check-code: find trailing whitespace

contrib/check-code.py

@@ -1,294 +1,295
 #!/usr/bin/env python
 #
 # check-code - a style and portability checker for Mercurial
 #
 # Copyright 2010 Matt Mackall <[e-mail]>
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 import re, glob, os, sys
 import optparse
 
 def repquote(m):
     t = re.sub(r"\w", "x", m.group('text'))
     t = re.sub(r"[^\sx]", "o", t)
     return m.group('quote') + t + m.group('quote')
 
 def reppython(m):
     comment = m.group('comment')
     if comment:
         return "#" * len(comment)
     return repquote(m)
 
 def repcomment(m):
     return m.group(1) + "#" * len(m.group(2))
 
 def repccomment(m):
     t = re.sub(r"((?<=\n) )|\S", "x", m.group(2))
     return m.group(1) + t + "*/"
 
 def repcallspaces(m):
     t = re.sub(r"\n\s+", "\n", m.group(2))
     return m.group(1) + t
 
 def repinclude(m):
     return m.group(1) + "<foo>"
 
 def rephere(m):
     t = re.sub(r"\S", "x", m.group(2))
     return m.group(1) + t
 
 
 testpats = [
     (r'(pushd|popd)', "don't use 'pushd' or 'popd', use 'cd'"),
     (r'\W\$?\(\([^\)]*\)\)', "don't use (()) or $(()), use 'expr'"),
     (r'^function', "don't use 'function', use old style"),
     (r'grep.*-q', "don't use 'grep -q', redirect to /dev/null"),
     (r'echo.*\\n', "don't use 'echo \\n', use printf"),
     (r'echo -n', "don't use 'echo -n', use printf"),
     (r'^diff.*-\w*N', "don't use 'diff -N'"),
     (r'(^| )wc[^|]*$', "filter wc output"),
     (r'head -c', "don't use 'head -c', use 'dd'"),
     (r'ls.*-\w*R', "don't use 'ls -R', use 'find'"),
     (r'printf.*\\\d\d\d', "don't use 'printf \NNN', use Python"),
     (r'printf.*\\x', "don't use printf \\x, use Python"),
     (r'\$\(.*\)', "don't use $(expr), use `expr`"),
     (r'rm -rf \*', "don't use naked rm -rf, target a directory"),
     (r'(^|\|\s*)grep (-\w\s+)*[^|]*[(|]\w',
      "use egrep for extended grep syntax"),
     (r'/bin/', "don't use explicit paths for tools"),
     (r'\$PWD', "don't use $PWD, use `pwd`"),
     (r'[^\n]\Z', "no trailing newline"),
     (r'export.*=', "don't export and assign at once"),
     ('^([^"\']|("[^"]*")|(\'[^\']*\'))*\\^', "^ must be quoted"),
     (r'^source\b', "don't use 'source', use '.'"),
     (r'touch -d', "don't use 'touch -d', use 'touch -t' instead"),
 ]
 
 testfilters = [
     (r"( *)(#([^\n]*\S)?)", repcomment),
     (r"<<(\S+)((.|\n)*?\n\1)", rephere),
 ]
 
 uprefix = r"^  \$ "
 uprefixc = r"^  > "
 utestpats = [
     (uprefix + r'.*\|\s*sed', "use regex test output patterns instead of sed"),
     (uprefix + r'(true|exit 0)', "explicit zero exit unnecessary"),
     (uprefix + r'.*\$\?', "explicit exit code checks unnecessary"),
     (uprefix + r'.*\|\| echo.*(fail|error)',
      "explicit exit code checks unnecessary"),
     (uprefix + r'set -e', "don't use set -e"),
     (uprefixc + r'( *)\t', "don't use tabs to indent"),
 ]
 
 for p, m in testpats:
     if p.startswith('^'):
         p = uprefix + p[1:]
     else:
         p = uprefix + p
     utestpats.append((p, m))
 
 utestfilters = [
     (r"( *)(#([^\n]*\S)?)", repcomment),
 ]
 
 pypats = [
     (r'^\s*def\s*\w+\s*\(.*,\s*\(',
      "tuple parameter unpacking not available in Python 3+"),
     (r'lambda\s*\(.*,.*\)',
      "tuple parameter unpacking not available in Python 3+"),
     (r'(?<!def)\s+(cmp)\(', "cmp is not available in Python 3+"),
     (r'\breduce\s*\(.*', "reduce is not available in Python 3+"),
     (r'\.has_key\b', "dict.has_key is not available in Python 3+"),
     (r'^\s*\t', "don't use tabs"),
     (r'\S;\s*\n', "semicolon"),
     (r'\w,\w', "missing whitespace after ,"),
     (r'\w[+/*\-<>]\w', "missing whitespace in expression"),
     (r'^\s+\w+=\w+[^,)]$', "missing whitespace in assignment"),
     (r'.{85}', "line too long"),
     (r'.{81}', "warning: line over 80 characters"),
     (r'[^\n]\Z', "no trailing newline"),
+    (r'(\S\s+|^\s+)\n', "trailing whitespace"),
 #    (r'^\s+[^_ ][^_. ]+_[^_]+\s*=', "don't use underbars in identifiers"),
 #    (r'\w*[a-z][A-Z]\w*\s*=', "don't use camelcase in identifiers"),
     (r'^\s*(if|while|def|class|except|try)\s[^[]*:\s*[^\]#\s]+',
      "linebreak after :"),
     (r'class\s[^(]:', "old-style class, use class foo(object)"),
     (r'^\s+del\(', "del isn't a function"),
     (r'^\s+except\(', "except isn't a function"),
     (r',]', "unneeded trailing ',' in list"),
 #    (r'class\s[A-Z][^\(]*\((?!Exception)',
 #     "don't capitalize non-exception classes"),
 #    (r'in range\(', "use xrange"),
 #    (r'^\s*print\s+', "avoid using print in core and extensions"),
     (r'[\x80-\xff]', "non-ASCII character literal"),
     (r'("\')\.format\(', "str.format() not available in Python 2.4"),
     (r'^\s*with\s+', "with not available in Python 2.4"),
     (r'(?<!def)\s+(any|all|format)\(',
      "any/all/format not available in Python 2.4"),
     (r'(?<!def)\s+(callable)\(',
      "callable not available in Python 3, use hasattr(f, '__call__')"),
     (r'if\s.*\selse', "if ... else form not available in Python 2.4"),
     (r'([\(\[]\s\S)|(\S\s[\)\]])', "gratuitous whitespace in () or []"),
 #    (r'\s\s=', "gratuitous whitespace before ="),
     (r'[^>< ](\+=|-=|!=|<>|<=|>=|<<=|>>=)\S',
      "missing whitespace around operator"),
     (r'[^>< ](\+=|-=|!=|<>|<=|>=|<<=|>>=)\s',
      "missing whitespace around operator"),
     (r'\s(\+=|-=|!=|<>|<=|>=|<<=|>>=)\S',
      "missing whitespace around operator"),
     (r'[^+=*!<>&| -](\s=|=\s)[^= ]',
      "wrong whitespace around ="),
     (r'raise Exception', "don't raise generic exceptions"),
     (r'ui\.(status|progress|write|note|warn)\([\'\"]x',
      "warning: unwrapped ui message"),
 ]
 
