merge with crew-stable

hgext/convert/cvs.py

@@ -1,291 +1,312 b''
 # CVS conversion code inspired by hg-cvs-import and git-cvsimport
 
 import os, locale, re, socket
 from cStringIO import StringIO
 from mercurial import util
 
 from common import NoRepo, commit, converter_source, checktool
 
 class convert_cvs(converter_source):
     def __init__(self, ui, path, rev=None):
         super(convert_cvs, self).__init__(ui, path, rev=rev)
 
         cvs = os.path.join(path, "CVS")
         if not os.path.exists(cvs):
             raise NoRepo("%s does not look like a CVS checkout" % path)
 
         for tool in ('cvsps', 'cvs'):
             checktool(tool)
 
         self.changeset = {}
         self.files = {}
         self.tags = {}
         self.lastbranch = {}
         self.parent = {}
         self.socket = None
         self.cvsroot = file(os.path.join(cvs, "Root")).read()[:-1]
         self.cvsrepo = file(os.path.join(cvs, "Repository")).read()[:-1]
         self.encoding = locale.getpreferredencoding()
         self._parse()
         self._connect()
 
     def _parse(self):
         if self.changeset:
             return
 
         maxrev = 0
         cmd = 'cvsps -A -u --cvs-direct -q'
         if self.rev:
             # TODO: handle tags
             try:
                 # patchset number?
                 maxrev = int(self.rev)
             except ValueError:
                 try:
                     # date
                     util.parsedate(self.rev, ['%Y/%m/%d %H:%M:%S'])
                     cmd = '%s -d "1970/01/01 00:00:01" -d "%s"' % (cmd, self.rev)
                 except util.Abort:
                     raise util.Abort('revision %s is not a patchset number or date' % self.rev)
 
         d = os.getcwd()
         try:
             os.chdir(self.path)
             id = None
             state = 0
+            filerevids = {}
             for l in util.popen(cmd):
                 if state == 0: # header
                     if l.startswith("PatchSet"):
                         id = l[9:-2]
                         if maxrev and int(id) > maxrev:
+                            # ignore everything
                             state = 3
                     elif l.startswith("Date"):
                         date = util.parsedate(l[6:-1], ["%Y/%m/%d %H:%M:%S"])
                         date = util.datestr(date)
                     elif l.startswith("Branch"):
                         branch = l[8:-1]
                         self.parent[id] = self.lastbranch.get(branch, 'bad')
                         self.lastbranch[branch] = id
                     elif l.startswith("Ancestor branch"):
                         ancestor = l[17:-1]
-                        self.parent[id] = self.lastbranch[ancestor]
+                        # figure out the parent later
+                        self.parent[id] = None
                     elif l.startswith("Author"):
                         author = self.recode(l[8:-1])
                     elif l.startswith("Tag:") or l.startswith("Tags:"):
                         t = l[l.index(':')+1:]
                         t = [ut.strip() for ut in t.split(',')]
                         if (len(t) > 1) or (t[0] and (t[0] != "(none)")):
                             self.tags.update(dict.fromkeys(t, id))
                     elif l.startswith("Log:"):
+                        # switch to gathering log
                         state = 1
                         log = ""
                 elif state == 1: # log
                     if l == "Members: \n":
+                        # switch to gathering members
                         files = {}
+                        oldrevs = []
                         log = self.recode(log[:-1])
                         state = 2
                     else:
+                        # gather log
                         log += l
-                elif state == 2:
-                    if l == "\n": #
+                elif state == 2: # members
+                    if l == "\n": # start of next entry
                         state = 0
                         p = [self.parent[id]]
                         if id == "1":
                             p = []
                         if branch == "HEAD":
                             branch = ""
+                        if branch and p[0] == None:
+                            latest = None
+                            # the last changeset that contains a base
+                            # file is our parent
+                            for r in oldrevs:
+                                latest = max(filerevids[r], latest)
+                            p = [latest]
+
+                        # add current commit to set
                         c = commit(author=author, date=date, parents=p,
                                    desc=log, branch=branch)
                         self.changeset[id] = c
                         self.files[id] = files
                     else:
                         colon = l.rfind(':')
                         file = l[1:colon]
                         rev = l[colon+1:-2]
-                        rev = rev.split("->")[1]
+                        oldrev, rev = rev.split("->")
                         files[file] = rev
+
+                        # save some information for identifying branch points
+                        oldrevs.append("%s:%s" % (oldrev, file))
+                        filerevids["%s:%s" % (rev, file)] = id
                 elif state == 3:
+                    # swallow all input
                     continue
 
             self.heads = self.lastbranch.values()
         finally:
             os.chdir(d)
 
     def _connect(self):
         root = self.cvsroot
         conntype = None
         user, host = None, None
         cmd = ['cvs', 'server']
 
         self.ui.status("connecting to %s\n" % root)
 
         if root.startswith(":pserver:"):
             root = root[9:]
             m = re.match(r'(?:(.*?)(?::(.*?))?@)?([^:\/]*)(?::(\d*))?(.*)',
                          root)
             if m:
                 conntype = "pserver"
                 user, passw, serv, port, root = m.groups()
                 if not user:
                     user = "anonymous"
                 if not port:
                     port = 2401
                 else:
                     port = int(port)
                 format0 = ":pserver:%s@%s:%s" % (user, serv, root)
                 format1 = ":pserver:%s@%s:%d%s" % (user, serv, port, root)
 
                 if not passw:
                     passw = "A"
                     pf = open(os.path.join(os.environ["HOME"], ".cvspass"))
                     for line in pf.read().splitlines():
                         part1, part2 = line.split(' ', 1)
                         if part1 == '/1':
                             # /1 :pserver:[e-mail]:2401/cvsroot/foo Ah<Z
                             part1, part2 = part2.split(' ', 1)
                             format = format1
                         else:
                             # :pserver:[e-mail]:/cvsroot/foo Ah<Z
                             format = format0
                         if part1 == format:
                             passw = part2
                             break
                     pf.close()
 
                 sck = socket.socket()
                 sck.connect((serv, port))
                 sck.send("\n".join(["BEGIN AUTH REQUEST", root, user, passw,
                                     "END AUTH REQUEST", ""]))
                 if sck.recv(128) != "I LOVE YOU\n":
                     raise util.Abort("CVS pserver authentication failed")
 
                 self.writep = self.readp = sck.makefile('r+')
 
         if not conntype and root.startswith(":local:"):
             conntype = "local"
             root = root[7:]
 
         if not conntype:
             # :ext:user@host/home/user/path/to/cvsroot
             if root.startswith(":ext:"):
                 root = root[5:]
             m = re.match(r'(?:([^@:/]+)@)?([^:/]+):?(.*)', root)
             # Do not take Windows path "c:\foo\bar" for a connection strings
             if os.path.isdir(root) or not m:
                 conntype = "local"
             else:
                 conntype = "rsh"
                 user, host, root = m.group(1), m.group(2), m.group(3)
 
         if conntype != "pserver":
             if conntype == "rsh":
                 rsh = os.environ.get("CVS_RSH") or "ssh"
                 if user:
                     cmd = [rsh, '-l', user, host] + cmd
                 else:
                     cmd = [rsh, host] + cmd
 
