revlog: fix caching of buffer objects

mercurial/changelog.py

@@ -1,198 +1,198 b''
 # changelog.py - changelog 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, incorporated herein by reference.
 
 from revlog import *
 from i18n import _
 import os, time, util
 
 def _string_escape(text):
     """
     >>> d = {'nl': chr(10), 'bs': chr(92), 'cr': chr(13), 'nul': chr(0)}
     >>> s = "ab%(nl)scd%(bs)s%(bs)sn%(nul)sab%(cr)scd%(bs)s%(nl)s" % d
     >>> s
     'ab\\ncd\\\\\\\\n\\x00ab\\rcd\\\\\\n'
     >>> res = _string_escape(s)
     >>> s == _string_unescape(res)
     True
     """
     # subset of the string_escape codec
     text = text.replace('\\', '\\\\').replace('\n', '\\n').replace('\r', '\\r')
     return text.replace('\0', '\\0')
 
 def _string_unescape(text):
     return text.decode('string_escape')
 
 class appender:
     '''the changelog index must be update last on disk, so we use this class
     to delay writes to it'''
     def __init__(self, fp, buf):
         self.data = buf
         self.fp = fp
         self.offset = fp.tell()
         self.size = util.fstat(fp).st_size
 
     def end(self):
         return self.size + len("".join(self.data))
     def tell(self):
         return self.offset
     def flush(self):
         pass
     def close(self):
         self.fp.close()
 
     def seek(self, offset, whence=0):
         '''virtual file offset spans real file and data'''
         if whence == 0:
             self.offset = offset
         elif whence == 1:
             self.offset += offset
         elif whence == 2:
             self.offset = self.end() + offset
         if self.offset < self.size:
             self.fp.seek(self.offset)
 
     def read(self, count=-1):
         '''only trick here is reads that span real file and data'''
         ret = ""
         if self.offset < self.size:
             s = self.fp.read(count)
             ret = s
             self.offset += len(s)
             if count > 0:
                 count -= len(s)
         if count != 0:
             doff = self.offset - self.size
             self.data.insert(0, "".join(self.data))
             del self.data[1:]
             s = self.data[0][doff:doff+count]
             self.offset += len(s)
             ret += s
         return ret
 
     def write(self, s):
-        self.data.append(s)
+        self.data.append(str(s))
         self.offset += len(s)
 
 class changelog(revlog):
     def __init__(self, opener):
         revlog.__init__(self, opener, "00changelog.i")
 
     def delayupdate(self):
         "delay visibility of index updates to other readers"
         self._realopener = self.opener
         self.opener = self._delayopener
         self._delaycount = self.count()
         self._delaybuf = []
         self._delayname = None
 
     def finalize(self, tr):
         "finalize index updates"
         self.opener = self._realopener
         # move redirected index data back into place
         if self._delayname:
             util.rename(self._delayname + ".a", self._delayname)
         elif self._delaybuf:
             fp = self.opener(self.indexfile, 'a')
             fp.write("".join(self._delaybuf))
             fp.close()
             del self._delaybuf
         # split when we're done
         self.checkinlinesize(tr)
 
     def _delayopener(self, name, mode='r'):
         fp = self._realopener(name, mode)
         # only divert the index
         if not name == self.indexfile:
             return fp
         # if we're doing an initial clone, divert to another file
         if self._delaycount == 0:
             self._delayname = fp.name
             return self._realopener(name + ".a", mode)
         # otherwise, divert to memory
         return appender(fp, self._delaybuf)
 
     def checkinlinesize(self, tr, fp=None):
         if self.opener == self._delayopener:
             return
         return revlog.checkinlinesize(self, tr, fp)
 
     def decode_extra(self, text):
         extra = {}
         for l in text.split('\0'):
             if not l:
                 continue
             k, v = _string_unescape(l).split(':', 1)
             extra[k] = v
         return extra
 
     def encode_extra(self, d):
         # keys must be sorted to produce a deterministic changelog entry
         keys = d.keys()
         keys.sort()
         items = [_string_escape('%s:%s' % (k, d[k])) for k in keys]
         return "\0".join(items)
 
     def extract(self, text):
         """
         format used:
         nodeid\n        : manifest node in ascii
         user\n          : user, no \n or \r allowed
         time tz extra\n : date (time is int or float, timezone is int)
                         : extra is metadatas, encoded and separated by '\0'
                         : older versions ignore it
         files\n\n       : files modified by the cset, no \n or \r allowed
         (.*)            : comment (free text, ideally utf-8)
 
         changelog v0 doesn't use extra
         """
         if not text:
             return (nullid, "", (0, 0), [], "", {'branch': 'default'})
         last = text.index("\n\n")
         desc = util.tolocal(text[last + 2:])
         l = text[:last].split('\n')
         manifest = bin(l[0])
         user = util.tolocal(l[1])
 
         extra_data = l[2].split(' ', 2)
         if len(extra_data) != 3:
             time = float(extra_data.pop(0))
             try:
                 # various tools did silly things with the time zone field.
                 timezone = int(extra_data[0])
             except:
                 timezone = 0
             extra = {}
         else:
             time, timezone, extra = extra_data
             time, timezone = float(time), int(timezone)
             extra = self.decode_extra(extra)
         if not extra.get('branch'):
             extra['branch'] = 'default'
         files = l[3:]
         return (manifest, user, (time, timezone), files, desc, extra)
 
     def read(self, node):
         return self.extract(self.revision(node))
 
     def add(self, manifest, list, desc, transaction, p1=None, p2=None,
                   user=None, date=None, extra={}):
 
         user, desc = util.fromlocal(user), util.fromlocal(desc)
 
         if date:
             parseddate = "%d %d" % util.parsedate(date)
         else:
             parseddate = "%d %d" % util.makedate()
         if extra and extra.get("branch") in ("default", ""):
             del extra["branch"]
         if extra:
             extra = self.encode_extra(extra)
             parseddate = "%s %s" % (parseddate, extra)
         list.sort()
         l = [hex(manifest), user, parseddate] + list + ["", desc]
         text = "\n".join(l)
         return self.addrevision(text, transaction, self.count(), p1, p2)

mercurial/revlog.py

@@ -1,1291 +1,1291 b''
 """
 revlog.py - storage back-end for mercurial
 
 This provides efficient delta storage with O(1) retrieve and append
 and O(changes) merge between branches
 
 Copyright 2005-2007 Matt Mackall <[e-mail]>
 
 This software may be used and distributed according to the terms
 of the GNU General Public License, incorporated herein by reference.
 """
 
 from node import *
 from i18n import _
 import binascii, changegroup, errno, ancestor, mdiff, os
 import sha, struct, util, zlib
 
 _pack = struct.pack
 _unpack = struct.unpack
 _compress = zlib.compress
 _decompress = zlib.decompress
 _sha = sha.new
 
 # revlog flags
 REVLOGV0 = 0
 REVLOGNG = 1
 REVLOGNGINLINEDATA = (1 << 16)
 REVLOG_DEFAULT_FLAGS = REVLOGNGINLINEDATA
 REVLOG_DEFAULT_FORMAT = REVLOGNG
 REVLOG_DEFAULT_VERSION = REVLOG_DEFAULT_FORMAT | REVLOG_DEFAULT_FLAGS
 
 class RevlogError(Exception):
     pass
 class LookupError(RevlogError):
     pass
 
 def getoffset(q):
     return int(q >> 16)
 
 def gettype(q):
     return int(q & 0xFFFF)
 
 def offset_type(offset, type):
     return long(long(offset) << 16 | type)
 
 def hash(text, p1, p2):
     """generate a hash from the given text and its parent hashes
 
     This hash combines both the current file contents and its history
     in a manner that makes it easy to distinguish nodes with the same
     content in the revision graph.
     """
     l = [p1, p2]
     l.sort()
     s = _sha(l[0])
     s.update(l[1])
     s.update(text)
     return s.digest()
 
 def compress(text):
     """ generate a possibly-compressed representation of text """
     if not text:
         return ("", text)
     if len(text) < 44:
         if text[0] == '\0':
             return ("", text)
         return ('u', text)
     bin = _compress(text)
     if len(bin) > len(text):
         if text[0] == '\0':
             return ("", text)
         return ('u', text)
     return ("", bin)
 
 def decompress(bin):
     """ decompress the given input """
     if not bin:
         return bin
     t = bin[0]
     if t == '\0':
         return bin
     if t == 'x':
         return _decompress(bin)
     if t == 'u':
         return bin[1:]
     raise RevlogError(_("unknown compression type %r") % t)
 
 class lazyparser(object):
     """
     this class avoids the need to parse the entirety of large indices
     """
 
     # lazyparser is not safe to use on windows if win32 extensions not
     # available. it keeps file handle open, which make it not possible
     # to break hardlinks on local cloned repos.
     safe_to_use = os.name != 'nt' or (not util.is_win_9x() and
                                       hasattr(util, 'win32api'))
 
     def __init__(self, dataf, size):
         self.dataf = dataf
         self.s = struct.calcsize(indexformatng)
         self.datasize = size
         self.l = size/self.s
         self.index = [None] * self.l
         self.map = {nullid: nullrev}
         self.allmap = 0
         self.all = 0
         self.mapfind_count = 0
 
     def loadmap(self):
         """
         during a commit, we need to make sure the rev being added is
         not a duplicate.  This requires loading the entire index,
         which is fairly slow.  loadmap can load up just the node map,
         which takes much less time.
         """
         if self.allmap:
             return
         end = self.datasize
         self.allmap = 1
         cur = 0
         count = 0
         blocksize = self.s * 256
         self.dataf.seek(0)
         while cur < end:
             data = self.dataf.read(blocksize)
             off = 0
             for x in xrange(256):
                 n = data[off + ngshaoffset:off + ngshaoffset + 20]
                 self.map[n] = count
                 count += 1
                 if count >= self.l:
                     break
                 off += self.s
             cur += blocksize
 
     def loadblock(self, blockstart, blocksize, data=None):
         if self.all:
             return
         if data is None:
             self.dataf.seek(blockstart)
             if blockstart + blocksize > self.datasize:
                 # the revlog may have grown since we've started running,
                 # but we don't have space in self.index for more entries.
                 # limit blocksize so that we don't get too much data.
                 blocksize = max(self.datasize - blockstart, 0)
             data = self.dataf.read(blocksize)
         lend = len(data) / self.s
         i = blockstart / self.s
         off = 0
         # lazyindex supports __delitem__
         if lend > len(self.index) - i:
             lend = len(self.index) - i
         for x in xrange(lend):
             if self.index[i + x] == None:
                 b = data[off : off + self.s]
                 self.index[i + x] = b
                 n = b[ngshaoffset:ngshaoffset + 20]
                 self.map[n] = i + x
             off += self.s
 
     def findnode(self, node):
         """search backwards through the index file for a specific node"""
         if self.allmap:
             return None
 
