Merge with crew-stable

hgext/convert/cvs.py

@@ -1,356 +1,353 b''
 # CVS conversion code inspired by hg-cvs-import and git-cvsimport
 
 import os, locale, re, socket, errno
 from cStringIO import StringIO
 from mercurial import util
 from mercurial.i18n import _
 
 from common import NoRepo, commit, converter_source, checktool
 import cvsps
 
 class convert_cvs(converter_source):
     def __init__(self, ui, path, rev=None):
         super(convert_cvs, self).__init__(ui, path, rev=rev)
 
         cvs = os.path.join(path, "CVS")
         if not os.path.exists(cvs):
             raise NoRepo("%s does not look like a CVS checkout" % path)
 
         checktool('cvs')
         self.cmd = ui.config('convert', 'cvsps', 'builtin')
         cvspsexe = self.cmd.split(None, 1)[0]
         self.builtin = cvspsexe == 'builtin'
 
         if not self.builtin:
             checktool(cvspsexe)
 
         self.changeset = {}
         self.files = {}
         self.tags = {}
         self.lastbranch = {}
         self.parent = {}
         self.socket = None
         self.cvsroot = file(os.path.join(cvs, "Root")).read()[:-1]
         self.cvsrepo = file(os.path.join(cvs, "Repository")).read()[:-1]
         self.encoding = locale.getpreferredencoding()
 
         self._parse(ui)
         self._connect()
 
     def _parse(self, ui):
         if self.changeset:
             return
 
         maxrev = 0
         cmd = self.cmd
         if self.rev:
             # TODO: handle tags
             try:
                 # patchset number?
                 maxrev = int(self.rev)
             except ValueError:
                 try:
                     # date
                     util.parsedate(self.rev, ['%Y/%m/%d %H:%M:%S'])
                     cmd = '%s -d "1970/01/01 00:00:01" -d "%s"' % (cmd, self.rev)
                 except util.Abort:
                     raise util.Abort(_('revision %s is not a patchset number or date') % self.rev)
 
         d = os.getcwd()
         try:
             os.chdir(self.path)
             id = None
             state = 0
             filerevids = {}
 
             if self.builtin:
                 # builtin cvsps code
                 ui.status(_('using builtin cvsps\n'))
 
                 db = cvsps.createlog(ui, cache='update')
                 db = cvsps.createchangeset(ui, db,
                       fuzz=int(ui.config('convert', 'cvsps.fuzz', 60)),
                       mergeto=ui.config('convert', 'cvsps.mergeto', None),
                       mergefrom=ui.config('convert', 'cvsps.mergefrom', None))
 
                 for cs in db:
                     if maxrev and cs.id>maxrev:
                         break
                     id = str(cs.id)
                     cs.author = self.recode(cs.author)
                     self.lastbranch[cs.branch] = id
                     cs.comment = self.recode(cs.comment)
                     date = util.datestr(cs.date)
                     self.tags.update(dict.fromkeys(cs.tags, id))
 
                     files = {}
                     for f in cs.entries:
                         files[f.file] = "%s%s" % ('.'.join([str(x) for x in f.revision]),
                                                   ['', '(DEAD)'][f.dead])
 
                     # add current commit to set
                     c = commit(author=cs.author, date=date,
                              parents=[str(p.id) for p in cs.parents],
                              desc=cs.comment, branch=cs.branch or '')
                     self.changeset[id] = c
                     self.files[id] = files
             else:
                 # external cvsps
                 for l in util.popen(cmd):
                     if state == 0: # header
                         if l.startswith("PatchSet"):
                             id = l[9:-2]
                             if maxrev and int(id) > maxrev:
                                 # ignore everything
                                 state = 3
                         elif l.startswith("Date:"):
                             date = util.parsedate(l[6:-1], ["%Y/%m/%d %H:%M:%S"])
                             date = util.datestr(date)
                         elif l.startswith("Branch:"):
                             branch = l[8:-1]
                             self.parent[id] = self.lastbranch.get(branch, 'bad')
                             self.lastbranch[branch] = id
                         elif l.startswith("Ancestor branch:"):
                             ancestor = l[17:-1]
                             # figure out the parent later
                             self.parent[id] = self.lastbranch[ancestor]
                         elif l.startswith("Author:"):
                             author = self.recode(l[8:-1])
                         elif l.startswith("Tag:") or l.startswith("Tags:"):
                             t = l[l.index(':')+1:]
                             t = [ut.strip() for ut in t.split(',')]
                             if (len(t) > 1) or (t[0] and (t[0] != "(none)")):
                                 self.tags.update(dict.fromkeys(t, id))
                         elif l.startswith("Log:"):
                             # switch to gathering log
                             state = 1
                             log = ""
                     elif state == 1: # log
                         if l == "Members: \n":
                             # switch to gathering members
                             files = {}
                             oldrevs = []
                             log = self.recode(log[:-1])
                             state = 2
                         else:
                             # gather log
                             log += l
                     elif state == 2: # members
                         if l == "\n": # start of next entry
                             state = 0
                             p = [self.parent[id]]
                             if id == "1":
                                 p = []
                             if branch == "HEAD":
                                 branch = ""
                             if branch:
                                 latest = 0
                                 # the last changeset that contains a base
                                 # file is our parent
                                 for r in oldrevs:
                                     latest = max(filerevids.get(r, 0), latest)
                                 if latest:
                                     p = [latest]
 
                             # add current commit to set
                             c = commit(author=author, date=date, parents=p,
                                        desc=log, branch=branch)
                             self.changeset[id] = c
                             self.files[id] = files
                         else:
                             colon = l.rfind(':')
                             file = l[1:colon]
                             rev = l[colon+1:-2]
                             oldrev, rev = rev.split("->")
                             files[file] = rev
 