             # popen2 does not support argument lists under Windows
             cmd = [util.shellquote(arg) for arg in cmd]
             cmd = util.quotecommand(' '.join(cmd))
             self.writep, self.readp = os.popen2(cmd, 'b')
 
         self.realroot = root
 
         self.writep.write("Root %s\n" % root)
         self.writep.write("Valid-responses ok error Valid-requests Mode"
                           " M Mbinary E Checked-in Created Updated"
                           " Merged Removed\n")
         self.writep.write("valid-requests\n")
         self.writep.flush()
         r = self.readp.readline()
         if not r.startswith("Valid-requests"):
             raise util.Abort("server sucks")
         if "UseUnchanged" in r:
             self.writep.write("UseUnchanged\n")
             self.writep.flush()
             r = self.readp.readline()
 
     def getheads(self):
         return self.heads
 
     def _getfile(self, name, rev):
 
         def chunkedread(fp, count):
             # file-objects returned by socked.makefile() do not handle
             # large read() requests very well.
             chunksize = 65536
             output = StringIO()
             while count > 0:
                 data = fp.read(min(count, chunksize))
                 if not data:
                     raise util.Abort("%d bytes missing from remote file" % count)
                 count -= len(data)
                 output.write(data)
             return output.getvalue()
 
         if rev.endswith("(DEAD)"):
             raise IOError
 
         args = ("-N -P -kk -r %s --" % rev).split()
         args.append(self.cvsrepo + '/' + name)
         for x in args:
             self.writep.write("Argument %s\n" % x)
         self.writep.write("Directory .\n%s\nco\n" % self.realroot)
         self.writep.flush()
 
         data = ""
         while 1:
             line = self.readp.readline()
             if line.startswith("Created ") or line.startswith("Updated "):
                 self.readp.readline() # path
                 self.readp.readline() # entries
                 mode = self.readp.readline()[:-1]
                 count = int(self.readp.readline()[:-1])
                 data = chunkedread(self.readp, count)
             elif line.startswith(" "):
                 data += line[1:]
             elif line.startswith("M "):
                 pass
             elif line.startswith("Mbinary "):
                 count = int(self.readp.readline()[:-1])
                 data = chunkedread(self.readp, count)
             else:
                 if line == "ok\n":
                     return (data, "x" in mode and "x" or "")
                 elif line.startswith("E "):
                     self.ui.warn("cvs server: %s\n" % line[2:])
                 elif line.startswith("Remove"):
                     l = self.readp.readline()
                     l = self.readp.readline()
                     if l != "ok\n":
                         raise util.Abort("unknown CVS response: %s" % l)
                 else:
                     raise util.Abort("unknown CVS response: %s" % line)
 
     def getfile(self, file, rev):
         data, mode = self._getfile(file, rev)
         self.modecache[(file, rev)] = mode
         return data
 
     def getmode(self, file, rev):
         return self.modecache[(file, rev)]
 
     def getchanges(self, rev):
         self.modecache = {}
         files = self.files[rev]
         cl = files.items()
         cl.sort()
         return (cl, {})
 
     def getcommit(self, rev):
         return self.changeset[rev]
 
     def gettags(self):
         return self.tags
 
     def getchangedfiles(self, rev, i):
         files = self.files[rev].keys()
         files.sort()
         return files
 

hgext/convert/hg.py

@@ -1,284 +1,288 b''
 # hg backend for convert extension
 
 # Notes for hg->hg conversion:
 #
 # * Old versions of Mercurial didn't trim the whitespace from the ends
 #   of commit messages, but new versions do.  Changesets created by
 #   those older versions, then converted, may thus have different
 #   hashes for changesets that are otherwise identical.
 #
 # * By default, the source revision is stored in the converted
 #   revision.  This will cause the converted revision to have a
 #   different identity than the source.  To avoid this, use the
 #   following option: "--config convert.hg.saverev=false"
 
 
 import os, time
 from mercurial.i18n import _
 from mercurial.node import *
 from mercurial import hg, lock, revlog, util
 
 from common import NoRepo, commit, converter_source, converter_sink
 
 class mercurial_sink(converter_sink):
     def __init__(self, ui, path):
         converter_sink.__init__(self, ui, path)
         self.branchnames = ui.configbool('convert', 'hg.usebranchnames', True)
         self.clonebranches = ui.configbool('convert', 'hg.clonebranches', False)
         self.tagsbranch = ui.config('convert', 'hg.tagsbranch', 'default')
         self.lastbranch = None
         if os.path.isdir(path) and len(os.listdir(path)) > 0:
             try:
                 self.repo = hg.repository(self.ui, path)
+                if not self.repo.local():
+                    raise NoRepo(_('%s is not a local Mercurial repo') % path)
             except hg.RepoError, err:
                 ui.print_exc()
                 raise NoRepo(err.args[0])
         else:
             try:
                 ui.status(_('initializing destination %s repository\n') % path)
                 self.repo = hg.repository(self.ui, path, create=True)
+                if not self.repo.local():
+                    raise NoRepo(_('%s is not a local Mercurial repo') % path)
                 self.created.append(path)
             except hg.RepoError, err:
                 ui.print_exc()
                 raise NoRepo("could not create hg repo %s as sink" % path)
         self.lock = None
         self.wlock = None
         self.filemapmode = False
 
     def before(self):
         self.ui.debug(_('run hg sink pre-conversion action\n'))
         self.wlock = self.repo.wlock()
         self.lock = self.repo.lock()
         self.repo.dirstate.clear()
 
     def after(self):
         self.ui.debug(_('run hg sink post-conversion action\n'))
         self.repo.dirstate.invalidate()
         self.lock = None
         self.wlock = None
 
     def revmapfile(self):
         return os.path.join(self.path, ".hg", "shamap")
 
     def authorfile(self):
         return os.path.join(self.path, ".hg", "authormap")
 
     def getheads(self):
         h = self.repo.changelog.heads()
         return [ hex(x) for x in h ]
 
     def putfile(self, f, e, data):
         self.repo.wwrite(f, data, e)
         if f not in self.repo.dirstate:
             self.repo.dirstate.normallookup(f)
 
     def copyfile(self, source, dest):
         self.repo.copy(source, dest)
 
     def delfile(self, f):
         try:
             util.unlink(self.repo.wjoin(f))
             #self.repo.remove([f])
         except OSError:
             pass
 
     def setbranch(self, branch, pbranch, parents):
         if (not self.clonebranches) or (branch == self.lastbranch):
             return
 
         self.lastbranch = branch
         self.after()
         if not branch:
             branch = 'default'
         if not pbranch:
             pbranch = 'default'
 
         branchpath = os.path.join(self.path, branch)
         try:
             self.repo = hg.repository(self.ui, branchpath)
         except:
             if not parents:
                 self.repo = hg.repository(self.ui, branchpath, create=True)
             else:
                 self.ui.note(_('cloning branch %s to %s\n') % (pbranch, branch))
                 hg.clone(self.ui, os.path.join(self.path, pbranch),
                          branchpath, rev=parents, update=False,
                          stream=True)
                 self.repo = hg.repository(self.ui, branchpath)
         self.before()
 
     def putcommit(self, files, parents, commit):
         seen = {}
         pl = []
         for p in parents:
             if p not in seen:
                 pl.append(p)
                 seen[p] = 1
         parents = pl
         nparents = len(parents)
         if self.filemapmode and nparents == 1:
             m1node = self.repo.changelog.read(bin(parents[0]))[0]
             parent = parents[0]
 
         if len(parents) < 2: parents.append("0" * 40)
         if len(parents) < 2: parents.append("0" * 40)
         p2 = parents.pop(0)
 
         text = commit.desc
         extra = commit.extra.copy()
         if self.branchnames and commit.branch:
             extra['branch'] = commit.branch
         if commit.rev:
             extra['convert_revision'] = commit.rev
 
         while parents:
             p1 = p2
             p2 = parents.pop(0)
             a = self.repo.rawcommit(files, text, commit.author, commit.date,
                                     bin(p1), bin(p2), extra=extra)
             self.repo.dirstate.clear()
             text = "(octopus merge fixup)\n"
             p2 = hg.hex(self.repo.changelog.tip())
 
         if self.filemapmode and nparents == 1:
             man = self.repo.manifest
             mnode = self.repo.changelog.read(bin(p2))[0]
             if not man.cmp(m1node, man.revision(mnode)):
                 self.repo.rollback()
                 self.repo.dirstate.clear()
                 return parent
         return p2
 
     def puttags(self, tags):
         try:
             old = self.repo.wfile(".hgtags").read()
             oldlines = old.splitlines(1)
             oldlines.sort()
         except:
             oldlines = []
 
         k = tags.keys()
         k.sort()
         newlines = []
         for tag in k:
             newlines.append("%s %s\n" % (tags[tag], tag))
 
         newlines.sort()
 