         # hg log will cause many many searches for the manifest
         # nodes.  After we get called a few times, just load the whole
         # thing.
         if self.mapfind_count > 8:
             self.loadmap()
             if node in self.map:
                 return node
             return None
         self.mapfind_count += 1
         last = self.l - 1
         while self.index[last] != None:
             if last == 0:
                 self.all = 1
                 self.allmap = 1
                 return None
             last -= 1
         end = (last + 1) * self.s
         blocksize = self.s * 256
         while end >= 0:
             start = max(end - blocksize, 0)
             self.dataf.seek(start)
             data = self.dataf.read(end - start)
             findend = end - start
             while True:
                 # we're searching backwards, so we have to make sure
                 # we don't find a changeset where this node is a parent
                 off = data.find(node, 0, findend)
                 findend = off
                 if off >= 0:
                     i = off / self.s
                     off = i * self.s
                     n = data[off + ngshaoffset:off + ngshaoffset + 20]
                     if n == node:
                         self.map[n] = i + start / self.s
                         return node
                 else:
                     break
             end -= blocksize
         return None
 
     def loadindex(self, i=None, end=None):
         if self.all:
             return
         all = False
         if i == None:
             blockstart = 0
             blocksize = (65536 / self.s) * self.s
             end = self.datasize
             all = True
         else:
             if end:
                 blockstart = i * self.s
                 end = end * self.s
                 blocksize = end - blockstart
             else:
                 blockstart = (i & ~1023) * self.s
                 blocksize = self.s * 1024
                 end = blockstart + blocksize
         while blockstart < end:
             self.loadblock(blockstart, blocksize)
             blockstart += blocksize
         if all:
             self.all = True
 
 class lazyindex(object):
     """a lazy version of the index array"""
     def __init__(self, parser):
         self.p = parser
     def __len__(self):
         return len(self.p.index)
     def load(self, pos):
         if pos < 0:
             pos += len(self.p.index)
         self.p.loadindex(pos)
         return self.p.index[pos]
     def __getitem__(self, pos):
         return _unpack(indexformatng, self.p.index[pos] or self.load(pos))
     def __setitem__(self, pos, item):
         self.p.index[pos] = _pack(indexformatng, *item)
     def __delitem__(self, pos):
         del self.p.index[pos]
     def insert(self, pos, e):
         self.p.index.insert(pos, _pack(indexformatng, *e))
     def append(self, e):
         self.p.index.append(_pack(indexformatng, *e))
 
 class lazymap(object):
     """a lazy version of the node map"""
     def __init__(self, parser):
         self.p = parser
     def load(self, key):
         n = self.p.findnode(key)
         if n == None:
             raise KeyError(key)
     def __contains__(self, key):
         if key in self.p.map:
             return True
         self.p.loadmap()
         return key in self.p.map
     def __iter__(self):
         yield nullid
         for i in xrange(self.p.l):
             ret = self.p.index[i]
             if not ret:
                 self.p.loadindex(i)
                 ret = self.p.index[i]
             if isinstance(ret, str):
                 ret = _unpack(indexformatng, ret)
             yield ret[7]
     def __getitem__(self, key):
         try:
             return self.p.map[key]
         except KeyError:
             try:
                 self.load(key)
                 return self.p.map[key]
             except KeyError:
                 raise KeyError("node " + hex(key))
     def __setitem__(self, key, val):
         self.p.map[key] = val
     def __delitem__(self, key):
         del self.p.map[key]
 
 indexformatv0 = ">4l20s20s20s"
 v0shaoffset = 56
 
 class revlogoldio(object):
     def __init__(self):
         self.size = struct.calcsize(indexformatv0)
 
     def parseindex(self, fp, inline):
         s = self.size
         index = []
         nodemap =  {nullid: nullrev}
         n = off = 0
         data = fp.read()
         l = len(data)
         while off + s <= l:
             cur = data[off:off + s]
             off += s
             e = _unpack(indexformatv0, cur)
             # transform to revlogv1 format
             e2 = (offset_type(e[0], 0), e[1], -1, e[2], e[3],
                   nodemap[e[4]], nodemap[e[5]], e[6])
             index.append(e2)
             nodemap[e[6]] = n
             n += 1
 
         return index, nodemap, None
 
     def packentry(self, entry, node, version, rev):
         e2 = (getoffset(entry[0]), entry[1], entry[3], entry[4],
               node(entry[5]), node(entry[6]), entry[7])
         return _pack(indexformatv0, *e2)
 
 # index ng:
 # 6 bytes offset
 # 2 bytes flags
 # 4 bytes compressed length
 # 4 bytes uncompressed length
 # 4 bytes: base rev
 # 4 bytes link rev
 # 4 bytes parent 1 rev
 # 4 bytes parent 2 rev
 # 32 bytes: nodeid
 indexformatng = ">Qiiiiii20s12x"
 ngshaoffset = 32
 versionformat = ">I"
 
 class revlogio(object):
     def __init__(self):
         self.size = struct.calcsize(indexformatng)
 
     def parseindex(self, fp, inline):
         try:
             size = util.fstat(fp).st_size
         except AttributeError:
             size = 0
 
         if lazyparser.safe_to_use and not inline and size > 1000000:
             # big index, let's parse it on demand
             parser = lazyparser(fp, size)
             index = lazyindex(parser)
             nodemap = lazymap(parser)
             e = list(index[0])
             type = gettype(e[0])
             e[0] = offset_type(0, type)
             index[0] = e
             return index, nodemap, None
 
         s = self.size
         cache = None
         index = []
         nodemap =  {nullid: nullrev}
         n = off = 0
         # if we're not using lazymap, always read the whole index
         data = fp.read()
         l = len(data) - s
         append = index.append
         if inline:
             cache = (0, data)
             while off <= l:
                 e = _unpack(indexformatng, data[off:off + s])
                 nodemap[e[7]] = n
                 append(e)
                 n += 1
                 if e[1] < 0:
                     break
                 off += e[1] + s
         else:
             while off <= l:
                 e = _unpack(indexformatng, data[off:off + s])
                 nodemap[e[7]] = n
                 append(e)
                 n += 1
                 off += s
 
         e = list(index[0])
         type = gettype(e[0])
         e[0] = offset_type(0, type)
         index[0] = e
 
         return index, nodemap, cache
 
     def packentry(self, entry, node, version, rev):
         p = _pack(indexformatng, *entry)
         if rev == 0:
             p = _pack(versionformat, version) + p[4:]
         return p
 
 class revlog(object):
     """
     the underlying revision storage object
 
     A revlog consists of two parts, an index and the revision data.
 
     The index is a file with a fixed record size containing
     information on each revision, includings its nodeid (hash), the
     nodeids of its parents, the position and offset of its data within
     the data file, and the revision it's based on. Finally, each entry
     contains a linkrev entry that can serve as a pointer to external
     data.
 
     The revision data itself is a linear collection of data chunks.
     Each chunk represents a revision and is usually represented as a
     delta against the previous chunk. To bound lookup time, runs of
     deltas are limited to about 2 times the length of the original
     version data. This makes retrieval of a version proportional to
     its size, or O(1) relative to the number of revisions.
 
     Both pieces of the revlog are written to in an append-only
     fashion, which means we never need to rewrite a file to insert or
     remove data, and can use some simple techniques to avoid the need
     for locking while reading.
     """
     def __init__(self, opener, indexfile):
         """
         create a revlog object
 
         opener is a function that abstracts the file opening operation
         and can be used to implement COW semantics or the like.
         """
         self.indexfile = indexfile
         self.datafile = indexfile[:-2] + ".d"
         self.opener = opener
         self._cache = None
         self._chunkcache = None
         self.nodemap = {nullid: nullrev}
         self.index = []
 
         v = REVLOG_DEFAULT_VERSION
         if hasattr(opener, "defversion"):
             v = opener.defversion
             if v & REVLOGNG:
                 v |= REVLOGNGINLINEDATA
 
         i = ""
         try:
             f = self.opener(self.indexfile)
             i = f.read(4)
             f.seek(0)
             if len(i) > 0:
                 v = struct.unpack(versionformat, i)[0]
         except IOError, inst:
             if inst.errno != errno.ENOENT:
                 raise
 
         self.version = v
         self._inline = v & REVLOGNGINLINEDATA
         flags = v & ~0xFFFF
         fmt = v & 0xFFFF
         if fmt == REVLOGV0 and flags:
             raise RevlogError(_("index %s unknown flags %#04x for format v0")
                               % (self.indexfile, flags >> 16))
         elif fmt == REVLOGNG and flags & ~REVLOGNGINLINEDATA:
             raise RevlogError(_("index %s unknown flags %#04x for revlogng")
                               % (self.indexfile, flags >> 16))
         elif fmt > REVLOGNG:
             raise RevlogError(_("index %s unknown format %d")
                               % (self.indexfile, fmt))
 
         self._io = revlogio()
         if self.version == REVLOGV0:
             self._io = revlogoldio()
         if i:
             d = self._io.parseindex(f, self._inline)
             self.index, self.nodemap, self._chunkcache = d
 
         # add the magic null revision at -1
         self.index.append((0, 0, 0, -1, -1, -1, -1, nullid))
 
     def _loadindex(self, start, end):
         """load a block of indexes all at once from the lazy parser"""
         if isinstance(self.index, lazyindex):
             self.index.p.loadindex(start, end)
 
     def _loadindexmap(self):
         """loads both the map and the index from the lazy parser"""
         if isinstance(self.index, lazyindex):
             p = self.index.p
             p.loadindex()
             self.nodemap = p.map
 
     def _loadmap(self):
         """loads the map from the lazy parser"""
         if isinstance(self.nodemap, lazymap):
             self.nodemap.p.loadmap()
             self.nodemap = self.nodemap.p.map
 
     def tip(self):
         return self.node(len(self.index) - 2)
     def count(self):
         return len(self.index) - 1
 
     def rev(self, node):
         try:
             return self.nodemap[node]
         except KeyError:
             raise LookupError(_('%s: no node %s') % (self.indexfile, hex(node)))
     def node(self, rev):
         return self.index[rev][7]
     def linkrev(self, node):
         return self.index[self.rev(node)][4]
     def parents(self, node):
         d = self.index[self.rev(node)][5:7]
         return (self.node(d[0]), self.node(d[1]))
     def parentrevs(self, rev):
         return self.index[rev][5:7]
     def start(self, rev):
         return int(self.index[rev][0] >> 16)
     def end(self, rev):
         return self.start(rev) + self.length(rev)
     def length(self, rev):
         return self.index[rev][1]
     def base(self, rev):
         return self.index[rev][3]
 
     def size(self, rev):
         """return the length of the uncompressed text for a given revision"""
         l = self.index[rev][2]
         if l >= 0:
             return l
 
         t = self.revision(self.node(rev))
         return len(t)
 
         # alternate implementation, The advantage to this code is it
         # will be faster for a single revision.  But, the results are not
         # cached, so finding the size of every revision will be slower.
         """
         if self.cache and self.cache[1] == rev:
             return len(self.cache[2])
 
         base = self.base(rev)
         if self.cache and self.cache[1] >= base and self.cache[1] < rev:
             base = self.cache[1]
             text = self.cache[2]
         else:
             text = self.revision(self.node(base))
 
         l = len(text)
         for x in xrange(base + 1, rev + 1):
             l = mdiff.patchedsize(l, self.chunk(x))
         return l
         """
 
     def reachable(self, node, stop=None):
         """return a hash of all nodes ancestral to a given node, including
          the node itself, stopping when stop is matched"""
         reachable = {}
         visit = [node]
         reachable[node] = 1
         if stop:
             stopn = self.rev(stop)
         else:
             stopn = 0
         while visit:
             n = visit.pop(0)
             if n == stop:
                 continue
             if n == nullid:
                 continue
             for p in self.parents(n):
                 if self.rev(p) < stopn:
                     continue
                 if p not in reachable:
                     reachable[p] = 1
                     visit.append(p)
         return reachable
 
     def nodesbetween(self, roots=None, heads=None):
         """Return a tuple containing three elements. Elements 1 and 2 contain
         a final list bases and heads after all the unreachable ones have been
         pruned.  Element 0 contains a topologically sorted list of all
 
         nodes that satisfy these constraints:
         1. All nodes must be descended from a node in roots (the nodes on
            roots are considered descended from themselves).
         2. All nodes must also be ancestors of a node in heads (the nodes in
            heads are considered to be their own ancestors).
 
