the copy to
     # be of the same type as the original, which would not happen with the
     # super type's copy().
     def copy(self):
         return self
 
 class excludeddirmanifestctx(treemanifestctx):
     """context wrapper for excludeddir - see that docstring for rationale"""
     def __init__(self, dir, node):
         self._dir = dir
         self._node = node
 
     def read(self):
         return excludeddir(self._dir, self._node)
 
     def write(self, *args):
         raise error.ProgrammingError(
             'attempt to write manifest from excluded dir %s' % self._dir)
 
 class excludedmanifestrevlog(manifestrevlog):
     """Stand-in for excluded treemanifest revlogs.
 
     When narrowing is active on a treemanifest repository, we'll have
     references to directories we can't see due to the revlog being
     skipped. This class exists to conform to the manifestrevlog
     interface for those directories and proactively prevent writes to
     outside the narrowspec.
     """
 
     def __init__(self, dir):
         self._dir = dir
 
     def __len__(self):
         raise error.ProgrammingError(
             'attempt to get length of excluded dir %s' % self._dir)
 
     def rev(self, node):
         raise error.ProgrammingError(
             'attempt to get rev from excluded dir %s' % self._dir)
 
     def linkrev(self, node):
         raise error.ProgrammingError(
             'attempt to get linkrev from excluded dir %s' % self._dir)
 
     def node(self, rev):
         raise error.ProgrammingError(
             'attempt to get node from excluded dir %s' % self._dir)
 
     def add(self, *args, **kwargs):
         # We should never write entries in dirlogs outside the narrow clone.
         # However, the method still gets called from writesubtree() in
         # _addtree(), so we need to handle it. We should possibly make that
         # avoid calling add() with a clean manifest (_dirty is always False
         # in excludeddir instances).
         pass

mercurial/statichttprepo.py

@@ -1,224 +1,225
 # statichttprepo.py - simple http repository class for mercurial
 #
 # This provides read-only repo access to repositories exported via static http
 #
 # Copyright 2005-2007 Matt Mackall <[e-mail]>
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 from __future__ import absolute_import
 
 import errno
 
 from .i18n import _
 from . import (
     changelog,
     error,
     localrepo,
     manifest,
     namespaces,
     pathutil,
     url,
     util,
     vfs as vfsmod,
 )
 
 urlerr = util.urlerr
 urlreq = util.urlreq
 
 class httprangereader(object):
     def __init__(self, url, opener):
         # we assume opener has HTTPRangeHandler
         self.url = url
         self.pos = 0
         self.opener = opener
         self.name = url
 
     def __enter__(self):
         return self
 
     def __exit__(self, exc_type, exc_value, traceback):
         self.close()
 
     def seek(self, pos):
         self.pos = pos
     def read(self, bytes=None):
         req = urlreq.request(self.url)
         end = ''
         if bytes:
             end = self.pos + bytes - 1
         if self.pos or end:
             req.add_header('Range', 'bytes=%d-%s' % (self.pos, end))
 
         try:
             f = self.opener.open(req)
             data = f.read()
             code = f.code
         except urlerr.httperror as inst:
             num = inst.code == 404 and errno.ENOENT or None
             raise IOError(num, inst)
         except urlerr.urlerror as inst:
             raise IOError(None, inst.reason[1])
 
         if code == 200:
             # HTTPRangeHandler does nothing if remote does not support
             # Range headers and returns the full entity. Let's slice it.
             if bytes:
                 data = data[self.pos:self.pos + bytes]
             else:
                 data = data[self.pos:]
         elif bytes:
             data = data[:bytes]
         self.pos += len(data)
         return data
     def readlines(self):
         return self.read().splitlines(True)
     def __iter__(self):
         return iter(self.readlines())
     def close(self):
         pass
 
 # _RangeError and _HTTPRangeHandler were originally in byterange.py,
 # which was itself extracted from urlgrabber. See the last version of
 # byterange.py from history if you need more information.
 class _RangeError(IOError):
     """Error raised when an unsatisfiable range is requested."""
 
 class _HTTPRangeHandler(urlreq.basehandler):
     """Handler that enables HTTP Range headers.
 
