dirstate: rename the filegenerator used for writing

We will need a different name in the next changesets. Changing the name is
actually not that trivial so we do it in its own changeset.

Differential Revision: https://phab.mercurial-scm.org/D12123

mercurial/dirstate.py

@@ -1,1430 +1,1430 b''
 # dirstate.py - working directory tracking for mercurial
 #
 # Copyright 2005-2007 Olivia Mackall <[e-mail]>
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 from __future__ import absolute_import
 
 import collections
 import contextlib
 import errno
 import os
 import stat
 
 from .i18n import _
 from .pycompat import delattr
 
 from hgdemandimport import tracing
 
 from . import (
     dirstatemap,
     encoding,
     error,
     match as matchmod,
     pathutil,
     policy,
     pycompat,
     scmutil,
     sparse,
     util,
 )
 
 from .dirstateutils import (
     timestamp,
 )
 
 from .interfaces import (
     dirstate as intdirstate,
     util as interfaceutil,
 )
 
 parsers = policy.importmod('parsers')
 rustmod = policy.importrust('dirstate')
 
 HAS_FAST_DIRSTATE_V2 = rustmod is not None
 
 propertycache = util.propertycache
 filecache = scmutil.filecache
 _rangemask = dirstatemap.rangemask
 
 DirstateItem = dirstatemap.DirstateItem
 
 
 class repocache(filecache):
     """filecache for files in .hg/"""
 
     def join(self, obj, fname):
         return obj._opener.join(fname)
 
 
 class rootcache(filecache):
     """filecache for files in the repository root"""
 
     def join(self, obj, fname):
         return obj._join(fname)
 
 
 def requires_parents_change(func):
     def wrap(self, *args, **kwargs):
         if not self.pendingparentchange():
             msg = 'calling `%s` outside of a parentchange context'
             msg %= func.__name__
             raise error.ProgrammingError(msg)
         return func(self, *args, **kwargs)
 
     return wrap
 
 
 def requires_no_parents_change(func):
     def wrap(self, *args, **kwargs):
         if self.pendingparentchange():
             msg = 'calling `%s` inside of a parentchange context'
             msg %= func.__name__
             raise error.ProgrammingError(msg)
         return func(self, *args, **kwargs)
 
     return wrap
 
 
 @interfaceutil.implementer(intdirstate.idirstate)
 class dirstate(object):
     def __init__(
         self,
         opener,
         ui,
         root,
         validate,
         sparsematchfn,
         nodeconstants,
         use_dirstate_v2,
     ):
         """Create a new dirstate object.
 
         opener is an open()-like callable that can be used to open the
         dirstate file; root is the root of the directory tracked by
         the dirstate.
         """
         self._use_dirstate_v2 = use_dirstate_v2
         self._nodeconstants = nodeconstants
         self._opener = opener
         self._validate = validate
         self._root = root
         self._sparsematchfn = sparsematchfn
         # ntpath.join(root, '') of Python 2.7.9 does not add sep if root is
         # UNC path pointing to root share (issue4557)
         self._rootdir = pathutil.normasprefix(root)
         self._dirty = False
         self._ui = ui
         self._filecache = {}
         self._parentwriters = 0
         self._filename = b'dirstate'
         self._pendingfilename = b'%s.pending' % self._filename
         self._plchangecallbacks = {}
         self._origpl = None
         self._mapcls = dirstatemap.dirstatemap
         # Access and cache cwd early, so we don't access it for the first time
         # after a working-copy update caused it to not exist (accessing it then
         # raises an exception).
         self._cwd
 
     def prefetch_parents(self):
         """make sure the parents are loaded
 
         Used to avoid a race condition.
         """
         self._pl
 
     @contextlib.contextmanager
     def parentchange(self):
         """Context manager for handling dirstate parents.
 
         If an exception occurs in the scope of the context manager,
         the incoherent dirstate won't be written when wlock is
         released.
         """
         self._parentwriters += 1
         yield
         # Typically we want the "undo" step of a context manager in a
         # finally block so it happens even when an exception
         # occurs. In this case, however, we only want to decrement
         # parentwriters if the code in the with statement exits
         # normally, so we don't have a try/finally here on purpose.
         self._parentwriters -= 1
 
     def pendingparentchange(self):
         """Returns true if the dirstate is in the middle of a set of changes
         that modify the dirstate parent.
         """
         return self._parentwriters > 0
 
     @propertycache
     def _map(self):
         """Return the dirstate contents (see documentation for dirstatemap)."""
         self._map = self._mapcls(
             self._ui,
             self._opener,
             self._root,
             self._nodeconstants,
             self._use_dirstate_v2,
         )
         return self._map
 
     @property
     def _sparsematcher(self):
         """The matcher for the sparse checkout.
 
         The working directory may not include every file from a manifest. The
         matcher obtained by this property will match a path if it is to be
         included in the working directory.
         """
         # TODO there is potential to cache this property. For now, the matcher
         # is resolved on every access. (But the called function does use a
         # cache to keep the lookup fast.)
         return self._sparsematchfn()
 
     @repocache(b'branch')
     def _branch(self):
         try:
             return self._opener.read(b"branch").strip() or b"default"
         except IOError as inst:
             if inst.errno != errno.ENOENT:
                 raise
             return b"default"
 
     @property
     def _pl(self):
         return self._map.parents()
 
     def hasdir(self, d):
         return self._map.hastrackeddir(d)
 
     @rootcache(b'.hgignore')
     def _ignore(self):
         files = self._ignorefiles()
         if not files:
             return matchmod.never()
 
         pats = [b'include:%s' % f for f in files]
         return matchmod.match(self._root, b'', [], pats, warn=self._ui.warn)
 
     @propertycache
     def _slash(self):
         return self._ui.configbool(b'ui', b'slash') and pycompat.ossep != b'/'
 
     @propertycache
     def _checklink(self):
         return util.checklink(self._root)
 
     @propertycache
     def _checkexec(self):
         return bool(util.checkexec(self._root))
 
     @propertycache
     def _checkcase(self):
         return not util.fscasesensitive(self._join(b'.hg'))
 
     def _join(self, f):
         # much faster than os.path.join()
         # it's safe because f is always a relative path
         return self._rootdir + f
 
     def flagfunc(self, buildfallback):
         """build a callable that returns flags associated with a filename
 
         The information is extracted from three possible layers:
         1. the file system if it supports the information
         2. the "fallback" information stored in the dirstate if any
         3. a more expensive mechanism inferring the flags from the parents.
         """
 
         # small hack to cache the result of buildfallback()
         fallback_func = []
 
         def get_flags(x):
             entry = None
             fallback_value = None
             try:
                 st = os.lstat(self._join(x))
             except OSError:
                 return b''
 
             if self._checklink:
                 if util.statislink(st):
                     return b'l'
             else:
                 entry = self.get_entry(x)
                 if entry.has_fallback_symlink:
                     if entry.fallback_symlink:
                         return b'l'
                 else:
                     if not fallback_func:
                         fallback_func.append(buildfallback())
                     fallback_value = fallback_func[0](x)
                     if b'l' in fallback_value:
                         return b'l'
 
             if self._checkexec:
                 if util.statisexec(st):
                     return b'x'
             else:
                 if entry is None:
                     entry = self.get_entry(x)
                 if entry.has_fallback_exec:
                     if entry.fallback_exec:
                         return b'x'
                 else:
                     if fallback_value is None:
                         if not fallback_func:
                             fallback_func.append(buildfallback())
                         fallback_value = fallback_func[0](x)
                     if b'x' in fallback_value:
                         return b'x'
             return b''
 
         return get_flags
 
     @propertycache
     def _cwd(self):
         # internal config: ui.forcecwd
         forcecwd = self._ui.config(b'ui', b'forcecwd')
         if forcecwd:
             return forcecwd
         return encoding.getcwd()
 
     def getcwd(self):
         """Return the path from which a canonical path is calculated.
 