         If roots is unspecified, nullid is assumed as the only root.
         If heads is unspecified, it is taken to be the output of the
         heads method (i.e. a list of all nodes in the repository that
         have no children)."""
         nonodes = ([], [], [])
         if roots is not None:
             roots = list(roots)
             if not roots:
                 return nonodes
             lowestrev = min([self.rev(n) for n in roots])
         else:
             roots = [nullid] # Everybody's a descendent of nullid
             lowestrev = nullrev
         if (lowestrev == nullrev) and (heads is None):
             # We want _all_ the nodes!
             return ([self.node(r) for r in xrange(0, self.count())],
                     [nullid], list(self.heads()))
         if heads is None:
             # All nodes are ancestors, so the latest ancestor is the last
             # node.
             highestrev = self.count() - 1
             # Set ancestors to None to signal that every node is an ancestor.
             ancestors = None
             # Set heads to an empty dictionary for later discovery of heads
             heads = {}
         else:
             heads = list(heads)
             if not heads:
                 return nonodes
             ancestors = {}
             # Turn heads into a dictionary so we can remove 'fake' heads.
             # Also, later we will be using it to filter out the heads we can't
             # find from roots.
             heads = dict.fromkeys(heads, 0)
             # Start at the top and keep marking parents until we're done.
             nodestotag = heads.keys()
             # Remember where the top was so we can use it as a limit later.
             highestrev = max([self.rev(n) for n in nodestotag])
             while nodestotag:
                 # grab a node to tag
                 n = nodestotag.pop()
                 # Never tag nullid
                 if n == nullid:
                     continue
                 # A node's revision number represents its place in a
                 # topologically sorted list of nodes.
                 r = self.rev(n)
                 if r >= lowestrev:
                     if n not in ancestors:
                         # If we are possibly a descendent of one of the roots
                         # and we haven't already been marked as an ancestor
                         ancestors[n] = 1 # Mark as ancestor
                         # Add non-nullid parents to list of nodes to tag.
                         nodestotag.extend([p for p in self.parents(n) if
                                            p != nullid])
                     elif n in heads: # We've seen it before, is it a fake head?
                         # So it is, real heads should not be the ancestors of
                         # any other heads.
                         heads.pop(n)
             if not ancestors:
                 return nonodes
             # Now that we have our set of ancestors, we want to remove any
             # roots that are not ancestors.
 
             # If one of the roots was nullid, everything is included anyway.
             if lowestrev > nullrev:
                 # But, since we weren't, let's recompute the lowest rev to not
                 # include roots that aren't ancestors.
 
                 # Filter out roots that aren't ancestors of heads
                 roots = [n for n in roots if n in ancestors]
                 # Recompute the lowest revision
                 if roots:
                     lowestrev = min([self.rev(n) for n in roots])
                 else:
                     # No more roots?  Return empty list
                     return nonodes
             else:
                 # We are descending from nullid, and don't need to care about
                 # any other roots.
                 lowestrev = nullrev
                 roots = [nullid]
         # Transform our roots list into a 'set' (i.e. a dictionary where the
         # values don't matter.
         descendents = dict.fromkeys(roots, 1)
         # Also, keep the original roots so we can filter out roots that aren't
         # 'real' roots (i.e. are descended from other roots).
         roots = descendents.copy()
         # Our topologically sorted list of output nodes.
         orderedout = []
         # Don't start at nullid since we don't want nullid in our output list,
         # and if nullid shows up in descedents, empty parents will look like
         # they're descendents.
         for r in xrange(max(lowestrev, 0), highestrev + 1):
             n = self.node(r)
             isdescendent = False
             if lowestrev == nullrev:  # Everybody is a descendent of nullid
                 isdescendent = True
             elif n in descendents:
                 # n is already a descendent
                 isdescendent = True
                 # This check only needs to be done here because all the roots
                 # will start being marked is descendents before the loop.
                 if n in roots:
                     # If n was a root, check if it's a 'real' root.
                     p = tuple(self.parents(n))
                     # If any of its parents are descendents, it's not a root.
                     if (p[0] in descendents) or (p[1] in descendents):
                         roots.pop(n)
             else:
                 p = tuple(self.parents(n))
                 # A node is a descendent if either of its parents are
                 # descendents.  (We seeded the dependents list with the roots
                 # up there, remember?)
                 if (p[0] in descendents) or (p[1] in descendents):
                     descendents[n] = 1
                     isdescendent = True
             if isdescendent and ((ancestors is None) or (n in ancestors)):
                 # Only include nodes that are both descendents and ancestors.
                 orderedout.append(n)
                 if (ancestors is not None) and (n in heads):
                     # We're trying to figure out which heads are reachable
                     # from roots.
                     # Mark this head as having been reached
                     heads[n] = 1
                 elif ancestors is None:
                     # Otherwise, we're trying to discover the heads.
                     # Assume this is a head because if it isn't, the next step
                     # will eventually remove it.
                     heads[n] = 1
                     # But, obviously its parents aren't.
                     for p in self.parents(n):
                         heads.pop(p, None)
         heads = [n for n in heads.iterkeys() if heads[n] != 0]
         roots = roots.keys()
         assert orderedout
         assert roots
         assert heads
         return (orderedout, roots, heads)
 
     def heads(self, start=None, stop=None):
         """return the list of all nodes that have no children
 
         if start is specified, only heads that are descendants of
         start will be returned
         if stop is specified, it will consider all the revs from stop
         as if they had no children
         """
         if start is None and stop is None:
             count = self.count()
             if not count:
                 return [nullid]
             ishead = [1] * (count + 1)
             index = self.index
             for r in xrange(count):
                 e = index[r]
                 ishead[e[5]] = ishead[e[6]] = 0
             return [self.node(r) for r in xrange(count) if ishead[r]]
 
         if start is None:
             start = nullid
         if stop is None:
             stop = []
         stoprevs = dict.fromkeys([self.rev(n) for n in stop])
         startrev = self.rev(start)
         reachable = {startrev: 1}
         heads = {startrev: 1}
 
         parentrevs = self.parentrevs
         for r in xrange(startrev + 1, self.count()):
             for p in parentrevs(r):
                 if p in reachable:
                     if r not in stoprevs:
                         reachable[r] = 1
                     heads[r] = 1
                 if p in heads and p not in stoprevs:
                     del heads[p]
 
         return [self.node(r) for r in heads]
 
     def children(self, node):
         """find the children of a given node"""
         c = []
         p = self.rev(node)
         for r in range(p + 1, self.count()):
             prevs = [pr for pr in self.parentrevs(r) if pr != nullrev]
             if prevs:
                 for pr in prevs:
                     if pr == p:
                         c.append(self.node(r))
             elif p == nullrev:
                 c.append(self.node(r))
         return c
 
     def _match(self, id):
         if isinstance(id, (long, int)):
             # rev
             return self.node(id)
         if len(id) == 20:
             # possibly a binary node
             # odds of a binary node being all hex in ASCII are 1 in 10**25
             try:
                 node = id
                 r = self.rev(node) # quick search the index
                 return node
             except LookupError:
                 pass # may be partial hex id
         try:
             # str(rev)
             rev = int(id)
             if str(rev) != id:
                 raise ValueError
             if rev < 0:
                 rev = self.count() + rev
             if rev < 0 or rev >= self.count():
                 raise ValueError
             return self.node(rev)
         except (ValueError, OverflowError):
             pass
         if len(id) == 40:
             try:
                 # a full hex nodeid?
                 node = bin(id)
                 r = self.rev(node)
                 return node
             except TypeError:
                 pass
 
     def _partialmatch(self, id):
         if len(id) < 40:
             try:
                 # hex(node)[:...]
                 bin_id = bin(id[:len(id) & ~1]) # grab an even number of digits
                 node = None
                 for n in self.nodemap:
                     if n.startswith(bin_id) and hex(n).startswith(id):
                         if node is not None:
                             raise LookupError(_("Ambiguous identifier"))
                         node = n
                 if node is not None:
                     return node
             except TypeError:
                 pass
 
     def lookup(self, id):
         """locate a node based on:
             - revision number or str(revision number)
             - nodeid or subset of hex nodeid
         """
         n = self._match(id)
         if n is not None:
             return n
         n = self._partialmatch(id)
         if n:
             return n
 
         raise LookupError(_("No match found"))
 
     def cmp(self, node, text):
         """compare text with a given file revision"""
         p1, p2 = self.parents(node)
         return hash(text, p1, p2) != node
 
     def chunk(self, rev, df=None):
         def loadcache(df):
             if not df:
                 if self._inline:
                     df = self.opener(self.indexfile)
                 else:
                     df = self.opener(self.datafile)
             df.seek(start)
             self._chunkcache = (start, df.read(cache_length))
 
         start, length = self.start(rev), self.length(rev)
         if self._inline:
             start += (rev + 1) * self._io.size
         end = start + length
 
         offset = 0
         if not self._chunkcache:
             cache_length = max(65536, length)
             loadcache(df)
         else:
             cache_start = self._chunkcache[0]
             cache_length = len(self._chunkcache[1])
             cache_end = cache_start + cache_length
             if start >= cache_start and end <= cache_end:
                 # it is cached
                 offset = start - cache_start
             else:
                 cache_length = max(65536, length)
                 loadcache(df)
 
         # avoid copying large chunks
         c = self._chunkcache[1]
         if cache_length != length:
             c = c[offset:offset + length]
 
         return decompress(c)
 
     def delta(self, node):
         """return or calculate a delta between a node and its predecessor"""
         r = self.rev(node)
         return self.revdiff(r - 1, r)
 
     def revdiff(self, rev1, rev2):
         """return or calculate a delta between two revisions"""
         if rev1 + 1 == rev2 and self.base(rev1) == self.base(rev2):
             return self.chunk(rev2)
 
         return mdiff.textdiff(self.revision(self.node(rev1)),
                               self.revision(self.node(rev2)))
 
     def revision(self, node):
         """return an uncompressed revision of a given"""
         if node == nullid:
             return ""
         if self._cache and self._cache[0] == node:
-            return self._cache[2]
+            return str(self._cache[2])
 
         # look up what we need to read
         text = None
         rev = self.rev(node)
         base = self.base(rev)
 
         # check rev flags
         if self.index[rev][0] & 0xFFFF:
             raise RevlogError(_('incompatible revision flag %x') %
                               (self.index[rev][0] & 0xFFFF))
 
         if self._inline:
             # we probably have the whole chunk cached
             df = None
         else:
             df = self.opener(self.datafile)
 
         # do we have useful data cached?
         if self._cache and self._cache[1] >= base and self._cache[1] < rev:
             base = self._cache[1]
-            text = self._cache[2]
+            text = str(self._cache[2])
             self._loadindex(base, rev + 1)
         else:
             self._loadindex(base, rev + 1)
             text = self.chunk(base, df=df)
 
         bins = [self.chunk(r, df) for r in xrange(base + 1, rev + 1)]
         text = mdiff.patches(text, bins)
         p1, p2 = self.parents(node)
         if node != hash(text, p1, p2):
             raise RevlogError(_("integrity check failed on %s:%d")
                               % (self.datafile, rev))
 
         self._cache = (node, rev, text)
         return text
 
     def checkinlinesize(self, tr, fp=None):
         if not self._inline:
             return
         if not fp:
             fp = self.opener(self.indexfile, 'r')
             fp.seek(0, 2)
         size = fp.tell()
         if size < 131072:
             return
         trinfo = tr.find(self.indexfile)
         if trinfo == None:
             raise RevlogError(_("%s not found in the transaction")
                               % self.indexfile)
 
         trindex = trinfo[2]
         dataoff = self.start(trindex)
 
         tr.add(self.datafile, dataoff)
         df = self.opener(self.datafile, 'w')
         calc = self._io.size
         for r in xrange(self.count()):
             start = self.start(r) + (r + 1) * calc
             length = self.length(r)
             fp.seek(start)
             d = fp.read(length)
             df.write(d)
         fp.close()
         df.close()
         fp = self.opener(self.indexfile, 'w', atomictemp=True)
         self.version &= ~(REVLOGNGINLINEDATA)
         self._inline = False
         for i in xrange(self.count()):
             e = self._io.packentry(self.index[i], self.node, self.version, i)
             fp.write(e)
 
