treemanifest: store submanifest revlog per directory

With this change, when tree manifests are enabled (in .hg/requires),
commits will be written with one manifest revlog per directory. The
manifest revlogs are stored in
.hg/store/meta/$dir/00manifest.[id].

Flat manifests can still be read and interacted with as usual (they
are also read into treemanifest instances). The functionality for
writing treemanifest as a flat manifest to disk is still left in the
code; tests still pass with '_treeinmem=True' hardcoded.

Exchange is not yet implemented.

tests/test-treemanifest.t

@@ -0,0 +1,278
+
+Set up repo
+
+  $ hg --config experimental.treemanifest=True init repo
+  $ cd repo
+
+Requirements get set on init
+
+  $ grep treemanifest .hg/requires
+  treemanifest
+
+Without directories, looks like any other repo
+
+  $ echo 0 > a
+  $ echo 0 > b
+  $ hg ci -Aqm initial
+  $ hg debugdata -m 0
+  a\x00362fef284ce2ca02aecc8de6d5e8a1c3af0556fe (esc)
+  b\x00362fef284ce2ca02aecc8de6d5e8a1c3af0556fe (esc)
+
+Submanifest is stored in separate revlog
+
+  $ mkdir dir1
+  $ echo 1 > dir1/a
+  $ echo 1 > dir1/b
+  $ echo 1 > e
+  $ hg ci -Aqm 'add dir1'
+  $ hg debugdata -m 1
+  a\x00362fef284ce2ca02aecc8de6d5e8a1c3af0556fe (esc)
+  b\x00362fef284ce2ca02aecc8de6d5e8a1c3af0556fe (esc)
+  dir1\x008b3ffd73f901e83304c83d33132c8e774ceac44ed (esc)
+  e\x00b8e02f6433738021a065f94175c7cd23db5f05be (esc)
+  $ hg debugdata .hg/store/meta/dir1/00manifest.i 0
+  a\x00b8e02f6433738021a065f94175c7cd23db5f05be (esc)
+  b\x00b8e02f6433738021a065f94175c7cd23db5f05be (esc)
+
+Can add nested directories
+
+  $ mkdir dir1/dir1
+  $ echo 2 > dir1/dir1/a
+  $ echo 2 > dir1/dir1/b
+  $ mkdir dir1/dir2
+  $ echo 2 > dir1/dir2/a
+  $ echo 2 > dir1/dir2/b
+  $ hg ci -Aqm 'add dir1/dir1'
+  $ hg files -r .
+  a
+  b
+  dir1/a
+  dir1/b
+  dir1/dir1/a
+  dir1/dir1/b
+  dir1/dir2/a
+  dir1/dir2/b
+  e
+
+Revision is not created for unchanged directory
+
+  $ mkdir dir2
+  $ echo 3 > dir2/a
+  $ hg add dir2
+  adding dir2/a
+  $ hg debugindex .hg/store/meta/dir1/00manifest.i > before
+  $ hg ci -qm 'add dir2'
+  $ hg debugindex .hg/store/meta/dir1/00manifest.i > after
+  $ diff before after
+  $ rm before after
+
+Removing directory does not create an revlog entry
+
+  $ hg rm dir1/dir1
+  removing dir1/dir1/a
+  removing dir1/dir1/b
+  $ hg debugindex .hg/store/meta/dir1/dir1/00manifest.i > before
+  $ hg ci -qm 'remove dir1/dir1'
+  $ hg debugindex .hg/store/meta/dir1/dir1/00manifest.i > after
+  $ diff before after
+  $ rm before after
+
+Check that hg files (calls treemanifest.walk()) works
+
+  $ hg co 'desc("add dir2")'
+  2 files updated, 0 files merged, 0 files removed, 0 files unresolved
+  $ hg files -r . dir1
+  dir1/a
+  dir1/b
+  dir1/dir1/a
+  dir1/dir1/b
+  dir1/dir2/a
+  dir1/dir2/b
+
+Check that status between revisions works (calls treemanifest.matches())
+
+  $ hg status --rev 'desc("add dir1")' --rev . dir1
+  A dir1/dir1/a
+  A dir1/dir1/b
+  A dir1/dir2/a
+  A dir1/dir2/b
+
+Merge creates 2-parent revision of directory revlog
+
+  $ echo 5 > dir1/a
+  $ hg ci -Aqm 'modify dir1/a'
+  $ hg co '.^'
+  1 files updated, 0 files merged, 0 files removed, 0 files unresolved
+  $ echo 6 > dir1/b
+  $ hg ci -Aqm 'modify dir1/b'
+  $ hg merge 'desc("modify dir1/a")'
+  1 files updated, 0 files merged, 0 files removed, 0 files unresolved
+  (branch merge, don't forget to commit)
+  $ hg ci -m 'conflict-free merge involving dir1/'
+  $ cat dir1/a
+  5
+  $ cat dir1/b
+  6
+  $ hg debugindex .hg/store/meta/dir1/00manifest.i
+     rev    offset  length   base linkrev nodeid       p1           p2
+       0         0      54      0       1 8b3ffd73f901 000000000000 000000000000
+       1        54      68      0       2 b66d046c644f 8b3ffd73f901 000000000000
+       2       122      12      0       4 b87265673c8a b66d046c644f 000000000000
+       3       134      95      0       5 aa5d3adcec72 b66d046c644f 000000000000
+       4       229      81      0       6 e29b066b91ad b66d046c644f 000000000000
+       5       310     107      5       7 a120ce2b83f5 e29b066b91ad aa5d3adcec72
+
+Merge keeping directory from parent 1 does not create revlog entry. (Note that
+dir1's manifest does change, but only because dir1/a's filelog changes.)
+
+  $ hg co 'desc("add dir2")'
+  2 files updated, 0 files merged, 0 files removed, 0 files unresolved
+  $ echo 8 > dir2/a
+  $ hg ci -m 'modify dir2/a'
+  created new head
+
+  $ hg debugindex .hg/store/meta/dir2/00manifest.i > before
+  $ hg merge 'desc("modify dir1/a")'
+  1 files updated, 0 files merged, 0 files removed, 0 files unresolved
+  (branch merge, don't forget to commit)
+  $ hg revert -r 'desc("modify dir2/a")' .
+  reverting dir1/a (glob)
+  $ hg ci -m 'merge, keeping parent 1'
+  $ hg debugindex .hg/store/meta/dir2/00manifest.i > after
+  $ diff before after
+  $ rm before after
+
+Merge keeping directory from parent 2 does not create revlog entry. (Note that
+dir2's manifest does change, but only because dir2/a's filelog changes.)
+
+  $ hg co 'desc("modify dir2/a")'
+  1 files updated, 0 files merged, 0 files removed, 0 files unresolved
+  $ hg debugindex .hg/store/meta/dir1/00manifest.i > before
+  $ hg merge 'desc("modify dir1/a")'
+  1 files updated, 0 files merged, 0 files removed, 0 files unresolved
+  (branch merge, don't forget to commit)
+  $ hg revert -r 'desc("modify dir1/a")' .
+  reverting dir2/a (glob)
+  $ hg ci -m 'merge, keeping parent 2'
+  created new head
+  $ hg debugindex .hg/store/meta/dir1/00manifest.i > after
+  $ diff before after
+  $ rm before after
+
+Create flat source repo for tests with mixed flat/tree manifests
+
+  $ cd ..
+  $ hg init repo-flat
+  $ cd repo-flat
+
+Create a few commits with flat manifest
+
+  $ echo 0 > a
+  $ echo 0 > b
+  $ echo 0 > e
+  $ for d in dir1 dir1/dir1 dir1/dir2 dir2
+  > do
+  >   mkdir $d
+  >   echo 0 > $d/a
+  >   echo 0 > $d/b
+  > done
+  $ hg ci -Aqm initial
+
+  $ echo 1 > a
+  $ echo 1 > dir1/a
+  $ echo 1 > dir1/dir1/a
+  $ hg ci -Aqm 'modify on branch 1'
+
+  $ hg co 0
+  3 files updated, 0 files merged, 0 files removed, 0 files unresolved
+  $ echo 2 > b
+  $ echo 2 > dir1/b
+  $ echo 2 > dir1/dir1/b
+  $ hg ci -Aqm 'modify on branch 2'
+
+  $ hg merge 1
+  3 files updated, 0 files merged, 0 files removed, 0 files unresolved
+  (branch merge, don't forget to commit)
+  $ hg ci -m 'merge of flat manifests to new flat manifest'
+
+Create clone with tree manifests enabled
+
+  $ cd ..
+  $ hg clone --pull --config experimental.treemanifest=1 repo-flat repo-mixed
+  requesting all changes
+  adding changesets
+  adding manifests
+  adding file changes
+  added 4 changesets with 17 changes to 11 files
+  updating to branch default
+  11 files updated, 0 files merged, 0 files removed, 0 files unresolved
+  $ cd repo-mixed
+  $ test -f .hg/store/meta
+  [1]
+  $ grep treemanifest .hg/requires
+  treemanifest
+
+Commit should store revlog per directory
+
+  $ hg co 1
+  3 files updated, 0 files merged, 0 files removed, 0 files unresolved
+  $ echo 3 > a
+  $ echo 3 > dir1/a
+  $ echo 3 > dir1/dir1/a
+  $ hg ci -m 'first tree'
+  created new head
+  $ find .hg/store/meta | sort
+  .hg/store/meta
+  .hg/store/meta/dir1
+  .hg/store/meta/dir1/00manifest.i
+  .hg/store/meta/dir1/dir1
+  .hg/store/meta/dir1/dir1/00manifest.i
+  .hg/store/meta/dir1/dir2
+  .hg/store/meta/dir1/dir2/00manifest.i
+  .hg/store/meta/dir2
+  .hg/store/meta/dir2/00manifest.i
+
+Merge of two trees
+
+  $ hg co 2
+  6 files updated, 0 files merged, 0 files removed, 0 files unresolved
+  $ hg merge 1
+  3 files updated, 0 files merged, 0 files removed, 0 files unresolved
+  (branch merge, don't forget to commit)
+  $ hg ci -m 'merge of flat manifests to new tree manifest'
+  created new head
+  $ hg diff -r 3
+
+Parent of tree root manifest should be flat manifest, and two for merge
+
+  $ hg debugindex -m
+     rev    offset  length   base linkrev nodeid       p1           p2
+       0         0      80      0       0 40536115ed9e 000000000000 000000000000
+       1        80      83      0       1 f3376063c255 40536115ed9e 000000000000
+       2       163     103      0       2 5d9b9da231a2 40536115ed9e 000000000000
+       3       266      83      0       3 d17d663cbd8a 5d9b9da231a2 f3376063c255
+       4       349     132      4       4 c05a51345f86 f3376063c255 000000000000
+       5       481     110      4       5 82594b1f557d 5d9b9da231a2 f3376063c255
+
+
+Status across flat/tree boundary should work
+
+  $ hg status --rev '.^' --rev .
+  M a
+  M dir1/a
+  M dir1/dir1/a
+
+
+Turning off treemanifest config has no effect
+
+  $ hg debugindex .hg/store/meta/dir1/00manifest.i
+     rev    offset  length   base linkrev nodeid       p1           p2
+       0         0     125      0       4 63c9c0557d24 000000000000 000000000000
+       1       125     109      0       5 23d12a1f6e0e 000000000000 000000000000
+  $ echo 2 > dir1/a
+  $ hg --config experimental.treemanifest=False ci -qm 'modify dir1/a'
+  $ hg debugindex .hg/store/meta/dir1/00manifest.i
+     rev    offset  length   base linkrev nodeid       p1           p2
+       0         0     125      0       4 63c9c0557d24 000000000000 000000000000
+       1       125     109      0       5 23d12a1f6e0e 000000000000 000000000000
+       2       234      55      0       6 3cb2d87b4250 23d12a1f6e0e 000000000000

