verify: check directory manifests

In repos with treemanifests, there is no specific verification of
directory manifest revlogs. It simply collects all file nodes by
reading each manifest delta. With treemanifests, that's means calling
the manifest._slowreaddelta(). If there are missing revlog entries in
a subdirectory revlog, 'hg verify' will simply report the exception
that occurred while trying to read the root manifest:


  manifest@0: reading delta 1700e2e92882: meta/b/00manifest.i@67688a370455: no node

This patch changes the verify code to load only the root manifest at
first and verify all revisions of it, then verify all revisions of
each direct subdirectory, and so on, recursively. The above message
becomes

  b/@0: parent-directory manifest refers to unknown revision 67688a370455

Since the new algorithm reads a single revlog at a time and in order,
'hg verify' on a treemanifest version of the hg core repo goes from
~50s to ~14s. As expected, there is no significant difference on a
repo with flat manifests.

mercurial/manifest.py

@@ -1,1052 +1,1072 b''
 # manifest.py - manifest revision class for mercurial
 #
 # Copyright 2005-2007 Matt Mackall <[e-mail]>
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 from __future__ import absolute_import
 
 import array
 import heapq
 import os
 import struct
 
 from .i18n import _
 from . import (
     error,
     mdiff,
     parsers,
     revlog,
     util,
 )
 
 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
             (match.prefix() and all(fn in self for fn in files))))
 
     def walk(self, match):
         '''Generates matching file names.
 
         Equivalent to manifest.matches(match).iterkeys(), but without creating
         an entirely new manifest.
 
         It also reports nonexistent files by marking them bad with match.bad().
         '''
         if match.always():
             for f in iter(self):
                 yield f
             return
 
         fset = set(match.files())
 
         # avoid the entire walk if we're only looking for specific files
         if self._filesfastpath(match):
             for fn in sorted(fset):
                 yield fn
             return
 
         for fn in self:
             if fn in fset:
                 # specified pattern is the exact name
                 fset.remove(fn)
             if match(fn):
                 yield fn
 
         # for dirstate.walk, files=['.'] means "walk the whole tree".
         # follow that here, too
         fset.discard('.')
 
         for fn in sorted(fset):
             if not self.hasdir(fn):
                 match.bad(fn, None)
 
     def matches(self, match):
         '''generate a new manifest filtered by the match argument'''
         if match.always():
             return self.copy()
 
         if self._filesfastpath(match):
             m = manifestdict()
             lm = self._lm
             for fn in match.files():
                 if fn in lm:
                     m._lm[fn] = lm[fn]
             return m
 
         m = manifestdict()
         m._lm = self._lm.filtercopy(match)
         return m
 
     def diff(self, m2, 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 iterentries(self):
+        return 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)
 
         changes = list(changes)
         if len(changes) < 1000:
             # start with a readonly loop that finds the offset of
             # each line and creates the deltas
             for f, todelete in changes:
                 # bs will either be the index of the item or the insert point
                 start, end = _msearch(addbuf, f, start)
                 if not todelete:
                     h, fl = self._lm[f]
                     l = "%s\0%s%s\n" % (f, 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)
         else:
             # For large changes, it's much cheaper to just build the text and
             # diff it.
             arraytext = array.array('c', self.text())
             deltatext = mdiff.textdiff(base, arraytext)
 
         return arraytext, deltatext
 
 def _msearch(m, s, lo=0, hi=None):
     '''return a tuple (start, end) that says where to find s within m.
 
     If the string is found m[start:end] are the line containing
     that string.  If start == end the string was not found and
     they indicate the proper sorted insertion point.
 
     m should be a buffer 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
 
 _noop = lambda s: None
 
 class treemanifest(object):
     def __init__(self, dir='', text=''):
         self._dir = dir
         self._node = revlog.nullid
         self._loadfunc = _noop
         self._copyfunc = _noop
         self._dirty = False
         self._dirs = {}
         # Using _lazymanifest here is a little slower than plain old dicts
         self._files = {}
         self._flags = {}
         if text:
             def readsubtree(subdir, subm):
                 raise AssertionError('treemanifest constructor only accepts '
                                      'flat manifests')
             self.parse(text, readsubtree)
             self._dirty = True # Mark flat manifest dirty after parsing
 
     def _subpath(self, path):
         return self._dir + path
 
     def __len__(self):
         self._load()
         size = len(self._files)
         for m in self._dirs.values():
             size += m.__len__()
         return size
 
     def _isempty(self):
         self._load() # for consistency; already loaded by all callers
         return (not self._files and (not self._dirs or
                 all(m._isempty() for m in self._dirs.values())))
 
     def __repr__(self):
         return ('<treemanifest dir=%s, node=%s, loaded=%s, dirty=%s at 0x%x>' %
                 (self._dir, revlog.hex(self._node),
                  bool(self._loadfunc is _noop),
                  self._dirty, id(self)))
 
     def dir(self):
         '''The directory that this tree manifest represents, including a
         trailing '/'. Empty string for the repo root directory.'''
         return self._dir
 
     def node(self):
         '''This node of this instance. nullid for unsaved instances. Should
         be updated when the instance is read or written from a revlog.
         '''
         assert not self._dirty
         return self._node
 
     def setnode(self, node):
         self._node = node
         self._dirty = False
 
     def iteritems(self):
         self._load()
         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):
         self._load()
         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
         self._load()
         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):
         self._load()
         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):
         self._load()
         dir, subpath = _splittopdir(f)
         if dir:
             return self._dirs[dir].__getitem__(subpath)
         else:
             return self._files[f]
 
     def flags(self, f):
         self._load()
         dir, subpath = _splittopdir(f)
         if dir:
             if dir 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):
         self._load()
         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):
         self._load()
         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]
         self._dirty = True
 
     def __setitem__(self, f, n):
         assert n is not None
         self._load()
         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
         self._dirty = True
 
     def _load(self):
         if self._loadfunc is not _noop:
             lf, self._loadfunc = self._loadfunc, _noop
             lf(self)
         elif self._copyfunc is not _noop:
             cf, self._copyfunc = self._copyfunc, _noop
             cf(self)
 
     def setflag(self, f, flags):
         """Set the flags (symlink, executable) for path f."""
         assert 't' not in flags
         self._load()
         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
         self._dirty = True
 