         This path should be used to resolve file patterns or to convert
         canonical paths back to file paths for display. It shouldn't be
         used to get real file paths. Use vfs functions instead.
         """
         cwd = self._cwd
         if cwd == self._root:
             return b''
         # self._root ends with a path separator if self._root is '/' or 'C:\'
         rootsep = self._root
         if not util.endswithsep(rootsep):
             rootsep += pycompat.ossep
         if cwd.startswith(rootsep):
             return cwd[len(rootsep) :]
         else:
             # we're outside the repo. return an absolute path.
             return cwd
 
     def pathto(self, f, cwd=None):
         if cwd is None:
             cwd = self.getcwd()
         path = util.pathto(self._root, cwd, f)
         if self._slash:
             return util.pconvert(path)
         return path
 
     def get_entry(self, path):
         """return a DirstateItem for the associated path"""
         entry = self._map.get(path)
         if entry is None:
             return DirstateItem()
         return entry
 
     def __contains__(self, key):
         return key in self._map
 
     def __iter__(self):
         return iter(sorted(self._map))
 
     def items(self):
         return pycompat.iteritems(self._map)
 
     iteritems = items
 
     def parents(self):
         return [self._validate(p) for p in self._pl]
 
     def p1(self):
         return self._validate(self._pl[0])
 
     def p2(self):
         return self._validate(self._pl[1])
 
     @property
     def in_merge(self):
         """True if a merge is in progress"""
         return self._pl[1] != self._nodeconstants.nullid
 
     def branch(self):
         return encoding.tolocal(self._branch)
 
     def setparents(self, p1, p2=None):
         """Set dirstate parents to p1 and p2.
 
         When moving from two parents to one, "merged" entries a
         adjusted to normal and previous copy records discarded and
         returned by the call.
 
         See localrepo.setparents()
         """
         if p2 is None:
             p2 = self._nodeconstants.nullid
         if self._parentwriters == 0:
             raise ValueError(
                 b"cannot set dirstate parent outside of "
                 b"dirstate.parentchange context manager"
             )
 
         self._dirty = True
         oldp2 = self._pl[1]
         if self._origpl is None:
             self._origpl = self._pl
         nullid = self._nodeconstants.nullid
         # True if we need to fold p2 related state back to a linear case
         fold_p2 = oldp2 != nullid and p2 == nullid
         return self._map.setparents(p1, p2, fold_p2=fold_p2)
 
     def setbranch(self, branch):
         self.__class__._branch.set(self, encoding.fromlocal(branch))
         f = self._opener(b'branch', b'w', atomictemp=True, checkambig=True)
         try:
             f.write(self._branch + b'\n')
             f.close()
 
             # make sure filecache has the correct stat info for _branch after
             # replacing the underlying file
             ce = self._filecache[b'_branch']
             if ce:
                 ce.refresh()
         except:  # re-raises
             f.discard()
             raise
 
     def invalidate(self):
         """Causes the next access to reread the dirstate.
 
         This is different from localrepo.invalidatedirstate() because it always
         rereads the dirstate. Use localrepo.invalidatedirstate() if you want to
         check whether the dirstate has changed before rereading it."""
 
         for a in ("_map", "_branch", "_ignore"):
             if a in self.__dict__:
                 delattr(self, a)
         self._dirty = False
         self._parentwriters = 0
         self._origpl = None
 
     def copy(self, source, dest):
         """Mark dest as a copy of source. Unmark dest if source is None."""
         if source == dest:
             return
         self._dirty = True
         if source is not None:
             self._map.copymap[dest] = source
         else:
             self._map.copymap.pop(dest, None)
 
     def copied(self, file):
         return self._map.copymap.get(file, None)
 
     def copies(self):
         return self._map.copymap
 
     @requires_no_parents_change
     def set_tracked(self, filename, reset_copy=False):
         """a "public" method for generic code to mark a file as tracked
 
         This function is to be called outside of "update/merge" case. For
         example by a command like `hg add X`.
 
         if reset_copy is set, any existing copy information will be dropped.
 
         return True the file was previously untracked, False otherwise.
         """
         self._dirty = True
         entry = self._map.get(filename)
         if entry is None or not entry.tracked:
             self._check_new_tracked_filename(filename)
         pre_tracked = self._map.set_tracked(filename)
         if reset_copy:
             self._map.copymap.pop(filename, None)
         return pre_tracked
 
     @requires_no_parents_change
     def set_untracked(self, filename):
         """a "public" method for generic code to mark a file as untracked
 
         This function is to be called outside of "update/merge" case. For
         example by a command like `hg remove X`.
 
         return True the file was previously tracked, False otherwise.
         """
         ret = self._map.set_untracked(filename)
         if ret:
             self._dirty = True
         return ret
 
     @requires_no_parents_change
     def set_clean(self, filename, parentfiledata):
         """record that the current state of the file on disk is known to be clean"""
         self._dirty = True
         if not self._map[filename].tracked:
             self._check_new_tracked_filename(filename)
         (mode, size, mtime) = parentfiledata
         self._map.set_clean(filename, mode, size, mtime)
 
     @requires_no_parents_change
     def set_possibly_dirty(self, filename):
         """record that the current state of the file on disk is unknown"""
         self._dirty = True
         self._map.set_possibly_dirty(filename)
 
     @requires_parents_change
     def update_file_p1(
         self,
         filename,
         p1_tracked,
     ):
         """Set a file as tracked in the parent (or not)
 
         This is to be called when adjust the dirstate to a new parent after an history
         rewriting operation.
 
         It should not be called during a merge (p2 != nullid) and only within
         a `with dirstate.parentchange():` context.
         """
         if self.in_merge:
             msg = b'update_file_reference should not be called when merging'
             raise error.ProgrammingError(msg)
         entry = self._map.get(filename)
         if entry is None:
             wc_tracked = False
         else:
             wc_tracked = entry.tracked
         if not (p1_tracked or wc_tracked):
             # the file is no longer relevant to anyone
             if self._map.get(filename) is not None:
                 self._map.reset_state(filename)
                 self._dirty = True
         elif (not p1_tracked) and wc_tracked:
             if entry is not None and entry.added:
                 return  # avoid dropping copy information (maybe?)
 
         self._map.reset_state(
             filename,
             wc_tracked,
             p1_tracked,
             # the underlying reference might have changed, we will have to
             # check it.
             has_meaningful_mtime=False,
         )
 
     @requires_parents_change
     def update_file(
         self,
         filename,
         wc_tracked,
         p1_tracked,
         p2_info=False,
         possibly_dirty=False,
         parentfiledata=None,
     ):
         """update the information about a file in the dirstate
 
         This is to be called when the direstates parent changes to keep track
         of what is the file situation in regards to the working copy and its parent.
 
         This function must be called within a `dirstate.parentchange` context.
 
         note: the API is at an early stage and we might need to adjust it
         depending of what information ends up being relevant and useful to
         other processing.
         """
 
         # note: I do not think we need to double check name clash here since we
         # are in a update/merge case that should already have taken care of
         # this. The test agrees
 
         self._dirty = True
 
         self._map.reset_state(
             filename,
             wc_tracked,
             p1_tracked,
             p2_info=p2_info,
             has_meaningful_mtime=not possibly_dirty,
             parentfiledata=parentfiledata,
         )
 
     def _check_new_tracked_filename(self, filename):
         scmutil.checkfilename(filename)
         if self._map.hastrackeddir(filename):
             msg = _(b'directory %r already in dirstate')
             msg %= pycompat.bytestr(filename)
             raise error.Abort(msg)
         # shadows
         for d in pathutil.finddirs(filename):
             if self._map.hastrackeddir(d):
                 break
             entry = self._map.get(d)
             if entry is not None and not entry.removed:
                 msg = _(b'file %r in dirstate clashes with %r')
                 msg %= (pycompat.bytestr(d), pycompat.bytestr(filename))
                 raise error.Abort(msg)
 
