config: add experimental argument to the config registrar

Until now, there are almost 28 config items which are considered as
`experimental` but, not present in the `experimental` section of
the registrar. This patch adds an `experimental` argument to the
config registrar to mark such config items.

Differential Revision: https://phab.mercurial-scm.org/D6728

contrib/perf.py

@@ -1,3063 +1,3068 b''
 # perf.py - performance test routines
 '''helper extension to measure performance
 
 Configurations
 ==============
 
 ``perf``
 --------
 
 ``all-timing``
     When set, additional statistics will be reported for each benchmark: best,
     worst, median average. If not set only the best timing is reported
     (default: off).
 
 ``presleep``
   number of second to wait before any group of runs (default: 1)
 
 ``pre-run``
   number of run to perform before starting measurement.
 
 ``profile-benchmark``
   Enable profiling for the benchmarked section.
   (The first iteration is benchmarked)
 
 ``run-limits``
   Control the number of runs each benchmark will perform. The option value
   should be a list of `<time>-<numberofrun>` pairs. After each run the
   conditions are considered in order with the following logic:
 
       If benchmark has been running for <time> seconds, and we have performed
       <numberofrun> iterations, stop the benchmark,
 
   The default value is: `3.0-100, 10.0-3`
 
 ``stub``
     When set, benchmarks will only be run once, useful for testing
     (default: off)
 '''
 
 # "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 contextlib
 import functools
 import gc
 import os
 import random
 import shutil
 import struct
 import sys
 import tempfile
 import threading
 import time
 from mercurial import (
     changegroup,
     cmdutil,
     commands,
     copies,
     error,
     extensions,
     hg,
     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.utils import repoviewutil # since 5.0
 except ImportError:
     repoviewutil = None
 try:
     from mercurial import scmutil # since 1.9 (or 8b252e826c68)
 except ImportError:
     pass
 try:
     from mercurial import setdiscovery # since 1.9 (or cb98fed52495)
 except ImportError:
     pass
 
 try:
     from mercurial import profiling
 except ImportError:
     profiling = None
 
 def identity(a):
     return a
 
 try:
     from mercurial import pycompat
     getargspec = pycompat.getargspec  # added to module after 4.5
     _byteskwargs = pycompat.byteskwargs  # since 4.1 (or fbc3f73dc802)
     _sysstr = pycompat.sysstr         # since 4.0 (or 2219f4f82ede)
     _xrange = pycompat.xrange         # since 4.8 (or 7eba8f83129b)
     fsencode = pycompat.fsencode      # since 3.9 (or f4a5e0e86a7e)
     if pycompat.ispy3:
         _maxint = sys.maxsize  # per py3 docs for replacing maxint
     else:
         _maxint = sys.maxint
 except (ImportError, AttributeError):
     import inspect
     getargspec = inspect.getargspec
     _byteskwargs = identity
     fsencode = identity               # no py3 support
     _maxint = sys.maxint              # no py3 support
     _sysstr = lambda x: x             # no py3 support
     _xrange = xrange
 
 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, _sysstr(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.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,
+        experimental=True,
     )
     configitem(b'perf', b'stub',
         default=mercurial.configitems.dynamicdefault,
+        experimental=True,
     )
     configitem(b'perf', b'parentscount',
         default=mercurial.configitems.dynamicdefault,
+        experimental=True,
     )
     configitem(b'perf', b'all-timing',
         default=mercurial.configitems.dynamicdefault,
+        experimental=True,
     )
     configitem(b'perf', b'pre-run',
         default=mercurial.configitems.dynamicdefault,
     )
     configitem(b'perf', b'profile-benchmark',
         default=mercurial.configitems.dynamicdefault,
     )
     configitem(b'perf', b'run-limits',
         default=mercurial.configitems.dynamicdefault,
+        experimental=True,
     )
 except (ImportError, AttributeError):
     pass
 
 def getlen(ui):
     if ui.configbool(b"perf", b"stub", False):
         return lambda x: 1
     return len
 
 class noop(object):
     """dummy context manager"""
     def __enter__(self):
         pass
     def __exit__(self, *args):
         pass
 
 NOOPCTX = noop()
 
 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)
 
     # experimental config: perf.run-limits
     limitspec = ui.configlist(b"perf", b"run-limits", [])
     limits = []
     for item in limitspec:
         parts = item.split(b'-', 1)
         if len(parts) < 2:
             ui.warn((b'malformatted run limit entry, missing "-": %s\n'
                      % item))
             continue
         try:
             time_limit = float(pycompat.sysstr(parts[0]))
         except ValueError as e:
             ui.warn((b'malformatted run limit entry, %s: %s\n'
                      % (pycompat.bytestr(e), item)))
             continue
         try:
             run_limit = int(pycompat.sysstr(parts[1]))
         except ValueError as e:
             ui.warn((b'malformatted run limit entry, %s: %s\n'
                      % (pycompat.bytestr(e), item)))
             continue
         limits.append((time_limit, run_limit))
     if not limits:
         limits = DEFAULTLIMITS
 
     profiler = None
     if profiling is not None:
         if ui.configbool(b"perf", b"profile-benchmark", False):
             profiler = profiling.profile(ui)
 
     prerun = getint(ui, b"perf", b"pre-run", 0)
     t = functools.partial(_timer, fm, displayall=displayall, limits=limits,
                           prerun=prerun, profiler=profiler)
     return t, fm
 
 def stub_timer(fm, func, setup=None, title=None):
     if setup is not None:
         setup()
     func()
 
 @contextlib.contextmanager
 def timeone():
     r = []
     ostart = os.times()
     cstart = util.timer()
     yield r
     cstop = util.timer()
     ostop = os.times()
     a, b = ostart, ostop
     r.append((cstop - cstart, b[0] - a[0], b[1]-a[1]))
 
 
 # list of stop condition (elapsed time, minimal run count)
 DEFAULTLIMITS = (
     (3.0, 100),
     (10.0, 3),
 )
 
 def _timer(fm, func, setup=None, title=None, displayall=False,
            limits=DEFAULTLIMITS, prerun=0, profiler=None):
     gc.collect()
     results = []
     begin = util.timer()
     count = 0
     if profiler is None:
         profiler = NOOPCTX
     for i in range(prerun):
         if setup is not None:
             setup()
         func()
     keepgoing = True
     while keepgoing:
         if setup is not None:
             setup()
         with profiler:
             with timeone() as item:
                 r = func()
         profiler = NOOPCTX
         count += 1
         results.append(item[0])
         cstop = util.timer()
         # Look for a stop condition.
         elapsed = cstop - begin
         for t, mincount in limits:
             if elapsed >= t and count >= mincount:
                 keepgoing = False
                 break
 
     formatone(fm, results, title=title, result=r,
               displayall=displayall)
 
 def formatone(fm, timings, title=None, result=None, displayall=False):
 
     count = len(timings)
 
     fm.startitem()
 
     if title:
         fm.write(b'title', b'! %s\n', title)
     if result:
         fm.write(b'result', b'! result: %s\n', result)
     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')
     timings.sort()
     min_val = timings[0]
     display(b'best', min_val)
     if displayall:
         max_val = timings[-1]
         display(b'max', max_val)
         avg = tuple([sum(x) / count for x in zip(*timings)])
         display(b'avg', avg)
         median = timings[len(timings) // 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, _sysstr(name))
     class attrutil(object):
         def set(self, newvalue):
             setattr(obj, _sysstr(name), newvalue)
         def restore(self):
             setattr(obj, _sysstr(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)
     # - repoviewutil since 5.0
     for mod in (branchmap, repoview, repoviewutil):
         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(obj, attrname):
     unfiltered = getattr(obj, 'unfiltered', None)
     if unfiltered is not None:
         obj = obj.unfiltered()
     if attrname in vars(obj):
         delattr(obj, attrname)
     obj._filecache.pop(attrname, None)
 
 def clearchangelog(repo):
     if repo is not repo.unfiltered():
         object.__setattr__(repo, r'_clcachekey', None)
         object.__setattr__(repo, r'_clcache', None)
     clearfilecache(repo.unfiltered(), 'changelog')
 
 # perf commands
 
 @command(b'perfwalk', formatteropts)
 def perfwalk(ui, repo, *pats, **opts):
     opts = _byteskwargs(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):
     opts = _byteskwargs(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):
     opts = _byteskwargs(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):
     opts = _byteskwargs(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
         if b'uipathfn' in getargspec(scmutil.addremove).args:
             uipathfn = scmutil.getuipathfn(repo)
             timer(lambda: scmutil.addremove(repo, matcher, b"", uipathfn, opts))
         else:
             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):
     """benchmark the computation of a changelog heads"""
     opts = _byteskwargs(opts)
     timer, fm = gettimer(ui, opts)
     cl = repo.changelog
     def s():
         clearcaches(cl)
     def d():
         len(cl.headrevs())
     timer(d, setup=s)
     fm.end()
 
 @command(b'perftags', formatteropts+
         [
             (b'', b'clear-revlogs', False, b'refresh changelog and manifest'),
         ])
 def perftags(ui, repo, **opts):
     opts = _byteskwargs(opts)
     timer, fm = gettimer(ui, opts)
     repocleartagscache = repocleartagscachefunc(repo)
     clearrevlogs = opts[b'clear_revlogs']
     def s():
         if clearrevlogs:
             clearchangelog(repo)
             clearfilecache(repo.unfiltered(), 'manifest')
         repocleartagscache()
     def t():
         return len(repo.tags())
     timer(t, setup=s)
     fm.end()
 
 @command(b'perfancestors', formatteropts)
 def perfancestors(ui, repo, **opts):
     opts = _byteskwargs(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):
     opts = _byteskwargs(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'perfdiscovery', formatteropts, b'PATH')
 def perfdiscovery(ui, repo, path, **opts):
     """benchmark discovery between local repo and the peer at given path
     """
     repos = [repo, None]
     timer, fm = gettimer(ui, opts)
     path = ui.expandpath(path)
 
     def s():
         repos[1] = hg.peer(ui, opts, path)
     def d():
         setdiscovery.findcommonheads(ui, *repos)
     timer(d, setup=s)
     fm.end()
 
 @command(b'perfbookmarks', formatteropts +
         [
             (b'', b'clear-revlogs', False, b'refresh changelog and manifest'),
         ])
 def perfbookmarks(ui, repo, **opts):
     """benchmark parsing bookmarks from disk to memory"""
     opts = _byteskwargs(opts)
     timer, fm = gettimer(ui, opts)
 
     clearrevlogs = opts[b'clear_revlogs']
     def s():
         if clearrevlogs:
             clearchangelog(repo)
         clearfilecache(repo, b'_bookmarks')
     def d():
         repo._bookmarks
     timer(d, setup=s)
     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,
     )
 
     opts = _byteskwargs(opts)
 
     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'cgversion', b'02', b'changegroup version'),
           (b'r', b'rev', b'', b'revisions to add to changegroup')])
 def perfchangegroupchangelog(ui, repo, cgversion=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.
     """
     opts = _byteskwargs(opts)
     cl = repo.changelog
     nodes = [cl.lookup(r) for r in repo.revs(rev or b'all()')]
     bundler = changegroup.getbundler(cgversion, 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):
     opts = _byteskwargs(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):
     opts = _byteskwargs(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):
     opts = _byteskwargs(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):
     opts = _byteskwargs(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):
     opts = _byteskwargs(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):
     opts = _byteskwargs(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()
 
 def _getmergerevs(repo, opts):
     """parse command argument to return rev involved in merge
 
     input: options dictionnary with `rev`, `from` and `bse`
     output: (localctx, otherctx, basectx)
     """
     if opts[b'from']:
         fromrev = scmutil.revsingle(repo, opts[b'from'])
         wctx = repo[fromrev]
     else:
         wctx = repo[None]
         # we don't want working dir files to be stat'd in the benchmark, so
         # prime that cache
         wctx.dirty()
     rctx = scmutil.revsingle(repo, opts[b'rev'], opts[b'rev'])
     if opts[b'base']:
         fromrev = scmutil.revsingle(repo, opts[b'base'])
         ancestor = repo[fromrev]
     else:
         ancestor = wctx.ancestor(rctx)
     return (wctx, rctx, ancestor)
 
 @command(b'perfmergecalculate',
          [
              (b'r', b'rev', b'.', b'rev to merge against'),
              (b'', b'from', b'', b'rev to merge from'),
              (b'', b'base', b'', b'the revision to use as base'),
          ] + formatteropts)
 def perfmergecalculate(ui, repo, **opts):
     opts = _byteskwargs(opts)
     timer, fm = gettimer(ui, opts)
 
     wctx, rctx, ancestor = _getmergerevs(repo, opts)
     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'perfmergecopies',
          [
              (b'r', b'rev', b'.', b'rev to merge against'),
              (b'', b'from', b'', b'rev to merge from'),
              (b'', b'base', b'', b'the revision to use as base'),
          ] + formatteropts)
 def perfmergecopies(ui, repo, **opts):
     """measure runtime of `copies.mergecopies`"""
     opts = _byteskwargs(opts)
     timer, fm = gettimer(ui, opts)
     wctx, rctx, ancestor = _getmergerevs(repo, opts)
     def d():
         # acceptremote is True because we don't want prompts in the middle of
         # our benchmark
         copies.mergecopies(repo, wctx, rctx, ancestor)
     timer(d)
     fm.end()
 
 @command(b'perfpathcopies', [], b"REV REV")
 def perfpathcopies(ui, repo, rev1, rev2, **opts):
     """benchmark the copy tracing logic"""
     opts = _byteskwargs(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"""
     opts = _byteskwargs(opts)
     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,
     )
     opts = _byteskwargs(opts)
     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
     ui.status((b'analysing phase of %s\n') % util.hidepassword(dest))
     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'),
          ] + formatteropts, 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."""
     opts = _byteskwargs(opts)
     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):
     opts = _byteskwargs(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'perfignore', formatteropts)
 def perfignore(ui, repo, **opts):
     """benchmark operation related to computing ignore"""
     opts = _byteskwargs(opts)
     timer, fm = gettimer(ui, opts)
     dirstate = repo.dirstate
 
     def setupone():
         dirstate.invalidate()
         clearfilecache(dirstate, b'_ignore')
 
     def runone():
         dirstate._ignore
 
     timer(runone, setup=setupone, title=b"load")
     fm.end()
 
 @command(b'perfindex', [
             (b'', b'rev', [], b'revision to be looked up (default tip)'),
             (b'', b'no-lookup', None, b'do not revision lookup post creation'),
          ] + formatteropts)
 def perfindex(ui, repo, **opts):
     """benchmark index creation time followed by a lookup
 
     The default is to look `tip` up. Depending on the index implementation,
     the revision looked up can matters. For example, an implementation
     scanning the index will have a faster lookup time for `--rev tip` than for
     `--rev 0`. The number of looked up revisions and their order can also
     matters.
 
     Example of useful set to test:
     * tip
     * 0
     * -10:
     * :10
     * -10: + :10
     * :10: + -10:
     * -10000:
     * -10000: + 0
 
     It is not currently possible to check for lookup of a missing node. For
     deeper lookup benchmarking, checkout the `perfnodemap` command."""
     import mercurial.revlog
     opts = _byteskwargs(opts)
     timer, fm = gettimer(ui, opts)
     mercurial.revlog._prereadsize = 2**24 # disable lazy parser in old hg
     if opts[b'no_lookup']:
         if opts['rev']:
             raise error.Abort('--no-lookup and --rev are mutually exclusive')
         nodes = []
     elif not opts[b'rev']:
         nodes = [repo[b"tip"].node()]
     else:
         revs = scmutil.revrange(repo, opts[b'rev'])
         cl = repo.changelog
         nodes = [cl.node(r) for r in revs]
 
     unfi = repo.unfiltered()
     # find the filecache func directly
     # This avoid polluting the benchmark with the filecache logic
     makecl = unfi.__class__.changelog.func
     def setup():
         # probably not necessary, but for good measure
         clearchangelog(unfi)
     def d():
         cl = makecl(unfi)
         for n in nodes:
             cl.rev(n)
     timer(d, setup=setup)
     fm.end()
 
 @command(b'perfnodemap', [
           (b'', b'rev', [], b'revision to be looked up (default tip)'),
           (b'', b'clear-caches', True, b'clear revlog cache between calls'),
     ] + formatteropts)
 def perfnodemap(ui, repo, **opts):
     """benchmark the time necessary to look up revision from a cold nodemap
 
     Depending on the implementation, the amount and order of revision we look
     up can varies. Example of useful set to test:
     * tip
     * 0
     * -10:
     * :10
     * -10: + :10
     * :10: + -10:
     * -10000:
     * -10000: + 0
 
     The command currently focus on valid binary lookup. Benchmarking for
     hexlookup, prefix lookup and missing lookup would also be valuable.
     """
     import mercurial.revlog
     opts = _byteskwargs(opts)
     timer, fm = gettimer(ui, opts)
     mercurial.revlog._prereadsize = 2**24 # disable lazy parser in old hg
 
     unfi = repo.unfiltered()
     clearcaches = opts['clear_caches']
     # find the filecache func directly
     # This avoid polluting the benchmark with the filecache logic
     makecl = unfi.__class__.changelog.func
     if not opts[b'rev']:
         raise error.Abort('use --rev to specify revisions to look up')
     revs = scmutil.revrange(repo, opts[b'rev'])
     cl = repo.changelog
     nodes = [cl.node(r) for r in revs]
 
     # use a list to pass reference to a nodemap from one closure to the next
     nodeget = [None]
     def setnodeget():
         # probably not necessary, but for good measure
         clearchangelog(unfi)
         nodeget[0] = makecl(unfi).nodemap.get
 
     def d():
         get = nodeget[0]
         for n in nodes:
             get(n)
 
     setup = None
     if clearcaches:
         def setup():
             setnodeget()
     else:
         setnodeget()
         d() # prewarm the data structure
     timer(d, setup=setup)
     fm.end()
 
 @command(b'perfstartup', formatteropts)
 def perfstartup(ui, repo, **opts):
     opts = _byteskwargs(opts)
     timer, fm = gettimer(ui, opts)
     def d():
         if os.name != r'nt':
             os.system(b"HGRCPATH= %s version -q > /dev/null" %
                       fsencode(sys.argv[0]))
         else:
             os.environ[r'HGRCPATH'] = r' '
             os.system(r"%s version -q > NUL" % sys.argv[0])
     timer(d)
     fm.end()
 
 @command(b'perfparents', formatteropts)
 def perfparents(ui, repo, **opts):
     """benchmark the time necessary to fetch one changeset's parents.
 
     The fetch is done using the `node identifier`, traversing all object layers
     from the repository object. The first N revisions will be used for this
     benchmark. N is controlled by the ``perf.parentscount`` config option
     (default: 1000).
     """
     opts = _byteskwargs(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):
     opts = _byteskwargs(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):
     opts = _byteskwargs(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):
     opts = _byteskwargs(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
 
     opts = _byteskwargs(opts)
 
     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):
     opts = _byteskwargs(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):
     opts = _byteskwargs(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):
     opts = _byteskwargs(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.
     """
     opts = _byteskwargs(opts)
     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")
 
     opts = _byteskwargs(opts)
 
     nullui = ui.copy()
     nullui.fout = open(os.devnull, r'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'perfhelper-mergecopies', formatteropts +
          [
           (b'r', b'revs', [], b'restrict search to these revisions'),
           (b'', b'timing', False, b'provides extra data (costly)'),
          ])
 def perfhelpermergecopies(ui, repo, revs=[], **opts):
     """find statistics about potential parameters for `perfmergecopies`
 
     This command find (base, p1, p2) triplet relevant for copytracing
     benchmarking in the context of a merge.  It reports values for some of the
     parameters that impact merge copy tracing time during merge.
 
     If `--timing` is set, rename detection is run and the associated timing
     will be reported. The extra details come at the cost of slower command
     execution.
 
     Since rename detection is only run once, other factors might easily
     affect the precision of the timing. However it should give a good
     approximation of which revision triplets are very costly.
     """
     opts = _byteskwargs(opts)
     fm = ui.formatter(b'perf', opts)
     dotiming = opts[b'timing']
 
     output_template = [
         ("base", "%(base)12s"),
         ("p1", "%(p1.node)12s"),
         ("p2", "%(p2.node)12s"),
         ("p1.nb-revs", "%(p1.nbrevs)12d"),
         ("p1.nb-files", "%(p1.nbmissingfiles)12d"),
         ("p1.renames", "%(p1.renamedfiles)12d"),
         ("p1.time", "%(p1.time)12.3f"),
         ("p2.nb-revs", "%(p2.nbrevs)12d"),
         ("p2.nb-files", "%(p2.nbmissingfiles)12d"),
         ("p2.renames", "%(p2.renamedfiles)12d"),
         ("p2.time", "%(p2.time)12.3f"),
         ("renames", "%(nbrenamedfiles)12d"),
         ("total.time", "%(time)12.3f"),
         ]
     if not dotiming:
         output_template = [i for i in output_template
                            if not ('time' in i[0] or 'renames' in i[0])]
     header_names = [h for (h, v) in output_template]
     output = ' '.join([v for (h, v) in output_template]) + '\n'
     header = ' '.join(['%12s'] * len(header_names)) + '\n'
     fm.plain(header % tuple(header_names))
 
     if not revs:
         revs = ['all()']
     revs = scmutil.revrange(repo, revs)
 
     roi = repo.revs('merge() and %ld', revs)
     for r in roi:
         ctx = repo[r]
         p1 = ctx.p1()
         p2 = ctx.p2()
         bases = repo.changelog._commonancestorsheads(p1.rev(), p2.rev())
         for b in bases:
             b = repo[b]
             p1missing = copies._computeforwardmissing(b, p1)
             p2missing = copies._computeforwardmissing(b, p2)
             data = {
                 b'base': b.hex(),
                 b'p1.node': p1.hex(),
                 b'p1.nbrevs': len(repo.revs('%d::%d', b.rev(), p1.rev())),
                 b'p1.nbmissingfiles': len(p1missing),
                 b'p2.node': p2.hex(),
                 b'p2.nbrevs': len(repo.revs('%d::%d', b.rev(), p2.rev())),
                 b'p2.nbmissingfiles': len(p2missing),
             }
             if dotiming:
                 begin = util.timer()
                 mergedata = copies.mergecopies(repo, p1, p2, b)
                 end = util.timer()
                 # not very stable timing since we did only one run
                 data['time'] = end - begin
                 # mergedata contains five dicts: "copy", "movewithdir",
                 # "diverge", "renamedelete" and "dirmove".
                 # The first 4 are about renamed file so lets count that.
                 renames = len(mergedata[0])
                 renames += len(mergedata[1])
                 renames += len(mergedata[2])
                 renames += len(mergedata[3])
                 data['nbrenamedfiles'] = renames
                 begin = util.timer()
                 p1renames = copies.pathcopies(b, p1)
                 end = util.timer()
                 data['p1.time'] = end - begin
                 begin = util.timer()
                 p2renames = copies.pathcopies(b, p2)
                 data['p2.time'] = end - begin
                 end = util.timer()
                 data['p1.renamedfiles'] = len(p1renames)
                 data['p2.renamedfiles'] = len(p2renames)
             fm.startitem()
             fm.data(**data)
             # make node pretty for the human output
             out = data.copy()
             out['base'] = fm.hexfunc(b.node())
             out['p1.node'] = fm.hexfunc(p1.node())
             out['p2.node'] = fm.hexfunc(p2.node())
             fm.plain(output % out)
 
     fm.end()
 
 @command(b'perfhelper-pathcopies', formatteropts +
          [
           (b'r', b'revs', [], b'restrict search to these revisions'),
           (b'', b'timing', False, b'provides extra data (costly)'),
          ])
 def perfhelperpathcopies(ui, repo, revs=[], **opts):
     """find statistic about potential parameters for the `perftracecopies`
 
     This command find source-destination pair relevant for copytracing testing.
     It report value for some of the parameters that impact copy tracing time.
 