     def copy(self):
         copy = treemanifest(self._dir)
         copy._node = self._node
         copy._dirty = self._dirty
         if self._copyfunc is _noop:
             def _copyfunc(s):
                 self._load()
                 for d in self._dirs:
                     s._dirs[d] = self._dirs[d].copy()
                 s._files = dict.copy(self._files)
                 s._flags = dict.copy(self._flags)
             if self._loadfunc is _noop:
                 _copyfunc(copy)
             else:
                 copy._copyfunc = _copyfunc
         else:
             copy._copyfunc = self._copyfunc
         return copy
 
     def filesnotin(self, m2):
         '''Set of files in this manifest that are not in the other'''
         files = set()
         def _filesnotin(t1, t2):
             if t1._node == t2._node and not t1._dirty and not t2._dirty:
                 return
             t1._load()
             t2._load()
             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):
         self._load()
         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):
         '''Recursively generates matching file names for walk().'''
         if not match.visitdir(self._dir[:-1] or '.'):
             return
 
         # yield this dir's files and walk its submanifests
         self._load()
         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):
                     yield f
 
     def matches(self, match):
         '''generate a new manifest filtered by the match argument'''
         if match.always():
             return self.copy()
 
         return self._matches(match)
 
     def _matches(self, match):
         '''recursively generate a new manifest filtered by the match argument.
         '''
 
         visit = match.visitdir(self._dir[:-1] or '.')
         if visit == 'all':
             return self.copy()
         ret = treemanifest(self._dir)
         if not visit:
             return ret
 
         self._load()
         for fn in self._files:
             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)
             if not m._isempty():
                 ret._dirs[dir] = m
 
         if not ret._isempty():
             ret._dirty = True
         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):
             if t1._node == t2._node and not t1._dirty and not t2._dirty:
                 return
             t1._load()
             t2._load()
             for d, m1 in t1._dirs.iteritems():
                 m2 = t2._dirs.get(d, emptytree)
                 _diff(m1, m2)
 
             for d, m2 in t2._dirs.iteritems():
                 if d not in t1._dirs:
                     _diff(emptytree, m2)
 
             for fn, n1 in t1._files.iteritems():
                 fl1 = t1._flags.get(fn, '')
                 n2 = t2._files.get(fn, None)
                 fl2 = t2._flags.get(fn, '')
                 if n1 != n2 or fl1 != fl2:
                     result[t1._subpath(fn)] = ((n1, fl1), (n2, fl2))
                 elif clean:
                     result[t1._subpath(fn)] = None
 
             for fn, n2 in t2._files.iteritems():
                 if fn not in t1._files:
                     fl2 = t2._flags.get(fn, '')
                     result[t2._subpath(fn)] = ((None, ''), (n2, fl2))
 
         _diff(self, m2)
         return result
 
     def unmodifiedsince(self, m2):
         return not self._dirty and not m2._dirty and self._node == m2._node
 
     def parse(self, text, readsubtree):
         for f, n, fl in _parse(text):
             if fl == 't':
                 f = f + '/'
                 self._dirs[f] = readsubtree(self._subpath(f), n)
             elif '/' in f:
                 # This is a flat manifest, so use __setitem__ and setflag rather
                 # than assigning directly to _files and _flags, so we can
                 # assign a path in a subdirectory, and to mark dirty (compared
                 # to nullid).
                 self[f] = n
                 if fl:
                     self.setflag(f, fl)
             else:
                 # Assigning to _files and _flags avoids marking as dirty,
                 # and should be a little faster.
                 self._files[f] = n
                 if fl:
                     self._flags[f] = fl
 
     def text(self, usemanifestv2=False):
         """Get the full data of this manifest as a bytestring."""
         self._load()
         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.
         """
         self._load()
         flags = self.flags
         dirs = [(d[:-1], self._dirs[d]._node, 't') for d in self._dirs]
         files = [(f, self._files[f], flags(f)) for f in self._files]
         return _text(sorted(dirs + files), usemanifestv2)
 
     def read(self, gettext, readsubtree):
         def _load_for_read(s):
             s.parse(gettext(), readsubtree)
             s._dirty = False
         self._loadfunc = _load_for_read
 
     def writesubtrees(self, m1, m2, writesubtree):
         self._load() # for consistency; should never have any effect here
         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, dir='', dirlogcache=None):
         '''The 'dir' and 'dirlogcache' arguments are for internal use by
         manifest.manifest only. External users should create a root manifest
         log with manifest.manifest(opener) and call dirlog() on it.
         '''
         # 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)
         self._treeinmem = usetreemanifest
         self._treeondisk = usetreemanifest
         self._usemanifestv2 = usemanifestv2
         indexfile = "00manifest.i"
         if dir:
             assert self._treeondisk
             if not dir.endswith('/'):
                 dir = dir + '/'
             indexfile = "meta/" + dir + "00manifest.i"
         revlog.revlog.__init__(self, opener, indexfile)
         self._dir = dir
         # The dirlogcache is kept on the root manifest log
         if dir:
             self._dirlogcache = dirlogcache
         else:
             self._dirlogcache = {'': self}
 
     def _newmanifest(self, data=''):
         if self._treeinmem:
             return treemanifest(self._dir, data)
         return manifestdict(data)
 
     def dirlog(self, dir):
-        assert self._treeondisk
+        if dir:
+            assert self._treeondisk
         if dir not in self._dirlogcache:
             self._dirlogcache[dir] = manifest(self.opener, dir,
                                               self._dirlogcache)
         return self._dirlogcache[dir]
 
     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 readshallowdelta(self, node):
+        '''For flat manifests, this is the same as readdelta(). For
+        treemanifests, this will read the delta for this revlog's directory,
+        without recursively reading subdirectory manifests. Instead, any
+        subdirectory entry will be reported as it appears in the manifests, i.e.
+        the subdirectory will be reported among files and distinguished only by
+        its 't' flag.'''
+        if not self._treeondisk:
+            return self.readdelta(node)
+        if self._usemanifestv2:
+            raise error.Abort(
+                "readshallowdelta() not implemented for manifestv2")
+        r = self.rev(node)
+        d = mdiff.patchtext(self.revdiff(self.deltaparent(r), r))
+        return manifestdict(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]
         if self._treeondisk:
             def gettext():
                 return self.revision(node)
             def readsubtree(dir, subm):
                 return self.dirlog(dir).read(subm)
             m = self._newmanifest()
             m.read(gettext, readsubtree)
             m.setnode(node)
             arraytext = None
         else:
             text = self.revision(node)
             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 sorted iterator
             work = heapq.merge([(x, False) for x in added],
                                [(x, True) for x in removed])
 