                             # save some information for identifying branch points
                             oldrevs.append("%s:%s" % (oldrev, file))
                             filerevids["%s:%s" % (rev, file)] = id
                     elif state == 3:
                         # swallow all input
                         continue
 
             self.heads = self.lastbranch.values()
         finally:
             os.chdir(d)
 
     def _connect(self):
         root = self.cvsroot
         conntype = None
         user, host = None, None
         cmd = ['cvs', 'server']
 
         self.ui.status(_("connecting to %s\n") % root)
 
         if root.startswith(":pserver:"):
             root = root[9:]
             m = re.match(r'(?:(.*?)(?::(.*?))?@)?([^:\/]*)(?::(\d*))?(.*)',
                          root)
             if m:
                 conntype = "pserver"
                 user, passw, serv, port, root = m.groups()
                 if not user:
                     user = "anonymous"
                 if not port:
                     port = 2401
                 else:
                     port = int(port)
                 format0 = ":pserver:%s@%s:%s" % (user, serv, root)
                 format1 = ":pserver:%s@%s:%d%s" % (user, serv, port, root)
 
                 if not passw:
                     passw = "A"
                     cvspass = os.path.expanduser("~/.cvspass")
                     try:
                         pf = open(cvspass)
                         for line in pf.read().splitlines():
                             part1, part2 = line.split(' ', 1)
                             if part1 == '/1':
                                 # /1 :pserver:[e-mail]:2401/cvsroot/foo Ah<Z
                                 part1, part2 = part2.split(' ', 1)
                                 format = format1
                             else:
                                 # :pserver:[e-mail]:/cvsroot/foo Ah<Z
                                 format = format0
                             if part1 == format:
                                 passw = part2
                                 break
                         pf.close()
                     except IOError, inst:
                         if inst.errno != errno.ENOENT:
                             if not getattr(inst, 'filename', None):
                                 inst.filename = cvspass
                             raise
 
                 sck = socket.socket()
                 sck.connect((serv, port))
                 sck.send("\n".join(["BEGIN AUTH REQUEST", root, user, passw,
                                     "END AUTH REQUEST", ""]))
                 if sck.recv(128) != "I LOVE YOU\n":
                     raise util.Abort(_("CVS pserver authentication failed"))
 
                 self.writep = self.readp = sck.makefile('r+')
 
         if not conntype and root.startswith(":local:"):
             conntype = "local"
             root = root[7:]
 
         if not conntype:
             # :ext:user@host/home/user/path/to/cvsroot
             if root.startswith(":ext:"):
                 root = root[5:]
             m = re.match(r'(?:([^@:/]+)@)?([^:/]+):?(.*)', root)
             # Do not take Windows path "c:\foo\bar" for a connection strings
             if os.path.isdir(root) or not m:
                 conntype = "local"
             else:
                 conntype = "rsh"
                 user, host, root = m.group(1), m.group(2), m.group(3)
 
         if conntype != "pserver":
             if conntype == "rsh":
                 rsh = os.environ.get("CVS_RSH") or "ssh"
                 if user:
                     cmd = [rsh, '-l', user, host] + cmd
                 else:
                     cmd = [rsh, host] + cmd
 
             # popen2 does not support argument lists under Windows
             cmd = [util.shellquote(arg) for arg in cmd]
             cmd = util.quotecommand(' '.join(cmd))
             self.writep, self.readp = util.popen2(cmd, 'b')
 
         self.realroot = root
 
         self.writep.write("Root %s\n" % root)
         self.writep.write("Valid-responses ok error Valid-requests Mode"
                           " M Mbinary E Checked-in Created Updated"
                           " Merged Removed\n")
         self.writep.write("valid-requests\n")
         self.writep.flush()
         r = self.readp.readline()
         if not r.startswith("Valid-requests"):
             raise util.Abort(_("server sucks"))
         if "UseUnchanged" in r:
             self.writep.write("UseUnchanged\n")
             self.writep.flush()
             r = self.readp.readline()
 
     def getheads(self):
         return self.heads
 
     def _getfile(self, name, rev):
 
         def chunkedread(fp, count):
             # file-objects returned by socked.makefile() do not handle
             # large read() requests very well.
             chunksize = 65536
             output = StringIO()
             while count > 0:
                 data = fp.read(min(count, chunksize))
                 if not data:
                     raise util.Abort(_("%d bytes missing from remote file") % count)
                 count -= len(data)
                 output.write(data)
             return output.getvalue()
 
         if rev.endswith("(DEAD)"):
             raise IOError
 
         args = ("-N -P -kk -r %s --" % rev).split()
         args.append(self.cvsrepo + '/' + name)
         for x in args:
             self.writep.write("Argument %s\n" % x)
         self.writep.write("Directory .\n%s\nco\n" % self.realroot)
         self.writep.flush()
 
         data = ""
         while 1:
             line = self.readp.readline()
             if line.startswith("Created ") or line.startswith("Updated "):
                 self.readp.readline() # path
                 self.readp.readline() # entries
                 mode = self.readp.readline()[:-1]
                 count = int(self.readp.readline()[:-1])
                 data = chunkedread(self.readp, count)
             elif line.startswith(" "):
                 data += line[1:]
             elif line.startswith("M "):
                 pass
             elif line.startswith("Mbinary "):
                 count = int(self.readp.readline()[:-1])
                 data = chunkedread(self.readp, count)
             else:
                 if line == "ok\n":
                     return (data, "x" in mode and "x" or "")
                 elif line.startswith("E "):
                     self.ui.warn(_("cvs server: %s\n") % line[2:])
                 elif line.startswith("Remove"):
                     l = self.readp.readline()
-                    l = self.readp.readline()
-                    if l != "ok\n":
-                        raise util.Abort(_("unknown CVS response: %s") % l)
                 else:
                     raise util.Abort(_("unknown CVS response: %s") % line)
 
     def getfile(self, file, rev):
         data, mode = self._getfile(file, rev)
         self.modecache[(file, rev)] = mode
         return data
 
     def getmode(self, file, rev):
         return self.modecache[(file, rev)]
 
     def getchanges(self, rev):
         self.modecache = {}
         return util.sort(self.files[rev].items()), {}
 
     def getcommit(self, rev):
         return self.changeset[rev]
 
     def gettags(self):
         return self.tags
 
     def getchangedfiles(self, rev, i):
         return util.sort(self.files[rev].keys())

mercurial/context.py

@@ -1,806 +1,806 b''
 # context.py - changeset and file context objects for mercurial
 #
 # Copyright 2006, 2007 Matt Mackall <[e-mail]>
 #
 # This software may be used and distributed according to the terms
 # of the GNU General Public License, incorporated herein by reference.
 
