Fix file-changed-to-dir and dir-to-file commits (issue660).

Allow adding to dirstate files that clash with previously existing
but marked for removal. Protect from reintroducing clashes by revert.

This change doesn't address related issues with update. Current
workaround is to do "clean" update by manually removing conflicting
files/dirs from working directory.

tests/test-issue660

@@ -0,0 +1,89 b''
+#!/bin/sh
+# http://www.selenic.com/mercurial/bts/issue660
+
+
+hg init a
+cd a
+echo a > a
+mkdir b
+echo b > b/b
+hg commit -A -m "a is file, b is dir"
+
+echo % file replaced with directory
+
+rm a
+mkdir a
+echo a > a/a
+
+echo % should fail - would corrupt dirstate
+hg add a/a 
+
+echo % removing shadow
+hg rm --after a
+
+echo % should succeed - shadow removed
+hg add a/a
+
+echo % directory replaced with file
+
+rm -r b
+echo b > b
+
+echo % should fail - would corrupt dirstate
+hg add b
+
+echo % removing shadow
+hg rm --after b/b
+
+echo % should succeed - shadow removed
+hg add b
+
+echo % look what we got
+hg st
+
+echo % revert reintroducing shadow - should fail
+rm -r a b
+hg revert b/b
+
+echo % revert all - should succeed
+hg revert --all
+hg st
+
+echo % addremove
+
+rm -r a b
+mkdir a
+echo a > a/a
+echo b > b
+
+hg addremove
+hg st
+
+echo % commit
+hg ci -A -m "a is dir, b is file"
+hg st --all
+
+echo % long directory replaced with file
+
+mkdir d
+mkdir d/d
+echo d > d/d/d
+hg commit -A -m "d is long directory"
+rm -r d
+echo d > d
+
+echo % should fail - would corrupt dirstate
+hg add d
+
+echo % removing shadow
+hg rm --after d/d/d
+
+echo % should succeed - shadow removed
+hg add d
+
+#echo % update should work
+#
+#hg up -r 0
+#hg up -r 1
+
+exit 0

tests/test-issue660.out

@@ -0,0 +1,42 b''
+adding a
+adding b/b
+% file replaced with directory
+% should fail - would corrupt dirstate
+abort: file 'a' in dirstate clashes with 'a/a'
+% removing shadow
+% should succeed - shadow removed
+% directory replaced with file
+% should fail - would corrupt dirstate
+abort: directory 'b' already in dirstate
+% removing shadow
+% should succeed - shadow removed
+% look what we got
+A a/a
+A b
+R a
+R b/b
+% revert reintroducing shadow - should fail
+abort: file 'b' in dirstate clashes with 'b/b'
+% revert all - should succeed
+undeleting a
+forgetting a/a
+forgetting b
+undeleting b/b
+% addremove
+removing a
+adding a/a
+adding b
+removing b/b
+A a/a
+A b
+R a
+R b/b
+% commit
+C a/a
+C b
+% long directory replaced with file
+adding d/d/d
+% should fail - would corrupt dirstate
+abort: directory 'd' already in dirstate
+% removing shadow
+% should succeed - shadow removed

mercurial/cmdutil.py

@@ -1,943 +1,944 b''
 # cmdutil.py - help for command processing in mercurial
 #
 # Copyright 2005-2007 Matt Mackall <[e-mail]>
 #
 # This software may be used and distributed according to the terms
 # of the GNU General Public License, incorporated herein by reference.
 
 from node import *
 from i18n import _
 import os, sys, bisect, stat
 import mdiff, bdiff, util, templater, patch
 
 revrangesep = ':'
 
 class UnknownCommand(Exception):
     """Exception raised if command is not in the command table."""
 class AmbiguousCommand(Exception):
     """Exception raised if command shortcut matches more than one command."""
 
 def findpossible(ui, cmd, table):
     """
     Return cmd -> (aliases, command table entry)
     for each matching command.
     Return debug commands (or their aliases) only if no normal command matches.
     """
     choice = {}
     debugchoice = {}
     for e in table.keys():
         aliases = e.lstrip("^").split("|")
         found = None
         if cmd in aliases:
             found = cmd
         elif not ui.config("ui", "strict"):
             for a in aliases:
                 if a.startswith(cmd):
                     found = a
                     break
         if found is not None:
             if aliases[0].startswith("debug") or found.startswith("debug"):
                 debugchoice[found] = (aliases, table[e])
             else:
                 choice[found] = (aliases, table[e])
 
     if not choice and debugchoice:
         choice = debugchoice
 
     return choice
 
 def findcmd(ui, cmd, table):
     """Return (aliases, command table entry) for command string."""
     choice = findpossible(ui, cmd, table)
 
     if choice.has_key(cmd):
         return choice[cmd]
 
     if len(choice) > 1:
         clist = choice.keys()
         clist.sort()
         raise AmbiguousCommand(cmd, clist)
 
     if choice:
         return choice.values()[0]
 
     raise UnknownCommand(cmd)
 
 def bail_if_changed(repo):
     modified, added, removed, deleted = repo.status()[:4]
     if modified or added or removed or deleted:
         raise util.Abort(_("outstanding uncommitted changes"))
 
 def logmessage(opts):
     """ get the log message according to -m and -l option """
     message = opts['message']
     logfile = opts['logfile']
 
     if message and logfile:
         raise util.Abort(_('options --message and --logfile are mutually '
                            'exclusive'))
     if not message and logfile:
         try:
             if logfile == '-':
                 message = sys.stdin.read()
             else:
                 message = open(logfile).read()
         except IOError, inst:
             raise util.Abort(_("can't read commit message '%s': %s") %
                              (logfile, inst.strerror))
     return message
 
 def setremoteconfig(ui, opts):
     "copy remote options to ui tree"
     if opts.get('ssh'):
         ui.setconfig("ui", "ssh", opts['ssh'])
     if opts.get('remotecmd'):
         ui.setconfig("ui", "remotecmd", opts['remotecmd'])
 
 def revpair(repo, revs):
     '''return pair of nodes, given list of revisions. second item can
     be None, meaning use working dir.'''
 
     def revfix(repo, val, defval):
         if not val and val != 0 and defval is not None:
             val = defval
         return repo.lookup(val)
 
     if not revs:
         return repo.dirstate.parents()[0], None
     end = None
     if len(revs) == 1:
         if revrangesep in revs[0]:
             start, end = revs[0].split(revrangesep, 1)
             start = revfix(repo, start, 0)
             end = revfix(repo, end, repo.changelog.count() - 1)
         else:
             start = revfix(repo, revs[0], None)
     elif len(revs) == 2:
         if revrangesep in revs[0] or revrangesep in revs[1]:
             raise util.Abort(_('too many revisions specified'))
         start = revfix(repo, revs[0], None)
         end = revfix(repo, revs[1], None)
     else:
         raise util.Abort(_('too many revisions specified'))
     return start, end
 
 def revrange(repo, revs):
     """Yield revision as strings from a list of revision specifications."""
 
     def revfix(repo, val, defval):
         if not val and val != 0 and defval is not None:
             return defval
         return repo.changelog.rev(repo.lookup(val))
 
     seen, l = {}, []
     for spec in revs:
         if revrangesep in spec:
             start, end = spec.split(revrangesep, 1)
             start = revfix(repo, start, 0)
             end = revfix(repo, end, repo.changelog.count() - 1)
             step = start > end and -1 or 1
             for rev in xrange(start, end+step, step):
                 if rev in seen:
                     continue
                 seen[rev] = 1
                 l.append(rev)
         else:
             rev = revfix(repo, spec, None)
             if rev in seen:
                 continue
             seen[rev] = 1
             l.append(rev)
 
     return l
 
 def make_filename(repo, pat, node,
                   total=None, seqno=None, revwidth=None, pathname=None):
     node_expander = {
         'H': lambda: hex(node),
         'R': lambda: str(repo.changelog.rev(node)),
         'h': lambda: short(node),
         }
     expander = {
         '%': lambda: '%',
         'b': lambda: os.path.basename(repo.root),
         }
 
     try:
         if node:
             expander.update(node_expander)
         if node:
             expander['r'] = (lambda:
                     str(repo.changelog.rev(node)).zfill(revwidth or 0))
         if total is not None:
             expander['N'] = lambda: str(total)
         if seqno is not None:
             expander['n'] = lambda: str(seqno)
         if total is not None and seqno is not None:
             expander['n'] = lambda: str(seqno).zfill(len(str(total)))
         if pathname is not None:
             expander['s'] = lambda: os.path.basename(pathname)
             expander['d'] = lambda: os.path.dirname(pathname) or '.'
             expander['p'] = lambda: pathname
 
         newname = []
         patlen = len(pat)
         i = 0
         while i < patlen:
             c = pat[i]
             if c == '%':
                 i += 1
                 c = pat[i]
                 c = expander[c]()
             newname.append(c)
             i += 1
         return ''.join(newname)
     except KeyError, inst:
         raise util.Abort(_("invalid format spec '%%%s' in output file name") %
                          inst.args[0])
 
 def make_file(repo, pat, node=None,
               total=None, seqno=None, revwidth=None, mode='wb', pathname=None):
     if not pat or pat == '-':
         return 'w' in mode and sys.stdout or sys.stdin
     if hasattr(pat, 'write') and 'w' in mode:
         return pat
     if hasattr(pat, 'read') and 'r' in mode:
         return pat
     return open(make_filename(repo, pat, node, total, seqno, revwidth,
                               pathname),
                 mode)
 
 def matchpats(repo, pats=[], opts={}, globbed=False, default=None):
     cwd = repo.getcwd()
     return util.cmdmatcher(repo.root, cwd, pats or [], opts.get('include'),
                            opts.get('exclude'), globbed=globbed,
                            default=default)
 
 def walk(repo, pats=[], opts={}, node=None, badmatch=None, globbed=False,
          default=None):
     files, matchfn, anypats = matchpats(repo, pats, opts, globbed=globbed,
                                         default=default)
     exact = dict.fromkeys(files)
     cwd = repo.getcwd()
     for src, fn in repo.walk(node=node, files=files, match=matchfn,
                              badmatch=badmatch):
         yield src, fn, repo.pathto(fn, cwd), fn in exact
 
 def findrenames(repo, added=None, removed=None, threshold=0.5):
     '''find renamed files -- yields (before, after, score) tuples'''
     if added is None or removed is None:
         added, removed = repo.status()[1:3]
     ctx = repo.changectx()
     for a in added:
         aa = repo.wread(a)
         bestname, bestscore = None, threshold
         for r in removed:
             rr = ctx.filectx(r).data()
 