mercurial/manifest.py

@@ -1,860 +1,944
 # manifest.py - manifest revision class for mercurial
 #
 # Copyright 2005-2007 Matt Mackall <[e-mail]>
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 from i18n import _
 import mdiff, parsers, error, revlog, util
 import array, struct
 import os
 
 propertycache = util.propertycache
 
 def _parsev1(data):
     # This method does a little bit of excessive-looking
     # precondition checking. This is so that the behavior of this
     # class exactly matches its C counterpart to try and help
     # prevent surprise breakage for anyone that develops against
     # the pure version.
     if data and data[-1] != '\n':
         raise ValueError('Manifest did not end in a newline.')
     prev = None
     for l in data.splitlines():
         if prev is not None and prev > l:
             raise ValueError('Manifest lines not in sorted order.')
         prev = l
         f, n = l.split('\0')
         if len(n) > 40:
             yield f, revlog.bin(n[:40]), n[40:]
         else:
             yield f, revlog.bin(n), ''
 
 def _parsev2(data):
     metadataend = data.find('\n')
     # Just ignore metadata for now
     pos = metadataend + 1
     prevf = ''
     while pos < len(data):
         end = data.find('\n', pos + 1) # +1 to skip stem length byte
         if end == -1:
             raise ValueError('Manifest ended with incomplete file entry.')
         stemlen = ord(data[pos])
         items = data[pos + 1:end].split('\0')
         f = prevf[:stemlen] + items[0]
         if prevf > f:
             raise ValueError('Manifest entries not in sorted order.')
         fl = items[1]
         # Just ignore metadata (items[2:] for now)
         n = data[end + 1:end + 21]
         yield f, n, fl
         pos = end + 22
         prevf = f
 
 def _parse(data):
     """Generates (path, node, flags) tuples from a manifest text"""
     if data.startswith('\0'):
         return iter(_parsev2(data))
     else:
         return iter(_parsev1(data))
 
 def _text(it, usemanifestv2):
     """Given an iterator over (path, node, flags) tuples, returns a manifest
     text"""
     if usemanifestv2:
         return _textv2(it)
     else:
         return _textv1(it)
 
 def _textv1(it):
     files = []
     lines = []
     _hex = revlog.hex
     for f, n, fl in it:
         files.append(f)
         # if this is changed to support newlines in filenames,
         # be sure to check the templates/ dir again (especially *-raw.tmpl)
         lines.append("%s\0%s%s\n" % (f, _hex(n), fl))
 
     _checkforbidden(files)
     return ''.join(lines)
 
 def _textv2(it):
     files = []
     lines = ['\0\n']
     prevf = ''
     for f, n, fl in it:
         files.append(f)
         stem = os.path.commonprefix([prevf, f])
         stemlen = min(len(stem), 255)
         lines.append("%c%s\0%s\n%s\n" % (stemlen, f[stemlen:], fl, n))
         prevf = f
     _checkforbidden(files)
     return ''.join(lines)
 
 class _lazymanifest(dict):
     """This is the pure implementation of lazymanifest.
 