 pyfilters = [
     (r"""(?msx)(?P<comment>\#.*?$)|
          ((?P<quote>('''|\"\"\"|(?<!')'(?!')|(?<!")"(?!")))
           (?P<text>(([^\\]|\\.)*?))
           (?P=quote))""", reppython),
 ]
 
 cpats = [
     (r'//', "don't use //-style comments"),
     (r'^  ', "don't use spaces to indent"),
     (r'\S\t', "don't use tabs except for indent"),
     (r'(\S\s+|^\s+)\n', "trailing whitespace"),
     (r'.{85}', "line too long"),
     (r'(while|if|do|for)\(', "use space after while/if/do/for"),
     (r'return\(', "return is not a function"),
     (r' ;', "no space before ;"),
     (r'\w+\* \w+', "use int *foo, not int* foo"),
     (r'\([^\)]+\) \w+', "use (int)foo, not (int) foo"),
     (r'\S+ (\+\+|--)', "use foo++, not foo ++"),
     (r'\w,\w', "missing whitespace after ,"),
     (r'\w[+/*]\w', "missing whitespace in expression"),
     (r'^#\s+\w', "use #foo, not # foo"),
     (r'[^\n]\Z', "no trailing newline"),
 ]
 
 cfilters = [
     (r'(/\*)(((\*(?!/))|[^*])*)\*/', repccomment),
     (r'''(?P<quote>(?<!")")(?P<text>([^"]|\\")+)"(?!")''', repquote),
     (r'''(#\s*include\s+<)([^>]+)>''', repinclude),
     (r'(\()([^)]+\))', repcallspaces),
 ]
 
 checks = [
     ('python', r'.*\.(py|cgi)$', pyfilters, pypats),
     ('test script', r'(.*/)?test-[^.~]*$', testfilters, testpats),
     ('c', r'.*\.c$', cfilters, cpats),
     ('unified test', r'.*\.t$', utestfilters, utestpats),
 ]
 
 class norepeatlogger(object):
     def __init__(self):
         self._lastseen = None
 
     def log(self, fname, lineno, line, msg, blame):
         """print error related a to given line of a given file.
 
         The faulty line will also be printed but only once in the case
         of multiple errors.
 
         :fname: filename
         :lineno: line number
         :line: actual content of the line
         :msg: error message
         """
         msgid = fname, lineno, line
         if msgid != self._lastseen:
             if blame:
                 print "%s:%d (%s):" % (fname, lineno, blame)
             else:
                 print "%s:%d:" % (fname, lineno)
             print " > %s" % line
             self._lastseen = msgid
         print " " + msg
 
 _defaultlogger = norepeatlogger()
 
 def getblame(f):
     lines = []
     for l in os.popen('hg annotate -un %s' % f):
         start, line = l.split(':', 1)
         user, rev = start.split()
         lines.append((line[1:-1], user, rev))
     return lines
 
 def checkfile(f, logfunc=_defaultlogger.log, maxerr=None, warnings=False,
               blame=False):
     """checks style and portability of a given file
 
     :f: filepath
     :logfunc: function used to report error
               logfunc(filename, linenumber, linecontent, errormessage)
     :maxerr: number of error to display before arborting.
              Set to None (default) to report all errors
 
     return True if no error is found, False otherwise.
     """
     blamecache = None
     result = True
     for name, match, filters, pats in checks:
         fc = 0
         if not re.match(match, f):
             continue
         pre = post = open(f).read()
         if "no-" + "check-code" in pre:
             break
         for p, r in filters:
             post = re.sub(p, r, post)
         # print post # uncomment to show filtered version
         z = enumerate(zip(pre.splitlines(), post.splitlines(True)))
         for n, l in z:
             if "check-code" + "-ignore" in l[0]:
                 continue
             for p, msg in pats:
                 if not warnings and msg.startswith("warning"):
                     continue
                 if re.search(p, l[1]):
                     bd = ""
                     if blame:
                         bd = 'working directory'
                         if not blamecache:
                             blamecache = getblame(f)
                         if n < len(blamecache):
                             bl, bu, br = blamecache[n]
                             if bl == l[0]:
                                 bd = '%s@%s' % (bu, br)
                     logfunc(f, n + 1, l[0], msg, bd)
                     fc += 1
                     result = False
             if maxerr is not None and fc >= maxerr:
                 print " (too many errors, giving up)"
                 break
         break
     return result
 