             # bdiff.blocks() returns blocks of matching lines
             # count the number of bytes in each
             equal = 0
             alines = mdiff.splitnewlines(aa)
             matches = bdiff.blocks(aa, rr)
             for x1,x2,y1,y2 in matches:
                 for line in alines[x1:x2]:
                     equal += len(line)
 
             lengths = len(aa) + len(rr)
             if lengths:
                 myscore = equal*2.0 / lengths
                 if myscore >= bestscore:
                     bestname, bestscore = r, myscore
         if bestname:
             yield bestname, a, bestscore
 
 def addremove(repo, pats=[], opts={}, dry_run=None, similarity=None):
     if dry_run is None:
         dry_run = opts.get('dry_run')
     if similarity is None:
         similarity = float(opts.get('similarity') or 0)
     add, remove = [], []
     mapping = {}
     for src, abs, rel, exact in walk(repo, pats, opts):
         target = repo.wjoin(abs)
         if src == 'f' and abs not in repo.dirstate:
             add.append(abs)
             mapping[abs] = rel, exact
             if repo.ui.verbose or not exact:
                 repo.ui.status(_('adding %s\n') % ((pats and rel) or abs))
-        if repo.dirstate[abs] != 'r' and not util.lexists(target):
+        if repo.dirstate[abs] != 'r' and (not util.lexists(target)
+            or (os.path.isdir(target) and not os.path.islink(target))):
             remove.append(abs)
             mapping[abs] = rel, exact
             if repo.ui.verbose or not exact:
                 repo.ui.status(_('removing %s\n') % ((pats and rel) or abs))
     if not dry_run:
+        repo.remove(remove)
         repo.add(add)
-        repo.remove(remove)
     if similarity > 0:
         for old, new, score in findrenames(repo, add, remove, similarity):
             oldrel, oldexact = mapping[old]
             newrel, newexact = mapping[new]
             if repo.ui.verbose or not oldexact or not newexact:
                 repo.ui.status(_('recording removal of %s as rename to %s '
                                  '(%d%% similar)\n') %
                                (oldrel, newrel, score * 100))
             if not dry_run:
                 repo.copy(old, new)
 
 def service(opts, parentfn=None, initfn=None, runfn=None):
     '''Run a command as a service.'''
 
     if opts['daemon'] and not opts['daemon_pipefds']:
         rfd, wfd = os.pipe()
         args = sys.argv[:]
         args.append('--daemon-pipefds=%d,%d' % (rfd, wfd))
         pid = os.spawnvp(os.P_NOWAIT | getattr(os, 'P_DETACH', 0),
                          args[0], args)
         os.close(wfd)
         os.read(rfd, 1)
         if parentfn:
             return parentfn(pid)
         else:
             os._exit(0)
 
     if initfn:
         initfn()
 
     if opts['pid_file']:
         fp = open(opts['pid_file'], 'w')
         fp.write(str(os.getpid()) + '\n')
         fp.close()
 
     if opts['daemon_pipefds']:
         rfd, wfd = [int(x) for x in opts['daemon_pipefds'].split(',')]
         os.close(rfd)
         try:
             os.setsid()
         except AttributeError:
             pass
         os.write(wfd, 'y')
         os.close(wfd)
         sys.stdout.flush()
         sys.stderr.flush()
         fd = os.open(util.nulldev, os.O_RDWR)
         if fd != 0: os.dup2(fd, 0)
         if fd != 1: os.dup2(fd, 1)
         if fd != 2: os.dup2(fd, 2)
         if fd not in (0, 1, 2): os.close(fd)
 
     if runfn:
         return runfn()
 
 class changeset_printer(object):
     '''show changeset information when templating not requested.'''
 
     def __init__(self, ui, repo, patch, buffered):
         self.ui = ui
         self.repo = repo
         self.buffered = buffered
         self.patch = patch
         self.header = {}
         self.hunk = {}
         self.lastheader = None
 
     def flush(self, rev):
         if rev in self.header:
             h = self.header[rev]
             if h != self.lastheader:
                 self.lastheader = h
                 self.ui.write(h)
             del self.header[rev]
         if rev in self.hunk:
             self.ui.write(self.hunk[rev])
             del self.hunk[rev]
             return 1
         return 0
 
     def show(self, rev=0, changenode=None, copies=(), **props):
         if self.buffered:
             self.ui.pushbuffer()
             self._show(rev, changenode, copies, props)
             self.hunk[rev] = self.ui.popbuffer()
         else:
             self._show(rev, changenode, copies, props)
 
     def _show(self, rev, changenode, copies, props):
         '''show a single changeset or file revision'''
         log = self.repo.changelog
         if changenode is None:
             changenode = log.node(rev)
         elif not rev:
             rev = log.rev(changenode)
 
         if self.ui.quiet:
             self.ui.write("%d:%s\n" % (rev, short(changenode)))
             return
 
         changes = log.read(changenode)
         date = util.datestr(changes[2])
         extra = changes[5]
         branch = extra.get("branch")
 
         hexfunc = self.ui.debugflag and hex or short
 
         parents = [(p, hexfunc(log.node(p)))
                    for p in self._meaningful_parentrevs(log, rev)]
 
         self.ui.write(_("changeset:   %d:%s\n") % (rev, hexfunc(changenode)))
 
         # don't show the default branch name
         if branch != 'default':
             branch = util.tolocal(branch)
             self.ui.write(_("branch:      %s\n") % branch)
         for tag in self.repo.nodetags(changenode):
             self.ui.write(_("tag:         %s\n") % tag)
         for parent in parents:
             self.ui.write(_("parent:      %d:%s\n") % parent)
 
         if self.ui.debugflag:
             self.ui.write(_("manifest:    %d:%s\n") %
                           (self.repo.manifest.rev(changes[0]), hex(changes[0])))
         self.ui.write(_("user:        %s\n") % changes[1])
         self.ui.write(_("date:        %s\n") % date)
 
         if self.ui.debugflag:
             files = self.repo.status(log.parents(changenode)[0], changenode)[:3]
             for key, value in zip([_("files:"), _("files+:"), _("files-:")],
                                   files):
                 if value:
                     self.ui.write("%-12s %s\n" % (key, " ".join(value)))
         elif changes[3] and self.ui.verbose:
             self.ui.write(_("files:       %s\n") % " ".join(changes[3]))
         if copies and self.ui.verbose:
             copies = ['%s (%s)' % c for c in copies]
             self.ui.write(_("copies:      %s\n") % ' '.join(copies))
 
         if extra and self.ui.debugflag:
             extraitems = extra.items()
             extraitems.sort()
             for key, value in extraitems:
                 self.ui.write(_("extra:       %s=%s\n")
                               % (key, value.encode('string_escape')))
 
         description = changes[4].strip()
         if description:
             if self.ui.verbose:
                 self.ui.write(_("description:\n"))
                 self.ui.write(description)
                 self.ui.write("\n\n")
             else:
                 self.ui.write(_("summary:     %s\n") %
                               description.splitlines()[0])
         self.ui.write("\n")
 
         self.showpatch(changenode)
 
     def showpatch(self, node):
         if self.patch:
             prev = self.repo.changelog.parents(node)[0]
             patch.diff(self.repo, prev, node, match=self.patch, fp=self.ui,
                        opts=patch.diffopts(self.ui))
             self.ui.write("\n")
 
     def _meaningful_parentrevs(self, log, rev):
         """Return list of meaningful (or all if debug) parentrevs for rev.
 
         For merges (two non-nullrev revisions) both parents are meaningful.
         Otherwise the first parent revision is considered meaningful if it
         is not the preceding revision.
         """
         parents = log.parentrevs(rev)
         if not self.ui.debugflag and parents[1] == nullrev:
             if parents[0] >= rev - 1:
                 parents = []
             else:
                 parents = [parents[0]]
         return parents
 
 
 class changeset_templater(changeset_printer):
     '''format changeset information.'''
 
     def __init__(self, ui, repo, patch, mapfile, buffered):
         changeset_printer.__init__(self, ui, repo, patch, buffered)
         filters = templater.common_filters.copy()
         filters['formatnode'] = (ui.debugflag and (lambda x: x)
                                  or (lambda x: x[:12]))
         self.t = templater.templater(mapfile, filters,
                                      cache={
                                          'parent': '{rev}:{node|formatnode} ',
                                          'manifest': '{rev}:{node|formatnode}',
                                          'filecopy': '{name} ({source})'})
 
     def use_template(self, t):
         '''set template string to use'''
         self.t.cache['changeset'] = t
 
     def _show(self, rev, changenode, copies, props):
         '''show a single changeset or file revision'''
         log = self.repo.changelog
         if changenode is None:
             changenode = log.node(rev)
         elif not rev:
             rev = log.rev(changenode)
 
         changes = log.read(changenode)
 
         def showlist(name, values, plural=None, **args):
             '''expand set of values.
             name is name of key in template map.
             values is list of strings or dicts.
             plural is plural of name, if not simply name + 's'.
 
             expansion works like this, given name 'foo'.
 
             if values is empty, expand 'no_foos'.
 
             if 'foo' not in template map, return values as a string,
             joined by space.
 
             expand 'start_foos'.
 
             for each value, expand 'foo'. if 'last_foo' in template
             map, expand it instead of 'foo' for last key.
 