         if newlines != oldlines:
             self.ui.status("updating tags\n")
             f = self.repo.wfile(".hgtags", "w")
             f.write("".join(newlines))
             f.close()
             if not oldlines: self.repo.add([".hgtags"])
             date = "%s 0" % int(time.mktime(time.gmtime()))
             extra = {}
             if self.tagsbranch != 'default':
                 extra['branch'] = self.tagsbranch
             try:
                 tagparent = self.repo.changectx(self.tagsbranch).node()
             except hg.RepoError, inst:
                 tagparent = nullid
             self.repo.rawcommit([".hgtags"], "update tags", "convert-repo",
                                 date, tagparent, nullid, extra=extra)
             return hex(self.repo.changelog.tip())
 
     def setfilemapmode(self, active):
         self.filemapmode = active
 
 class mercurial_source(converter_source):
     def __init__(self, ui, path, rev=None):
         converter_source.__init__(self, ui, path, rev)
         self.saverev = ui.configbool('convert', 'hg.saverev', True)
         try:
             self.repo = hg.repository(self.ui, path)
             # try to provoke an exception if this isn't really a hg
             # repo, but some other bogus compatible-looking url
             if not self.repo.local():
                 raise hg.RepoError()
         except hg.RepoError:
             ui.print_exc()
             raise NoRepo("%s is not a local Mercurial repo" % path)
         self.lastrev = None
         self.lastctx = None
         self._changescache = None
         self.convertfp = None
 
     def changectx(self, rev):
         if self.lastrev != rev:
             self.lastctx = self.repo.changectx(rev)
             self.lastrev = rev
         return self.lastctx
 
     def getheads(self):
         if self.rev:
             return [hex(self.repo.changectx(self.rev).node())]
         else:
             return [hex(node) for node in self.repo.heads()]
 
     def getfile(self, name, rev):
         try:
             return self.changectx(rev).filectx(name).data()
         except revlog.LookupError, err:
             raise IOError(err)
 
     def getmode(self, name, rev):
         m = self.changectx(rev).manifest()
         return (m.execf(name) and 'x' or '') + (m.linkf(name) and 'l' or '')
 
     def getchanges(self, rev):
         ctx = self.changectx(rev)
         if self._changescache and self._changescache[0] == rev:
             m, a, r = self._changescache[1]
         else:
             m, a, r = self.repo.status(ctx.parents()[0].node(), ctx.node())[:3]
         changes = [(name, rev) for name in m + a + r]
         changes.sort()
         return (changes, self.getcopies(ctx, m + a))
 
     def getcopies(self, ctx, files):
         copies = {}
         for name in files:
             try:
                 copies[name] = ctx.filectx(name).renamed()[0]
             except TypeError:
                 pass
         return copies
 
     def getcommit(self, rev):
         ctx = self.changectx(rev)
         parents = [hex(p.node()) for p in ctx.parents() if p.node() != nullid]
         if self.saverev:
             crev = rev
         else:
             crev = None
         return commit(author=ctx.user(), date=util.datestr(ctx.date()),
                       desc=ctx.description(), rev=crev, parents=parents,
                       branch=ctx.branch(), extra=ctx.extra())
 
     def gettags(self):
         tags = [t for t in self.repo.tagslist() if t[0] != 'tip']
         return dict([(name, hex(node)) for name, node in tags])
 
     def getchangedfiles(self, rev, i):
         ctx = self.changectx(rev)
         i = i or 0
         changes = self.repo.status(ctx.parents()[i].node(), ctx.node())[:3]
 
         if i == 0:
             self._changescache = (rev, changes)
 
         return changes[0] + changes[1] + changes[2]
 
     def converted(self, rev, destrev):
         if self.convertfp is None:
             self.convertfp = open(os.path.join(self.path, '.hg', 'shamap'),
                                   'a')
         self.convertfp.write('%s %s\n' % (destrev, rev))
         self.convertfp.flush()
 
     def before(self):
         self.ui.debug(_('run hg source pre-conversion action\n'))
 
     def after(self):
         self.ui.debug(_('run hg source post-conversion action\n'))

hgext/convert/subversion.py

@@ -1,984 +1,984 b''
 # Subversion 1.4/1.5 Python API backend
 #
 # Copyright(C) 2007 Daniel Holth et al
 #
 # Configuration options:
 #
 # convert.svn.trunk
 #   Relative path to the trunk (default: "trunk")
 # convert.svn.branches
 #   Relative path to tree of branches (default: "branches")
 # convert.svn.tags
 #   Relative path to tree of tags (default: "tags")
 #
 # Set these in a hgrc, or on the command line as follows:
 #
 #   hg convert --config convert.svn.trunk=wackoname [...]
 
 import locale
 import os
 import re
 import sys
 import cPickle as pickle
 import tempfile
 
 from mercurial import strutil, util
 from mercurial.i18n import _
 
 # Subversion stuff. Works best with very recent Python SVN bindings
 # e.g. SVN 1.5 or backports. Thanks to the bzr folks for enhancing
 # these bindings.
 
 from cStringIO import StringIO
 
 from common import NoRepo, commit, converter_source, encodeargs, decodeargs
 from common import commandline, converter_sink, mapfile
 
 try:
     from svn.core import SubversionException, Pool
     import svn
     import svn.client
     import svn.core
     import svn.ra
     import svn.delta
     import transport
 except ImportError:
     pass
 
 def geturl(path):
     try:
         return svn.client.url_from_path(svn.core.svn_path_canonicalize(path))
     except SubversionException:
         pass
     if os.path.isdir(path):
         path = os.path.normpath(os.path.abspath(path))
         if os.name == 'nt':
             path = '/' + util.normpath(path)
         return 'file://%s' % path
     return path
 
 def optrev(number):
     optrev = svn.core.svn_opt_revision_t()
     optrev.kind = svn.core.svn_opt_revision_number
     optrev.value.number = number
     return optrev
 
 class changedpath(object):
     def __init__(self, p):
         self.copyfrom_path = p.copyfrom_path
         self.copyfrom_rev = p.copyfrom_rev
         self.action = p.action
 
 def get_log_child(fp, url, paths, start, end, limit=0, discover_changed_paths=True,
                     strict_node_history=False):
     protocol = -1
     def receiver(orig_paths, revnum, author, date, message, pool):
         if orig_paths is not None:
             for k, v in orig_paths.iteritems():
                 orig_paths[k] = changedpath(v)
         pickle.dump((orig_paths, revnum, author, date, message),
                     fp, protocol)
 
     try:
         # Use an ra of our own so that our parent can consume
         # our results without confusing the server.
         t = transport.SvnRaTransport(url=url)
         svn.ra.get_log(t.ra, paths, start, end, limit,
                        discover_changed_paths,
                        strict_node_history,
                        receiver)
     except SubversionException, (inst, num):
         pickle.dump(num, fp, protocol)
     except IOError:
         # Caller may interrupt the iteration
         pickle.dump(None, fp, protocol)
     else:
         pickle.dump(None, fp, protocol)
     fp.close()
 
 def debugsvnlog(ui, **opts):
     """Fetch SVN log in a subprocess and channel them back to parent to
     avoid memory collection issues.
     """
     util.set_binary(sys.stdin)
     util.set_binary(sys.stdout)
     args = decodeargs(sys.stdin.read())
     get_log_child(sys.stdout, *args)
 
 class logstream:
     """Interruptible revision log iterator."""
     def __init__(self, stdout):
         self._stdout = stdout
 
     def __iter__(self):
         while True:
             entry = pickle.load(self._stdout)
             try:
                 orig_paths, revnum, author, date, message = entry
             except:
                 if entry is None:
                     break
                 raise SubversionException("child raised exception", entry)
             yield entry
 
     def close(self):
         if self._stdout:
             self._stdout.close()
             self._stdout = None
 
 def get_log(url, paths, start, end, limit=0, discover_changed_paths=True,
                 strict_node_history=False):
     args = [url, paths, start, end, limit, discover_changed_paths,
             strict_node_history]
     arg = encodeargs(args)
     hgexe = util.hgexecutable()
     cmd = '%s debugsvnlog' % util.shellquote(hgexe)
     stdin, stdout = os.popen2(cmd, 'b')
     stdin.write(arg)
     stdin.close()
     return logstream(stdout)
 