 if __name__ == "__main__":
     parser = optparse.OptionParser("%prog [options] [files]")
     parser.add_option("-w", "--warnings", action="store_true",
                       help="include warning-level checks")
     parser.add_option("-p", "--per-file", type="int",
                       help="max warnings per file")
     parser.add_option("-b", "--blame", action="store_true",
                       help="use annotate to generate blame info")
 
     parser.set_defaults(per_file=15, warnings=False, blame=False)
     (options, args) = parser.parse_args()
 
     if len(args) == 0:
         check = glob.glob("*")
     else:
         check = args
 
     for f in check:
         ret = 0
         if not checkfile(f, maxerr=options.per_file, warnings=options.warnings,
                          blame=options.blame):
             ret = 1
     sys.exit(ret)

hgext/convert/subversion.py

@@ -1,1161 +1,1161
 # Subversion 1.4/1.5 Python API backend
 #
 # Copyright(C) 2007 Daniel Holth et al
 
 import os
 import re
 import sys
 import cPickle as pickle
 import tempfile
 import urllib
 import urllib2
 
 from mercurial import strutil, util, encoding
 from mercurial.i18n import _
 
 # Subversion stuff. Works best with very recent Python SVN bindings
 # e.g. SVN 1.5 or backports. Thanks to the bzr folks for enhancing
 # these bindings.
 
 from cStringIO import StringIO
 
 from common import NoRepo, MissingTool, commit, encodeargs, decodeargs
 from common import commandline, converter_source, converter_sink, mapfile
 
 try:
     from svn.core import SubversionException, Pool
     import svn
     import svn.client
     import svn.core
     import svn.ra
     import svn.delta
     import transport
     import warnings
     warnings.filterwarnings('ignore',
             module='svn.core',
             category=DeprecationWarning)
 
 except ImportError:
     pass
 
 class SvnPathNotFound(Exception):
     pass
 
 def geturl(path):
     try:
         return svn.client.url_from_path(svn.core.svn_path_canonicalize(path))
     except SubversionException:
         pass
     if os.path.isdir(path):
         path = os.path.normpath(os.path.abspath(path))
         if os.name == 'nt':
             path = '/' + util.normpath(path)
         # Module URL is later compared with the repository URL returned
         # by svn API, which is UTF-8.
         path = encoding.tolocal(path)
         return 'file://%s' % urllib.quote(path)
     return path
 
 def optrev(number):
     optrev = svn.core.svn_opt_revision_t()
     optrev.kind = svn.core.svn_opt_revision_number
     optrev.value.number = number
     return optrev
 
 class changedpath(object):
     def __init__(self, p):
         self.copyfrom_path = p.copyfrom_path
         self.copyfrom_rev = p.copyfrom_rev
         self.action = p.action
 
 def get_log_child(fp, url, paths, start, end, limit=0, discover_changed_paths=True,
                     strict_node_history=False):
     protocol = -1
     def receiver(orig_paths, revnum, author, date, message, pool):
         if orig_paths is not None:
             for k, v in orig_paths.iteritems():
                 orig_paths[k] = changedpath(v)
         pickle.dump((orig_paths, revnum, author, date, message),
                     fp, protocol)
 
     try:
         # Use an ra of our own so that our parent can consume
         # our results without confusing the server.
         t = transport.SvnRaTransport(url=url)
         svn.ra.get_log(t.ra, paths, start, end, limit,
                        discover_changed_paths,
                        strict_node_history,
                        receiver)
     except SubversionException, (inst, num):
         pickle.dump(num, fp, protocol)
     except IOError:
         # Caller may interrupt the iteration
         pickle.dump(None, fp, protocol)
     else:
         pickle.dump(None, fp, protocol)
     fp.close()
     # With large history, cleanup process goes crazy and suddenly
     # consumes *huge* amount of memory. The output file being closed,
     # there is no need for clean termination.
     os._exit(0)
 
 def debugsvnlog(ui, **opts):
     """Fetch SVN log in a subprocess and channel them back to parent to
     avoid memory collection issues.
     """
     util.set_binary(sys.stdin)
     util.set_binary(sys.stdout)
     args = decodeargs(sys.stdin.read())
     get_log_child(sys.stdout, *args)
 
 class logstream(object):
     """Interruptible revision log iterator."""
     def __init__(self, stdout):
         self._stdout = stdout
 
     def __iter__(self):
         while True:
             try:
                 entry = pickle.load(self._stdout)
             except EOFError:
                 raise util.Abort(_('Mercurial failed to run itself, check'
                                    ' hg executable is in PATH'))
             try:
                 orig_paths, revnum, author, date, message = entry
             except:
                 if entry is None:
                     break
                 raise SubversionException("child raised exception", entry)
             yield entry
 
     def close(self):
         if self._stdout:
             self._stdout.close()
             self._stdout = None
 
 
 # Check to see if the given path is a local Subversion repo. Verify this by
 # looking for several svn-specific files and directories in the given
 # directory.
 def filecheck(ui, path, proto):
     for x in ('locks', 'hooks', 'format', 'db'):
         if not os.path.exists(os.path.join(path, x)):
             return False
     return True
 
 # Check to see if a given path is the root of an svn repo over http. We verify
 # this by requesting a version-controlled URL we know can't exist and looking
 # for the svn-specific "not found" XML.
 def httpcheck(ui, path, proto):
     try:
         opener = urllib2.build_opener()
         rsp = opener.open('%s://%s/!svn/ver/0/.svn' % (proto, path))
         data = rsp.read()
     except urllib2.HTTPError, inst:
         if inst.code != 404:
             # Except for 404 we cannot know for sure this is not an svn repo
             ui.warn(_('svn: cannot probe remote repository, assume it could '
                       'be a subversion repository. Use --source-type if you '
                       'know better.\n'))
             return True
         data = inst.fp.read()
     except:
         # Could be urllib2.URLError if the URL is invalid or anything else.
         return False
     return '<m:human-readable errcode="160013">' in data
 
 protomap = {'http': httpcheck,
             'https': httpcheck,
             'file': filecheck,
             }
 def issvnurl(ui, url):
     try:
         proto, path = url.split('://', 1)
         if proto == 'file':
             path = urllib.url2pathname(path)
     except ValueError:
         proto = 'file'
         path = os.path.abspath(url)
     if proto == 'file':
         path = path.replace(os.sep, '/')
     check = protomap.get(proto, lambda *args: False)
     while '/' in path:
         if check(ui, path, proto):
             return True
         path = path.rsplit('/', 1)[0]
     return False
 