     If `--timing` is set, rename detection is run and the associated timing
     will be reported. The extra details comes at the cost of a slower command
     execution.
 
     Since the rename detection is only run once, other factors might easily
     affect the precision of the timing. However it should give a good
     approximation of which revision pairs are very costly.
     """
     opts = _byteskwargs(opts)
     fm = ui.formatter(b'perf', opts)
     dotiming = opts[b'timing']
 
     if dotiming:
         header = '%12s %12s %12s %12s %12s %12s\n'
         output = ("%(source)12s %(destination)12s "
                   "%(nbrevs)12d %(nbmissingfiles)12d "
                   "%(nbrenamedfiles)12d %(time)18.5f\n")
         header_names = ("source", "destination", "nb-revs", "nb-files",
                         "nb-renames", "time")
         fm.plain(header % header_names)
     else:
         header = '%12s %12s %12s %12s\n'
         output = ("%(source)12s %(destination)12s "
                   "%(nbrevs)12d %(nbmissingfiles)12d\n")
         fm.plain(header % ("source", "destination", "nb-revs", "nb-files"))
 
     if not revs:
         revs = ['all()']
     revs = scmutil.revrange(repo, revs)
 
     roi = repo.revs('merge() and %ld', revs)
     for r in roi:
         ctx = repo[r]
         p1 = ctx.p1().rev()
         p2 = ctx.p2().rev()
         bases = repo.changelog._commonancestorsheads(p1, p2)
         for p in (p1, p2):
             for b in bases:
                 base = repo[b]
                 parent = repo[p]
                 missing = copies._computeforwardmissing(base, parent)
                 if not missing:
                     continue
                 data = {
                     b'source': base.hex(),
                     b'destination': parent.hex(),
                     b'nbrevs': len(repo.revs('%d::%d', b, p)),
                     b'nbmissingfiles': len(missing),
                 }
                 if dotiming:
                     begin = util.timer()
                     renames = copies.pathcopies(base, parent)
                     end = util.timer()
                     # not very stable timing since we did only one run
                     data['time'] = end - begin
                     data['nbrenamedfiles'] = len(renames)
                 fm.startitem()
                 fm.data(**data)
                 out = data.copy()
                 out['source'] = fm.hexfunc(base.node())
                 out['destination'] = fm.hexfunc(parent.node())
                 fm.plain(output % out)
 
     fm.end()
 
 @command(b'perfcca', formatteropts)
 def perfcca(ui, repo, **opts):
     opts = _byteskwargs(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):
     opts = _byteskwargs(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):
     opts = _byteskwargs(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):
     opts = _byteskwargs(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 = _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).
     """
     opts = _byteskwargs(opts)
     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"""
     opts = _byteskwargs(opts)
     timer, fm = gettimer(ui, opts)
     options = {
         'w': 'ignore_all_space',
         'b': 'ignore_space_change',
         'B': 'ignore_blank_lines',
         }
 
     for diffopt in ('', 'w', 'b', 'B', 'wB'):
         opts = dict((options[c], b'1') for c in diffopt)
         def d():
             ui.pushbuffer()
             commands.diff(ui, repo, **opts)
             ui.popbuffer()
         diffopt = diffopt.encode('ascii')
         title = b'diffopts: %s' % (diffopt and (b'-' + diffopt) or b'none')
         timer(d, title=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.
     """
 
     opts = _byteskwargs(opts)
 
     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``.
     """
     opts = _byteskwargs(opts)
 
     rl = cmdutil.openrevlog(repo, b'perfrevlogrevisions', file_, opts)
     rllen = getlen(ui)(rl)
 
     if startrev < 0:
         startrev = rllen + startrev
 
     def d():
         rl.clearcaches()
 
         beginrev = startrev
         endrev = rllen
         dist = opts[b'dist']
 
         if reverse:
             beginrev, endrev = endrev - 1, beginrev - 1
             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'perfrevlogwrite', revlogopts + formatteropts +
          [(b's', b'startrev', 1000, b'revision to start writing at'),
           (b'', b'stoprev', -1, b'last revision to write'),
           (b'', b'count', 3, b'number of passes to perform'),
           (b'', b'details', False, b'print timing for every revisions tested'),
           (b'', b'source', b'full', b'the kind of data feed in the revlog'),
           (b'', b'lazydeltabase', True, b'try the provided delta first'),
           (b'', b'clear-caches', True, b'clear revlog cache between calls'),
          ],
          b'-c|-m|FILE')
 def perfrevlogwrite(ui, repo, file_=None, startrev=1000, stoprev=-1, **opts):
     """Benchmark writing a series of revisions to a revlog.
 
     Possible source values are:
     * `full`: add from a full text (default).
     * `parent-1`: add from a delta to the first parent
     * `parent-2`: add from a delta to the second parent if it exists
                   (use a delta from the first parent otherwise)
     * `parent-smallest`: add from the smallest delta (either p1 or p2)
     * `storage`: add from the existing precomputed deltas
 
     Note: This performance command measures performance in a custom way. As a
     result some of the global configuration of the 'perf' command does not
     apply to it:
 
     * ``pre-run``: disabled
 
     * ``profile-benchmark``: disabled
 
     * ``run-limits``: disabled use --count instead
     """
     opts = _byteskwargs(opts)
 
     rl = cmdutil.openrevlog(repo, b'perfrevlogwrite', file_, opts)
     rllen = getlen(ui)(rl)
     if startrev < 0:
         startrev = rllen + startrev
     if stoprev < 0:
         stoprev = rllen + stoprev
 
     lazydeltabase = opts['lazydeltabase']
     source = opts['source']
     clearcaches = opts['clear_caches']
     validsource = (b'full', b'parent-1', b'parent-2', b'parent-smallest',
                    b'storage')
     if source not in validsource:
         raise error.Abort('invalid source type: %s' % source)
 
     ### actually gather results
     count = opts['count']
     if count <= 0:
         raise error.Abort('invalide run count: %d' % count)
     allresults = []
     for c in range(count):
         timing = _timeonewrite(ui, rl, source, startrev, stoprev, c + 1,
                                lazydeltabase=lazydeltabase,
                                clearcaches=clearcaches)
         allresults.append(timing)
 
     ### consolidate the results in a single list
     results = []
     for idx, (rev, t) in enumerate(allresults[0]):
         ts = [t]
         for other in allresults[1:]:
             orev, ot = other[idx]
             assert orev == rev
             ts.append(ot)
         results.append((rev, ts))
     resultcount = len(results)
 
     ### Compute and display relevant statistics
 
     # get a formatter
     fm = ui.formatter(b'perf', opts)
     displayall = ui.configbool(b"perf", b"all-timing", False)
 
     # print individual details if requested
     if opts['details']:
         for idx, item in enumerate(results, 1):
             rev, data = item
             title = 'revisions #%d of %d, rev %d' % (idx, resultcount, rev)
             formatone(fm, data, title=title, displayall=displayall)
 
     # sorts results by median time
     results.sort(key=lambda x: sorted(x[1])[len(x[1]) // 2])
     # list of (name, index) to display)
     relevants = [
         ("min", 0),
         ("10%", resultcount * 10 // 100),
         ("25%", resultcount * 25 // 100),
         ("50%", resultcount * 70 // 100),
         ("75%", resultcount * 75 // 100),
         ("90%", resultcount * 90 // 100),
         ("95%", resultcount * 95 // 100),
         ("99%", resultcount * 99 // 100),
         ("99.9%", resultcount * 999 // 1000),
         ("99.99%", resultcount * 9999 // 10000),
         ("99.999%", resultcount * 99999 // 100000),
         ("max", -1),
     ]
     if not ui.quiet:
         for name, idx in relevants:
             data = results[idx]
             title = '%s of %d, rev %d' % (name, resultcount, data[0])
             formatone(fm, data[1], title=title, displayall=displayall)
 
     # XXX summing that many float will not be very precise, we ignore this fact
     # for now
     totaltime = []
     for item in allresults:
         totaltime.append((sum(x[1][0] for x in item),
                           sum(x[1][1] for x in item),
                           sum(x[1][2] for x in item),)
         )
     formatone(fm, totaltime, title="total time (%d revs)" % resultcount,
               displayall=displayall)
     fm.end()
 
 class _faketr(object):
     def add(s, x, y, z=None):
         return None
 
 def _timeonewrite(ui, orig, source, startrev, stoprev, runidx=None,
                   lazydeltabase=True, clearcaches=True):
     timings = []
     tr = _faketr()
     with _temprevlog(ui, orig, startrev) as dest:
         dest._lazydeltabase = lazydeltabase
         revs = list(orig.revs(startrev, stoprev))
         total = len(revs)
         topic = 'adding'
         if runidx is not None:
             topic += ' (run #%d)' % runidx
          # Support both old and new progress API
         if util.safehasattr(ui, 'makeprogress'):
             progress = ui.makeprogress(topic, unit='revs', total=total)
             def updateprogress(pos):
                 progress.update(pos)
             def completeprogress():
                 progress.complete()
         else:
             def updateprogress(pos):
                 ui.progress(topic, pos, unit='revs', total=total)
             def completeprogress():
                 ui.progress(topic, None, unit='revs', total=total)
 
         for idx, rev in enumerate(revs):
             updateprogress(idx)
             addargs, addkwargs = _getrevisionseed(orig, rev, tr, source)
             if clearcaches:
                 dest.index.clearcaches()
                 dest.clearcaches()
             with timeone() as r:
                 dest.addrawrevision(*addargs, **addkwargs)
             timings.append((rev, r[0]))
         updateprogress(total)
         completeprogress()
     return timings
 
 def _getrevisionseed(orig, rev, tr, source):
     from mercurial.node import nullid
 
     linkrev = orig.linkrev(rev)
     node = orig.node(rev)
     p1, p2 = orig.parents(node)
     flags = orig.flags(rev)
     cachedelta = None
     text = None
 
     if source == b'full':
         text = orig.revision(rev)
     elif source == b'parent-1':
         baserev = orig.rev(p1)
         cachedelta = (baserev, orig.revdiff(p1, rev))
     elif source == b'parent-2':
         parent = p2
         if p2 == nullid:
             parent = p1
         baserev = orig.rev(parent)
         cachedelta = (baserev, orig.revdiff(parent, rev))
     elif source == b'parent-smallest':
         p1diff = orig.revdiff(p1, rev)
         parent = p1
         diff = p1diff
         if p2 != nullid:
             p2diff = orig.revdiff(p2, rev)
             if len(p1diff) > len(p2diff):
                 parent = p2
                 diff = p2diff
         baserev = orig.rev(parent)
         cachedelta = (baserev, diff)
     elif source == b'storage':
         baserev = orig.deltaparent(rev)
         cachedelta = (baserev, orig.revdiff(orig.node(baserev), rev))
 
     return ((text, tr, linkrev, p1, p2),
             {'node': node, 'flags': flags, 'cachedelta': cachedelta})
 
 @contextlib.contextmanager
 def _temprevlog(ui, orig, truncaterev):
     from mercurial import vfs as vfsmod
 
     if orig._inline:
         raise error.Abort('not supporting inline revlog (yet)')
     revlogkwargs = {}
     k = 'upperboundcomp'
     if util.safehasattr(orig, k):
         revlogkwargs[k] = getattr(orig, k)
 
     origindexpath = orig.opener.join(orig.indexfile)
     origdatapath = orig.opener.join(orig.datafile)
     indexname = 'revlog.i'
     dataname = 'revlog.d'
 
     tmpdir = tempfile.mkdtemp(prefix='tmp-hgperf-')
     try:
         # copy the data file in a temporary directory
         ui.debug('copying data in %s\n' % tmpdir)
         destindexpath = os.path.join(tmpdir, 'revlog.i')
         destdatapath = os.path.join(tmpdir, 'revlog.d')
         shutil.copyfile(origindexpath, destindexpath)
         shutil.copyfile(origdatapath, destdatapath)
 
         # remove the data we want to add again
         ui.debug('truncating data to be rewritten\n')
         with open(destindexpath, 'ab') as index:
             index.seek(0)
             index.truncate(truncaterev * orig._io.size)
         with open(destdatapath, 'ab') as data:
             data.seek(0)
             data.truncate(orig.start(truncaterev))
 
         # instantiate a new revlog from the temporary copy
         ui.debug('truncating adding to be rewritten\n')
         vfs = vfsmod.vfs(tmpdir)
         vfs.options = getattr(orig.opener, 'options', None)
 
         dest = revlog.revlog(vfs,
                              indexfile=indexname,
                              datafile=dataname, **revlogkwargs)
         if dest._inline:
             raise error.Abort('not supporting inline revlog (yet)')
         # make sure internals are initialized
         dest.revision(len(dest) - 1)
         yield dest
         del dest, vfs
     finally:
         shutil.rmtree(tmpdir, True)
 
 @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``.
     """
     opts = _byteskwargs(opts)
 
     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. Slice the delta chain if applicable
     3. Obtain the raw chunks for that delta chain
     4. Decompress each raw chunk
     5. Apply binary patches to obtain fulltext
     6. Verify hash of fulltext
 
     This command measures the time spent in each of these phases.
     """
     opts = _byteskwargs(opts)
 
     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
 
         chunks = []
         ladd = chunks.append
         for idx, item in enumerate(chain):
             offset = start(item[0])
             bits = data[idx]
             for rev in item:
                 chunkstart = start(rev)
                 if inline:
                     chunkstart += (rev + 1) * iosize
                 chunklength = length(rev)
                 ladd(buffer(bits, chunkstart - offset, chunklength))
 
         return chunks
 
     def dodeltachain(rev):
         if not cache:
             r.clearcaches()
         r._deltachain(rev)
 
     def doread(chain):
         if not cache:
             r.clearcaches()
         for item in slicedchain:
             segmentforrevs(item[0], item[-1])
 
     def doslice(r, chain, size):
         for s in slicechunk(r, chain, targetsize=size):
             pass
 
     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)
 
     try:
         from mercurial.revlogutils.deltas import slicechunk
     except ImportError:
         slicechunk = getattr(revlog, '_slicechunk', None)
 
     size = r.length(rev)
     chain = r._deltachain(rev)[0]
     if not getattr(r, '_withsparseread', False):
         slicedchain = (chain,)
     else:
         slicedchain = tuple(slicechunk(r, chain, targetsize=size))
     data = [segmentforrevs(seg[0], seg[-1])[1] for seg in slicedchain]
     rawchunks = getrawchunks(data, slicedchain)
     bins = r._chunks(chain)
     text = bytes(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'),
     ]
 
     if getattr(r, '_withsparseread', False):
         slicing = (lambda: doslice(r, chain, size), b'slice-sparse-chain')
         benches.append(slicing)
 
     benches.extend([
         (lambda: dorawchunks(data, slicedchain), b'rawchunks'),
         (lambda: dodecompress(rawchunks), b'decompress'),
         (lambda: dopatch(text, bins), b'patch'),
         (lambda: dohash(text), b'hash'),
     ])
 
     timer, fm = gettimer(ui, opts)
     for fn, title in benches:
         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."""
     opts = _byteskwargs(opts)
 
     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."""
     opts = _byteskwargs(opts)
     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
     """
     opts = _byteskwargs(opts)
     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)
         if util.safehasattr(view._branchcaches, '_per_filter'):
             filtered = view._branchcaches._per_filter
         else:
             # older versions
             filtered = view._branchcaches
         def d():
             if clear_revbranch:
                 repo.revbranchcache()._clear()
             if full:
                 view._branchcaches.clear()
             else:
                 filtered.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)
 
     if util.safehasattr(branchmap.branchcache, 'fromfile'):
         branchcacheread = safeattrsetter(branchmap.branchcache, b'fromfile')
         branchcacheread.set(classmethod(lambda *args: None))
     else:
         # older versions
         branchcacheread = safeattrsetter(branchmap, b'read')
         branchcacheread.set(lambda *args: None)
     branchcachewrite = safeattrsetter(branchmap.branchcache, b'write')
     branchcachewrite.set(lambda *args: 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'perfbranchmapupdate', [
      (b'', b'base', [], b'subset of revision to start from'),
      (b'', b'target', [], b'subset of revision to end with'),
      (b'', b'clear-caches', False, b'clear cache between each runs')
     ] + formatteropts)
 def perfbranchmapupdate(ui, repo, base=(), target=(), **opts):
     """benchmark branchmap update from for <base> revs to <target> revs
 
     If `--clear-caches` is passed, the following items will be reset before
     each update:
         * the changelog instance and associated indexes
         * the rev-branch-cache instance
 
     Examples:
 
        # update for the one last revision
        $ hg perfbranchmapupdate --base 'not tip' --target 'tip'
 
        $ update for change coming with a new branch
        $ hg perfbranchmapupdate --base 'stable' --target 'default'
     """
     from mercurial import branchmap
     from mercurial import repoview
     opts = _byteskwargs(opts)
     timer, fm = gettimer(ui, opts)
     clearcaches = opts[b'clear_caches']
     unfi = repo.unfiltered()
     x = [None] # used to pass data between closure
 
     # we use a `list` here to avoid possible side effect from smartset
     baserevs = list(scmutil.revrange(repo, base))
     targetrevs = list(scmutil.revrange(repo, target))
     if not baserevs:
         raise error.Abort(b'no revisions selected for --base')
     if not targetrevs:
         raise error.Abort(b'no revisions selected for --target')
 
     # make sure the target branchmap also contains the one in the base
     targetrevs = list(set(baserevs) | set(targetrevs))
     targetrevs.sort()
 
     cl = repo.changelog
     allbaserevs = list(cl.ancestors(baserevs, inclusive=True))
     allbaserevs.sort()
     alltargetrevs = frozenset(cl.ancestors(targetrevs, inclusive=True))
 
     newrevs = list(alltargetrevs.difference(allbaserevs))
     newrevs.sort()
 
     allrevs = frozenset(unfi.changelog.revs())
     basefilterrevs = frozenset(allrevs.difference(allbaserevs))
     targetfilterrevs = frozenset(allrevs.difference(alltargetrevs))
 
     def basefilter(repo, visibilityexceptions=None):
         return basefilterrevs
 
     def targetfilter(repo, visibilityexceptions=None):
         return targetfilterrevs
 
     msg = b'benchmark of branchmap with %d revisions with %d new ones\n'
     ui.status(msg % (len(allbaserevs), len(newrevs)))
     if targetfilterrevs:
         msg = b'(%d revisions still filtered)\n'
         ui.status(msg % len(targetfilterrevs))
 
     try:
         repoview.filtertable[b'__perf_branchmap_update_base'] = basefilter
         repoview.filtertable[b'__perf_branchmap_update_target'] = targetfilter
 
         baserepo = repo.filtered(b'__perf_branchmap_update_base')
         targetrepo = repo.filtered(b'__perf_branchmap_update_target')
 
         # try to find an existing branchmap to reuse
         subsettable = getbranchmapsubsettable()
         candidatefilter = subsettable.get(None)
         while candidatefilter is not None:
             candidatebm = repo.filtered(candidatefilter).branchmap()
             if candidatebm.validfor(baserepo):
                 filtered = repoview.filterrevs(repo, candidatefilter)
                 missing = [r for r in allbaserevs if r in filtered]
                 base = candidatebm.copy()
                 base.update(baserepo, missing)
                 break
             candidatefilter = subsettable.get(candidatefilter)
         else:
             # no suitable subset where found
             base = branchmap.branchcache()
             base.update(baserepo, allbaserevs)
 
         def setup():
             x[0] = base.copy()
             if clearcaches:
                 unfi._revbranchcache = None
                 clearchangelog(repo)
 
         def bench():
             x[0].update(targetrepo, newrevs)
 
         timer(bench, setup=setup)
         fm.end()
     finally:
         repoview.filtertable.pop(b'__perf_branchmap_update_base', None)
         repoview.filtertable.pop(b'__perf_branchmap_update_target', None)
 
 @command(b'perfbranchmapload', [
      (b'f', b'filter', b'', b'Specify repoview filter'),
      (b'', b'list', False, b'List brachmap filter caches'),
      (b'', b'clear-revlogs', False, b'refresh changelog and manifest'),
 
     ] + formatteropts)
 def perfbranchmapload(ui, repo, filter=b'', list=False, **opts):
     """benchmark reading the branchmap"""
     opts = _byteskwargs(opts)
     clearrevlogs = opts[b'clear_revlogs']
 
     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 not filter:
         filter = None
     subsettable = getbranchmapsubsettable()
     if filter is None:
         repo = repo.unfiltered()
     else:
         repo = repoview.repoview(repo, filter)
 
     repo.branchmap() # make sure we have a relevant, up to date branchmap
 
     try:
         fromfile = branchmap.branchcache.fromfile
     except AttributeError:
         # older versions
         fromfile = branchmap.read
 
     currentfilter = filter
     # try once without timer, the filter may not be cached
     while fromfile(repo) is None:
         currentfilter = subsettable.get(currentfilter)
         if currentfilter is None:
             raise error.Abort(b'No branchmap cached for %s repo'
                               % (filter or b'unfiltered'))
         repo = repo.filtered(currentfilter)
     timer, fm = gettimer(ui, opts)
     def setup():
         if clearrevlogs:
             clearchangelog(repo)
     def bench():
         fromfile(repo)
     timer(bench, setup=setup)
     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):
     opts = _byteskwargs(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, _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, _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
     """
     opts = _byteskwargs(opts)
 
     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)
 
 @command(b'perfprogress', formatteropts + [
     (b'', b'topic', b'topic', b'topic for progress messages'),
     (b'c', b'total', 1000000, b'total value we are progressing to'),
 ], norepo=True)
 def perfprogress(ui, topic=None, total=None, **opts):
     """printing of progress bars"""
     opts = _byteskwargs(opts)
 
     timer, fm = gettimer(ui, opts)
 
     def doprogress():
         with ui.makeprogress(topic, total=total) as progress:
             for i in pycompat.xrange(total):
                 progress.increment()
 
     timer(doprogress)
     fm.end()

hgext/fsmonitor/__init__.py

@@ -1,850 +1,851 b''
 # __init__.py - fsmonitor initialization and overrides
 #
 # Copyright 2013-2016 Facebook, Inc.
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 '''Faster status operations with the Watchman file monitor (EXPERIMENTAL)
 
 Integrates the file-watching program Watchman with Mercurial to produce faster
 status results.
 
 On a particular Linux system, for a real-world repository with over 400,000
 files hosted on ext4, vanilla `hg status` takes 1.3 seconds. On the same
 system, with fsmonitor it takes about 0.3 seconds.
 
 fsmonitor requires no configuration -- it will tell Watchman about your
 repository as necessary. You'll need to install Watchman from
 https://facebook.github.io/watchman/ and make sure it is in your PATH.
 
 fsmonitor is incompatible with the largefiles and eol extensions, and
 will disable itself if any of those are active.
 
 The following configuration options exist:
 
 ::
 
     [fsmonitor]
     mode = {off, on, paranoid}
 
 When `mode = off`, fsmonitor will disable itself (similar to not loading the
 extension at all). When `mode = on`, fsmonitor will be enabled (the default).
 When `mode = paranoid`, fsmonitor will query both Watchman and the filesystem,
 and ensure that the results are consistent.
 
 ::
 
     [fsmonitor]
     timeout = (float)
 
 A value, in seconds, that determines how long fsmonitor will wait for Watchman
 to return results. Defaults to `2.0`.
 
 ::
 
     [fsmonitor]
     blacklistusers = (list of userids)
 
 A list of usernames for which fsmonitor will disable itself altogether.
 
 ::
 
     [fsmonitor]
     walk_on_invalidate = (boolean)
 
 Whether or not to walk the whole repo ourselves when our cached state has been
 invalidated, for example when Watchman has been restarted or .hgignore rules
 have been changed. Walking the repo in that case can result in competing for
 I/O with Watchman. For large repos it is recommended to set this value to
 false. You may wish to set this to true if you have a very fast filesystem
 that can outpace the IPC overhead of getting the result data for the full repo
 from Watchman. Defaults to false.
 