     def _discoverpath(self, path, normed, ignoremissing, exists, storemap):
         if exists is None:
             exists = os.path.lexists(os.path.join(self._root, path))
         if not exists:
             # Maybe a path component exists
             if not ignoremissing and b'/' in path:
                 d, f = path.rsplit(b'/', 1)
                 d = self._normalize(d, False, ignoremissing, None)
                 folded = d + b"/" + f
             else:
                 # No path components, preserve original case
                 folded = path
         else:
             # recursively normalize leading directory components
             # against dirstate
             if b'/' in normed:
                 d, f = normed.rsplit(b'/', 1)
                 d = self._normalize(d, False, ignoremissing, True)
                 r = self._root + b"/" + d
                 folded = d + b"/" + util.fspath(f, r)
             else:
                 folded = util.fspath(normed, self._root)
             storemap[normed] = folded
 
         return folded
 
     def _normalizefile(self, path, isknown, ignoremissing=False, exists=None):
         normed = util.normcase(path)
         folded = self._map.filefoldmap.get(normed, None)
         if folded is None:
             if isknown:
                 folded = path
             else:
                 folded = self._discoverpath(
                     path, normed, ignoremissing, exists, self._map.filefoldmap
                 )
         return folded
 
     def _normalize(self, path, isknown, ignoremissing=False, exists=None):
         normed = util.normcase(path)
         folded = self._map.filefoldmap.get(normed, None)
         if folded is None:
             folded = self._map.dirfoldmap.get(normed, None)
         if folded is None:
             if isknown:
                 folded = path
             else:
                 # store discovered result in dirfoldmap so that future
                 # normalizefile calls don't start matching directories
                 folded = self._discoverpath(
                     path, normed, ignoremissing, exists, self._map.dirfoldmap
                 )
         return folded
 
     def normalize(self, path, isknown=False, ignoremissing=False):
         """
         normalize the case of a pathname when on a casefolding filesystem
 
         isknown specifies whether the filename came from walking the
         disk, to avoid extra filesystem access.
 
         If ignoremissing is True, missing path are returned
         unchanged. Otherwise, we try harder to normalize possibly
         existing path components.
 
         The normalized case is determined based on the following precedence:
 
         - version of name already stored in the dirstate
         - version of name stored on disk
         - version provided via command arguments
         """
 
         if self._checkcase:
             return self._normalize(path, isknown, ignoremissing)
         return path
 
     def clear(self):
         self._map.clear()
         self._dirty = True
 
     def rebuild(self, parent, allfiles, changedfiles=None):
         if changedfiles is None:
             # Rebuild entire dirstate
             to_lookup = allfiles
             to_drop = []
             self.clear()
         elif len(changedfiles) < 10:
             # Avoid turning allfiles into a set, which can be expensive if it's
             # large.
             to_lookup = []
             to_drop = []
             for f in changedfiles:
                 if f in allfiles:
                     to_lookup.append(f)
                 else:
                     to_drop.append(f)
         else:
             changedfilesset = set(changedfiles)
             to_lookup = changedfilesset & set(allfiles)
             to_drop = changedfilesset - to_lookup
 
         if self._origpl is None:
             self._origpl = self._pl
         self._map.setparents(parent, self._nodeconstants.nullid)
 
         for f in to_lookup:
 
             if self.in_merge:
                 self.set_tracked(f)
             else:
                 self._map.reset_state(
                     f,
                     wc_tracked=True,
                     p1_tracked=True,
                 )
         for f in to_drop:
             self._map.reset_state(f)
 
         self._dirty = True
 
     def identity(self):
         """Return identity of dirstate itself to detect changing in storage
 
         If identity of previous dirstate is equal to this, writing
         changes based on the former dirstate out can keep consistency.
         """
         return self._map.identity
 
     def write(self, tr):
         if not self._dirty:
             return
 
         filename = self._filename
         if tr:
             # delay writing in-memory changes out
             tr.addfilegenerator(
-                b'dirstate',
+                b'dirstate-1-main',
                 (self._filename,),
                 lambda f: self._writedirstate(tr, f),
                 location=b'plain',
             )
             return
 
         file = lambda f: self._opener(f, b"w", atomictemp=True, checkambig=True)
         with file(self._filename) as f:
             self._writedirstate(tr, f)
 
     def addparentchangecallback(self, category, callback):
         """add a callback to be called when the wd parents are changed
 
         Callback will be called with the following arguments:
             dirstate, (oldp1, oldp2), (newp1, newp2)
 
         Category is a unique identifier to allow overwriting an old callback
         with a newer callback.
         """
         self._plchangecallbacks[category] = callback
 
     def _writedirstate(self, tr, st):
         # notify callbacks about parents change
         if self._origpl is not None and self._origpl != self._pl:
             for c, callback in sorted(
                 pycompat.iteritems(self._plchangecallbacks)
             ):
                 callback(self, self._origpl, self._pl)
             self._origpl = None
 
         self._map.write(tr, st)
         self._dirty = False
 
     def _dirignore(self, f):
         if self._ignore(f):
             return True
         for p in pathutil.finddirs(f):
             if self._ignore(p):
                 return True
         return False
 
     def _ignorefiles(self):
         files = []
         if os.path.exists(self._join(b'.hgignore')):
             files.append(self._join(b'.hgignore'))
         for name, path in self._ui.configitems(b"ui"):
             if name == b'ignore' or name.startswith(b'ignore.'):
                 # we need to use os.path.join here rather than self._join
                 # because path is arbitrary and user-specified
                 files.append(os.path.join(self._rootdir, util.expandpath(path)))
         return files
 
     def _ignorefileandline(self, f):
         files = collections.deque(self._ignorefiles())
         visited = set()
         while files:
             i = files.popleft()
             patterns = matchmod.readpatternfile(
                 i, self._ui.warn, sourceinfo=True
             )
             for pattern, lineno, line in patterns:
                 kind, p = matchmod._patsplit(pattern, b'glob')
                 if kind == b"subinclude":
                     if p not in visited:
                         files.append(p)
                     continue
                 m = matchmod.match(
                     self._root, b'', [], [pattern], warn=self._ui.warn
                 )
                 if m(f):
                     return (i, lineno, line)
             visited.add(i)
         return (None, -1, b"")
 
     def _walkexplicit(self, match, subrepos):
         """Get stat data about the files explicitly specified by match.
 
         Return a triple (results, dirsfound, dirsnotfound).
         - results is a mapping from filename to stat result. It also contains
           listings mapping subrepos and .hg to None.
         - dirsfound is a list of files found to be directories.
         - dirsnotfound is a list of files that the dirstate thinks are
           directories and that were not found."""
 
         def badtype(mode):
             kind = _(b'unknown')
             if stat.S_ISCHR(mode):
                 kind = _(b'character device')
             elif stat.S_ISBLK(mode):
                 kind = _(b'block device')
             elif stat.S_ISFIFO(mode):
                 kind = _(b'fifo')
             elif stat.S_ISSOCK(mode):
                 kind = _(b'socket')
             elif stat.S_ISDIR(mode):
                 kind = _(b'directory')
             return _(b'unsupported file type (type is %s)') % kind
 
         badfn = match.bad
         dmap = self._map
         lstat = os.lstat
         getkind = stat.S_IFMT
         dirkind = stat.S_IFDIR
         regkind = stat.S_IFREG
         lnkkind = stat.S_IFLNK
         join = self._join
         dirsfound = []
         foundadd = dirsfound.append
         dirsnotfound = []
         notfoundadd = dirsnotfound.append
 
         if not match.isexact() and self._checkcase:
             normalize = self._normalize
         else:
             normalize = None
 
         files = sorted(match.files())
         subrepos.sort()
         i, j = 0, 0
         while i < len(files) and j < len(subrepos):
             subpath = subrepos[j] + b"/"
             if files[i] < subpath:
                 i += 1
                 continue
             while i < len(files) and files[i].startswith(subpath):
                 del files[i]
             j += 1
 
         if not files or b'' in files:
             files = [b'']
             # constructing the foldmap is expensive, so don't do it for the
             # common case where files is ['']
             normalize = None
         results = dict.fromkeys(subrepos)
         results[b'.hg'] = None
 
         for ff in files:
             if normalize:
                 nf = normalize(ff, False, True)
             else:
                 nf = ff
             if nf in results:
                 continue
 
             try:
                 st = lstat(join(nf))
                 kind = getkind(st.st_mode)
                 if kind == dirkind:
                     if nf in dmap:
                         # file replaced by dir on disk but still in dirstate
                         results[nf] = None
                     foundadd((nf, ff))
                 elif kind == regkind or kind == lnkkind:
                     results[nf] = st
                 else:
                     badfn(ff, badtype(kind))
                     if nf in dmap:
                         results[nf] = None
             except OSError as inst:  # nf not found on disk - it is dirstate only
                 if nf in dmap:  # does it exactly match a missing file?
                     results[nf] = None
                 else:  # does it match a missing directory?
                     if self._map.hasdir(nf):
                         notfoundadd(nf)
                     else:
                         badfn(ff, encoding.strtolocal(inst.strerror))
 