 # SVN conversion code stolen from bzr-svn and tailor
 #
 # Subversion looks like a versioned filesystem, branches structures
 # are defined by conventions and not enforced by the tool. First,
 # we define the potential branches (modules) as "trunk" and "branches"
 # children directories. Revisions are then identified by their
 # module and revision number (and a repository identifier).
 #
 # The revision graph is really a tree (or a forest). By default, a
 # revision parent is the previous revision in the same module. If the
 # module directory is copied/moved from another module then the
 # revision is the module root and its parent the source revision in
 # the parent module. A revision has at most one parent.
 #
 class svn_source(converter_source):
     def __init__(self, ui, url, rev=None):
         super(svn_source, self).__init__(ui, url, rev=rev)
 
         if not (url.startswith('svn://') or url.startswith('svn+ssh://') or
                 (os.path.exists(url) and
                  os.path.exists(os.path.join(url, '.svn'))) or
                 issvnurl(ui, url)):
             raise NoRepo(_("%s does not look like a Subversion repository")
                          % url)
 
         try:
             SubversionException
         except NameError:
             raise MissingTool(_('Subversion python bindings could not be loaded'))
 
         try:
             version = svn.core.SVN_VER_MAJOR, svn.core.SVN_VER_MINOR
             if version < (1, 4):
                 raise MissingTool(_('Subversion python bindings %d.%d found, '
                                     '1.4 or later required') % version)
         except AttributeError:
             raise MissingTool(_('Subversion python bindings are too old, 1.4 '
                                 'or later required'))
 
         self.lastrevs = {}
 
         latest = None
         try:
             # Support file://path@rev syntax. Useful e.g. to convert
             # deleted branches.
             at = url.rfind('@')
             if at >= 0:
                 latest = int(url[at + 1:])
                 url = url[:at]
         except ValueError:
             pass
         self.url = geturl(url)
         self.encoding = 'UTF-8' # Subversion is always nominal UTF-8
         try:
             self.transport = transport.SvnRaTransport(url=self.url)
             self.ra = self.transport.ra
             self.ctx = self.transport.client
             self.baseurl = svn.ra.get_repos_root(self.ra)
             # Module is either empty or a repository path starting with
             # a slash and not ending with a slash.
             self.module = urllib.unquote(self.url[len(self.baseurl):])
             self.prevmodule = None
             self.rootmodule = self.module
             self.commits = {}
             self.paths = {}
             self.uuid = svn.ra.get_uuid(self.ra)
         except SubversionException:
             ui.traceback()
             raise NoRepo(_("%s does not look like a Subversion repository")
                          % self.url)
 
         if rev:
             try:
                 latest = int(rev)
             except ValueError:
                 raise util.Abort(_('svn: revision %s is not an integer') % rev)
 
         self.startrev = self.ui.config('convert', 'svn.startrev', default=0)
         try:
             self.startrev = int(self.startrev)
             if self.startrev < 0:
                 self.startrev = 0
         except ValueError:
             raise util.Abort(_('svn: start revision %s is not an integer')
                              % self.startrev)
 
         self.head = self.latest(self.module, latest)
         if not self.head:
             raise util.Abort(_('no revision found in module %s')
                              % self.module)
         self.last_changed = self.revnum(self.head)
 
         self._changescache = None
 
         if os.path.exists(os.path.join(url, '.svn/entries')):
             self.wc = url
         else:
             self.wc = None
         self.convertfp = None
 
     def setrevmap(self, revmap):
         lastrevs = {}
         for revid in revmap.iterkeys():
             uuid, module, revnum = self.revsplit(revid)
             lastrevnum = lastrevs.setdefault(module, revnum)
             if revnum > lastrevnum:
                 lastrevs[module] = revnum
         self.lastrevs = lastrevs
 
     def exists(self, path, optrev):
         try:
             svn.client.ls(self.url.rstrip('/') + '/' + urllib.quote(path),
                                  optrev, False, self.ctx)
             return True
         except SubversionException:
             return False
 
     def getheads(self):
 
         def isdir(path, revnum):
             kind = self._checkpath(path, revnum)
             return kind == svn.core.svn_node_dir
 
         def getcfgpath(name, rev):
             cfgpath = self.ui.config('convert', 'svn.' + name)
             if cfgpath is not None and cfgpath.strip() == '':
                 return None
             path = (cfgpath or name).strip('/')
             if not self.exists(path, rev):
                 if cfgpath:
                     raise util.Abort(_('expected %s to be at %r, but not found')
                                  % (name, path))
                 return None
             self.ui.note(_('found %s at %r\n') % (name, path))
             return path
 
         rev = optrev(self.last_changed)
         oldmodule = ''
         trunk = getcfgpath('trunk', rev)
         self.tags = getcfgpath('tags', rev)
         branches = getcfgpath('branches', rev)
 
         # If the project has a trunk or branches, we will extract heads
         # from them. We keep the project root otherwise.
         if trunk:
             oldmodule = self.module or ''
             self.module += '/' + trunk
             self.head = self.latest(self.module, self.last_changed)
             if not self.head:
                 raise util.Abort(_('no revision found in module %s')
                                  % self.module)
 
         # First head in the list is the module's head
         self.heads = [self.head]
         if self.tags is not None:
             self.tags = '%s/%s' % (oldmodule , (self.tags or 'tags'))
 