 ::
 
     [fsmonitor]
     warn_when_unused = (boolean)
 
 Whether to print a warning during certain operations when fsmonitor would be
 beneficial to performance but isn't enabled.
 
 ::
 
     [fsmonitor]
     warn_update_file_count = (integer)
 
 If ``warn_when_unused`` is set and fsmonitor isn't enabled, a warning will
 be printed during working directory updates if this many files will be
 created.
 '''
 
 # Platforms Supported
 # ===================
 #
 # **Linux:** *Stable*. Watchman and fsmonitor are both known to work reliably,
 #   even under severe loads.
 #
 # **Mac OS X:** *Stable*. The Mercurial test suite passes with fsmonitor
 #   turned on, on case-insensitive HFS+. There has been a reasonable amount of
 #   user testing under normal loads.
 #
 # **Solaris, BSD:** *Alpha*. watchman and fsmonitor are believed to work, but
 #   very little testing has been done.
 #
 # **Windows:** *Alpha*. Not in a release version of watchman or fsmonitor yet.
 #
 # Known Issues
 # ============
 #
 # * fsmonitor will disable itself if any of the following extensions are
 #   enabled: largefiles, inotify, eol; or if the repository has subrepos.
 # * fsmonitor will produce incorrect results if nested repos that are not
 #   subrepos exist. *Workaround*: add nested repo paths to your `.hgignore`.
 #
 # The issues related to nested repos and subrepos are probably not fundamental
 # ones. Patches to fix them are welcome.
 
 from __future__ import absolute_import
 
 import codecs
 import hashlib
 import os
 import stat
 import sys
 import tempfile
 import weakref
 
 from mercurial.i18n import _
 from mercurial.node import (
     hex,
 )
 
 from mercurial import (
     context,
     encoding,
     error,
     extensions,
     localrepo,
     merge,
     pathutil,
     pycompat,
     registrar,
     scmutil,
     util,
 )
 from mercurial import match as matchmod
 
 from . import (
     pywatchman,
     state,
     watchmanclient,
 )
 
 # Note for extension authors: ONLY specify testedwith = 'ships-with-hg-core' for
 # extensions which SHIP WITH MERCURIAL. Non-mainline extensions should
 # be specifying the version(s) of Mercurial they are tested with, or
 # leave the attribute unspecified.
 testedwith = 'ships-with-hg-core'
 
 configtable = {}
 configitem = registrar.configitem(configtable)
 
 configitem('fsmonitor', 'mode',
     default='on',
 )
 configitem('fsmonitor', 'walk_on_invalidate',
     default=False,
 )
 configitem('fsmonitor', 'timeout',
     default='2',
 )
 configitem('fsmonitor', 'blacklistusers',
     default=list,
 )
 configitem('fsmonitor', 'watchman_exe',
     default='watchman',
 )
 configitem('fsmonitor', 'verbose',
     default=True,
+    experimental=True,
 )
 configitem('experimental', 'fsmonitor.transaction_notify',
     default=False,
 )
 
 # This extension is incompatible with the following blacklisted extensions
 # and will disable itself when encountering one of these:
 _blacklist = ['largefiles', 'eol']
 
 def debuginstall(ui, fm):
     fm.write("fsmonitor-watchman",
              _("fsmonitor checking for watchman binary... (%s)\n"),
                ui.configpath("fsmonitor", "watchman_exe"))
     root = tempfile.mkdtemp()
     c = watchmanclient.client(ui, root)
     err = None
     try:
         v = c.command("version")
         fm.write("fsmonitor-watchman-version",
                  _(" watchman binary version %s\n"), v["version"])
     except watchmanclient.Unavailable as e:
         err = str(e)
     fm.condwrite(err, "fsmonitor-watchman-error",
                  _(" watchman binary missing or broken: %s\n"), err)
     return 1 if err else 0
 
 def _handleunavailable(ui, state, ex):
     """Exception handler for Watchman interaction exceptions"""
     if isinstance(ex, watchmanclient.Unavailable):
         # experimental config: fsmonitor.verbose
         if ex.warn and ui.configbool('fsmonitor', 'verbose'):
             if 'illegal_fstypes' not in str(ex):
                 ui.warn(str(ex) + '\n')
         if ex.invalidate:
             state.invalidate()
         # experimental config: fsmonitor.verbose
         if ui.configbool('fsmonitor', 'verbose'):
             ui.log('fsmonitor', 'Watchman unavailable: %s\n', ex.msg)
     else:
         ui.log('fsmonitor', 'Watchman exception: %s\n', ex)
 
 def _hashignore(ignore):
     """Calculate hash for ignore patterns and filenames
 
     If this information changes between Mercurial invocations, we can't
     rely on Watchman information anymore and have to re-scan the working
     copy.
 
     """
     sha1 = hashlib.sha1()
     sha1.update(repr(ignore))
     return sha1.hexdigest()
 
 _watchmanencoding = pywatchman.encoding.get_local_encoding()
 _fsencoding = sys.getfilesystemencoding() or sys.getdefaultencoding()
 _fixencoding = codecs.lookup(_watchmanencoding) != codecs.lookup(_fsencoding)
 
 def _watchmantofsencoding(path):
     """Fix path to match watchman and local filesystem encoding
 
     watchman's paths encoding can differ from filesystem encoding. For example,
     on Windows, it's always utf-8.
     """
     try:
         decoded = path.decode(_watchmanencoding)
     except UnicodeDecodeError as e:
         raise error.Abort(str(e), hint='watchman encoding error')
 
     try:
         encoded = decoded.encode(_fsencoding, 'strict')
     except UnicodeEncodeError as e:
         raise error.Abort(str(e))
 
     return encoded
 
 def overridewalk(orig, self, match, subrepos, unknown, ignored, full=True):
     '''Replacement for dirstate.walk, hooking into Watchman.
 
     Whenever full is False, ignored is False, and the Watchman client is
     available, use Watchman combined with saved state to possibly return only a
     subset of files.'''
     def bail(reason):
         self._ui.debug('fsmonitor: fallback to core status, %s\n' % reason)
         return orig(match, subrepos, unknown, ignored, full=True)
 
     if full:
         return bail('full rewalk requested')
     if ignored:
         return bail('listing ignored files')
     if not self._watchmanclient.available():
         return bail('client unavailable')
     state = self._fsmonitorstate
     clock, ignorehash, notefiles = state.get()
     if not clock:
         if state.walk_on_invalidate:
             return bail('no clock')
         # Initial NULL clock value, see
         # https://facebook.github.io/watchman/docs/clockspec.html
         clock = 'c:0:0'
         notefiles = []
 
     ignore = self._ignore
     dirignore = self._dirignore
     if unknown:
         if _hashignore(ignore) != ignorehash and clock != 'c:0:0':
             # ignore list changed -- can't rely on Watchman state any more
             if state.walk_on_invalidate:
                 return bail('ignore rules changed')
             notefiles = []
             clock = 'c:0:0'
     else:
         # always ignore
         ignore = util.always
         dirignore = util.always
 
     matchfn = match.matchfn
     matchalways = match.always()
     dmap = self._map
     if util.safehasattr(dmap, '_map'):
         # for better performance, directly access the inner dirstate map if the
         # standard dirstate implementation is in use.
         dmap = dmap._map
     nonnormalset = self._map.nonnormalset
 
     copymap = self._map.copymap
     getkind = stat.S_IFMT
     dirkind = stat.S_IFDIR
     regkind = stat.S_IFREG
     lnkkind = stat.S_IFLNK
     join = self._join
     normcase = util.normcase
     fresh_instance = False
 
     exact = skipstep3 = False
     if match.isexact():  # match.exact
         exact = True
         dirignore = util.always  # skip step 2
     elif match.prefix():  # match.match, no patterns
         skipstep3 = True
 
     if not exact and self._checkcase:
         # note that even though we could receive directory entries, we're only
         # interested in checking if a file with the same name exists. So only
         # normalize files if possible.
         normalize = self._normalizefile
         skipstep3 = False
     else:
         normalize = None
 
     # step 1: find all explicit files
     results, work, dirsnotfound = self._walkexplicit(match, subrepos)
 
     skipstep3 = skipstep3 and not (work or dirsnotfound)
     work = [d for d in work if not dirignore(d[0])]
 
     if not work and (exact or skipstep3):
         for s in subrepos:
             del results[s]
         del results['.hg']
         return results
 
     # step 2: query Watchman
     try:
         # Use the user-configured timeout for the query.
         # Add a little slack over the top of the user query to allow for
         # overheads while transferring the data
         self._watchmanclient.settimeout(state.timeout + 0.1)
         result = self._watchmanclient.command('query', {
             'fields': ['mode', 'mtime', 'size', 'exists', 'name'],
             'since': clock,
             'expression': [
                 'not', [
                     'anyof', ['dirname', '.hg'],
                     ['name', '.hg', 'wholename']
                 ]
             ],
             'sync_timeout': int(state.timeout * 1000),
             'empty_on_fresh_instance': state.walk_on_invalidate,
         })
     except Exception as ex:
         _handleunavailable(self._ui, state, ex)
         self._watchmanclient.clearconnection()
         return bail('exception during run')
     else:
         # We need to propagate the last observed clock up so that we
         # can use it for our next query
         state.setlastclock(result['clock'])
         if result['is_fresh_instance']:
             if state.walk_on_invalidate:
                 state.invalidate()
                 return bail('fresh instance')
             fresh_instance = True
             # Ignore any prior noteable files from the state info
             notefiles = []
 
     # for file paths which require normalization and we encounter a case
     # collision, we store our own foldmap
     if normalize:
         foldmap = dict((normcase(k), k) for k in results)
 
     switch_slashes = pycompat.ossep == '\\'
     # The order of the results is, strictly speaking, undefined.
     # For case changes on a case insensitive filesystem we may receive
     # two entries, one with exists=True and another with exists=False.
     # The exists=True entries in the same response should be interpreted
     # as being happens-after the exists=False entries due to the way that
     # Watchman tracks files.  We use this property to reconcile deletes
     # for name case changes.
     for entry in result['files']:
         fname = entry['name']
         if _fixencoding:
             fname = _watchmantofsencoding(fname)
         if switch_slashes:
             fname = fname.replace('\\', '/')
         if normalize:
             normed = normcase(fname)
             fname = normalize(fname, True, True)
             foldmap[normed] = fname
         fmode = entry['mode']
         fexists = entry['exists']
         kind = getkind(fmode)
 
         if '/.hg/' in fname or fname.endswith('/.hg'):
             return bail('nested-repo-detected')
 
         if not fexists:
             # if marked as deleted and we don't already have a change
             # record, mark it as deleted.  If we already have an entry
             # for fname then it was either part of walkexplicit or was
             # an earlier result that was a case change
             if fname not in results and fname in dmap and (
                     matchalways or matchfn(fname)):
                 results[fname] = None
         elif kind == dirkind:
             if fname in dmap and (matchalways or matchfn(fname)):
                 results[fname] = None
         elif kind == regkind or kind == lnkkind:
             if fname in dmap:
                 if matchalways or matchfn(fname):
                     results[fname] = entry
             elif (matchalways or matchfn(fname)) and not ignore(fname):
                 results[fname] = entry
         elif fname in dmap and (matchalways or matchfn(fname)):
             results[fname] = None
 
     # step 3: query notable files we don't already know about
     # XXX try not to iterate over the entire dmap
     if normalize:
         # any notable files that have changed case will already be handled
         # above, so just check membership in the foldmap
         notefiles = set((normalize(f, True, True) for f in notefiles
                          if normcase(f) not in foldmap))
     visit = set((f for f in notefiles if (f not in results and matchfn(f)
                                           and (f in dmap or not ignore(f)))))
 
     if not fresh_instance:
         if matchalways:
             visit.update(f for f in nonnormalset if f not in results)
             visit.update(f for f in copymap if f not in results)
         else:
             visit.update(f for f in nonnormalset
                          if f not in results and matchfn(f))
             visit.update(f for f in copymap
                          if f not in results and matchfn(f))
     else:
         if matchalways:
             visit.update(f for f, st in dmap.iteritems() if f not in results)
             visit.update(f for f in copymap if f not in results)
         else:
             visit.update(f for f, st in dmap.iteritems()
                          if f not in results and matchfn(f))
             visit.update(f for f in copymap
                          if f not in results and matchfn(f))
 
     audit = pathutil.pathauditor(self._root, cached=True).check
     auditpass = [f for f in visit if audit(f)]
     auditpass.sort()
     auditfail = visit.difference(auditpass)
     for f in auditfail:
         results[f] = None
 
     nf = iter(auditpass).next
     for st in util.statfiles([join(f) for f in auditpass]):
         f = nf()
         if st or f in dmap:
             results[f] = st
 
     for s in subrepos:
         del results[s]
     del results['.hg']
     return results
 
 def overridestatus(
         orig, self, node1='.', node2=None, match=None, ignored=False,
         clean=False, unknown=False, listsubrepos=False):
     listignored = ignored
     listclean = clean
     listunknown = unknown
 
     def _cmpsets(l1, l2):
         try:
             if 'FSMONITOR_LOG_FILE' in encoding.environ:
                 fn = encoding.environ['FSMONITOR_LOG_FILE']
                 f = open(fn, 'wb')
             else:
                 fn = 'fsmonitorfail.log'
                 f = self.vfs.open(fn, 'wb')
         except (IOError, OSError):
             self.ui.warn(_('warning: unable to write to %s\n') % fn)
             return
 
         try:
             for i, (s1, s2) in enumerate(zip(l1, l2)):
                 if set(s1) != set(s2):
                     f.write('sets at position %d are unequal\n' % i)
                     f.write('watchman returned: %s\n' % s1)
                     f.write('stat returned: %s\n' % s2)
         finally:
             f.close()
 
     if isinstance(node1, context.changectx):
         ctx1 = node1
     else:
         ctx1 = self[node1]
     if isinstance(node2, context.changectx):
         ctx2 = node2
     else:
         ctx2 = self[node2]
 
     working = ctx2.rev() is None
     parentworking = working and ctx1 == self['.']
     match = match or matchmod.always()
 
     # Maybe we can use this opportunity to update Watchman's state.
     # Mercurial uses workingcommitctx and/or memctx to represent the part of
     # the workingctx that is to be committed. So don't update the state in
     # that case.
     # HG_PENDING is set in the environment when the dirstate is being updated
     # in the middle of a transaction; we must not update our state in that
     # case, or we risk forgetting about changes in the working copy.
     updatestate = (parentworking and match.always() and
                    not isinstance(ctx2, (context.workingcommitctx,
                                          context.memctx)) and
                    'HG_PENDING' not in encoding.environ)
 
     try:
         if self._fsmonitorstate.walk_on_invalidate:
             # Use a short timeout to query the current clock.  If that
             # takes too long then we assume that the service will be slow
             # to answer our query.
             # walk_on_invalidate indicates that we prefer to walk the
             # tree ourselves because we can ignore portions that Watchman
             # cannot and we tend to be faster in the warmer buffer cache
             # cases.
             self._watchmanclient.settimeout(0.1)
         else:
             # Give Watchman more time to potentially complete its walk
             # and return the initial clock.  In this mode we assume that
             # the filesystem will be slower than parsing a potentially
             # very large Watchman result set.
             self._watchmanclient.settimeout(
                 self._fsmonitorstate.timeout + 0.1)
         startclock = self._watchmanclient.getcurrentclock()
     except Exception as ex:
         self._watchmanclient.clearconnection()
         _handleunavailable(self.ui, self._fsmonitorstate, ex)
         # boo, Watchman failed. bail
         return orig(node1, node2, match, listignored, listclean,
                     listunknown, listsubrepos)
 
     if updatestate:
         # We need info about unknown files. This may make things slower the
         # first time, but whatever.
         stateunknown = True
     else:
         stateunknown = listunknown
 
     if updatestate:
         ps = poststatus(startclock)
         self.addpostdsstatus(ps)
 
     r = orig(node1, node2, match, listignored, listclean, stateunknown,
              listsubrepos)
     modified, added, removed, deleted, unknown, ignored, clean = r
 
     if not listunknown:
         unknown = []
 
     # don't do paranoid checks if we're not going to query Watchman anyway
     full = listclean or match.traversedir is not None
     if self._fsmonitorstate.mode == 'paranoid' and not full:
         # run status again and fall back to the old walk this time
         self.dirstate._fsmonitordisable = True
 
         # shut the UI up
         quiet = self.ui.quiet
         self.ui.quiet = True
         fout, ferr = self.ui.fout, self.ui.ferr
         self.ui.fout = self.ui.ferr = open(os.devnull, 'wb')
 
         try:
             rv2 = orig(
                 node1, node2, match, listignored, listclean, listunknown,
                 listsubrepos)
         finally:
             self.dirstate._fsmonitordisable = False
             self.ui.quiet = quiet
             self.ui.fout, self.ui.ferr = fout, ferr
 
         # clean isn't tested since it's set to True above
         with self.wlock():
             _cmpsets(
                 [modified, added, removed, deleted, unknown, ignored, clean],
                 rv2)
         modified, added, removed, deleted, unknown, ignored, clean = rv2
 
     return scmutil.status(
         modified, added, removed, deleted, unknown, ignored, clean)
 
 class poststatus(object):
     def __init__(self, startclock):
         self._startclock = startclock
 
     def __call__(self, wctx, status):
         clock = wctx.repo()._fsmonitorstate.getlastclock() or self._startclock
         hashignore = _hashignore(wctx.repo().dirstate._ignore)
         notefiles = (status.modified + status.added + status.removed +
                      status.deleted + status.unknown)
         wctx.repo()._fsmonitorstate.set(clock, hashignore, notefiles)
 
 def makedirstate(repo, dirstate):
     class fsmonitordirstate(dirstate.__class__):
         def _fsmonitorinit(self, repo):
             # _fsmonitordisable is used in paranoid mode
             self._fsmonitordisable = False
             self._fsmonitorstate = repo._fsmonitorstate
             self._watchmanclient = repo._watchmanclient
             self._repo = weakref.proxy(repo)
 
         def walk(self, *args, **kwargs):
             orig = super(fsmonitordirstate, self).walk
             if self._fsmonitordisable:
                 return orig(*args, **kwargs)
             return overridewalk(orig, self, *args, **kwargs)
 
         def rebuild(self, *args, **kwargs):
             self._fsmonitorstate.invalidate()
             return super(fsmonitordirstate, self).rebuild(*args, **kwargs)
 
         def invalidate(self, *args, **kwargs):
             self._fsmonitorstate.invalidate()
             return super(fsmonitordirstate, self).invalidate(*args, **kwargs)
 
     dirstate.__class__ = fsmonitordirstate
     dirstate._fsmonitorinit(repo)
 
 def wrapdirstate(orig, self):
     ds = orig(self)
     # only override the dirstate when Watchman is available for the repo
     if util.safehasattr(self, '_fsmonitorstate'):
         makedirstate(self, ds)
     return ds
 
 def extsetup(ui):
     extensions.wrapfilecache(
         localrepo.localrepository, 'dirstate', wrapdirstate)
     if pycompat.isdarwin:
         # An assist for avoiding the dangling-symlink fsevents bug
         extensions.wrapfunction(os, 'symlink', wrapsymlink)
 
     extensions.wrapfunction(merge, 'update', wrapupdate)
 
 def wrapsymlink(orig, source, link_name):
     ''' if we create a dangling symlink, also touch the parent dir
     to encourage fsevents notifications to work more correctly '''
     try:
         return orig(source, link_name)
     finally:
         try:
             os.utime(os.path.dirname(link_name), None)
         except OSError:
             pass
 
 class state_update(object):
     ''' This context manager is responsible for dispatching the state-enter
         and state-leave signals to the watchman service. The enter and leave
         methods can be invoked manually (for scenarios where context manager
         semantics are not possible). If parameters oldnode and newnode are None,
         they will be populated based on current working copy in enter and
         leave, respectively. Similarly, if the distance is none, it will be
         calculated based on the oldnode and newnode in the leave method.'''
 
     def __init__(self, repo, name, oldnode=None, newnode=None, distance=None,
                  partial=False):
         self.repo = repo.unfiltered()
         self.name = name
         self.oldnode = oldnode
         self.newnode = newnode
         self.distance = distance
         self.partial = partial
         self._lock = None
         self.need_leave = False
 
     def __enter__(self):
         self.enter()
 
     def enter(self):
         # Make sure we have a wlock prior to sending notifications to watchman.
         # We don't want to race with other actors. In the update case,
         # merge.update is going to take the wlock almost immediately. We are
         # effectively extending the lock around several short sanity checks.
         if self.oldnode is None:
             self.oldnode = self.repo['.'].node()
 
         if self.repo.currentwlock() is None:
             if util.safehasattr(self.repo, 'wlocknostateupdate'):
                 self._lock = self.repo.wlocknostateupdate()
             else:
                 self._lock = self.repo.wlock()
         self.need_leave = self._state(
             'state-enter',
             hex(self.oldnode))
         return self
 
     def __exit__(self, type_, value, tb):
         abort = True if type_ else False
         self.exit(abort=abort)
 
     def exit(self, abort=False):
         try:
             if self.need_leave:
                 status = 'failed' if abort else 'ok'
                 if self.newnode is None:
                     self.newnode = self.repo['.'].node()
                 if self.distance is None:
                     self.distance = calcdistance(
                         self.repo, self.oldnode, self.newnode)
                 self._state(
                     'state-leave',
                     hex(self.newnode),
                     status=status)
         finally:
             self.need_leave = False
             if self._lock:
                 self._lock.release()
 
     def _state(self, cmd, commithash, status='ok'):
         if not util.safehasattr(self.repo, '_watchmanclient'):
             return False
         try:
             self.repo._watchmanclient.command(cmd, {
                 'name': self.name,
                 'metadata': {
                     # the target revision
                     'rev': commithash,
                     # approximate number of commits between current and target
                     'distance': self.distance if self.distance else 0,
                     # success/failure (only really meaningful for state-leave)
                     'status': status,
                     # whether the working copy parent is changing
                     'partial': self.partial,
             }})
             return True
         except Exception as e:
             # Swallow any errors; fire and forget
             self.repo.ui.log(
                 'watchman', 'Exception %s while running %s\n', e, cmd)
             return False
 
 # Estimate the distance between two nodes
 def calcdistance(repo, oldnode, newnode):
     anc = repo.changelog.ancestor(oldnode, newnode)
     ancrev = repo[anc].rev()
     distance = (abs(repo[oldnode].rev() - ancrev)
         + abs(repo[newnode].rev() - ancrev))
     return distance
 
 # Bracket working copy updates with calls to the watchman state-enter
 # and state-leave commands.  This allows clients to perform more intelligent
 # settling during bulk file change scenarios
 # https://facebook.github.io/watchman/docs/cmd/subscribe.html#advanced-settling
 def wrapupdate(orig, repo, node, branchmerge, force, ancestor=None,
                mergeancestor=False, labels=None, matcher=None, **kwargs):
 
     distance = 0
     partial = True
     oldnode = repo['.'].node()
     newnode = repo[node].node()
     if matcher is None or matcher.always():
         partial = False
         distance = calcdistance(repo.unfiltered(), oldnode, newnode)
 
     with state_update(repo, name="hg.update", oldnode=oldnode, newnode=newnode,
                       distance=distance, partial=partial):
         return orig(
             repo, node, branchmerge, force, ancestor, mergeancestor,
             labels, matcher, **kwargs)
 
 def repo_has_depth_one_nested_repo(repo):
     for f in repo.wvfs.listdir():
         if os.path.isdir(os.path.join(repo.root, f, '.hg')):
             msg = 'fsmonitor: sub-repository %r detected, fsmonitor disabled\n'
             repo.ui.debug(msg % f)
             return True
     return False
 
 def reposetup(ui, repo):
     # We don't work with largefiles or inotify
     exts = extensions.enabled()
     for ext in _blacklist:
         if ext in exts:
             ui.warn(_('The fsmonitor extension is incompatible with the %s '
                       'extension and has been disabled.\n') % ext)
             return
 
     if repo.local():
         # We don't work with subrepos either.
         #
         # if repo[None].substate can cause a dirstate parse, which is too
         # slow. Instead, look for a file called hgsubstate,
         if repo.wvfs.exists('.hgsubstate') or repo.wvfs.exists('.hgsub'):
             return
 
         if repo_has_depth_one_nested_repo(repo):
             return
 
         fsmonitorstate = state.state(repo)
         if fsmonitorstate.mode == 'off':
             return
 
         try:
             client = watchmanclient.client(repo.ui, repo._root)
         except Exception as ex:
             _handleunavailable(ui, fsmonitorstate, ex)
             return
 
         repo._fsmonitorstate = fsmonitorstate
         repo._watchmanclient = client
 
         dirstate, cached = localrepo.isfilecached(repo, 'dirstate')
         if cached:
             # at this point since fsmonitorstate wasn't present,
             # repo.dirstate is not a fsmonitordirstate
             makedirstate(repo, dirstate)
 
         class fsmonitorrepo(repo.__class__):
             def status(self, *args, **kwargs):
                 orig = super(fsmonitorrepo, self).status
                 return overridestatus(orig, self, *args, **kwargs)
 
             def wlocknostateupdate(self, *args, **kwargs):
                 return super(fsmonitorrepo, self).wlock(*args, **kwargs)
 
             def wlock(self, *args, **kwargs):
                 l = super(fsmonitorrepo, self).wlock(*args, **kwargs)
                 if not ui.configbool(
                     "experimental", "fsmonitor.transaction_notify"):
                     return l
                 if l.held != 1:
                     return l
                 origrelease = l.releasefn
 
                 def staterelease():
                     if origrelease:
                         origrelease()
                     if l.stateupdate:
                         l.stateupdate.exit()
                         l.stateupdate = None
 
                 try:
                     l.stateupdate = None
                     l.stateupdate = state_update(self, name="hg.transaction")
                     l.stateupdate.enter()
                     l.releasefn = staterelease
                 except Exception as e:
                     # Swallow any errors; fire and forget
                     self.ui.log(
                         'watchman', 'Exception in state update %s\n', e)
                 return l
 
         repo.__class__ = fsmonitorrepo

hgext/remotefilelog/__init__.py

@@ -1,1113 +1,1113 b''
 # __init__.py - remotefilelog extension
 #
 # Copyright 2013 Facebook, Inc.
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 """remotefilelog causes Mercurial to lazilly fetch file contents (EXPERIMENTAL)
 
 This extension is HIGHLY EXPERIMENTAL. There are NO BACKWARDS COMPATIBILITY
 GUARANTEES. This means that repositories created with this extension may
 only be usable with the exact version of this extension/Mercurial that was
 used. The extension attempts to enforce this in order to prevent repository
 corruption.
 
 remotefilelog works by fetching file contents lazily and storing them
 in a cache on the client rather than in revlogs. This allows enormous
 histories to be transferred only partially, making them easier to
 operate on.
 