             expand 'end_foos'.
             '''
             if plural: names = plural
             else: names = name + 's'
             if not values:
                 noname = 'no_' + names
                 if noname in self.t:
                     yield self.t(noname, **args)
                 return
             if name not in self.t:
                 if isinstance(values[0], str):
                     yield ' '.join(values)
                 else:
                     for v in values:
                         yield dict(v, **args)
                 return
             startname = 'start_' + names
             if startname in self.t:
                 yield self.t(startname, **args)
             vargs = args.copy()
             def one(v, tag=name):
                 try:
                     vargs.update(v)
                 except (AttributeError, ValueError):
                     try:
                         for a, b in v:
                             vargs[a] = b
                     except ValueError:
                         vargs[name] = v
                 return self.t(tag, **vargs)
             lastname = 'last_' + name
             if lastname in self.t:
                 last = values.pop()
             else:
                 last = None
             for v in values:
                 yield one(v)
             if last is not None:
                 yield one(last, tag=lastname)
             endname = 'end_' + names
             if endname in self.t:
                 yield self.t(endname, **args)
 
         def showbranches(**args):
             branch = changes[5].get("branch")
             if branch != 'default':
                 branch = util.tolocal(branch)
                 return showlist('branch', [branch], plural='branches', **args)
 
         def showparents(**args):
             parents = [[('rev', p), ('node', hex(log.node(p)))]
                        for p in self._meaningful_parentrevs(log, rev)]
             return showlist('parent', parents, **args)
 
         def showtags(**args):
             return showlist('tag', self.repo.nodetags(changenode), **args)
 
         def showextras(**args):
             extras = changes[5].items()
             extras.sort()
             for key, value in extras:
                 args = args.copy()
                 args.update(dict(key=key, value=value))
                 yield self.t('extra', **args)
 
         def showcopies(**args):
             c = [{'name': x[0], 'source': x[1]} for x in copies]
             return showlist('file_copy', c, plural='file_copies', **args)
 
         if self.ui.debugflag:
             files = self.repo.status(log.parents(changenode)[0], changenode)[:3]
             def showfiles(**args):
                 return showlist('file', files[0], **args)
             def showadds(**args):
                 return showlist('file_add', files[1], **args)
             def showdels(**args):
                 return showlist('file_del', files[2], **args)
             def showmanifest(**args):
                 args = args.copy()
                 args.update(dict(rev=self.repo.manifest.rev(changes[0]),
                                  node=hex(changes[0])))
                 return self.t('manifest', **args)
         else:
             def showfiles(**args):
                 return showlist('file', changes[3], **args)
             showadds = ''
             showdels = ''
             showmanifest = ''
 
         defprops = {
             'author': changes[1],
             'branches': showbranches,
             'date': changes[2],
             'desc': changes[4].strip(),
             'file_adds': showadds,
             'file_dels': showdels,
             'files': showfiles,
             'file_copies': showcopies,
             'manifest': showmanifest,
             'node': hex(changenode),
             'parents': showparents,
             'rev': rev,
             'tags': showtags,
             'extras': showextras,
             }
         props = props.copy()
         props.update(defprops)
 
         try:
             if self.ui.debugflag and 'header_debug' in self.t:
                 key = 'header_debug'
             elif self.ui.quiet and 'header_quiet' in self.t:
                 key = 'header_quiet'
             elif self.ui.verbose and 'header_verbose' in self.t:
                 key = 'header_verbose'
             elif 'header' in self.t:
                 key = 'header'
             else:
                 key = ''
             if key:
                 h = templater.stringify(self.t(key, **props))
                 if self.buffered:
                     self.header[rev] = h
                 else:
                     self.ui.write(h)
             if self.ui.debugflag and 'changeset_debug' in self.t:
                 key = 'changeset_debug'
             elif self.ui.quiet and 'changeset_quiet' in self.t:
                 key = 'changeset_quiet'
             elif self.ui.verbose and 'changeset_verbose' in self.t:
                 key = 'changeset_verbose'
             else:
                 key = 'changeset'
             self.ui.write(templater.stringify(self.t(key, **props)))
             self.showpatch(changenode)
         except KeyError, inst:
             raise util.Abort(_("%s: no key named '%s'") % (self.t.mapfile,
                                                            inst.args[0]))
         except SyntaxError, inst:
             raise util.Abort(_('%s: %s') % (self.t.mapfile, inst.args[0]))
 
 def show_changeset(ui, repo, opts, buffered=False, matchfn=False):
     """show one changeset using template or regular display.
 
     Display format will be the first non-empty hit of:
     1. option 'template'
     2. option 'style'
     3. [ui] setting 'logtemplate'
     4. [ui] setting 'style'
     If all of these values are either the unset or the empty string,
     regular display via changeset_printer() is done.
     """
     # options
     patch = False
     if opts.get('patch'):
         patch = matchfn or util.always
 
     tmpl = opts.get('template')
     mapfile = None
     if tmpl:
         tmpl = templater.parsestring(tmpl, quoted=False)
     else:
         mapfile = opts.get('style')
         # ui settings
         if not mapfile:
             tmpl = ui.config('ui', 'logtemplate')
             if tmpl:
                 tmpl = templater.parsestring(tmpl)
             else:
                 mapfile = ui.config('ui', 'style')
 
     if tmpl or mapfile:
         if mapfile:
             if not os.path.split(mapfile)[0]:
                 mapname = (templater.templatepath('map-cmdline.' + mapfile)
                            or templater.templatepath(mapfile))
                 if mapname: mapfile = mapname
         try:
             t = changeset_templater(ui, repo, patch, mapfile, buffered)
         except SyntaxError, inst:
             raise util.Abort(inst.args[0])
         if tmpl: t.use_template(tmpl)
         return t
     return changeset_printer(ui, repo, patch, buffered)
 
 def finddate(ui, repo, date):
     """Find the tipmost changeset that matches the given date spec"""
     df = util.matchdate(date + " to " + date)
     get = util.cachefunc(lambda r: repo.changectx(r).changeset())
     changeiter, matchfn = walkchangerevs(ui, repo, [], get, {'rev':None})
     results = {}
     for st, rev, fns in changeiter:
         if st == 'add':
             d = get(rev)[2]
             if df(d[0]):
                 results[rev] = d
         elif st == 'iter':
             if rev in results:
                 ui.status("Found revision %s from %s\n" %
                           (rev, util.datestr(results[rev])))
                 return str(rev)
 
     raise util.Abort(_("revision matching date not found"))
 
 def walkchangerevs(ui, repo, pats, change, opts):
     '''Iterate over files and the revs they changed in.
 
     Callers most commonly need to iterate backwards over the history
     it is interested in.  Doing so has awful (quadratic-looking)
     performance, so we use iterators in a "windowed" way.
 
     We walk a window of revisions in the desired order.  Within the
     window, we first walk forwards to gather data, then in the desired
     order (usually backwards) to display it.
 
     This function returns an (iterator, matchfn) tuple. The iterator
     yields 3-tuples. They will be of one of the following forms:
 
     "window", incrementing, lastrev: stepping through a window,
     positive if walking forwards through revs, last rev in the
     sequence iterated over - use to reset state for the current window
 
     "add", rev, fns: out-of-order traversal of the given file names
     fns, which changed during revision rev - use to gather data for
     possible display
 
     "iter", rev, None: in-order traversal of the revs earlier iterated
     over with "add" - use to display data'''
 
     def increasing_windows(start, end, windowsize=8, sizelimit=512):
         if start < end:
             while start < end:
                 yield start, min(windowsize, end-start)
                 start += windowsize
                 if windowsize < sizelimit:
                     windowsize *= 2
         else:
             while start > end:
                 yield start, min(windowsize, start-end-1)
                 start -= windowsize
                 if windowsize < sizelimit:
                     windowsize *= 2
 
     files, matchfn, anypats = matchpats(repo, pats, opts)
     follow = opts.get('follow') or opts.get('follow_first')
 
     if repo.changelog.count() == 0:
         return [], matchfn
 
     if follow:
         defrange = '%s:0' % repo.changectx().rev()
     else:
         defrange = 'tip:0'
     revs = revrange(repo, opts['rev'] or [defrange])
     wanted = {}
     slowpath = anypats or opts.get('removed')
     fncache = {}
 
     if not slowpath and not files:
         # No files, no patterns.  Display all revs.
         wanted = dict.fromkeys(revs)
     copies = []
     if not slowpath:
         # Only files, no patterns.  Check the history of each file.
         def filerevgen(filelog, node):
             cl_count = repo.changelog.count()
             if node is None:
                 last = filelog.count() - 1
             else:
                 last = filelog.rev(node)
             for i, window in increasing_windows(last, nullrev):
                 revs = []
                 for j in xrange(i - window, i + 1):
                     n = filelog.node(j)
                     revs.append((filelog.linkrev(n),
                                  follow and filelog.renamed(n)))
                 revs.reverse()
                 for rev in revs:
                     # only yield rev for which we have the changelog, it can
                     # happen while doing "hg log" during a pull or commit
                     if rev[0] < cl_count:
                         yield rev
         def iterfiles():
             for filename in files:
                 yield filename, None
             for filename_node in copies:
                 yield filename_node
         minrev, maxrev = min(revs), max(revs)
         for file_, node in iterfiles():
             filelog = repo.file(file_)
             # A zero count may be a directory or deleted file, so
             # try to find matching entries on the slow path.
             if filelog.count() == 0:
                 slowpath = True
                 break
             for rev, copied in filerevgen(filelog, node):
                 if rev <= maxrev:
                     if rev < minrev:
                         break
                     fncache.setdefault(rev, [])
                     fncache[rev].append(file_)
                     wanted[rev] = 1
                     if follow and copied:
                         copies.append(copied)
     if slowpath:
         if follow:
             raise util.Abort(_('can only follow copies/renames for explicit '
                                'file names'))
 