     It has not been optimized *at all* and is not lazy.
     """
 
     def __init__(self, data):
         dict.__init__(self)
         for f, n, fl in _parse(data):
             self[f] = n, fl
 
     def __setitem__(self, k, v):
         node, flag = v
         assert node is not None
         if len(node) > 21:
             node = node[:21] # match c implementation behavior
         dict.__setitem__(self, k, (node, flag))
 
     def __iter__(self):
         return iter(sorted(dict.keys(self)))
 
     def iterkeys(self):
         return iter(sorted(dict.keys(self)))
 
     def iterentries(self):
         return ((f, e[0], e[1]) for f, e in sorted(self.iteritems()))
 
     def copy(self):
         c = _lazymanifest('')
         c.update(self)
         return c
 
     def diff(self, m2, clean=False):
         '''Finds changes between the current manifest and m2.'''
         diff = {}
 
         for fn, e1 in self.iteritems():
             if fn not in m2:
                 diff[fn] = e1, (None, '')
             else:
                 e2 = m2[fn]
                 if e1 != e2:
                     diff[fn] = e1, e2
                 elif clean:
                     diff[fn] = None
 
         for fn, e2 in m2.iteritems():
             if fn not in self:
                 diff[fn] = (None, ''), e2
 
         return diff
 
     def filtercopy(self, filterfn):
         c = _lazymanifest('')
         for f, n, fl in self.iterentries():
             if filterfn(f):
                 c[f] = n, fl
         return c
 
     def text(self):
         """Get the full data of this manifest as a bytestring."""
         return _textv1(self.iterentries())
 
 try:
     _lazymanifest = parsers.lazymanifest
 except AttributeError:
     pass
 
 class manifestdict(object):
     def __init__(self, data=''):
         if data.startswith('\0'):
             #_lazymanifest can not parse v2
             self._lm = _lazymanifest('')
             for f, n, fl in _parsev2(data):
                 self._lm[f] = n, fl
         else:
             self._lm = _lazymanifest(data)
 
     def __getitem__(self, key):
         return self._lm[key][0]
 
     def find(self, key):
         return self._lm[key]
 
     def __len__(self):
         return len(self._lm)
 
     def __setitem__(self, key, node):
         self._lm[key] = node, self.flags(key, '')
 
     def __contains__(self, key):
         return key in self._lm
 
     def __delitem__(self, key):
         del self._lm[key]
 
     def __iter__(self):
         return self._lm.__iter__()
 
     def iterkeys(self):
         return self._lm.iterkeys()
 
     def keys(self):
         return list(self.iterkeys())
 
     def filesnotin(self, m2):
         '''Set of files in this manifest that are not in the other'''
         files = set(self)
         files.difference_update(m2)
         return files
 
     @propertycache
     def _dirs(self):
         return util.dirs(self)
 
     def dirs(self):
         return self._dirs
 
     def hasdir(self, dir):
         return dir in self._dirs
 
     def _filesfastpath(self, match):
         '''Checks whether we can correctly and quickly iterate over matcher
         files instead of over manifest files.'''
         files = match.files()
         return (len(files) < 100 and (match.isexact() or
             (not match.anypats() and util.all(fn in self for fn in files))))
 
     def walk(self, match):
         '''Generates matching file names.
 
         Equivalent to manifest.matches(match).iterkeys(), but without creating
         an entirely new manifest.
 
         It also reports nonexistent files by marking them bad with match.bad().
         '''
         if match.always():
             for f in iter(self):
                 yield f
             return
 
         fset = set(match.files())
 
         # avoid the entire walk if we're only looking for specific files
         if self._filesfastpath(match):
             for fn in sorted(fset):
                 yield fn
             return
 
         for fn in self:
             if fn in fset:
                 # specified pattern is the exact name
                 fset.remove(fn)
             if match(fn):
                 yield fn
 
         # for dirstate.walk, files=['.'] means "walk the whole tree".
         # follow that here, too
         fset.discard('.')
 
         for fn in sorted(fset):
             if not self.hasdir(fn):
                 match.bad(fn, None)
 
     def matches(self, match):
         '''generate a new manifest filtered by the match argument'''
         if match.always():
             return self.copy()
 
         if self._filesfastpath(match):
             m = manifestdict()
             lm = self._lm
             for fn in match.files():
                 if fn in lm:
                     m._lm[fn] = lm[fn]
             return m
 
         m = manifestdict()
         m._lm = self._lm.filtercopy(match)
         return m
 
     def diff(self, m2, clean=False):
         '''Finds changes between the current manifest and m2.
 
         Args:
           m2: the manifest to which this manifest should be compared.
           clean: if true, include files unchanged between these manifests
                  with a None value in the returned dictionary.
 
         The result is returned as a dict with filename as key and
         values of the form ((n1,fl1),(n2,fl2)), where n1/n2 is the
         nodeid in the current/other manifest and fl1/fl2 is the flag
         in the current/other manifest. Where the file does not exist,
         the nodeid will be None and the flags will be the empty
         string.
         '''
         return self._lm.diff(m2._lm, clean)
 
     def setflag(self, key, flag):
         self._lm[key] = self[key], flag
 
     def get(self, key, default=None):
         try:
             return self._lm[key][0]
         except KeyError:
             return default
 
     def flags(self, key, default=''):
         try:
             return self._lm[key][1]
         except KeyError:
             return default
 
     def copy(self):
         c = manifestdict()
         c._lm = self._lm.copy()
         return c
 
     def iteritems(self):
         return (x[:2] for x in self._lm.iterentries())
 
     def text(self, usemanifestv2=False):
         if usemanifestv2:
             return _textv2(self._lm.iterentries())
         else:
             # use (probably) native version for v1
             return self._lm.text()
 
     def fastdelta(self, base, changes):
         """Given a base manifest text as an array.array and a list of changes
         relative to that text, compute a delta that can be used by revlog.
         """
         delta = []
         dstart = None
         dend = None
         dline = [""]
         start = 0
         # zero copy representation of base as a buffer
         addbuf = util.buffer(base)
 
         # start with a readonly loop that finds the offset of
         # each line and creates the deltas
         for f, todelete in changes:
             # bs will either be the index of the item or the insert point
             start, end = _msearch(addbuf, f, start)
             if not todelete:
                 h, fl = self._lm[f]
                 l = "%s\0%s%s\n" % (f, revlog.hex(h), fl)
             else:
                 if start == end:
                     # item we want to delete was not found, error out
                     raise AssertionError(
                             _("failed to remove %s from manifest") % f)
                 l = ""
             if dstart is not None and dstart <= start and dend >= start:
                 if dend < end:
                     dend = end
                 if l:
                     dline.append(l)
             else:
                 if dstart is not None:
                     delta.append([dstart, dend, "".join(dline)])
                 dstart = start
                 dend = end
                 dline = [l]
 
         if dstart is not None:
             delta.append([dstart, dend, "".join(dline)])
         # apply the delta to the base, and get a delta for addrevision
         deltatext, arraytext = _addlistdelta(base, delta)
         return arraytext, deltatext
 
 def _msearch(m, s, lo=0, hi=None):
     '''return a tuple (start, end) that says where to find s within m.
 