 Configs:
 
     ``packs.maxchainlen`` specifies the maximum delta chain length in pack files
 
     ``packs.maxpacksize`` specifies the maximum pack file size
 
     ``packs.maxpackfilecount`` specifies the maximum number of packs in the
       shared cache (trees only for now)
 
     ``remotefilelog.backgroundprefetch`` runs prefetch in background when True
 
     ``remotefilelog.bgprefetchrevs`` specifies revisions to fetch on commit and
       update, and on other commands that use them. Different from pullprefetch.
 
     ``remotefilelog.gcrepack`` does garbage collection during repack when True
 
     ``remotefilelog.nodettl`` specifies maximum TTL of a node in seconds before
       it is garbage collected
 
     ``remotefilelog.repackonhggc`` runs repack on hg gc when True
 
     ``remotefilelog.prefetchdays`` specifies the maximum age of a commit in
       days after which it is no longer prefetched.
 
     ``remotefilelog.prefetchdelay`` specifies delay between background
       prefetches in seconds after operations that change the working copy parent
 
     ``remotefilelog.data.gencountlimit`` constraints the minimum number of data
       pack files required to be considered part of a generation. In particular,
       minimum number of packs files > gencountlimit.
 
     ``remotefilelog.data.generations`` list for specifying the lower bound of
       each generation of the data pack files. For example, list ['100MB','1MB']
       or ['1MB', '100MB'] will lead to three generations: [0, 1MB), [
       1MB, 100MB) and [100MB, infinity).
 
     ``remotefilelog.data.maxrepackpacks`` the maximum number of pack files to
       include in an incremental data repack.
 
     ``remotefilelog.data.repackmaxpacksize`` the maximum size of a pack file for
       it to be considered for an incremental data repack.
 
     ``remotefilelog.data.repacksizelimit`` the maximum total size of pack files
       to include in an incremental data repack.
 
     ``remotefilelog.history.gencountlimit`` constraints the minimum number of
       history pack files required to be considered part of a generation. In
       particular, minimum number of packs files > gencountlimit.
 
     ``remotefilelog.history.generations`` list for specifying the lower bound of
       each generation of the history pack files. For example, list [
       '100MB', '1MB'] or ['1MB', '100MB'] will lead to three generations: [
       0, 1MB), [1MB, 100MB) and [100MB, infinity).
 
     ``remotefilelog.history.maxrepackpacks`` the maximum number of pack files to
       include in an incremental history repack.
 
     ``remotefilelog.history.repackmaxpacksize`` the maximum size of a pack file
       for it to be considered for an incremental history repack.
 
     ``remotefilelog.history.repacksizelimit`` the maximum total size of pack
       files to include in an incremental history repack.
 
     ``remotefilelog.backgroundrepack`` automatically consolidate packs in the
       background
 
     ``remotefilelog.cachepath`` path to cache
 
     ``remotefilelog.cachegroup`` if set, make cache directory sgid to this
       group
 
     ``remotefilelog.cacheprocess`` binary to invoke for fetching file data
 
     ``remotefilelog.debug`` turn on remotefilelog-specific debug output
 
     ``remotefilelog.excludepattern`` pattern of files to exclude from pulls
 
     ``remotefilelog.includepattern`` pattern of files to include in pulls
 
     ``remotefilelog.fetchwarning``: message to print when too many
       single-file fetches occur
 
     ``remotefilelog.getfilesstep`` number of files to request in a single RPC
 
     ``remotefilelog.getfilestype`` if set to 'threaded' use threads to fetch
       files, otherwise use optimistic fetching
 
     ``remotefilelog.pullprefetch`` revset for selecting files that should be
       eagerly downloaded rather than lazily
 
     ``remotefilelog.reponame`` name of the repo. If set, used to partition
       data from other repos in a shared store.
 
     ``remotefilelog.server`` if true, enable server-side functionality
 
     ``remotefilelog.servercachepath`` path for caching blobs on the server
 
     ``remotefilelog.serverexpiration`` number of days to keep cached server
       blobs
 
     ``remotefilelog.validatecache`` if set, check cache entries for corruption
       before returning blobs
 
     ``remotefilelog.validatecachelog`` if set, check cache entries for
       corruption before returning metadata
 
 """
 from __future__ import absolute_import
 
 import os
 import time
 import traceback
 
 from mercurial.node import hex
 from mercurial.i18n import _
 from mercurial import (
     changegroup,
     changelog,
     cmdutil,
     commands,
     configitems,
     context,
     copies,
     debugcommands as hgdebugcommands,
     dispatch,
     error,
     exchange,
     extensions,
     hg,
     localrepo,
     match,
     merge,
     node as nodemod,
     patch,
     pycompat,
     registrar,
     repair,
     repoview,
     revset,
     scmutil,
     smartset,
     streamclone,
     util,
 )
 from . import (
     constants,
     debugcommands,
     fileserverclient,
     remotefilectx,
     remotefilelog,
     remotefilelogserver,
     repack as repackmod,
     shallowbundle,
     shallowrepo,
     shallowstore,
     shallowutil,
     shallowverifier,
 )
 
 # ensures debug commands are registered
 hgdebugcommands.command
 
 cmdtable = {}
 command = registrar.command(cmdtable)
 
 configtable = {}
 configitem = registrar.configitem(configtable)
 
 configitem('remotefilelog', 'debug', default=False)
 
 configitem('remotefilelog', 'reponame', default='')
 configitem('remotefilelog', 'cachepath', default=None)
 configitem('remotefilelog', 'cachegroup', default=None)
 configitem('remotefilelog', 'cacheprocess', default=None)
 configitem('remotefilelog', 'cacheprocess.includepath', default=None)
 configitem("remotefilelog", "cachelimit", default="1000 GB")
 
 configitem('remotefilelog', 'fallbackpath', default=configitems.dynamicdefault,
            alias=[('remotefilelog', 'fallbackrepo')])
 
 configitem('remotefilelog', 'validatecachelog', default=None)
 configitem('remotefilelog', 'validatecache', default='on')
 configitem('remotefilelog', 'server', default=None)
 configitem('remotefilelog', 'servercachepath', default=None)
 configitem("remotefilelog", "serverexpiration", default=30)
 configitem('remotefilelog', 'backgroundrepack', default=False)
 configitem('remotefilelog', 'bgprefetchrevs', default=None)
 configitem('remotefilelog', 'pullprefetch', default=None)
 configitem('remotefilelog', 'backgroundprefetch', default=False)
 configitem('remotefilelog', 'prefetchdelay', default=120)
 configitem('remotefilelog', 'prefetchdays', default=14)
 
 configitem('remotefilelog', 'getfilesstep', default=10000)
 configitem('remotefilelog', 'getfilestype', default='optimistic')
 configitem('remotefilelog', 'batchsize', configitems.dynamicdefault)
 configitem('remotefilelog', 'fetchwarning', default='')
 
 configitem('remotefilelog', 'includepattern', default=None)
 configitem('remotefilelog', 'excludepattern', default=None)
 
 configitem('remotefilelog', 'gcrepack', default=False)
 configitem('remotefilelog', 'repackonhggc', default=False)
-configitem('repack', 'chainorphansbysize', default=True)
+configitem('repack', 'chainorphansbysize', default=True, experimental=True)
 
 configitem('packs', 'maxpacksize', default=0)
 configitem('packs', 'maxchainlen', default=1000)
 
 #  default TTL limit is 30 days
 _defaultlimit = 60 * 60 * 24 * 30
 configitem('remotefilelog', 'nodettl', default=_defaultlimit)
 
 configitem('remotefilelog', 'data.gencountlimit', default=2),
 configitem('remotefilelog', 'data.generations',
            default=['1GB', '100MB', '1MB'])
 configitem('remotefilelog', 'data.maxrepackpacks', default=50)
 configitem('remotefilelog', 'data.repackmaxpacksize', default='4GB')
 configitem('remotefilelog', 'data.repacksizelimit', default='100MB')
 
 configitem('remotefilelog', 'history.gencountlimit', default=2),
 configitem('remotefilelog', 'history.generations', default=['100MB'])
 configitem('remotefilelog', 'history.maxrepackpacks', default=50)
 configitem('remotefilelog', 'history.repackmaxpacksize', default='400MB')
 configitem('remotefilelog', 'history.repacksizelimit', default='100MB')
 
 # Note for extension authors: ONLY specify testedwith = 'ships-with-hg-core' for
 # extensions which SHIP WITH MERCURIAL. Non-mainline extensions should
 # be specifying the version(s) of Mercurial they are tested with, or
 # leave the attribute unspecified.
 testedwith = 'ships-with-hg-core'
 
 repoclass = localrepo.localrepository
 repoclass._basesupported.add(constants.SHALLOWREPO_REQUIREMENT)
 
 isenabled = shallowutil.isenabled
 
 def uisetup(ui):
     """Wraps user facing Mercurial commands to swap them out with shallow
     versions.
     """
     hg.wirepeersetupfuncs.append(fileserverclient.peersetup)
 
     entry = extensions.wrapcommand(commands.table, 'clone', cloneshallow)
     entry[1].append(('', 'shallow', None,
                      _("create a shallow clone which uses remote file "
                        "history")))
 
     extensions.wrapcommand(commands.table, 'debugindex',
         debugcommands.debugindex)
     extensions.wrapcommand(commands.table, 'debugindexdot',
         debugcommands.debugindexdot)
     extensions.wrapcommand(commands.table, 'log', log)
     extensions.wrapcommand(commands.table, 'pull', pull)
 
     # Prevent 'hg manifest --all'
     def _manifest(orig, ui, repo, *args, **opts):
         if (isenabled(repo) and opts.get(r'all')):
             raise error.Abort(_("--all is not supported in a shallow repo"))
 
         return orig(ui, repo, *args, **opts)
     extensions.wrapcommand(commands.table, "manifest", _manifest)
 
     # Wrap remotefilelog with lfs code
     def _lfsloaded(loaded=False):
         lfsmod = None
         try:
             lfsmod = extensions.find('lfs')
         except KeyError:
             pass
         if lfsmod:
             lfsmod.wrapfilelog(remotefilelog.remotefilelog)
             fileserverclient._lfsmod = lfsmod
     extensions.afterloaded('lfs', _lfsloaded)
 
     # debugdata needs remotefilelog.len to work
     extensions.wrapcommand(commands.table, 'debugdata', debugdatashallow)
 
     changegroup.cgpacker = shallowbundle.shallowcg1packer
 
     extensions.wrapfunction(changegroup, '_addchangegroupfiles',
                             shallowbundle.addchangegroupfiles)
     extensions.wrapfunction(
         changegroup, 'makechangegroup', shallowbundle.makechangegroup)
     extensions.wrapfunction(localrepo, 'makestore', storewrapper)
     extensions.wrapfunction(exchange, 'pull', exchangepull)
     extensions.wrapfunction(merge, 'applyupdates', applyupdates)
     extensions.wrapfunction(merge, '_checkunknownfiles', checkunknownfiles)
     extensions.wrapfunction(context.workingctx, '_checklookup', checklookup)
     extensions.wrapfunction(scmutil, '_findrenames', findrenames)
     extensions.wrapfunction(copies, '_computeforwardmissing',
                             computeforwardmissing)
     extensions.wrapfunction(dispatch, 'runcommand', runcommand)
     extensions.wrapfunction(repair, '_collectbrokencsets', _collectbrokencsets)
     extensions.wrapfunction(context.changectx, 'filectx', filectx)
     extensions.wrapfunction(context.workingctx, 'filectx', workingfilectx)
     extensions.wrapfunction(patch, 'trydiff', trydiff)
     extensions.wrapfunction(hg, 'verify', _verify)
     scmutil.fileprefetchhooks.add('remotefilelog', _fileprefetchhook)
 
     # disappointing hacks below
     extensions.wrapfunction(scmutil, 'getrenamedfn', getrenamedfn)
     extensions.wrapfunction(revset, 'filelog', filelogrevset)
     revset.symbols['filelog'] = revset.filelog
     extensions.wrapfunction(cmdutil, 'walkfilerevs', walkfilerevs)
 
 
 def cloneshallow(orig, ui, repo, *args, **opts):
     if opts.get(r'shallow'):
         repos = []
         def pull_shallow(orig, self, *args, **kwargs):
             if not isenabled(self):
                 repos.append(self.unfiltered())
                 # set up the client hooks so the post-clone update works
                 setupclient(self.ui, self.unfiltered())
 
                 # setupclient fixed the class on the repo itself
                 # but we also need to fix it on the repoview
                 if isinstance(self, repoview.repoview):
                     self.__class__.__bases__ = (self.__class__.__bases__[0],
                                                 self.unfiltered().__class__)
                 self.requirements.add(constants.SHALLOWREPO_REQUIREMENT)
                 self._writerequirements()
 
                 # Since setupclient hadn't been called, exchange.pull was not
                 # wrapped. So we need to manually invoke our version of it.
                 return exchangepull(orig, self, *args, **kwargs)
             else:
                 return orig(self, *args, **kwargs)
         extensions.wrapfunction(exchange, 'pull', pull_shallow)
 
         # Wrap the stream logic to add requirements and to pass include/exclude
         # patterns around.
         def setup_streamout(repo, remote):
             # Replace remote.stream_out with a version that sends file
             # patterns.
             def stream_out_shallow(orig):
                 caps = remote.capabilities()
                 if constants.NETWORK_CAP_LEGACY_SSH_GETFILES in caps:
                     opts = {}
                     if repo.includepattern:
                         opts[r'includepattern'] = '\0'.join(repo.includepattern)
                     if repo.excludepattern:
                         opts[r'excludepattern'] = '\0'.join(repo.excludepattern)
                     return remote._callstream('stream_out_shallow', **opts)
                 else:
                     return orig()
             extensions.wrapfunction(remote, 'stream_out', stream_out_shallow)
         def stream_wrap(orig, op):
             setup_streamout(op.repo, op.remote)
             return orig(op)
         extensions.wrapfunction(
             streamclone, 'maybeperformlegacystreamclone', stream_wrap)
 
         def canperformstreamclone(orig, pullop, bundle2=False):
             # remotefilelog is currently incompatible with the
             # bundle2 flavor of streamclones, so force us to use
             # v1 instead.
             if 'v2' in pullop.remotebundle2caps.get('stream', []):
                 pullop.remotebundle2caps['stream'] = [
                     c for c in pullop.remotebundle2caps['stream']
                     if c != 'v2']
             if bundle2:
                 return False, None
             supported, requirements = orig(pullop, bundle2=bundle2)
             if requirements is not None:
                 requirements.add(constants.SHALLOWREPO_REQUIREMENT)
             return supported, requirements
         extensions.wrapfunction(
             streamclone, 'canperformstreamclone', canperformstreamclone)
 
     try:
         orig(ui, repo, *args, **opts)
     finally:
         if opts.get(r'shallow'):
             for r in repos:
                 if util.safehasattr(r, 'fileservice'):
                     r.fileservice.close()
 
 def debugdatashallow(orig, *args, **kwds):
     oldlen = remotefilelog.remotefilelog.__len__
     try:
         remotefilelog.remotefilelog.__len__ = lambda x: 1
         return orig(*args, **kwds)
     finally:
         remotefilelog.remotefilelog.__len__ = oldlen
 
 def reposetup(ui, repo):
     if not repo.local():
         return
 
     # put here intentionally bc doesnt work in uisetup
     ui.setconfig('hooks', 'update.prefetch', wcpprefetch)
     ui.setconfig('hooks', 'commit.prefetch', wcpprefetch)
 
     isserverenabled = ui.configbool('remotefilelog', 'server')
     isshallowclient = isenabled(repo)
 
     if isserverenabled and isshallowclient:
         raise RuntimeError("Cannot be both a server and shallow client.")
 
     if isshallowclient:
         setupclient(ui, repo)
 
     if isserverenabled:
         remotefilelogserver.setupserver(ui, repo)
 
 def setupclient(ui, repo):
     if not isinstance(repo, localrepo.localrepository):
         return
 
     # Even clients get the server setup since they need to have the
     # wireprotocol endpoints registered.
     remotefilelogserver.onetimesetup(ui)
     onetimeclientsetup(ui)
 
     shallowrepo.wraprepo(repo)
     repo.store = shallowstore.wrapstore(repo.store)
 
 def storewrapper(orig, requirements, path, vfstype):
     s = orig(requirements, path, vfstype)
     if constants.SHALLOWREPO_REQUIREMENT in requirements:
         s = shallowstore.wrapstore(s)
 
     return s
 
 # prefetch files before update
 def applyupdates(orig, repo, actions, wctx, mctx, overwrite, wantfiledata,
                  labels=None):
     if isenabled(repo):
         manifest = mctx.manifest()
         files = []
         for f, args, msg in actions['g']:
             files.append((f, hex(manifest[f])))
         # batch fetch the needed files from the server
         repo.fileservice.prefetch(files)
     return orig(repo, actions, wctx, mctx, overwrite, wantfiledata,
                 labels=labels)
 
 # Prefetch merge checkunknownfiles
 def checkunknownfiles(orig, repo, wctx, mctx, force, actions,
     *args, **kwargs):
     if isenabled(repo):
         files = []
         sparsematch = repo.maybesparsematch(mctx.rev())
         for f, (m, actionargs, msg) in actions.iteritems():
             if sparsematch and not sparsematch(f):
                 continue
             if m in ('c', 'dc', 'cm'):
                 files.append((f, hex(mctx.filenode(f))))
             elif m == 'dg':
                 f2 = actionargs[0]
                 files.append((f2, hex(mctx.filenode(f2))))
         # batch fetch the needed files from the server
         repo.fileservice.prefetch(files)
     return orig(repo, wctx, mctx, force, actions, *args, **kwargs)
 
 # Prefetch files before status attempts to look at their size and contents
 def checklookup(orig, self, files):
     repo = self._repo
     if isenabled(repo):
         prefetchfiles = []
         for parent in self._parents:
             for f in files:
                 if f in parent:
                     prefetchfiles.append((f, hex(parent.filenode(f))))
         # batch fetch the needed files from the server
         repo.fileservice.prefetch(prefetchfiles)
     return orig(self, files)
 
 # Prefetch the logic that compares added and removed files for renames
 def findrenames(orig, repo, matcher, added, removed, *args, **kwargs):
     if isenabled(repo):
         files = []
         pmf = repo['.'].manifest()
         for f in removed:
             if f in pmf:
                 files.append((f, hex(pmf[f])))
         # batch fetch the needed files from the server
         repo.fileservice.prefetch(files)
     return orig(repo, matcher, added, removed, *args, **kwargs)
 
 # prefetch files before pathcopies check
 def computeforwardmissing(orig, a, b, match=None):
     missing = orig(a, b, match=match)
     repo = a._repo
     if isenabled(repo):
         mb = b.manifest()
 
         files = []
         sparsematch = repo.maybesparsematch(b.rev())
         if sparsematch:
             sparsemissing = set()
             for f in missing:
                 if sparsematch(f):
                     files.append((f, hex(mb[f])))
                     sparsemissing.add(f)
             missing = sparsemissing
 
         # batch fetch the needed files from the server
         repo.fileservice.prefetch(files)
     return missing
 
 # close cache miss server connection after the command has finished
 def runcommand(orig, lui, repo, *args, **kwargs):
     fileservice = None
     # repo can be None when running in chg:
     # - at startup, reposetup was called because serve is not norepo
     # - a norepo command like "help" is called
     if repo and isenabled(repo):
         fileservice = repo.fileservice
     try:
         return orig(lui, repo, *args, **kwargs)
     finally:
         if fileservice:
             fileservice.close()
 
 # prevent strip from stripping remotefilelogs
 def _collectbrokencsets(orig, repo, files, striprev):
     if isenabled(repo):
         files = list([f for f in files if not repo.shallowmatch(f)])
     return orig(repo, files, striprev)
 
 # changectx wrappers
 def filectx(orig, self, path, fileid=None, filelog=None):
     if fileid is None:
         fileid = self.filenode(path)
     if (isenabled(self._repo) and self._repo.shallowmatch(path)):
         return remotefilectx.remotefilectx(self._repo, path, fileid=fileid,
                                            changectx=self, filelog=filelog)
     return orig(self, path, fileid=fileid, filelog=filelog)
 
 def workingfilectx(orig, self, path, filelog=None):
     if (isenabled(self._repo) and self._repo.shallowmatch(path)):
         return remotefilectx.remoteworkingfilectx(self._repo, path,
                                                   workingctx=self,
                                                   filelog=filelog)
     return orig(self, path, filelog=filelog)
 
 # prefetch required revisions before a diff
 def trydiff(orig, repo, revs, ctx1, ctx2, modified, added, removed,
     copy, getfilectx, *args, **kwargs):
     if isenabled(repo):
         prefetch = []
         mf1 = ctx1.manifest()
         for fname in modified + added + removed:
             if fname in mf1:
                 fnode = getfilectx(fname, ctx1).filenode()
                 # fnode can be None if it's a edited working ctx file
                 if fnode:
                     prefetch.append((fname, hex(fnode)))
             if fname not in removed:
                 fnode = getfilectx(fname, ctx2).filenode()
                 if fnode:
                     prefetch.append((fname, hex(fnode)))
 
         repo.fileservice.prefetch(prefetch)
 
     return orig(repo, revs, ctx1, ctx2, modified, added, removed, copy,
                 getfilectx, *args, **kwargs)
 