             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.
             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):
         # If the manifest is unchanged compared to one parent,
         # don't write a new revision
         if m.unmodifiedsince(m1) or m.unmodifiedsince(m2):
             return m.node()
         def writesubtree(subm, subp1, subp2):
             sublog = self.dirlog(subm.dir())
             sublog.add(subm, transaction, link, subp1, subp2, None, None)
         m.writesubtrees(m1, m2, writesubtree)
         text = m.dirtext(self._usemanifestv2)
         # Double-check whether contents are unchanged to one parent
         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
 
     def clearcaches(self):
         super(manifest, self).clearcaches()
         self._mancache.clear()
         self._dirlogcache = {'': self}

mercurial/verify.py

@@ -1,381 +1,408 b''
 # verify.py - repository integrity checking for Mercurial
 #
 # Copyright 2006, 2007 Matt Mackall <[e-mail]>
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 from __future__ import absolute_import
 
 import os
 
 from .i18n import _
 from .node import (
     nullid,
     short,
 )
 
 from . import (
     error,
     revlog,
     util,
 )
 
 def verify(repo):
     with repo.lock():
         return verifier(repo).verify()
 
 def _normpath(f):
     # under hg < 2.4, convert didn't sanitize paths properly, so a
     # converted repo may contain repeated slashes
     while '//' in f:
         f = f.replace('//', '/')
     return f
 
 def _validpath(repo, path):
     """Returns False if a path should NOT be treated as part of a repo.
 
     For all in-core cases, this returns True, as we have no way for a
     path to be mentioned in the history but not actually be
     relevant. For narrow clones, this is important because many
     filelogs will be missing, and changelog entries may mention
     modified files that are outside the narrow scope.
     """
     return True
 
 class verifier(object):
     def __init__(self, repo):
         self.repo = repo.unfiltered()
         self.ui = repo.ui
         self.badrevs = set()
         self.errors = 0
         self.warnings = 0
         self.havecl = len(repo.changelog) > 0
         self.havemf = len(repo.manifest) > 0
         self.revlogv1 = repo.changelog.version != revlog.REVLOGV0
         self.lrugetctx = util.lrucachefunc(repo.changectx)
         self.refersmf = False
         self.fncachewarned = False
 
     def warn(self, msg):
         self.ui.warn(msg + "\n")
         self.warnings += 1
 
     def err(self, linkrev, msg, filename=None):
         if linkrev is not None:
             self.badrevs.add(linkrev)
         else:
             linkrev = '?'
         msg = "%s: %s" % (linkrev, msg)
         if filename:
             msg = "%s@%s" % (filename, msg)
         self.ui.warn(" " + msg + "\n")
         self.errors += 1
 
     def exc(self, linkrev, msg, inst, filename=None):
         if not str(inst):
             inst = repr(inst)
         self.err(linkrev, "%s: %s" % (msg, inst), filename)
 
     def checklog(self, obj, name, linkrev):
         if not len(obj) and (self.havecl or self.havemf):
             self.err(linkrev, _("empty or missing %s") % name)
             return
 
         d = obj.checksize()
         if d[0]:
             self.err(None, _("data length off by %d bytes") % d[0], name)
         if d[1]:
             self.err(None, _("index contains %d extra bytes") % d[1], name)
 
         if obj.version != revlog.REVLOGV0:
             if not self.revlogv1:
                 self.warn(_("warning: `%s' uses revlog format 1") % name)
         elif self.revlogv1:
             self.warn(_("warning: `%s' uses revlog format 0") % name)
 
     def checkentry(self, obj, i, node, seen, linkrevs, f):
         lr = obj.linkrev(obj.rev(node))
         if lr < 0 or (self.havecl and lr not in linkrevs):
             if lr < 0 or lr >= len(self.repo.changelog):
                 msg = _("rev %d points to nonexistent changeset %d")
             else:
                 msg = _("rev %d points to unexpected changeset %d")
             self.err(None, msg % (i, lr), f)
             if linkrevs:
                 if f and len(linkrevs) > 1:
                     try:
                         # attempt to filter down to real linkrevs
                         linkrevs = [l for l in linkrevs
                                     if self.lrugetctx(l)[f].filenode() == node]
                     except Exception:
                         pass
                 self.warn(_(" (expected %s)") % " ".join(map(str, linkrevs)))
             lr = None # can't be trusted
 
         try:
             p1, p2 = obj.parents(node)
             if p1 not in seen and p1 != nullid:
                 self.err(lr, _("unknown parent 1 %s of %s") %
                     (short(p1), short(node)), f)
             if p2 not in seen and p2 != nullid:
                 self.err(lr, _("unknown parent 2 %s of %s") %
                     (short(p2), short(node)), f)
         except Exception as inst:
             self.exc(lr, _("checking parents of %s") % short(node), inst, f)
 
         if node in seen:
             self.err(lr, _("duplicate revision %d (%d)") % (i, seen[node]), f)
         seen[node] = i
         return lr
 
     def verify(self):
         repo = self.repo
 
         ui = repo.ui
 
         if not repo.url().startswith('file:'):
             raise error.Abort(_("cannot verify bundle or remote repos"))
 
         if os.path.exists(repo.sjoin("journal")):
             ui.warn(_("abandoned transaction found - run hg recover\n"))
 
         if ui.verbose or not self.revlogv1:
             ui.status(_("repository uses revlog format %d\n") %
                            (self.revlogv1 and 1 or 0))
 
         mflinkrevs, filelinkrevs = self._verifychangelog()
 
         filenodes = self._verifymanifest(mflinkrevs)
         del mflinkrevs
 
         self._crosscheckfiles(filelinkrevs, filenodes)
 
         totalfiles, filerevisions = self._verifyfiles(filenodes, filelinkrevs)
 