     If the string is found m[start:end] are the line containing
     that string.  If start == end the string was not found and
     they indicate the proper sorted insertion point.
 
     m should be a buffer or a string
     s is a string'''
     def advance(i, c):
         while i < lenm and m[i] != c:
             i += 1
         return i
     if not s:
         return (lo, lo)
     lenm = len(m)
     if not hi:
         hi = lenm
     while lo < hi:
         mid = (lo + hi) // 2
         start = mid
         while start > 0 and m[start - 1] != '\n':
             start -= 1
         end = advance(start, '\0')
         if m[start:end] < s:
             # we know that after the null there are 40 bytes of sha1
             # this translates to the bisect lo = mid + 1
             lo = advance(end + 40, '\n') + 1
         else:
             # this translates to the bisect hi = mid
             hi = start
     end = advance(lo, '\0')
     found = m[lo:end]
     if s == found:
         # we know that after the null there are 40 bytes of sha1
         end = advance(end + 40, '\n')
         return (lo, end + 1)
     else:
         return (lo, lo)
 
 def _checkforbidden(l):
     """Check filenames for illegal characters."""
     for f in l:
         if '\n' in f or '\r' in f:
             raise error.RevlogError(
                 _("'\\n' and '\\r' disallowed in filenames: %r") % f)
 
 
 # apply the changes collected during the bisect loop to our addlist
 # return a delta suitable for addrevision
 def _addlistdelta(addlist, x):
     # for large addlist arrays, building a new array is cheaper
     # than repeatedly modifying the existing one
     currentposition = 0
     newaddlist = array.array('c')
 
     for start, end, content in x:
         newaddlist += addlist[currentposition:start]
         if content:
             newaddlist += array.array('c', content)
 
         currentposition = end
 
     newaddlist += addlist[currentposition:]
 
     deltatext = "".join(struct.pack(">lll", start, end, len(content))
                    + content for start, end, content in x)
     return deltatext, newaddlist
 
 def _splittopdir(f):
     if '/' in f:
         dir, subpath = f.split('/', 1)
         return dir + '/', subpath
     else:
         return '', f
 
 class treemanifest(object):
     def __init__(self, dir='', text=''):
         self._dir = dir
+        self._node = revlog.nullid
         self._dirs = {}
         # Using _lazymanifest here is a little slower than plain old dicts
         self._files = {}
         self._flags = {}
-        self.parse(text)
+        def readsubtree(subdir, subm):
+            raise AssertionError('treemanifest constructor only accepts '
+                                 'flat manifests')
+        self.parse(text, readsubtree)
 
     def _subpath(self, path):
         return self._dir + path
 
     def __len__(self):
         size = len(self._files)
         for m in self._dirs.values():
             size += m.__len__()
         return size
 
     def _isempty(self):
         return (not self._files and (not self._dirs or
                 util.all(m._isempty() for m in self._dirs.values())))
 
     def __str__(self):
-        return '<treemanifest dir=%s>' % self._dir
+        return ('<treemanifest dir=%s, node=%s>' %
+                (self._dir, revlog.hex(self._node)))
+
+    def dir(self):
+        '''The directory that this tree manifest represents, including a
+        trailing '/'. Empty string for the repo root directory.'''
+        return self._dir
+
+    def node(self):
+        '''This node of this instance. nullid for unsaved instances. Should
+        be updated when the instance is read or written from a revlog.
+        '''
+        return self._node
+
+    def setnode(self, node):
+        self._node = node
 
     def iteritems(self):
         for p, n in sorted(self._dirs.items() + self._files.items()):
             if p in self._files:
                 yield self._subpath(p), n
             else:
                 for f, sn in n.iteritems():
                     yield f, sn
 
     def iterkeys(self):
         for p in sorted(self._dirs.keys() + self._files.keys()):
             if p in self._files:
                 yield self._subpath(p)
             else:
                 for f in self._dirs[p].iterkeys():
                     yield f
 
     def keys(self):
         return list(self.iterkeys())
 
     def __iter__(self):
         return self.iterkeys()
 
     def __contains__(self, f):
         if f is None:
             return False
         dir, subpath = _splittopdir(f)
         if dir:
             if dir not in self._dirs:
                 return False
             return self._dirs[dir].__contains__(subpath)
         else:
             return f in self._files
 
     def get(self, f, default=None):
         dir, subpath = _splittopdir(f)
         if dir:
             if dir not in self._dirs:
                 return default
             return self._dirs[dir].get(subpath, default)
         else:
             return self._files.get(f, default)
 
     def __getitem__(self, f):
         dir, subpath = _splittopdir(f)
         if dir:
             return self._dirs[dir].__getitem__(subpath)
         else:
             return self._files[f]
 
     def flags(self, f):
         dir, subpath = _splittopdir(f)
         if dir:
             if dir not in self._dirs:
                 return ''
             return self._dirs[dir].flags(subpath)
         else:
             if f in self._dirs:
                 return ''
             return self._flags.get(f, '')
 
     def find(self, f):
         dir, subpath = _splittopdir(f)
         if dir:
             return self._dirs[dir].find(subpath)
         else:
             return self._files[f], self._flags.get(f, '')
 
     def __delitem__(self, f):
         dir, subpath = _splittopdir(f)
         if dir:
             self._dirs[dir].__delitem__(subpath)
             # If the directory is now empty, remove it
             if self._dirs[dir]._isempty():
                 del self._dirs[dir]
         else:
             del self._files[f]
             if f in self._flags:
                 del self._flags[f]
 
     def __setitem__(self, f, n):
         assert n is not None
         dir, subpath = _splittopdir(f)
         if dir:
             if dir not in self._dirs:
                 self._dirs[dir] = treemanifest(self._subpath(dir))
             self._dirs[dir].__setitem__(subpath, n)
         else:
             self._files[f] = n[:21] # to match manifestdict's behavior
 
     def setflag(self, f, flags):
         """Set the flags (symlink, executable) for path f."""
+        assert 'd' not in flags
         dir, subpath = _splittopdir(f)
         if dir:
             if dir not in self._dirs:
                 self._dirs[dir] = treemanifest(self._subpath(dir))
             self._dirs[dir].setflag(subpath, flags)
         else:
             self._flags[f] = flags
 
     def copy(self):
         copy = treemanifest(self._dir)
+        copy._node = self._node
         for d in self._dirs:
             copy._dirs[d] = self._dirs[d].copy()
         copy._files = dict.copy(self._files)
         copy._flags = dict.copy(self._flags)
         return copy
 
     def filesnotin(self, m2):
         '''Set of files in this manifest that are not in the other'''
         files = set()
         def _filesnotin(t1, t2):
             for d, m1 in t1._dirs.iteritems():
                 if d in t2._dirs:
                     m2 = t2._dirs[d]
                     _filesnotin(m1, m2)
                 else:
                     files.update(m1.iterkeys())
 
             for fn in t1._files.iterkeys():
                 if fn not in t2._files:
                     files.add(t1._subpath(fn))
 
         _filesnotin(self, m2)
         return files
 
     @propertycache
     def _alldirs(self):
         return util.dirs(self)
 
     def dirs(self):
         return self._alldirs
 
     def hasdir(self, dir):
         topdir, subdir = _splittopdir(dir)
         if topdir:
             if topdir in self._dirs:
                 return self._dirs[topdir].hasdir(subdir)
             return False
         return (dir + '/') in self._dirs
 
     def walk(self, match):
         '''Generates matching file names.
 