         # The slow path checks files modified in every changeset.
         def changerevgen():
             for i, window in increasing_windows(repo.changelog.count()-1,
                                                 nullrev):
                 for j in xrange(i - window, i + 1):
                     yield j, change(j)[3]
 
         for rev, changefiles in changerevgen():
             matches = filter(matchfn, changefiles)
             if matches:
                 fncache[rev] = matches
                 wanted[rev] = 1
 
     class followfilter:
         def __init__(self, onlyfirst=False):
             self.startrev = nullrev
             self.roots = []
             self.onlyfirst = onlyfirst
 
         def match(self, rev):
             def realparents(rev):
                 if self.onlyfirst:
                     return repo.changelog.parentrevs(rev)[0:1]
                 else:
                     return filter(lambda x: x != nullrev,
                                   repo.changelog.parentrevs(rev))
 
             if self.startrev == nullrev:
                 self.startrev = rev
                 return True
 
             if rev > self.startrev:
                 # forward: all descendants
                 if not self.roots:
                     self.roots.append(self.startrev)
                 for parent in realparents(rev):
                     if parent in self.roots:
                         self.roots.append(rev)
                         return True
             else:
                 # backwards: all parents
                 if not self.roots:
                     self.roots.extend(realparents(self.startrev))
                 if rev in self.roots:
                     self.roots.remove(rev)
                     self.roots.extend(realparents(rev))
                     return True
 
             return False
 
     # it might be worthwhile to do this in the iterator if the rev range
     # is descending and the prune args are all within that range
     for rev in opts.get('prune', ()):
         rev = repo.changelog.rev(repo.lookup(rev))
         ff = followfilter()
         stop = min(revs[0], revs[-1])
         for x in xrange(rev, stop-1, -1):
             if ff.match(x) and x in wanted:
                 del wanted[x]
 
     def iterate():
         if follow and not files:
             ff = followfilter(onlyfirst=opts.get('follow_first'))
             def want(rev):
                 if ff.match(rev) and rev in wanted:
                     return True
                 return False
         else:
             def want(rev):
                 return rev in wanted
 
         for i, window in increasing_windows(0, len(revs)):
             yield 'window', revs[0] < revs[-1], revs[-1]
             nrevs = [rev for rev in revs[i:i+window] if want(rev)]
             srevs = list(nrevs)
             srevs.sort()
             for rev in srevs:
                 fns = fncache.get(rev)
                 if not fns:
                     def fns_generator():
                         for f in change(rev)[3]:
                             if matchfn(f):
                                 yield f
                     fns = fns_generator()
                 yield 'add', rev, fns
             for rev in nrevs:
                 yield 'iter', rev, None
     return iterate(), matchfn
 
 def commit(ui, repo, commitfunc, pats, opts):
     '''commit the specified files or all outstanding changes'''
     message = logmessage(opts)
 
     if opts['addremove']:
         addremove(repo, pats, opts)
     fns, match, anypats = matchpats(repo, pats, opts)
     if pats:
         status = repo.status(files=fns, match=match)
         modified, added, removed, deleted, unknown = status[:5]
         files = modified + added + removed
         slist = None
         for f in fns:
             if f == '.':
                 continue
             if f not in files:
                 rf = repo.wjoin(f)
                 try:
                     mode = os.lstat(rf)[stat.ST_MODE]
                 except OSError:
                     raise util.Abort(_("file %s not found!") % rf)
                 if stat.S_ISDIR(mode):
                     name = f + '/'
                     if slist is None:
                         slist = list(files)
                         slist.sort()
                     i = bisect.bisect(slist, name)
                     if i >= len(slist) or not slist[i].startswith(name):
                         raise util.Abort(_("no match under directory %s!")
                                          % rf)
                 elif not (stat.S_ISREG(mode) or stat.S_ISLNK(mode)):
                     raise util.Abort(_("can't commit %s: "
                                        "unsupported file type!") % rf)
                 elif f not in repo.dirstate:
                     raise util.Abort(_("file %s not tracked!") % rf)
     else:
         files = []
     try:
         return commitfunc(ui, repo, files, message, match, opts)
     except ValueError, inst:
         raise util.Abort(str(inst))

mercurial/dirstate.py

@@ -1,557 +1,574 b''
 """
 dirstate.py - working directory tracking for mercurial
 
 Copyright 2005-2007 Matt Mackall <[e-mail]>
 
 This software may be used and distributed according to the terms
 of the GNU General Public License, incorporated herein by reference.
 """
 
 from node import *
 from i18n import _
 import struct, os, time, bisect, stat, strutil, util, re, errno, ignore
 import cStringIO, osutil
 
 _unknown = ('?', 0, 0, 0)
 _format = ">cllll"
 
 class dirstate(object):
 
     def __init__(self, opener, ui, root):
         self._opener = opener
         self._root = root
         self._dirty = False
         self._dirtypl = False
         self._ui = ui
 
     def __getattr__(self, name):
         if name == '_map':
             self._read()
             return self._map
         elif name == '_copymap':
             self._read()
             return self._copymap
         elif name == '_branch':
             try:
                 self._branch = (self._opener("branch").read().strip()
                                 or "default")
             except IOError:
                 self._branch = "default"
             return self._branch
         elif name == '_pl':
             self._pl = [nullid, nullid]
             try:
                 st = self._opener("dirstate").read(40)
                 if len(st) == 40:
                     self._pl = st[:20], st[20:40]
             except IOError, err:
                 if err.errno != errno.ENOENT: raise
             return self._pl
         elif name == '_dirs':
             self._dirs = {}
             for f in self._map:
-                self._incpath(f)
+                if self[f] != 'r':
+                    self._incpath(f)
             return self._dirs
         elif name == '_ignore':
             files = [self._join('.hgignore')]
             for name, path in self._ui.configitems("ui"):
                 if name == 'ignore' or name.startswith('ignore.'):
                     files.append(os.path.expanduser(path))
             self._ignore = ignore.ignore(self._root, files, self._ui.warn)
             return self._ignore
         elif name == '_slash':
             self._slash = self._ui.configbool('ui', 'slash') and os.sep != '/'
             return self._slash
         else:
             raise AttributeError, name
 
     def _join(self, f):
         return os.path.join(self._root, f)
 
     def getcwd(self):
         cwd = os.getcwd()
         if cwd == self._root: return ''
         # self._root ends with a path separator if self._root is '/' or 'C:\'
         rootsep = self._root
         if not rootsep.endswith(os.sep):
             rootsep += os.sep
         if cwd.startswith(rootsep):
             return cwd[len(rootsep):]
         else:
             # we're outside the repo. return an absolute path.
             return cwd
 
     def pathto(self, f, cwd=None):
         if cwd is None:
             cwd = self.getcwd()
         path = util.pathto(self._root, cwd, f)
         if self._slash:
             return path.replace(os.sep, '/')
         return path
 
     def __getitem__(self, key):
         ''' current states:
         n  normal
         m  needs merging
         r  marked for removal
         a  marked for addition
         ?  not tracked'''
         return self._map.get(key, ("?",))[0]
 
     def __contains__(self, key):
         return key in self._map
 
     def __iter__(self):
         a = self._map.keys()
         a.sort()
         for x in a:
             yield x
 
     def parents(self):
         return self._pl
 
     def branch(self):
         return self._branch
 
     def setparents(self, p1, p2=nullid):
         self._dirty = self._dirtypl = True
         self._pl = p1, p2
 
     def setbranch(self, branch):
         self._branch = branch
         self._opener("branch", "w").write(branch + '\n')
 
     def _read(self):
         self._map = {}
         self._copymap = {}
         if not self._dirtypl:
             self._pl = [nullid, nullid]
         try:
             st = self._opener("dirstate").read()
         except IOError, err:
             if err.errno != errno.ENOENT: raise
             return
         if not st:
             return
 
         if not self._dirtypl:
             self._pl = [st[:20], st[20: 40]]
 
         # deref fields so they will be local in loop
         dmap = self._map
         copymap = self._copymap
         unpack = struct.unpack
         e_size = struct.calcsize(_format)
         pos1 = 40
         l = len(st)
 
         # the inner loop
         while pos1 < l:
             pos2 = pos1 + e_size
             e = unpack(">cllll", st[pos1:pos2]) # a literal here is faster
             pos1 = pos2 + e[4]
             f = st[pos2:pos1]
             if '\0' in f:
                 f, c = f.split('\0')
                 copymap[f] = c
             dmap[f] = e # we hold onto e[4] because making a subtuple is slow
 
     def invalidate(self):
         for a in "_map _copymap _branch _pl _dirs _ignore".split():
             if a in self.__dict__:
                 delattr(self, a)
         self._dirty = False
 
     def copy(self, source, dest):
         self._dirty = True
         self._copymap[dest] = source
 
     def copied(self, file):
         return self._copymap.get(file, None)
 
     def copies(self):
         return self._copymap
 
     def _incpath(self, path):
         c = path.rfind('/')
         if c >= 0:
             dirs = self._dirs
             base = path[:c]
             if base not in dirs:
                 self._incpath(base)
                 dirs[base] = 1
             else:
                 dirs[base] += 1
 
     def _decpath(self, path):
         if "_dirs" in self.__dict__:
             c = path.rfind('/')
             if c >= 0:
                 base = path[:c]
                 dirs = self._dirs
                 if dirs[base] == 1:
                     del dirs[base]
                     self._decpath(base)
                 else:
                     dirs[base] -= 1
 
     def _incpathcheck(self, f):
         if '\r' in f or '\n' in f:
             raise util.Abort(_("'\\n' and '\\r' disallowed in filenames"))
         # shadows
         if f in self._dirs:
             raise util.Abort(_('directory %r already in dirstate') % f)
         for c in strutil.rfindall(f, '/'):
             d = f[:c]
             if d in self._dirs:
                 break
-            if d in self._map:
+            if d in self._map and self[d] != 'r':
                 raise util.Abort(_('file %r in dirstate clashes with %r') %
                                  (d, f))
         self._incpath(f)
 
+    def _changepath(self, f, newstate):
+        # handle upcoming path changes
+        oldstate = self[f]
+        if oldstate not in "?r" and newstate in "?r":
+            self._decpath(f)
+            return
+        if oldstate in "?r" and newstate not in "?r":
+            self._incpathcheck(f)
+            return
+
     def normal(self, f):
         'mark a file normal and clean'
         self._dirty = True
+        self._changepath(f, 'n')
         s = os.lstat(self._join(f))
         self._map[f] = ('n', s.st_mode, s.st_size, s.st_mtime, 0)
         if self._copymap.has_key(f):
             del self._copymap[f]
 
     def normallookup(self, f):
         'mark a file normal, but possibly dirty'
         self._dirty = True
+        self._changepath(f, 'n')
         self._map[f] = ('n', 0, -1, -1, 0)
         if f in self._copymap:
             del self._copymap[f]
 
     def normaldirty(self, f):
         'mark a file normal, but dirty'
         self._dirty = True
+        self._changepath(f, 'n')
         self._map[f] = ('n', 0, -2, -1, 0)
         if f in self._copymap:
             del self._copymap[f]
 
     def add(self, f):
         'mark a file added'
         self._dirty = True
-        self._incpathcheck(f)
+        self._changepath(f, 'a')
         self._map[f] = ('a', 0, -1, -1, 0)
         if f in self._copymap:
             del self._copymap[f]
 
     def remove(self, f):
         'mark a file removed'
         self._dirty = True
+        self._changepath(f, 'r')
         self._map[f] = ('r', 0, 0, 0, 0)
-        self._decpath(f)
         if f in self._copymap:
             del self._copymap[f]
 
     def merge(self, f):
         'mark a file merged'
         self._dirty = True
         s = os.lstat(self._join(f))
+        self._changepath(f, 'm')
         self._map[f] = ('m', s.st_mode, s.st_size, s.st_mtime, 0)
         if f in self._copymap:
             del self._copymap[f]
 