         # match.files() may contain explicitly-specified paths that shouldn't
         # be taken; drop them from the list of files found. dirsfound/notfound
         # aren't filtered here because they will be tested later.
         if match.anypats():
             for f in list(results):
                 if f == b'.hg' or f in subrepos:
                     # keep sentinel to disable further out-of-repo walks
                     continue
                 if not match(f):
                     del results[f]
 
         # Case insensitive filesystems cannot rely on lstat() failing to detect
         # a case-only rename.  Prune the stat object for any file that does not
         # match the case in the filesystem, if there are multiple files that
         # normalize to the same path.
         if match.isexact() and self._checkcase:
             normed = {}
 
             for f, st in pycompat.iteritems(results):
                 if st is None:
                     continue
 
                 nc = util.normcase(f)
                 paths = normed.get(nc)
 
                 if paths is None:
                     paths = set()
                     normed[nc] = paths
 
                 paths.add(f)
 
             for norm, paths in pycompat.iteritems(normed):
                 if len(paths) > 1:
                     for path in paths:
                         folded = self._discoverpath(
                             path, norm, True, None, self._map.dirfoldmap
                         )
                         if path != folded:
                             results[path] = None
 
         return results, dirsfound, dirsnotfound
 
     def walk(self, match, subrepos, unknown, ignored, full=True):
         """
         Walk recursively through the directory tree, finding all files
         matched by match.
 
         If full is False, maybe skip some known-clean files.
 
         Return a dict mapping filename to stat-like object (either
         mercurial.osutil.stat instance or return value of os.stat()).
 
         """
         # full is a flag that extensions that hook into walk can use -- this
         # implementation doesn't use it at all. This satisfies the contract
         # because we only guarantee a "maybe".
 
         if ignored:
             ignore = util.never
             dirignore = util.never
         elif unknown:
             ignore = self._ignore
             dirignore = self._dirignore
         else:
             # if not unknown and not ignored, drop dir recursion and step 2
             ignore = util.always
             dirignore = util.always
 
         matchfn = match.matchfn
         matchalways = match.always()
         matchtdir = match.traversedir
         dmap = self._map
         listdir = util.listdir
         lstat = os.lstat
         dirkind = stat.S_IFDIR
         regkind = stat.S_IFREG
         lnkkind = stat.S_IFLNK
         join = self._join
 
         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:
             normalize = self._normalize
             normalizefile = self._normalizefile
             skipstep3 = False
         else:
             normalize = self._normalize
             normalizefile = None
 
         # step 1: find all explicit files
         results, work, dirsnotfound = self._walkexplicit(match, subrepos)
         if matchtdir:
             for d in work:
                 matchtdir(d[0])
             for d in dirsnotfound:
                 matchtdir(d)
 
         skipstep3 = skipstep3 and not (work or dirsnotfound)
         work = [d for d in work if not dirignore(d[0])]
 
         # step 2: visit subdirectories
         def traverse(work, alreadynormed):
             wadd = work.append
             while work:
                 tracing.counter('dirstate.walk work', len(work))
                 nd = work.pop()
                 visitentries = match.visitchildrenset(nd)
                 if not visitentries:
                     continue
                 if visitentries == b'this' or visitentries == b'all':
                     visitentries = None
                 skip = None
                 if nd != b'':
                     skip = b'.hg'
                 try:
                     with tracing.log('dirstate.walk.traverse listdir %s', nd):
                         entries = listdir(join(nd), stat=True, skip=skip)
                 except OSError as inst:
                     if inst.errno in (errno.EACCES, errno.ENOENT):
                         match.bad(
                             self.pathto(nd), encoding.strtolocal(inst.strerror)
                         )
                         continue
                     raise
                 for f, kind, st in entries:
                     # Some matchers may return files in the visitentries set,
                     # instead of 'this', if the matcher explicitly mentions them
                     # and is not an exactmatcher. This is acceptable; we do not
                     # make any hard assumptions about file-or-directory below
                     # based on the presence of `f` in visitentries. If
                     # visitchildrenset returned a set, we can always skip the
                     # entries *not* in the set it provided regardless of whether
                     # they're actually a file or a directory.
                     if visitentries and f not in visitentries:
                         continue
                     if normalizefile:
                         # even though f might be a directory, we're only
                         # interested in comparing it to files currently in the
                         # dmap -- therefore normalizefile is enough
                         nf = normalizefile(
                             nd and (nd + b"/" + f) or f, True, True
                         )
                     else:
                         nf = nd and (nd + b"/" + f) or f
                     if nf not in results:
                         if kind == dirkind:
                             if not ignore(nf):
                                 if matchtdir:
                                     matchtdir(nf)
                                 wadd(nf)
                             if nf in dmap and (matchalways or matchfn(nf)):
                                 results[nf] = None
                         elif kind == regkind or kind == lnkkind:
                             if nf in dmap:
                                 if matchalways or matchfn(nf):
                                     results[nf] = st
                             elif (matchalways or matchfn(nf)) and not ignore(
                                 nf
                             ):
                                 # unknown file -- normalize if necessary
                                 if not alreadynormed:
                                     nf = normalize(nf, False, True)
                                 results[nf] = st
                         elif nf in dmap and (matchalways or matchfn(nf)):
                             results[nf] = None
 
         for nd, d in work:
             # alreadynormed means that processwork doesn't have to do any
             # expensive directory normalization
             alreadynormed = not normalize or nd == d
             traverse([d], alreadynormed)
 
         for s in subrepos:
             del results[s]
         del results[b'.hg']
 
         # step 3: visit remaining files from dmap
         if not skipstep3 and not exact:
             # If a dmap file is not in results yet, it was either
             # a) not matching matchfn b) ignored, c) missing, or d) under a
             # symlink directory.
             if not results and matchalways:
                 visit = [f for f in dmap]
             else:
                 visit = [f for f in dmap if f not in results and matchfn(f)]
             visit.sort()
 
             if unknown:
                 # unknown == True means we walked all dirs under the roots
                 # that wasn't ignored, and everything that matched was stat'ed
                 # and is already in results.
                 # The rest must thus be ignored or under a symlink.
                 audit_path = pathutil.pathauditor(self._root, cached=True)
 
                 for nf in iter(visit):
                     # If a stat for the same file was already added with a
                     # different case, don't add one for this, since that would
                     # make it appear as if the file exists under both names
                     # on disk.
                     if (
                         normalizefile
                         and normalizefile(nf, True, True) in results
                     ):
                         results[nf] = None
                     # Report ignored items in the dmap as long as they are not
                     # under a symlink directory.
                     elif audit_path.check(nf):
                         try:
                             results[nf] = lstat(join(nf))
                             # file was just ignored, no links, and exists
                         except OSError:
                             # file doesn't exist
                             results[nf] = None
                     else:
                         # It's either missing or under a symlink directory
                         # which we in this case report as missing
                         results[nf] = None
             else:
                 # We may not have walked the full directory tree above,
                 # so stat and check everything we missed.
                 iv = iter(visit)
                 for st in util.statfiles([join(i) for i in visit]):
                     results[next(iv)] = st
         return results
 
     def _rust_status(self, matcher, list_clean, list_ignored, list_unknown):
         # Force Rayon (Rust parallelism library) to respect the number of
         # workers. This is a temporary workaround until Rust code knows
         # how to read the config file.
         numcpus = self._ui.configint(b"worker", b"numcpus")
         if numcpus is not None:
             encoding.environ.setdefault(b'RAYON_NUM_THREADS', b'%d' % numcpus)
 
         workers_enabled = self._ui.configbool(b"worker", b"enabled", True)
         if not workers_enabled:
             encoding.environ[b"RAYON_NUM_THREADS"] = b"1"
 