         Equivalent to manifest.matches(match).iterkeys(), but without creating
         an entirely new manifest.
 
         It also reports nonexistent files by marking them bad with match.bad().
         '''
         if match.always():
             for f in iter(self):
                 yield f
             return
 
         fset = set(match.files())
 
         for fn in self._walk(match):
             if fn in fset:
                 # specified pattern is the exact name
                 fset.remove(fn)
             yield fn
 
         # for dirstate.walk, files=['.'] means "walk the whole tree".
         # follow that here, too
         fset.discard('.')
 
         for fn in sorted(fset):
             if not self.hasdir(fn):
                 match.bad(fn, None)
 
     def _walk(self, match, alldirs=False):
         '''Recursively generates matching file names for walk().
 
         Will visit all subdirectories if alldirs is True, otherwise it will
         only visit subdirectories for which match.visitdir is True.'''
 
         if not alldirs:
             # substring to strip trailing slash
             visit = match.visitdir(self._dir[:-1] or '.')
             if not visit:
                 return
             alldirs = (visit == 'all')
 
         # yield this dir's files and walk its submanifests
         for p in sorted(self._dirs.keys() + self._files.keys()):
             if p in self._files:
                 fullp = self._subpath(p)
                 if match(fullp):
                     yield fullp
             else:
                 for f in self._dirs[p]._walk(match, alldirs):
                     yield f
 
     def matches(self, match):
         '''generate a new manifest filtered by the match argument'''
         if match.always():
             return self.copy()
 
         return self._matches(match)
 
     def _matches(self, match, alldirs=False):
         '''recursively generate a new manifest filtered by the match argument.
 
         Will visit all subdirectories if alldirs is True, otherwise it will
         only visit subdirectories for which match.visitdir is True.'''
 
         ret = treemanifest(self._dir)
         if not alldirs:
             # substring to strip trailing slash
             visit = match.visitdir(self._dir[:-1] or '.')
             if not visit:
                 return ret
             alldirs = (visit == 'all')
 
         for fn in self._files:
             fullp = self._subpath(fn)
             if not match(fullp):
                 continue
             ret._files[fn] = self._files[fn]
             if fn in self._flags:
                 ret._flags[fn] = self._flags[fn]
 
         for dir, subm in self._dirs.iteritems():
             m = subm._matches(match, alldirs)
             if not m._isempty():
                 ret._dirs[dir] = m
 
         return ret
 
     def diff(self, m2, clean=False):
         '''Finds changes between the current manifest and m2.
 
         Args:
           m2: the manifest to which this manifest should be compared.
           clean: if true, include files unchanged between these manifests
                  with a None value in the returned dictionary.
 
         The result is returned as a dict with filename as key and
         values of the form ((n1,fl1),(n2,fl2)), where n1/n2 is the
         nodeid in the current/other manifest and fl1/fl2 is the flag
         in the current/other manifest. Where the file does not exist,
         the nodeid will be None and the flags will be the empty
         string.
         '''
         result = {}
         emptytree = treemanifest()
         def _diff(t1, t2):
             for d, m1 in t1._dirs.iteritems():
                 m2 = t2._dirs.get(d, emptytree)
                 _diff(m1, m2)
 
             for d, m2 in t2._dirs.iteritems():
                 if d not in t1._dirs:
                     _diff(emptytree, m2)
 
             for fn, n1 in t1._files.iteritems():
                 fl1 = t1._flags.get(fn, '')
                 n2 = t2._files.get(fn, None)
                 fl2 = t2._flags.get(fn, '')
                 if n1 != n2 or fl1 != fl2:
                     result[t1._subpath(fn)] = ((n1, fl1), (n2, fl2))
                 elif clean:
                     result[t1._subpath(fn)] = None
 
             for fn, n2 in t2._files.iteritems():
                 if fn not in t1._files:
                     fl2 = t2._flags.get(fn, '')
                     result[t2._subpath(fn)] = ((None, ''), (n2, fl2))
 
         _diff(self, m2)
         return result
 
-    def parse(self, text):
+    def parse(self, text, readsubtree):
         for f, n, fl in _parse(text):
-            self[f] = n
-            if fl:
-                self.setflag(f, fl)
+            if fl == 'd':
+                f = f + '/'
+                self._dirs[f] = readsubtree(self._subpath(f), n)
+            else:
+                # Use __setitem__ and setflag rather than assigning directly
+                # to _files and _flags, thereby letting us parse flat manifests
+                # as well as tree manifests.
+                self[f] = n
+                if fl:
+                    self.setflag(f, fl)
 
     def text(self, usemanifestv2=False):
         """Get the full data of this manifest as a bytestring."""
         flags = self.flags
         return _text(((f, self[f], flags(f)) for f in self.keys()),
                      usemanifestv2)
 
+    def dirtext(self, usemanifestv2=False):
+        """Get the full data of this directory as a bytestring. Make sure that
+        any submanifests have been written first, so their nodeids are correct.
+        """
+        flags = self.flags
+        dirs = [(d[:-1], self._dirs[d]._node, 'd') for d in self._dirs]
+        files = [(f, self._files[f], flags(f)) for f in self._files]
+        return _text(sorted(dirs + files), usemanifestv2)
+
+    def writesubtrees(self, m1, m2, writesubtree):
+        emptytree = treemanifest()
+        for d, subm in self._dirs.iteritems():
+            subp1 = m1._dirs.get(d, emptytree)._node
+            subp2 = m2._dirs.get(d, emptytree)._node
+            if subp1 == revlog.nullid:
+                subp1, subp2 = subp2, subp1
+            writesubtree(subm, subp1, subp2)
+
 class manifest(revlog.revlog):
-    def __init__(self, opener):
+    def __init__(self, opener, dir=''):
         # During normal operations, we expect to deal with not more than four
         # revs at a time (such as during commit --amend). When rebasing large
         # stacks of commits, the number can go up, hence the config knob below.
         cachesize = 4
         usetreemanifest = False
         usemanifestv2 = False
         opts = getattr(opener, 'options', None)
         if opts is not None:
             cachesize = opts.get('manifestcachesize', cachesize)
             usetreemanifest = opts.get('treemanifest', usetreemanifest)
             usemanifestv2 = opts.get('manifestv2', usemanifestv2)
         self._mancache = util.lrucachedict(cachesize)
-        revlog.revlog.__init__(self, opener, "00manifest.i")
         self._treeinmem = usetreemanifest
         self._treeondisk = usetreemanifest
         self._usemanifestv2 = usemanifestv2
+        indexfile = "00manifest.i"
+        if dir:
+            assert self._treeondisk
+            indexfile = "meta/" + dir + "00manifest.i"
+        revlog.revlog.__init__(self, opener, indexfile)
+        self._dir = dir
 
     def _newmanifest(self, data=''):
         if self._treeinmem:
-            return treemanifest('', data)
+            return treemanifest(self._dir, data)
         return manifestdict(data)
 
     def _slowreaddelta(self, node):
         r0 = self.deltaparent(self.rev(node))
         m0 = self.read(self.node(r0))
         m1 = self.read(node)
         md = self._newmanifest()
         for f, ((n0, fl0), (n1, fl1)) in m0.diff(m1).iteritems():
             if n1:
                 md[f] = n1
                 if fl1:
                     md.setflag(f, fl1)
         return md
 