         ui.status(_("%d files, %d changesets, %d total revisions\n") %
                        (totalfiles, len(repo.changelog), filerevisions))
         if self.warnings:
             ui.warn(_("%d warnings encountered!\n") % self.warnings)
         if self.fncachewarned:
             ui.warn(_('hint: run "hg debugrebuildfncache" to recover from '
                       'corrupt fncache\n'))
         if self.errors:
             ui.warn(_("%d integrity errors encountered!\n") % self.errors)
             if self.badrevs:
                 ui.warn(_("(first damaged changeset appears to be %d)\n")
                         % min(self.badrevs))
             return 1
 
     def _verifychangelog(self):
         ui = self.ui
         repo = self.repo
         cl = repo.changelog
 
         ui.status(_("checking changesets\n"))
         mflinkrevs = {}
         filelinkrevs = {}
         seen = {}
         self.checklog(cl, "changelog", 0)
         total = len(repo)
         for i in repo:
             ui.progress(_('checking'), i, total=total, unit=_('changesets'))
             n = cl.node(i)
             self.checkentry(cl, i, n, seen, [i], "changelog")
 
             try:
                 changes = cl.read(n)
                 if changes[0] != nullid:
                     mflinkrevs.setdefault(changes[0], []).append(i)
                     self.refersmf = True
                 for f in changes[3]:
                     if _validpath(repo, f):
                         filelinkrevs.setdefault(_normpath(f), []).append(i)
             except Exception as inst:
                 self.refersmf = True
                 self.exc(i, _("unpacking changeset %s") % short(n), inst)
         ui.progress(_('checking'), None)
         return mflinkrevs, filelinkrevs
 
-    def _verifymanifest(self, mflinkrevs):
+    def _verifymanifest(self, mflinkrevs, dir=""):
         repo = self.repo
         ui = self.ui
-        mf = self.repo.manifest
+        mf = self.repo.manifest.dirlog(dir)
 
-        ui.status(_("checking manifests\n"))
+        if not dir:
+            self.ui.status(_("checking manifests\n"))
+
         filenodes = {}
+        subdirnodes = {}
         seen = {}
         label = "manifest"
+        if dir:
+            label = dir
         if self.refersmf:
             # Do not check manifest if there are only changelog entries with
             # null manifests.
             self.checklog(mf, label, 0)
         total = len(mf)
         for i in mf:
-            ui.progress(_('checking'), i, total=total, unit=_('manifests'))
+            if not dir:
+                ui.progress(_('checking'), i, total=total, unit=_('manifests'))
             n = mf.node(i)
             lr = self.checkentry(mf, i, n, seen, mflinkrevs.get(n, []), label)
             if n in mflinkrevs:
                 del mflinkrevs[n]
+            elif dir:
+                self.err(lr, _("%s not in parent-directory manifest") %
+                         short(n), label)
             else:
                 self.err(lr, _("%s not in changesets") % short(n), label)
 
             try:
-                for f, fn in mf.readdelta(n).iteritems():
+                for f, fn, fl in mf.readshallowdelta(n).iterentries():
                     if not f:
-                        self.err(lr, _("file without name in manifest"))
-                    elif f != "/dev/null": # ignore this in very old repos
-                        if _validpath(repo, f):
-                            filenodes.setdefault(
-                                _normpath(f), {}).setdefault(fn, lr)
+                        self.err(lr, _("entry without name in manifest"))
+                    elif f == "/dev/null":  # ignore this in very old repos
+                        continue
+                    fullpath = dir + _normpath(f)
+                    if not _validpath(repo, fullpath):
+                        continue
+                    if fl == 't':
+                        subdirnodes.setdefault(fullpath + '/', {}).setdefault(
+                            fn, []).append(lr)
+                    else:
+                        filenodes.setdefault(fullpath, {}).setdefault(fn, lr)
             except Exception as inst:
                 self.exc(lr, _("reading delta %s") % short(n), inst, label)
-        ui.progress(_('checking'), None)
+        if not dir:
+            ui.progress(_('checking'), None)
 
         if self.havemf:
             for c, m in sorted([(c, m) for m in mflinkrevs
                         for c in mflinkrevs[m]]):
-                self.err(c, _("changeset refers to unknown revision %s") %
-                         short(m), label)
+                if dir:
+                    self.err(c, _("parent-directory manifest refers to unknown "
+                                  "revision %s") % short(m), label)
+                else:
+                    self.err(c, _("changeset refers to unknown revision %s") %
+                             short(m), label)
+
+        if not dir and subdirnodes:
+            self.ui.status(_("checking directory manifests\n"))
+        for subdir, linkrevs in subdirnodes.iteritems():
+            subdirfilenodes = self._verifymanifest(linkrevs, subdir)
+            for f, onefilenodes in subdirfilenodes.iteritems():
+                filenodes.setdefault(f, {}).update(onefilenodes)
 
         return filenodes
 
     def _crosscheckfiles(self, filelinkrevs, filenodes):
         repo = self.repo
         ui = self.ui
         ui.status(_("crosschecking files in changesets and manifests\n"))
 
         total = len(filelinkrevs) + len(filenodes)
         count = 0
         if self.havemf:
             for f in sorted(filelinkrevs):
                 count += 1
                 ui.progress(_('crosschecking'), count, total=total)
                 if f not in filenodes:
                     lr = filelinkrevs[f][0]
                     self.err(lr, _("in changeset but not in manifest"), f)
 
         if self.havecl:
             for f in sorted(filenodes):
                 count += 1
                 ui.progress(_('crosschecking'), count, total=total)
                 if f not in filelinkrevs:
                     try:
                         fl = repo.file(f)
                         lr = min([fl.linkrev(fl.rev(n)) for n in filenodes[f]])
                     except Exception:
                         lr = None
                     self.err(lr, _("in manifest but not in changeset"), f)
 
         ui.progress(_('crosschecking'), None)
 
     def _verifyfiles(self, filenodes, filelinkrevs):
         repo = self.repo
         ui = self.ui
         lrugetctx = self.lrugetctx
         revlogv1 = self.revlogv1
         havemf = self.havemf
         ui.status(_("checking files\n"))
 