         (
             lookup,
             modified,
             added,
             removed,
             deleted,
             clean,
             ignored,
             unknown,
             warnings,
             bad,
             traversed,
             dirty,
         ) = rustmod.status(
             self._map._map,
             matcher,
             self._rootdir,
             self._ignorefiles(),
             self._checkexec,
             bool(list_clean),
             bool(list_ignored),
             bool(list_unknown),
             bool(matcher.traversedir),
         )
 
         self._dirty |= dirty
 
         if matcher.traversedir:
             for dir in traversed:
                 matcher.traversedir(dir)
 
         if self._ui.warn:
             for item in warnings:
                 if isinstance(item, tuple):
                     file_path, syntax = item
                     msg = _(b"%s: ignoring invalid syntax '%s'\n") % (
                         file_path,
                         syntax,
                     )
                     self._ui.warn(msg)
                 else:
                     msg = _(b"skipping unreadable pattern file '%s': %s\n")
                     self._ui.warn(
                         msg
                         % (
                             pathutil.canonpath(
                                 self._rootdir, self._rootdir, item
                             ),
                             b"No such file or directory",
                         )
                     )
 
         for (fn, message) in bad:
             matcher.bad(fn, encoding.strtolocal(message))
 
         status = scmutil.status(
             modified=modified,
             added=added,
             removed=removed,
             deleted=deleted,
             unknown=unknown,
             ignored=ignored,
             clean=clean,
         )
         return (lookup, status)
 
     def status(self, match, subrepos, ignored, clean, unknown):
         """Determine the status of the working copy relative to the
         dirstate and return a pair of (unsure, status), where status is of type
         scmutil.status and:
 
           unsure:
             files that might have been modified since the dirstate was
             written, but need to be read to be sure (size is the same
             but mtime differs)
           status.modified:
             files that have definitely been modified since the dirstate
             was written (different size or mode)
           status.clean:
             files that have definitely not been modified since the
             dirstate was written
         """
         listignored, listclean, listunknown = ignored, clean, unknown
         lookup, modified, added, unknown, ignored = [], [], [], [], []
         removed, deleted, clean = [], [], []
 
         dmap = self._map
         dmap.preload()
 
         use_rust = True
 
         allowed_matchers = (
             matchmod.alwaysmatcher,
             matchmod.exactmatcher,
             matchmod.includematcher,
         )
 
         if rustmod is None:
             use_rust = False
         elif self._checkcase:
             # Case-insensitive filesystems are not handled yet
             use_rust = False
         elif subrepos:
             use_rust = False
         elif sparse.enabled:
             use_rust = False
         elif not isinstance(match, allowed_matchers):
             # Some matchers have yet to be implemented
             use_rust = False
 
         # Get the time from the filesystem so we can disambiguate files that
         # appear modified in the present or future.
         try:
             mtime_boundary = timestamp.get_fs_now(self._opener)
         except OSError:
             # In largefiles or readonly context
             mtime_boundary = None
 
         if use_rust:
             try:
                 res = self._rust_status(
                     match, listclean, listignored, listunknown
                 )
                 return res + (mtime_boundary,)
             except rustmod.FallbackError:
                 pass
 
         def noop(f):
             pass
 
         dcontains = dmap.__contains__
         dget = dmap.__getitem__
         ladd = lookup.append  # aka "unsure"
         madd = modified.append
         aadd = added.append
         uadd = unknown.append if listunknown else noop
         iadd = ignored.append if listignored else noop
         radd = removed.append
         dadd = deleted.append
         cadd = clean.append if listclean else noop
         mexact = match.exact
         dirignore = self._dirignore
         checkexec = self._checkexec
         checklink = self._checklink
         copymap = self._map.copymap
 
         # We need to do full walks when either
         # - we're listing all clean files, or
         # - match.traversedir does something, because match.traversedir should
         #   be called for every dir in the working dir
         full = listclean or match.traversedir is not None
         for fn, st in pycompat.iteritems(
             self.walk(match, subrepos, listunknown, listignored, full=full)
         ):
             if not dcontains(fn):
                 if (listignored or mexact(fn)) and dirignore(fn):
                     if listignored:
                         iadd(fn)
                 else:
                     uadd(fn)
                 continue
 
             t = dget(fn)
             mode = t.mode
             size = t.size
 
             if not st and t.tracked:
                 dadd(fn)
             elif t.p2_info:
                 madd(fn)
             elif t.added:
                 aadd(fn)
             elif t.removed:
                 radd(fn)
             elif t.tracked:
                 if not checklink and t.has_fallback_symlink:
                     # If the file system does not support symlink, the mode
                     # might not be correctly stored in the dirstate, so do not
                     # trust it.
                     ladd(fn)
                 elif not checkexec and t.has_fallback_exec:
                     # If the file system does not support exec bits, the mode
                     # might not be correctly stored in the dirstate, so do not
                     # trust it.
                     ladd(fn)
                 elif (
                     size >= 0
                     and (
                         (size != st.st_size and size != st.st_size & _rangemask)
                         or ((mode ^ st.st_mode) & 0o100 and checkexec)
                     )
                     or fn in copymap
                 ):
                     if stat.S_ISLNK(st.st_mode) and size != st.st_size:
                         # issue6456: Size returned may be longer due to
                         # encryption on EXT-4 fscrypt, undecided.
                         ladd(fn)
                     else:
                         madd(fn)
                 elif not t.mtime_likely_equal_to(timestamp.mtime_of(st)):
                     # There might be a change in the future if for example the
                     # internal clock is off, but this is a case where the issues
                     # the user would face would be a lot worse and there is
                     # nothing we can really do.
                     ladd(fn)
                 elif listclean:
                     cadd(fn)
         status = scmutil.status(
             modified, added, removed, deleted, unknown, ignored, clean
         )
         return (lookup, status, mtime_boundary)
 
     def matches(self, match):
         """
         return files in the dirstate (in whatever state) filtered by match
         """
         dmap = self._map
         if rustmod is not None:
             dmap = self._map._map
 
         if match.always():
             return dmap.keys()
         files = match.files()
         if match.isexact():
             # fast path -- filter the other way around, since typically files is
             # much smaller than dmap
             return [f for f in files if f in dmap]
         if match.prefix() and all(fn in dmap for fn in files):
             # fast path -- all the values are known to be files, so just return
             # that
             return list(files)
         return [f for f in dmap if match(f)]
 
     def _actualfilename(self, tr):
         if tr:
             return self._pendingfilename
         else:
             return self._filename
 
     def savebackup(self, tr, backupname):
         '''Save current dirstate into backup file'''
         filename = self._actualfilename(tr)
         assert backupname != filename
 
         # use '_writedirstate' instead of 'write' to write changes certainly,
         # because the latter omits writing out if transaction is running.
         # output file will be used to create backup of dirstate at this point.
         if self._dirty or not self._opener.exists(filename):
             self._writedirstate(
                 tr,
                 self._opener(filename, b"w", atomictemp=True, checkambig=True),
             )
 
         if tr:
             # ensure that subsequent tr.writepending returns True for
             # changes written out above, even if dirstate is never
             # changed after this
             tr.addfilegenerator(
-                b'dirstate',
+                b'dirstate-1-main',
                 (self._filename,),
                 lambda f: self._writedirstate(tr, f),
                 location=b'plain',
             )
 