 # Prevent verify from processing files
 # a stub for mercurial.hg.verify()
 def _verify(orig, repo, level=None):
     lock = repo.lock()
     try:
         return shallowverifier.shallowverifier(repo).verify()
     finally:
         lock.release()
 
 
 clientonetime = False
 def onetimeclientsetup(ui):
     global clientonetime
     if clientonetime:
         return
     clientonetime = True
 
     # Don't commit filelogs until we know the commit hash, since the hash
     # is present in the filelog blob.
     # This violates Mercurial's filelog->manifest->changelog write order,
     # but is generally fine for client repos.
     pendingfilecommits = []
     def addrawrevision(orig, self, rawtext, transaction, link, p1, p2, node,
                        flags, cachedelta=None, _metatuple=None):
         if isinstance(link, int):
             pendingfilecommits.append(
                 (self, rawtext, transaction, link, p1, p2, node, flags,
                  cachedelta, _metatuple))
             return node
         else:
             return orig(self, rawtext, transaction, link, p1, p2, node, flags,
                         cachedelta, _metatuple=_metatuple)
     extensions.wrapfunction(
         remotefilelog.remotefilelog, 'addrawrevision', addrawrevision)
 
     def changelogadd(orig, self, *args):
         oldlen = len(self)
         node = orig(self, *args)
         newlen = len(self)
         if oldlen != newlen:
             for oldargs in pendingfilecommits:
                 log, rt, tr, link, p1, p2, n, fl, c, m = oldargs
                 linknode = self.node(link)
                 if linknode == node:
                     log.addrawrevision(rt, tr, linknode, p1, p2, n, fl, c, m)
                 else:
                     raise error.ProgrammingError(
                         'pending multiple integer revisions are not supported')
         else:
             # "link" is actually wrong here (it is set to len(changelog))
             # if changelog remains unchanged, skip writing file revisions
             # but still do a sanity check about pending multiple revisions
             if len(set(x[3] for x in pendingfilecommits)) > 1:
                 raise error.ProgrammingError(
                     'pending multiple integer revisions are not supported')
         del pendingfilecommits[:]
         return node
     extensions.wrapfunction(changelog.changelog, 'add', changelogadd)
 
 def getrenamedfn(orig, repo, endrev=None):
     if not isenabled(repo) or copies.usechangesetcentricalgo(repo):
         return orig(repo, endrev)
 
     rcache = {}
 
     def getrenamed(fn, rev):
         '''looks up all renames for a file (up to endrev) the first
         time the file is given. It indexes on the changerev and only
         parses the manifest if linkrev != changerev.
         Returns rename info for fn at changerev rev.'''
         if rev in rcache.setdefault(fn, {}):
             return rcache[fn][rev]
 
         try:
             fctx = repo[rev].filectx(fn)
             for ancestor in fctx.ancestors():
                 if ancestor.path() == fn:
                     renamed = ancestor.renamed()
                     rcache[fn][ancestor.rev()] = renamed and renamed[0]
 
             renamed = fctx.renamed()
             return renamed and renamed[0]
         except error.LookupError:
             return None
 
     return getrenamed
 
 def walkfilerevs(orig, repo, match, follow, revs, fncache):
     if not isenabled(repo):
         return orig(repo, match, follow, revs, fncache)
 
     # remotefilelog's can't be walked in rev order, so throw.
     # The caller will see the exception and walk the commit tree instead.
     if not follow:
         raise cmdutil.FileWalkError("Cannot walk via filelog")
 
     wanted = set()
     minrev, maxrev = min(revs), max(revs)
 
     pctx = repo['.']
     for filename in match.files():
         if filename not in pctx:
             raise error.Abort(_('cannot follow file not in parent '
                                'revision: "%s"') % filename)
         fctx = pctx[filename]
 
         linkrev = fctx.linkrev()
         if linkrev >= minrev and linkrev <= maxrev:
             fncache.setdefault(linkrev, []).append(filename)
             wanted.add(linkrev)
 
         for ancestor in fctx.ancestors():
             linkrev = ancestor.linkrev()
             if linkrev >= minrev and linkrev <= maxrev:
                 fncache.setdefault(linkrev, []).append(ancestor.path())
                 wanted.add(linkrev)
 
     return wanted
 
 def filelogrevset(orig, repo, subset, x):
     """``filelog(pattern)``
     Changesets connected to the specified filelog.
 
     For performance reasons, ``filelog()`` does not show every changeset
     that affects the requested file(s). See :hg:`help log` for details. For
     a slower, more accurate result, use ``file()``.
     """
 
     if not isenabled(repo):
         return orig(repo, subset, x)
 
     # i18n: "filelog" is a keyword
     pat = revset.getstring(x, _("filelog requires a pattern"))
     m = match.match(repo.root, repo.getcwd(), [pat], default='relpath',
                        ctx=repo[None])
     s = set()
 
     if not match.patkind(pat):
         # slow
         for r in subset:
             ctx = repo[r]
             cfiles = ctx.files()
             for f in m.files():
                 if f in cfiles:
                     s.add(ctx.rev())
                     break
     else:
         # partial
         files = (f for f in repo[None] if m(f))
         for f in files:
             fctx = repo[None].filectx(f)
             s.add(fctx.linkrev())
             for actx in fctx.ancestors():
                 s.add(actx.linkrev())
 
     return smartset.baseset([r for r in subset if r in s])
 
 @command('gc', [], _('hg gc [REPO...]'), norepo=True)
 def gc(ui, *args, **opts):
     '''garbage collect the client and server filelog caches
     '''
     cachepaths = set()
 
     # get the system client cache
     systemcache = shallowutil.getcachepath(ui, allowempty=True)
     if systemcache:
         cachepaths.add(systemcache)
 
     # get repo client and server cache
     repopaths = []
     pwd = ui.environ.get('PWD')
     if pwd:
         repopaths.append(pwd)
 
     repopaths.extend(args)
     repos = []
     for repopath in repopaths:
         try:
             repo = hg.peer(ui, {}, repopath)
             repos.append(repo)
 
             repocache = shallowutil.getcachepath(repo.ui, allowempty=True)
             if repocache:
                 cachepaths.add(repocache)
         except error.RepoError:
             pass
 
     # gc client cache
     for cachepath in cachepaths:
         gcclient(ui, cachepath)
 
     # gc server cache
     for repo in repos:
         remotefilelogserver.gcserver(ui, repo._repo)
 
 def gcclient(ui, cachepath):
     # get list of repos that use this cache
     repospath = os.path.join(cachepath, 'repos')
     if not os.path.exists(repospath):
         ui.warn(_("no known cache at %s\n") % cachepath)
         return
 
     reposfile = open(repospath, 'rb')
     repos = {r[:-1] for r in reposfile.readlines()}
     reposfile.close()
 
     # build list of useful files
     validrepos = []
     keepkeys = set()
 
     sharedcache = None
     filesrepacked = False
 
     count = 0
     progress = ui.makeprogress(_("analyzing repositories"), unit="repos",
                                total=len(repos))
     for path in repos:
         progress.update(count)
         count += 1
         try:
             path = ui.expandpath(os.path.normpath(path))
         except TypeError as e:
             ui.warn(_("warning: malformed path: %r:%s\n") % (path, e))
             traceback.print_exc()
             continue
         try:
             peer = hg.peer(ui, {}, path)
             repo = peer._repo
         except error.RepoError:
             continue
 
         validrepos.append(path)
 
         # Protect against any repo or config changes that have happened since
         # this repo was added to the repos file. We'd rather this loop succeed
         # and too much be deleted, than the loop fail and nothing gets deleted.
         if not isenabled(repo):
             continue
 
         if not util.safehasattr(repo, 'name'):
             ui.warn(_("repo %s is a misconfigured remotefilelog repo\n") % path)
             continue
 
         # If garbage collection on repack and repack on hg gc are enabled
         # then loose files are repacked and garbage collected.
         # Otherwise regular garbage collection is performed.
         repackonhggc = repo.ui.configbool('remotefilelog', 'repackonhggc')
         gcrepack = repo.ui.configbool('remotefilelog', 'gcrepack')
         if repackonhggc and gcrepack:
             try:
                 repackmod.incrementalrepack(repo)
                 filesrepacked = True
                 continue
             except (IOError, repackmod.RepackAlreadyRunning):
                 # If repack cannot be performed due to not enough disk space
                 # continue doing garbage collection of loose files w/o repack
                 pass
 
         reponame = repo.name
         if not sharedcache:
             sharedcache = repo.sharedstore
 
         # Compute a keepset which is not garbage collected
         def keyfn(fname, fnode):
             return fileserverclient.getcachekey(reponame, fname, hex(fnode))
         keepkeys = repackmod.keepset(repo, keyfn=keyfn, lastkeepkeys=keepkeys)
 
     progress.complete()
 
     # write list of valid repos back
     oldumask = os.umask(0o002)
     try:
         reposfile = open(repospath, 'wb')
         reposfile.writelines([("%s\n" % r) for r in validrepos])
         reposfile.close()
     finally:
         os.umask(oldumask)
 
     # prune cache
     if sharedcache is not None:
         sharedcache.gc(keepkeys)
     elif not filesrepacked:
         ui.warn(_("warning: no valid repos in repofile\n"))
 
 def log(orig, ui, repo, *pats, **opts):
     if not isenabled(repo):
         return orig(ui, repo, *pats, **opts)
 
     follow = opts.get(r'follow')
     revs = opts.get(r'rev')
     if pats:
         # Force slowpath for non-follow patterns and follows that start from
         # non-working-copy-parent revs.
         if not follow or revs:
             # This forces the slowpath
             opts[r'removed'] = True
 
         # If this is a non-follow log without any revs specified, recommend that
         # the user add -f to speed it up.
         if not follow and not revs:
             match = scmutil.match(repo['.'], pats, pycompat.byteskwargs(opts))
             isfile = not match.anypats()
             if isfile:
                 for file in match.files():
                     if not os.path.isfile(repo.wjoin(file)):
                         isfile = False
                         break
 
             if isfile:
                 ui.warn(_("warning: file log can be slow on large repos - " +
                           "use -f to speed it up\n"))
 
     return orig(ui, repo, *pats, **opts)
 
 def revdatelimit(ui, revset):
     """Update revset so that only changesets no older than 'prefetchdays' days
     are included. The default value is set to 14 days. If 'prefetchdays' is set
     to zero or negative value then date restriction is not applied.
     """
     days = ui.configint('remotefilelog', 'prefetchdays')
     if days > 0:
         revset = '(%s) & date(-%s)' % (revset, days)
     return revset
 
 def readytofetch(repo):
     """Check that enough time has passed since the last background prefetch.
     This only relates to prefetches after operations that change the working
     copy parent. Default delay between background prefetches is 2 minutes.
     """
     timeout = repo.ui.configint('remotefilelog', 'prefetchdelay')
     fname = repo.vfs.join('lastprefetch')
 
     ready = False
     with open(fname, 'a'):
         # the with construct above is used to avoid race conditions
         modtime = os.path.getmtime(fname)
         if (time.time() - modtime) > timeout:
             os.utime(fname, None)
             ready = True
 
     return ready
 
 def wcpprefetch(ui, repo, **kwargs):
     """Prefetches in background revisions specified by bgprefetchrevs revset.
     Does background repack if backgroundrepack flag is set in config.
     """
     shallow = isenabled(repo)
     bgprefetchrevs = ui.config('remotefilelog', 'bgprefetchrevs')
     isready = readytofetch(repo)
 
     if not (shallow and bgprefetchrevs and isready):
         return
 
     bgrepack = repo.ui.configbool('remotefilelog', 'backgroundrepack')
     # update a revset with a date limit
     bgprefetchrevs = revdatelimit(ui, bgprefetchrevs)
 
     def anon():
         if util.safehasattr(repo, 'ranprefetch') and repo.ranprefetch:
             return
         repo.ranprefetch = True
         repo.backgroundprefetch(bgprefetchrevs, repack=bgrepack)
 
     repo._afterlock(anon)
 
 def pull(orig, ui, repo, *pats, **opts):
     result = orig(ui, repo, *pats, **opts)
 
     if isenabled(repo):
         # prefetch if it's configured
         prefetchrevset = ui.config('remotefilelog', 'pullprefetch')
         bgrepack = repo.ui.configbool('remotefilelog', 'backgroundrepack')
         bgprefetch = repo.ui.configbool('remotefilelog', 'backgroundprefetch')
 
         if prefetchrevset:
             ui.status(_("prefetching file contents\n"))
             revs = scmutil.revrange(repo, [prefetchrevset])
             base = repo['.'].rev()
             if bgprefetch:
                 repo.backgroundprefetch(prefetchrevset, repack=bgrepack)
             else:
                 repo.prefetch(revs, base=base)
                 if bgrepack:
                     repackmod.backgroundrepack(repo, incremental=True)
         elif bgrepack:
             repackmod.backgroundrepack(repo, incremental=True)
 
     return result
 
 def exchangepull(orig, repo, remote, *args, **kwargs):
     # Hook into the callstream/getbundle to insert bundle capabilities
     # during a pull.
     def localgetbundle(orig, source, heads=None, common=None, bundlecaps=None,
                        **kwargs):
         if not bundlecaps:
             bundlecaps = set()
         bundlecaps.add(constants.BUNDLE2_CAPABLITY)
         return orig(source, heads=heads, common=common, bundlecaps=bundlecaps,
                     **kwargs)
 
     if util.safehasattr(remote, '_callstream'):
         remote._localrepo = repo
     elif util.safehasattr(remote, 'getbundle'):
         extensions.wrapfunction(remote, 'getbundle', localgetbundle)
 
     return orig(repo, remote, *args, **kwargs)
 
 def _fileprefetchhook(repo, revs, match):
     if isenabled(repo):
         allfiles = []
         for rev in revs:
             if rev == nodemod.wdirrev or rev is None:
                 continue
             ctx = repo[rev]
             mf = ctx.manifest()
             sparsematch = repo.maybesparsematch(ctx.rev())
             for path in ctx.walk(match):
                 if (not sparsematch or sparsematch(path)) and path in mf:
                     allfiles.append((path, hex(mf[path])))
         repo.fileservice.prefetch(allfiles)
 
 @command('debugremotefilelog', [
     ('d', 'decompress', None, _('decompress the filelog first')),
     ], _('hg debugremotefilelog <path>'), norepo=True)
 def debugremotefilelog(ui, path, **opts):
     return debugcommands.debugremotefilelog(ui, path, **opts)
 
 @command('verifyremotefilelog', [
     ('d', 'decompress', None, _('decompress the filelogs first')),
     ], _('hg verifyremotefilelogs <directory>'), norepo=True)
 def verifyremotefilelog(ui, path, **opts):
     return debugcommands.verifyremotefilelog(ui, path, **opts)
 
 @command('debugdatapack', [
     ('', 'long', None, _('print the long hashes')),
     ('', 'node', '', _('dump the contents of node'), 'NODE'),
     ], _('hg debugdatapack <paths>'), norepo=True)
 def debugdatapack(ui, *paths, **opts):
     return debugcommands.debugdatapack(ui, *paths, **opts)
 
 @command('debughistorypack', [
     ], _('hg debughistorypack <path>'), norepo=True)
 def debughistorypack(ui, path, **opts):
     return debugcommands.debughistorypack(ui, path)
 
 @command('debugkeepset', [
     ], _('hg debugkeepset'))
 def debugkeepset(ui, repo, **opts):
     # The command is used to measure keepset computation time
     def keyfn(fname, fnode):
         return fileserverclient.getcachekey(repo.name, fname, hex(fnode))
     repackmod.keepset(repo, keyfn)
     return
 
 @command('debugwaitonrepack', [
     ], _('hg debugwaitonrepack'))
 def debugwaitonrepack(ui, repo, **opts):
     return debugcommands.debugwaitonrepack(repo)
 
 @command('debugwaitonprefetch', [
     ], _('hg debugwaitonprefetch'))
 def debugwaitonprefetch(ui, repo, **opts):
     return debugcommands.debugwaitonprefetch(repo)
 
 def resolveprefetchopts(ui, opts):
     if not opts.get('rev'):
         revset = ['.', 'draft()']
 
         prefetchrevset = ui.config('remotefilelog', 'pullprefetch', None)
         if prefetchrevset:
             revset.append('(%s)' % prefetchrevset)
         bgprefetchrevs = ui.config('remotefilelog', 'bgprefetchrevs', None)
         if bgprefetchrevs:
             revset.append('(%s)' % bgprefetchrevs)
         revset = '+'.join(revset)
 
         # update a revset with a date limit
         revset = revdatelimit(ui, revset)
 
         opts['rev'] = [revset]
 
     if not opts.get('base'):
         opts['base'] = None
 
     return opts
 
 @command('prefetch', [
     ('r', 'rev', [], _('prefetch the specified revisions'), _('REV')),
     ('', 'repack', False, _('run repack after prefetch')),
     ('b', 'base', '', _("rev that is assumed to already be local")),
     ] + commands.walkopts, _('hg prefetch [OPTIONS] [FILE...]'))
 def prefetch(ui, repo, *pats, **opts):
     """prefetch file revisions from the server
 
     Prefetchs file revisions for the specified revs and stores them in the
     local remotefilelog cache.  If no rev is specified, the default rev is
     used which is the union of dot, draft, pullprefetch and bgprefetchrev.
     File names or patterns can be used to limit which files are downloaded.
 
     Return 0 on success.
     """
     opts = pycompat.byteskwargs(opts)
     if not isenabled(repo):
         raise error.Abort(_("repo is not shallow"))
 
     opts = resolveprefetchopts(ui, opts)
     revs = scmutil.revrange(repo, opts.get('rev'))
     repo.prefetch(revs, opts.get('base'), pats, opts)
 
     # Run repack in background
     if opts.get('repack'):
         repackmod.backgroundrepack(repo, incremental=True)
 
 @command('repack', [
      ('', 'background', None, _('run in a background process'), None),
      ('', 'incremental', None, _('do an incremental repack'), None),
      ('', 'packsonly', None, _('only repack packs (skip loose objects)'), None),
     ], _('hg repack [OPTIONS]'))
 def repack_(ui, repo, *pats, **opts):
     if opts.get(r'background'):
         repackmod.backgroundrepack(repo, incremental=opts.get(r'incremental'),
                                    packsonly=opts.get(r'packsonly', False))
         return
 
     options = {'packsonly': opts.get(r'packsonly')}
 
     try:
         if opts.get(r'incremental'):
             repackmod.incrementalrepack(repo, options=options)
         else:
             repackmod.fullrepack(repo, options=options)
     except repackmod.RepackAlreadyRunning as ex:
         # Don't propogate the exception if the repack is already in
         # progress, since we want the command to exit 0.
         repo.ui.warn('%s\n' % ex)

hgext/sqlitestore.py

@@ -1,1174 +1,1175 b''
 # sqlitestore.py - Storage backend that uses SQLite
 #
 # Copyright 2018 Gregory Szorc <[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.
 
 """store repository data in SQLite (EXPERIMENTAL)
 
 The sqlitestore extension enables the storage of repository data in SQLite.
 
 This extension is HIGHLY EXPERIMENTAL. There are NO BACKWARDS COMPATIBILITY
 GUARANTEES. This means that repositories created with this extension may
 only be usable with the exact version of this extension/Mercurial that was
 used. The extension attempts to enforce this in order to prevent repository
 corruption.
 
 In addition, several features are not yet supported or have known bugs:
 
 * Only some data is stored in SQLite. Changeset, manifest, and other repository
   data is not yet stored in SQLite.
 * Transactions are not robust. If the process is aborted at the right time
   during transaction close/rollback, the repository could be in an inconsistent
   state. This problem will diminish once all repository data is tracked by
   SQLite.
 * Bundle repositories do not work (the ability to use e.g.
   `hg -R <bundle-file> log` to automatically overlay a bundle on top of the
   existing repository).
 * Various other features don't work.
 
 This extension should work for basic clone/pull, update, and commit workflows.
 Some history rewriting operations may fail due to lack of support for bundle
 repositories.
 
 To use, activate the extension and set the ``storage.new-repo-backend`` config
 option to ``sqlite`` to enable new repositories to use SQLite for storage.
 """
 
 # To run the test suite with repos using SQLite by default, execute the
 # following:
 #
 # HGREPOFEATURES="sqlitestore" run-tests.py \
 #     --extra-config-opt extensions.sqlitestore= \
 #     --extra-config-opt storage.new-repo-backend=sqlite
 
 from __future__ import absolute_import
 
 import hashlib
 import sqlite3
 import struct
 import threading
 import zlib
 
 from mercurial.i18n import _
 from mercurial.node import (
     nullid,
     nullrev,
     short,
 )
 from mercurial.thirdparty import (
     attr,
 )
 from mercurial import (
     ancestor,
     dagop,
     encoding,
     error,
     extensions,
     localrepo,
     mdiff,
     pycompat,
     registrar,
     repository,
     util,
     verify,
 )
 from mercurial.utils import (
     interfaceutil,
     storageutil,
 )
 
 try:
     from mercurial import zstd
     zstd.__version__
 except ImportError:
     zstd = None
 
 configtable = {}
 configitem = registrar.configitem(configtable)
 
 # experimental config: storage.sqlite.compression
 configitem('storage', 'sqlite.compression',
-           default='zstd' if zstd else 'zlib')
+           default='zstd' if zstd else 'zlib',
+           experimental=True)
 
 # Note for extension authors: ONLY specify testedwith = 'ships-with-hg-core' for
 # extensions which SHIP WITH MERCURIAL. Non-mainline extensions should
 # be specifying the version(s) of Mercurial they are tested with, or
 # leave the attribute unspecified.
 testedwith = 'ships-with-hg-core'
 
 REQUIREMENT = b'exp-sqlite-001'
 REQUIREMENT_ZSTD = b'exp-sqlite-comp-001=zstd'
 REQUIREMENT_ZLIB = b'exp-sqlite-comp-001=zlib'
 REQUIREMENT_NONE = b'exp-sqlite-comp-001=none'
 REQUIREMENT_SHALLOW_FILES = b'exp-sqlite-shallow-files'
 
 CURRENT_SCHEMA_VERSION = 1
 
 COMPRESSION_NONE = 1
 COMPRESSION_ZSTD = 2
 COMPRESSION_ZLIB = 3
 
 FLAG_CENSORED = 1
 FLAG_MISSING_P1 = 2
 FLAG_MISSING_P2 = 4
 
 CREATE_SCHEMA = [
     # Deltas are stored as content-indexed blobs.
     # compression column holds COMPRESSION_* constant for how the
     # delta is encoded.
 
     r'CREATE TABLE delta ('
     r'    id INTEGER PRIMARY KEY, '
     r'    compression INTEGER NOT NULL, '
     r'    hash BLOB UNIQUE ON CONFLICT ABORT, '
     r'    delta BLOB NOT NULL '
     r')',
 
     # Tracked paths are denormalized to integers to avoid redundant
     # storage of the path name.
     r'CREATE TABLE filepath ('
     r'    id INTEGER PRIMARY KEY, '
     r'    path BLOB NOT NULL '
     r')',
 
     r'CREATE UNIQUE INDEX filepath_path '
     r'    ON filepath (path)',
 
     # We have a single table for all file revision data.
     # Each file revision is uniquely described by a (path, rev) and
     # (path, node).
     #
     # Revision data is stored as a pointer to the delta producing this
     # revision and the file revision whose delta should be applied before
     # that one. One can reconstruct the delta chain by recursively following
     # the delta base revision pointers until one encounters NULL.
     #
     # flags column holds bitwise integer flags controlling storage options.
     # These flags are defined by the FLAG_* constants.
     r'CREATE TABLE fileindex ('
     r'    id INTEGER PRIMARY KEY, '
     r'    pathid INTEGER REFERENCES filepath(id), '
     r'    revnum INTEGER NOT NULL, '
     r'    p1rev INTEGER NOT NULL, '
     r'    p2rev INTEGER NOT NULL, '
     r'    linkrev INTEGER NOT NULL, '
     r'    flags INTEGER NOT NULL, '
     r'    deltaid INTEGER REFERENCES delta(id), '
     r'    deltabaseid INTEGER REFERENCES fileindex(id), '
     r'    node BLOB NOT NULL '
     r')',
 
     r'CREATE UNIQUE INDEX fileindex_pathrevnum '
     r'    ON fileindex (pathid, revnum)',
 
     r'CREATE UNIQUE INDEX fileindex_pathnode '
     r'    ON fileindex (pathid, node)',
 
     # Provide a view over all file data for convenience.
     r'CREATE VIEW filedata AS '
     r'SELECT '
     r'    fileindex.id AS id, '
     r'    filepath.id AS pathid, '
     r'    filepath.path AS path, '
     r'    fileindex.revnum AS revnum, '
     r'    fileindex.node AS node, '
     r'    fileindex.p1rev AS p1rev, '
     r'    fileindex.p2rev AS p2rev, '
     r'    fileindex.linkrev AS linkrev, '
     r'    fileindex.flags AS flags, '
     r'    fileindex.deltaid AS deltaid, '
     r'    fileindex.deltabaseid AS deltabaseid '
     r'FROM filepath, fileindex '
     r'WHERE fileindex.pathid=filepath.id',
 
     r'PRAGMA user_version=%d' % CURRENT_SCHEMA_VERSION,
 ]
 
 def resolvedeltachain(db, pathid, node, revisioncache,
                       stoprids, zstddctx=None):
     """Resolve a delta chain for a file node."""
 