         # Check if branches bring a few more heads to the list
         if branches:
             rpath = self.url.strip('/')
             branchnames = svn.client.ls(rpath + '/' + urllib.quote(branches),
                                         rev, False, self.ctx)
             for branch in branchnames.keys():
                 module = '%s/%s/%s' % (oldmodule, branches, branch)
                 if not isdir(module, self.last_changed):
                     continue
                 brevid = self.latest(module, self.last_changed)
                 if not brevid:
                     self.ui.note(_('ignoring empty branch %s\n') % branch)
                     continue
                 self.ui.note(_('found branch %s at %d\n') %
                              (branch, self.revnum(brevid)))
                 self.heads.append(brevid)
 
         if self.startrev and self.heads:
             if len(self.heads) > 1:
                 raise util.Abort(_('svn: start revision is not supported '
                                    'with more than one branch'))
             revnum = self.revnum(self.heads[0])
             if revnum < self.startrev:
                 raise util.Abort(
                     _('svn: no revision found after start revision %d')
                                  % self.startrev)
 
         return self.heads
 
     def getchanges(self, rev):
         if self._changescache and self._changescache[0] == rev:
             return self._changescache[1]
         self._changescache = None
         (paths, parents) = self.paths[rev]
         if parents:
             files, self.removed, copies = self.expandpaths(rev, paths, parents)
         else:
             # Perform a full checkout on roots
             uuid, module, revnum = self.revsplit(rev)
             entries = svn.client.ls(self.baseurl + urllib.quote(module),
                                     optrev(revnum), True, self.ctx)
             files = [n for n, e in entries.iteritems()
                      if e.kind == svn.core.svn_node_file]
             copies = {}
             self.removed = set()
 
         files.sort()
         files = zip(files, [rev] * len(files))
 
         # caller caches the result, so free it here to release memory
         del self.paths[rev]
         return (files, copies)
 
     def getchangedfiles(self, rev, i):
         changes = self.getchanges(rev)
         self._changescache = (rev, changes)
         return [f[0] for f in changes[0]]
 
     def getcommit(self, rev):
         if rev not in self.commits:
             uuid, module, revnum = self.revsplit(rev)
             self.module = module
             self.reparent(module)
             # We assume that:
             # - requests for revisions after "stop" come from the
             # revision graph backward traversal. Cache all of them
             # down to stop, they will be used eventually.
             # - requests for revisions before "stop" come to get
             # isolated branches parents. Just fetch what is needed.
             stop = self.lastrevs.get(module, 0)
             if revnum < stop:
                 stop = revnum + 1
             self._fetch_revisions(revnum, stop)
         commit = self.commits[rev]
         # caller caches the result, so free it here to release memory
         del self.commits[rev]
         return commit
 
     def gettags(self):
         tags = {}
         if self.tags is None:
             return tags
 
         # svn tags are just a convention, project branches left in a
         # 'tags' directory. There is no other relationship than
         # ancestry, which is expensive to discover and makes them hard
         # to update incrementally.  Worse, past revisions may be
         # referenced by tags far away in the future, requiring a deep
         # history traversal on every calculation.  Current code
         # performs a single backward traversal, tracking moves within
         # the tags directory (tag renaming) and recording a new tag
         # everytime a project is copied from outside the tags
         # directory. It also lists deleted tags, this behaviour may
         # change in the future.
         pendings = []
         tagspath = self.tags
         start = svn.ra.get_latest_revnum(self.ra)
         stream = self._getlog([self.tags], start, self.startrev)
         try:
             for entry in stream:
                 origpaths, revnum, author, date, message = entry
                 copies = [(e.copyfrom_path, e.copyfrom_rev, p) for p, e
                           in origpaths.iteritems() if e.copyfrom_path]
                 # Apply moves/copies from more specific to general
                 copies.sort(reverse=True)
 
                 srctagspath = tagspath
                 if copies and copies[-1][2] == tagspath:
                     # Track tags directory moves
                     srctagspath = copies.pop()[0]
 
                 for source, sourcerev, dest in copies:
                     if not dest.startswith(tagspath + '/'):
                         continue
                     for tag in pendings:
                         if tag[0].startswith(dest):
                             tagpath = source + tag[0][len(dest):]
                             tag[:2] = [tagpath, sourcerev]
                             break
                     else:
                         pendings.append([source, sourcerev, dest])
 
                 # Filter out tags with children coming from different
                 # parts of the repository like:
                 # /tags/tag.1 (from /trunk:10)
                 # /tags/tag.1/foo (from /branches/foo:12)
                 # Here/tags/tag.1 discarded as well as its children.
                 # It happens with tools like cvs2svn. Such tags cannot
                 # be represented in mercurial.
                 addeds = dict((p, e.copyfrom_path) for p, e
                               in origpaths.iteritems()
                               if e.action == 'A' and e.copyfrom_path)
                 badroots = set()
                 for destroot in addeds:
                     for source, sourcerev, dest in pendings:
                         if (not dest.startswith(destroot + '/')
                             or source.startswith(addeds[destroot] + '/')):
                             continue
                         badroots.add(destroot)
                         break
 
                 for badroot in badroots:
                     pendings = [p for p in pendings if p[2] != badroot
                                 and not p[2].startswith(badroot + '/')]
 