 # SVN conversion code stolen from bzr-svn and tailor
 #
 # Subversion looks like a versioned filesystem, branches structures
 # are defined by conventions and not enforced by the tool. First,
 # we define the potential branches (modules) as "trunk" and "branches"
 # children directories. Revisions are then identified by their
 # module and revision number (and a repository identifier).
 #
 # The revision graph is really a tree (or a forest). By default, a
 # revision parent is the previous revision in the same module. If the
 # module directory is copied/moved from another module then the
 # revision is the module root and its parent the source revision in
 # the parent module. A revision has at most one parent.
 #
 class svn_source(converter_source):
     def __init__(self, ui, url, rev=None):
         super(svn_source, self).__init__(ui, url, rev=rev)
 
         try:
             SubversionException
         except NameError:
             raise NoRepo('Subversion python bindings could not be loaded')
 
         self.encoding = locale.getpreferredencoding()
         self.lastrevs = {}
 
         latest = None
         try:
             # Support file://path@rev syntax. Useful e.g. to convert
             # deleted branches.
             at = url.rfind('@')
             if at >= 0:
                 latest = int(url[at+1:])
                 url = url[:at]
         except ValueError, e:
             pass
         self.url = geturl(url)
         self.encoding = 'UTF-8' # Subversion is always nominal UTF-8
         try:
             self.transport = transport.SvnRaTransport(url=self.url)
             self.ra = self.transport.ra
             self.ctx = self.transport.client
             self.base = svn.ra.get_repos_root(self.ra)
             self.module = self.url[len(self.base):]
             self.commits = {}
             self.paths = {}
             self.uuid = svn.ra.get_uuid(self.ra).decode(self.encoding)
         except SubversionException, e:
             ui.print_exc()
             raise NoRepo("%s does not look like a Subversion repo" % self.url)
 
         if rev:
             try:
                 latest = int(rev)
             except ValueError:
                 raise util.Abort('svn: revision %s is not an integer' % rev)
 
         try:
             self.get_blacklist()
         except IOError, e:
             pass
 
         self.last_changed = self.latest(self.module, latest)
 
         self.head = self.revid(self.last_changed)
         self._changescache = None
 
         if os.path.exists(os.path.join(url, '.svn/entries')):
             self.wc = url
         else:
             self.wc = None
         self.convertfp = None
 
     def setrevmap(self, revmap):
         lastrevs = {}
         for revid in revmap.iterkeys():
             uuid, module, revnum = self.revsplit(revid)
             lastrevnum = lastrevs.setdefault(module, revnum)
             if revnum > lastrevnum:
                 lastrevs[module] = revnum
         self.lastrevs = lastrevs
 
     def exists(self, path, optrev):
         try:
             svn.client.ls(self.url.rstrip('/') + '/' + path,
                                  optrev, False, self.ctx)
             return True
         except SubversionException, err:
             return False
 
     def getheads(self):
 
         def getcfgpath(name, rev):
             cfgpath = self.ui.config('convert', 'svn.' + name)
             path = (cfgpath or name).strip('/')
             if not self.exists(path, rev):
                 if cfgpath:
                     raise util.Abort(_('expected %s to be at %r, but not found')
                                  % (name, path))
                 return None
             self.ui.note(_('found %s at %r\n') % (name, path))
             return path
 
         rev = optrev(self.last_changed)
         oldmodule = ''
         trunk = getcfgpath('trunk', rev)
         tags = getcfgpath('tags', rev)
         branches = getcfgpath('branches', rev)
 
         # If the project has a trunk or branches, we will extract heads
         # from them. We keep the project root otherwise.
         if trunk:
             oldmodule = self.module or ''
             self.module += '/' + trunk
             lt = self.latest(self.module, self.last_changed)
             self.head = self.revid(lt)
 
         # First head in the list is the module's head
         self.heads = [self.head]
         self.tags = '%s/%s' % (oldmodule , (tags or 'tags'))
 
         # Check if branches bring a few more heads to the list
         if branches:
             rpath = self.url.strip('/')
             branchnames = svn.client.ls(rpath + '/' + branches, rev, False,
                                         self.ctx)
             for branch in branchnames.keys():
                 module = '%s/%s/%s' % (oldmodule, branches, branch)
                 brevnum = self.latest(module, self.last_changed)
                 brev = self.revid(brevnum, module)
                 self.ui.note('found branch %s at %d\n' % (branch, brevnum))
                 self.heads.append(brev)
 
         return self.heads
 
     def getfile(self, file, rev):
         data, mode = self._getfile(file, rev)
         self.modecache[(file, rev)] = mode
         return data
 
     def getmode(self, file, rev):
         return self.modecache[(file, rev)]
 
     def getchanges(self, rev):
         if self._changescache and self._changescache[0] == rev:
             return self._changescache[1]
         self._changescache = None
         self.modecache = {}
         (paths, parents) = self.paths[rev]
         files, copies = self.expandpaths(rev, paths, parents)
         files.sort()
         files = zip(files, [rev] * len(files))
 
         # caller caches the result, so free it here to release memory
         del self.paths[rev]
         return (files, copies)
 
     def getchangedfiles(self, rev, i):
         changes = self.getchanges(rev)
         self._changescache = (rev, changes)
         return [f[0] for f in changes[0]]
 
     def getcommit(self, rev):
         if rev not in self.commits:
             uuid, module, revnum = self.revsplit(rev)
             self.module = module
             self.reparent(module)
             # We assume that:
             # - requests for revisions after "stop" come from the
             # revision graph backward traversal. Cache all of them
             # down to stop, they will be used eventually.
             # - requests for revisions before "stop" come to get
             # isolated branches parents. Just fetch what is needed.
             stop = self.lastrevs.get(module, 0)
             if revnum < stop:
                 stop = revnum + 1
             self._fetch_revisions(revnum, stop)
         commit = self.commits[rev]
         # caller caches the result, so free it here to release memory
         del self.commits[rev]
         return commit
 
     def gettags(self):
         tags = {}
         start = self.revnum(self.head)
         try:
             for entry in get_log(self.url, [self.tags], 0, start):
                 orig_paths, revnum, author, date, message = entry
                 for path in orig_paths:
                     if not path.startswith(self.tags+'/'):
                         continue
                     ent = orig_paths[path]
                     source = ent.copyfrom_path
                     rev = ent.copyfrom_rev
                     tag = path.split('/')[-1]
                     tags[tag] = self.revid(rev, module=source)
         except SubversionException, (inst, num):
             self.ui.note('no tags found at revision %d\n' % start)
         return tags
 
     def converted(self, rev, destrev):
         if not self.wc:
             return
         if self.convertfp is None:
             self.convertfp = open(os.path.join(self.wc, '.svn', 'hg-shamap'),
                                   'a')
         self.convertfp.write('%s %d\n' % (destrev, self.revnum(rev)))
         self.convertfp.flush()
 
     # -- helper functions --
 
     def revid(self, revnum, module=None):
         if not module:
             module = self.module
         return u"svn:%s%s@%s" % (self.uuid, module.decode(self.encoding),
                                  revnum)
 
     def revnum(self, rev):
         return int(rev.split('@')[-1])
 
     def revsplit(self, rev):
         url, revnum = rev.encode(self.encoding).split('@', 1)
         revnum = int(revnum)
         parts = url.split('/', 1)
         uuid = parts.pop(0)[4:]
         mod = ''
         if parts:
             mod = '/' + parts[0]
         return uuid, mod, revnum
 
     def latest(self, path, stop=0):
         'find the latest revision affecting path, up to stop'
         if not stop:
             stop = svn.ra.get_latest_revnum(self.ra)
         try:
             self.reparent('')
             dirent = svn.ra.stat(self.ra, path.strip('/'), stop)
             self.reparent(self.module)
         except SubversionException:
             dirent = None
         if not dirent:
             raise util.Abort('%s not found up to revision %d' % (path, stop))
 