 from node import nullid, nullrev, short, hex
 from i18n import _
 import ancestor, bdiff, revlog, util, os, errno
 
 class propertycache(object):
     def __init__(self, func):
         self.func = func
         self.name = func.__name__
     def __get__(self, obj, type=None):
         result = self.func(obj)
         setattr(obj, self.name, result)
         return result
 
 class changectx(object):
     """A changecontext object makes access to data related to a particular
     changeset convenient."""
     def __init__(self, repo, changeid=''):
         """changeid is a revision number, node, or tag"""
         if changeid == '':
             changeid = '.'
         self._repo = repo
         if isinstance(changeid, (long, int)):
             self._rev = changeid
             self._node = self._repo.changelog.node(changeid)
         else:
             self._node = self._repo.lookup(changeid)
             self._rev = self._repo.changelog.rev(self._node)
 
     def __str__(self):
         return short(self.node())
 
     def __int__(self):
         return self.rev()
 
     def __repr__(self):
         return "<changectx %s>" % str(self)
 
     def __hash__(self):
         try:
             return hash(self._rev)
         except AttributeError:
             return id(self)
 
     def __eq__(self, other):
         try:
             return self._rev == other._rev
         except AttributeError:
             return False
 
     def __ne__(self, other):
         return not (self == other)
 
     def __nonzero__(self):
         return self._rev != nullrev
 
     def _changeset(self):
         return self._repo.changelog.read(self.node())
     _changeset = propertycache(_changeset)
 
     def _manifest(self):
         return self._repo.manifest.read(self._changeset[0])
     _manifest = propertycache(_manifest)
 
     def _manifestdelta(self):
         return self._repo.manifest.readdelta(self._changeset[0])
     _manifestdelta = propertycache(_manifestdelta)
 
     def _parents(self):
         p = self._repo.changelog.parentrevs(self._rev)
         if p[1] == nullrev:
             p = p[:-1]
         return [changectx(self._repo, x) for x in p]
     _parents = propertycache(_parents)
 
     def __contains__(self, key):
         return key in self._manifest
 
     def __getitem__(self, key):
         return self.filectx(key)
 
     def __iter__(self):
         for f in util.sort(self._manifest):
             yield f
 
     def changeset(self): return self._changeset
     def manifest(self): return self._manifest
 
     def rev(self): return self._rev
     def node(self): return self._node
     def hex(self): return hex(self._node)
     def user(self): return self._changeset[1]
     def date(self): return self._changeset[2]
     def files(self): return self._changeset[3]
     def description(self): return self._changeset[4]
     def branch(self): return self._changeset[5].get("branch")
     def extra(self): return self._changeset[5]
     def tags(self): return self._repo.nodetags(self._node)
 
     def parents(self):
         """return contexts for each parent changeset"""
         return self._parents
 
     def children(self):
         """return contexts for each child changeset"""
         c = self._repo.changelog.children(self._node)
         return [changectx(self._repo, x) for x in c]
 
     def ancestors(self):
         for a in self._repo.changelog.ancestors(self._rev):
             yield changectx(self._repo, a)
 
     def descendants(self):
         for d in self._repo.changelog.descendants(self._rev):
             yield changectx(self._repo, d)
 
     def _fileinfo(self, path):
         if '_manifest' in self.__dict__:
             try:
                 return self._manifest[path], self._manifest.flags(path)
             except KeyError:
                 raise revlog.LookupError(self._node, path,
                                          _('not found in manifest'))
         if '_manifestdelta' in self.__dict__ or path in self.files():
             if path in self._manifestdelta:
                 return self._manifestdelta[path], self._manifestdelta.flags(path)
         node, flag = self._repo.manifest.find(self._changeset[0], path)
         if not node:
             raise revlog.LookupError(self._node, path,
                                      _('not found in manifest'))
 
         return node, flag
 
     def filenode(self, path):
         return self._fileinfo(path)[0]
 
     def flags(self, path):
         try:
             return self._fileinfo(path)[1]
         except revlog.LookupError:
             return ''
 
     def filectx(self, path, fileid=None, filelog=None):
         """get a file context from this changeset"""
         if fileid is None:
             fileid = self.filenode(path)
         return filectx(self._repo, path, fileid=fileid,
                        changectx=self, filelog=filelog)
 
     def ancestor(self, c2):
         """
         return the ancestor context of self and c2
         """
         n = self._repo.changelog.ancestor(self._node, c2._node)
         return changectx(self._repo, n)
 
     def walk(self, match):
         fdict = dict.fromkeys(match.files())
         # for dirstate.walk, files=['.'] means "walk the whole tree".
         # follow that here, too
         fdict.pop('.', None)
         for fn in self:
             for ffn in fdict:
                 # match if the file is the exact name or a directory
                 if ffn == fn or fn.startswith("%s/" % ffn):
                     del fdict[ffn]
                     break
             if match(fn):
                 yield fn
         for fn in util.sort(fdict):
             if match.bad(fn, 'No such file in rev ' + str(self)) and match(fn):
                 yield fn
 
 class filectx(object):
     """A filecontext object makes access to data related to a particular
        filerevision convenient."""
     def __init__(self, repo, path, changeid=None, fileid=None,
                  filelog=None, changectx=None):
         """changeid can be a changeset revision, node, or tag.
            fileid can be a file revision or node."""
         self._repo = repo
         self._path = path
 
         assert (changeid is not None
                 or fileid is not None
                 or changectx is not None)
 
         if filelog:
             self._filelog = filelog
 
         if changeid is not None:
             self._changeid = changeid
         if changectx is not None:
             self._changectx = changectx
         if fileid is not None:
             self._fileid = fileid
 
     def _changectx(self):
         return changectx(self._repo, self._changeid)
     _changectx = propertycache(_changectx)
 
     def _filelog(self):
         return self._repo.file(self._path)
     _filelog = propertycache(_filelog)
 
     def _changeid(self):
         if '_changectx' in self.__dict__:
             return self._changectx.rev()
         else:
             return self._filelog.linkrev(self._filerev)
     _changeid = propertycache(_changeid)
 
     def _filenode(self):
         if '_fileid' in self.__dict__:
             return self._filelog.lookup(self._fileid)
         else:
             return self._changectx.filenode(self._path)
     _filenode = propertycache(_filenode)
 