     This was extremely simple. The Range header is a HTTP feature to
     begin with so all this class does is tell urllib2 that the
     "206 Partial Content" response from the HTTP server is what we
     expected.
     """
 
     def http_error_206(self, req, fp, code, msg, hdrs):
         # 206 Partial Content Response
         r = urlreq.addinfourl(fp, hdrs, req.get_full_url())
         r.code = code
         r.msg = msg
         return r
 
     def http_error_416(self, req, fp, code, msg, hdrs):
         # HTTP's Range Not Satisfiable error
         raise _RangeError('Requested Range Not Satisfiable')
 
 def build_opener(ui, authinfo):
     # urllib cannot handle URLs with embedded user or passwd
     urlopener = url.opener(ui, authinfo)
     urlopener.add_handler(_HTTPRangeHandler())
 
     class statichttpvfs(vfsmod.abstractvfs):
         def __init__(self, base):
             self.base = base
 
         def __call__(self, path, mode='r', *args, **kw):
             if mode not in ('r', 'rb'):
                 raise IOError('Permission denied')
             f = "/".join((self.base, urlreq.quote(path)))
             return httprangereader(f, urlopener)
 
         def join(self, path):
             if path:
                 return pathutil.join(self.base, path)
             else:
                 return self.base
 
     return statichttpvfs
 
 class statichttppeer(localrepo.localpeer):
     def local(self):
         return None
     def canpush(self):
         return False
 
 class statichttprepository(localrepo.localrepository):
     supported = localrepo.localrepository._basesupported
 
     def __init__(self, ui, path):
         self._url = path
         self.ui = ui
 
         self.root = path
         u = util.url(path.rstrip('/') + "/.hg")
         self.path, authinfo = u.authinfo()
 
         vfsclass = build_opener(ui, authinfo)
         self.vfs = vfsclass(self.path)
         self.cachevfs = vfsclass(self.vfs.join('cache'))
         self._phasedefaults = []
 
         self.names = namespaces.namespaces()
         self.filtername = None
 
         try:
             requirements = set(self.vfs.read(b'requires').splitlines())
         except IOError as inst:
             if inst.errno != errno.ENOENT:
                 raise
             requirements = set()
 
             # check if it is a non-empty old-style repository
             try:
                 fp = self.vfs("00changelog.i")
                 fp.read(1)
                 fp.close()
             except IOError as inst:
                 if inst.errno != errno.ENOENT:
                     raise
                 # we do not care about empty old-style repositories here
                 msg = _("'%s' does not appear to be an hg repository") % path
                 raise error.RepoError(msg)
 
         supportedrequirements = localrepo.gathersupportedrequirements(ui)
         localrepo.ensurerequirementsrecognized(requirements,
                                                supportedrequirements)
         localrepo.ensurerequirementscompatible(ui, requirements)
 
         # setup store
         self.store = localrepo.makestore(requirements, self.path, vfsclass)
         self.spath = self.store.path
         self.svfs = self.store.opener
         self.sjoin = self.store.join
         self._filecache = {}
         self.requirements = requirements
 
-        self.manifestlog = manifest.manifestlog(self.svfs, self)
+        rootmanifest = manifest.manifestrevlog(self.svfs)
+        self.manifestlog = manifest.manifestlog(self.svfs, self, rootmanifest)
         self.changelog = changelog.changelog(self.svfs)
         self._tags = None
         self.nodetagscache = None
         self._branchcaches = {}
         self._revbranchcache = None
         self.encodepats = None
         self.decodepats = None
         self._transref = None
 
     def _restrictcapabilities(self, caps):
         caps = super(statichttprepository, self)._restrictcapabilities(caps)
         return caps.difference(["pushkey"])
 
     def url(self):
         return self._url
 
     def local(self):
         return False
 
     def peer(self):
         return statichttppeer(self)
 
     def wlock(self, wait=True):
         raise error.LockUnavailable(0, _('lock not available'), 'lock',
                                     _('cannot lock static-http repository'))
 
     def lock(self, wait=True):
         raise error.Abort(_('cannot lock static-http repository'))
 
     def _writecaches(self):
         pass # statichttprepository are read only
 
 def instance(ui, path, create, intents=None, createopts=None):
     if create:
         raise error.Abort(_('cannot create new static-http repository'))
     return statichttprepository(ui, path[7:])

mercurial/unionrepo.py

@@ -1,284 +1,290
 # unionrepo.py - repository class for viewing union of repository changesets
 #
 # Derived from bundlerepo.py
 # Copyright 2006, 2007 Benoit Boissinot <[e-mail]>
 # Copyright 2013 Unity Technologies, Mads Kiilerich <[e-mail]>
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 """Repository class for "in-memory pull" of one local repository to another,
 allowing operations like diff and log with revsets.
 """
 