     def forget(self, f):
         'forget a file'
         self._dirty = True
         try:
+            self._changepath(f, '?')
             del self._map[f]
-            self._decpath(f)
         except KeyError:
             self._ui.warn(_("not in dirstate: %s!\n") % f)
 
     def clear(self):
         self._map = {}
+        if "_dirs" in self.__dict__:
+            delattr(self, "_dirs");
         self._copymap = {}
         self._pl = [nullid, nullid]
         self._dirty = True
 
     def rebuild(self, parent, files):
         self.clear()
         for f in files:
             if files.execf(f):
                 self._map[f] = ('n', 0777, -1, 0, 0)
             else:
                 self._map[f] = ('n', 0666, -1, 0, 0)
         self._pl = (parent, nullid)
         self._dirty = True
 
     def write(self):
         if not self._dirty:
             return
         cs = cStringIO.StringIO()
         copymap = self._copymap
         pack = struct.pack
         write = cs.write
         write("".join(self._pl))
         for f, e in self._map.iteritems():
             if f in copymap:
                 f = "%s\0%s" % (f, copymap[f])
             e = pack(_format, e[0], e[1], e[2], e[3], len(f))
             write(e)
             write(f)
         st = self._opener("dirstate", "w", atomictemp=True)
         st.write(cs.getvalue())
         st.rename()
         self._dirty = self._dirtypl = False
 
     def _filter(self, files):
         ret = {}
         unknown = []
 
         for x in files:
             if x == '.':
                 return self._map.copy()
             if x not in self._map:
                 unknown.append(x)
             else:
                 ret[x] = self._map[x]
 
         if not unknown:
             return ret
 
         b = self._map.keys()
         b.sort()
         blen = len(b)
 
         for x in unknown:
             bs = bisect.bisect(b, "%s%s" % (x, '/'))
             while bs < blen:
                 s = b[bs]
                 if len(s) > len(x) and s.startswith(x):
                     ret[s] = self._map[s]
                 else:
                     break
                 bs += 1
         return ret
 
     def _supported(self, f, mode, verbose=False):
         if stat.S_ISREG(mode) or stat.S_ISLNK(mode):
             return True
         if verbose:
             kind = 'unknown'
             if stat.S_ISCHR(mode): kind = _('character device')
             elif stat.S_ISBLK(mode): kind = _('block device')
             elif stat.S_ISFIFO(mode): kind = _('fifo')
             elif stat.S_ISSOCK(mode): kind = _('socket')
             elif stat.S_ISDIR(mode): kind = _('directory')
             self._ui.warn(_('%s: unsupported file type (type is %s)\n')
                           % (self.pathto(f), kind))
         return False
 
     def walk(self, files=None, match=util.always, badmatch=None):
         # filter out the stat
         for src, f, st in self.statwalk(files, match, badmatch=badmatch):
             yield src, f
 
     def statwalk(self, files=None, match=util.always, ignored=False,
                  badmatch=None, directories=False):
         '''
         walk recursively through the directory tree, finding all files
         matched by the match function
 
         results are yielded in a tuple (src, filename, st), where src
         is one of:
         'f' the file was found in the directory tree
         'd' the file is a directory of the tree
         'm' the file was only in the dirstate and not in the tree
         'b' file was not found and matched badmatch
 
         and st is the stat result if the file was found in the directory.
         '''
 
         # walk all files by default
         if not files:
             files = ['.']
             dc = self._map.copy()
         else:
             files = util.unique(files)
             dc = self._filter(files)
 
         def imatch(file_):
             if file_ not in dc and self._ignore(file_):
                 return False
             return match(file_)
 
         ignore = self._ignore
         if ignored:
             imatch = match
             ignore = util.never
 
         # self._root may end with a path separator when self._root == '/'
         common_prefix_len = len(self._root)
         if not self._root.endswith(os.sep):
             common_prefix_len += 1
 
         normpath = util.normpath
         listdir = osutil.listdir
         lstat = os.lstat
         bisect_left = bisect.bisect_left
         isdir = os.path.isdir
         pconvert = util.pconvert
         join = os.path.join
         s_isdir = stat.S_ISDIR
         supported = self._supported
         _join = self._join
         known = {'.hg': 1}
 
         # recursion free walker, faster than os.walk.
         def findfiles(s):
             work = [s]
             wadd = work.append
             found = []
             add = found.append
             if directories:
                 add((normpath(s[common_prefix_len:]), 'd', lstat(s)))
             while work:
                 top = work.pop()
                 entries = listdir(top, stat=True)
                 # nd is the top of the repository dir tree
                 nd = normpath(top[common_prefix_len:])
                 if nd == '.':
                     nd = ''
                 else:
                     # do not recurse into a repo contained in this
                     # one. use bisect to find .hg directory so speed
                     # is good on big directory.
                     names = [e[0] for e in entries]
                     hg = bisect_left(names, '.hg')
                     if hg < len(names) and names[hg] == '.hg':
                         if isdir(join(top, '.hg')):
                             continue
                 for f, kind, st in entries:
                     np = pconvert(join(nd, f))
                     if np in known:
                         continue
                     known[np] = 1
                     p = join(top, f)
                     # don't trip over symlinks
                     if kind == stat.S_IFDIR:
                         if not ignore(np):
                             wadd(p)
                             if directories:
                                 add((np, 'd', st))
                         if np in dc and match(np):
                             add((np, 'm', st))
                     elif imatch(np):
                         if supported(np, st.st_mode):
                             add((np, 'f', st))
                         elif np in dc:
                             add((np, 'm', st))
             found.sort()
             return found
 
         # step one, find all files that match our criteria
         files.sort()
         for ff in files:
             nf = normpath(ff)
             f = _join(ff)
             try:
                 st = lstat(f)
             except OSError, inst:
                 found = False
                 for fn in dc:
                     if nf == fn or (fn.startswith(nf) and fn[len(nf)] == '/'):
                         found = True
                         break
                 if not found:
                     if inst.errno != errno.ENOENT or not badmatch:
                         self._ui.warn('%s: %s\n' %
                                       (self.pathto(ff), inst.strerror))
                     elif badmatch and badmatch(ff) and imatch(nf):
                         yield 'b', ff, None
                 continue
             if s_isdir(st.st_mode):
                 for f, src, st in findfiles(f):
                     yield src, f, st
             else:
                 if nf in known:
                     continue
                 known[nf] = 1
                 if match(nf):
                     if supported(ff, st.st_mode, verbose=True):
                         yield 'f', nf, st
                     elif ff in dc:
                         yield 'm', nf, st
 
         # step two run through anything left in the dc hash and yield
         # if we haven't already seen it
         ks = dc.keys()
         ks.sort()
         for k in ks:
             if k in known:
                 continue
             known[k] = 1
             if imatch(k):
                 yield 'm', k, None
 
     def status(self, files, match, list_ignored, list_clean):
         lookup, modified, added, unknown, ignored = [], [], [], [], []
         removed, deleted, clean = [], [], []
 
         _join = self._join
         lstat = os.lstat
         cmap = self._copymap
         dmap = self._map
         ladd = lookup.append
         madd = modified.append
         aadd = added.append
         uadd = unknown.append
         iadd = ignored.append
         radd = removed.append
         dadd = deleted.append
         cadd = clean.append
 
         for src, fn, st in self.statwalk(files, match, ignored=list_ignored):
             if fn in dmap:
                 type_, mode, size, time, foo = dmap[fn]
             else:
                 if list_ignored and self._ignore(fn):
                     iadd(fn)
                 else:
                     uadd(fn)
                 continue
             if src == 'm':
                 nonexistent = True
                 if not st:
                     try:
                         st = lstat(_join(fn))
                     except OSError, inst:
-                        if inst.errno != errno.ENOENT:
+                        if inst.errno not in (errno.ENOENT, errno.ENOTDIR):
                             raise
                         st = None
                     # We need to re-check that it is a valid file
                     if st and self._supported(fn, st.st_mode):
                         nonexistent = False
                 # XXX: what to do with file no longer present in the fs
                 # who are not removed in the dirstate ?
                 if nonexistent and type_ in "nm":
                     dadd(fn)
                     continue
             # check the common case first
             if type_ == 'n':
                 if not st:
                     st = lstat(_join(fn))
                 if (size >= 0 and (size != st.st_size
                                    or (mode ^ st.st_mode) & 0100)
                     or size == -2
                     or fn in self._copymap):
                     madd(fn)
                 elif time != int(st.st_mtime):
                     ladd(fn)
                 elif list_clean:
                     cadd(fn)
             elif type_ == 'm':
                 madd(fn)
             elif type_ == 'a':
                 aadd(fn)
             elif type_ == 'r':
                 radd(fn)
 
         return (lookup, modified, added, removed, deleted, unknown, ignored,
                 clean)

mercurial/util.py

@@ -1,1700 +1,1700 b''
 """
 util.py - Mercurial utility functions and platform specfic implementations
 
  Copyright 2005 K. Thananchayan <[e-mail]>
  Copyright 2005-2007 Matt Mackall <[e-mail]>
  Copyright 2006 Vadim Gelfer <[e-mail]>
 
 This software may be used and distributed according to the terms
 of the GNU General Public License, incorporated herein by reference.
 