     def _filerev(self):
         return self._filelog.rev(self._filenode)
     _filerev = propertycache(_filerev)
 
     def _repopath(self):
         return self._path
     _repopath = propertycache(_repopath)
 
     def __nonzero__(self):
         try:
             n = self._filenode
             return True
         except revlog.LookupError:
             # file is missing
             return False
 
     def __str__(self):
         return "%s@%s" % (self.path(), short(self.node()))
 
     def __repr__(self):
         return "<filectx %s>" % str(self)
 
     def __hash__(self):
         try:
             return hash((self._path, self._fileid))
         except AttributeError:
             return id(self)
 
     def __eq__(self, other):
         try:
             return (self._path == other._path
                     and self._fileid == other._fileid)
         except AttributeError:
             return False
 
     def __ne__(self, other):
         return not (self == other)
 
     def filectx(self, fileid):
         '''opens an arbitrary revision of the file without
         opening a new filelog'''
         return filectx(self._repo, self._path, fileid=fileid,
                        filelog=self._filelog)
 
     def filerev(self): return self._filerev
     def filenode(self): return self._filenode
     def flags(self): return self._changectx.flags(self._path)
     def filelog(self): return self._filelog
 
     def rev(self):
         if '_changectx' in self.__dict__:
             return self._changectx.rev()
         if '_changeid' in self.__dict__:
             return self._changectx.rev()
         return self._filelog.linkrev(self._filerev)
 
     def linkrev(self): return self._filelog.linkrev(self._filerev)
     def node(self): return self._changectx.node()
     def user(self): return self._changectx.user()
     def date(self): return self._changectx.date()
     def files(self): return self._changectx.files()
     def description(self): return self._changectx.description()
     def branch(self): return self._changectx.branch()
     def manifest(self): return self._changectx.manifest()
     def changectx(self): return self._changectx
 
     def data(self): return self._filelog.read(self._filenode)
     def path(self): return self._path
     def size(self): return self._filelog.size(self._filerev)
 
     def cmp(self, text): return self._filelog.cmp(self._filenode, text)
 
     def renamed(self):
         """check if file was actually renamed in this changeset revision
 
         If rename logged in file revision, we report copy for changeset only
         if file revisions linkrev points back to the changeset in question
         or both changeset parents contain different file revisions.
         """
 
         renamed = self._filelog.renamed(self._filenode)
         if not renamed:
             return renamed
 
         if self.rev() == self.linkrev():
             return renamed
 
         name = self.path()
         fnode = self._filenode
         for p in self._changectx.parents():
             try:
                 if fnode == p.filenode(name):
                     return None
             except revlog.LookupError:
                 pass
         return renamed
 
     def parents(self):
         p = self._path
         fl = self._filelog
         pl = [(p, n, fl) for n in self._filelog.parents(self._filenode)]
 
         r = self._filelog.renamed(self._filenode)
         if r:
             pl[0] = (r[0], r[1], None)
 
         return [filectx(self._repo, p, fileid=n, filelog=l)
                 for p,n,l in pl if n != nullid]
 
     def children(self):
         # hard for renames
         c = self._filelog.children(self._filenode)
         return [filectx(self._repo, self._path, fileid=x,
                         filelog=self._filelog) for x in c]
 
     def annotate(self, follow=False, linenumber=None):
         '''returns a list of tuples of (ctx, line) for each line
         in the file, where ctx is the filectx of the node where
         that line was last changed.
         This returns tuples of ((ctx, linenumber), line) for each line,
         if "linenumber" parameter is NOT "None".
         In such tuples, linenumber means one at the first appearance
         in the managed file.
         To reduce annotation cost,
         this returns fixed value(False is used) as linenumber,
         if "linenumber" parameter is "False".'''
 
         def decorate_compat(text, rev):
             return ([rev] * len(text.splitlines()), text)
 
         def without_linenumber(text, rev):
             return ([(rev, False)] * len(text.splitlines()), text)
 
         def with_linenumber(text, rev):
             size = len(text.splitlines())
             return ([(rev, i) for i in xrange(1, size + 1)], text)
 
         decorate = (((linenumber is None) and decorate_compat) or
                     (linenumber and with_linenumber) or
                     without_linenumber)
 
         def pair(parent, child):
             for a1, a2, b1, b2 in bdiff.blocks(parent[1], child[1]):
                 child[0][b1:b2] = parent[0][a1:a2]
             return child
 
         getlog = util.cachefunc(lambda x: self._repo.file(x))
         def getctx(path, fileid):
             log = path == self._path and self._filelog or getlog(path)
             return filectx(self._repo, path, fileid=fileid, filelog=log)
         getctx = util.cachefunc(getctx)
 
         def parents(f):
             # we want to reuse filectx objects as much as possible
             p = f._path
             if f._filerev is None: # working dir
                 pl = [(n.path(), n.filerev()) for n in f.parents()]
             else:
                 pl = [(p, n) for n in f._filelog.parentrevs(f._filerev)]
 
             if follow:
                 r = f.renamed()
                 if r:
                     pl[0] = (r[0], getlog(r[0]).rev(r[1]))
 
             return [getctx(p, n) for p, n in pl if n != nullrev]
 
         # use linkrev to find the first changeset where self appeared
         if self.rev() != self.linkrev():
             base = self.filectx(self.filerev())
         else:
             base = self
 
         # find all ancestors
         needed = {base: 1}
         visit = [base]
         files = [base._path]
         while visit:
             f = visit.pop(0)
             for p in parents(f):
                 if p not in needed:
                     needed[p] = 1
                     visit.append(p)
                     if p._path not in files:
                         files.append(p._path)
                 else:
                     # count how many times we'll use this
                     needed[p] += 1
 