             # ensure that pending file written above is unlinked at
             # failure, even if tr.writepending isn't invoked until the
             # end of this transaction
             tr.registertmp(filename, location=b'plain')
 
         self._opener.tryunlink(backupname)
         # hardlink backup is okay because _writedirstate is always called
         # with an "atomictemp=True" file.
         util.copyfile(
             self._opener.join(filename),
             self._opener.join(backupname),
             hardlink=True,
         )
 
     def restorebackup(self, tr, backupname):
         '''Restore dirstate by backup file'''
         # this "invalidate()" prevents "wlock.release()" from writing
         # changes of dirstate out after restoring from backup file
         self.invalidate()
         filename = self._actualfilename(tr)
         o = self._opener
         if util.samefile(o.join(backupname), o.join(filename)):
             o.unlink(backupname)
         else:
             o.rename(backupname, filename, checkambig=True)
 
     def clearbackup(self, tr, backupname):
         '''Clear backup file'''
         self._opener.unlink(backupname)
 
     def verify(self, m1, m2):
         """check the dirstate content again the parent manifest and yield errors"""
         missing_from_p1 = b"%s in state %s, but not in manifest1\n"
         unexpected_in_p1 = b"%s in state %s, but also in manifest1\n"
         missing_from_ps = b"%s in state %s, but not in either manifest\n"
         missing_from_ds = b"%s in manifest1, but listed as state %s\n"
         for f, entry in self.items():
             state = entry.state
             if state in b"nr" and f not in m1:
                 yield (missing_from_p1, f, state)
             if state in b"a" and f in m1:
                 yield (unexpected_in_p1, f, state)
             if state in b"m" and f not in m1 and f not in m2:
                 yield (missing_from_ps, f, state)
         for f in m1:
             state = self.get_entry(f).state
             if state not in b"nrm":
                 yield (missing_from_ds, f, state)

mercurial/transaction.py

@@ -1,768 +1,771 b''
 # transaction.py - simple journaling scheme for mercurial
 #
 # This transaction scheme is intended to gracefully handle program
 # errors and interruptions. More serious failures like system crashes
 # can be recovered with an fsck-like tool. As the whole repository is
 # effectively log-structured, this should amount to simply truncating
 # anything that isn't referenced in the changelog.
 #
 # Copyright 2005, 2006 Olivia Mackall <[e-mail]>
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 from __future__ import absolute_import
 
 import errno
 
 from .i18n import _
 from . import (
     error,
     pycompat,
     util,
 )
 from .utils import stringutil
 
 version = 2
 
 # These are the file generators that should only be executed after the
 # finalizers are done, since they rely on the output of the finalizers (like
 # the changelog having been written).
-postfinalizegenerators = {b'bookmarks', b'dirstate'}
+postfinalizegenerators = {
+    b'bookmarks',
+    b'dirstate-1-main',
+}
 
 GEN_GROUP_ALL = b'all'
 GEN_GROUP_PRE_FINALIZE = b'prefinalize'
 GEN_GROUP_POST_FINALIZE = b'postfinalize'
 
 
 def active(func):
     def _active(self, *args, **kwds):
         if self._count == 0:
             raise error.ProgrammingError(
                 b'cannot use transaction when it is already committed/aborted'
             )
         return func(self, *args, **kwds)
 
     return _active
 
 
 def _playback(
     journal,
     report,
     opener,
     vfsmap,
     entries,
     backupentries,
     unlink=True,
     checkambigfiles=None,
 ):
     for f, o in sorted(dict(entries).items()):
         if o or not unlink:
             checkambig = checkambigfiles and (f, b'') in checkambigfiles
             try:
                 fp = opener(f, b'a', checkambig=checkambig)
                 if fp.tell() < o:
                     raise error.Abort(
                         _(
                             b"attempted to truncate %s to %d bytes, but it was "
                             b"already %d bytes\n"
                         )
                         % (f, o, fp.tell())
                     )
                 fp.truncate(o)
                 fp.close()
             except IOError:
                 report(_(b"failed to truncate %s\n") % f)
                 raise
         else:
             try:
                 opener.unlink(f)
             except (IOError, OSError) as inst:
                 if inst.errno != errno.ENOENT:
                     raise
 
     backupfiles = []
     for l, f, b, c in backupentries:
         if l not in vfsmap and c:
             report(b"couldn't handle %s: unknown cache location %s\n" % (b, l))
         vfs = vfsmap[l]
         try:
             if f and b:
                 filepath = vfs.join(f)
                 backuppath = vfs.join(b)
                 checkambig = checkambigfiles and (f, l) in checkambigfiles
                 try:
                     util.copyfile(backuppath, filepath, checkambig=checkambig)
                     backupfiles.append(b)
                 except IOError as exc:
                     e_msg = stringutil.forcebytestr(exc)
                     report(_(b"failed to recover %s (%s)\n") % (f, e_msg))
             else:
                 target = f or b
                 try:
                     vfs.unlink(target)
                 except (IOError, OSError) as inst:
                     if inst.errno != errno.ENOENT:
                         raise
         except (IOError, OSError, error.Abort):
             if not c:
                 raise
 
     backuppath = b"%s.backupfiles" % journal
     if opener.exists(backuppath):
         opener.unlink(backuppath)
     opener.unlink(journal)
     try:
         for f in backupfiles:
             if opener.exists(f):
                 opener.unlink(f)
     except (IOError, OSError, error.Abort):
         # only pure backup file remains, it is sage to ignore any error
         pass
 
 
 class transaction(util.transactional):
     def __init__(
         self,
         report,
         opener,
         vfsmap,
         journalname,
         undoname=None,
         after=None,
         createmode=None,
         validator=None,
         releasefn=None,
         checkambigfiles=None,
         name='<unnamed>',
     ):
         """Begin a new transaction
 
         Begins a new transaction that allows rolling back writes in the event of
         an exception.
 
         * `after`: called after the transaction has been committed
         * `createmode`: the mode of the journal file that will be created
         * `releasefn`: called after releasing (with transaction and result)
 
         `checkambigfiles` is a set of (path, vfs-location) tuples,
         which determine whether file stat ambiguity should be avoided
         for corresponded files.
         """
         self._count = 1
         self._usages = 1
         self._report = report
         # a vfs to the store content
         self._opener = opener
         # a map to access file in various {location -> vfs}
         vfsmap = vfsmap.copy()
         vfsmap[b''] = opener  # set default value
         self._vfsmap = vfsmap
         self._after = after
         self._offsetmap = {}
         self._newfiles = set()
         self._journal = journalname
         self._undoname = undoname
         self._queue = []
         # A callback to do something just after releasing transaction.
         if releasefn is None:
             releasefn = lambda tr, success: None
         self._releasefn = releasefn
 
         self._checkambigfiles = set()
         if checkambigfiles:
             self._checkambigfiles.update(checkambigfiles)
 
         self._names = [name]
 
         # A dict dedicated to precisely tracking the changes introduced in the
         # transaction.
         self.changes = {}
 
         # a dict of arguments to be passed to hooks
         self.hookargs = {}
         self._file = opener.open(self._journal, b"w+")
 
         # a list of ('location', 'path', 'backuppath', cache) entries.
         # - if 'backuppath' is empty, no file existed at backup time
         # - if 'path' is empty, this is a temporary transaction file
         # - if 'location' is not empty, the path is outside main opener reach.
         #   use 'location' value as a key in a vfsmap to find the right 'vfs'
         # (cache is currently unused)
         self._backupentries = []
         self._backupmap = {}
         self._backupjournal = b"%s.backupfiles" % self._journal
         self._backupsfile = opener.open(self._backupjournal, b'w')
         self._backupsfile.write(b'%d\n' % version)
 
         if createmode is not None:
             opener.chmod(self._journal, createmode & 0o666)
             opener.chmod(self._backupjournal, createmode & 0o666)
 
         # hold file generations to be performed on commit
         self._filegenerators = {}
         # hold callback to write pending data for hooks
         self._pendingcallback = {}
         # True is any pending data have been written ever
         self._anypending = False
         # holds callback to call when writing the transaction
         self._finalizecallback = {}
         # holds callback to call when validating the transaction
         # should raise exception if anything is wrong
         self._validatecallback = {}
         if validator is not None:
             self._validatecallback[b'001-userhooks'] = validator
         # hold callback for post transaction close
         self._postclosecallback = {}
         # holds callbacks to call during abort
         self._abortcallback = {}
 
     def __repr__(self):
         name = '/'.join(self._names)
         return '<transaction name=%s, count=%d, usages=%d>' % (
             name,
             self._count,
             self._usages,
         )
 
     def __del__(self):
         if self._journal:
             self._abort()
 
     @property
     def finalized(self):
         return self._finalizecallback is None
 
     @active
     def startgroup(self):
         """delay registration of file entry
 
         This is used by strip to delay vision of strip offset. The transaction
         sees either none or all of the strip actions to be done."""
         self._queue.append([])
 
     @active
     def endgroup(self):
         """apply delayed registration of file entry.
 