         # if we don't call rename, the temp file will never replace the
         # real index
         fp.rename()
 
         tr.replace(self.indexfile, trindex * calc)
         self._chunkcache = None
 
     def addrevision(self, text, transaction, link, p1, p2, d=None):
         """add a revision to the log
 
         text - the revision data to add
         transaction - the transaction object used for rollback
         link - the linkrev data to add
         p1, p2 - the parent nodeids of the revision
         d - an optional precomputed delta
         """
         dfh = None
         if not self._inline:
             dfh = self.opener(self.datafile, "a")
         ifh = self.opener(self.indexfile, "a+")
         return self._addrevision(text, transaction, link, p1, p2, d, ifh, dfh)
 
     def _addrevision(self, text, transaction, link, p1, p2, d, ifh, dfh):
         node = hash(text, p1, p2)
         if node in self.nodemap:
             return node
 
         curr = self.count()
         prev = curr - 1
         base = self.base(prev)
         offset = self.end(prev)
 
         if curr:
             if not d:
                 ptext = self.revision(self.node(prev))
                 d = mdiff.textdiff(ptext, text)
             data = compress(d)
             l = len(data[1]) + len(data[0])
             dist = l + offset - self.start(base)
 
         # full versions are inserted when the needed deltas
         # become comparable to the uncompressed text
         if not curr or dist > len(text) * 2:
             data = compress(text)
             l = len(data[1]) + len(data[0])
             base = curr
 
         e = (offset_type(offset, 0), l, len(text),
              base, link, self.rev(p1), self.rev(p2), node)
         self.index.insert(-1, e)
         self.nodemap[node] = curr
 
         entry = self._io.packentry(e, self.node, self.version, curr)
         if not self._inline:
             transaction.add(self.datafile, offset)
             transaction.add(self.indexfile, curr * len(entry))
             if data[0]:
                 dfh.write(data[0])
             dfh.write(data[1])
             dfh.flush()
             ifh.write(entry)
         else:
             offset += curr * self._io.size
             transaction.add(self.indexfile, offset, curr)
             ifh.write(entry)
             ifh.write(data[0])
             ifh.write(data[1])
             self.checkinlinesize(transaction, ifh)
 
         self._cache = (node, curr, text)
         return node
 
     def ancestor(self, a, b):
         """calculate the least common ancestor of nodes a and b"""
 
         def parents(rev):
             return [p for p in self.parentrevs(rev) if p != nullrev]
 
         c = ancestor.ancestor(self.rev(a), self.rev(b), parents)
         if c is None:
             return nullid
 
         return self.node(c)
 
     def group(self, nodelist, lookup, infocollect=None):
         """calculate a delta group
 
         Given a list of changeset revs, return a set of deltas and
         metadata corresponding to nodes. the first delta is
         parent(nodes[0]) -> nodes[0] the receiver is guaranteed to
         have this parent as it has all history before these
         changesets. parent is parent[0]
         """
         revs = [self.rev(n) for n in nodelist]
 
         # if we don't have any revisions touched by these changesets, bail
         if not revs:
             yield changegroup.closechunk()
             return
 
         # add the parent of the first rev
         p = self.parents(self.node(revs[0]))[0]
         revs.insert(0, self.rev(p))
 
         # build deltas
         for d in xrange(0, len(revs) - 1):
             a, b = revs[d], revs[d + 1]
             nb = self.node(b)
 
             if infocollect is not None:
                 infocollect(nb)
 
             p = self.parents(nb)
             meta = nb + p[0] + p[1] + lookup(nb)
             if a == -1:
                 d = self.revision(nb)
                 meta += mdiff.trivialdiffheader(len(d))
             else:
                 d = self.revdiff(a, b)
             yield changegroup.chunkheader(len(meta) + len(d))
             yield meta
             if len(d) > 2**20:
                 pos = 0
                 while pos < len(d):
                     pos2 = pos + 2 ** 18
                     yield d[pos:pos2]
                     pos = pos2
             else:
                 yield d
 
         yield changegroup.closechunk()
 
     def addgroup(self, revs, linkmapper, transaction, unique=0):
         """
         add a delta group
 
         given a set of deltas, add them to the revision log. the
         first delta is against its parent, which should be in our
         log, the rest are against the previous delta.
         """
 
         #track the base of the current delta log
         r = self.count()
         t = r - 1
         node = None
 
         base = prev = nullrev
         start = end = textlen = 0
         if r:
             end = self.end(t)
 
         ifh = self.opener(self.indexfile, "a+")
         isize = r * self._io.size
         if self._inline:
             transaction.add(self.indexfile, end + isize, r)
             dfh = None
         else:
             transaction.add(self.indexfile, isize, r)
             transaction.add(self.datafile, end)
             dfh = self.opener(self.datafile, "a")
 
         # loop through our set of deltas
         chain = None
         for chunk in revs:
             node, p1, p2, cs = struct.unpack("20s20s20s20s", chunk[:80])
             link = linkmapper(cs)
             if node in self.nodemap:
                 # this can happen if two branches make the same change
                 # if unique:
                 #    raise RevlogError(_("already have %s") % hex(node[:4]))
                 chain = node
                 continue
             delta = buffer(chunk, 80)
             del chunk
 
             for p in (p1, p2):
                 if not p in self.nodemap:
                     raise LookupError(_("unknown parent %s") % short(p))
 
             if not chain:
                 # retrieve the parent revision of the delta chain
                 chain = p1
                 if not chain in self.nodemap:
                     raise LookupError(_("unknown base %s") % short(chain[:4]))
 
             # full versions are inserted when the needed deltas become
             # comparable to the uncompressed text or when the previous
             # version is not the one we have a delta against. We use
             # the size of the previous full rev as a proxy for the
             # current size.
 
             if chain == prev:
                 cdelta = compress(delta)
                 cdeltalen = len(cdelta[0]) + len(cdelta[1])
                 textlen = mdiff.patchedsize(textlen, delta)
 
             if chain != prev or (end - start + cdeltalen) > textlen * 2:
                 # flush our writes here so we can read it in revision
                 if dfh:
                     dfh.flush()
                 ifh.flush()
                 text = self.revision(chain)
                 if len(text) == 0:
                     # skip over trivial delta header
                     text = buffer(delta, 12)
                 else:
                     text = mdiff.patches(text, [delta])
                 del delta
                 chk = self._addrevision(text, transaction, link, p1, p2, None,
                                         ifh, dfh)
                 if not dfh and not self._inline:
                     # addrevision switched from inline to conventional
                     # reopen the index
                     dfh = self.opener(self.datafile, "a")
                     ifh = self.opener(self.indexfile, "a")
                 if chk != node:
                     raise RevlogError(_("consistency error adding group"))
                 textlen = len(text)
             else:
                 e = (offset_type(end, 0), cdeltalen, textlen, base,
                      link, self.rev(p1), self.rev(p2), node)
                 self.index.insert(-1, e)
                 self.nodemap[node] = r
                 entry = self._io.packentry(e, self.node, self.version, r)
                 if self._inline:
                     ifh.write(entry)
                     ifh.write(cdelta[0])
                     ifh.write(cdelta[1])
                     self.checkinlinesize(transaction, ifh)
                     if not self._inline:
                         dfh = self.opener(self.datafile, "a")
                         ifh = self.opener(self.indexfile, "a")
                 else:
                     dfh.write(cdelta[0])
                     dfh.write(cdelta[1])
                     ifh.write(entry)
 
             t, r, chain, prev = r, r + 1, node, node
             base = self.base(t)
             start = self.start(base)
             end = self.end(t)
 
         return node
 
     def strip(self, rev, minlink):
         if self.count() == 0 or rev >= self.count():
             return
 
         if isinstance(self.index, lazyindex):
             self._loadindexmap()
 
         # When stripping away a revision, we need to make sure it
         # does not actually belong to an older changeset.
         # The minlink parameter defines the oldest revision
         # we're allowed to strip away.
         while minlink > self.index[rev][4]:
             rev += 1
             if rev >= self.count():
                 return
 
         # first truncate the files on disk
         end = self.start(rev)
         if not self._inline:
             df = self.opener(self.datafile, "a")
             df.truncate(end)
             end = rev * self._io.size
         else:
             end += rev * self._io.size
 
         indexf = self.opener(self.indexfile, "a")
         indexf.truncate(end)
 
         # then reset internal state in memory to forget those revisions
         self._cache = None
         self._chunkcache = None
         for x in xrange(rev, self.count()):
             del self.nodemap[self.node(x)]
 
         del self.index[rev:-1]
 
     def checksize(self):
         expected = 0
         if self.count():
             expected = max(0, self.end(self.count() - 1))
 
         try:
             f = self.opener(self.datafile)
             f.seek(0, 2)
             actual = f.tell()
             dd = actual - expected
         except IOError, inst:
             if inst.errno != errno.ENOENT:
                 raise
             dd = 0
 
         try:
             f = self.opener(self.indexfile)
             f.seek(0, 2)
             actual = f.tell()
             s = self._io.size
             i = max(0, actual / s)
             di = actual - (i * s)
             if self._inline:
                 databytes = 0
                 for r in xrange(self.count()):
                     databytes += max(0, self.length(r))
                 dd = 0
                 di = actual - self.count() * s - databytes
         except IOError, inst:
             if inst.errno != errno.ENOENT:
                 raise
             di = 0
 
         return (dd, di)

mercurial/util.py

@@ -1,1690 +1,1690 b''
 """
 util.py - Mercurial utility functions and platform specfic implementations
 
  Copyright 2005 K. Thananchayan <[e-mail]>
  Copyright 2005-2007 Matt Mackall <[e-mail]>
  Copyright 2006 Vadim Gelfer <[e-mail]>
 
 This software may be used and distributed according to the terms
 of the GNU General Public License, incorporated herein by reference.
 
 This contains helper routines that are independent of the SCM core and hide
 platform-specific details from the core.
 """
 
 from i18n import _
 import cStringIO, errno, getpass, popen2, re, shutil, sys, tempfile, strutil
 import os, stat, threading, time, calendar, ConfigParser, locale, glob, osutil
 
 try:
     set = set
     frozenset = frozenset
 except NameError:
     from sets import Set as set, ImmutableSet as frozenset
 
 try:
     _encoding = os.environ.get("HGENCODING")
     if sys.platform == 'darwin' and not _encoding:
         # On darwin, getpreferredencoding ignores the locale environment and
         # always returns mac-roman. We override this if the environment is
         # not C (has been customized by the user).
         locale.setlocale(locale.LC_CTYPE, '')
         _encoding = locale.getlocale()[1]
     if not _encoding:
         _encoding = locale.getpreferredencoding() or 'ascii'
 except locale.Error:
     _encoding = 'ascii'
 _encodingmode = os.environ.get("HGENCODINGMODE", "strict")
 _fallbackencoding = 'ISO-8859-1'
 
 def tolocal(s):
     """
     Convert a string from internal UTF-8 to local encoding
 
     All internal strings should be UTF-8 but some repos before the
     implementation of locale support may contain latin1 or possibly
     other character sets. We attempt to decode everything strictly
     using UTF-8, then Latin-1, and failing that, we use UTF-8 and
     replace unknown characters.
     """
     for e in ('UTF-8', _fallbackencoding):
         try:
             u = s.decode(e) # attempt strict decoding
             return u.encode(_encoding, "replace")
         except LookupError, k:
             raise Abort(_("%s, please check your locale settings") % k)
         except UnicodeDecodeError:
             pass
     u = s.decode("utf-8", "replace") # last ditch
     return u.encode(_encoding, "replace")
 
 def fromlocal(s):
     """
     Convert a string from the local character encoding to UTF-8
 