         # sort by revision (per file) which is a topological order
         visit = []
         for f in files:
             fn = [(n.rev(), n) for n in needed if n._path == f]
             visit.extend(fn)
 
         hist = {}
         for r, f in util.sort(visit):
             curr = decorate(f.data(), f)
             for p in parents(f):
                 if p != nullid:
                     curr = pair(hist[p], curr)
                     # trim the history of unneeded revs
                     needed[p] -= 1
                     if not needed[p]:
                         del hist[p]
             hist[f] = curr
 
         return zip(hist[f][0], hist[f][1].splitlines(1))
 
     def ancestor(self, fc2):
         """
         find the common ancestor file context, if any, of self, and fc2
         """
 
         acache = {}
 
         # prime the ancestor cache for the working directory
         for c in (self, fc2):
             if c._filerev == None:
                 pl = [(n.path(), n.filenode()) for n in c.parents()]
                 acache[(c._path, None)] = pl
 
         flcache = {self._repopath:self._filelog, fc2._repopath:fc2._filelog}
         def parents(vertex):
             if vertex in acache:
                 return acache[vertex]
             f, n = vertex
             if f not in flcache:
                 flcache[f] = self._repo.file(f)
             fl = flcache[f]
             pl = [(f, p) for p in fl.parents(n) if p != nullid]
             re = fl.renamed(n)
             if re:
                 pl.append(re)
             acache[vertex] = pl
             return pl
 
         a, b = (self._path, self._filenode), (fc2._path, fc2._filenode)
         v = ancestor.ancestor(a, b, parents)
         if v:
             f, n = v
             return filectx(self._repo, f, fileid=n, filelog=flcache[f])
 
         return None
 
 class workingctx(changectx):
     """A workingctx object makes access to data related to
     the current working directory convenient.
     parents - a pair of parent nodeids, or None to use the dirstate.
     date - any valid date string or (unixtime, offset), or None.
     user - username string, or None.
     extra - a dictionary of extra values, or None.
     changes - a list of file lists as returned by localrepo.status()
                or None to use the repository status.
     """
     def __init__(self, repo, parents=None, text="", user=None, date=None,
                  extra=None, changes=None):
         self._repo = repo
         self._rev = None
         self._node = None
         self._text = text
         if date:
             self._date = util.parsedate(date)
         if user:
             self._user = user
         if parents:
             self._parents = [changectx(self._repo, p) for p in parents]
         if changes:
             self._status = list(changes)
 
         self._extra = {}
         if extra:
             self._extra = extra.copy()
         if 'branch' not in self._extra:
             branch = self._repo.dirstate.branch()
             try:
                 branch = branch.decode('UTF-8').encode('UTF-8')
             except UnicodeDecodeError:
                 raise util.Abort(_('branch name not in UTF-8!'))
             self._extra['branch'] = branch
         if self._extra['branch'] == '':
             self._extra['branch'] = 'default'
 
     def __str__(self):
         return str(self._parents[0]) + "+"
 
     def __nonzero__(self):
         return True
 
     def __contains__(self, key):
         return self._dirstate[key] not in "?r"
 
     def _manifest(self):
         """generate a manifest corresponding to the working directory"""
 
         man = self._parents[0].manifest().copy()
         copied = self._repo.dirstate.copies()
         cf = lambda x: man.flags(copied.get(x, x))
         ff = self._repo.dirstate.flagfunc(cf)
         modified, added, removed, deleted, unknown = self._status[:5]
         for i, l in (("a", added), ("m", modified), ("u", unknown)):
             for f in l:
                 man[f] = man.get(copied.get(f, f), nullid) + i
                 try:
                     man.set(f, ff(f))
                 except OSError:
                     pass
 
         for f in deleted + removed:
             if f in man:
                 del man[f]
 
         return man
     _manifest = propertycache(_manifest)
 
     def _status(self):
         return self._repo.status(unknown=True)
     _status = propertycache(_status)
 
     def _user(self):
         return self._repo.ui.username()
     _user = propertycache(_user)
 
     def _date(self):
         return util.makedate()
     _date = propertycache(_date)
 
     def _parents(self):
         p = self._repo.dirstate.parents()
         if p[1] == nullid:
             p = p[:-1]
         self._parents = [changectx(self._repo, x) for x in p]
         return self._parents
     _parents = propertycache(_parents)
 
     def manifest(self): return self._manifest
 
     def user(self): return self._user or self._repo.ui.username()
     def date(self): return self._date
     def description(self): return self._text
     def files(self):
         return util.sort(self._status[0] + self._status[1] + self._status[2])
 
     def modified(self): return self._status[0]
     def added(self): return self._status[1]
     def removed(self): return self._status[2]
     def deleted(self): return self._status[3]
     def unknown(self): return self._status[4]
     def clean(self): return self._status[5]
     def branch(self): return self._extra['branch']
     def extra(self): return self._extra
 
     def tags(self):
         t = []
         [t.extend(p.tags()) for p in self.parents()]
         return t
 
     def children(self):
         return []
 
     def flags(self, path):
         if '_manifest' in self.__dict__:
             try:
                 return self._manifest.flags(path)
             except KeyError:
                 return ''
 
         pnode = self._parents[0].changeset()[0]
         orig = self._repo.dirstate.copies().get(path, path)
         node, flag = self._repo.manifest.find(pnode, orig)
         try:
             ff = self._repo.dirstate.flagfunc(lambda x: flag or '')
             return ff(path)
         except OSError:
             pass
 
         if not node or path in self.deleted() or path in self.removed():
             return ''
         return flag
 
     def filectx(self, path, filelog=None):
         """get a file context from the working directory"""
         return workingfilectx(self._repo, path, workingctx=self,
                               filelog=filelog)
 
     def ancestor(self, c2):
         """return the ancestor context of self and c2"""
         return self._parents[0].ancestor(c2) # punt on two parents for now
 
     def walk(self, match):
         return util.sort(self._repo.dirstate.walk(match, True, False).keys())
 
 class workingfilectx(filectx):
     """A workingfilectx object makes access to data related to a particular
        file in the working directory convenient."""
     def __init__(self, repo, path, filelog=None, workingctx=None):
         """changeid can be a changeset revision, node, or tag.
            fileid can be a file revision or node."""
         self._repo = repo
         self._path = path
         self._changeid = None
         self._filerev = self._filenode = None
 
         if filelog:
             self._filelog = filelog
         if workingctx:
             self._changectx = workingctx
 
     def _changectx(self):
         return workingctx(self._repo)
     _changectx = propertycache(_changectx)
 
     def _repopath(self):
         return self._repo.dirstate.copied(self._path) or self._path
     _repopath = propertycache(_repopath)
 
     def _filelog(self):
         return self._repo.file(self._repopath)
     _filelog = propertycache(_filelog)
 
     def __nonzero__(self):
         return True
 
     def __str__(self):
         return "%s@%s" % (self.path(), self._changectx)
 