     # TODO the "not in ({stops})" here is possibly slowing down the query
     # because it needs to perform the lookup on every recursive invocation.
     # This could possibly be faster if we created a temporary query with
     # baseid "poisoned" to null and limited the recursive filter to
     # "is not null".
     res = db.execute(
         r'WITH RECURSIVE '
         r'    deltachain(deltaid, baseid) AS ('
         r'        SELECT deltaid, deltabaseid FROM fileindex '
         r'            WHERE pathid=? AND node=? '
         r'        UNION ALL '
         r'        SELECT fileindex.deltaid, deltabaseid '
         r'            FROM fileindex, deltachain '
         r'            WHERE '
         r'                fileindex.id=deltachain.baseid '
         r'                AND deltachain.baseid IS NOT NULL '
         r'                AND fileindex.id NOT IN ({stops}) '
         r'    ) '
         r'SELECT deltachain.baseid, compression, delta '
         r'FROM deltachain, delta '
         r'WHERE delta.id=deltachain.deltaid'.format(
             stops=r','.join([r'?'] * len(stoprids))),
         tuple([pathid, node] + list(stoprids.keys())))
 
     deltas = []
     lastdeltabaseid = None
 
     for deltabaseid, compression, delta in res:
         lastdeltabaseid = deltabaseid
 
         if compression == COMPRESSION_ZSTD:
             delta = zstddctx.decompress(delta)
         elif compression == COMPRESSION_NONE:
             delta = delta
         elif compression == COMPRESSION_ZLIB:
             delta = zlib.decompress(delta)
         else:
             raise SQLiteStoreError('unhandled compression type: %d' %
                                    compression)
 
         deltas.append(delta)
 
     if lastdeltabaseid in stoprids:
         basetext = revisioncache[stoprids[lastdeltabaseid]]
     else:
         basetext = deltas.pop()
 
     deltas.reverse()
     fulltext = mdiff.patches(basetext, deltas)
 
     # SQLite returns buffer instances for blob columns on Python 2. This
     # type can propagate through the delta application layer. Because
     # downstream callers assume revisions are bytes, cast as needed.
     if not isinstance(fulltext, bytes):
         fulltext = bytes(delta)
 
     return fulltext
 
 def insertdelta(db, compression, hash, delta):
     try:
         return db.execute(
             r'INSERT INTO delta (compression, hash, delta) '
             r'VALUES (?, ?, ?)',
             (compression, hash, delta)).lastrowid
     except sqlite3.IntegrityError:
         return db.execute(
             r'SELECT id FROM delta WHERE hash=?',
             (hash,)).fetchone()[0]
 
 class SQLiteStoreError(error.StorageError):
     pass
 
 @attr.s
 class revisionentry(object):
     rid = attr.ib()
     rev = attr.ib()
     node = attr.ib()
     p1rev = attr.ib()
     p2rev = attr.ib()
     p1node = attr.ib()
     p2node = attr.ib()
     linkrev = attr.ib()
     flags = attr.ib()
 
 @interfaceutil.implementer(repository.irevisiondelta)
 @attr.s(slots=True)
 class sqliterevisiondelta(object):
     node = attr.ib()
     p1node = attr.ib()
     p2node = attr.ib()
     basenode = attr.ib()
     flags = attr.ib()
     baserevisionsize = attr.ib()
     revision = attr.ib()
     delta = attr.ib()
     linknode = attr.ib(default=None)
 
 @interfaceutil.implementer(repository.iverifyproblem)
 @attr.s(frozen=True)
 class sqliteproblem(object):
     warning = attr.ib(default=None)
     error = attr.ib(default=None)
     node = attr.ib(default=None)
 
 @interfaceutil.implementer(repository.ifilestorage)
 class sqlitefilestore(object):
     """Implements storage for an individual tracked path."""
 
     def __init__(self, db, path, compression):
         self._db = db
         self._path = path
 
         self._pathid = None
 
         # revnum -> node
         self._revtonode = {}
         # node -> revnum
         self._nodetorev = {}
         # node -> data structure
         self._revisions = {}
 
         self._revisioncache = util.lrucachedict(10)
 
         self._compengine = compression
 
         if compression == 'zstd':
             self._cctx = zstd.ZstdCompressor(level=3)
             self._dctx = zstd.ZstdDecompressor()
         else:
             self._cctx = None
             self._dctx = None
 
         self._refreshindex()
 
     def _refreshindex(self):
         self._revtonode = {}
         self._nodetorev = {}
         self._revisions = {}
 
         res = list(self._db.execute(
             r'SELECT id FROM filepath WHERE path=?', (self._path,)))
 
         if not res:
             self._pathid = None
             return
 
         self._pathid = res[0][0]
 
         res = self._db.execute(
             r'SELECT id, revnum, node, p1rev, p2rev, linkrev, flags '
             r'FROM fileindex '
             r'WHERE pathid=? '
             r'ORDER BY revnum ASC',
             (self._pathid,))
 
         for i, row in enumerate(res):
             rid, rev, node, p1rev, p2rev, linkrev, flags = row
 
             if i != rev:
                 raise SQLiteStoreError(_('sqlite database has inconsistent '
                                          'revision numbers'))
 
             if p1rev == nullrev:
                 p1node = nullid
             else:
                 p1node = self._revtonode[p1rev]
 
             if p2rev == nullrev:
                 p2node = nullid
             else:
                 p2node = self._revtonode[p2rev]
 
             entry = revisionentry(
                 rid=rid,
                 rev=rev,
                 node=node,
                 p1rev=p1rev,
                 p2rev=p2rev,
                 p1node=p1node,
                 p2node=p2node,
                 linkrev=linkrev,
                 flags=flags)
 
             self._revtonode[rev] = node
             self._nodetorev[node] = rev
             self._revisions[node] = entry
 
     # Start of ifileindex interface.
 
     def __len__(self):
         return len(self._revisions)
 
     def __iter__(self):
         return iter(pycompat.xrange(len(self._revisions)))
 
     def hasnode(self, node):
         if node == nullid:
             return False
 
         return node in self._nodetorev
 
     def revs(self, start=0, stop=None):
         return storageutil.iterrevs(len(self._revisions), start=start,
                                     stop=stop)
 
     def parents(self, node):
         if node == nullid:
             return nullid, nullid
 
         if node not in self._revisions:
             raise error.LookupError(node, self._path, _('no node'))
 
         entry = self._revisions[node]
         return entry.p1node, entry.p2node
 
     def parentrevs(self, rev):
         if rev == nullrev:
             return nullrev, nullrev
 
         if rev not in self._revtonode:
             raise IndexError(rev)
 
         entry = self._revisions[self._revtonode[rev]]
         return entry.p1rev, entry.p2rev
 
     def rev(self, node):
         if node == nullid:
             return nullrev
 
         if node not in self._nodetorev:
             raise error.LookupError(node, self._path, _('no node'))
 
         return self._nodetorev[node]
 
     def node(self, rev):
         if rev == nullrev:
             return nullid
 
         if rev not in self._revtonode:
             raise IndexError(rev)
 
         return self._revtonode[rev]
 
     def lookup(self, node):
         return storageutil.fileidlookup(self, node, self._path)
 
     def linkrev(self, rev):
         if rev == nullrev:
             return nullrev
 
         if rev not in self._revtonode:
             raise IndexError(rev)
 
         entry = self._revisions[self._revtonode[rev]]
         return entry.linkrev
 
     def iscensored(self, rev):
         if rev == nullrev:
             return False
 
         if rev not in self._revtonode:
             raise IndexError(rev)
 
         return self._revisions[self._revtonode[rev]].flags & FLAG_CENSORED
 
     def commonancestorsheads(self, node1, node2):
         rev1 = self.rev(node1)
         rev2 = self.rev(node2)
 
         ancestors = ancestor.commonancestorsheads(self.parentrevs, rev1, rev2)
         return pycompat.maplist(self.node, ancestors)
 
     def descendants(self, revs):
         # TODO we could implement this using a recursive SQL query, which
         # might be faster.
         return dagop.descendantrevs(revs, self.revs, self.parentrevs)
 
     def heads(self, start=None, stop=None):
         if start is None and stop is None:
             if not len(self):
                 return [nullid]
 
         startrev = self.rev(start) if start is not None else nullrev
         stoprevs = {self.rev(n) for n in stop or []}
 
         revs = dagop.headrevssubset(self.revs, self.parentrevs,
                                     startrev=startrev, stoprevs=stoprevs)
 
         return [self.node(rev) for rev in revs]
 
     def children(self, node):
         rev = self.rev(node)
 
         res = self._db.execute(
             r'SELECT'
             r'  node '
             r'  FROM filedata '
             r'  WHERE path=? AND (p1rev=? OR p2rev=?) '
             r'  ORDER BY revnum ASC',
             (self._path, rev, rev))
 
         return [row[0] for row in res]
 
     # End of ifileindex interface.
 
     # Start of ifiledata interface.
 
     def size(self, rev):
         if rev == nullrev:
             return 0
 
         if rev not in self._revtonode:
             raise IndexError(rev)
 
         node = self._revtonode[rev]
 
         if self.renamed(node):
             return len(self.read(node))
 
         return len(self.revision(node))
 
     def revision(self, node, raw=False, _verifyhash=True):
         if node in (nullid, nullrev):
             return b''
 
         if isinstance(node, int):
             node = self.node(node)
 
         if node not in self._nodetorev:
             raise error.LookupError(node, self._path, _('no node'))
 
         if node in self._revisioncache:
             return self._revisioncache[node]
 
         # Because we have a fulltext revision cache, we are able to
         # short-circuit delta chain traversal and decompression as soon as
         # we encounter a revision in the cache.
 
         stoprids = {self._revisions[n].rid: n
                     for n in self._revisioncache}
 
         if not stoprids:
             stoprids[-1] = None
 
         fulltext = resolvedeltachain(self._db, self._pathid, node,
                                      self._revisioncache, stoprids,
                                      zstddctx=self._dctx)
 
         # Don't verify hashes if parent nodes were rewritten, as the hash
         # wouldn't verify.
         if self._revisions[node].flags & (FLAG_MISSING_P1 | FLAG_MISSING_P2):
             _verifyhash = False
 
         if _verifyhash:
             self._checkhash(fulltext, node)
             self._revisioncache[node] = fulltext
 
         return fulltext
 
     def rawdata(self, *args, **kwargs):
         return self.revision(*args, **kwargs)
 
     def read(self, node):
         return storageutil.filtermetadata(self.revision(node))
 
     def renamed(self, node):
         return storageutil.filerevisioncopied(self, node)
 
     def cmp(self, node, fulltext):
         return not storageutil.filedataequivalent(self, node, fulltext)
 
     def emitrevisions(self, nodes, nodesorder=None, revisiondata=False,
                       assumehaveparentrevisions=False,
                       deltamode=repository.CG_DELTAMODE_STD):
         if nodesorder not in ('nodes', 'storage', 'linear', None):
             raise error.ProgrammingError('unhandled value for nodesorder: %s' %
                                          nodesorder)
 
         nodes = [n for n in nodes if n != nullid]
 
         if not nodes:
             return
 
         # TODO perform in a single query.
         res = self._db.execute(
             r'SELECT revnum, deltaid FROM fileindex '
             r'WHERE pathid=? '
             r'    AND node in (%s)' % (r','.join([r'?'] * len(nodes))),
             tuple([self._pathid] + nodes))
 
         deltabases = {}
 
         for rev, deltaid in res:
             res = self._db.execute(
                 r'SELECT revnum from fileindex WHERE pathid=? AND deltaid=?',
                 (self._pathid, deltaid))
             deltabases[rev] = res.fetchone()[0]
 
         # TODO define revdifffn so we can use delta from storage.
         for delta in storageutil.emitrevisions(
             self, nodes, nodesorder, sqliterevisiondelta,
             deltaparentfn=deltabases.__getitem__,
             revisiondata=revisiondata,
             assumehaveparentrevisions=assumehaveparentrevisions,
             deltamode=deltamode):
 
             yield delta
 
     # End of ifiledata interface.
 
     # Start of ifilemutation interface.
 
     def add(self, filedata, meta, transaction, linkrev, p1, p2):
         if meta or filedata.startswith(b'\x01\n'):
             filedata = storageutil.packmeta(meta, filedata)
 
         return self.addrevision(filedata, transaction, linkrev, p1, p2)
 
     def addrevision(self, revisiondata, transaction, linkrev, p1, p2, node=None,
                     flags=0, cachedelta=None):
         if flags:
             raise SQLiteStoreError(_('flags not supported on revisions'))
 
         validatehash = node is not None
         node = node or storageutil.hashrevisionsha1(revisiondata, p1, p2)
 
         if validatehash:
             self._checkhash(revisiondata, node, p1, p2)
 
         if node in self._nodetorev:
             return node
 
         node = self._addrawrevision(node, revisiondata, transaction, linkrev,
                                     p1, p2)
 
         self._revisioncache[node] = revisiondata
         return node
 
     def addgroup(self, deltas, linkmapper, transaction, addrevisioncb=None,
                  maybemissingparents=False):
         nodes = []
 
         for node, p1, p2, linknode, deltabase, delta, wireflags in deltas:
             storeflags = 0
 
             if wireflags & repository.REVISION_FLAG_CENSORED:
                 storeflags |= FLAG_CENSORED
 
             if wireflags & ~repository.REVISION_FLAG_CENSORED:
                 raise SQLiteStoreError('unhandled revision flag')
 
             if maybemissingparents:
                 if p1 != nullid and not self.hasnode(p1):
                     p1 = nullid
                     storeflags |= FLAG_MISSING_P1
 
                 if p2 != nullid and not self.hasnode(p2):
                     p2 = nullid
                     storeflags |= FLAG_MISSING_P2
 
             baserev = self.rev(deltabase)
 
             # If base is censored, delta must be full replacement in a single
             # patch operation.
             if baserev != nullrev and self.iscensored(baserev):
                 hlen = struct.calcsize('>lll')
                 oldlen = len(self.revision(deltabase, raw=True,
                                            _verifyhash=False))
                 newlen = len(delta) - hlen
 
                 if delta[:hlen] != mdiff.replacediffheader(oldlen, newlen):
                     raise error.CensoredBaseError(self._path,
                                                   deltabase)
 
             if (not (storeflags & FLAG_CENSORED)
                 and storageutil.deltaiscensored(
                     delta, baserev, lambda x: len(self.revision(x, raw=True)))):
                 storeflags |= FLAG_CENSORED
 
             linkrev = linkmapper(linknode)
 
             nodes.append(node)
 
             if node in self._revisions:
                 # Possibly reset parents to make them proper.
                 entry = self._revisions[node]
 
                 if entry.flags & FLAG_MISSING_P1 and p1 != nullid:
                     entry.p1node = p1
                     entry.p1rev = self._nodetorev[p1]
                     entry.flags &= ~FLAG_MISSING_P1
 
                     self._db.execute(
                         r'UPDATE fileindex SET p1rev=?, flags=? '
                         r'WHERE id=?',
                         (self._nodetorev[p1], entry.flags, entry.rid))
 
                 if entry.flags & FLAG_MISSING_P2 and p2 != nullid:
                     entry.p2node = p2
                     entry.p2rev = self._nodetorev[p2]
                     entry.flags &= ~FLAG_MISSING_P2
 
                     self._db.execute(
                         r'UPDATE fileindex SET p2rev=?, flags=? '
                         r'WHERE id=?',
                         (self._nodetorev[p1], entry.flags, entry.rid))
 
                 continue
 
             if deltabase == nullid:
                 text = mdiff.patch(b'', delta)
                 storedelta = None
             else:
                 text = None
                 storedelta = (deltabase, delta)
 
             self._addrawrevision(node, text, transaction, linkrev, p1, p2,
                                  storedelta=storedelta, flags=storeflags)
 
             if addrevisioncb:
                 addrevisioncb(self, node)
 
         return nodes
 
     def censorrevision(self, tr, censornode, tombstone=b''):
         tombstone = storageutil.packmeta({b'censored': tombstone}, b'')
 
         # This restriction is cargo culted from revlogs and makes no sense for
         # SQLite, since columns can be resized at will.
         if len(tombstone) > len(self.revision(censornode, raw=True)):
             raise error.Abort(_('censor tombstone must be no longer than '
                                 'censored data'))
 
         # We need to replace the censored revision's data with the tombstone.
         # But replacing that data will have implications for delta chains that
         # reference it.
         #
         # While "better," more complex strategies are possible, we do something
         # simple: we find delta chain children of the censored revision and we
         # replace those incremental deltas with fulltexts of their corresponding
         # revision. Then we delete the now-unreferenced delta and original
         # revision and insert a replacement.
 
         # Find the delta to be censored.
         censoreddeltaid = self._db.execute(
             r'SELECT deltaid FROM fileindex WHERE id=?',
             (self._revisions[censornode].rid,)).fetchone()[0]
 
         # Find all its delta chain children.
         # TODO once we support storing deltas for !files, we'll need to look
         # for those delta chains too.
         rows = list(self._db.execute(
             r'SELECT id, pathid, node FROM fileindex '
             r'WHERE deltabaseid=? OR deltaid=?',
             (censoreddeltaid, censoreddeltaid)))
 
         for row in rows:
             rid, pathid, node = row
 
             fulltext = resolvedeltachain(self._db, pathid, node, {}, {-1: None},
                                          zstddctx=self._dctx)
 
             deltahash = hashlib.sha1(fulltext).digest()
 
             if self._compengine == 'zstd':
                 deltablob = self._cctx.compress(fulltext)
                 compression = COMPRESSION_ZSTD
             elif self._compengine == 'zlib':
                 deltablob = zlib.compress(fulltext)
                 compression = COMPRESSION_ZLIB
             elif self._compengine == 'none':
                 deltablob = fulltext
                 compression = COMPRESSION_NONE
             else:
                 raise error.ProgrammingError('unhandled compression engine: %s'
                                              % self._compengine)
 
             if len(deltablob) >= len(fulltext):
                 deltablob = fulltext
                 compression = COMPRESSION_NONE
 
             deltaid = insertdelta(self._db, compression, deltahash, deltablob)
 
             self._db.execute(
                 r'UPDATE fileindex SET deltaid=?, deltabaseid=NULL '
                 r'WHERE id=?', (deltaid, rid))
 
         # Now create the tombstone delta and replace the delta on the censored
         # node.
         deltahash = hashlib.sha1(tombstone).digest()
         tombstonedeltaid = insertdelta(self._db, COMPRESSION_NONE,
                                        deltahash, tombstone)
 
         flags = self._revisions[censornode].flags
         flags |= FLAG_CENSORED
 
         self._db.execute(
             r'UPDATE fileindex SET flags=?, deltaid=?, deltabaseid=NULL '
             r'WHERE pathid=? AND node=?',
             (flags, tombstonedeltaid, self._pathid, censornode))
 
         self._db.execute(
             r'DELETE FROM delta WHERE id=?', (censoreddeltaid,))
 
         self._refreshindex()
         self._revisioncache.clear()
 
     def getstrippoint(self, minlink):
         return storageutil.resolvestripinfo(minlink, len(self) - 1,
                                             [self.rev(n) for n in self.heads()],
                                             self.linkrev,
                                             self.parentrevs)
 
     def strip(self, minlink, transaction):
         if not len(self):
             return
 
         rev, _ignored = self.getstrippoint(minlink)
 
         if rev == len(self):
             return
 
         for rev in self.revs(rev):
             self._db.execute(
                 r'DELETE FROM fileindex WHERE pathid=? AND node=?',
                 (self._pathid, self.node(rev)))
 
         # TODO how should we garbage collect data in delta table?
 
         self._refreshindex()
 
     # End of ifilemutation interface.
 
     # Start of ifilestorage interface.
 
     def files(self):
         return []
 
     def storageinfo(self, exclusivefiles=False, sharedfiles=False,
                     revisionscount=False, trackedsize=False,
                     storedsize=False):
         d = {}
 
         if exclusivefiles:
             d['exclusivefiles'] = []
 
         if sharedfiles:
             # TODO list sqlite file(s) here.
             d['sharedfiles'] = []
 
         if revisionscount:
             d['revisionscount'] = len(self)
 
         if trackedsize:
             d['trackedsize'] = sum(len(self.revision(node))
                                        for node in self._nodetorev)
 
         if storedsize:
             # TODO implement this?
             d['storedsize'] = None
 
         return d
 
     def verifyintegrity(self, state):
         state['skipread'] = set()
 
         for rev in self:
             node = self.node(rev)
 
             try:
                 self.revision(node)
             except Exception as e:
                 yield sqliteproblem(
                     error=_('unpacking %s: %s') % (short(node), e),
                     node=node)
 
                 state['skipread'].add(node)
 
     # End of ifilestorage interface.
 
     def _checkhash(self, fulltext, node, p1=None, p2=None):
         if p1 is None and p2 is None:
             p1, p2 = self.parents(node)
 
         if node == storageutil.hashrevisionsha1(fulltext, p1, p2):
             return
 
         try:
             del self._revisioncache[node]
         except KeyError:
             pass
 
         if storageutil.iscensoredtext(fulltext):
             raise error.CensoredNodeError(self._path, node, fulltext)
 
         raise SQLiteStoreError(_('integrity check failed on %s') %
                                self._path)
 
     def _addrawrevision(self, node, revisiondata, transaction, linkrev,
                         p1, p2, storedelta=None, flags=0):
         if self._pathid is None:
             res = self._db.execute(
                 r'INSERT INTO filepath (path) VALUES (?)', (self._path,))
             self._pathid = res.lastrowid
 
         # For simplicity, always store a delta against p1.
         # TODO we need a lot more logic here to make behavior reasonable.
 
         if storedelta:
             deltabase, delta = storedelta
 
             if isinstance(deltabase, int):
                 deltabase = self.node(deltabase)
 
         else:
             assert revisiondata is not None
             deltabase = p1
 
             if deltabase == nullid:
                 delta = revisiondata
             else:
                 delta = mdiff.textdiff(self.revision(self.rev(deltabase)),
                                        revisiondata)
 
         # File index stores a pointer to its delta and the parent delta.
         # The parent delta is stored via a pointer to the fileindex PK.
         if deltabase == nullid:
             baseid = None
         else:
             baseid = self._revisions[deltabase].rid
 