 from __future__ import absolute_import
 
 from .i18n import _
 from .node import nullid
 
 from . import (
     changelog,
     cmdutil,
     error,
     filelog,
     localrepo,
     manifest,
     mdiff,
     pathutil,
     pycompat,
     revlog,
     util,
     vfs as vfsmod,
 )
 
 class unionrevlog(revlog.revlog):
     def __init__(self, opener, indexfile, revlog2, linkmapper):
         # How it works:
         # To retrieve a revision, we just need to know the node id so we can
         # look it up in revlog2.
         #
         # To differentiate a rev in the second revlog from a rev in the revlog,
         # we check revision against repotiprev.
         opener = vfsmod.readonlyvfs(opener)
         revlog.revlog.__init__(self, opener, indexfile)
         self.revlog2 = revlog2
 
         n = len(self)
         self.repotiprev = n - 1
         self.bundlerevs = set() # used by 'bundle()' revset expression
         for rev2 in self.revlog2:
             rev = self.revlog2.index[rev2]
             # rev numbers - in revlog2, very different from self.rev
             _start, _csize, rsize, base, linkrev, p1rev, p2rev, node = rev
             flags = _start & 0xFFFF
 
             if linkmapper is None: # link is to same revlog
                 assert linkrev == rev2 # we never link back
                 link = n
             else: # rev must be mapped from repo2 cl to unified cl by linkmapper
                 link = linkmapper(linkrev)
 
             if linkmapper is not None: # link is to same revlog
                 base = linkmapper(base)
 
             if node in self.nodemap:
                 # this happens for the common revlog revisions
                 self.bundlerevs.add(self.nodemap[node])
                 continue
 
             p1node = self.revlog2.node(p1rev)
             p2node = self.revlog2.node(p2rev)
 
             # TODO: it's probably wrong to set compressed length to None, but
             # I have no idea if csize is valid in the base revlog context.
             e = (flags, None, rsize, base,
                  link, self.rev(p1node), self.rev(p2node), node)
             self.index.append(e)
             self.nodemap[node] = n
             self.bundlerevs.add(n)
             n += 1
 
     def _chunk(self, rev):
         if rev <= self.repotiprev:
             return revlog.revlog._chunk(self, rev)
         return self.revlog2._chunk(self.node(rev))
 
     def revdiff(self, rev1, rev2):
         """return or calculate a delta between two revisions"""
         if rev1 > self.repotiprev and rev2 > self.repotiprev:
             return self.revlog2.revdiff(
                 self.revlog2.rev(self.node(rev1)),
                 self.revlog2.rev(self.node(rev2)))
         elif rev1 <= self.repotiprev and rev2 <= self.repotiprev:
             return self.baserevdiff(rev1, rev2)
 
         return mdiff.textdiff(self.revision(rev1), self.revision(rev2))
 
     def revision(self, nodeorrev, _df=None, raw=False):
         """return an uncompressed revision of a given node or revision
         number.
         """
         if isinstance(nodeorrev, int):
             rev = nodeorrev
             node = self.node(rev)
         else:
             node = nodeorrev
             rev = self.rev(node)
 
         if node == nullid:
             return ""
 
         if rev > self.repotiprev:
             text = self.revlog2.revision(node)
             self._cache = (node, rev, text)
         else:
             text = self.baserevision(rev)
             # already cached
         return text
 
     def baserevision(self, nodeorrev):
         # Revlog subclasses may override 'revision' method to modify format of
         # content retrieved from revlog. To use unionrevlog with such class one
         # needs to override 'baserevision' and make more specific call here.
         return revlog.revlog.revision(self, nodeorrev)
 
     def baserevdiff(self, rev1, rev2):
         # Exists for the same purpose as baserevision.
         return revlog.revlog.revdiff(self, rev1, rev2)
 
     def addrevision(self, text, transaction, link, p1=None, p2=None, d=None):
         raise NotImplementedError
     def addgroup(self, deltas, transaction, addrevisioncb=None):
         raise NotImplementedError
     def strip(self, rev, minlink):
         raise NotImplementedError
     def checksize(self):
         raise NotImplementedError
 
 class unionchangelog(unionrevlog, changelog.changelog):
     def __init__(self, opener, opener2):
         changelog.changelog.__init__(self, opener)
         linkmapper = None
         changelog2 = changelog.changelog(opener2)
         unionrevlog.__init__(self, opener, self.indexfile, changelog2,
                              linkmapper)
 