     def filectx(self, fileid):
         '''opens an arbitrary revision of the file without
         opening a new filelog'''
         return filectx(self._repo, self._repopath, fileid=fileid,
                        filelog=self._filelog)
 
     def rev(self):
         if '_changectx' in self.__dict__:
             return self._changectx.rev()
         return self._filelog.linkrev(self._filerev)
 
     def data(self): return self._repo.wread(self._path)
     def renamed(self):
         rp = self._repopath
         if rp == self._path:
             return None
         return rp, self._changectx._parents[0]._manifest.get(rp, nullid)
 
     def parents(self):
         '''return parent filectxs, following copies if necessary'''
         p = self._path
         rp = self._repopath
         pcl = self._changectx._parents
         fl = self._filelog
         pl = [(rp, pcl[0]._manifest.get(rp, nullid), fl)]
         if len(pcl) > 1:
             if rp != p:
                 fl = None
             pl.append((p, pcl[1]._manifest.get(p, nullid), fl))
 
         return [filectx(self._repo, p, fileid=n, filelog=l)
                 for p,n,l in pl if n != nullid]
 
     def children(self):
         return []
 
     def size(self): return os.stat(self._repo.wjoin(self._path)).st_size
     def date(self):
         t, tz = self._changectx.date()
         try:
             return (int(os.lstat(self._repo.wjoin(self._path)).st_mtime), tz)
         except OSError, err:
             if err.errno != errno.ENOENT: raise
             return (t, tz)
 
     def cmp(self, text): return self._repo.wread(self._path) == text
 
 class memctx(object):
     """Use memctx to perform in-memory commits via localrepo.commitctx().
 
     Revision information is supplied at initialization time while
     related files data and is made available through a callback
     mechanism.  'repo' is the current localrepo, 'parents' is a
     sequence of two parent revisions identifiers (pass None for every
     missing parent), 'text' is the commit message and 'files' lists
     names of files touched by the revision (normalized and relative to
     repository root).
 
     filectxfn(repo, memctx, path) is a callable receiving the
     repository, the current memctx object and the normalized path of
     requested file, relative to repository root. It is fired by the
     commit function for every file in 'files', but calls order is
     undefined. If the file is available in the revision being
     committed (updated or added), filectxfn returns a memfilectx
     object. If the file was removed, filectxfn raises an
     IOError. Moved files are represented by marking the source file
     removed and the new file added with copy information (see
     memfilectx).
 
     user receives the committer name and defaults to current
     repository username, date is the commit date in any format
     supported by util.parsedate() and defaults to current date, extra
     is a dictionary of metadata or is left empty.
     """
     def __init__(self, repo, parents, text, files, filectxfn, user=None,
                  date=None, extra=None):
         self._repo = repo
         self._rev = None
         self._node = None
         self._text = text
         self._date = date and util.parsedate(date) or util.makedate()
         self._user = user
         parents = [(p or nullid) for p in parents]
         p1, p2 = parents
         self._parents = [changectx(self._repo, p) for p in (p1, p2)]
-        files = util.sort(list(files))
+        files = util.sort(util.unique(files))
         self._status = [files, [], [], [], []]
         self._filectxfn = filectxfn
 
         self._extra = extra and extra.copy() or {}
         if 'branch' not in self._extra:
             self._extra['branch'] = 'default'
         elif self._extra.get('branch') == '':
             self._extra['branch'] = 'default'
 
     def __str__(self):
         return str(self._parents[0]) + "+"
 
     def __int__(self):
         return self._rev
 
     def __nonzero__(self):
         return True
 
     def user(self): return self._user or self._repo.ui.username()
     def date(self): return self._date
     def description(self): return self._text
     def files(self): return self.modified()
     def modified(self): return self._status[0]
     def added(self): return self._status[1]
     def removed(self): return self._status[2]
     def deleted(self): return self._status[3]
     def unknown(self): return self._status[4]
     def clean(self): return self._status[5]
     def branch(self): return self._extra['branch']
     def extra(self): return self._extra
     def flags(self, f): return self[f].flags()
 
     def parents(self):
         """return contexts for each parent changeset"""
         return self._parents
 
     def filectx(self, path, filelog=None):
         """get a file context from the working directory"""
         return self._filectxfn(self._repo, self, path)
 
 class memfilectx(object):
     """memfilectx represents an in-memory file to commit.
 
     See memctx for more details.
     """
     def __init__(self, path, data, islink, isexec, copied):
         """
         path is the normalized file path relative to repository root.
         data is the file content as a string.
         islink is True if the file is a symbolic link.
         isexec is True if the file is executable.
         copied is the source file path if current file was copied in the
         revision being committed, or None."""
         self._path = path
         self._data = data
         self._flags = (islink and 'l' or '') + (isexec and 'x' or '')
         self._copied = None
         if copied:
             self._copied = (copied, nullid)
 
     def __nonzero__(self): return True
     def __str__(self): return "%s@%s" % (self.path(), self._changectx)
     def path(self): return self._path
     def data(self): return self._data
     def flags(self): return self._flags
     def isexec(self): return 'x' in self._flags
     def islink(self): return 'l' in self._flags
     def renamed(self): return self._copied
 

mercurial/store.py

@@ -1,291 +1,297 b''
 # store.py - repository store handling for Mercurial
 #
 # Copyright 2008 Matt Mackall <[e-mail]>
 #
 # This software may be used and distributed according to the terms
 # of the GNU General Public License, incorporated herein by reference.
 
 from i18n import _
 import os, stat, osutil, util
 
 _sha = util.sha1
 
 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 _build_lower_encodefun():
     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):
         cmap[chr(x)] = chr(x).lower()
     return lambda s: "".join([cmap[c] for c in s])
 
 lowerencode = _build_lower_encodefun()
 
 _windows_reserved_filenames = '''con prn aux nul
     com1 com2 com3 com4 com5 com6 com7 com8 com9
     lpt1 lpt2 lpt3 lpt4 lpt5 lpt6 lpt7 lpt8 lpt9'''.split()
 def auxencode(path):
     res = []
     for n in path.split('/'):
         if n:
             base = n.split('.')[0]
             if base and (base in _windows_reserved_filenames):
                 # encode third letter ('aux' -> 'au~78')
                 ec = "~%02x" % ord(n[2])
                 n = n[0:2] + ec + n[3:]
+            if n[-1] in '. ':
+                # encode last period or space ('foo...' -> 'foo..~2e')
+                n = n[:-1] + "~%02x" % ord(n[-1])
         res.append(n)
     return '/'.join(res)
 
 MAX_PATH_LEN_IN_HGSTORE = 120
 DIR_PREFIX_LEN = 8
 _MAX_SHORTENED_DIRS_LEN = 8 * (DIR_PREFIX_LEN + 1) - 4
 def hybridencode(path):
     '''encodes path with a length limit
 
     Encodes all paths that begin with 'data/', according to the following.
 