         storefiles = set()
         for f, f2, size in repo.store.datafiles():
             if not f:
                 self.err(None, _("cannot decode filename '%s'") % f2)
             elif (size > 0 or not revlogv1) and f.startswith('data/'):
                 storefiles.add(_normpath(f))
 
         files = sorted(set(filenodes) | set(filelinkrevs))
         total = len(files)
         revisions = 0
         for i, f in enumerate(files):
             ui.progress(_('checking'), i, item=f, total=total)
             try:
                 linkrevs = filelinkrevs[f]
             except KeyError:
                 # in manifest but not in changelog
                 linkrevs = []
 
             if linkrevs:
                 lr = linkrevs[0]
             else:
                 lr = None
 
             try:
                 fl = repo.file(f)
             except error.RevlogError as e:
                 self.err(lr, _("broken revlog! (%s)") % e, f)
                 continue
 
             for ff in fl.files():
                 try:
                     storefiles.remove(ff)
                 except KeyError:
                     self.warn(_(" warning: revlog '%s' not in fncache!") % ff)
                     self.fncachewarned = True
 
             self.checklog(fl, f, lr)
             seen = {}
             rp = None
             for i in fl:
                 revisions += 1
                 n = fl.node(i)
                 lr = self.checkentry(fl, i, n, seen, linkrevs, f)
                 if f in filenodes:
                     if havemf and n not in filenodes[f]:
                         self.err(lr, _("%s not in manifests") % (short(n)), f)
                     else:
                         del filenodes[f][n]
 
                 # verify contents
                 try:
                     l = len(fl.read(n))
                     rp = fl.renamed(n)
                     if l != fl.size(i):
                         if len(fl.revision(n)) != fl.size(i):
                             self.err(lr, _("unpacked size is %s, %s expected") %
                                      (l, fl.size(i)), f)
                 except error.CensoredNodeError:
                     # experimental config: censor.policy
                     if ui.config("censor", "policy", "abort") == "abort":
                         self.err(lr, _("censored file data"), f)
                 except Exception as inst:
                     self.exc(lr, _("unpacking %s") % short(n), inst, f)
 
                 # check renames
                 try:
                     if rp:
                         if lr is not None and ui.verbose:
                             ctx = lrugetctx(lr)
                             found = False
                             for pctx in ctx.parents():
                                 if rp[0] in pctx:
                                     found = True
                                     break
                             if not found:
                                 self.warn(_("warning: copy source of '%s' not"
                                             " in parents of %s") % (f, ctx))
                         fl2 = repo.file(rp[0])
                         if not len(fl2):
                             self.err(lr, _("empty or missing copy source "
                                      "revlog %s:%s") % (rp[0], short(rp[1])), f)
                         elif rp[1] == nullid:
                             ui.note(_("warning: %s@%s: copy source"
                                       " revision is nullid %s:%s\n")
                                 % (f, lr, rp[0], short(rp[1])))
                         else:
                             fl2.rev(rp[1])
                 except Exception as inst:
                     self.exc(lr, _("checking rename of %s") % short(n), inst, f)
 
             # cross-check
             if f in filenodes:
                 fns = [(lr, n) for n, lr in filenodes[f].iteritems()]
                 for lr, node in sorted(fns):
                     self.err(lr, _("manifest refers to unknown revision %s") %
                              short(node), f)
         ui.progress(_('checking'), None)
 
         for f in storefiles:
             self.warn(_("warning: orphan revlog '%s'") % f)
 
         return len(files), revisions

tests/test-treemanifest.t

@@ -1,666 +1,724 b''
 #require killdaemons
 
   $ cat << EOF >> $HGRCPATH
   > [format]
   > usegeneraldelta=yes
   > [ui]
   > ssh=python "$TESTDIR/dummyssh"
   > EOF
 
 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\x008b3ffd73f901e83304c83d33132c8e774ceac44et (esc)
   e\x00b8e02f6433738021a065f94175c7cd23db5f05be (esc)
   $ hg debugdata --dir dir1 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 (glob)
   dir1/b (glob)
   dir1/dir1/a (glob)
   dir1/dir1/b (glob)
   dir1/dir2/a (glob)
   dir1/dir2/b (glob)
   e
 
 Revision is not created for unchanged directory
 
   $ mkdir dir2
   $ echo 3 > dir2/a
   $ hg add dir2
   adding dir2/a (glob)
   $ hg debugindex --dir dir1 > before
   $ hg ci -qm 'add dir2'
   $ hg debugindex --dir dir1 > after
   $ diff before after
   $ rm before after
 
 Removing directory does not create an revlog entry
 
   $ hg rm dir1/dir1
   removing dir1/dir1/a (glob)
   removing dir1/dir1/b (glob)
   $ hg debugindex --dir dir1/dir1 > before
   $ hg ci -qm 'remove dir1/dir1'
   $ hg debugindex --dir dir1/dir1 > after
   $ diff before after
   $ rm before after
 
 Check that hg files (calls treemanifest.walk()) works
 without loading all directory revlogs
 
   $ hg co 'desc("add dir2")'
   2 files updated, 0 files merged, 0 files removed, 0 files unresolved
   $ mv .hg/store/meta/dir2 .hg/store/meta/dir2-backup
   $ hg files -r . dir1
   dir1/a (glob)
   dir1/b (glob)
   dir1/dir1/a (glob)
   dir1/dir1/b (glob)
   dir1/dir2/a (glob)
   dir1/dir2/b (glob)
 
 Check that status between revisions works (calls treemanifest.matches())
 without loading all directory revlogs
 
   $ hg status --rev 'desc("add dir1")' --rev . dir1
   A dir1/dir1/a
   A dir1/dir1/b
   A dir1/dir2/a
   A dir1/dir2/b
   $ mv .hg/store/meta/dir2-backup .hg/store/meta/dir2
 
 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 --dir dir1
      rev    offset  length  delta linkrev nodeid       p1           p2
        0         0      54     -1       1 8b3ffd73f901 000000000000 000000000000
        1        54      68      0       2 68e9d057c5a8 8b3ffd73f901 000000000000
        2       122      12      1       4 4698198d2624 68e9d057c5a8 000000000000
        3       134      55      1       5 44844058ccce 68e9d057c5a8 000000000000
        4       189      55      1       6 bf3d9b744927 68e9d057c5a8 000000000000
        5       244      55      4       7 dde7c0af2a03 bf3d9b744927 44844058ccce
 