 This contains helper routines that are independent of the SCM core and hide
 platform-specific details from the core.
 """
 
 from i18n import _
 import cStringIO, errno, getpass, popen2, re, shutil, sys, tempfile, strutil
 import os, stat, threading, time, calendar, ConfigParser, locale, glob, osutil
 
 try:
     set = set
     frozenset = frozenset
 except NameError:
     from sets import Set as set, ImmutableSet as frozenset
 
 try:
     _encoding = os.environ.get("HGENCODING")
     if sys.platform == 'darwin' and not _encoding:
         # On darwin, getpreferredencoding ignores the locale environment and
         # always returns mac-roman. We override this if the environment is
         # not C (has been customized by the user).
         locale.setlocale(locale.LC_CTYPE, '')
         _encoding = locale.getlocale()[1]
     if not _encoding:
         _encoding = locale.getpreferredencoding() or 'ascii'
 except locale.Error:
     _encoding = 'ascii'
 _encodingmode = os.environ.get("HGENCODINGMODE", "strict")
 _fallbackencoding = 'ISO-8859-1'
 
 def tolocal(s):
     """
     Convert a string from internal UTF-8 to local encoding
 
     All internal strings should be UTF-8 but some repos before the
     implementation of locale support may contain latin1 or possibly
     other character sets. We attempt to decode everything strictly
     using UTF-8, then Latin-1, and failing that, we use UTF-8 and
     replace unknown characters.
     """
     for e in ('UTF-8', _fallbackencoding):
         try:
             u = s.decode(e) # attempt strict decoding
             return u.encode(_encoding, "replace")
         except LookupError, k:
             raise Abort(_("%s, please check your locale settings") % k)
         except UnicodeDecodeError:
             pass
     u = s.decode("utf-8", "replace") # last ditch
     return u.encode(_encoding, "replace")
 
 def fromlocal(s):
     """
     Convert a string from the local character encoding to UTF-8
 
     We attempt to decode strings using the encoding mode set by
     HGENCODINGMODE, which defaults to 'strict'. In this mode, unknown
     characters will cause an error message. Other modes include
     'replace', which replaces unknown characters with a special
     Unicode character, and 'ignore', which drops the character.
     """
     try:
         return s.decode(_encoding, _encodingmode).encode("utf-8")
     except UnicodeDecodeError, inst:
         sub = s[max(0, inst.start-10):inst.start+10]
         raise Abort("decoding near '%s': %s!" % (sub, inst))
     except LookupError, k:
         raise Abort(_("%s, please check your locale settings") % k)
 
 def locallen(s):
     """Find the length in characters of a local string"""
     return len(s.decode(_encoding, "replace"))
 
 def localsub(s, a, b=None):
     try:
         u = s.decode(_encoding, _encodingmode)
         if b is not None:
             u = u[a:b]
         else:
             u = u[:a]
         return u.encode(_encoding, _encodingmode)
     except UnicodeDecodeError, inst:
         sub = s[max(0, inst.start-10), inst.start+10]
         raise Abort(_("decoding near '%s': %s!") % (sub, inst))
 
 # used by parsedate
 defaultdateformats = (
     '%Y-%m-%d %H:%M:%S',
     '%Y-%m-%d %I:%M:%S%p',
     '%Y-%m-%d %H:%M',
     '%Y-%m-%d %I:%M%p',
     '%Y-%m-%d',
     '%m-%d',
     '%m/%d',
     '%m/%d/%y',
     '%m/%d/%Y',
     '%a %b %d %H:%M:%S %Y',
     '%a %b %d %I:%M:%S%p %Y',
     '%a, %d %b %Y %H:%M:%S',        #  GNU coreutils "/bin/date --rfc-2822"
     '%b %d %H:%M:%S %Y',
     '%b %d %I:%M:%S%p %Y',
     '%b %d %H:%M:%S',
     '%b %d %I:%M:%S%p',
     '%b %d %H:%M',
     '%b %d %I:%M%p',
     '%b %d %Y',
     '%b %d',
     '%H:%M:%S',
     '%I:%M:%SP',
     '%H:%M',
     '%I:%M%p',
 )
 
 extendeddateformats = defaultdateformats + (
     "%Y",
     "%Y-%m",
     "%b",
     "%b %Y",
     )
 
 class SignalInterrupt(Exception):
     """Exception raised on SIGTERM and SIGHUP."""
 
 # differences from SafeConfigParser:
 # - case-sensitive keys
 # - allows values that are not strings (this means that you may not
 #   be able to save the configuration to a file)
 class configparser(ConfigParser.SafeConfigParser):
     def optionxform(self, optionstr):
         return optionstr
 
     def set(self, section, option, value):
         return ConfigParser.ConfigParser.set(self, section, option, value)
 
     def _interpolate(self, section, option, rawval, vars):
         if not isinstance(rawval, basestring):
             return rawval
         return ConfigParser.SafeConfigParser._interpolate(self, section,
                                                           option, rawval, vars)
 
 def cachefunc(func):
     '''cache the result of function calls'''
     # XXX doesn't handle keywords args
     cache = {}
     if func.func_code.co_argcount == 1:
         # we gain a small amount of time because
         # we don't need to pack/unpack the list
         def f(arg):
             if arg not in cache:
                 cache[arg] = func(arg)
             return cache[arg]
     else:
         def f(*args):
             if args not in cache:
                 cache[args] = func(*args)
             return cache[args]
 
     return f
 
 def pipefilter(s, cmd):
     '''filter string S through command CMD, returning its output'''
     (pin, pout) = os.popen2(cmd, 'b')
     def writer():
         try:
             pin.write(s)
             pin.close()
         except IOError, inst:
             if inst.errno != errno.EPIPE:
                 raise
 
     # we should use select instead on UNIX, but this will work on most
     # systems, including Windows
     w = threading.Thread(target=writer)
     w.start()
     f = pout.read()
     pout.close()
     w.join()
     return f
 
 def tempfilter(s, cmd):
     '''filter string S through a pair of temporary files with CMD.
     CMD is used as a template to create the real command to be run,
     with the strings INFILE and OUTFILE replaced by the real names of
     the temporary files generated.'''
     inname, outname = None, None
     try:
         infd, inname = tempfile.mkstemp(prefix='hg-filter-in-')
         fp = os.fdopen(infd, 'wb')
         fp.write(s)
         fp.close()
         outfd, outname = tempfile.mkstemp(prefix='hg-filter-out-')
         os.close(outfd)
         cmd = cmd.replace('INFILE', inname)
         cmd = cmd.replace('OUTFILE', outname)
         code = os.system(cmd)
         if sys.platform == 'OpenVMS' and code & 1:
             code = 0
         if code: raise Abort(_("command '%s' failed: %s") %
                              (cmd, explain_exit(code)))
         return open(outname, 'rb').read()
     finally:
         try:
             if inname: os.unlink(inname)
         except: pass
         try:
             if outname: os.unlink(outname)
         except: pass
 
 filtertable = {
     'tempfile:': tempfilter,
     'pipe:': pipefilter,
     }
 
 def filter(s, cmd):
     "filter a string through a command that transforms its input to its output"
     for name, fn in filtertable.iteritems():
         if cmd.startswith(name):
             return fn(s, cmd[len(name):].lstrip())
     return pipefilter(s, cmd)
 
 def binary(s):
     """return true if a string is binary data using diff's heuristic"""
     if s and '\0' in s[:4096]:
         return True
     return False
 
 def unique(g):
     """return the uniq elements of iterable g"""
     seen = {}
     l = []
     for f in g:
         if f not in seen:
             seen[f] = 1
             l.append(f)
     return l
 
 class Abort(Exception):
     """Raised if a command needs to print an error and exit."""
 
 class UnexpectedOutput(Abort):
     """Raised to print an error with part of output and exit."""
 
 def always(fn): return True
 def never(fn): return False
 
 def expand_glob(pats):
     '''On Windows, expand the implicit globs in a list of patterns'''
     if os.name != 'nt':
         return list(pats)
     ret = []
     for p in pats:
         kind, name = patkind(p, None)
         if kind is None:
             globbed = glob.glob(name)
             if globbed:
                 ret.extend(globbed)
                 continue
             # if we couldn't expand the glob, just keep it around
         ret.append(p)
     return ret
 
 def patkind(name, dflt_pat='glob'):
     """Split a string into an optional pattern kind prefix and the
     actual pattern."""
     for prefix in 're', 'glob', 'path', 'relglob', 'relpath', 'relre':
         if name.startswith(prefix + ':'): return name.split(':', 1)
     return dflt_pat, name
 
 def globre(pat, head='^', tail='$'):
     "convert a glob pattern into a regexp"
     i, n = 0, len(pat)
     res = ''
     group = False
     def peek(): return i < n and pat[i]
     while i < n:
         c = pat[i]
         i = i+1
         if c == '*':
             if peek() == '*':
                 i += 1
                 res += '.*'
             else:
                 res += '[^/]*'
         elif c == '?':
             res += '.'
         elif c == '[':
             j = i
             if j < n and pat[j] in '!]':
                 j += 1
             while j < n and pat[j] != ']':
                 j += 1
             if j >= n:
                 res += '\\['
             else:
                 stuff = pat[i:j].replace('\\','\\\\')
                 i = j + 1
                 if stuff[0] == '!':
                     stuff = '^' + stuff[1:]
                 elif stuff[0] == '^':
                     stuff = '\\' + stuff
                 res = '%s[%s]' % (res, stuff)
         elif c == '{':
             group = True
             res += '(?:'
         elif c == '}' and group:
             res += ')'
             group = False
         elif c == ',' and group:
             res += '|'
         elif c == '\\':
             p = peek()
             if p:
                 i += 1
                 res += re.escape(p)
             else:
                 res += re.escape(c)
         else:
             res += re.escape(c)
     return head + res + tail
 
 _globchars = {'[': 1, '{': 1, '*': 1, '?': 1}
 
 def pathto(root, n1, n2):
     '''return the relative path from one place to another.
     root should use os.sep to separate directories
     n1 should use os.sep to separate directories
     n2 should use "/" to separate directories
     returns an os.sep-separated path.
 