     Default encoding (reversible):
 
     Encodes all uppercase letters 'X' as '_x'. All reserved or illegal
     characters are encoded as '~xx', where xx is the two digit hex code
     of the character (see encodefilename).
     Relevant path components consisting of Windows reserved filenames are
     masked by encoding the third character ('aux' -> 'au~78', see auxencode).
 
     Hashed encoding (not reversible):
 
     If the default-encoded path is longer than MAX_PATH_LEN_IN_HGSTORE, a
     non-reversible hybrid hashing of the path is done instead.
     This encoding uses up to DIR_PREFIX_LEN characters of all directory
     levels of the lowerencoded path, but not more levels than can fit into
     _MAX_SHORTENED_DIRS_LEN.
     Then follows the filler followed by the sha digest of the full path.
     The filler is the beginning of the basename of the lowerencoded path
     (the basename is everything after the last path separator). The filler
     is as long as possible, filling in characters from the basename until
     the encoded path has MAX_PATH_LEN_IN_HGSTORE characters (or all chars
     of the basename have been taken).
     The extension (e.g. '.i' or '.d') is preserved.
 
     The string 'data/' at the beginning is replaced with 'dh/', if the hashed
     encoding was used.
     '''
     if not path.startswith('data/'):
         return path
     ndpath = path[len('data/'):]
     res = 'data/' + auxencode(encodefilename(ndpath))
     if len(res) > MAX_PATH_LEN_IN_HGSTORE:
         digest = _sha(path).hexdigest()
         aep = auxencode(lowerencode(ndpath))
         _root, ext = os.path.splitext(aep)
         parts = aep.split('/')
         basename = parts[-1]
         sdirs = []
         for p in parts[:-1]:
             d = p[:DIR_PREFIX_LEN]
+            if d[-1] in '. ':
+                # Windows can't access dirs ending in period or space
+                d = d[:-1] + '_'
             t = '/'.join(sdirs) + '/' + d
             if len(t) > _MAX_SHORTENED_DIRS_LEN:
                 break
             sdirs.append(d)
         dirs = '/'.join(sdirs)
         if len(dirs) > 0:
             dirs += '/'
         res = 'dh/' + dirs + digest + ext
         space_left = MAX_PATH_LEN_IN_HGSTORE - len(res)
         if space_left > 0:
             filler = basename[:space_left]
             res = 'dh/' + dirs + filler + digest + ext
     return res
 
 def _calcmode(path):
     try:
         # files in .hg/ will be created using this mode
         mode = os.stat(path).st_mode
             # avoid some useless chmods
         if (0777 & ~util._umask) == (0777 & mode):
             mode = None
     except OSError:
         mode = None
     return mode
 
 _data = 'data 00manifest.d 00manifest.i 00changelog.d  00changelog.i'
 
 class basicstore:
     '''base class for local repository stores'''
     def __init__(self, path, opener, pathjoiner):
         self.pathjoiner = pathjoiner
         self.path = path
         self.createmode = _calcmode(path)
         self.opener = opener(self.path)
         self.opener.createmode = self.createmode
 
     def join(self, f):
         return self.pathjoiner(self.path, f)
 
     def _walk(self, relpath, recurse):
         '''yields (unencoded, encoded, size)'''
         path = self.pathjoiner(self.path, relpath)
         striplen = len(self.path) + len(os.sep)
         l = []
         if os.path.isdir(path):
             visit = [path]
             while visit:
                 p = visit.pop()
                 for f, kind, st in osutil.listdir(p, stat=True):
                     fp = self.pathjoiner(p, f)
                     if kind == stat.S_IFREG and f[-2:] in ('.d', '.i'):
                         n = util.pconvert(fp[striplen:])
                         l.append((n, n, st.st_size))
                     elif kind == stat.S_IFDIR and recurse:
                         visit.append(fp)
         return util.sort(l)
 
     def datafiles(self):
         return self._walk('data', True)
 
     def walk(self):
         '''yields (unencoded, encoded, size)'''
         # yield data files first
         for x in self.datafiles():
             yield x
         # yield manifest before changelog
         meta = self._walk('', False)
         meta.reverse()
         for x in meta:
             yield x
 
     def copylist(self):
         return ['requires'] + _data.split()
 
 class encodedstore(basicstore):
     def __init__(self, path, opener, pathjoiner):
         self.pathjoiner = pathjoiner
         self.path = self.pathjoiner(path, 'store')
         self.createmode = _calcmode(self.path)
         op = opener(self.path)
         op.createmode = self.createmode
         self.opener = lambda f, *args, **kw: op(encodefilename(f), *args, **kw)
 
     def datafiles(self):
         for a, b, size in self._walk('data', True):
             try:
                 a = decodefilename(a)
             except KeyError:
                 a = None
             yield a, b, size
 
     def join(self, f):
         return self.pathjoiner(self.path, encodefilename(f))
 
     def copylist(self):
         return (['requires', '00changelog.i'] +
                 [self.pathjoiner('store', f) for f in _data.split()])
 
 def fncache(opener):
     '''yields the entries in the fncache file'''
     try:
         fp = opener('fncache', mode='rb')
     except IOError:
         # skip nonexistent file
         return
     for n, line in enumerate(fp):
         if (len(line) < 2) or (line[-1] != '\n'):
             t = _('invalid entry in fncache, line %s') % (n + 1)
             raise util.Abort(t)
         yield line[:-1]
     fp.close()
 
 class fncacheopener(object):
     def __init__(self, opener):
         self.opener = opener
         self.entries = None
 