     def baserevision(self, nodeorrev):
         # Although changelog doesn't override 'revision' method, some extensions
         # may replace this class with another that does. Same story with
         # manifest and filelog classes.
         return changelog.changelog.revision(self, nodeorrev)
 
     def baserevdiff(self, rev1, rev2):
         return changelog.changelog.revdiff(self, rev1, rev2)
 
 class unionmanifest(unionrevlog, manifest.manifestrevlog):
     def __init__(self, opener, opener2, linkmapper):
         manifest.manifestrevlog.__init__(self, opener)
         manifest2 = manifest.manifestrevlog(opener2)
         unionrevlog.__init__(self, opener, self.indexfile, manifest2,
                              linkmapper)
 
     def baserevision(self, nodeorrev):
         return manifest.manifestrevlog.revision(self, nodeorrev)
 
     def baserevdiff(self, rev1, rev2):
         return manifest.manifestrevlog.revdiff(self, rev1, rev2)
 
 class unionfilelog(filelog.filelog):
     def __init__(self, opener, path, opener2, linkmapper, repo):
         filelog.filelog.__init__(self, opener, path)
         filelog2 = filelog.filelog(opener2, path)
         self._revlog = unionrevlog(opener, self.indexfile,
                                    filelog2._revlog, linkmapper)
         self._repo = repo
         self.repotiprev = self._revlog.repotiprev
         self.revlog2 = self._revlog.revlog2
 
     def baserevision(self, nodeorrev):
         return filelog.filelog.revision(self, nodeorrev)
 
     def baserevdiff(self, rev1, rev2):
         return filelog.filelog.revdiff(self, rev1, rev2)
 
     def iscensored(self, rev):
         """Check if a revision is censored."""
         if rev <= self.repotiprev:
             return filelog.filelog.iscensored(self, rev)
         node = self.node(rev)
         return self.revlog2.iscensored(self.revlog2.rev(node))
 
 class unionpeer(localrepo.localpeer):
     def canpush(self):
         return False
 
 class unionrepository(object):
     """Represents the union of data in 2 repositories.
 
     Instances are not usable if constructed directly. Use ``instance()``
     or ``makeunionrepository()`` to create a usable instance.
     """
     def __init__(self, repo2, url):
         self.repo2 = repo2
         self._url = url
 
         self.ui.setconfig('phases', 'publish', False, 'unionrepo')
 
     @localrepo.unfilteredpropertycache
     def changelog(self):
         return unionchangelog(self.svfs, self.repo2.svfs)
 
+    @localrepo.unfilteredpropertycache
+    def manifestlog(self):
+        rootstore = unionmanifest(self.svfs, self.repo2.svfs,
+                                  self.unfiltered()._clrev)
+        return manifest.manifestlog(self.svfs, self, rootstore)
+
     def _clrev(self, rev2):
         """map from repo2 changelog rev to temporary rev in self.changelog"""
         node = self.repo2.changelog.node(rev2)
         return self.changelog.rev(node)
 
     def _constructmanifest(self):
         return unionmanifest(self.svfs, self.repo2.svfs,
                              self.unfiltered()._clrev)
 
     def url(self):
         return self._url
 
     def file(self, f):
         return unionfilelog(self.svfs, f, self.repo2.svfs,
                             self.unfiltered()._clrev, self)
 
     def close(self):
         self.repo2.close()
 
     def cancopy(self):
         return False
 
     def peer(self):
         return unionpeer(self)
 
     def getcwd(self):
         return pycompat.getcwd() # always outside the repo
 
 def instance(ui, path, create, intents=None, createopts=None):
     if create:
         raise error.Abort(_('cannot create new union repository'))
     parentpath = ui.config("bundle", "mainreporoot")
     if not parentpath:
         # try to find the correct path to the working directory repo
         parentpath = cmdutil.findrepo(pycompat.getcwd())
         if parentpath is None:
             parentpath = ''
     if parentpath:
         # Try to make the full path relative so we get a nice, short URL.
         # In particular, we don't want temp dir names in test outputs.
         cwd = pycompat.getcwd()
         if parentpath == cwd:
             parentpath = ''
         else:
             cwd = pathutil.normasprefix(cwd)
             if parentpath.startswith(cwd):
                 parentpath = parentpath[len(cwd):]
     if path.startswith('union:'):
         s = path.split(":", 1)[1].split("+", 1)
         if len(s) == 1:
             repopath, repopath2 = parentpath, s[0]
         else:
             repopath, repopath2 = s
     else:
         repopath, repopath2 = parentpath, path
 
     return makeunionrepository(ui, repopath, repopath2)
 