         return dirent.created_rev
 
     def get_blacklist(self):
         """Avoid certain revision numbers.
         It is not uncommon for two nearby revisions to cancel each other
         out, e.g. 'I copied trunk into a subdirectory of itself instead
         of making a branch'. The converted repository is significantly
         smaller if we ignore such revisions."""
         self.blacklist = util.set()
         blacklist = self.blacklist
         for line in file("blacklist.txt", "r"):
             if not line.startswith("#"):
                 try:
                     svn_rev = int(line.strip())
                     blacklist.add(svn_rev)
                 except ValueError, e:
                     pass # not an integer or a comment
 
     def is_blacklisted(self, svn_rev):
         return svn_rev in self.blacklist
 
     def reparent(self, module):
         svn_url = self.base + module
         self.ui.debug("reparent to %s\n" % svn_url.encode(self.encoding))
         svn.ra.reparent(self.ra, svn_url.encode(self.encoding))
 
     def expandpaths(self, rev, paths, parents):
         def get_entry_from_path(path, module=self.module):
             # Given the repository url of this wc, say
             #   "http://server/plone/CMFPlone/branches/Plone-2_0-branch"
             # extract the "entry" portion (a relative path) from what
             # svn log --xml says, ie
             #   "/CMFPlone/branches/Plone-2_0-branch/tests/PloneTestCase.py"
             # that is to say "tests/PloneTestCase.py"
             if path.startswith(module):
                 relative = path[len(module):]
                 if relative.startswith('/'):
                     return relative[1:]
                 else:
                     return relative
 
             # The path is outside our tracked tree...
             self.ui.debug('%r is not under %r, ignoring\n' % (path, module))
             return None
 
         entries = []
         copyfrom = {} # Map of entrypath, revision for finding source of deleted revisions.
         copies = {}
 
         new_module, revnum = self.revsplit(rev)[1:]
         if new_module != self.module:
             self.module = new_module
             self.reparent(self.module)
 
         for path, ent in paths:
             entrypath = get_entry_from_path(path, module=self.module)
             entry = entrypath.decode(self.encoding)
 
             kind = svn.ra.check_path(self.ra, entrypath, revnum)
             if kind == svn.core.svn_node_file:
                 if ent.copyfrom_path:
                     copyfrom_path = get_entry_from_path(ent.copyfrom_path)
                     if copyfrom_path:
                         self.ui.debug("Copied to %s from %s@%s\n" %
                                       (entrypath, copyfrom_path,
                                        ent.copyfrom_rev))
                         # It's probably important for hg that the source
                         # exists in the revision's parent, not just the
                         # ent.copyfrom_rev
                         fromkind = svn.ra.check_path(self.ra, copyfrom_path, ent.copyfrom_rev)
                         if fromkind != 0:
                             copies[self.recode(entry)] = self.recode(copyfrom_path)
                 entries.append(self.recode(entry))
             elif kind == 0: # gone, but had better be a deleted *file*
                 self.ui.debug("gone from %s\n" % ent.copyfrom_rev)
 
                 # if a branch is created but entries are removed in the same
                 # changeset, get the right fromrev
                 # parents cannot be empty here, you cannot remove things from
                 # a root revision.
                 uuid, old_module, fromrev = self.revsplit(parents[0])
 
                 basepath = old_module + "/" + get_entry_from_path(path, module=self.module)
                 entrypath = old_module + "/" + get_entry_from_path(path, module=self.module)
 
                 def lookup_parts(p):
                     rc = None
                     parts = p.split("/")
                     for i in range(len(parts)):
                         part = "/".join(parts[:i])
                         info = part, copyfrom.get(part, None)
                         if info[1] is not None:
                             self.ui.debug("Found parent directory %s\n" % info[1])
                             rc = info
                     return rc
 
                 self.ui.debug("base, entry %s %s\n" % (basepath, entrypath))
 
                 frompath, froment = lookup_parts(entrypath) or (None, revnum - 1)
 
                 # need to remove fragment from lookup_parts and replace with copyfrom_path
                 if frompath is not None:
                     self.ui.debug("munge-o-matic\n")
                     self.ui.debug(entrypath + '\n')
                     self.ui.debug(entrypath[len(frompath):] + '\n')
                     entrypath = froment.copyfrom_path + entrypath[len(frompath):]
                     fromrev = froment.copyfrom_rev
                     self.ui.debug("Info: %s %s %s %s\n" % (frompath, froment, ent, entrypath))
 
                 # We can avoid the reparent calls if the module has not changed
                 # but it probably does not worth the pain.
                 self.reparent('')
                 fromkind = svn.ra.check_path(self.ra, entrypath.strip('/'), fromrev)
                 self.reparent(self.module)
                 
                 if fromkind == svn.core.svn_node_file:   # a deleted file
                     entries.append(self.recode(entry))
                 elif fromkind == svn.core.svn_node_dir:
                     # print "Deleted/moved non-file:", revnum, path, ent
                     # children = self._find_children(path, revnum - 1)
                     # print "find children %s@%d from %d action %s" % (path, revnum, ent.copyfrom_rev, ent.action)
                     # Sometimes this is tricky. For example: in
                     # The Subversion Repository revision 6940 a dir
                     # was copied and one of its files was deleted
                     # from the new location in the same commit. This
                     # code can't deal with that yet.
                     if ent.action == 'C':
                         children = self._find_children(path, fromrev)
                     else:
                         oroot = entrypath.strip('/')
                         nroot = path.strip('/')
                         children = self._find_children(oroot, fromrev)
                         children = [s.replace(oroot,nroot) for s in children]
                     # Mark all [files, not directories] as deleted.
                     for child in children:
                         # Can we move a child directory and its
                         # parent in the same commit? (probably can). Could
                         # cause problems if instead of revnum -1,
                         # we have to look in (copyfrom_path, revnum - 1)
                         entrypath = get_entry_from_path("/" + child, module=old_module)
                         if entrypath:
                             entry = self.recode(entrypath.decode(self.encoding))
                             if entry in copies:
                                 # deleted file within a copy
                                 del copies[entry]
                             else:
                                 entries.append(entry)
                 else:
                     self.ui.debug('unknown path in revision %d: %s\n' % \
                                   (revnum, path))
             elif kind == svn.core.svn_node_dir:
                 # Should probably synthesize normal file entries
                 # and handle as above to clean up copy/rename handling.
 
                 # If the directory just had a prop change,
                 # then we shouldn't need to look for its children.
                 if ent.action == 'M':
                     continue
 
                 # Also this could create duplicate entries. Not sure
                 # whether this will matter. Maybe should make entries a set.
                 # print "Changed directory", revnum, path, ent.action, ent.copyfrom_path, ent.copyfrom_rev
                 # This will fail if a directory was copied
                 # from another branch and then some of its files
                 # were deleted in the same transaction.
                 children = self._find_children(path, revnum)
                 children.sort()
                 for child in children:
                     # Can we move a child directory and its
                     # parent in the same commit? (probably can). Could
                     # cause problems if instead of revnum -1,
                     # we have to look in (copyfrom_path, revnum - 1)
                     entrypath = get_entry_from_path("/" + child, module=self.module)
                     # print child, self.module, entrypath
                     if entrypath:
                         # Need to filter out directories here...
                         kind = svn.ra.check_path(self.ra, entrypath, revnum)
                         if kind != svn.core.svn_node_dir:
                             entries.append(self.recode(entrypath))
 
                 # Copies here (must copy all from source)
                 # Probably not a real problem for us if
                 # source does not exist
 
                 # Can do this with the copy command "hg copy"
                 # if ent.copyfrom_path:
                 #     copyfrom_entry = get_entry_from_path(ent.copyfrom_path.decode(self.encoding),
                 #             module=self.module)
                 #     copyto_entry = entrypath
                 #
                 #     print "copy directory", copyfrom_entry, 'to', copyto_entry
                 #
                 #     copies.append((copyfrom_entry, copyto_entry))
 
                 if ent.copyfrom_path:
                     copyfrom_path = ent.copyfrom_path.decode(self.encoding)
                     copyfrom_entry = get_entry_from_path(copyfrom_path, module=self.module)
                     if copyfrom_entry:
                         copyfrom[path] = ent
                         self.ui.debug("mark %s came from %s\n" % (path, copyfrom[path]))
 