     def loadfncache(self):
         self.entries = {}
         for f in fncache(self.opener):
             self.entries[f] = True
 
     def __call__(self, path, mode='r', *args, **kw):
         if mode not in ('r', 'rb') and path.startswith('data/'):
             if self.entries is None:
                 self.loadfncache()
             if path not in self.entries:
                 self.opener('fncache', 'ab').write(path + '\n')
                 # fncache may contain non-existent files after rollback / strip
                 self.entries[path] = True
         return self.opener(hybridencode(path), mode, *args, **kw)
 
 class fncachestore(basicstore):
     def __init__(self, path, opener, pathjoiner):
         self.pathjoiner = pathjoiner
         self.path = self.pathjoiner(path, 'store')
         self.createmode = _calcmode(self.path)
         self._op = opener(self.path)
         self._op.createmode = self.createmode
         self.opener = fncacheopener(self._op)
 
     def join(self, f):
         return self.pathjoiner(self.path, hybridencode(f))
 
     def datafiles(self):
         rewrite = False
         existing = []
         pjoin = self.pathjoiner
         spath = self.path
         for f in fncache(self._op):
             ef = hybridencode(f)
             try:
                 st = os.stat(pjoin(spath, ef))
                 yield f, ef, st.st_size
                 existing.append(f)
             except OSError:
                 # nonexistent entry
                 rewrite = True
         if rewrite:
             # rewrite fncache to remove nonexistent entries
             # (may be caused by rollback / strip)
             fp = self._op('fncache', mode='wb')
             for p in existing:
                 fp.write(p + '\n')
             fp.close()
 
     def copylist(self):
         d = _data + ' dh fncache'
         return (['requires', '00changelog.i'] +
                 [self.pathjoiner('store', f) for f in d.split()])
 
 def store(requirements, path, opener, pathjoiner=None):
     pathjoiner = pathjoiner or os.path.join
     if 'store' in requirements:
         if 'fncache' in requirements:
             return fncachestore(path, opener, pathjoiner)
         return encodedstore(path, opener, pathjoiner)
     return basicstore(path, opener, pathjoiner)

tests/test-hybridencode.py

@@ -1,16 +1,19 b''
 #!/usr/bin/env python
 
 from mercurial import store
 
 enc = store.hybridencode # used for fncache repo format
 
 def show(s):
     print "A = '%s'" % s
     print "B = '%s'" % enc(s)
     print
 
 show('data/aux.bla/bla.aux/prn/PRN/lpt/com3/nul/coma/foo.NUL/normal.c.i')
 
 show('data/AUX/SECOND/X.PRN/FOURTH/FI:FTH/SIXTH/SEVENTH/EIGHTH/NINETH/TENTH/ELEVENTH/LOREMIPSUM.TXT.i')
 show('data/enterprise/openesbaddons/contrib-imola/corba-bc/netbeansplugin/wsdlExtension/src/main/java/META-INF/services/org.netbeans.modules.xml.wsdl.bindingsupport.spi.ExtensibilityElementTemplateProvider.i')
 show('data/AUX.THE-QUICK-BROWN-FOX-JU:MPS-OVER-THE-LAZY-DOG-THE-QUICK-BROWN-FOX-JUMPS-OVER-THE-LAZY-DOG.TXT.i')
+show('data/Project Planning/Resources/AnotherLongDirectoryName/Followedbyanother/AndAnother/AndThenAnExtremelyLongFileName.txt')
+show('data/Project.Planning/Resources/AnotherLongDirectoryName/Followedbyanother/AndAnother/AndThenAnExtremelyLongFileName.txt')
+show('data/foo.../foo   / /a./_. /__/.x../    bla/something.i')

tests/test-hybridencode.py.out

@@ -1,12 +1,21 b''
 A = 'data/aux.bla/bla.aux/prn/PRN/lpt/com3/nul/coma/foo.NUL/normal.c.i'
 B = 'data/au~78.bla/bla.aux/pr~6e/_p_r_n/lpt/co~6d3/nu~6c/coma/foo._n_u_l/normal.c.i'
 
 A = 'data/AUX/SECOND/X.PRN/FOURTH/FI:FTH/SIXTH/SEVENTH/EIGHTH/NINETH/TENTH/ELEVENTH/LOREMIPSUM.TXT.i'
 B = 'dh/au~78/second/x.prn/fourth/fi~3afth/sixth/seventh/eighth/nineth/tenth/loremia20419e358ddff1bf8751e38288aff1d7c32ec05.i'
 
 A = 'data/enterprise/openesbaddons/contrib-imola/corba-bc/netbeansplugin/wsdlExtension/src/main/java/META-INF/services/org.netbeans.modules.xml.wsdl.bindingsupport.spi.ExtensibilityElementTemplateProvider.i'
 B = 'dh/enterpri/openesba/contrib-/corba-bc/netbeans/wsdlexte/src/main/java/org.net7018f27961fdf338a598a40c4683429e7ffb9743.i'
 
 A = 'data/AUX.THE-QUICK-BROWN-FOX-JU:MPS-OVER-THE-LAZY-DOG-THE-QUICK-BROWN-FOX-JUMPS-OVER-THE-LAZY-DOG.TXT.i'
 B = 'dh/au~78.the-quick-brown-fox-ju~3amps-over-the-lazy-dog-the-quick-brown-fox-jud4dcadd033000ab2b26eb66bae1906bcb15d4a70.i'
 
+A = 'data/Project Planning/Resources/AnotherLongDirectoryName/Followedbyanother/AndAnother/AndThenAnExtremelyLongFileName.txt'
+B = 'dh/project_/resource/anotherl/followed/andanoth/andthenanextremelylongfilenaf93030515d9849cfdca52937c2204d19f83913e5.txt'
+
+A = 'data/Project.Planning/Resources/AnotherLongDirectoryName/Followedbyanother/AndAnother/AndThenAnExtremelyLongFileName.txt'
+B = 'dh/project_/resource/anotherl/followed/andanoth/andthenanextremelylongfilena0fd7c506f5c9d58204444fc67e9499006bd2d445.txt'
+
+A = 'data/foo.../foo   / /a./_. /__/.x../    bla/something.i'
+B = 'data/foo..~2e/foo  ~20/~20/a~2e/__.~20/____/.x.~2e/    bla/something.i'
+