     def readdelta(self, node):
         if self._usemanifestv2 or self._treeondisk:
             return self._slowreaddelta(node)
         r = self.rev(node)
         d = mdiff.patchtext(self.revdiff(self.deltaparent(r), r))
         return self._newmanifest(d)
 
     def readfast(self, node):
         '''use the faster of readdelta or read
 
         This will return a manifest which is either only the files
         added/modified relative to p1, or all files in the
         manifest. Which one is returned depends on the codepath used
         to retrieve the data.
         '''
         r = self.rev(node)
         deltaparent = self.deltaparent(r)
         if deltaparent != revlog.nullrev and deltaparent in self.parentrevs(r):
             return self.readdelta(node)
         return self.read(node)
 
     def read(self, node):
         if node == revlog.nullid:
             return self._newmanifest() # don't upset local cache
         if node in self._mancache:
             return self._mancache[node][0]
         text = self.revision(node)
-        arraytext = array.array('c', text)
-        m = self._newmanifest(text)
+        if self._treeondisk:
+            def readsubtree(dir, subm):
+                sublog = manifest(self.opener, dir)
+                return sublog.read(subm)
+            m = self._newmanifest()
+            m.parse(text, readsubtree)
+            m.setnode(node)
+            arraytext = None
+        else:
+            m = self._newmanifest(text)
+            arraytext = array.array('c', text)
         self._mancache[node] = (m, arraytext)
         return m
 
     def find(self, node, f):
         '''look up entry for a single file efficiently.
         return (node, flags) pair if found, (None, None) if not.'''
         m = self.read(node)
         try:
             return m.find(f)
         except KeyError:
             return None, None
 
     def add(self, m, transaction, link, p1, p2, added, removed):
         if (p1 in self._mancache and not self._treeinmem
             and not self._usemanifestv2):
             # If our first parent is in the manifest cache, we can
             # compute a delta here using properties we know about the
             # manifest up-front, which may save time later for the
             # revlog layer.
 
             _checkforbidden(added)
             # combine the changed lists into one list for sorting
             work = [(x, False) for x in added]
             work.extend((x, True) for x in removed)
             # this could use heapq.merge() (from Python 2.6+) or equivalent
             # since the lists are already sorted
             work.sort()
 
             arraytext, deltatext = m.fastdelta(self._mancache[p1][1], work)
             cachedelta = self.rev(p1), deltatext
             text = util.buffer(arraytext)
             n = self.addrevision(text, transaction, link, p1, p2, cachedelta)
         else:
             # The first parent manifest isn't already loaded, so we'll
             # just encode a fulltext of the manifest and pass that
             # through to the revlog layer, and let it handle the delta
             # process.
-            text = m.text(self._usemanifestv2)
-            arraytext = array.array('c', text)
-            n = self.addrevision(text, transaction, link, p1, p2)
+            if self._treeondisk:
+                m1 = self.read(p1)
+                m2 = self.read(p2)
+                n = self._addtree(m, transaction, link, m1, m2)
+                arraytext = None
+            else:
+                text = m.text(self._usemanifestv2)
+                n = self.addrevision(text, transaction, link, p1, p2)
+                arraytext = array.array('c', text)
 
         self._mancache[n] = (m, arraytext)
 
         return n
+
+    def _addtree(self, m, transaction, link, m1, m2):
+        def writesubtree(subm, subp1, subp2):
+            sublog = manifest(self.opener, subm.dir())
+            sublog.add(subm, transaction, link, subp1, subp2, None, None)
+        m.writesubtrees(m1, m2, writesubtree)
+        text = m.dirtext(self._usemanifestv2)
+        # If the manifest is unchanged compared to one parent,
+        # don't write a new revision
+        if text == m1.dirtext(self._usemanifestv2):
+            n = m1.node()
+        elif text == m2.dirtext(self._usemanifestv2):
+            n = m2.node()
+        else:
+            n = self.addrevision(text, transaction, link, m1.node(), m2.node())
+        # Save nodeid so parent manifest can calculate its nodeid
+        m.setnode(n)
+        return n

mercurial/store.py

@@ -1,542 +1,542
 # store.py - repository store handling for Mercurial
 #
 # Copyright 2008 Matt Mackall <[e-mail]>
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 from i18n import _
 import scmutil, util, parsers
 import os, stat, errno
 
 _sha = util.sha1
 
 # This avoids a collision between a file named foo and a dir named
 # foo.i or foo.d
 def _encodedir(path):
     '''
     >>> _encodedir('data/foo.i')
     'data/foo.i'
     >>> _encodedir('data/foo.i/bla.i')
     'data/foo.i.hg/bla.i'
     >>> _encodedir('data/foo.i.hg/bla.i')
     'data/foo.i.hg.hg/bla.i'
     >>> _encodedir('data/foo.i\\ndata/foo.i/bla.i\\ndata/foo.i.hg/bla.i\\n')
     'data/foo.i\\ndata/foo.i.hg/bla.i\\ndata/foo.i.hg.hg/bla.i\\n'
     '''
     return (path
             .replace(".hg/", ".hg.hg/")
             .replace(".i/", ".i.hg/")
             .replace(".d/", ".d.hg/"))
 
 encodedir = getattr(parsers, 'encodedir', _encodedir)
 
 def decodedir(path):
     '''
     >>> decodedir('data/foo.i')
     'data/foo.i'
     >>> decodedir('data/foo.i.hg/bla.i')
     'data/foo.i/bla.i'
     >>> decodedir('data/foo.i.hg.hg/bla.i')
     'data/foo.i.hg/bla.i'
     '''
     if ".hg/" not in path:
         return path
     return (path
             .replace(".d.hg/", ".d/")
             .replace(".i.hg/", ".i/")
             .replace(".hg.hg/", ".hg/"))
 
 def _buildencodefun():
     '''
     >>> enc, dec = _buildencodefun()
 
     >>> enc('nothing/special.txt')
     'nothing/special.txt'
     >>> dec('nothing/special.txt')
     'nothing/special.txt'
 
     >>> enc('HELLO')
     '_h_e_l_l_o'
     >>> dec('_h_e_l_l_o')
     'HELLO'
 
     >>> enc('hello:world?')
     'hello~3aworld~3f'
     >>> dec('hello~3aworld~3f')
     'hello:world?'
 