                 # Tell tag renamings from tag creations
                 remainings = []
                 for source, sourcerev, dest in pendings:
                     tagname = dest.split('/')[-1]
                     if source.startswith(srctagspath):
                         remainings.append([source, sourcerev, tagname])
                         continue
                     if tagname in tags:
                         # Keep the latest tag value
                         continue
                     # From revision may be fake, get one with changes
                     try:
                         tagid = self.latest(source, sourcerev)
                         if tagid and tagname not in tags:
                             tags[tagname] = tagid
                     except SvnPathNotFound:
                         # It happens when we are following directories
                         # we assumed were copied with their parents
                         # but were really created in the tag
                         # directory.
                         pass
                 pendings = remainings
                 tagspath = srctagspath
         finally:
             stream.close()
         return tags
 
     def converted(self, rev, destrev):
         if not self.wc:
             return
         if self.convertfp is None:
             self.convertfp = open(os.path.join(self.wc, '.svn', 'hg-shamap'),
                                   'a')
         self.convertfp.write('%s %d\n' % (destrev, self.revnum(rev)))
         self.convertfp.flush()
 
     def revid(self, revnum, module=None):
         return 'svn:%s%s@%s' % (self.uuid, module or self.module, revnum)
 
     def revnum(self, rev):
         return int(rev.split('@')[-1])
 
     def revsplit(self, rev):
         url, revnum = rev.rsplit('@', 1)
         revnum = int(revnum)
         parts = url.split('/', 1)
         uuid = parts.pop(0)[4:]
         mod = ''
         if parts:
             mod = '/' + parts[0]
         return uuid, mod, revnum
 
     def latest(self, path, stop=0):
         """Find the latest revid affecting path, up to stop. It may return
         a revision in a different module, since a branch may be moved without
         a change being reported. Return None if computed module does not
         belong to rootmodule subtree.
         """
         if not path.startswith(self.rootmodule):
             # Requests on foreign branches may be forbidden at server level
             self.ui.debug('ignoring foreign branch %r\n' % path)
             return None
 
         if not stop:
             stop = svn.ra.get_latest_revnum(self.ra)
         try:
             prevmodule = self.reparent('')
             dirent = svn.ra.stat(self.ra, path.strip('/'), stop)
             self.reparent(prevmodule)
         except SubversionException:
             dirent = None
         if not dirent:
             raise SvnPathNotFound(_('%s not found up to revision %d')
                                   % (path, stop))
 
         # stat() gives us the previous revision on this line of
         # development, but it might be in *another module*. Fetch the
         # log and detect renames down to the latest revision.
         stream = self._getlog([path], stop, dirent.created_rev)
         try:
             for entry in stream:
                 paths, revnum, author, date, message = entry
                 if revnum <= dirent.created_rev:
                     break
 
                 for p in paths:
                     if not path.startswith(p) or not paths[p].copyfrom_path:
                         continue
                     newpath = paths[p].copyfrom_path + path[len(p):]
                     self.ui.debug("branch renamed from %s to %s at %d\n" %
                                   (path, newpath, revnum))
                     path = newpath
                     break
         finally:
             stream.close()
 
         if not path.startswith(self.rootmodule):
             self.ui.debug('ignoring foreign branch %r\n' % path)
             return None
         return self.revid(dirent.created_rev, path)
 
     def reparent(self, module):
         """Reparent the svn transport and return the previous parent."""
         if self.prevmodule == module:
             return module
         svnurl = self.baseurl + urllib.quote(module)
         prevmodule = self.prevmodule
         if prevmodule is None:
             prevmodule = ''
         self.ui.debug("reparent to %s\n" % svnurl)
         svn.ra.reparent(self.ra, svnurl)
         self.prevmodule = module
         return prevmodule
 
     def expandpaths(self, rev, paths, parents):
         changed, removed = set(), set()
         copies = {}
 
         new_module, revnum = self.revsplit(rev)[1:]
         if new_module != self.module:
             self.module = new_module
             self.reparent(self.module)
 
         for i, (path, ent) in enumerate(paths):
             self.ui.progress(_('scanning paths'), i, item=path,
                              total=len(paths))
             entrypath = self.getrelpath(path)
 
             kind = self._checkpath(entrypath, revnum)
             if kind == svn.core.svn_node_file:
                 changed.add(self.recode(entrypath))
                 if not ent.copyfrom_path or not parents:
                     continue
                 # Copy sources not in parent revisions cannot be
                 # represented, ignore their origin for now
                 pmodule, prevnum = self.revsplit(parents[0])[1:]
                 if ent.copyfrom_rev < prevnum:
                     continue
                 copyfrom_path = self.getrelpath(ent.copyfrom_path, pmodule)
                 if not copyfrom_path:
                     continue
                 self.ui.debug("copied to %s from %s@%s\n" %
                               (entrypath, copyfrom_path, ent.copyfrom_rev))
                 copies[self.recode(entrypath)] = self.recode(copyfrom_path)
             elif kind == 0: # gone, but had better be a deleted *file*
                 self.ui.debug("gone from %s\n" % ent.copyfrom_rev)
                 pmodule, prevnum = self.revsplit(parents[0])[1:]
                 parentpath = pmodule + "/" + entrypath
                 fromkind = self._checkpath(entrypath, prevnum, pmodule)
 
                 if fromkind == svn.core.svn_node_file:
                     removed.add(self.recode(entrypath))
                 elif fromkind == svn.core.svn_node_dir:
                     oroot = parentpath.strip('/')
                     nroot = path.strip('/')
                     children = self._iterfiles(oroot, prevnum)
                     for childpath in children:
                         childpath = childpath.replace(oroot, nroot)
                         childpath = self.getrelpath("/" + childpath, pmodule)
                         if childpath:
                             removed.add(self.recode(childpath))
                 else:
                     self.ui.debug('unknown path in revision %d: %s\n' % \
                                   (revnum, path))
-            elif kind == svn.core.svn_node_dir:                
+            elif kind == svn.core.svn_node_dir:
                 if ent.action == 'M':
                     # If the directory just had a prop change,
                     # then we shouldn't need to look for its children.
                     continue
                 elif ent.action == 'R' and parents:
                     # If a directory is replacing a file, mark the previous
                     # file as deleted
                     pmodule, prevnum = self.revsplit(parents[0])[1:]
                     pkind = self._checkpath(entrypath, prevnum, pmodule)
                     if pkind == svn.core.svn_node_file:
                         removed.add(self.recode(entrypath))
 
                 for childpath in self._iterfiles(path, revnum):
                     childpath = self.getrelpath("/" + childpath)
                     if childpath:
                         changed.add(self.recode(childpath))
 