                         # Good, /probably/ a regular copy. Really should check
                         # to see whether the parent revision actually contains
                         # the directory in question.
                         children = self._find_children(self.recode(copyfrom_path), ent.copyfrom_rev)
                         children.sort()
                         for child in children:
                             entrypath = get_entry_from_path("/" + child, module=self.module)
                             if entrypath:
                                 entry = entrypath.decode(self.encoding)
                                 # print "COPY COPY From", copyfrom_entry, entry
                                 copyto_path = path + entry[len(copyfrom_entry):]
                                 copyto_entry =  get_entry_from_path(copyto_path, module=self.module)
                                 # print "COPY", entry, "COPY To", copyto_entry
                                 copies[self.recode(copyto_entry)] = self.recode(entry)
                                 # copy from quux splort/quuxfile
 
         return (util.unique(entries), copies)
 
     def _fetch_revisions(self, from_revnum, to_revnum):
         if from_revnum < to_revnum:
             from_revnum, to_revnum = to_revnum, from_revnum
 
         self.child_cset = None
         def parselogentry(orig_paths, revnum, author, date, message):
             """Return the parsed commit object or None, and True if 
             the revision is a branch root.
             """
             self.ui.debug("parsing revision %d (%d changes)\n" %
                           (revnum, len(orig_paths)))
 
             rev = self.revid(revnum)
             # branch log might return entries for a parent we already have
 
             if (rev in self.commits or revnum < to_revnum):
                 return None, False
 
             parents = []
             # check whether this revision is the start of a branch
             if self.module in orig_paths:
                 ent = orig_paths[self.module]
                 if ent.copyfrom_path:
                     # ent.copyfrom_rev may not be the actual last revision
                     prev = self.latest(ent.copyfrom_path, ent.copyfrom_rev)
                     parents = [self.revid(prev, ent.copyfrom_path)]
                     self.ui.note('found parent of branch %s at %d: %s\n' % \
                                      (self.module, prev, ent.copyfrom_path))
                 else:
                     self.ui.debug("No copyfrom path, don't know what to do.\n")
 
             orig_paths = orig_paths.items()
             orig_paths.sort()
             paths = []
             # filter out unrelated paths
             for path, ent in orig_paths:
                 if not path.startswith(self.module):
                     self.ui.debug("boring@%s: %s\n" % (revnum, path))
                     continue
                 paths.append((path, ent))
 
             # Example SVN datetime. Includes microseconds.
             # ISO-8601 conformant
             # '2007-01-04T17:35:00.902377Z'
             date = util.parsedate(date[:19] + " UTC", ["%Y-%m-%dT%H:%M:%S"])
 
-            log = message and self.recode(message)
+            log = message and self.recode(message) or ''
             author = author and self.recode(author) or ''
             try:
                 branch = self.module.split("/")[-1]
                 if branch == 'trunk':
                     branch = ''
             except IndexError:
                 branch = None
 
             cset = commit(author=author,
                           date=util.datestr(date),
                           desc=log,
                           parents=parents,
                           branch=branch,
                           rev=rev.encode('utf-8'))
 
             self.commits[rev] = cset
             # The parents list is *shared* among self.paths and the
             # commit object. Both will be updated below.
             self.paths[rev] = (paths, cset.parents)
             if self.child_cset and not self.child_cset.parents:
                 self.child_cset.parents[:] = [rev]
             self.child_cset = cset
             return cset, len(parents) > 0
 
         self.ui.note('fetching revision log for "%s" from %d to %d\n' %
                      (self.module, from_revnum, to_revnum))
 
         try:
             firstcset = None
             stream = get_log(self.url, [self.module], from_revnum, to_revnum)
             try:
                 for entry in stream:
                     paths, revnum, author, date, message = entry
                     if self.is_blacklisted(revnum):
                         self.ui.note('skipping blacklisted revision %d\n' 
                                      % revnum)
                         continue
                     if paths is None:
                         self.ui.debug('revision %d has no entries\n' % revnum)
                         continue
                     cset, branched = parselogentry(paths, revnum, author, 
                                                    date, message)
                     if cset:
                         firstcset = cset
                     if branched:
                         break
             finally:
                 stream.close()
 
             if firstcset and not firstcset.parents:
                 # The first revision of the sequence (the last fetched one)
                 # has invalid parents if not a branch root. Find the parent
                 # revision now, if any.
                 try:
                     firstrevnum = self.revnum(firstcset.rev)
                     if firstrevnum > 1:
                         latest = self.latest(self.module, firstrevnum - 1)
                         firstcset.parents.append(self.revid(latest))
                 except util.Abort:
                     pass
         except SubversionException, (inst, num):
             if num == svn.core.SVN_ERR_FS_NO_SUCH_REVISION:
                 raise NoSuchRevision(branch=self,
                     revision="Revision number %d" % to_revnum)
             raise
 
     def _getfile(self, file, rev):
         io = StringIO()
         # TODO: ra.get_file transmits the whole file instead of diffs.
         mode = ''
         try:
             new_module, revnum = self.revsplit(rev)[1:]
             if self.module != new_module:
                 self.module = new_module
                 self.reparent(self.module)
             info = svn.ra.get_file(self.ra, file, revnum, io)
             if isinstance(info, list):
                 info = info[-1]
             mode = ("svn:executable" in info) and 'x' or ''
             mode = ("svn:special" in info) and 'l' or mode
         except SubversionException, e:
             notfound = (svn.core.SVN_ERR_FS_NOT_FOUND,
                 svn.core.SVN_ERR_RA_DAV_PATH_NOT_FOUND)
             if e.apr_err in notfound: # File not found
                 raise IOError()
             raise
         data = io.getvalue()
         if mode == 'l':
             link_prefix = "link "
             if data.startswith(link_prefix):
                 data = data[len(link_prefix):]
         return data, mode
 
     def _find_children(self, path, revnum):
         path = path.strip('/')
         pool = Pool()
         rpath = '/'.join([self.base, path]).strip('/')
         return ['%s/%s' % (path, x) for x in svn.client.ls(rpath, optrev(revnum), True, self.ctx, pool).keys()]
 
 pre_revprop_change = '''#!/bin/sh
 
 REPOS="$1"
 REV="$2"
 USER="$3"
 PROPNAME="$4"
 ACTION="$5"
 
 if [ "$ACTION" = "M" -a "$PROPNAME" = "svn:log" ]; then exit 0; fi
 if [ "$ACTION" = "A" -a "$PROPNAME" = "hg:convert-branch" ]; then exit 0; fi
 if [ "$ACTION" = "A" -a "$PROPNAME" = "hg:convert-rev" ]; then exit 0; fi
 
 echo "Changing prohibited revision property" >&2
 exit 1
 '''
 
 class svn_sink(converter_sink, commandline):
     commit_re = re.compile(r'Committed revision (\d+).', re.M)
 
     def prerun(self):
         if self.wc:
             os.chdir(self.wc)
 
     def postrun(self):
         if self.wc:
             os.chdir(self.cwd)
 
     def join(self, name):
         return os.path.join(self.wc, '.svn', name)
 
     def revmapfile(self):
         return self.join('hg-shamap')
 
     def authorfile(self):
         return self.join('hg-authormap')
 
     def __init__(self, ui, path):
         converter_sink.__init__(self, ui, path)
         commandline.__init__(self, ui, 'svn')
         self.delete = []
         self.setexec = []
         self.delexec = []
         self.copies = []
         self.wc = None
         self.cwd = os.getcwd()
 
         path = os.path.realpath(path)
 
         created = False
         if os.path.isfile(os.path.join(path, '.svn', 'entries')):
             self.wc = path
             self.run0('update')
         else:
             wcpath = os.path.join(os.getcwd(), os.path.basename(path) + '-wc')
 
             if os.path.isdir(os.path.dirname(path)):
                 if not os.path.exists(os.path.join(path, 'db', 'fs-type')):
                     ui.status(_('initializing svn repo %r\n') %
                               os.path.basename(path))
                     commandline(ui, 'svnadmin').run0('create', path)
                     created = path
                 path = util.normpath(path)
                 if not path.startswith('/'):
                     path = '/' + path
                 path = 'file://' + path
 