     If n1 is a relative path, it's assumed it's
     relative to root.
     n2 should always be relative to root.
     '''
     if not n1: return localpath(n2)
     if os.path.isabs(n1):
         if os.path.splitdrive(root)[0] != os.path.splitdrive(n1)[0]:
             return os.path.join(root, localpath(n2))
         n2 = '/'.join((pconvert(root), n2))
     a, b = n1.split(os.sep), n2.split('/')
     a.reverse()
     b.reverse()
     while a and b and a[-1] == b[-1]:
         a.pop()
         b.pop()
     b.reverse()
     return os.sep.join((['..'] * len(a)) + b)
 
 def canonpath(root, cwd, myname):
     """return the canonical path of myname, given cwd and root"""
     if root == os.sep:
         rootsep = os.sep
     elif root.endswith(os.sep):
         rootsep = root
     else:
         rootsep = root + os.sep
     name = myname
     if not os.path.isabs(name):
         name = os.path.join(root, cwd, name)
     name = os.path.normpath(name)
     audit_path = path_auditor(root)
     if name != rootsep and name.startswith(rootsep):
         name = name[len(rootsep):]
         audit_path(name)
         return pconvert(name)
     elif name == root:
         return ''
     else:
         # Determine whether `name' is in the hierarchy at or beneath `root',
         # by iterating name=dirname(name) until that causes no change (can't
         # check name == '/', because that doesn't work on windows).  For each
         # `name', compare dev/inode numbers.  If they match, the list `rel'
         # holds the reversed list of components making up the relative file
         # name we want.
         root_st = os.stat(root)
         rel = []
         while True:
             try:
                 name_st = os.stat(name)
             except OSError:
                 break
             if samestat(name_st, root_st):
                 if not rel:
                     # name was actually the same as root (maybe a symlink)
                     return ''
                 rel.reverse()
                 name = os.path.join(*rel)
                 audit_path(name)
                 return pconvert(name)
             dirname, basename = os.path.split(name)
             rel.append(basename)
             if dirname == name:
                 break
             name = dirname
 
         raise Abort('%s not under root' % myname)
 
 def matcher(canonroot, cwd='', names=[], inc=[], exc=[], src=None):
     return _matcher(canonroot, cwd, names, inc, exc, 'glob', src)
 
 def cmdmatcher(canonroot, cwd='', names=[], inc=[], exc=[], src=None,
                globbed=False, default=None):
     default = default or 'relpath'
     if default == 'relpath' and not globbed:
         names = expand_glob(names)
     return _matcher(canonroot, cwd, names, inc, exc, default, src)
 
 def _matcher(canonroot, cwd, names, inc, exc, dflt_pat, src):
     """build a function to match a set of file patterns
 
     arguments:
     canonroot - the canonical root of the tree you're matching against
     cwd - the current working directory, if relevant
     names - patterns to find
     inc - patterns to include
     exc - patterns to exclude
     dflt_pat - if a pattern in names has no explicit type, assume this one
     src - where these patterns came from (e.g. .hgignore)
 
     a pattern is one of:
     'glob:<glob>' - a glob relative to cwd
     're:<regexp>' - a regular expression
     'path:<path>' - a path relative to canonroot
     'relglob:<glob>' - an unrooted glob (*.c matches C files in all dirs)
     'relpath:<path>' - a path relative to cwd
     'relre:<regexp>' - a regexp that doesn't have to match the start of a name
     '<something>' - one of the cases above, selected by the dflt_pat argument
 
     returns:
     a 3-tuple containing
     - list of roots (places where one should start a recursive walk of the fs);
       this often matches the explicit non-pattern names passed in, but also
       includes the initial part of glob: patterns that has no glob characters
     - a bool match(filename) function
     - a bool indicating if any patterns were passed in
     """
 
     # a common case: no patterns at all
     if not names and not inc and not exc:
         return [], always, False
 
     def contains_glob(name):
         for c in name:
             if c in _globchars: return True
         return False
 
     def regex(kind, name, tail):
         '''convert a pattern into a regular expression'''
         if not name:
             return ''
         if kind == 're':
             return name
         elif kind == 'path':
             return '^' + re.escape(name) + '(?:/|$)'
         elif kind == 'relglob':
             return globre(name, '(?:|.*/)', tail)
         elif kind == 'relpath':
             return re.escape(name) + '(?:/|$)'
         elif kind == 'relre':
             if name.startswith('^'):
                 return name
             return '.*' + name
         return globre(name, '', tail)
 
     def matchfn(pats, tail):
         """build a matching function from a set of patterns"""
         if not pats:
             return
         try:
             pat = '(?:%s)' % '|'.join([regex(k, p, tail) for (k, p) in pats])
             return re.compile(pat).match
         except OverflowError:
             # We're using a Python with a tiny regex engine and we
             # made it explode, so we'll divide the pattern list in two
             # until it works
             l = len(pats)
             if l < 2:
                 raise
             a, b = matchfn(pats[:l//2], tail), matchfn(pats[l//2:], tail)
             return lambda s: a(s) or b(s)
         except re.error:
             for k, p in pats:
                 try:
                     re.compile('(?:%s)' % regex(k, p, tail))
                 except re.error:
                     if src:
                         raise Abort("%s: invalid pattern (%s): %s" %
                                     (src, k, p))
                     else:
                         raise Abort("invalid pattern (%s): %s" % (k, p))
             raise Abort("invalid pattern")
 
     def globprefix(pat):
         '''return the non-glob prefix of a path, e.g. foo/* -> foo'''
         root = []
         for p in pat.split('/'):
             if contains_glob(p): break
             root.append(p)
         return '/'.join(root) or '.'
 
     def normalizepats(names, default):
         pats = []
         roots = []
         anypats = False
         for kind, name in [patkind(p, default) for p in names]:
             if kind in ('glob', 'relpath'):
                 name = canonpath(canonroot, cwd, name)
             elif kind in ('relglob', 'path'):
                 name = normpath(name)
 
             pats.append((kind, name))
 
             if kind in ('glob', 're', 'relglob', 'relre'):
                 anypats = True
 
             if kind == 'glob':
                 root = globprefix(name)
                 roots.append(root)
             elif kind in ('relpath', 'path'):
                 roots.append(name or '.')
             elif kind == 'relglob':
                 roots.append('.')
         return roots, pats, anypats
 
     roots, pats, anypats = normalizepats(names, dflt_pat)
 
     patmatch = matchfn(pats, '$') or always
     incmatch = always
     if inc:
         dummy, inckinds, dummy = normalizepats(inc, 'glob')
         incmatch = matchfn(inckinds, '(?:/|$)')
     excmatch = lambda fn: False
     if exc:
         dummy, exckinds, dummy = normalizepats(exc, 'glob')
         excmatch = matchfn(exckinds, '(?:/|$)')
 
     if not names and inc and not exc:
         # common case: hgignore patterns
         match = incmatch
     else:
         match = lambda fn: incmatch(fn) and not excmatch(fn) and patmatch(fn)
 
     return (roots, match, (inc or exc or anypats) and True)
 
 _hgexecutable = None
 
 def hgexecutable():
     """return location of the 'hg' executable.
 
     Defaults to $HG or 'hg' in the search path.
     """
     if _hgexecutable is None:
         set_hgexecutable(os.environ.get('HG') or find_exe('hg', 'hg'))
     return _hgexecutable
 
 def set_hgexecutable(path):
     """set location of the 'hg' executable"""
     global _hgexecutable
     _hgexecutable = path
 
 def system(cmd, environ={}, cwd=None, onerr=None, errprefix=None):
     '''enhanced shell command execution.
     run with environment maybe modified, maybe in different dir.
 
     if command fails and onerr is None, return status.  if ui object,
     print error message and return status, else raise onerr object as
     exception.'''
     def py2shell(val):
         'convert python object into string that is useful to shell'
         if val in (None, False):
             return '0'
         if val == True:
             return '1'
         return str(val)
     oldenv = {}
     for k in environ:
         oldenv[k] = os.environ.get(k)
     if cwd is not None:
         oldcwd = os.getcwd()
     origcmd = cmd
     if os.name == 'nt':
         cmd = '"%s"' % cmd
     try:
         for k, v in environ.iteritems():
             os.environ[k] = py2shell(v)
         os.environ['HG'] = hgexecutable()
         if cwd is not None and oldcwd != cwd:
             os.chdir(cwd)
         rc = os.system(cmd)
         if sys.platform == 'OpenVMS' and rc & 1:
             rc = 0
         if rc and onerr:
             errmsg = '%s %s' % (os.path.basename(origcmd.split(None, 1)[0]),
                                 explain_exit(rc)[0])
             if errprefix:
                 errmsg = '%s: %s' % (errprefix, errmsg)
             try:
                 onerr.warn(errmsg + '\n')
             except AttributeError:
                 raise onerr(errmsg)
         return rc
     finally:
         for k, v in oldenv.iteritems():
             if v is None:
                 del os.environ[k]
             else:
                 os.environ[k] = v
         if cwd is not None and oldcwd != cwd:
             os.chdir(oldcwd)
 
 # os.path.lexists is not available on python2.3
 def lexists(filename):
     "test whether a file with this name exists. does not follow symlinks"
     try:
         os.lstat(filename)
     except:
         return False
     return True
 
 def rename(src, dst):
     """forcibly rename a file"""
     try:
         os.rename(src, dst)
     except OSError, err: # FIXME: check err (EEXIST ?)
         # on windows, rename to existing file is not allowed, so we
         # must delete destination first. but if file is open, unlink
         # schedules it for delete but does not delete it. rename
         # happens immediately even for open files, so we create
         # temporary file, delete it, rename destination to that name,
         # then delete that. then rename is safe to do.
         fd, temp = tempfile.mkstemp(dir=os.path.dirname(dst) or '.')
         os.close(fd)
         os.unlink(temp)
         os.rename(dst, temp)
         os.unlink(temp)
         os.rename(src, dst)
 
 def unlink(f):
     """unlink and remove the directory if it is empty"""
     os.unlink(f)
     # try removing directories that might now be empty
     try:
         os.removedirs(os.path.dirname(f))
     except OSError:
         pass
 
 def copyfile(src, dest):
     "copy a file, preserving mode"
     if os.path.islink(src):
         try:
             os.unlink(dest)
         except:
             pass
         os.symlink(os.readlink(src), dest)
     else:
         try:
             shutil.copyfile(src, dest)
             shutil.copymode(src, dest)
         except shutil.Error, inst:
             raise Abort(str(inst))
 
 def copyfiles(src, dst, hardlink=None):
     """Copy a directory tree using hardlinks if possible"""
 
     if hardlink is None:
         hardlink = (os.stat(src).st_dev ==
                     os.stat(os.path.dirname(dst)).st_dev)
 
     if os.path.isdir(src):
         os.mkdir(dst)
         for name, kind in osutil.listdir(src):
             srcname = os.path.join(src, name)
             dstname = os.path.join(dst, name)
             copyfiles(srcname, dstname, hardlink)
     else:
         if hardlink:
             try:
                 os_link(src, dst)
             except (IOError, OSError):
                 hardlink = False
                 shutil.copy(src, dst)
         else:
             shutil.copy(src, dst)
 
 class path_auditor(object):
     '''ensure that a filesystem path contains no banned components.
     the following properties of a path are checked:
 