                 # Handle directory copies
                 if not ent.copyfrom_path or not parents:
                     continue
                 # Copy sources not in parent revisions cannot be
                 # represented, ignore their origin for now
                 pmodule, prevnum = self.revsplit(parents[0])[1:]
                 if ent.copyfrom_rev < prevnum:
                     continue
                 copyfrompath = self.getrelpath(ent.copyfrom_path, pmodule)
                 if not copyfrompath:
                     continue
                 self.ui.debug("mark %s came from %s:%d\n"
                               % (path, copyfrompath, ent.copyfrom_rev))
                 children = self._iterfiles(ent.copyfrom_path, ent.copyfrom_rev)
                 for childpath in children:
                     childpath = self.getrelpath("/" + childpath, pmodule)
                     if not childpath:
                         continue
                     copytopath = path + childpath[len(copyfrompath):]
                     copytopath = self.getrelpath(copytopath)
                     copies[self.recode(copytopath)] = self.recode(childpath)
 
         self.ui.progress(_('scanning paths'), None)
         changed.update(removed)
         return (list(changed), removed, copies)
 
     def _fetch_revisions(self, from_revnum, to_revnum):
         if from_revnum < to_revnum:
             from_revnum, to_revnum = to_revnum, from_revnum
 
         self.child_cset = None
 
         def parselogentry(orig_paths, revnum, author, date, message):
             """Return the parsed commit object or None, and True if
             the revision is a branch root.
             """
             self.ui.debug("parsing revision %d (%d changes)\n" %
                           (revnum, len(orig_paths)))
 
             branched = False
             rev = self.revid(revnum)
             # branch log might return entries for a parent we already have
 
             if rev in self.commits or revnum < to_revnum:
                 return None, branched
 
             parents = []
             # check whether this revision is the start of a branch or part
             # of a branch renaming
             orig_paths = sorted(orig_paths.iteritems())
             root_paths = [(p, e) for p, e in orig_paths
                           if self.module.startswith(p)]
             if root_paths:
                 path, ent = root_paths[-1]
                 if ent.copyfrom_path:
                     branched = True
                     newpath = ent.copyfrom_path + self.module[len(path):]
                     # ent.copyfrom_rev may not be the actual last revision
                     previd = self.latest(newpath, ent.copyfrom_rev)
                     if previd is not None:
                         prevmodule, prevnum = self.revsplit(previd)[1:]
                         if prevnum >= self.startrev:
                             parents = [previd]
                             self.ui.note(
                                 _('found parent of branch %s at %d: %s\n') %
                                 (self.module, prevnum, prevmodule))
                 else:
                     self.ui.debug("no copyfrom path, don't know what to do.\n")
 
             paths = []
             # filter out unrelated paths
             for path, ent in orig_paths:
                 if self.getrelpath(path) is None:
                     continue
                 paths.append((path, ent))
 
             # Example SVN datetime. Includes microseconds.
             # ISO-8601 conformant
             # '2007-01-04T17:35:00.902377Z'
             date = util.parsedate(date[:19] + " UTC", ["%Y-%m-%dT%H:%M:%S"])
 
             log = message and self.recode(message) or ''
             author = author and self.recode(author) or ''
             try:
                 branch = self.module.split("/")[-1]
                 if branch == 'trunk':
                     branch = ''
             except IndexError:
                 branch = None
 
             cset = commit(author=author,
                           date=util.datestr(date),
                           desc=log,
                           parents=parents,
                           branch=branch,
                           rev=rev)
 
             self.commits[rev] = cset
             # The parents list is *shared* among self.paths and the
             # commit object. Both will be updated below.
             self.paths[rev] = (paths, cset.parents)
             if self.child_cset and not self.child_cset.parents:
                 self.child_cset.parents[:] = [rev]
             self.child_cset = cset
             return cset, branched
 
         self.ui.note(_('fetching revision log for "%s" from %d to %d\n') %
                      (self.module, from_revnum, to_revnum))
 
         try:
             firstcset = None
             lastonbranch = False
             stream = self._getlog([self.module], from_revnum, to_revnum)
             try:
                 for entry in stream:
                     paths, revnum, author, date, message = entry
                     if revnum < self.startrev:
                         lastonbranch = True
                         break
                     if not paths:
                         self.ui.debug('revision %d has no entries\n' % revnum)
                         # If we ever leave the loop on an empty
                         # revision, do not try to get a parent branch
                         lastonbranch = lastonbranch or revnum == 0
                         continue
                     cset, lastonbranch = parselogentry(paths, revnum, author,
                                                        date, message)
                     if cset:
                         firstcset = cset
                     if lastonbranch:
                         break
             finally:
                 stream.close()
 
             if not lastonbranch and firstcset and not firstcset.parents:
                 # The first revision of the sequence (the last fetched one)
                 # has invalid parents if not a branch root. Find the parent
                 # revision now, if any.
                 try:
                     firstrevnum = self.revnum(firstcset.rev)
                     if firstrevnum > 1:
                         latest = self.latest(self.module, firstrevnum - 1)
                         if latest:
                             firstcset.parents.append(latest)
                 except SvnPathNotFound:
                     pass
         except SubversionException, (inst, num):
             if num == svn.core.SVN_ERR_FS_NO_SUCH_REVISION:
                 raise util.Abort(_('svn: branch has no revision %s') % to_revnum)
             raise
 
     def getfile(self, file, rev):
         # TODO: ra.get_file transmits the whole file instead of diffs.
         if file in self.removed:
-            raise IOError()         
+            raise IOError()
         mode = ''
         try:
             new_module, revnum = self.revsplit(rev)[1:]
             if self.module != new_module:
                 self.module = new_module
                 self.reparent(self.module)
             io = StringIO()
             info = svn.ra.get_file(self.ra, file, revnum, io)
             data = io.getvalue()
             # ra.get_files() seems to keep a reference on the input buffer
             # preventing collection. Release it explicitely.
             io.close()
             if isinstance(info, list):
                 info = info[-1]
             mode = ("svn:executable" in info) and 'x' or ''
             mode = ("svn:special" in info) and 'l' or mode
         except SubversionException, e:
             notfound = (svn.core.SVN_ERR_FS_NOT_FOUND,
                 svn.core.SVN_ERR_RA_DAV_PATH_NOT_FOUND)
             if e.apr_err in notfound: # File not found
                 raise IOError()
             raise
         if mode == 'l':
             link_prefix = "link "
             if data.startswith(link_prefix):
                 data = data[len(link_prefix):]
         return data, mode
 