             ui.status(_('initializing svn wc %r\n') % os.path.basename(wcpath))
             self.run0('checkout', path, wcpath)
 
             self.wc = wcpath
         self.opener = util.opener(self.wc)
         self.wopener = util.opener(self.wc)
         self.childmap = mapfile(ui, self.join('hg-childmap'))
         self.is_exec = util.checkexec(self.wc) and util.is_exec or None
 
         if created:
             hook = os.path.join(created, 'hooks', 'pre-revprop-change')
             fp = open(hook, 'w')
             fp.write(pre_revprop_change)
             fp.close()
             util.set_flags(hook, "x")
 
         xport = transport.SvnRaTransport(url=geturl(path))
         self.uuid = svn.ra.get_uuid(xport.ra)
 
     def wjoin(self, *names):
         return os.path.join(self.wc, *names)
 
     def putfile(self, filename, flags, data):
         if 'l' in flags:
             self.wopener.symlink(data, filename)
         else:
             try:
                 if os.path.islink(self.wjoin(filename)):
                     os.unlink(filename)
             except OSError:
                 pass
             self.wopener(filename, 'w').write(data)
 
             if self.is_exec:
                 was_exec = self.is_exec(self.wjoin(filename))
             else:
                 # On filesystems not supporting execute-bit, there is no way
                 # to know if it is set but asking subversion. Setting it
                 # systematically is just as expensive and much simpler.
                 was_exec = 'x' not in flags
 
             util.set_flags(self.wjoin(filename), flags)
             if was_exec:
                 if 'x' not in flags:
                     self.delexec.append(filename)
             else:
                 if 'x' in flags:
                     self.setexec.append(filename)
 
     def delfile(self, name):
         self.delete.append(name)
 
     def copyfile(self, source, dest):
         self.copies.append([source, dest])
 
     def _copyfile(self, source, dest):
         # SVN's copy command pukes if the destination file exists, but
         # our copyfile method expects to record a copy that has
         # already occurred.  Cross the semantic gap.
         wdest = self.wjoin(dest)
         exists = os.path.exists(wdest)
         if exists:
             fd, tempname = tempfile.mkstemp(
                 prefix='hg-copy-', dir=os.path.dirname(wdest))
             os.close(fd)
             os.unlink(tempname)
             os.rename(wdest, tempname)
         try:
             self.run0('copy', source, dest)
         finally:
             if exists:
                 try:
                     os.unlink(wdest)
                 except OSError:
                     pass
                 os.rename(tempname, wdest)
 
     def dirs_of(self, files):
         dirs = set()
         for f in files:
             if os.path.isdir(self.wjoin(f)):
                 dirs.add(f)
             for i in strutil.rfindall(f, '/'):
                 dirs.add(f[:i])
         return dirs
 
     def add_dirs(self, files):
         add_dirs = [d for d in self.dirs_of(files)
                     if not os.path.exists(self.wjoin(d, '.svn', 'entries'))]
         if add_dirs:
             add_dirs.sort()
             self.xargs(add_dirs, 'add', non_recursive=True, quiet=True)
         return add_dirs
 
     def add_files(self, files):
         if files:
             self.xargs(files, 'add', quiet=True)
         return files
 
     def tidy_dirs(self, names):
         dirs = list(self.dirs_of(names))
         dirs.sort(reverse=True)
         deleted = []
         for d in dirs:
             wd = self.wjoin(d)
             if os.listdir(wd) == '.svn':
                 self.run0('delete', d)
                 deleted.append(d)
         return deleted
 
     def addchild(self, parent, child):
         self.childmap[parent] = child
 
     def revid(self, rev):
         return u"svn:%s@%s" % (self.uuid, rev)
 
     def putcommit(self, files, parents, commit):
         for parent in parents:
             try:
                 return self.revid(self.childmap[parent])
             except KeyError:
                 pass
         entries = set(self.delete)
         files = util.frozenset(files)
         entries.update(self.add_dirs(files.difference(entries)))
         if self.copies:
             for s, d in self.copies:
                 self._copyfile(s, d)
             self.copies = []
         if self.delete:
             self.xargs(self.delete, 'delete')
             self.delete = []
         entries.update(self.add_files(files.difference(entries)))
         entries.update(self.tidy_dirs(entries))
         if self.delexec:
             self.xargs(self.delexec, 'propdel', 'svn:executable')
             self.delexec = []
         if self.setexec:
             self.xargs(self.setexec, 'propset', 'svn:executable', '*')
             self.setexec = []
 
         fd, messagefile = tempfile.mkstemp(prefix='hg-convert-')
         fp = os.fdopen(fd, 'w')
         fp.write(commit.desc)
         fp.close()
         try:
             output = self.run0('commit',
                                username=util.shortuser(commit.author),
                                file=messagefile,
                                encoding='utf-8')
             try:
                 rev = self.commit_re.search(output).group(1)
             except AttributeError:
                 self.ui.warn(_('unexpected svn output:\n'))
                 self.ui.warn(output)
                 raise util.Abort(_('unable to cope with svn output'))
             if commit.rev:
                 self.run('propset', 'hg:convert-rev', commit.rev,
                          revprop=True, revision=rev)
             if commit.branch and commit.branch != 'default':
                 self.run('propset', 'hg:convert-branch', commit.branch,
                          revprop=True, revision=rev)
             for parent in parents:
                 self.addchild(parent, rev)
             return self.revid(rev)
         finally:
             os.unlink(messagefile)
 
     def puttags(self, tags):
         self.ui.warn(_('XXX TAGS NOT IMPLEMENTED YET\n'))

mercurial/util.py

@@ -1,1733 +1,1733 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
-import re, urlparse
+import urlparse
 
 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"))
 
 # 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"""
     return dict.fromkeys(g).keys()
 
 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 = splitpath(n1), 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 endswithsep(root):
         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 = splitpath(normpath)
         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.ENOTDIR, 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))
         parts.pop()
         prefixes = []
         for n in range(len(parts)):
             prefix = os.sep.join(parts)
             if prefix in self.auditeddir:
                 break
             check(prefix)
             prefixes.append(prefix)
             parts.pop()
 
         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 openhardlinks():
     '''return true if it is safe to hold open file handles to hardlinks'''
     return True
 
 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)
     """
 
     # VFAT on some Linux versions can flip mode but it doesn't persist
     # a FS remount. Frequently we can detect it if files are created
     # with exec bit on.
 
     try:
         EXECFLAGS = stat.S_IXUSR | stat.S_IXGRP | stat.S_IXOTH
         fh, fn = tempfile.mkstemp("", "", path)
         try:
             os.close(fh)
             m = os.stat(fn).st_mode & 0777
             new_file_has_exec = m & EXECFLAGS
             os.chmod(fn, m ^ EXECFLAGS)
             exec_flags_cannot_flip = ((os.stat(fn).st_mode & 0777) == m)
         finally:
             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'
 
 def endswithsep(path):
     '''Check path ends with os.sep or os.altsep.'''
     return path.endswith(os.sep) or os.altsep and path.endswith(os.altsep)
 
 def splitpath(path):
     '''Split path by os.sep.
     Note that this function does not use os.altsep because this is
     an alternative of simple "xxx.split(os.sep)".
     It is recommended to use os.path.normpath() before using this
     function if need.'''
     return path.split(os.sep)
 
 # 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:
                 # This is workaround for "Not enough space" error on
                 # writing large size of data to console.
                 limit = 16000
                 l = len(s)
                 start = 0
                 while start < l:
                     end = start + limit
                     self.fp.write(s[start:end])
                     start = end
             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 _is_win_9x():
         '''return true if run on windows 95, 98 or me.'''
         try:
             return sys.getwindowsversion()[3] == 1
         except AttributeError:
             return 'command' in os.environ.get('comspec', '')
 
     def openhardlinks():
         return not _is_win_9x and "win32api" in locals()
 