 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 --dir dir2 > 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 --dir dir2 > 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 --dir dir1 > 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 --dir dir1 > 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'
 
   $ hg serve -p $HGPORT -d --pid-file=hg.pid --errorlog=errors.log
   $ cat hg.pid >> $DAEMON_PIDS
 
 Create clone with tree manifests enabled
 
   $ cd ..
   $ hg clone --config experimental.treemanifest=1 \
   >   http://localhost:$HGPORT repo-mixed -r 1
   adding changesets
   adding manifests
   adding file changes
   added 2 changesets with 14 changes to 11 files
   updating to branch default
   11 files updated, 0 files merged, 0 files removed, 0 files unresolved
   $ cd repo-mixed
   $ test -d .hg/store/meta
   [1]
   $ grep treemanifest .hg/requires
   treemanifest
 
 Should be possible to push updates from flat to tree manifest repo
 
   $ hg -R ../repo-flat push ssh://user@dummy/repo-mixed
   pushing to ssh://user@dummy/repo-mixed
   searching for changes
   remote: adding changesets
   remote: adding manifests
   remote: adding file changes
   remote: added 2 changesets with 3 changes to 3 files
 
 Commit should store revlog per directory
 
   $ hg co 1
   0 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  delta linkrev nodeid       p1           p2
        0         0      80     -1       0 40536115ed9e 000000000000 000000000000
        1        80      83      0       1 f3376063c255 40536115ed9e 000000000000
        2       163      89      0       2 5d9b9da231a2 40536115ed9e 000000000000
        3       252      83      2       3 d17d663cbd8a 5d9b9da231a2 f3376063c255
        4       335     124      1       4 51e32a8c60ee f3376063c255 000000000000
        5       459     126      2       5 cc5baa78b230 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 --dir dir1
      rev    offset  length  delta linkrev nodeid       p1           p2
        0         0     127     -1       4 064927a0648a 000000000000 000000000000
        1       127     111      0       5 25ecb8cb8618 000000000000 000000000000
   $ echo 2 > dir1/a
   $ hg --config experimental.treemanifest=False ci -qm 'modify dir1/a'
   $ hg debugindex --dir dir1
      rev    offset  length  delta linkrev nodeid       p1           p2
        0         0     127     -1       4 064927a0648a 000000000000 000000000000
        1       127     111      0       5 25ecb8cb8618 000000000000 000000000000
        2       238      55      1       6 5b16163a30c6 25ecb8cb8618 000000000000
 
 Stripping and recovering changes should work
 
   $ hg st --change tip
   M dir1/a
   $ hg --config extensions.strip= strip tip
   1 files updated, 0 files merged, 0 files removed, 0 files unresolved
   saved backup bundle to $TESTTMP/repo-mixed/.hg/strip-backup/51cfd7b1e13b-78a2f3ed-backup.hg (glob)
   $ hg unbundle -q .hg/strip-backup/*
   $ hg st --change tip
   M dir1/a
 
 Shelving and unshelving should work
 
   $ echo foo >> dir1/a
   $ hg --config extensions.shelve= shelve
   shelved as default
   1 files updated, 0 files merged, 0 files removed, 0 files unresolved
   $ hg --config extensions.shelve= unshelve
   unshelving change 'default'
   $ hg diff --nodates
   diff -r 708a273da119 dir1/a
   --- a/dir1/a
   +++ b/dir1/a
   @@ -1,1 +1,2 @@
    1
   +foo
 
 Pushing from treemanifest repo to an empty repo makes that a treemanifest repo
 
   $ cd ..
   $ hg init empty-repo
   $ cat << EOF >> empty-repo/.hg/hgrc
   > [experimental]
   > changegroup3=yes
   > EOF
   $ grep treemanifest empty-repo/.hg/requires
   [1]
   $ hg push -R repo -r 0 empty-repo
   pushing to empty-repo
   searching for changes
   adding changesets
   adding manifests
   adding file changes
   added 1 changesets with 2 changes to 2 files
   $ grep treemanifest empty-repo/.hg/requires
   treemanifest
 
 Pushing to an empty repo works
 
   $ hg --config experimental.treemanifest=1 init clone
   $ grep treemanifest clone/.hg/requires
   treemanifest
   $ hg push -R repo clone
   pushing to clone
   searching for changes
   adding changesets
   adding manifests
   adding file changes
   added 11 changesets with 15 changes to 10 files (+3 heads)
   $ grep treemanifest clone/.hg/requires
   treemanifest
 
 Create deeper repo with tree manifests.
 
   $ hg --config experimental.treemanifest=True init deeprepo
   $ cd deeprepo
 
   $ mkdir .A
   $ mkdir b
   $ mkdir b/bar
   $ mkdir b/bar/orange
   $ mkdir b/bar/orange/fly
   $ mkdir b/foo
   $ mkdir b/foo/apple
   $ mkdir b/foo/apple/bees
 
   $ touch .A/one.txt
   $ touch .A/two.txt
   $ touch b/bar/fruits.txt
   $ touch b/bar/orange/fly/gnat.py
   $ touch b/bar/orange/fly/housefly.txt
   $ touch b/foo/apple/bees/flower.py
   $ touch c.txt
   $ touch d.py
 
   $ hg ci -Aqm 'initial'
 
 We'll see that visitdir works by removing some treemanifest revlogs and running
 the files command with various parameters.
 
 Test files from the root.
 
   $ hg files -r .
   .A/one.txt (glob)
   .A/two.txt (glob)
   b/bar/fruits.txt (glob)
   b/bar/orange/fly/gnat.py (glob)
   b/bar/orange/fly/housefly.txt (glob)
   b/foo/apple/bees/flower.py (glob)
   c.txt
   d.py
 
 Excludes with a glob should not exclude everything from the glob's root
 
   $ hg files -r . -X 'b/fo?' b
   b/bar/fruits.txt (glob)
   b/bar/orange/fly/gnat.py (glob)
   b/bar/orange/fly/housefly.txt (glob)
   $ cp -r .hg/store .hg/store-copy
 
 Test files for a subdirectory.
 