         This is used by strip to delay vision of strip offset. The transaction
         sees either none or all of the strip actions to be done."""
         q = self._queue.pop()
         for f, o in q:
             self._addentry(f, o)
 
     @active
     def add(self, file, offset):
         """record the state of an append-only file before update"""
         if (
             file in self._newfiles
             or file in self._offsetmap
             or file in self._backupmap
         ):
             return
         if self._queue:
             self._queue[-1].append((file, offset))
             return
 
         self._addentry(file, offset)
 
     def _addentry(self, file, offset):
         """add a append-only entry to memory and on-disk state"""
         if (
             file in self._newfiles
             or file in self._offsetmap
             or file in self._backupmap
         ):
             return
         if offset:
             self._offsetmap[file] = offset
         else:
             self._newfiles.add(file)
         # add enough data to the journal to do the truncate
         self._file.write(b"%s\0%d\n" % (file, offset))
         self._file.flush()
 
     @active
     def addbackup(self, file, hardlink=True, location=b''):
         """Adds a backup of the file to the transaction
 
         Calling addbackup() creates a hardlink backup of the specified file
         that is used to recover the file in the event of the transaction
         aborting.
 
         * `file`: the file path, relative to .hg/store
         * `hardlink`: use a hardlink to quickly create the backup
         """
         if self._queue:
             msg = b'cannot use transaction.addbackup inside "group"'
             raise error.ProgrammingError(msg)
 
         if (
             file in self._newfiles
             or file in self._offsetmap
             or file in self._backupmap
         ):
             return
         vfs = self._vfsmap[location]
         dirname, filename = vfs.split(file)
         backupfilename = b"%s.backup.%s" % (self._journal, filename)
         backupfile = vfs.reljoin(dirname, backupfilename)
         if vfs.exists(file):
             filepath = vfs.join(file)
             backuppath = vfs.join(backupfile)
             util.copyfile(filepath, backuppath, hardlink=hardlink)
         else:
             backupfile = b''
 
         self._addbackupentry((location, file, backupfile, False))
 
     def _addbackupentry(self, entry):
         """register a new backup entry and write it to disk"""
         self._backupentries.append(entry)
         self._backupmap[entry[1]] = len(self._backupentries) - 1
         self._backupsfile.write(b"%s\0%s\0%s\0%d\n" % entry)
         self._backupsfile.flush()
 
     @active
     def registertmp(self, tmpfile, location=b''):
         """register a temporary transaction file
 
         Such files will be deleted when the transaction exits (on both
         failure and success).
         """
         self._addbackupentry((location, b'', tmpfile, False))
 
     @active
     def addfilegenerator(
         self, genid, filenames, genfunc, order=0, location=b''
     ):
         """add a function to generates some files at transaction commit
 
         The `genfunc` argument is a function capable of generating proper
         content of each entry in the `filename` tuple.
 
         At transaction close time, `genfunc` will be called with one file
         object argument per entries in `filenames`.
 
         The transaction itself is responsible for the backup, creation and
         final write of such file.
 
         The `genid` argument is used to ensure the same set of file is only
         generated once. Call to `addfilegenerator` for a `genid` already
         present will overwrite the old entry.
 
         The `order` argument may be used to control the order in which multiple
         generator will be executed.
 
         The `location` arguments may be used to indicate the files are located
         outside of the the standard directory for transaction. It should match
         one of the key of the `transaction.vfsmap` dictionary.
         """
         # For now, we are unable to do proper backup and restore of custom vfs
         # but for bookmarks that are handled outside this mechanism.
         self._filegenerators[genid] = (order, filenames, genfunc, location)
 
     @active
     def removefilegenerator(self, genid):
         """reverse of addfilegenerator, remove a file generator function"""
         if genid in self._filegenerators:
             del self._filegenerators[genid]
 
     def _generatefiles(self, suffix=b'', group=GEN_GROUP_ALL):
         # write files registered for generation
         any = False
 
         if group == GEN_GROUP_ALL:
             skip_post = skip_pre = False
         else:
             skip_pre = group == GEN_GROUP_POST_FINALIZE
             skip_post = group == GEN_GROUP_PRE_FINALIZE
 
         for id, entry in sorted(pycompat.iteritems(self._filegenerators)):
             any = True
             order, filenames, genfunc, location = entry
 
             # for generation at closing, check if it's before or after finalize
             is_post = id in postfinalizegenerators
             if skip_post and is_post:
                 continue
             elif skip_pre and not is_post:
                 continue
 
             vfs = self._vfsmap[location]
             files = []
             try:
                 for name in filenames:
                     name += suffix
                     if suffix:
                         self.registertmp(name, location=location)
                         checkambig = False
                     else:
                         self.addbackup(name, location=location)
                         checkambig = (name, location) in self._checkambigfiles
                     files.append(
                         vfs(name, b'w', atomictemp=True, checkambig=checkambig)
                     )
                 genfunc(*files)
                 for f in files:
                     f.close()
                 # skip discard() loop since we're sure no open file remains
                 del files[:]
             finally:
                 for f in files:
                     f.discard()
         return any
 
     @active
     def findoffset(self, file):
         if file in self._newfiles:
             return 0
         return self._offsetmap.get(file)
 
     @active
     def readjournal(self):
         self._file.seek(0)
         entries = []
         for l in self._file.readlines():
             file, troffset = l.split(b'\0')
             entries.append((file, int(troffset)))
         return entries
 
     @active
     def replace(self, file, offset):
         """
         replace can only replace already committed entries
         that are not pending in the queue
         """
         if file in self._newfiles:
             if not offset:
                 return
             self._newfiles.remove(file)
             self._offsetmap[file] = offset
         elif file in self._offsetmap:
             if not offset:
                 del self._offsetmap[file]
                 self._newfiles.add(file)
             else:
                 self._offsetmap[file] = offset
         else:
             raise KeyError(file)
         self._file.write(b"%s\0%d\n" % (file, offset))
         self._file.flush()
 
     @active
     def nest(self, name='<unnamed>'):
         self._count += 1
         self._usages += 1
         self._names.append(name)
         return self
 
     def release(self):
         if self._count > 0:
             self._usages -= 1
         if self._names:
             self._names.pop()
         # if the transaction scopes are left without being closed, fail
         if self._count > 0 and self._usages == 0:
             self._abort()
 
     def running(self):
         return self._count > 0
 
     def addpending(self, category, callback):
         """add a callback to be called when the transaction is pending
 
         The transaction will be given as callback's first argument.
 
         Category is a unique identifier to allow overwriting an old callback
         with a newer callback.
         """
         self._pendingcallback[category] = callback
 
     @active
     def writepending(self):
         """write pending file to temporary version
 
         This is used to allow hooks to view a transaction before commit"""
         categories = sorted(self._pendingcallback)
         for cat in categories:
             # remove callback since the data will have been flushed
             any = self._pendingcallback.pop(cat)(self)
             self._anypending = self._anypending or any
         self._anypending |= self._generatefiles(suffix=b'.pending')
         return self._anypending
 
     @active
     def hasfinalize(self, category):
         """check is a callback already exist for a category"""
         return category in self._finalizecallback
 
     @active
     def addfinalize(self, category, callback):
         """add a callback to be called when the transaction is closed
 
         The transaction will be given as callback's first argument.
 