     - under top-level .hg
     - starts at the root of a windows drive
     - contains ".."
     - traverses a symlink (e.g. a/symlink_here/b)
     - inside a nested repository'''
 
     def __init__(self, root):
         self.audited = set()
         self.auditeddir = set()
         self.root = root
 
     def __call__(self, path):
         if path in self.audited:
             return
         normpath = os.path.normcase(path)
         parts = normpath.split(os.sep)
         if (os.path.splitdrive(path)[0] or parts[0] in ('.hg', '')
             or os.pardir in parts):
             raise Abort(_("path contains illegal component: %s") % path)
         def check(prefix):
             curpath = os.path.join(self.root, prefix)
             try:
                 st = os.lstat(curpath)
             except OSError, err:
                 # EINVAL can be raised as invalid path syntax under win32.
                 # They must be ignored for patterns can be checked too.
-                if err.errno not in (errno.ENOENT, errno.EINVAL):
+                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))
 
         prefixes = []
         for c in strutil.rfindall(normpath, os.sep):
             prefix = normpath[:c]
             if prefix in self.auditeddir:
                 break
             check(prefix)
             prefixes.append(prefix)
 
         self.audited.add(path)
         # only add prefixes to the cache after checking everything: we don't
         # want to add "foo/bar/baz" before checking if there's a "foo/.hg"
         self.auditeddir.update(prefixes)
 
 def _makelock_file(info, pathname):
     ld = os.open(pathname, os.O_CREAT | os.O_WRONLY | os.O_EXCL)
     os.write(ld, info)
     os.close(ld)
 
 def _readlock_file(pathname):
     return posixfile(pathname).read()
 
 def nlinks(pathname):
     """Return number of hardlinks for the given file."""
     return os.lstat(pathname).st_nlink
 
 if hasattr(os, 'link'):
     os_link = os.link
 else:
     def os_link(src, dst):
         raise OSError(0, _("Hardlinks not supported"))
 
 def fstat(fp):
     '''stat file object that may not have fileno method.'''
     try:
         return os.fstat(fp.fileno())
     except AttributeError:
         return os.stat(fp.name)
 
 posixfile = file
 
 def is_win_9x():
     '''return true if run on windows 95, 98 or me.'''
     try:
         return sys.getwindowsversion()[3] == 1
     except AttributeError:
         return os.name == 'nt' and 'command' in os.environ.get('comspec', '')
 
 getuser_fallback = None
 
 def getuser():
     '''return name of current user'''
     try:
         return getpass.getuser()
     except ImportError:
         # import of pwd will fail on windows - try fallback
         if getuser_fallback:
             return getuser_fallback()
     # raised if win32api not available
     raise Abort(_('user name not available - set USERNAME '
                   'environment variable'))
 
 def username(uid=None):
     """Return the name of the user with the given uid.
 
     If uid is None, return the name of the current user."""
     try:
         import pwd
         if uid is None:
             uid = os.getuid()
         try:
             return pwd.getpwuid(uid)[0]
         except KeyError:
             return str(uid)
     except ImportError:
         return None
 
 def groupname(gid=None):
     """Return the name of the group with the given gid.
 
     If gid is None, return the name of the current group."""
     try:
         import grp
         if gid is None:
             gid = os.getgid()
         try:
             return grp.getgrgid(gid)[0]
         except KeyError:
             return str(gid)
     except ImportError:
         return None
 
 # File system features
 
 def checkfolding(path):
     """
     Check whether the given path is on a case-sensitive filesystem
 
     Requires a path (like /foo/.hg) ending with a foldable final
     directory component.
     """
     s1 = os.stat(path)
     d, b = os.path.split(path)
     p2 = os.path.join(d, b.upper())
     if path == p2:
         p2 = os.path.join(d, b.lower())
     try:
         s2 = os.stat(p2)
         if s2 == s1:
             return False
         return True
     except:
         return True
 
 def checkexec(path):
     """
     Check whether the given path is on a filesystem with UNIX-like exec flags
 
     Requires a directory (like /foo/.hg)
     """
     try:
         EXECFLAGS = stat.S_IXUSR | stat.S_IXGRP | stat.S_IXOTH
         fh, fn = tempfile.mkstemp("", "", path)
         os.close(fh)
         m = os.stat(fn).st_mode
         # VFAT on Linux can flip mode but it doesn't persist a FS remount.
         # frequently we can detect it if files are created with exec bit on.
         new_file_has_exec = m & EXECFLAGS
         os.chmod(fn, m ^ EXECFLAGS)
         exec_flags_cannot_flip = (os.stat(fn).st_mode == m)
         os.unlink(fn)
     except (IOError,OSError):
         # we don't care, the user probably won't be able to commit anyway
         return False
     return not (new_file_has_exec or exec_flags_cannot_flip)
 
 def execfunc(path, fallback):
     '''return an is_exec() function with default to fallback'''
     if checkexec(path):
         return lambda x: is_exec(os.path.join(path, x))
     return fallback
 
 def checklink(path):
     """check whether the given path is on a symlink-capable filesystem"""
     # mktemp is not racy because symlink creation will fail if the
     # file already exists
     name = tempfile.mktemp(dir=path)
     try:
         os.symlink(".", name)
         os.unlink(name)
         return True
     except (OSError, AttributeError):
         return False
 
 def linkfunc(path, fallback):
     '''return an is_link() function with default to fallback'''
     if checklink(path):
         return lambda x: os.path.islink(os.path.join(path, x))
     return fallback
 
 _umask = os.umask(0)
 os.umask(_umask)
 
 def needbinarypatch():
     """return True if patches should be applied in binary mode by default."""
     return os.name == 'nt'
 
 # Platform specific variants
 if os.name == 'nt':
     import msvcrt
     nulldev = 'NUL:'
 
     class winstdout:
         '''stdout on windows misbehaves if sent through a pipe'''
 
         def __init__(self, fp):
             self.fp = fp
 
         def __getattr__(self, key):
             return getattr(self.fp, key)
 
         def close(self):
             try:
                 self.fp.close()
             except: pass
 
         def write(self, s):
             try:
                 return self.fp.write(s)
             except IOError, inst:
                 if inst.errno != 0: raise
                 self.close()
                 raise IOError(errno.EPIPE, 'Broken pipe')
 
         def flush(self):
             try:
                 return self.fp.flush()
             except IOError, inst:
                 if inst.errno != errno.EINVAL: raise
                 self.close()
                 raise IOError(errno.EPIPE, 'Broken pipe')
 
     sys.stdout = winstdout(sys.stdout)
 
     def system_rcpath():
         try:
             return system_rcpath_win32()
         except:
             return [r'c:\mercurial\mercurial.ini']
 
     def user_rcpath():
         '''return os-specific hgrc search path to the user dir'''
         try:
             userrc = user_rcpath_win32()
         except:
             userrc = os.path.join(os.path.expanduser('~'), 'mercurial.ini')
         path = [userrc]
         userprofile = os.environ.get('USERPROFILE')
         if userprofile:
             path.append(os.path.join(userprofile, 'mercurial.ini'))
         return path
 
     def parse_patch_output(output_line):
         """parses the output produced by patch and returns the file name"""
         pf = output_line[14:]
         if pf[0] == '`':
             pf = pf[1:-1] # Remove the quotes
         return pf
 
     def testpid(pid):
         '''return False if pid dead, True if running or not known'''
         return True
 
     def set_exec(f, mode):
         pass
 
     def set_link(f, mode):
         pass
 
     def set_binary(fd):
         msvcrt.setmode(fd.fileno(), os.O_BINARY)
 
     def pconvert(path):
         return path.replace("\\", "/")
 
     def localpath(path):
         return path.replace('/', '\\')
 
     def normpath(path):
         return pconvert(os.path.normpath(path))
 
     makelock = _makelock_file
     readlock = _readlock_file
 
     def samestat(s1, s2):
         return False
 
     # A sequence of backslashes is special iff it precedes a double quote:
     # - if there's an even number of backslashes, the double quote is not
     #   quoted (i.e. it ends the quoted region)
     # - if there's an odd number of backslashes, the double quote is quoted
     # - in both cases, every pair of backslashes is unquoted into a single
     #   backslash
     # (See http://msdn2.microsoft.com/en-us/library/a1y7w461.aspx )
     # So, to quote a string, we must surround it in double quotes, double
     # the number of backslashes that preceed double quotes and add another
     # backslash before every double quote (being careful with the double
     # quote we've appended to the end)
     _quotere = None
     def shellquote(s):
         global _quotere
         if _quotere is None:
             _quotere = re.compile(r'(\\*)("|\\$)')
         return '"%s"' % _quotere.sub(r'\1\1\\\2', s)
 
     def quotecommand(cmd):
         """Build a command string suitable for os.popen* calls."""
         # The extra quotes are needed because popen* runs the command
         # through the current COMSPEC. cmd.exe suppress enclosing quotes.
         return '"' + cmd + '"'
 
     def 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 is_exec(f):
         """check whether a file is executable"""
         return (os.lstat(f).st_mode & 0100 != 0)
 
     def set_exec(f, mode):
         s = os.lstat(f).st_mode
         if stat.S_ISLNK(s) or (s & 0100 != 0) == mode:
             return
         if mode:
             # Turn on +x for every +r bit when making a file executable
             # and obey umask.
             os.chmod(f, s | (s & 0444) >> 2 & ~_umask)
         else:
             os.chmod(f, s & 0666)
 
     def set_link(f, mode):
         """make a file a symbolic link/regular file
 
         if a file is changed to a link, its contents become the link data
         if a link is changed to a file, its link data become its contents
         """
 
         m = os.path.islink(f)
         if m == bool(mode):
             return
 
         if mode: # switch file to link
             data = file(f).read()
             os.unlink(f)
             os.symlink(data, f)
         else:
             data = os.readlink(f)
             os.unlink(f)
             file(f, "w").write(data)
 
     def set_binary(fd):
         pass
 
     def pconvert(path):
         return path
 
     def localpath(path):
         return path
 
     normpath = os.path.normpath
     samestat = os.path.samestat
 
     def makelock(info, pathname):
         try:
             os.symlink(info, pathname)
         except OSError, why:
             if why.errno == errno.EEXIST:
                 raise
             else:
                 _makelock_file(info, pathname)
 
     def readlock(pathname):
         try:
             return os.readlink(pathname)
         except OSError, why:
             if why.errno in (errno.EINVAL, errno.ENOSYS):
                 return _readlock_file(pathname)
             else:
                 raise
 
     def shellquote(s):
         if os.sys.platform == 'OpenVMS':
             return '"%s"' % s
         else:
             return "'%s'" % s.replace("'", "'\\''")
 
     def quotecommand(cmd):
         return cmd
 
     def 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
     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('\\\\', '\\')