         # Deltas are stored with a hash of their content. This allows
         # us to de-duplicate. The table is configured to ignore conflicts
         # and it is faster to just insert and silently noop than to look
         # first.
         deltahash = hashlib.sha1(delta).digest()
 
         if self._compengine == 'zstd':
             deltablob = self._cctx.compress(delta)
             compression = COMPRESSION_ZSTD
         elif self._compengine == 'zlib':
             deltablob = zlib.compress(delta)
             compression = COMPRESSION_ZLIB
         elif self._compengine == 'none':
             deltablob = delta
             compression = COMPRESSION_NONE
         else:
             raise error.ProgrammingError('unhandled compression engine: %s' %
                                          self._compengine)
 
         # Don't store compressed data if it isn't practical.
         if len(deltablob) >= len(delta):
             deltablob = delta
             compression = COMPRESSION_NONE
 
         deltaid = insertdelta(self._db, compression, deltahash, deltablob)
 
         rev = len(self)
 
         if p1 == nullid:
             p1rev = nullrev
         else:
             p1rev = self._nodetorev[p1]
 
         if p2 == nullid:
             p2rev = nullrev
         else:
             p2rev = self._nodetorev[p2]
 
         rid = self._db.execute(
             r'INSERT INTO fileindex ('
             r'    pathid, revnum, node, p1rev, p2rev, linkrev, flags, '
             r'    deltaid, deltabaseid) '
             r'    VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?)',
             (self._pathid, rev, node, p1rev, p2rev, linkrev, flags,
              deltaid, baseid)
         ).lastrowid
 
         entry = revisionentry(
             rid=rid,
             rev=rev,
             node=node,
             p1rev=p1rev,
             p2rev=p2rev,
             p1node=p1,
             p2node=p2,
             linkrev=linkrev,
             flags=flags)
 
         self._nodetorev[node] = rev
         self._revtonode[rev] = node
         self._revisions[node] = entry
 
         return node
 
 class sqliterepository(localrepo.localrepository):
     def cancopy(self):
         return False
 
     def transaction(self, *args, **kwargs):
         current = self.currenttransaction()
 
         tr = super(sqliterepository, self).transaction(*args, **kwargs)
 
         if current:
             return tr
 
         self._dbconn.execute(r'BEGIN TRANSACTION')
 
         def committransaction(_):
             self._dbconn.commit()
 
         tr.addfinalize('sqlitestore', committransaction)
 
         return tr
 
     @property
     def _dbconn(self):
         # SQLite connections can only be used on the thread that created
         # them. In most cases, this "just works." However, hgweb uses
         # multiple threads.
         tid = threading.current_thread().ident
 
         if self._db:
             if self._db[0] == tid:
                 return self._db[1]
 
         db = makedb(self.svfs.join('db.sqlite'))
         self._db = (tid, db)
 
         return db
 
 def makedb(path):
     """Construct a database handle for a database at path."""
 
     db = sqlite3.connect(encoding.strfromlocal(path))
     db.text_factory = bytes
 
     res = db.execute(r'PRAGMA user_version').fetchone()[0]
 
     # New database.
     if res == 0:
         for statement in CREATE_SCHEMA:
             db.execute(statement)
 
         db.commit()
 
     elif res == CURRENT_SCHEMA_VERSION:
         pass
 
     else:
         raise error.Abort(_('sqlite database has unrecognized version'))
 
     db.execute(r'PRAGMA journal_mode=WAL')
 
     return db
 
 def featuresetup(ui, supported):
     supported.add(REQUIREMENT)
 
     if zstd:
         supported.add(REQUIREMENT_ZSTD)
 
     supported.add(REQUIREMENT_ZLIB)
     supported.add(REQUIREMENT_NONE)
     supported.add(REQUIREMENT_SHALLOW_FILES)
     supported.add(repository.NARROW_REQUIREMENT)
 
 def newreporequirements(orig, ui, createopts):
     if createopts['backend'] != 'sqlite':
         return orig(ui, createopts)
 
     # This restriction can be lifted once we have more confidence.
     if 'sharedrepo' in createopts:
         raise error.Abort(_('shared repositories not supported with SQLite '
                             'store'))
 
     # This filtering is out of an abundance of caution: we want to ensure
     # we honor creation options and we do that by annotating exactly the
     # creation options we recognize.
     known = {
         'narrowfiles',
         'backend',
         'shallowfilestore',
     }
 
     unsupported = set(createopts) - known
     if unsupported:
         raise error.Abort(_('SQLite store does not support repo creation '
                             'option: %s') % ', '.join(sorted(unsupported)))
 
     # Since we're a hybrid store that still relies on revlogs, we fall back
     # to using the revlogv1 backend's storage requirements then adding our
     # own requirement.
     createopts['backend'] = 'revlogv1'
     requirements = orig(ui, createopts)
     requirements.add(REQUIREMENT)
 
     compression = ui.config('storage', 'sqlite.compression')
 
     if compression == 'zstd' and not zstd:
         raise error.Abort(_('storage.sqlite.compression set to "zstd" but '
                             'zstandard compression not available to this '
                             'Mercurial install'))
 
     if compression == 'zstd':
         requirements.add(REQUIREMENT_ZSTD)
     elif compression == 'zlib':
         requirements.add(REQUIREMENT_ZLIB)
     elif compression == 'none':
         requirements.add(REQUIREMENT_NONE)
     else:
         raise error.Abort(_('unknown compression engine defined in '
                             'storage.sqlite.compression: %s') % compression)
 
     if createopts.get('shallowfilestore'):
         requirements.add(REQUIREMENT_SHALLOW_FILES)
 
     return requirements
 
 @interfaceutil.implementer(repository.ilocalrepositoryfilestorage)
 class sqlitefilestorage(object):
     """Repository file storage backed by SQLite."""
     def file(self, path):
         if path[0] == b'/':
             path = path[1:]
 
         if REQUIREMENT_ZSTD in self.requirements:
             compression = 'zstd'
         elif REQUIREMENT_ZLIB in self.requirements:
             compression = 'zlib'
         elif REQUIREMENT_NONE in self.requirements:
             compression = 'none'
         else:
             raise error.Abort(_('unable to determine what compression engine '
                                 'to use for SQLite storage'))
 
         return sqlitefilestore(self._dbconn, path, compression)
 
 def makefilestorage(orig, requirements, features, **kwargs):
     """Produce a type conforming to ``ilocalrepositoryfilestorage``."""
     if REQUIREMENT in requirements:
         if REQUIREMENT_SHALLOW_FILES in requirements:
             features.add(repository.REPO_FEATURE_SHALLOW_FILE_STORAGE)
 
         return sqlitefilestorage
     else:
         return orig(requirements=requirements, features=features, **kwargs)
 
 def makemain(orig, ui, requirements, **kwargs):
     if REQUIREMENT in requirements:
         if REQUIREMENT_ZSTD in requirements and not zstd:
             raise error.Abort(_('repository uses zstandard compression, which '
                                 'is not available to this Mercurial install'))
 
         return sqliterepository
 
     return orig(requirements=requirements, **kwargs)
 
 def verifierinit(orig, self, *args, **kwargs):
     orig(self, *args, **kwargs)
 
     # We don't care that files in the store don't align with what is
     # advertised. So suppress these warnings.
     self.warnorphanstorefiles = False
 
 def extsetup(ui):
     localrepo.featuresetupfuncs.add(featuresetup)
     extensions.wrapfunction(localrepo, 'newreporequirements',
                             newreporequirements)
     extensions.wrapfunction(localrepo, 'makefilestorage',
                             makefilestorage)
     extensions.wrapfunction(localrepo, 'makemain',
                             makemain)
     extensions.wrapfunction(verify.verifier, '__init__',
                             verifierinit)
 
 def reposetup(ui, repo):
     if isinstance(repo, sqliterepository):
         repo._db = None
 
     # TODO check for bundlerepository?

mercurial/configitems.py

@@ -1,1505 +1,1521 b''
 # configitems.py - centralized declaration of configuration option
 #
 #  Copyright 2017 Pierre-Yves David <[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 functools
 import re
 
 from . import (
     encoding,
     error,
 )
 
 def loadconfigtable(ui, extname, configtable):
     """update config item known to the ui with the extension ones"""
     for section, items in sorted(configtable.items()):
         knownitems = ui._knownconfig.setdefault(section, itemregister())
         knownkeys = set(knownitems)
         newkeys = set(items)
         for key in sorted(knownkeys & newkeys):
             msg = "extension '%s' overwrite config item '%s.%s'"
             msg %= (extname, section, key)
             ui.develwarn(msg, config='warn-config')
 
         knownitems.update(items)
 
 class configitem(object):
     """represent a known config item
 
     :section: the official config section where to find this item,
        :name: the official name within the section,
     :default: default value for this item,
     :alias: optional list of tuples as alternatives,
     :generic: this is a generic definition, match name using regular expression.
     """
 
     def __init__(self, section, name, default=None, alias=(),
-                 generic=False, priority=0):
+                 generic=False, priority=0, experimental=False):
         self.section = section
         self.name = name
         self.default = default
         self.alias = list(alias)
         self.generic = generic
         self.priority = priority
+        self.experimental = experimental
         self._re = None
         if generic:
             self._re = re.compile(self.name)
 
 class itemregister(dict):
     """A specialized dictionary that can handle wild-card selection"""
 
     def __init__(self):
         super(itemregister, self).__init__()
         self._generics = set()
 
     def update(self, other):
         super(itemregister, self).update(other)
         self._generics.update(other._generics)
 
     def __setitem__(self, key, item):
         super(itemregister, self).__setitem__(key, item)
         if item.generic:
             self._generics.add(item)
 
     def get(self, key):
         baseitem = super(itemregister, self).get(key)
         if baseitem is not None and not baseitem.generic:
             return baseitem
 
         # search for a matching generic item
         generics = sorted(self._generics, key=(lambda x: (x.priority, x.name)))
         for item in generics:
             # we use 'match' instead of 'search' to make the matching simpler
             # for people unfamiliar with regular expression. Having the match
             # rooted to the start of the string will produce less surprising
             # result for user writing simple regex for sub-attribute.
             #
             # For example using "color\..*" match produces an unsurprising
             # result, while using search could suddenly match apparently
             # unrelated configuration that happens to contains "color."
             # anywhere. This is a tradeoff where we favor requiring ".*" on
             # some match to avoid the need to prefix most pattern with "^".
             # The "^" seems more error prone.
             if item._re.match(key):
                 return item
 
         return None
 
 coreitems = {}
 
 def _register(configtable, *args, **kwargs):
     item = configitem(*args, **kwargs)
     section = configtable.setdefault(item.section, itemregister())
     if item.name in section:
         msg = "duplicated config item registration for '%s.%s'"
         raise error.ProgrammingError(msg % (item.section, item.name))
     section[item.name] = item
 
 # special value for case where the default is derived from other values
 dynamicdefault = object()
 
 # Registering actual config items
 
 def getitemregister(configtable):
     f = functools.partial(_register, configtable)
     # export pseudo enum as configitem.*
     f.dynamicdefault = dynamicdefault
     return f
 
 coreconfigitem = getitemregister(coreitems)
 
 def _registerdiffopts(section, configprefix=''):
     coreconfigitem(section, configprefix + 'nodates',
         default=False,
     )
     coreconfigitem(section, configprefix + 'showfunc',
         default=False,
     )
     coreconfigitem(section, configprefix + 'unified',
         default=None,
     )
     coreconfigitem(section, configprefix + 'git',
         default=False,
     )
     coreconfigitem(section, configprefix + 'ignorews',
         default=False,
     )
     coreconfigitem(section, configprefix + 'ignorewsamount',
         default=False,
     )
     coreconfigitem(section, configprefix + 'ignoreblanklines',
         default=False,
     )
     coreconfigitem(section, configprefix + 'ignorewseol',
         default=False,
     )
     coreconfigitem(section, configprefix + 'nobinary',
         default=False,
     )
     coreconfigitem(section, configprefix + 'noprefix',
         default=False,
     )
     coreconfigitem(section, configprefix + 'word-diff',
         default=False,
     )
 
 coreconfigitem('alias', '.*',
     default=dynamicdefault,
     generic=True,
 )
 coreconfigitem('auth', 'cookiefile',
     default=None,
 )
 _registerdiffopts(section='annotate')
 # bookmarks.pushing: internal hack for discovery
 coreconfigitem('bookmarks', 'pushing',
     default=list,
 )
 # bundle.mainreporoot: internal hack for bundlerepo
 coreconfigitem('bundle', 'mainreporoot',
     default='',
 )
 coreconfigitem('censor', 'policy',
     default='abort',
+    experimental=True,
 )
 coreconfigitem('chgserver', 'idletimeout',
     default=3600,
 )
 coreconfigitem('chgserver', 'skiphash',
     default=False,
 )
 coreconfigitem('cmdserver', 'log',
     default=None,
 )
 coreconfigitem('cmdserver', 'max-log-files',
     default=7,
 )
 coreconfigitem('cmdserver', 'max-log-size',
     default='1 MB',
 )
 coreconfigitem('cmdserver', 'max-repo-cache',
     default=0,
+    experimental=True,
 )
 coreconfigitem('cmdserver', 'message-encodings',
     default=list,
+    experimental=True,
 )
 coreconfigitem('cmdserver', 'track-log',
     default=lambda: ['chgserver', 'cmdserver', 'repocache'],
 )
 coreconfigitem('color', '.*',
     default=None,
     generic=True,
 )
 coreconfigitem('color', 'mode',
     default='auto',
 )
 coreconfigitem('color', 'pagermode',
     default=dynamicdefault,
 )
 _registerdiffopts(section='commands', configprefix='commit.interactive.')
 coreconfigitem('commands', 'commit.post-status',
     default=False,
 )
 coreconfigitem('commands', 'grep.all-files',
     default=False,
+    experimental=True,
 )
 coreconfigitem('commands', 'resolve.confirm',
     default=False,
 )
 coreconfigitem('commands', 'resolve.explicit-re-merge',
     default=False,
 )
 coreconfigitem('commands', 'resolve.mark-check',
     default='none',
 )
 _registerdiffopts(section='commands', configprefix='revert.interactive.')
 coreconfigitem('commands', 'show.aliasprefix',
     default=list,
 )
 coreconfigitem('commands', 'status.relative',
     default=False,
 )
 coreconfigitem('commands', 'status.skipstates',
     default=[],
+    experimental=True,
 )
 coreconfigitem('commands', 'status.terse',
     default='',
 )
 coreconfigitem('commands', 'status.verbose',
     default=False,
 )
 coreconfigitem('commands', 'update.check',
     default=None,
 )
 coreconfigitem('commands', 'update.requiredest',
     default=False,
 )
 coreconfigitem('committemplate', '.*',
     default=None,
     generic=True,
 )
 coreconfigitem('convert', 'bzr.saverev',
     default=True,
 )
 coreconfigitem('convert', 'cvsps.cache',
     default=True,
 )
 coreconfigitem('convert', 'cvsps.fuzz',
     default=60,
 )
 coreconfigitem('convert', 'cvsps.logencoding',
     default=None,
 )
 coreconfigitem('convert', 'cvsps.mergefrom',
     default=None,
 )
 coreconfigitem('convert', 'cvsps.mergeto',
     default=None,
 )
 coreconfigitem('convert', 'git.committeractions',
     default=lambda: ['messagedifferent'],
 )
 coreconfigitem('convert', 'git.extrakeys',
     default=list,
 )
 coreconfigitem('convert', 'git.findcopiesharder',
     default=False,
 )
 coreconfigitem('convert', 'git.remoteprefix',
     default='remote',
 )
 coreconfigitem('convert', 'git.renamelimit',
     default=400,
 )
 coreconfigitem('convert', 'git.saverev',
     default=True,
 )
 coreconfigitem('convert', 'git.similarity',
     default=50,
 )
 coreconfigitem('convert', 'git.skipsubmodules',
     default=False,
 )
 coreconfigitem('convert', 'hg.clonebranches',
     default=False,
 )
 coreconfigitem('convert', 'hg.ignoreerrors',
     default=False,
 )
 coreconfigitem('convert', 'hg.preserve-hash',
     default=False,
 )
 coreconfigitem('convert', 'hg.revs',
     default=None,
 )
 coreconfigitem('convert', 'hg.saverev',
     default=False,
 )
 coreconfigitem('convert', 'hg.sourcename',
     default=None,
 )
 coreconfigitem('convert', 'hg.startrev',
     default=None,
 )
 coreconfigitem('convert', 'hg.tagsbranch',
     default='default',
 )
 coreconfigitem('convert', 'hg.usebranchnames',
     default=True,
 )
 coreconfigitem('convert', 'ignoreancestorcheck',
     default=False,
+    experimental=True,
 )
 coreconfigitem('convert', 'localtimezone',
     default=False,
 )
 coreconfigitem('convert', 'p4.encoding',
     default=dynamicdefault,
 )
 coreconfigitem('convert', 'p4.startrev',
     default=0,
 )
 coreconfigitem('convert', 'skiptags',
     default=False,
 )
 coreconfigitem('convert', 'svn.debugsvnlog',
     default=True,
 )
 coreconfigitem('convert', 'svn.trunk',
     default=None,
 )
 coreconfigitem('convert', 'svn.tags',
     default=None,
 )
 coreconfigitem('convert', 'svn.branches',
     default=None,
 )
 coreconfigitem('convert', 'svn.startrev',
     default=0,
 )
 coreconfigitem('debug', 'dirstate.delaywrite',
     default=0,
 )
 coreconfigitem('defaults', '.*',
     default=None,
     generic=True,
 )
 coreconfigitem('devel', 'all-warnings',
     default=False,
 )
 coreconfigitem('devel', 'bundle2.debug',
     default=False,
 )
 coreconfigitem('devel', 'bundle.delta',
     default='',
 )
 coreconfigitem('devel', 'cache-vfs',
     default=None,
 )
 coreconfigitem('devel', 'check-locks',
     default=False,
 )
 coreconfigitem('devel', 'check-relroot',
     default=False,
 )
 coreconfigitem('devel', 'default-date',
     default=None,
 )
 coreconfigitem('devel', 'deprec-warn',
     default=False,
 )
 coreconfigitem('devel', 'disableloaddefaultcerts',
     default=False,
 )
 coreconfigitem('devel', 'warn-empty-changegroup',
     default=False,
 )
 coreconfigitem('devel', 'legacy.exchange',
     default=list,
 )
 coreconfigitem('devel', 'servercafile',
     default='',
 )
 coreconfigitem('devel', 'serverexactprotocol',
     default='',
 )
 coreconfigitem('devel', 'serverrequirecert',
     default=False,
 )
 coreconfigitem('devel', 'strip-obsmarkers',
     default=True,
 )
 coreconfigitem('devel', 'warn-config',
     default=None,
 )
 coreconfigitem('devel', 'warn-config-default',
     default=None,
 )
 coreconfigitem('devel', 'user.obsmarker',
     default=None,
 )
 coreconfigitem('devel', 'warn-config-unknown',
     default=None,
 )
 coreconfigitem('devel', 'debug.copies',
     default=False,
 )
 coreconfigitem('devel', 'debug.extensions',
     default=False,
 )
 coreconfigitem('devel', 'debug.peer-request',
     default=False,
 )
 coreconfigitem('devel', 'discovery.randomize',
     default=True,
 )
 _registerdiffopts(section='diff')
 coreconfigitem('email', 'bcc',
     default=None,
 )
 coreconfigitem('email', 'cc',
     default=None,
 )
 coreconfigitem('email', 'charsets',
     default=list,
 )
 coreconfigitem('email', 'from',
     default=None,
 )
 coreconfigitem('email', 'method',
     default='smtp',
 )
 coreconfigitem('email', 'reply-to',
     default=None,
 )
 coreconfigitem('email', 'to',
     default=None,
 )
 coreconfigitem('experimental', 'archivemetatemplate',
     default=dynamicdefault,
 )
 coreconfigitem('experimental', 'auto-publish',
     default='publish',
 )
 coreconfigitem('experimental', 'bundle-phases',
     default=False,
 )
 coreconfigitem('experimental', 'bundle2-advertise',
     default=True,
 )
 coreconfigitem('experimental', 'bundle2-output-capture',
     default=False,
 )
 coreconfigitem('experimental', 'bundle2.pushback',
     default=False,
 )
 coreconfigitem('experimental', 'bundle2lazylocking',
     default=False,
 )
 coreconfigitem('experimental', 'bundlecomplevel',
     default=None,
 )
 coreconfigitem('experimental', 'bundlecomplevel.bzip2',
     default=None,
 )
 coreconfigitem('experimental', 'bundlecomplevel.gzip',
     default=None,
 )
 coreconfigitem('experimental', 'bundlecomplevel.none',
     default=None,
 )
 coreconfigitem('experimental', 'bundlecomplevel.zstd',
     default=None,
 )
 coreconfigitem('experimental', 'changegroup3',
     default=False,
 )
 coreconfigitem('experimental', 'cleanup-as-archived',
     default=False,
 )
 coreconfigitem('experimental', 'clientcompressionengines',
     default=list,
 )
 coreconfigitem('experimental', 'copytrace',
     default='on',
 )
 coreconfigitem('experimental', 'copytrace.movecandidateslimit',
     default=100,
 )
 coreconfigitem('experimental', 'copytrace.sourcecommitlimit',
     default=100,
 )
 coreconfigitem('experimental', 'copies.read-from',
     default="filelog-only",
 )
 coreconfigitem('experimental', 'copies.write-to',
     default='filelog-only',
 )
 coreconfigitem('experimental', 'crecordtest',
     default=None,
 )
 coreconfigitem('experimental', 'directaccess',
     default=False,
 )
 coreconfigitem('experimental', 'directaccess.revnums',
     default=False,
 )
 coreconfigitem('experimental', 'editortmpinhg',
     default=False,
 )
 coreconfigitem('experimental', 'evolution',
     default=list,
 )
 coreconfigitem('experimental', 'evolution.allowdivergence',
     default=False,
     alias=[('experimental', 'allowdivergence')]
 )
 coreconfigitem('experimental', 'evolution.allowunstable',
     default=None,
 )
 coreconfigitem('experimental', 'evolution.createmarkers',
     default=None,
 )
 coreconfigitem('experimental', 'evolution.effect-flags',
     default=True,
     alias=[('experimental', 'effect-flags')]
 )
 coreconfigitem('experimental', 'evolution.exchange',
     default=None,
 )
 coreconfigitem('experimental', 'evolution.bundle-obsmarker',
     default=False,
 )
 coreconfigitem('experimental', 'log.topo',
     default=False,
 )
 coreconfigitem('experimental', 'evolution.report-instabilities',
     default=True,
 )
 coreconfigitem('experimental', 'evolution.track-operation',
     default=True,
 )
 # repo-level config to exclude a revset visibility
 #
 # The target use case is to use `share` to expose different subset of the same
 # repository, especially server side. See also `server.view`.
 coreconfigitem('experimental', 'extra-filter-revs',
     default=None,
 )
 coreconfigitem('experimental', 'maxdeltachainspan',
     default=-1,
 )
 coreconfigitem('experimental', 'mergetempdirprefix',
     default=None,
 )
 coreconfigitem('experimental', 'mmapindexthreshold',
     default=None,
 )
 coreconfigitem('experimental', 'narrow',
     default=False,
 )
 coreconfigitem('experimental', 'nonnormalparanoidcheck',
     default=False,
 )
 coreconfigitem('experimental', 'exportableenviron',
     default=list,
 )
 coreconfigitem('experimental', 'extendedheader.index',
     default=None,
 )
 coreconfigitem('experimental', 'extendedheader.similarity',
     default=False,
 )
 coreconfigitem('experimental', 'graphshorten',
     default=False,
 )
 coreconfigitem('experimental', 'graphstyle.parent',
     default=dynamicdefault,
 )
 coreconfigitem('experimental', 'graphstyle.missing',
     default=dynamicdefault,
 )
 coreconfigitem('experimental', 'graphstyle.grandparent',
     default=dynamicdefault,
 )
 coreconfigitem('experimental', 'hook-track-tags',
     default=False,
 )
 coreconfigitem('experimental', 'httppeer.advertise-v2',
     default=False,
 )
 coreconfigitem('experimental', 'httppeer.v2-encoder-order',
     default=None,
 )
 coreconfigitem('experimental', 'httppostargs',
     default=False,
 )
 coreconfigitem('experimental', 'mergedriver',
     default=None,
 )
 coreconfigitem('experimental', 'nointerrupt', default=False)
 coreconfigitem('experimental', 'nointerrupt-interactiveonly', default=True)
 