     >>> enc('the\x07quick\xADshot')
     'the~07quick~adshot'
     >>> dec('the~07quick~adshot')
     'the\\x07quick\\xadshot'
     '''
     e = '_'
     winreserved = [ord(x) for x in '\\:*?"<>|']
     cmap = dict([(chr(x), chr(x)) for x in xrange(127)])
     for x in (range(32) + range(126, 256) + winreserved):
         cmap[chr(x)] = "~%02x" % x
     for x in range(ord("A"), ord("Z") + 1) + [ord(e)]:
         cmap[chr(x)] = e + chr(x).lower()
     dmap = {}
     for k, v in cmap.iteritems():
         dmap[v] = k
     def decode(s):
         i = 0
         while i < len(s):
             for l in xrange(1, 4):
                 try:
                     yield dmap[s[i:i + l]]
                     i += l
                     break
                 except KeyError:
                     pass
             else:
                 raise KeyError
     return (lambda s: ''.join([cmap[c] for c in s]),
             lambda s: ''.join(list(decode(s))))
 
 _encodefname, _decodefname = _buildencodefun()
 
 def encodefilename(s):
     '''
     >>> encodefilename('foo.i/bar.d/bla.hg/hi:world?/HELLO')
     'foo.i.hg/bar.d.hg/bla.hg.hg/hi~3aworld~3f/_h_e_l_l_o'
     '''
     return _encodefname(encodedir(s))
 
 def decodefilename(s):
     '''
     >>> decodefilename('foo.i.hg/bar.d.hg/bla.hg.hg/hi~3aworld~3f/_h_e_l_l_o')
     'foo.i/bar.d/bla.hg/hi:world?/HELLO'
     '''
     return decodedir(_decodefname(s))
 
 def _buildlowerencodefun():
     '''
     >>> f = _buildlowerencodefun()
     >>> f('nothing/special.txt')
     'nothing/special.txt'
     >>> f('HELLO')
     'hello'
     >>> f('hello:world?')
     'hello~3aworld~3f'
     >>> f('the\x07quick\xADshot')
     'the~07quick~adshot'
     '''
     winreserved = [ord(x) for x in '\\:*?"<>|']
     cmap = dict([(chr(x), chr(x)) for x in xrange(127)])
     for x in (range(32) + range(126, 256) + winreserved):
         cmap[chr(x)] = "~%02x" % x
     for x in range(ord("A"), ord("Z") + 1):
         cmap[chr(x)] = chr(x).lower()
     return lambda s: "".join([cmap[c] for c in s])
 
 lowerencode = getattr(parsers, 'lowerencode', None) or _buildlowerencodefun()
 
 # Windows reserved names: con, prn, aux, nul, com1..com9, lpt1..lpt9
 _winres3 = ('aux', 'con', 'prn', 'nul') # length 3
 _winres4 = ('com', 'lpt')               # length 4 (with trailing 1..9)
 def _auxencode(path, dotencode):
     '''
     Encodes filenames containing names reserved by Windows or which end in
     period or space. Does not touch other single reserved characters c.
     Specifically, c in '\\:*?"<>|' or ord(c) <= 31 are *not* encoded here.
     Additionally encodes space or period at the beginning, if dotencode is
     True. Parameter path is assumed to be all lowercase.
     A segment only needs encoding if a reserved name appears as a
     basename (e.g. "aux", "aux.foo"). A directory or file named "foo.aux"
     doesn't need encoding.
 
     >>> s = '.foo/aux.txt/txt.aux/con/prn/nul/foo.'
     >>> _auxencode(s.split('/'), True)
     ['~2efoo', 'au~78.txt', 'txt.aux', 'co~6e', 'pr~6e', 'nu~6c', 'foo~2e']
     >>> s = '.com1com2/lpt9.lpt4.lpt1/conprn/com0/lpt0/foo.'
     >>> _auxencode(s.split('/'), False)
     ['.com1com2', 'lp~749.lpt4.lpt1', 'conprn', 'com0', 'lpt0', 'foo~2e']
     >>> _auxencode(['foo. '], True)
     ['foo.~20']
     >>> _auxencode([' .foo'], True)
     ['~20.foo']
     '''
     for i, n in enumerate(path):
         if not n:
             continue
         if dotencode and n[0] in '. ':
             n = "~%02x" % ord(n[0]) + n[1:]
             path[i] = n
         else:
             l = n.find('.')
             if l == -1:
                 l = len(n)
             if ((l == 3 and n[:3] in _winres3) or
                 (l == 4 and n[3] <= '9' and n[3] >= '1'
                         and n[:3] in _winres4)):
                 # encode third letter ('aux' -> 'au~78')
                 ec = "~%02x" % ord(n[2])
                 n = n[0:2] + ec + n[3:]
                 path[i] = n
         if n[-1] in '. ':
             # encode last period or space ('foo...' -> 'foo..~2e')
             path[i] = n[:-1] + "~%02x" % ord(n[-1])
     return path
 
 _maxstorepathlen = 120
 _dirprefixlen = 8
 _maxshortdirslen = 8 * (_dirprefixlen + 1) - 4
 
 def _hashencode(path, dotencode):
     digest = _sha(path).hexdigest()
-    le = lowerencode(path[5:]).split('/') # skips prefix 'data/'
+    le = lowerencode(path[5:]).split('/') # skips prefix 'data/' or 'meta/'
     parts = _auxencode(le, dotencode)
     basename = parts[-1]
     _root, ext = os.path.splitext(basename)
     sdirs = []
     sdirslen = 0
     for p in parts[:-1]:
         d = p[:_dirprefixlen]
         if d[-1] in '. ':
             # Windows can't access dirs ending in period or space
             d = d[:-1] + '_'
         if sdirslen == 0:
             t = len(d)
         else:
             t = sdirslen + 1 + len(d)
             if t > _maxshortdirslen:
                 break
         sdirs.append(d)
         sdirslen = t
     dirs = '/'.join(sdirs)
     if len(dirs) > 0:
         dirs += '/'
     res = 'dh/' + dirs + digest + ext
     spaceleft = _maxstorepathlen - len(res)
     if spaceleft > 0:
         filler = basename[:spaceleft]
         res = 'dh/' + dirs + filler + digest + ext
     return res
 
 def _hybridencode(path, dotencode):
     '''encodes path with a length limit
 
     Encodes all paths that begin with 'data/', according to the following.
 
     Default encoding (reversible):
 
     Encodes all uppercase letters 'X' as '_x'. All reserved or illegal
     characters are encoded as '~xx', where xx is the two digit hex code
     of the character (see encodefilename).
     Relevant path components consisting of Windows reserved filenames are
     masked by encoding the third character ('aux' -> 'au~78', see _auxencode).
 
     Hashed encoding (not reversible):
 
     If the default-encoded path is longer than _maxstorepathlen, a
     non-reversible hybrid hashing of the path is done instead.
     This encoding uses up to _dirprefixlen characters of all directory
     levels of the lowerencoded path, but not more levels than can fit into
     _maxshortdirslen.
     Then follows the filler followed by the sha digest of the full path.
     The filler is the beginning of the basename of the lowerencoded path
     (the basename is everything after the last path separator). The filler
     is as long as possible, filling in characters from the basename until
     the encoded path has _maxstorepathlen characters (or all chars of the
     basename have been taken).
     The extension (e.g. '.i' or '.d') is preserved.
 