   $ rm -r .hg/store/meta/~2e_a
   $ hg files -r . b
   b/bar/fruits.txt (glob)
   b/bar/orange/fly/gnat.py (glob)
   b/bar/orange/fly/housefly.txt (glob)
   b/foo/apple/bees/flower.py (glob)
   $ cp -r .hg/store-copy/* .hg/store
 
 Test files with just includes and excludes.
 
   $ rm -r .hg/store/meta/~2e_a
   $ rm -r .hg/store/meta/b/bar/orange/fly
   $ rm -r .hg/store/meta/b/foo/apple/bees
   $ hg files -r . -I path:b/bar -X path:b/bar/orange/fly -I path:b/foo -X path:b/foo/apple/bees
   b/bar/fruits.txt (glob)
   $ cp -r .hg/store-copy/* .hg/store
 
 Test files for a subdirectory, excluding a directory within it.
 
   $ rm -r .hg/store/meta/~2e_a
   $ rm -r .hg/store/meta/b/foo
   $ hg files -r . -X path:b/foo b
   b/bar/fruits.txt (glob)
   b/bar/orange/fly/gnat.py (glob)
   b/bar/orange/fly/housefly.txt (glob)
   $ cp -r .hg/store-copy/* .hg/store
 
 Test files for a sub directory, including only a directory within it, and
 including an unrelated directory.
 
   $ rm -r .hg/store/meta/~2e_a
   $ rm -r .hg/store/meta/b/foo
   $ hg files -r . -I path:b/bar/orange -I path:a b
   b/bar/orange/fly/gnat.py (glob)
   b/bar/orange/fly/housefly.txt (glob)
   $ cp -r .hg/store-copy/* .hg/store
 
 Test files for a pattern, including a directory, and excluding a directory
 within that.
 
   $ rm -r .hg/store/meta/~2e_a
   $ rm -r .hg/store/meta/b/foo
   $ rm -r .hg/store/meta/b/bar/orange
   $ hg files -r . glob:**.txt -I path:b/bar -X path:b/bar/orange
   b/bar/fruits.txt (glob)
   $ cp -r .hg/store-copy/* .hg/store
 
 Add some more changes to the deep repo
   $ echo narf >> b/bar/fruits.txt
   $ hg ci -m narf
   $ echo troz >> b/bar/orange/fly/gnat.py
   $ hg ci -m troz
 
 Verify works
   $ hg verify
   checking changesets
   checking manifests
+  checking directory manifests
   crosschecking files in changesets and manifests
   checking files
   8 files, 3 changesets, 10 total revisions
 
 Dirlogs are included in fncache
   $ grep meta/.A/00manifest.i .hg/store/fncache
   meta/.A/00manifest.i
 
 Rebuilt fncache includes dirlogs
   $ rm .hg/store/fncache
   $ hg debugrebuildfncache
   adding data/.A/one.txt.i
   adding data/.A/two.txt.i
   adding data/b/bar/fruits.txt.i
   adding data/b/bar/orange/fly/gnat.py.i
   adding data/b/bar/orange/fly/housefly.txt.i
   adding data/b/foo/apple/bees/flower.py.i
   adding data/c.txt.i
   adding data/d.py.i
   adding meta/.A/00manifest.i
   adding meta/b/00manifest.i
   adding meta/b/bar/00manifest.i
   adding meta/b/bar/orange/00manifest.i
   adding meta/b/bar/orange/fly/00manifest.i
   adding meta/b/foo/00manifest.i
   adding meta/b/foo/apple/00manifest.i
   adding meta/b/foo/apple/bees/00manifest.i
   16 items added, 0 removed from fncache
 
 Finish first server
   $ killdaemons.py
 
+Back up the recently added revlogs
+  $ cp -r .hg/store .hg/store-newcopy
+
+Verify reports missing dirlog
+  $ rm .hg/store/meta/b/00manifest.*
+  $ hg verify
+  checking changesets
+  checking manifests
+  checking directory manifests
+   0: empty or missing b/
+   b/@0: parent-directory manifest refers to unknown revision 67688a370455
+   b/@1: parent-directory manifest refers to unknown revision f38e85d334c5
+   b/@2: parent-directory manifest refers to unknown revision 99c9792fd4b0
+  crosschecking files in changesets and manifests
+   b/bar/fruits.txt@0: in changeset but not in manifest
+   b/bar/orange/fly/gnat.py@0: in changeset but not in manifest
+   b/bar/orange/fly/housefly.txt@0: in changeset but not in manifest
+   b/foo/apple/bees/flower.py@0: in changeset but not in manifest
+  checking files
+  8 files, 3 changesets, 10 total revisions
+  8 integrity errors encountered!
+  (first damaged changeset appears to be 0)
+  [1]
+  $ cp -rT .hg/store-newcopy .hg/store
+
+Verify reports missing dirlog entry
+  $ mv -f .hg/store-copy/meta/b/00manifest.* .hg/store/meta/b/
+  $ hg verify
+  checking changesets
+  checking manifests
+  checking directory manifests
+   b/@1: parent-directory manifest refers to unknown revision f38e85d334c5
+   b/@2: parent-directory manifest refers to unknown revision 99c9792fd4b0
+   b/bar/@?: rev 1 points to unexpected changeset 1
+   b/bar/@?: 5e03c4ee5e4a not in parent-directory manifest
+   b/bar/@?: rev 2 points to unexpected changeset 2
+   b/bar/@?: 1b16940d66d6 not in parent-directory manifest
+   b/bar/orange/@?: rev 1 points to unexpected changeset 2
+   (expected None)
+   b/bar/orange/fly/@?: rev 1 points to unexpected changeset 2
+   (expected None)
+  crosschecking files in changesets and manifests
+  checking files
+  8 files, 3 changesets, 10 total revisions
+  2 warnings encountered!
+  8 integrity errors encountered!
+  (first damaged changeset appears to be 1)
+  [1]
+  $ cp -rT .hg/store-newcopy .hg/store
+
 Test cloning a treemanifest repo over http.
   $ hg serve -p $HGPORT -d --pid-file=hg.pid --errorlog=errors.log
   $ cat hg.pid >> $DAEMON_PIDS
   $ cd ..
 We can clone even with the knob turned off and we'll get a treemanifest repo.
   $ hg clone --config experimental.treemanifest=False \
   >   --config experimental.changegroup3=True \
   >   http://localhost:$HGPORT deepclone
   requesting all changes
   adding changesets
   adding manifests
   adding file changes
   added 3 changesets with 10 changes to 8 files
   updating to branch default
   8 files updated, 0 files merged, 0 files removed, 0 files unresolved
 No server errors.
   $ cat deeprepo/errors.log
 requires got updated to include treemanifest
   $ cat deepclone/.hg/requires | grep treemanifest
   treemanifest
 Tree manifest revlogs exist.
   $ find deepclone/.hg/store/meta | sort
   deepclone/.hg/store/meta
   deepclone/.hg/store/meta/b
   deepclone/.hg/store/meta/b/00manifest.i
   deepclone/.hg/store/meta/b/bar
   deepclone/.hg/store/meta/b/bar/00manifest.i
   deepclone/.hg/store/meta/b/bar/orange
   deepclone/.hg/store/meta/b/bar/orange/00manifest.i
   deepclone/.hg/store/meta/b/bar/orange/fly
   deepclone/.hg/store/meta/b/bar/orange/fly/00manifest.i
   deepclone/.hg/store/meta/b/foo
   deepclone/.hg/store/meta/b/foo/00manifest.i
   deepclone/.hg/store/meta/b/foo/apple
   deepclone/.hg/store/meta/b/foo/apple/00manifest.i
   deepclone/.hg/store/meta/b/foo/apple/bees
   deepclone/.hg/store/meta/b/foo/apple/bees/00manifest.i
   deepclone/.hg/store/meta/~2e_a
   deepclone/.hg/store/meta/~2e_a/00manifest.i
 Verify passes.
   $ cd deepclone
   $ hg verify
   checking changesets
   checking manifests
+  checking directory manifests
   crosschecking files in changesets and manifests
   checking files
   8 files, 3 changesets, 10 total revisions
   $ cd ..
 