     We attempt to decode strings using the encoding mode set by
     HGENCODINGMODE, which defaults to 'strict'. In this mode, unknown
     characters will cause an error message. Other modes include
     'replace', which replaces unknown characters with a special
     Unicode character, and 'ignore', which drops the character.
     """
     try:
         return s.decode(_encoding, _encodingmode).encode("utf-8")
     except UnicodeDecodeError, inst:
         sub = s[max(0, inst.start-10):inst.start+10]
         raise Abort("decoding near '%s': %s!" % (sub, inst))
     except LookupError, k:
         raise Abort(_("%s, please check your locale settings") % k)
 
 def locallen(s):
     """Find the length in characters of a local string"""
     return len(s.decode(_encoding, "replace"))
 
 def localsub(s, a, b=None):
     try:
         u = s.decode(_encoding, _encodingmode)
         if b is not None:
             u = u[a:b]
         else:
             u = u[:a]
         return u.encode(_encoding, _encodingmode)
     except UnicodeDecodeError, inst:
         sub = s[max(0, inst.start-10), inst.start+10]
         raise Abort(_("decoding near '%s': %s!") % (sub, inst))
 
 # used by parsedate
 defaultdateformats = (
     '%Y-%m-%d %H:%M:%S',
     '%Y-%m-%d %I:%M:%S%p',
     '%Y-%m-%d %H:%M',
     '%Y-%m-%d %I:%M%p',
     '%Y-%m-%d',
     '%m-%d',
     '%m/%d',
     '%m/%d/%y',
     '%m/%d/%Y',
     '%a %b %d %H:%M:%S %Y',
     '%a %b %d %I:%M:%S%p %Y',
     '%a, %d %b %Y %H:%M:%S',        #  GNU coreutils "/bin/date --rfc-2822"
     '%b %d %H:%M:%S %Y',
     '%b %d %I:%M:%S%p %Y',
     '%b %d %H:%M:%S',
     '%b %d %I:%M:%S%p',
     '%b %d %H:%M',
     '%b %d %I:%M%p',
     '%b %d %Y',
     '%b %d',
     '%H:%M:%S',
     '%I:%M:%SP',
     '%H:%M',
     '%I:%M%p',
 )
 
 extendeddateformats = defaultdateformats + (
     "%Y",
     "%Y-%m",
     "%b",
     "%b %Y",
     )
 
 class SignalInterrupt(Exception):
     """Exception raised on SIGTERM and SIGHUP."""
 
 # differences from SafeConfigParser:
 # - case-sensitive keys
 # - allows values that are not strings (this means that you may not
 #   be able to save the configuration to a file)
 class configparser(ConfigParser.SafeConfigParser):
     def optionxform(self, optionstr):
         return optionstr
 
     def set(self, section, option, value):
         return ConfigParser.ConfigParser.set(self, section, option, value)
 
     def _interpolate(self, section, option, rawval, vars):
         if not isinstance(rawval, basestring):
             return rawval
         return ConfigParser.SafeConfigParser._interpolate(self, section,
                                                           option, rawval, vars)
 
 def cachefunc(func):
     '''cache the result of function calls'''
     # XXX doesn't handle keywords args
     cache = {}
     if func.func_code.co_argcount == 1:
         # we gain a small amount of time because
         # we don't need to pack/unpack the list
         def f(arg):
             if arg not in cache:
                 cache[arg] = func(arg)
             return cache[arg]
     else:
         def f(*args):
             if args not in cache:
                 cache[args] = func(*args)
             return cache[args]
 
     return f
 
 def pipefilter(s, cmd):
     '''filter string S through command CMD, returning its output'''
     (pin, pout) = os.popen2(cmd, 'b')
     def writer():
         try:
             pin.write(s)
             pin.close()
         except IOError, inst:
             if inst.errno != errno.EPIPE:
                 raise
 
     # we should use select instead on UNIX, but this will work on most
     # systems, including Windows
     w = threading.Thread(target=writer)
     w.start()
     f = pout.read()
     pout.close()
     w.join()
     return f
 
 def tempfilter(s, cmd):
     '''filter string S through a pair of temporary files with CMD.
     CMD is used as a template to create the real command to be run,
     with the strings INFILE and OUTFILE replaced by the real names of
     the temporary files generated.'''
     inname, outname = None, None
     try:
         infd, inname = tempfile.mkstemp(prefix='hg-filter-in-')
         fp = os.fdopen(infd, 'wb')
         fp.write(s)
         fp.close()
         outfd, outname = tempfile.mkstemp(prefix='hg-filter-out-')
         os.close(outfd)
         cmd = cmd.replace('INFILE', inname)
         cmd = cmd.replace('OUTFILE', outname)
         code = os.system(cmd)
         if sys.platform == 'OpenVMS' and code & 1:
             code = 0
         if code: raise Abort(_("command '%s' failed: %s") %
                              (cmd, explain_exit(code)))
         return open(outname, 'rb').read()
     finally:
         try:
             if inname: os.unlink(inname)
         except: pass
         try:
             if outname: os.unlink(outname)
         except: pass
 
 filtertable = {
     'tempfile:': tempfilter,
     'pipe:': pipefilter,
     }
 
 def filter(s, cmd):
     "filter a string through a command that transforms its input to its output"
     for name, fn in filtertable.iteritems():
         if cmd.startswith(name):
             return fn(s, cmd[len(name):].lstrip())
     return pipefilter(s, cmd)
 
 def binary(s):
     """return true if a string is binary data using diff's heuristic"""
     if s and '\0' in s[:4096]:
         return True
     return False
 
 def unique(g):
     """return the uniq elements of iterable g"""
     seen = {}
     l = []
     for f in g:
         if f not in seen:
             seen[f] = 1
             l.append(f)
     return l
 
 class Abort(Exception):
     """Raised if a command needs to print an error and exit."""
 
 class UnexpectedOutput(Abort):
     """Raised to print an error with part of output and exit."""
 
 def always(fn): return True
 def never(fn): return False
 
 def expand_glob(pats):
     '''On Windows, expand the implicit globs in a list of patterns'''
     if os.name != 'nt':
         return list(pats)
     ret = []
     for p in pats:
         kind, name = patkind(p, None)
         if kind is None:
             globbed = glob.glob(name)
             if globbed:
                 ret.extend(globbed)
                 continue
             # if we couldn't expand the glob, just keep it around
         ret.append(p)
     return ret
 
 def patkind(name, dflt_pat='glob'):
     """Split a string into an optional pattern kind prefix and the
     actual pattern."""
     for prefix in 're', 'glob', 'path', 'relglob', 'relpath', 'relre':
         if name.startswith(prefix + ':'): return name.split(':', 1)
     return dflt_pat, name
 
 def globre(pat, head='^', tail='$'):
     "convert a glob pattern into a regexp"
     i, n = 0, len(pat)
     res = ''
     group = False
     def peek(): return i < n and pat[i]
     while i < n:
         c = pat[i]
         i = i+1
         if c == '*':
             if peek() == '*':
                 i += 1
                 res += '.*'
             else:
                 res += '[^/]*'
         elif c == '?':
             res += '.'
         elif c == '[':
             j = i
             if j < n and pat[j] in '!]':
                 j += 1
             while j < n and pat[j] != ']':
                 j += 1
             if j >= n:
                 res += '\\['
             else:
                 stuff = pat[i:j].replace('\\','\\\\')
                 i = j + 1
                 if stuff[0] == '!':
                     stuff = '^' + stuff[1:]
                 elif stuff[0] == '^':
                     stuff = '\\' + stuff
                 res = '%s[%s]' % (res, stuff)
         elif c == '{':
             group = True
             res += '(?:'
         elif c == '}' and group:
             res += ')'
             group = False
         elif c == ',' and group:
             res += '|'
         elif c == '\\':
             p = peek()
             if p:
                 i += 1
                 res += re.escape(p)
             else:
                 res += re.escape(c)
         else:
             res += re.escape(c)
     return head + res + tail
 
 _globchars = {'[': 1, '{': 1, '*': 1, '?': 1}
 
 def pathto(root, n1, n2):
     '''return the relative path from one place to another.
     root should use os.sep to separate directories
     n1 should use os.sep to separate directories
     n2 should use "/" to separate directories
     returns an os.sep-separated path.
 
     If n1 is a relative path, it's assumed it's
     relative to root.
     n2 should always be relative to root.
     '''
     if not n1: return localpath(n2)
     if os.path.isabs(n1):
         if os.path.splitdrive(root)[0] != os.path.splitdrive(n1)[0]:
             return os.path.join(root, localpath(n2))
         n2 = '/'.join((pconvert(root), n2))
     a, b = n1.split(os.sep), n2.split('/')
     a.reverse()
     b.reverse()
     while a and b and a[-1] == b[-1]:
         a.pop()
         b.pop()
     b.reverse()
     return os.sep.join((['..'] * len(a)) + b)
 
 def canonpath(root, cwd, myname):
     """return the canonical path of myname, given cwd and root"""
     if root == os.sep:
         rootsep = os.sep
     elif root.endswith(os.sep):
         rootsep = root
     else:
         rootsep = root + os.sep
     name = myname
     if not os.path.isabs(name):
         name = os.path.join(root, cwd, name)
     name = os.path.normpath(name)
     audit_path = path_auditor(root)
     if name != rootsep and name.startswith(rootsep):
         name = name[len(rootsep):]
         audit_path(name)
         return pconvert(name)
     elif name == root:
         return ''
     else:
         # Determine whether `name' is in the hierarchy at or beneath `root',
         # by iterating name=dirname(name) until that causes no change (can't
         # check name == '/', because that doesn't work on windows).  For each
         # `name', compare dev/inode numbers.  If they match, the list `rel'
         # holds the reversed list of components making up the relative file
         # name we want.
         root_st = os.stat(root)
         rel = []
         while True:
             try:
                 name_st = os.stat(name)
             except OSError:
                 break
             if samestat(name_st, root_st):
                 if not rel:
                     # name was actually the same as root (maybe a symlink)
                     return ''
                 rel.reverse()
                 name = os.path.join(*rel)
                 audit_path(name)
                 return pconvert(name)
             dirname, basename = os.path.split(name)
             rel.append(basename)
             if dirname == name:
                 break
             name = dirname
 
         raise Abort('%s not under root' % myname)
 
 def matcher(canonroot, cwd='', names=[], inc=[], exc=[], src=None):
     return _matcher(canonroot, cwd, names, inc, exc, 'glob', src)
 
 def cmdmatcher(canonroot, cwd='', names=[], inc=[], exc=[], src=None,
                globbed=False, default=None):
     default = default or 'relpath'
     if default == 'relpath' and not globbed:
         names = expand_glob(names)
     return _matcher(canonroot, cwd, names, inc, exc, default, src)
 
 def _matcher(canonroot, cwd, names, inc, exc, dflt_pat, src):
     """build a function to match a set of file patterns
 
     arguments:
     canonroot - the canonical root of the tree you're matching against
     cwd - the current working directory, if relevant
     names - patterns to find
     inc - patterns to include
     exc - patterns to exclude
     dflt_pat - if a pattern in names has no explicit type, assume this one
     src - where these patterns came from (e.g. .hgignore)
 
     a pattern is one of:
     'glob:<glob>' - a glob relative to cwd
     're:<regexp>' - a regular expression
     'path:<path>' - a path relative to canonroot
     'relglob:<glob>' - an unrooted glob (*.c matches C files in all dirs)
     'relpath:<path>' - a path relative to cwd
     'relre:<regexp>' - a regexp that doesn't have to match the start of a name
     '<something>' - one of the cases above, selected by the dflt_pat argument
 
     returns:
     a 3-tuple containing
     - list of roots (places where one should start a recursive walk of the fs);
       this often matches the explicit non-pattern names passed in, but also
       includes the initial part of glob: patterns that has no glob characters
     - a bool match(filename) function
     - a bool indicating if any patterns were passed in
     """
 