         Category is a unique identifier to allow overwriting old callbacks with
         newer callbacks.
         """
         self._finalizecallback[category] = callback
 
     @active
     def addpostclose(self, category, callback):
         """add or replace a callback to be called after the transaction closed
 
         The transaction will be given as callback's first argument.
 
         Category is a unique identifier to allow overwriting an old callback
         with a newer callback.
         """
         self._postclosecallback[category] = callback
 
     @active
     def getpostclose(self, category):
         """return a postclose callback added before, or None"""
         return self._postclosecallback.get(category, None)
 
     @active
     def addabort(self, category, callback):
         """add a callback to be called when the transaction is aborted.
 
         The transaction will be given as the first argument to the callback.
 
         Category is a unique identifier to allow overwriting an old callback
         with a newer callback.
         """
         self._abortcallback[category] = callback
 
     @active
     def addvalidator(self, category, callback):
         """adds a callback to be called when validating the transaction.
 
         The transaction will be given as the first argument to the callback.
 
         callback should raise exception if to abort transaction"""
         self._validatecallback[category] = callback
 
     @active
     def close(self):
         '''commit the transaction'''
         if self._count == 1:
             for category in sorted(self._validatecallback):
                 self._validatecallback[category](self)
             self._validatecallback = None  # Help prevent cycles.
             self._generatefiles(group=GEN_GROUP_PRE_FINALIZE)
             while self._finalizecallback:
                 callbacks = self._finalizecallback
                 self._finalizecallback = {}
                 categories = sorted(callbacks)
                 for cat in categories:
                     callbacks[cat](self)
             # Prevent double usage and help clear cycles.
             self._finalizecallback = None
             self._generatefiles(group=GEN_GROUP_POST_FINALIZE)
 
         self._count -= 1
         if self._count != 0:
             return
         self._file.close()
         self._backupsfile.close()
         # cleanup temporary files
         for l, f, b, c in self._backupentries:
             if l not in self._vfsmap and c:
                 self._report(
                     b"couldn't remove %s: unknown cache location %s\n" % (b, l)
                 )
                 continue
             vfs = self._vfsmap[l]
             if not f and b and vfs.exists(b):
                 try:
                     vfs.unlink(b)
                 except (IOError, OSError, error.Abort) as inst:
                     if not c:
                         raise
                     # Abort may be raise by read only opener
                     self._report(
                         b"couldn't remove %s: %s\n" % (vfs.join(b), inst)
                     )
         self._offsetmap = {}
         self._newfiles = set()
         self._writeundo()
         if self._after:
             self._after()
             self._after = None  # Help prevent cycles.
         if self._opener.isfile(self._backupjournal):
             self._opener.unlink(self._backupjournal)
         if self._opener.isfile(self._journal):
             self._opener.unlink(self._journal)
         for l, _f, b, c in self._backupentries:
             if l not in self._vfsmap and c:
                 self._report(
                     b"couldn't remove %s: unknown cache location"
                     b"%s\n" % (b, l)
                 )
                 continue
             vfs = self._vfsmap[l]
             if b and vfs.exists(b):
                 try:
                     vfs.unlink(b)
                 except (IOError, OSError, error.Abort) as inst:
                     if not c:
                         raise
                     # Abort may be raise by read only opener
                     self._report(
                         b"couldn't remove %s: %s\n" % (vfs.join(b), inst)
                     )
         self._backupentries = []
         self._journal = None
 
         self._releasefn(self, True)  # notify success of closing transaction
         self._releasefn = None  # Help prevent cycles.
 
         # run post close action
         categories = sorted(self._postclosecallback)
         for cat in categories:
             self._postclosecallback[cat](self)
         # Prevent double usage and help clear cycles.
         self._postclosecallback = None
 
     @active
     def abort(self):
         """abort the transaction (generally called on error, or when the
         transaction is not explicitly committed before going out of
         scope)"""
         self._abort()
 
     def _writeundo(self):
         """write transaction data for possible future undo call"""
         if self._undoname is None:
             return
 
         undo_backup_path = b"%s.backupfiles" % self._undoname
         undobackupfile = self._opener.open(undo_backup_path, b'w')
         undobackupfile.write(b'%d\n' % version)
         for l, f, b, c in self._backupentries:
             if not f:  # temporary file
                 continue
             if not b:
                 u = b''
             else:
                 if l not in self._vfsmap and c:
                     self._report(
                         b"couldn't remove %s: unknown cache location"
                         b"%s\n" % (b, l)
                     )
                     continue
                 vfs = self._vfsmap[l]
                 base, name = vfs.split(b)
                 assert name.startswith(self._journal), name
                 uname = name.replace(self._journal, self._undoname, 1)
                 u = vfs.reljoin(base, uname)
                 util.copyfile(vfs.join(b), vfs.join(u), hardlink=True)
             undobackupfile.write(b"%s\0%s\0%s\0%d\n" % (l, f, u, c))
         undobackupfile.close()
 
     def _abort(self):
         entries = self.readjournal()
         self._count = 0
         self._usages = 0
         self._file.close()
         self._backupsfile.close()
 
         try:
             if not entries and not self._backupentries:
                 if self._backupjournal:
                     self._opener.unlink(self._backupjournal)
                 if self._journal:
                     self._opener.unlink(self._journal)
                 return
 
             self._report(_(b"transaction abort!\n"))
 
             try:
                 for cat in sorted(self._abortcallback):
                     self._abortcallback[cat](self)
                 # Prevent double usage and help clear cycles.
                 self._abortcallback = None
                 _playback(
                     self._journal,
                     self._report,
                     self._opener,
                     self._vfsmap,
                     entries,
                     self._backupentries,
                     False,
                     checkambigfiles=self._checkambigfiles,
                 )
                 self._report(_(b"rollback completed\n"))
             except BaseException as exc:
                 self._report(_(b"rollback failed - please run hg recover\n"))
                 self._report(
                     _(b"(failure reason: %s)\n") % stringutil.forcebytestr(exc)
                 )
         finally:
             self._journal = None
             self._releasefn(self, False)  # notify failure of transaction
             self._releasefn = None  # Help prevent cycles.
 
 
 BAD_VERSION_MSG = _(
     b"journal was created by a different version of Mercurial\n"
 )
 
 
 def rollback(opener, vfsmap, file, report, checkambigfiles=None):
     """Rolls back the transaction contained in the given file
 
     Reads the entries in the specified file, and the corresponding
     '*.backupfiles' file, to recover from an incomplete transaction.
 
     * `file`: a file containing a list of entries, specifying where
     to truncate each file.  The file should contain a list of
     file\0offset pairs, delimited by newlines. The corresponding
     '*.backupfiles' file should contain a list of file\0backupfile
     pairs, delimited by \0.
 
     `checkambigfiles` is a set of (path, vfs-location) tuples,
     which determine whether file stat ambiguity should be avoided at
     restoring corresponded files.
     """
     entries = []
     backupentries = []
 
     with opener.open(file) as fp:
         lines = fp.readlines()
     for l in lines:
         try:
             f, o = l.split(b'\0')
             entries.append((f, int(o)))
         except ValueError:
             report(
                 _(b"couldn't read journal entry %r!\n") % pycompat.bytestr(l)
             )
 
     backupjournal = b"%s.backupfiles" % file
     if opener.exists(backupjournal):
         fp = opener.open(backupjournal)
         lines = fp.readlines()
         if lines:
             ver = lines[0][:-1]
             if ver != (b'%d' % version):
                 report(BAD_VERSION_MSG)
             else:
                 for line in lines[1:]:
                     if line:
                         # Shave off the trailing newline
                         line = line[:-1]
                         l, f, b, c = line.split(b'\0')
                         backupentries.append((l, f, b, bool(c)))
 
     _playback(
         file,
         report,
         opener,
         vfsmap,
         entries,
         backupentries,
         checkambigfiles=checkambigfiles,
     )