     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 sshargs(sshcmd, host, user, port):
         '''Build argument list for ssh or Plink'''
         pflag = 'plink' in sshcmd.lower() and '-P' or '-p'
         args = user and ("%s@%s" % (user, host)) or host
         return port and ("%s %s %s" % (args, pflag, port)) or args
 
     def testpid(pid):
         '''return False if pid dead, True if running or not known'''
         return True
 
     def set_flags(f, flags):
         pass
 
     def set_binary(fd):
         msvcrt.setmode(fd.fileno(), os.O_BINARY)
 
     def pconvert(path):
         return '/'.join(splitpath(path))
 
     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 popen(command):
         # Work around "popen spawned process may not write to stdout
         # under windows"
         # http://bugs.python.org/issue1366
         command += " 2> %s" % nulldev
         return os.popen(quotecommand(command))
 
     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 sshargs(sshcmd, host, user, port):
         '''Build argument list for ssh'''
         args = user and ("%s@%s" % (user, host)) or host
         return port and ("%s -p %s" % (args, port)) or args
 
     def is_exec(f):
         """check whether a file is executable"""
         return (os.lstat(f).st_mode & 0100 != 0)
 
     def set_flags(f, flags):
         s = os.lstat(f).st_mode
         x = "x" in flags
         l = "l" in flags
         if l:
             if not stat.S_ISLNK(s):
                 # switch file to link
                 data = file(f).read()
                 os.unlink(f)
                 os.symlink(data, f)
             # no chmod needed at this point
             return
         if stat.S_ISLNK(s):
             # switch link to file
             data = os.readlink(f)
             os.unlink(f)
             file(f, "w").write(data)
             s = 0666 & ~_umask # avoid restatting for chmod
 
         sx = s & 0100
         if x and not sx:
             # Turn on +x for every +r bit when making a file executable
             # and obey umask.
             os.chmod(f, s | (s & 0444) >> 2 & ~_umask)
         elif not x and sx:
             # Turn off all +x bits
             os.chmod(f, s & 0666)
 
     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 popen(command):
         return os.popen(command)
 
     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 & 0777
     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 = 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('\\\\', '\\')
 
 def hidepassword(url):
     '''hide user credential in a url string'''
     scheme, netloc, path, params, query, fragment = urlparse.urlparse(url)
     netloc = re.sub('([^:]*):([^@]*)@(.*)', r'\1:***@\3', netloc)
     return urlparse.urlunparse((scheme, netloc, path, params, query, fragment))
 
 def removeauth(url):
     '''remove all authentication information from a url string'''
     scheme, netloc, path, params, query, fragment = urlparse.urlparse(url)
     netloc = netloc[netloc.find('@')+1:]
     return urlparse.urlunparse((scheme, netloc, path, params, query, fragment))

tests/test-convert-cvs

@@ -1,73 +1,95 b''
 #!/bin/sh
 
 "$TESTDIR/hghave" cvs cvsps || exit 80
 
 cvscall()
 {
     cvs -f $@
 }
 
 echo "[extensions]" >> $HGRCPATH
 echo "convert = " >> $HGRCPATH
 
 echo % create cvs repository
 mkdir cvsrepo
 cd cvsrepo
 export CVSROOT=`pwd`
 export CVS_OPTIONS=-f
 cd ..
 
 cvscall -q -d "$CVSROOT" init
 
 echo % create source directory
 mkdir src-temp
 cd src-temp
 echo a > a
 mkdir b
 cd b
 echo c > c
 cd ..
 
 echo % import source directory
 cvscall -q import -m import src INITIAL start
 cd ..
 
 echo % checkout source directory
 cvscall -q checkout src
 
 echo % commit a new revision changing b/c
 cd src
 echo c >> b/c
 cvscall -q commit -mci0 . | grep '<--' |\
     sed -e 's:.*src/\(.*\),v.*:checking in src/\1,v:g'
 cd ..
 
 echo % convert fresh repo
 hg convert src src-hg | sed -e 's/connecting to.*cvsrepo/connecting to cvsrepo/g'
 cat src-hg/a
 cat src-hg/b/c
 
 echo % convert fresh repo with --filemap
 echo include b/c > filemap
 hg convert --filemap filemap src src-filemap | sed -e 's/connecting to.*cvsrepo/connecting to cvsrepo/g'
 cat src-hg/b/c
 hg -R src-filemap log --template '#rev# #desc# files: #files#\n'
 
 echo % commit new file revisions
 cd src
 echo a >> a
 echo c >> b/c
 cvscall -q commit -mci1 . | grep '<--' |\
     sed -e 's:.*src/\(.*\),v.*:checking in src/\1,v:g'
 cd ..
 
 echo % convert again
 hg convert src src-hg | sed -e 's/connecting to.*cvsrepo/connecting to cvsrepo/g'
 cat src-hg/a
 cat src-hg/b/c
 
 echo % convert again with --filemap
 hg convert --filemap filemap src src-filemap | sed -e 's/connecting to.*cvsrepo/connecting to cvsrepo/g'
 cat src-hg/b/c
 hg -R src-filemap log --template '#rev# #desc# files: #files#\n'
 
+echo % commit branch
+cd src
+cvs -q update -r1.1 b/c
+cvs -q tag -b branch
+cvs -q update -r branch
+echo d >> b/c
+cvs -q commit -mci2 . | grep '<--' |\
+    sed -e 's:.*src/\(.*\),v.*:checking in src/\1,v:g'
+cd ..
+
+echo % convert again
+hg convert src src-hg | sed -e 's/connecting to.*cvsrepo/connecting to cvsrepo/g'
+cat src-hg/a
+cat src-hg/b/c
+
+echo % convert again with --filemap
+hg convert --filemap filemap src src-filemap | sed -e 's/connecting to.*cvsrepo/connecting to cvsrepo/g'
+cat src-hg/b/c
+hg -R src-filemap log --template '#rev# #desc# files: #files#\n'
+
+echo "graphlog = " >> $HGRCPATH
+hg -R src-hg glog --template '#rev# (#branches#) #desc# files: #files#\n'

tests/test-convert-cvs.out

@@ -1,69 +1,109 b''
 % create cvs repository
 % create source directory
 % import source directory
 N src/a
 N src/b/c
 
 No conflicts created by this import
 
 % checkout source directory
 U src/a
 U src/b/c
 % commit a new revision changing b/c
 checking in src/b/c,v
 % convert fresh repo
 initializing destination src-hg repository
 connecting to cvsrepo
 scanning source...
 sorting...
 converting...
 2 Initial revision
 1 import
 0 ci0
 updating tags
 a
 c
 c
 % convert fresh repo with --filemap
 initializing destination src-filemap repository
 connecting to cvsrepo
 scanning source...
 sorting...
 converting...
 2 Initial revision
 1 import
 rolling back last transaction
 0 ci0
 updating tags
 c
 c
 2 update tags files: .hgtags
 1 ci0 files: b/c
 0 Initial revision files: b/c
 % commit new file revisions
 checking in src/a,v
 checking in src/b/c,v
 % convert again
 connecting to cvsrepo
 scanning source...
 sorting...
 converting...
 0 ci1
 a
 a
 c
 c
 c
 % convert again with --filemap
 connecting to cvsrepo
 scanning source...
 sorting...
 converting...
 0 ci1
 c
 c
 c
 3 ci1 files: b/c
 2 update tags files: .hgtags
 1 ci0 files: b/c
 0 Initial revision files: b/c
+% commit branch
+U b/c
+T a
+T b/c
+checking in src/b/c,v
+% convert again
+connecting to cvsrepo
+scanning source...
+sorting...
+converting...
+0 ci2
+a
+a
+c
+d
+% convert again with --filemap
+connecting to cvsrepo
+scanning source...
+sorting...
+converting...
+0 ci2
+c
+d
+4 ci2 files: b/c
+3 ci1 files: b/c
+2 update tags files: .hgtags
+1 ci0 files: b/c
+0 Initial revision files: b/c
+o  5 (branch) ci2 files: b/c
+|
+| o  4 () ci1 files: a b/c
+| |
+| o  3 () update tags files: .hgtags
+| |
+| o  2 () ci0 files: b/c
+|/
+| o  1 (INITIAL) import files:
+|/
+o  0 () Initial revision files: a b/c
+