     # a common case: no patterns at all
     if not names and not inc and not exc:
         return [], always, False
 
     def contains_glob(name):
         for c in name:
             if c in _globchars: return True
         return False
 
     def regex(kind, name, tail):
         '''convert a pattern into a regular expression'''
         if not name:
             return ''
         if kind == 're':
             return name
         elif kind == 'path':
             return '^' + re.escape(name) + '(?:/|$)'
         elif kind == 'relglob':
             return globre(name, '(?:|.*/)', tail)
         elif kind == 'relpath':
             return re.escape(name) + '(?:/|$)'
         elif kind == 'relre':
             if name.startswith('^'):
                 return name
             return '.*' + name
         return globre(name, '', tail)
 
     def matchfn(pats, tail):
         """build a matching function from a set of patterns"""
         if not pats:
             return
         try:
             pat = '(?:%s)' % '|'.join([regex(k, p, tail) for (k, p) in pats])
             return re.compile(pat).match
         except OverflowError:
             # We're using a Python with a tiny regex engine and we
             # made it explode, so we'll divide the pattern list in two
             # until it works
             l = len(pats)
             if l < 2:
                 raise
             a, b = matchfn(pats[:l/2], tail), matchfn(pats[l/2:], tail)
             return lambda s: a(s) or b(s)
         except re.error:
             for k, p in pats:
                 try:
                     re.compile('(?:%s)' % regex(k, p, tail))
                 except re.error:
                     if src:
                         raise Abort("%s: invalid pattern (%s): %s" %
                                     (src, k, p))
                     else:
                         raise Abort("invalid pattern (%s): %s" % (k, p))
             raise Abort("invalid pattern")
 
     def globprefix(pat):
         '''return the non-glob prefix of a path, e.g. foo/* -> foo'''
         root = []
         for p in pat.split('/'):
             if contains_glob(p): break
             root.append(p)
         return '/'.join(root) or '.'
 
     def normalizepats(names, default):
         pats = []
         roots = []
         anypats = False
         for kind, name in [patkind(p, default) for p in names]:
             if kind in ('glob', 'relpath'):
                 name = canonpath(canonroot, cwd, name)
             elif kind in ('relglob', 'path'):
                 name = normpath(name)
 
             pats.append((kind, name))
 
             if kind in ('glob', 're', 'relglob', 'relre'):
                 anypats = True
 
             if kind == 'glob':
                 root = globprefix(name)
                 roots.append(root)
             elif kind in ('relpath', 'path'):
                 roots.append(name or '.')
             elif kind == 'relglob':
                 roots.append('.')
         return roots, pats, anypats
 
     roots, pats, anypats = normalizepats(names, dflt_pat)
 
     patmatch = matchfn(pats, '$') or always
     incmatch = always
     if inc:
         dummy, inckinds, dummy = normalizepats(inc, 'glob')
         incmatch = matchfn(inckinds, '(?:/|$)')
     excmatch = lambda fn: False
     if exc:
         dummy, exckinds, dummy = normalizepats(exc, 'glob')
         excmatch = matchfn(exckinds, '(?:/|$)')
 
     if not names and inc and not exc:
         # common case: hgignore patterns
         match = incmatch
     else:
         match = lambda fn: incmatch(fn) and not excmatch(fn) and patmatch(fn)
 
     return (roots, match, (inc or exc or anypats) and True)
 
 _hgexecutable = None
 
 def hgexecutable():
     """return location of the 'hg' executable.
 
     Defaults to $HG or 'hg' in the search path.
     """
     if _hgexecutable is None:
         set_hgexecutable(os.environ.get('HG') or find_exe('hg', 'hg'))
     return _hgexecutable
 
 def set_hgexecutable(path):
     """set location of the 'hg' executable"""
     global _hgexecutable
     _hgexecutable = path
 
 def system(cmd, environ={}, cwd=None, onerr=None, errprefix=None):
     '''enhanced shell command execution.
     run with environment maybe modified, maybe in different dir.
 
     if command fails and onerr is None, return status.  if ui object,
     print error message and return status, else raise onerr object as
     exception.'''
     def py2shell(val):
         'convert python object into string that is useful to shell'
         if val in (None, False):
             return '0'
         if val == True:
             return '1'
         return str(val)
     oldenv = {}
     for k in environ:
         oldenv[k] = os.environ.get(k)
     if cwd is not None:
         oldcwd = os.getcwd()
     origcmd = cmd
     if os.name == 'nt':
         cmd = '"%s"' % cmd
     try:
         for k, v in environ.iteritems():
             os.environ[k] = py2shell(v)
         os.environ['HG'] = hgexecutable()
         if cwd is not None and oldcwd != cwd:
             os.chdir(cwd)
         rc = os.system(cmd)
         if sys.platform == 'OpenVMS' and rc & 1:
             rc = 0
         if rc and onerr:
             errmsg = '%s %s' % (os.path.basename(origcmd.split(None, 1)[0]),
                                 explain_exit(rc)[0])
             if errprefix:
                 errmsg = '%s: %s' % (errprefix, errmsg)
             try:
                 onerr.warn(errmsg + '\n')
             except AttributeError:
                 raise onerr(errmsg)
         return rc
     finally:
         for k, v in oldenv.iteritems():
             if v is None:
                 del os.environ[k]
             else:
                 os.environ[k] = v
         if cwd is not None and oldcwd != cwd:
             os.chdir(oldcwd)
 
 # os.path.lexists is not available on python2.3
 def lexists(filename):
     "test whether a file with this name exists. does not follow symlinks"
     try:
         os.lstat(filename)
     except:
         return False
     return True
 
 def rename(src, dst):
     """forcibly rename a file"""
     try:
         os.rename(src, dst)
     except OSError, err: # FIXME: check err (EEXIST ?)
         # on windows, rename to existing file is not allowed, so we
         # must delete destination first. but if file is open, unlink
         # schedules it for delete but does not delete it. rename
         # happens immediately even for open files, so we create
         # temporary file, delete it, rename destination to that name,
         # then delete that. then rename is safe to do.
         fd, temp = tempfile.mkstemp(dir=os.path.dirname(dst) or '.')
         os.close(fd)
         os.unlink(temp)
         os.rename(dst, temp)
         os.unlink(temp)
         os.rename(src, dst)
 
 def unlink(f):
     """unlink and remove the directory if it is empty"""
     os.unlink(f)
     # try removing directories that might now be empty
     try:
         os.removedirs(os.path.dirname(f))
     except OSError:
         pass
 
 def copyfile(src, dest):
     "copy a file, preserving mode"
     if os.path.islink(src):
         try:
             os.unlink(dest)
         except:
             pass
         os.symlink(os.readlink(src), dest)
     else:
         try:
             shutil.copyfile(src, dest)
             shutil.copymode(src, dest)
         except shutil.Error, inst:
             raise Abort(str(inst))
 
 def copyfiles(src, dst, hardlink=None):
     """Copy a directory tree using hardlinks if possible"""
 
     if hardlink is None:
         hardlink = (os.stat(src).st_dev ==
                     os.stat(os.path.dirname(dst)).st_dev)
 
     if os.path.isdir(src):
         os.mkdir(dst)
         for name, kind in osutil.listdir(src):
             srcname = os.path.join(src, name)
             dstname = os.path.join(dst, name)
             copyfiles(srcname, dstname, hardlink)
     else:
         if hardlink:
             try:
                 os_link(src, dst)
             except (IOError, OSError):
                 hardlink = False
                 shutil.copy(src, dst)
         else:
             shutil.copy(src, dst)
 
 class path_auditor(object):
     '''ensure that a filesystem path contains no banned components.
     the following properties of a path are checked:
 
     - under top-level .hg
     - starts at the root of a windows drive
     - contains ".."
     - traverses a symlink (e.g. a/symlink_here/b)
     - inside a nested repository'''
 
     def __init__(self, root):
         self.audited = set()
         self.auditeddir = set()
         self.root = root
 
     def __call__(self, path):
         if path in self.audited:
             return
         normpath = os.path.normcase(path)
         parts = normpath.split(os.sep)
         if (os.path.splitdrive(path)[0] or parts[0] in ('.hg', '')
             or os.pardir in parts):
             raise Abort(_("path contains illegal component: %s") % path)
         def check(prefix):
             curpath = os.path.join(self.root, prefix)
             try:
                 st = os.lstat(curpath)
             except OSError, err:
                 # EINVAL can be raised as invalid path syntax under win32.
                 # They must be ignored for patterns can be checked too.
                 if err.errno not in (errno.ENOENT, errno.EINVAL):
                     raise
             else:
                 if stat.S_ISLNK(st.st_mode):
                     raise Abort(_('path %r traverses symbolic link %r') %
                                 (path, prefix))
                 elif (stat.S_ISDIR(st.st_mode) and
                       os.path.isdir(os.path.join(curpath, '.hg'))):
                     raise Abort(_('path %r is inside repo %r') %
                                 (path, prefix))
 
         prefixes = []
         for c in strutil.rfindall(normpath, os.sep):
             prefix = normpath[:c]
             if prefix in self.auditeddir:
                 break
             check(prefix)
             prefixes.append(prefix)
 
         self.audited.add(path)
         # only add prefixes to the cache after checking everything: we don't
         # want to add "foo/bar/baz" before checking if there's a "foo/.hg"
         self.auditeddir.update(prefixes)
 
 def _makelock_file(info, pathname):
     ld = os.open(pathname, os.O_CREAT | os.O_WRONLY | os.O_EXCL)
     os.write(ld, info)
     os.close(ld)
 
 def _readlock_file(pathname):
     return posixfile(pathname).read()
 
 def nlinks(pathname):
     """Return number of hardlinks for the given file."""
     return os.lstat(pathname).st_nlink
 
 if hasattr(os, 'link'):
     os_link = os.link
 else:
     def os_link(src, dst):
         raise OSError(0, _("Hardlinks not supported"))
 
 def fstat(fp):
     '''stat file object that may not have fileno method.'''
     try:
         return os.fstat(fp.fileno())
     except AttributeError:
         return os.stat(fp.name)
 
 posixfile = file
 
 def is_win_9x():
     '''return true if run on windows 95, 98 or me.'''
     try:
         return sys.getwindowsversion()[3] == 1
     except AttributeError:
         return os.name == 'nt' and 'command' in os.environ.get('comspec', '')
 
 getuser_fallback = None
 
 def getuser():
     '''return name of current user'''
     try:
         return getpass.getuser()
     except ImportError:
         # import of pwd will fail on windows - try fallback
         if getuser_fallback:
             return getuser_fallback()
     # raised if win32api not available
     raise Abort(_('user name not available - set USERNAME '
                   'environment variable'))
 
 def username(uid=None):
     """Return the name of the user with the given uid.
 
     If uid is None, return the name of the current user."""
     try:
         import pwd
         if uid is None:
             uid = os.getuid()
         try:
             return pwd.getpwuid(uid)[0]
         except KeyError:
             return str(uid)
     except ImportError:
         return None
 
 def groupname(gid=None):
     """Return the name of the group with the given gid.
 