 def makeunionrepository(ui, repopath1, repopath2):
     """Make a union repository object from 2 local repo paths."""
     repo1 = localrepo.instance(ui, repopath1, create=False)
     repo2 = localrepo.instance(ui, repopath2, create=False)
 
     url = 'union:%s+%s' % (util.expandpath(repopath1),
                            util.expandpath(repopath2))
 
     class derivedunionrepository(unionrepository, repo1.__class__):
         pass
 
     repo = repo1
     repo.__class__ = derivedunionrepository
     unionrepository.__init__(repo1, repo2, url)
 
     return repo

tests/test-check-interfaces.py

@@ -1,230 +1,230
 # Test that certain objects conform to well-defined interfaces.
 
 from __future__ import absolute_import, print_function
 
 from mercurial import encoding
 encoding.environ[b'HGREALINTERFACES'] = b'1'
 
 import os
 import subprocess
 import sys
 
 # Only run if tests are run in a repo
 if subprocess.call(['python', '%s/hghave' % os.environ['TESTDIR'],
                     'test-repo']):
     sys.exit(80)
 
 from mercurial.thirdparty.zope import (
     interface as zi,
 )
 from mercurial.thirdparty.zope.interface import (
     verify as ziverify,
 )
 from mercurial import (
     changegroup,
     bundlerepo,
     filelog,
     httppeer,
     localrepo,
     manifest,
     pycompat,
     repository,
     revlog,
     sshpeer,
     statichttprepo,
     ui as uimod,
     unionrepo,
     vfs as vfsmod,
     wireprotoserver,
     wireprototypes,
     wireprotov1peer,
     wireprotov2server,
 )
 
 rootdir = pycompat.fsencode(
     os.path.normpath(os.path.join(os.path.dirname(__file__), '..')))
 
 def checkzobject(o, allowextra=False):
     """Verify an object with a zope interface."""
     ifaces = zi.providedBy(o)
     if not ifaces:
         print('%r does not provide any zope interfaces' % o)
         return
 
     # Run zope.interface's built-in verification routine. This verifies that
     # everything that is supposed to be present is present.
     for iface in ifaces:
         ziverify.verifyObject(iface, o)
 
     if allowextra:
         return
 
     # Now verify that the object provides no extra public attributes that
     # aren't declared as part of interfaces.
     allowed = set()
     for iface in ifaces:
         allowed |= set(iface.names(all=True))
 
     public = {a for a in dir(o) if not a.startswith('_')}
 
     for attr in sorted(public - allowed):
         print('public attribute not declared in interfaces: %s.%s' % (
             o.__class__.__name__, attr))
 
 # Facilitates testing localpeer.
 class dummyrepo(object):
     def __init__(self):
         self.ui = uimod.ui()
     def filtered(self, name):
         pass
     def _restrictcapabilities(self, caps):
         pass
 
 class dummyopener(object):
     handlers = []
 
 # Facilitates testing sshpeer without requiring a server.
 class badpeer(httppeer.httppeer):
     def __init__(self):
         super(badpeer, self).__init__(None, None, None, dummyopener(), None,
                                       None)
         self.badattribute = True
 
     def badmethod(self):
         pass
 
 class dummypipe(object):
     def close(self):
         pass
 
 def main():
     ui = uimod.ui()
     # Needed so we can open a local repo with obsstore without a warning.
     ui.setconfig('experimental', 'evolution.createmarkers', True)
 
     checkzobject(badpeer())
 
     ziverify.verifyClass(repository.ipeerbase, httppeer.httppeer)
     checkzobject(httppeer.httppeer(None, None, None, dummyopener(), None, None))
 
     ziverify.verifyClass(repository.ipeerconnection,
                          httppeer.httpv2peer)
     ziverify.verifyClass(repository.ipeercapabilities,
                          httppeer.httpv2peer)
     checkzobject(httppeer.httpv2peer(None, b'', b'', None, None, None))
 
     ziverify.verifyClass(repository.ipeerbase,
                          localrepo.localpeer)
     checkzobject(localrepo.localpeer(dummyrepo()))
 
     ziverify.verifyClass(repository.ipeercommandexecutor,
                          localrepo.localcommandexecutor)
     checkzobject(localrepo.localcommandexecutor(None))
 
     ziverify.verifyClass(repository.ipeercommandexecutor,
                          wireprotov1peer.peerexecutor)
     checkzobject(wireprotov1peer.peerexecutor(None))
 