 coreconfigitem('experimental', 'obsmarkers-exchange-debug',
     default=False,
 )
 coreconfigitem('experimental', 'remotenames',
     default=False,
 )
 coreconfigitem('experimental', 'removeemptydirs',
     default=True,
 )
 coreconfigitem('experimental', 'revert.interactive.select-to-keep',
     default=False,
 )
 coreconfigitem('experimental', 'revisions.prefixhexnode',
     default=False,
 )
 coreconfigitem('experimental', 'revlogv2',
     default=None,
 )
 coreconfigitem('experimental', 'revisions.disambiguatewithin',
     default=None,
 )
 coreconfigitem('experimental', 'server.filesdata.recommended-batch-size',
     default=50000,
 )
 coreconfigitem('experimental', 'server.manifestdata.recommended-batch-size',
     default=100000,
 )
 coreconfigitem('experimental', 'server.stream-narrow-clones',
     default=False,
 )
 coreconfigitem('experimental', 'single-head-per-branch',
     default=False,
 )
 coreconfigitem('experimental', 'sshserver.support-v2',
     default=False,
 )
 coreconfigitem('experimental', 'sparse-read',
     default=False,
 )
 coreconfigitem('experimental', 'sparse-read.density-threshold',
     default=0.50,
 )
 coreconfigitem('experimental', 'sparse-read.min-gap-size',
     default='65K',
 )
 coreconfigitem('experimental', 'treemanifest',
     default=False,
 )
 coreconfigitem('experimental', 'update.atomic-file',
     default=False,
 )
 coreconfigitem('experimental', 'sshpeer.advertise-v2',
     default=False,
 )
 coreconfigitem('experimental', 'web.apiserver',
     default=False,
 )
 coreconfigitem('experimental', 'web.api.http-v2',
     default=False,
 )
 coreconfigitem('experimental', 'web.api.debugreflect',
     default=False,
 )
 coreconfigitem('experimental', 'worker.wdir-get-thread-safe',
     default=False,
 )
 coreconfigitem('experimental', 'xdiff',
     default=False,
 )
 coreconfigitem('extensions', '.*',
     default=None,
     generic=True,
 )
 coreconfigitem('extdata', '.*',
     default=None,
     generic=True,
 )
 coreconfigitem('format', 'bookmarks-in-store',
     default=False,
 )
 coreconfigitem('format', 'chunkcachesize',
     default=None,
+    experimental=True,
 )
 coreconfigitem('format', 'dotencode',
     default=True,
 )
 coreconfigitem('format', 'generaldelta',
     default=False,
+    experimental=True,
 )
 coreconfigitem('format', 'manifestcachesize',
     default=None,
+    experimental=True,
 )
 coreconfigitem('format', 'maxchainlen',
     default=dynamicdefault,
+    experimental=True,
 )
 coreconfigitem('format', 'obsstore-version',
     default=None,
 )
 coreconfigitem('format', 'sparse-revlog',
     default=True,
 )
 coreconfigitem('format', 'revlog-compression',
     default='zlib',
     alias=[('experimental', 'format.compression')]
 )
 coreconfigitem('format', 'usefncache',
     default=True,
 )
 coreconfigitem('format', 'usegeneraldelta',
     default=True,
 )
 coreconfigitem('format', 'usestore',
     default=True,
 )
 coreconfigitem('format', 'internal-phase',
     default=False,
+    experimental=True,
 )
 coreconfigitem('fsmonitor', 'warn_when_unused',
     default=True,
 )
 coreconfigitem('fsmonitor', 'warn_update_file_count',
     default=50000,
 )
 coreconfigitem('help', br'hidden-command\..*',
     default=False,
     generic=True,
 )
 coreconfigitem('help', br'hidden-topic\..*',
     default=False,
     generic=True,
 )
 coreconfigitem('hooks', '.*',
     default=dynamicdefault,
     generic=True,
 )
 coreconfigitem('hgweb-paths', '.*',
     default=list,
     generic=True,
 )
 coreconfigitem('hostfingerprints', '.*',
     default=list,
     generic=True,
 )
 coreconfigitem('hostsecurity', 'ciphers',
     default=None,
 )
 coreconfigitem('hostsecurity', 'disabletls10warning',
     default=False,
 )
 coreconfigitem('hostsecurity', 'minimumprotocol',
     default=dynamicdefault,
 )
 coreconfigitem('hostsecurity', '.*:minimumprotocol$',
     default=dynamicdefault,
     generic=True,
 )
 coreconfigitem('hostsecurity', '.*:ciphers$',
     default=dynamicdefault,
     generic=True,
 )
 coreconfigitem('hostsecurity', '.*:fingerprints$',
     default=list,
     generic=True,
 )
 coreconfigitem('hostsecurity', '.*:verifycertsfile$',
     default=None,
     generic=True,
 )
 
 coreconfigitem('http_proxy', 'always',
     default=False,
 )
 coreconfigitem('http_proxy', 'host',
     default=None,
 )
 coreconfigitem('http_proxy', 'no',
     default=list,
 )
 coreconfigitem('http_proxy', 'passwd',
     default=None,
 )
 coreconfigitem('http_proxy', 'user',
     default=None,
 )
 
 coreconfigitem('http', 'timeout',
     default=None,
 )
 
 coreconfigitem('logtoprocess', 'commandexception',
     default=None,
 )
 coreconfigitem('logtoprocess', 'commandfinish',
     default=None,
 )
 coreconfigitem('logtoprocess', 'command',
     default=None,
 )
 coreconfigitem('logtoprocess', 'develwarn',
     default=None,
 )
 coreconfigitem('logtoprocess', 'uiblocked',
     default=None,
 )
 coreconfigitem('merge', 'checkunknown',
     default='abort',
 )
 coreconfigitem('merge', 'checkignored',
     default='abort',
 )
 coreconfigitem('experimental', 'merge.checkpathconflicts',
     default=False,
 )
 coreconfigitem('merge', 'followcopies',
     default=True,
 )
 coreconfigitem('merge', 'on-failure',
     default='continue',
 )
 coreconfigitem('merge', 'preferancestor',
         default=lambda: ['*'],
+        experimental=True,
 )
 coreconfigitem('merge', 'strict-capability-check',
     default=False,
 )
 coreconfigitem('merge-tools', '.*',
     default=None,
     generic=True,
 )
 coreconfigitem('merge-tools', br'.*\.args$',
     default="$local $base $other",
     generic=True,
     priority=-1,
 )
 coreconfigitem('merge-tools', br'.*\.binary$',
     default=False,
     generic=True,
     priority=-1,
 )
 coreconfigitem('merge-tools', br'.*\.check$',
     default=list,
     generic=True,
     priority=-1,
 )
 coreconfigitem('merge-tools', br'.*\.checkchanged$',
     default=False,
     generic=True,
     priority=-1,
 )
 coreconfigitem('merge-tools', br'.*\.executable$',
     default=dynamicdefault,
     generic=True,
     priority=-1,
 )
 coreconfigitem('merge-tools', br'.*\.fixeol$',
     default=False,
     generic=True,
     priority=-1,
 )
 coreconfigitem('merge-tools', br'.*\.gui$',
     default=False,
     generic=True,
     priority=-1,
 )
 coreconfigitem('merge-tools', br'.*\.mergemarkers$',
     default='basic',
     generic=True,
     priority=-1,
 )
 coreconfigitem('merge-tools', br'.*\.mergemarkertemplate$',
     default=dynamicdefault,  # take from ui.mergemarkertemplate
     generic=True,
     priority=-1,
 )
 coreconfigitem('merge-tools', br'.*\.priority$',
     default=0,
     generic=True,
     priority=-1,
 )
 coreconfigitem('merge-tools', br'.*\.premerge$',
     default=dynamicdefault,
     generic=True,
     priority=-1,
 )
 coreconfigitem('merge-tools', br'.*\.symlink$',
     default=False,
     generic=True,
     priority=-1,
 )
 coreconfigitem('pager', 'attend-.*',
     default=dynamicdefault,
     generic=True,
 )
 coreconfigitem('pager', 'ignore',
     default=list,
 )
 coreconfigitem('pager', 'pager',
     default=dynamicdefault,
 )
 coreconfigitem('patch', 'eol',
     default='strict',
 )
 coreconfigitem('patch', 'fuzz',
     default=2,
 )
 coreconfigitem('paths', 'default',
     default=None,
 )
 coreconfigitem('paths', 'default-push',
     default=None,
 )
 coreconfigitem('paths', '.*',
     default=None,
     generic=True,
 )
 coreconfigitem('phases', 'checksubrepos',
     default='follow',
 )
 coreconfigitem('phases', 'new-commit',
     default='draft',
 )
 coreconfigitem('phases', 'publish',
     default=True,
 )
 coreconfigitem('profiling', 'enabled',
     default=False,
 )
 coreconfigitem('profiling', 'format',
     default='text',
 )
 coreconfigitem('profiling', 'freq',
     default=1000,
 )
 coreconfigitem('profiling', 'limit',
     default=30,
 )
 coreconfigitem('profiling', 'nested',
     default=0,
 )
 coreconfigitem('profiling', 'output',
     default=None,
 )
 coreconfigitem('profiling', 'showmax',
     default=0.999,
 )
 coreconfigitem('profiling', 'showmin',
     default=dynamicdefault,
 )
 coreconfigitem('profiling', 'showtime',
     default=True,
 )
 coreconfigitem('profiling', 'sort',
     default='inlinetime',
 )
 coreconfigitem('profiling', 'statformat',
     default='hotpath',
 )
 coreconfigitem('profiling', 'time-track',
     default=dynamicdefault,
 )
 coreconfigitem('profiling', 'type',
     default='stat',
 )
 coreconfigitem('progress', 'assume-tty',
     default=False,
 )
 coreconfigitem('progress', 'changedelay',
     default=1,
 )
 coreconfigitem('progress', 'clear-complete',
     default=True,
 )
 coreconfigitem('progress', 'debug',
     default=False,
 )
 coreconfigitem('progress', 'delay',
     default=3,
 )
 coreconfigitem('progress', 'disable',
     default=False,
 )
 coreconfigitem('progress', 'estimateinterval',
     default=60.0,
 )
 coreconfigitem('progress', 'format',
     default=lambda: ['topic', 'bar', 'number', 'estimate'],
 )
 coreconfigitem('progress', 'refresh',
     default=0.1,
 )
 coreconfigitem('progress', 'width',
     default=dynamicdefault,
 )
 coreconfigitem('push', 'pushvars.server',
     default=False,
 )
 coreconfigitem('rewrite', 'backup-bundle',
     default=True,
     alias=[('ui', 'history-editing-backup')],
 )
 coreconfigitem('rewrite', 'update-timestamp',
     default=False,
 )
 coreconfigitem('storage', 'new-repo-backend',
     default='revlogv1',
+    experimental=True,
 )
 coreconfigitem('storage', 'revlog.optimize-delta-parent-choice',
     default=True,
     alias=[('format', 'aggressivemergedeltas')],
 )
 coreconfigitem('storage', 'revlog.reuse-external-delta',
     default=True,
 )
 coreconfigitem('storage', 'revlog.reuse-external-delta-parent',
     default=None,
 )
 coreconfigitem('storage', 'revlog.zlib.level',
     default=None,
 )
 coreconfigitem('storage', 'revlog.zstd.level',
     default=None,
 )
 coreconfigitem('server', 'bookmarks-pushkey-compat',
     default=True,
 )
 coreconfigitem('server', 'bundle1',
     default=True,
 )
 coreconfigitem('server', 'bundle1gd',
     default=None,
 )
 coreconfigitem('server', 'bundle1.pull',
     default=None,
 )
 coreconfigitem('server', 'bundle1gd.pull',
     default=None,
 )
 coreconfigitem('server', 'bundle1.push',
     default=None,
 )
 coreconfigitem('server', 'bundle1gd.push',
     default=None,
 )
 coreconfigitem('server', 'bundle2.stream',
     default=True,
     alias=[('experimental', 'bundle2.stream')]
 )
 coreconfigitem('server', 'compressionengines',
     default=list,
 )
 coreconfigitem('server', 'concurrent-push-mode',
     default='strict',
 )
 coreconfigitem('server', 'disablefullbundle',
     default=False,
 )
 coreconfigitem('server', 'maxhttpheaderlen',
     default=1024,
 )
 coreconfigitem('server', 'pullbundle',
     default=False,
 )
 coreconfigitem('server', 'preferuncompressed',
     default=False,
 )
 coreconfigitem('server', 'streamunbundle',
     default=False,
 )
 coreconfigitem('server', 'uncompressed',
     default=True,
 )
 coreconfigitem('server', 'uncompressedallowsecret',
     default=False,
 )
 coreconfigitem('server', 'view',
     default='served',
 )
 coreconfigitem('server', 'validate',
     default=False,
 )
 coreconfigitem('server', 'zliblevel',
     default=-1,
 )
 coreconfigitem('server', 'zstdlevel',
     default=3,
 )
 coreconfigitem('share', 'pool',
     default=None,
 )
 coreconfigitem('share', 'poolnaming',
     default='identity',
 )
 coreconfigitem('shelve','maxbackups',
     default=10,
 )
 coreconfigitem('smtp', 'host',
     default=None,
 )
 coreconfigitem('smtp', 'local_hostname',
     default=None,
 )
 coreconfigitem('smtp', 'password',
     default=None,
 )
 coreconfigitem('smtp', 'port',
     default=dynamicdefault,
 )
 coreconfigitem('smtp', 'tls',
     default='none',
 )
 coreconfigitem('smtp', 'username',
     default=None,
 )
 coreconfigitem('sparse', 'missingwarning',
     default=True,
+    experimental=True,
 )
 coreconfigitem('subrepos', 'allowed',
     default=dynamicdefault,  # to make backporting simpler
 )
 coreconfigitem('subrepos', 'hg:allowed',
     default=dynamicdefault,
 )
 coreconfigitem('subrepos', 'git:allowed',
     default=dynamicdefault,
 )
 coreconfigitem('subrepos', 'svn:allowed',
     default=dynamicdefault,
 )
 coreconfigitem('templates', '.*',
     default=None,
     generic=True,
 )
 coreconfigitem('templateconfig', '.*',
     default=dynamicdefault,
     generic=True,
 )
 coreconfigitem('trusted', 'groups',
     default=list,
 )
 coreconfigitem('trusted', 'users',
     default=list,
 )
 coreconfigitem('ui', '_usedassubrepo',
     default=False,
 )
 coreconfigitem('ui', 'allowemptycommit',
     default=False,
 )
 coreconfigitem('ui', 'archivemeta',
     default=True,
 )
 coreconfigitem('ui', 'askusername',
     default=False,
 )
 coreconfigitem('ui', 'clonebundlefallback',
     default=False,
 )
 coreconfigitem('ui', 'clonebundleprefers',
     default=list,
 )
 coreconfigitem('ui', 'clonebundles',
     default=True,
 )
 coreconfigitem('ui', 'color',
     default='auto',
 )
 coreconfigitem('ui', 'commitsubrepos',
     default=False,
 )
 coreconfigitem('ui', 'debug',
     default=False,
 )
 coreconfigitem('ui', 'debugger',
     default=None,
 )
 coreconfigitem('ui', 'editor',
     default=dynamicdefault,
 )
 coreconfigitem('ui', 'fallbackencoding',
     default=None,
 )
 coreconfigitem('ui', 'forcecwd',
     default=None,
 )
 coreconfigitem('ui', 'forcemerge',
     default=None,
 )
 coreconfigitem('ui', 'formatdebug',
     default=False,
 )
 coreconfigitem('ui', 'formatjson',
     default=False,
 )
 coreconfigitem('ui', 'formatted',
     default=None,
 )
 coreconfigitem('ui', 'graphnodetemplate',
     default=None,
 )
 coreconfigitem('ui', 'interactive',
     default=None,
 )
 coreconfigitem('ui', 'interface',
     default=None,
 )
 coreconfigitem('ui', 'interface.chunkselector',
     default=None,
 )
 coreconfigitem('ui', 'large-file-limit',
     default=10000000,
 )
 coreconfigitem('ui', 'logblockedtimes',
     default=False,
 )
 coreconfigitem('ui', 'logtemplate',
     default=None,
 )
 coreconfigitem('ui', 'merge',
     default=None,
 )
 coreconfigitem('ui', 'mergemarkers',
     default='basic',
 )
 coreconfigitem('ui', 'mergemarkertemplate',
     default=('{node|short} '
             '{ifeq(tags, "tip", "", '
             'ifeq(tags, "", "", "{tags} "))}'
             '{if(bookmarks, "{bookmarks} ")}'
             '{ifeq(branch, "default", "", "{branch} ")}'
             '- {author|user}: {desc|firstline}')
 )
 coreconfigitem('ui', 'message-output',
     default='stdio',
 )
 coreconfigitem('ui', 'nontty',
     default=False,
 )
 coreconfigitem('ui', 'origbackuppath',
     default=None,
 )
 coreconfigitem('ui', 'paginate',
     default=True,
 )
 coreconfigitem('ui', 'patch',
     default=None,
 )
 coreconfigitem('ui', 'pre-merge-tool-output-template',
     default=None,
 )
 coreconfigitem('ui', 'portablefilenames',
     default='warn',
 )
 coreconfigitem('ui', 'promptecho',
     default=False,
 )
 coreconfigitem('ui', 'quiet',
     default=False,
 )
 coreconfigitem('ui', 'quietbookmarkmove',
     default=False,
 )
 coreconfigitem('ui', 'relative-paths',
     default='legacy',
 )
 coreconfigitem('ui', 'remotecmd',
     default='hg',
 )
 coreconfigitem('ui', 'report_untrusted',
     default=True,
 )
 coreconfigitem('ui', 'rollback',
     default=True,
 )
 coreconfigitem('ui', 'signal-safe-lock',
     default=True,
 )
 coreconfigitem('ui', 'slash',
     default=False,
 )
 coreconfigitem('ui', 'ssh',
     default='ssh',
 )
 coreconfigitem('ui', 'ssherrorhint',
     default=None,
 )
 coreconfigitem('ui', 'statuscopies',
     default=False,
 )
 coreconfigitem('ui', 'strict',
     default=False,
 )
 coreconfigitem('ui', 'style',
     default='',
 )
 coreconfigitem('ui', 'supportcontact',
     default=None,
 )
 coreconfigitem('ui', 'textwidth',
     default=78,
 )
 coreconfigitem('ui', 'timeout',
     default='600',
 )
 coreconfigitem('ui', 'timeout.warn',
     default=0,
 )
 coreconfigitem('ui', 'traceback',
     default=False,
 )
 coreconfigitem('ui', 'tweakdefaults',
     default=False,
 )
 coreconfigitem('ui', 'username',
     alias=[('ui', 'user')]
 )
 coreconfigitem('ui', 'verbose',
     default=False,
 )
 coreconfigitem('verify', 'skipflags',
     default=None,
 )
 coreconfigitem('web', 'allowbz2',
     default=False,
 )
 coreconfigitem('web', 'allowgz',
     default=False,
 )
 coreconfigitem('web', 'allow-pull',
     alias=[('web', 'allowpull')],
     default=True,
 )
 coreconfigitem('web', 'allow-push',
     alias=[('web', 'allow_push')],
     default=list,
 )
 coreconfigitem('web', 'allowzip',
     default=False,
 )
 coreconfigitem('web', 'archivesubrepos',
     default=False,
 )
 coreconfigitem('web', 'cache',
     default=True,
 )
 coreconfigitem('web', 'comparisoncontext',
     default=5,
 )
 coreconfigitem('web', 'contact',
     default=None,
 )
 coreconfigitem('web', 'deny_push',
     default=list,
 )
 coreconfigitem('web', 'guessmime',
     default=False,
 )
 coreconfigitem('web', 'hidden',
     default=False,
 )
 coreconfigitem('web', 'labels',
     default=list,
 )
 coreconfigitem('web', 'logoimg',
     default='hglogo.png',
 )
 coreconfigitem('web', 'logourl',
     default='https://mercurial-scm.org/',
 )
 coreconfigitem('web', 'accesslog',
     default='-',
 )
 coreconfigitem('web', 'address',
     default='',
 )
 coreconfigitem('web', 'allow-archive',
     alias=[('web', 'allow_archive')],
     default=list,
 )
 coreconfigitem('web', 'allow_read',
     default=list,
 )
 coreconfigitem('web', 'baseurl',
     default=None,
 )
 coreconfigitem('web', 'cacerts',
     default=None,
 )
 coreconfigitem('web', 'certificate',
     default=None,
 )
 coreconfigitem('web', 'collapse',
     default=False,
 )
 coreconfigitem('web', 'csp',
     default=None,
 )
 coreconfigitem('web', 'deny_read',
     default=list,
 )
 coreconfigitem('web', 'descend',
     default=True,
 )
 coreconfigitem('web', 'description',
     default="",
 )
 coreconfigitem('web', 'encoding',
     default=lambda: encoding.encoding,
 )
 coreconfigitem('web', 'errorlog',
     default='-',
 )
 coreconfigitem('web', 'ipv6',
     default=False,
 )
 coreconfigitem('web', 'maxchanges',
     default=10,
 )
 coreconfigitem('web', 'maxfiles',
     default=10,
 )
 coreconfigitem('web', 'maxshortchanges',
     default=60,
 )
 coreconfigitem('web', 'motd',
     default='',
 )
 coreconfigitem('web', 'name',
     default=dynamicdefault,
 )
 coreconfigitem('web', 'port',
     default=8000,
 )
 coreconfigitem('web', 'prefix',
     default='',
 )
 coreconfigitem('web', 'push_ssl',
     default=True,
 )
 coreconfigitem('web', 'refreshinterval',
     default=20,
 )
 coreconfigitem('web', 'server-header',
     default=None,
 )
 coreconfigitem('web', 'static',
     default=None,
 )
 coreconfigitem('web', 'staticurl',
     default=None,
 )
 coreconfigitem('web', 'stripes',
     default=1,
 )
 coreconfigitem('web', 'style',
     default='paper',
 )
 coreconfigitem('web', 'templates',
     default=None,
 )
 coreconfigitem('web', 'view',
     default='served',
+    experimental=True,
 )
 coreconfigitem('worker', 'backgroundclose',
     default=dynamicdefault,
 )
 # Windows defaults to a limit of 512 open files. A buffer of 128
 # should give us enough headway.
 coreconfigitem('worker', 'backgroundclosemaxqueue',
     default=384,
 )
 coreconfigitem('worker', 'backgroundcloseminfilecount',
     default=2048,
 )
 coreconfigitem('worker', 'backgroundclosethreadcount',
     default=4,
 )
 coreconfigitem('worker', 'enabled',
     default=True,
 )
 coreconfigitem('worker', 'numcpus',
     default=None,
 )
 
 # Rebase related configuration moved to core because other extension are doing
 # strange things. For example, shelve import the extensions to reuse some bit
 # without formally loading it.
 coreconfigitem('commands', 'rebase.requiredest',
             default=False,
 )
 coreconfigitem('experimental', 'rebaseskipobsolete',
     default=True,
 )
 coreconfigitem('rebase', 'singletransaction',
     default=False,
 )
 coreconfigitem('rebase', 'experimental.inmemory',
     default=False,
 )