     The string 'data/' at the beginning is replaced with 'dh/', if the hashed
     encoding was used.
     '''
     path = encodedir(path)
     ef = _encodefname(path).split('/')
     res = '/'.join(_auxencode(ef, dotencode))
     if len(res) > _maxstorepathlen:
         res = _hashencode(path, dotencode)
     return res
 
 def _pathencode(path):
     de = encodedir(path)
     if len(path) > _maxstorepathlen:
         return _hashencode(de, True)
     ef = _encodefname(de).split('/')
     res = '/'.join(_auxencode(ef, True))
     if len(res) > _maxstorepathlen:
         return _hashencode(de, True)
     return res
 
 _pathencode = getattr(parsers, 'pathencode', _pathencode)
 
 def _plainhybridencode(f):
     return _hybridencode(f, False)
 
 def _calcmode(vfs):
     try:
         # files in .hg/ will be created using this mode
         mode = vfs.stat().st_mode
             # avoid some useless chmods
         if (0777 & ~util.umask) == (0777 & mode):
             mode = None
     except OSError:
         mode = None
     return mode
 
 _data = ('data 00manifest.d 00manifest.i 00changelog.d 00changelog.i'
          ' phaseroots obsstore')
 
 class basicstore(object):
     '''base class for local repository stores'''
     def __init__(self, path, vfstype):
         vfs = vfstype(path)
         self.path = vfs.base
         self.createmode = _calcmode(vfs)
         vfs.createmode = self.createmode
         self.rawvfs = vfs
         self.vfs = scmutil.filtervfs(vfs, encodedir)
         self.opener = self.vfs
 
     def join(self, f):
         return self.path + '/' + encodedir(f)
 
     def _walk(self, relpath, recurse):
         '''yields (unencoded, encoded, size)'''
         path = self.path
         if relpath:
             path += '/' + relpath
         striplen = len(self.path) + 1
         l = []
         if self.rawvfs.isdir(path):
             visit = [path]
             readdir = self.rawvfs.readdir
             while visit:
                 p = visit.pop()
                 for f, kind, st in readdir(p, stat=True):
                     fp = p + '/' + f
                     if kind == stat.S_IFREG and f[-2:] in ('.d', '.i'):
                         n = util.pconvert(fp[striplen:])
                         l.append((decodedir(n), n, st.st_size))
                     elif kind == stat.S_IFDIR and recurse:
                         visit.append(fp)
         l.sort()
         return l
 
     def datafiles(self):
         return self._walk('data', True)
 
     def topfiles(self):
         # yield manifest before changelog
         return reversed(self._walk('', False))
 
     def walk(self):
         '''yields (unencoded, encoded, size)'''
         # yield data files first
         for x in self.datafiles():
             yield x
         for x in self.topfiles():
             yield x
 
     def copylist(self):
         return ['requires'] + _data.split()
 
     def write(self, tr):
         pass
 
     def invalidatecaches(self):
         pass
 
     def markremoved(self, fn):
         pass
 
     def __contains__(self, path):
         '''Checks if the store contains path'''
         path = "/".join(("data", path))
         # file?
         if self.vfs.exists(path + ".i"):
             return True
         # dir?
         if not path.endswith("/"):
             path = path + "/"
         return self.vfs.exists(path)
 
 class encodedstore(basicstore):
     def __init__(self, path, vfstype):
         vfs = vfstype(path + '/store')
         self.path = vfs.base
         self.createmode = _calcmode(vfs)
         vfs.createmode = self.createmode
         self.rawvfs = vfs
         self.vfs = scmutil.filtervfs(vfs, encodefilename)
         self.opener = self.vfs
 
     def datafiles(self):
         for a, b, size in self._walk('data', True):
             try:
                 a = decodefilename(a)
             except KeyError:
                 a = None
             yield a, b, size
 
     def join(self, f):
         return self.path + '/' + encodefilename(f)
 
     def copylist(self):
         return (['requires', '00changelog.i'] +
                 ['store/' + f for f in _data.split()])
 
 class fncache(object):
     # the filename used to be partially encoded
     # hence the encodedir/decodedir dance
     def __init__(self, vfs):
         self.vfs = vfs
         self.entries = None
         self._dirty = False
 
     def _load(self):
         '''fill the entries from the fncache file'''
         self._dirty = False
         try:
             fp = self.vfs('fncache', mode='rb')
         except IOError:
             # skip nonexistent file
             self.entries = set()
             return
         self.entries = set(decodedir(fp.read()).splitlines())
         if '' in self.entries:
             fp.seek(0)
             for n, line in enumerate(fp):
                 if not line.rstrip('\n'):
                     t = _('invalid entry in fncache, line %s') % (n + 1)
                     raise util.Abort(t)
         fp.close()
 
     def write(self, tr):
         if self._dirty:
             tr.addbackup('fncache')
             fp = self.vfs('fncache', mode='wb', atomictemp=True)
             if self.entries:
                 fp.write(encodedir('\n'.join(self.entries) + '\n'))
             fp.close()
             self._dirty = False
 
     def add(self, fn):
         if self.entries is None:
             self._load()
         if fn not in self.entries:
             self._dirty = True
             self.entries.add(fn)
 
     def remove(self, fn):
         if self.entries is None:
             self._load()
         try:
             self.entries.remove(fn)
             self._dirty = True
         except KeyError:
             pass
 
     def __contains__(self, fn):
         if self.entries is None:
             self._load()
         return fn in self.entries
 
     def __iter__(self):
         if self.entries is None:
             self._load()
         return iter(self.entries)
 
 class _fncachevfs(scmutil.abstractvfs, scmutil.auditvfs):
     def __init__(self, vfs, fnc, encode):
         scmutil.auditvfs.__init__(self, vfs)
         self.fncache = fnc
         self.encode = encode
 
     def __call__(self, path, mode='r', *args, **kw):
         if mode not in ('r', 'rb') and path.startswith('data/'):
             self.fncache.add(path)
         return self.vfs(self.encode(path), mode, *args, **kw)
 
     def join(self, path):
         if path:
             return self.vfs.join(self.encode(path))
         else:
             return self.vfs.join(path)
 
 class fncachestore(basicstore):
     def __init__(self, path, vfstype, dotencode):
         if dotencode:
             encode = _pathencode
         else:
             encode = _plainhybridencode
         self.encode = encode
         vfs = vfstype(path + '/store')
         self.path = vfs.base
         self.pathsep = self.path + '/'
         self.createmode = _calcmode(vfs)
         vfs.createmode = self.createmode
         self.rawvfs = vfs
         fnc = fncache(vfs)
         self.fncache = fnc
         self.vfs = _fncachevfs(vfs, fnc, encode)
         self.opener = self.vfs
 
     def join(self, f):
         return self.pathsep + self.encode(f)
 
     def getsize(self, path):
         return self.rawvfs.stat(path).st_size
 
     def datafiles(self):
         for f in sorted(self.fncache):
             ef = self.encode(f)
             try:
                 yield f, ef, self.getsize(ef)
             except OSError, err:
                 if err.errno != errno.ENOENT:
                     raise
 
     def copylist(self):
         d = ('data dh fncache phaseroots obsstore'
              ' 00manifest.d 00manifest.i 00changelog.d 00changelog.i')
         return (['requires', '00changelog.i'] +
                 ['store/' + f for f in d.split()])
 
     def write(self, tr):
         self.fncache.write(tr)
 
     def invalidatecaches(self):
         self.fncache.entries = None
 
     def markremoved(self, fn):
         self.fncache.remove(fn)
 
     def _exists(self, f):
         ef = self.encode(f)
         try:
             self.getsize(ef)
             return True
         except OSError, err:
             if err.errno != errno.ENOENT:
                 raise
             # nonexistent entry
             return False
 
     def __contains__(self, path):
         '''Checks if the store contains path'''
         path = "/".join(("data", path))
         # check for files (exact match)
         e = path + '.i'
         if e in self.fncache and self._exists(e):
             return True
         # now check for directories (prefix match)
         if not path.endswith('/'):
             path += '/'
         for e in self.fncache:
             if e.startswith(path) and self._exists(e):
                 return True
         return False
 
 def store(requirements, path, vfstype):
     if 'store' in requirements:
         if 'fncache' in requirements:
             return fncachestore(path, vfstype, 'dotencode' in requirements)
         return encodedstore(path, vfstype)
     return basicstore(path, vfstype)