     If gid is None, return the name of the current group."""
     try:
         import grp
         if gid is None:
             gid = os.getgid()
         try:
             return grp.getgrgid(gid)[0]
         except KeyError:
             return str(gid)
     except ImportError:
         return None
 
 # File system features
 
 def checkfolding(path):
     """
     Check whether the given path is on a case-sensitive filesystem
 
     Requires a path (like /foo/.hg) ending with a foldable final
     directory component.
     """
     s1 = os.stat(path)
     d, b = os.path.split(path)
     p2 = os.path.join(d, b.upper())
     if path == p2:
         p2 = os.path.join(d, b.lower())
     try:
         s2 = os.stat(p2)
         if s2 == s1:
             return False
         return True
     except:
         return True
 
 def checkexec(path):
     """
     Check whether the given path is on a filesystem with UNIX-like exec flags
 
     Requires a directory (like /foo/.hg)
     """
     try:
         EXECFLAGS = stat.S_IXUSR | stat.S_IXGRP | stat.S_IXOTH
         fh, fn = tempfile.mkstemp("", "", path)
         os.close(fh)
         m = os.stat(fn).st_mode
         # VFAT on Linux can flip mode but it doesn't persist a FS remount.
         # frequently we can detect it if files are created with exec bit on.
         new_file_has_exec = m & EXECFLAGS
         os.chmod(fn, m ^ EXECFLAGS)
         exec_flags_cannot_flip = (os.stat(fn).st_mode == m)
         os.unlink(fn)
     except (IOError,OSError):
         # we don't care, the user probably won't be able to commit anyway
         return False
     return not (new_file_has_exec or exec_flags_cannot_flip)
 
 def execfunc(path, fallback):
     '''return an is_exec() function with default to fallback'''
     if checkexec(path):
         return lambda x: is_exec(os.path.join(path, x))
     return fallback
 
 def checklink(path):
     """check whether the given path is on a symlink-capable filesystem"""
     # mktemp is not racy because symlink creation will fail if the
     # file already exists
     name = tempfile.mktemp(dir=path)
     try:
         os.symlink(".", name)
         os.unlink(name)
         return True
     except (OSError, AttributeError):
         return False
 
 def linkfunc(path, fallback):
     '''return an is_link() function with default to fallback'''
     if checklink(path):
         return lambda x: os.path.islink(os.path.join(path, x))
     return fallback
 
 _umask = os.umask(0)
 os.umask(_umask)
 
 def needbinarypatch():
     """return True if patches should be applied in binary mode by default."""
     return os.name == 'nt'
 
 # Platform specific variants
 if os.name == 'nt':
     import msvcrt
     nulldev = 'NUL:'
 
     class winstdout:
         '''stdout on windows misbehaves if sent through a pipe'''
 
         def __init__(self, fp):
             self.fp = fp
 
         def __getattr__(self, key):
             return getattr(self.fp, key)
 
         def close(self):
             try:
                 self.fp.close()
             except: pass
 
         def write(self, s):
             try:
                 return self.fp.write(s)
             except IOError, inst:
                 if inst.errno != 0: raise
                 self.close()
                 raise IOError(errno.EPIPE, 'Broken pipe')
 
         def flush(self):
             try:
                 return self.fp.flush()
             except IOError, inst:
                 if inst.errno != errno.EINVAL: raise
                 self.close()
                 raise IOError(errno.EPIPE, 'Broken pipe')
 
     sys.stdout = winstdout(sys.stdout)
 
     def system_rcpath():
         try:
             return system_rcpath_win32()
         except:
             return [r'c:\mercurial\mercurial.ini']
 
     def user_rcpath():
         '''return os-specific hgrc search path to the user dir'''
         try:
             userrc = user_rcpath_win32()
         except:
             userrc = os.path.join(os.path.expanduser('~'), 'mercurial.ini')
         path = [userrc]
         userprofile = os.environ.get('USERPROFILE')
         if userprofile:
             path.append(os.path.join(userprofile, 'mercurial.ini'))
         return path
 
     def parse_patch_output(output_line):
         """parses the output produced by patch and returns the file name"""
         pf = output_line[14:]
         if pf[0] == '`':
             pf = pf[1:-1] # Remove the quotes
         return pf
 
     def testpid(pid):
         '''return False if pid dead, True if running or not known'''
         return True
 
     def set_exec(f, mode):
         pass
 
     def set_link(f, mode):
         pass
 
     def set_binary(fd):
         msvcrt.setmode(fd.fileno(), os.O_BINARY)
 
     def pconvert(path):
         return path.replace("\\", "/")
 
     def localpath(path):
         return path.replace('/', '\\')
 
     def normpath(path):
         return pconvert(os.path.normpath(path))
 
     makelock = _makelock_file
     readlock = _readlock_file
 
     def samestat(s1, s2):
         return False
 
     # A sequence of backslashes is special iff it precedes a double quote:
     # - if there's an even number of backslashes, the double quote is not
     #   quoted (i.e. it ends the quoted region)
     # - if there's an odd number of backslashes, the double quote is quoted
     # - in both cases, every pair of backslashes is unquoted into a single
     #   backslash
     # (See http://msdn2.microsoft.com/en-us/library/a1y7w461.aspx )
     # So, to quote a string, we must surround it in double quotes, double
     # the number of backslashes that preceed double quotes and add another
     # backslash before every double quote (being careful with the double
     # quote we've appended to the end)
     _quotere = None
     def shellquote(s):
         global _quotere
         if _quotere is None:
             _quotere = re.compile(r'(\\*)("|\\$)')
         return '"%s"' % _quotere.sub(r'\1\1\\\2', s)
 
     def quotecommand(cmd):
         """Build a command string suitable for os.popen* calls."""
         # The extra quotes are needed because popen* runs the command
         # through the current COMSPEC. cmd.exe suppress enclosing quotes.
         return '"' + cmd + '"'
 
     def explain_exit(code):
         return _("exited with status %d") % code, code
 
     # if you change this stub into a real check, please try to implement the
     # username and groupname functions above, too.
     def isowner(fp, st=None):
         return True
 
     def find_in_path(name, path, default=None):
         '''find name in search path. path can be string (will be split
         with os.pathsep), or iterable thing that returns strings.  if name
         found, return path to name. else return default. name is looked up
         using cmd.exe rules, using PATHEXT.'''
         if isinstance(path, str):
             path = path.split(os.pathsep)
 
         pathext = os.environ.get('PATHEXT', '.COM;.EXE;.BAT;.CMD')
         pathext = pathext.lower().split(os.pathsep)
         isexec = os.path.splitext(name)[1].lower() in pathext
 
         for p in path:
             p_name = os.path.join(p, name)
 
             if isexec and os.path.exists(p_name):
                 return p_name
 
             for ext in pathext:
                 p_name_ext = p_name + ext
                 if os.path.exists(p_name_ext):
                     return p_name_ext
         return default
 
     def set_signal_handler():
         try:
             set_signal_handler_win32()
         except NameError:
             pass
 
     try:
         # override functions with win32 versions if possible
         from util_win32 import *
         if not is_win_9x():
             posixfile = posixfile_nt
     except ImportError:
         pass
 
 else:
     nulldev = '/dev/null'
 
     def rcfiles(path):
         rcs = [os.path.join(path, 'hgrc')]
         rcdir = os.path.join(path, 'hgrc.d')
         try:
             rcs.extend([os.path.join(rcdir, f)
                         for f, kind in osutil.listdir(rcdir)
                         if f.endswith(".rc")])
         except OSError:
             pass
         return rcs
 
     def system_rcpath():
         path = []
         # old mod_python does not set sys.argv
         if len(getattr(sys, 'argv', [])) > 0:
             path.extend(rcfiles(os.path.dirname(sys.argv[0]) +
                                   '/../etc/mercurial'))
         path.extend(rcfiles('/etc/mercurial'))
         return path
 
     def user_rcpath():
         return [os.path.expanduser('~/.hgrc')]
 
     def parse_patch_output(output_line):
         """parses the output produced by patch and returns the file name"""
         pf = output_line[14:]
         if os.sys.platform == 'OpenVMS':
             if pf[0] == '`':
                 pf = pf[1:-1] # Remove the quotes
         else:
            if pf.startswith("'") and pf.endswith("'") and " " in pf:
                 pf = pf[1:-1] # Remove the quotes
         return pf
 
     def is_exec(f):
         """check whether a file is executable"""
         return (os.lstat(f).st_mode & 0100 != 0)
 
     def set_exec(f, mode):
         s = os.lstat(f).st_mode
         if stat.S_ISLNK(s) or (s & 0100 != 0) == mode:
             return
         if mode:
             # Turn on +x for every +r bit when making a file executable
             # and obey umask.
             os.chmod(f, s | (s & 0444) >> 2 & ~_umask)
         else:
             os.chmod(f, s & 0666)
 
     def set_link(f, mode):
         """make a file a symbolic link/regular file
 
         if a file is changed to a link, its contents become the link data
         if a link is changed to a file, its link data become its contents
         """
 
         m = os.path.islink(f)
         if m == bool(mode):
             return
 
         if mode: # switch file to link
             data = file(f).read()
             os.unlink(f)
             os.symlink(data, f)
         else:
             data = os.readlink(f)
             os.unlink(f)
             file(f, "w").write(data)
 
     def set_binary(fd):
         pass
 
     def pconvert(path):
         return path
 
     def localpath(path):
         return path
 
     normpath = os.path.normpath
     samestat = os.path.samestat
 
     def makelock(info, pathname):
         try:
             os.symlink(info, pathname)
         except OSError, why:
             if why.errno == errno.EEXIST:
                 raise
             else:
                 _makelock_file(info, pathname)
 
     def readlock(pathname):
         try:
             return os.readlink(pathname)
         except OSError, why:
             if why.errno in (errno.EINVAL, errno.ENOSYS):
                 return _readlock_file(pathname)
             else:
                 raise
 
     def shellquote(s):
         if os.sys.platform == 'OpenVMS':
             return '"%s"' % s
         else:
             return "'%s'" % s.replace("'", "'\\''")
 
     def quotecommand(cmd):
         return cmd
 
     def testpid(pid):
         '''return False if pid dead, True if running or not sure'''
         if os.sys.platform == 'OpenVMS':
             return True
         try:
             os.kill(pid, 0)
             return True
         except OSError, inst:
             return inst.errno != errno.ESRCH
 
     def explain_exit(code):
         """return a 2-tuple (desc, code) describing a process's status"""
         if os.WIFEXITED(code):
             val = os.WEXITSTATUS(code)
             return _("exited with status %d") % val, val
         elif os.WIFSIGNALED(code):
             val = os.WTERMSIG(code)
             return _("killed by signal %d") % val, val
         elif os.WIFSTOPPED(code):
             val = os.WSTOPSIG(code)
             return _("stopped by signal %d") % val, val
         raise ValueError(_("invalid exit code"))
 
     def isowner(fp, st=None):
         """Return True if the file object f belongs to the current user.
 
         The return value of a util.fstat(f) may be passed as the st argument.
         """
         if st is None:
             st = fstat(fp)
         return st.st_uid == os.getuid()
 
     def find_in_path(name, path, default=None):
         '''find name in search path. path can be string (will be split
         with os.pathsep), or iterable thing that returns strings.  if name
         found, return path to name. else return default.'''
         if isinstance(path, str):
             path = path.split(os.pathsep)
         for p in path:
             p_name = os.path.join(p, name)
             if os.path.exists(p_name):
                 return p_name
         return default
 
     def set_signal_handler():
         pass
 
 def find_exe(name, default=None):
     '''find path of an executable.
     if name contains a path component, return it as is.  otherwise,
     use normal executable search path.'''
 
     if os.sep in name or sys.platform == 'OpenVMS':
         # don't check the executable bit.  if the file isn't
         # executable, whoever tries to actually run it will give a
         # much more useful error message.
         return name
     return find_in_path(name, os.environ.get('PATH', ''), default=default)
 
 def _buildencodefun():
     e = '_'
     win_reserved = [ord(x) for x in '\\:*?"<>|']
     cmap = dict([ (chr(x), chr(x)) for x in xrange(127) ])
     for x in (range(32) + range(126, 256) + win_reserved):
         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))))
 
 encodefilename, decodefilename = _buildencodefun()
 
 def encodedopener(openerfn, fn):
     def o(path, *args, **kw):
         return openerfn(fn(path), *args, **kw)
     return o
 
 def mktempcopy(name, emptyok=False):
     """Create a temporary file with the same contents from name
 
     The permission bits are copied from the original file.
 
     If the temporary file is going to be truncated immediately, you
     can use emptyok=True as an optimization.
 