 Create clones using old repo formats to use in later tests
   $ hg clone --config format.usestore=False \
   >   --config experimental.changegroup3=True \
   >   http://localhost:$HGPORT deeprepo-basicstore
   requesting all changes
   adding changesets
   adding manifests
   adding file changes
   added 3 changesets with 10 changes to 8 files
   updating to branch default
   8 files updated, 0 files merged, 0 files removed, 0 files unresolved
   $ cd deeprepo-basicstore
   $ grep store .hg/requires
   [1]
   $ hg serve -p $HGPORT1 -d --pid-file=hg.pid --errorlog=errors.log
   $ cat hg.pid >> $DAEMON_PIDS
   $ cd ..
   $ hg clone --config format.usefncache=False \
   >   --config experimental.changegroup3=True \
   >   http://localhost:$HGPORT deeprepo-encodedstore
   requesting all changes
   adding changesets
   adding manifests
   adding file changes
   added 3 changesets with 10 changes to 8 files
   updating to branch default
   8 files updated, 0 files merged, 0 files removed, 0 files unresolved
   $ cd deeprepo-encodedstore
   $ grep fncache .hg/requires
   [1]
   $ hg serve -p $HGPORT2 -d --pid-file=hg.pid --errorlog=errors.log
   $ cat hg.pid >> $DAEMON_PIDS
   $ cd ..
 
 Local clone with basicstore
   $ hg clone -U deeprepo-basicstore local-clone-basicstore
   $ hg -R local-clone-basicstore verify
   checking changesets
   checking manifests
+  checking directory manifests
   crosschecking files in changesets and manifests
   checking files
   8 files, 3 changesets, 10 total revisions
 
 Local clone with encodedstore
   $ hg clone -U deeprepo-encodedstore local-clone-encodedstore
   $ hg -R local-clone-encodedstore verify
   checking changesets
   checking manifests
+  checking directory manifests
   crosschecking files in changesets and manifests
   checking files
   8 files, 3 changesets, 10 total revisions
 
 Local clone with fncachestore
   $ hg clone -U deeprepo local-clone-fncachestore
   $ hg -R local-clone-fncachestore verify
   checking changesets
   checking manifests
+  checking directory manifests
   crosschecking files in changesets and manifests
   checking files
   8 files, 3 changesets, 10 total revisions
 
 Stream clone with basicstore
   $ hg clone --config experimental.changegroup3=True --uncompressed -U \
   >   http://localhost:$HGPORT1 stream-clone-basicstore
   streaming all changes
   18 files to transfer, * of data (glob)
   transferred * in * seconds (*) (glob)
   searching for changes
   no changes found
   $ hg -R stream-clone-basicstore verify
   checking changesets
   checking manifests
+  checking directory manifests
   crosschecking files in changesets and manifests
   checking files
   8 files, 3 changesets, 10 total revisions
 
 Stream clone with encodedstore
   $ hg clone --config experimental.changegroup3=True --uncompressed -U \
   >   http://localhost:$HGPORT2 stream-clone-encodedstore
   streaming all changes
   18 files to transfer, * of data (glob)
   transferred * in * seconds (*) (glob)
   searching for changes
   no changes found
   $ hg -R stream-clone-encodedstore verify
   checking changesets
   checking manifests
+  checking directory manifests
   crosschecking files in changesets and manifests
   checking files
   8 files, 3 changesets, 10 total revisions
 
 Stream clone with fncachestore
   $ hg clone --config experimental.changegroup3=True --uncompressed -U \
   >   http://localhost:$HGPORT stream-clone-fncachestore
   streaming all changes
   18 files to transfer, * of data (glob)
   transferred * in * seconds (*) (glob)
   searching for changes
   no changes found
   $ hg -R stream-clone-fncachestore verify
   checking changesets
   checking manifests
+  checking directory manifests
   crosschecking files in changesets and manifests
   checking files
   8 files, 3 changesets, 10 total revisions
 
 Packed bundle
   $ hg -R deeprepo debugcreatestreamclonebundle repo-packed.hg
   writing 3349 bytes for 18 files
   bundle requirements: generaldelta, revlogv1, treemanifest
   $ hg debugbundle --spec repo-packed.hg
   none-packed1;requirements%3Dgeneraldelta%2Crevlogv1%2Ctreemanifest