     ziverify.verifyClass(repository.ipeerbase, sshpeer.sshv1peer)
     checkzobject(sshpeer.sshv1peer(ui, b'ssh://localhost/foo', b'', dummypipe(),
                                    dummypipe(), None, None))
 
     ziverify.verifyClass(repository.ipeerbase, sshpeer.sshv2peer)
     checkzobject(sshpeer.sshv2peer(ui, b'ssh://localhost/foo', b'', dummypipe(),
                                    dummypipe(), None, None))
 
     ziverify.verifyClass(repository.ipeerbase, bundlerepo.bundlepeer)
     checkzobject(bundlerepo.bundlepeer(dummyrepo()))
 
     ziverify.verifyClass(repository.ipeerbase, statichttprepo.statichttppeer)
     checkzobject(statichttprepo.statichttppeer(dummyrepo()))
 
     ziverify.verifyClass(repository.ipeerbase, unionrepo.unionpeer)
     checkzobject(unionrepo.unionpeer(dummyrepo()))
 
     ziverify.verifyClass(repository.completelocalrepository,
                          localrepo.localrepository)
     repo = localrepo.makelocalrepository(ui, rootdir)
     checkzobject(repo)
 
     ziverify.verifyClass(wireprototypes.baseprotocolhandler,
                          wireprotoserver.sshv1protocolhandler)
     ziverify.verifyClass(wireprototypes.baseprotocolhandler,
                          wireprotoserver.sshv2protocolhandler)
     ziverify.verifyClass(wireprototypes.baseprotocolhandler,
                          wireprotoserver.httpv1protocolhandler)
     ziverify.verifyClass(wireprototypes.baseprotocolhandler,
                          wireprotov2server.httpv2protocolhandler)
 
     sshv1 = wireprotoserver.sshv1protocolhandler(None, None, None)
     checkzobject(sshv1)
     sshv2 = wireprotoserver.sshv2protocolhandler(None, None, None)
     checkzobject(sshv2)
 
     httpv1 = wireprotoserver.httpv1protocolhandler(None, None, None)
     checkzobject(httpv1)
     httpv2 = wireprotov2server.httpv2protocolhandler(None, None)
     checkzobject(httpv2)
 
     ziverify.verifyClass(repository.ifilestorage, filelog.filelog)
     ziverify.verifyClass(repository.imanifestdict, manifest.manifestdict)
     ziverify.verifyClass(repository.imanifestrevisionstored,
                          manifest.manifestctx)
     ziverify.verifyClass(repository.imanifestrevisionwritable,
                          manifest.memmanifestctx)
     ziverify.verifyClass(repository.imanifestrevisionstored,
                          manifest.treemanifestctx)
     ziverify.verifyClass(repository.imanifestrevisionwritable,
                          manifest.memtreemanifestctx)
     ziverify.verifyClass(repository.imanifestlog, manifest.manifestlog)
     ziverify.verifyClass(repository.imanifeststorage, manifest.manifestrevlog)
 
     vfs = vfsmod.vfs(b'.')
     fl = filelog.filelog(vfs, b'dummy.i')
     checkzobject(fl, allowextra=True)
 
     # Conforms to imanifestlog.
-    ml = manifest.manifestlog(vfs, repo)
+    ml = manifest.manifestlog(vfs, repo, manifest.manifestrevlog(repo.svfs))
     checkzobject(ml)
     checkzobject(repo.manifestlog)
 
     # Conforms to imanifestrevision.
     mctx = ml[repo[0].manifestnode()]
     checkzobject(mctx)
 
     # Conforms to imanifestrevisionwritable.
     checkzobject(mctx.new())
     checkzobject(mctx.copy())
 
     # Conforms to imanifestdict.
     checkzobject(mctx.read())
 
     mrl = manifest.manifestrevlog(vfs)
     checkzobject(mrl)
 
     ziverify.verifyClass(repository.irevisiondelta,
                          revlog.revlogrevisiondelta)
     ziverify.verifyClass(repository.irevisiondeltarequest,
                          changegroup.revisiondeltarequest)
 
     rd = revlog.revlogrevisiondelta(
         node=b'',
         p1node=b'',
         p2node=b'',
         basenode=b'',
         linknode=b'',
         flags=b'',
         baserevisionsize=None,
         revision=b'',
         delta=None)
     checkzobject(rd)
 
     rdr = changegroup.revisiondeltarequest(
         node=b'',
         linknode=b'',
         p1node=b'',
         p2node=b'',
         basenode=b'')
     checkzobject(rdr)
 
 main()