     Returns the name of the temporary file.
     """
     d, fn = os.path.split(name)
     fd, temp = tempfile.mkstemp(prefix='.%s-' % fn, dir=d)
     os.close(fd)
     # Temporary files are created with mode 0600, which is usually not
     # what we want.  If the original file already exists, just copy
     # its mode.  Otherwise, manually obey umask.
     try:
         st_mode = os.lstat(name).st_mode
     except OSError, inst:
         if inst.errno != errno.ENOENT:
             raise
         st_mode = 0666 & ~_umask
     os.chmod(temp, st_mode)
     if emptyok:
         return temp
     try:
         try:
             ifp = posixfile(name, "rb")
         except IOError, inst:
             if inst.errno == errno.ENOENT:
                 return temp
             if not getattr(inst, 'filename', None):
                 inst.filename = name
             raise
         ofp = posixfile(temp, "wb")
         for chunk in filechunkiter(ifp):
             ofp.write(chunk)
         ifp.close()
         ofp.close()
     except:
         try: os.unlink(temp)
         except: pass
         raise
     return temp
 
 class atomictempfile(posixfile):
     """file-like object that atomically updates a file
 
     All writes will be redirected to a temporary copy of the original
     file.  When rename is called, the copy is renamed to the original
     name, making the changes visible.
     """
     def __init__(self, name, mode):
         self.__name = name
         self.temp = mktempcopy(name, emptyok=('w' in mode))
         posixfile.__init__(self, self.temp, mode)
 
     def rename(self):
         if not self.closed:
             posixfile.close(self)
             rename(self.temp, localpath(self.__name))
 
     def __del__(self):
         if not self.closed:
             try:
                 os.unlink(self.temp)
             except: pass
             posixfile.close(self)
 
 class opener(object):
     """Open files relative to a base directory
 
     This class is used to hide the details of COW semantics and
     remote file access from higher level code.
     """
     def __init__(self, base, audit=True):
         self.base = base
         if audit:
             self.audit_path = path_auditor(base)
         else:
             self.audit_path = always
 
     def __getattr__(self, name):
         if name == '_can_symlink':
             self._can_symlink = checklink(self.base)
             return self._can_symlink
         raise AttributeError(name)
 
     def __call__(self, path, mode="r", text=False, atomictemp=False):
         self.audit_path(path)
         f = os.path.join(self.base, path)
 
         if not text and "b" not in mode:
             mode += "b" # for that other OS
 
         if mode[0] != "r":
             try:
                 nlink = nlinks(f)
             except OSError:
                 nlink = 0
                 d = os.path.dirname(f)
                 if not os.path.isdir(d):
                     os.makedirs(d)
             if atomictemp:
                 return atomictempfile(f, mode)
             if nlink > 1:
                 rename(mktempcopy(f), f)
         return posixfile(f, mode)
 
     def symlink(self, src, dst):
         self.audit_path(dst)
         linkname = os.path.join(self.base, dst)
         try:
             os.unlink(linkname)
         except OSError:
             pass
 
         dirname = os.path.dirname(linkname)
         if not os.path.exists(dirname):
             os.makedirs(dirname)
 
         if self._can_symlink:
             try:
                 os.symlink(src, linkname)
             except OSError, err:
                 raise OSError(err.errno, _('could not symlink to %r: %s') %
                               (src, err.strerror), linkname)
         else:
             f = self(dst, "w")
             f.write(src)
             f.close()
 
 class chunkbuffer(object):
     """Allow arbitrary sized chunks of data to be efficiently read from an
     iterator over chunks of arbitrary size."""
 
     def __init__(self, in_iter):
         """in_iter is the iterator that's iterating over the input chunks.
         targetsize is how big a buffer to try to maintain."""
         self.iter = iter(in_iter)
         self.buf = ''
         self.targetsize = 2**16
 
     def read(self, l):
         """Read L bytes of data from the iterator of chunks of data.
         Returns less than L bytes if the iterator runs dry."""
         if l > len(self.buf) and self.iter:
             # Clamp to a multiple of self.targetsize
             targetsize = max(l, self.targetsize)
             collector = cStringIO.StringIO()
             collector.write(self.buf)
             collected = len(self.buf)
             for chunk in self.iter:
                 collector.write(chunk)
                 collected += len(chunk)
                 if collected >= targetsize:
                     break
             if collected < targetsize:
                 self.iter = False
             self.buf = collector.getvalue()
         if len(self.buf) == l:
-            s, self.buf = self.buf, ''
+            s, self.buf = str(self.buf), ''
         else:
             s, self.buf = self.buf[:l], buffer(self.buf, l)
         return s
 
 def filechunkiter(f, size=65536, limit=None):
     """Create a generator that produces the data in the file size
     (default 65536) bytes at a time, up to optional limit (default is
     to read all data).  Chunks may be less than size bytes if the
     chunk is the last chunk in the file, or the file is a socket or
     some other type of file that sometimes reads less data than is
     requested."""
     assert size >= 0
     assert limit is None or limit >= 0
     while True:
         if limit is None: nbytes = size
         else: nbytes = min(limit, size)
         s = nbytes and f.read(nbytes)
         if not s: break
         if limit: limit -= len(s)
         yield s
 
 def makedate():
     lt = time.localtime()
     if lt[8] == 1 and time.daylight:
         tz = time.altzone
     else:
         tz = time.timezone
     return time.mktime(lt), tz
 
 def datestr(date=None, format='%a %b %d %H:%M:%S %Y', timezone=True, timezone_format=" %+03d%02d"):
     """represent a (unixtime, offset) tuple as a localized time.
     unixtime is seconds since the epoch, and offset is the time zone's
     number of seconds away from UTC. if timezone is false, do not
     append time zone to string."""
     t, tz = date or makedate()
     s = time.strftime(format, time.gmtime(float(t) - tz))
     if timezone:
         s += timezone_format % (-tz / 3600, ((-tz % 3600) / 60))
     return s
 
 def strdate(string, format, defaults=[]):
     """parse a localized time string and return a (unixtime, offset) tuple.
     if the string cannot be parsed, ValueError is raised."""
     def timezone(string):
         tz = string.split()[-1]
         if tz[0] in "+-" and len(tz) == 5 and tz[1:].isdigit():
             tz = int(tz)
             offset = - 3600 * (tz / 100) - 60 * (tz % 100)
             return offset
         if tz == "GMT" or tz == "UTC":
             return 0
         return None
 
     # NOTE: unixtime = localunixtime + offset
     offset, date = timezone(string), string
     if offset != None:
         date = " ".join(string.split()[:-1])
 
     # add missing elements from defaults
     for part in defaults:
         found = [True for p in part if ("%"+p) in format]
         if not found:
             date += "@" + defaults[part]
             format += "@%" + part[0]
 
     timetuple = time.strptime(date, format)
     localunixtime = int(calendar.timegm(timetuple))
     if offset is None:
         # local timezone
         unixtime = int(time.mktime(timetuple))
         offset = unixtime - localunixtime
     else:
         unixtime = localunixtime + offset
     return unixtime, offset
 
 def parsedate(string, formats=None, defaults=None):
     """parse a localized time string and return a (unixtime, offset) tuple.
     The date may be a "unixtime offset" string or in one of the specified
     formats."""
     if not string:
         return 0, 0
     if not formats:
         formats = defaultdateformats
     string = string.strip()
     try:
         when, offset = map(int, string.split(' '))
     except ValueError:
         # fill out defaults
         if not defaults:
             defaults = {}
         now = makedate()
         for part in "d mb yY HI M S".split():
             if part not in defaults:
                 if part[0] in "HMS":
                     defaults[part] = "00"
                 elif part[0] in "dm":
                     defaults[part] = "1"
                 else:
                     defaults[part] = datestr(now, "%" + part[0], False)
 
         for format in formats:
             try:
                 when, offset = strdate(string, format, defaults)
             except ValueError:
                 pass
             else:
                 break
         else:
             raise Abort(_('invalid date: %r ') % string)
     # validate explicit (probably user-specified) date and
     # time zone offset. values must fit in signed 32 bits for
     # current 32-bit linux runtimes. timezones go from UTC-12
     # to UTC+14
     if abs(when) > 0x7fffffff:
         raise Abort(_('date exceeds 32 bits: %d') % when)
     if offset < -50400 or offset > 43200:
         raise Abort(_('impossible time zone offset: %d') % offset)
     return when, offset
 
 def matchdate(date):
     """Return a function that matches a given date match specifier
 
     Formats include:
 
     '{date}' match a given date to the accuracy provided
 
     '<{date}' on or before a given date
 
     '>{date}' on or after a given date
 
     """
 
     def lower(date):
         return parsedate(date, extendeddateformats)[0]
 
     def upper(date):
         d = dict(mb="12", HI="23", M="59", S="59")
         for days in "31 30 29".split():
             try:
                 d["d"] = days
                 return parsedate(date, extendeddateformats, d)[0]
             except:
                 pass
         d["d"] = "28"
         return parsedate(date, extendeddateformats, d)[0]
 
     if date[0] == "<":
         when = upper(date[1:])
         return lambda x: x <= when
     elif date[0] == ">":
         when = lower(date[1:])
         return lambda x: x >= when
     elif date[0] == "-":
         try:
             days = int(date[1:])
         except ValueError:
             raise Abort(_("invalid day spec: %s") % date[1:])
         when = makedate()[0] - days * 3600 * 24
         return lambda x: x >= when
     elif " to " in date:
         a, b = date.split(" to ")
         start, stop = lower(a), upper(b)
         return lambda x: x >= start and x <= stop
     else:
         start, stop = lower(date), upper(date)
         return lambda x: x >= start and x <= stop
 
 def shortuser(user):
     """Return a short representation of a user name or email address."""
     f = user.find('@')
     if f >= 0:
         user = user[:f]
     f = user.find('<')
     if f >= 0:
         user = user[f+1:]
     f = user.find(' ')
     if f >= 0:
         user = user[:f]
     f = user.find('.')
     if f >= 0:
         user = user[:f]
     return user
 
 def ellipsis(text, maxlength=400):
     """Trim string to at most maxlength (default: 400) characters."""
     if len(text) <= maxlength:
         return text
     else:
         return "%s..." % (text[:maxlength-3])
 
 def walkrepos(path):
     '''yield every hg repository under path, recursively.'''
     def errhandler(err):
         if err.filename == path:
             raise err
 
     for root, dirs, files in os.walk(path, onerror=errhandler):
         for d in dirs:
             if d == '.hg':
                 yield root
                 dirs[:] = []
                 break
 
 _rcpath = None
 
 def os_rcpath():
     '''return default os-specific hgrc search path'''
     path = system_rcpath()
     path.extend(user_rcpath())
     path = [os.path.normpath(f) for f in path]
     return path
 
 def rcpath():
     '''return hgrc search path. if env var HGRCPATH is set, use it.
     for each item in path, if directory, use files ending in .rc,
     else use item.
     make HGRCPATH empty to only look in .hg/hgrc of current repo.
     if no HGRCPATH, use default os-specific path.'''
     global _rcpath
     if _rcpath is None:
         if 'HGRCPATH' in os.environ:
             _rcpath = []
             for p in os.environ['HGRCPATH'].split(os.pathsep):
                 if not p: continue
                 if os.path.isdir(p):
                     for f, kind in osutil.listdir(p):
                         if f.endswith('.rc'):
                             _rcpath.append(os.path.join(p, f))
                 else:
                     _rcpath.append(p)
         else:
             _rcpath = os_rcpath()
     return _rcpath
 
 def bytecount(nbytes):
     '''return byte count formatted as readable string, with units'''
 
     units = (
         (100, 1<<30, _('%.0f GB')),
         (10, 1<<30, _('%.1f GB')),
         (1, 1<<30, _('%.2f GB')),
         (100, 1<<20, _('%.0f MB')),
         (10, 1<<20, _('%.1f MB')),
         (1, 1<<20, _('%.2f MB')),
         (100, 1<<10, _('%.0f KB')),
         (10, 1<<10, _('%.1f KB')),
         (1, 1<<10, _('%.2f KB')),
         (1, 1, _('%.0f bytes')),
         )
 
     for multiplier, divisor, format in units:
         if nbytes >= divisor * multiplier:
             return format % (nbytes / float(divisor))
     return units[-1][2] % nbytes
 
 def drop_scheme(scheme, path):
     sc = scheme + ':'
     if path.startswith(sc):
         path = path[len(sc):]
         if path.startswith('//'):
             path = path[2:]
     return path
 
 def uirepr(s):
     # Avoid double backslash in Windows path repr()
     return repr(s).replace('\\\\', '\\')