     def _iterfiles(self, path, revnum):
         """Enumerate all files in path at revnum, recursively."""
         path = path.strip('/')
         pool = Pool()
         rpath = '/'.join([self.baseurl, urllib.quote(path)]).strip('/')
         entries = svn.client.ls(rpath, optrev(revnum), True, self.ctx, pool)
         return ((path + '/' + p) for p, e in entries.iteritems()
                 if e.kind == svn.core.svn_node_file)
 
     def getrelpath(self, path, module=None):
         if module is None:
             module = self.module
         # Given the repository url of this wc, say
         #   "http://server/plone/CMFPlone/branches/Plone-2_0-branch"
         # extract the "entry" portion (a relative path) from what
         # svn log --xml says, ie
         #   "/CMFPlone/branches/Plone-2_0-branch/tests/PloneTestCase.py"
         # that is to say "tests/PloneTestCase.py"
         if path.startswith(module):
             relative = path.rstrip('/')[len(module):]
             if relative.startswith('/'):
                 return relative[1:]
             elif relative == '':
                 return relative
 
         # The path is outside our tracked tree...
         self.ui.debug('%r is not under %r, ignoring\n' % (path, module))
         return None
 
     def _checkpath(self, path, revnum, module=None):
         if module is not None:
             prevmodule = self.reparent('')
             path = module + '/' + path
         try:
             # ra.check_path does not like leading slashes very much, it leads
             # to PROPFIND subversion errors
             return svn.ra.check_path(self.ra, path.strip('/'), revnum)
         finally:
             if module is not None:
                 self.reparent(prevmodule)
-    
+
     def _getlog(self, paths, start, end, limit=0, discover_changed_paths=True,
                 strict_node_history=False):
         # Normalize path names, svn >= 1.5 only wants paths relative to
         # supplied URL
         relpaths = []
         for p in paths:
             if not p.startswith('/'):
                 p = self.module + '/' + p
             relpaths.append(p.strip('/'))
         args = [self.baseurl, relpaths, start, end, limit, discover_changed_paths,
                 strict_node_history]
         arg = encodeargs(args)
         hgexe = util.hgexecutable()
         cmd = '%s debugsvnlog' % util.shellquote(hgexe)
         stdin, stdout = util.popen2(cmd)
         stdin.write(arg)
         try:
             stdin.close()
         except IOError:
             raise util.Abort(_('Mercurial failed to run itself, check'
                                ' hg executable is in PATH'))
         return logstream(stdout)
 
 pre_revprop_change = '''#!/bin/sh
 
 REPOS="$1"
 REV="$2"
 USER="$3"
 PROPNAME="$4"
 ACTION="$5"
 
 if [ "$ACTION" = "M" -a "$PROPNAME" = "svn:log" ]; then exit 0; fi
 if [ "$ACTION" = "A" -a "$PROPNAME" = "hg:convert-branch" ]; then exit 0; fi
 if [ "$ACTION" = "A" -a "$PROPNAME" = "hg:convert-rev" ]; then exit 0; fi
 
 echo "Changing prohibited revision property" >&2
 exit 1
 '''
 
 class svn_sink(converter_sink, commandline):
     commit_re = re.compile(r'Committed revision (\d+).', re.M)
 
     def prerun(self):
         if self.wc:
             os.chdir(self.wc)
 
     def postrun(self):
         if self.wc:
             os.chdir(self.cwd)
 
     def join(self, name):
         return os.path.join(self.wc, '.svn', name)
 
     def revmapfile(self):
         return self.join('hg-shamap')
 
     def authorfile(self):
         return self.join('hg-authormap')
 
     def __init__(self, ui, path):
         converter_sink.__init__(self, ui, path)
         commandline.__init__(self, ui, 'svn')
         self.delete = []
         self.setexec = []
         self.delexec = []
         self.copies = []
         self.wc = None
         self.cwd = os.getcwd()
 
         path = os.path.realpath(path)
 
         created = False
         if os.path.isfile(os.path.join(path, '.svn', 'entries')):
             self.wc = path
             self.run0('update')
         else:
             wcpath = os.path.join(os.getcwd(), os.path.basename(path) + '-wc')
 
             if os.path.isdir(os.path.dirname(path)):
                 if not os.path.exists(os.path.join(path, 'db', 'fs-type')):
                     ui.status(_('initializing svn repository %r\n') %
                               os.path.basename(path))
                     commandline(ui, 'svnadmin').run0('create', path)
                     created = path
                 path = util.normpath(path)
                 if not path.startswith('/'):
                     path = '/' + path
                 path = 'file://' + path
 
             ui.status(_('initializing svn working copy %r\n')
                       % os.path.basename(wcpath))
             self.run0('checkout', path, wcpath)
 
             self.wc = wcpath
         self.opener = util.opener(self.wc)
         self.wopener = util.opener(self.wc)
         self.childmap = mapfile(ui, self.join('hg-childmap'))
         self.is_exec = util.checkexec(self.wc) and util.is_exec or None
 
         if created:
             hook = os.path.join(created, 'hooks', 'pre-revprop-change')
             fp = open(hook, 'w')
             fp.write(pre_revprop_change)
             fp.close()
             util.set_flags(hook, False, True)
 
         xport = transport.SvnRaTransport(url=geturl(path))
         self.uuid = svn.ra.get_uuid(xport.ra)
 
     def wjoin(self, *names):
         return os.path.join(self.wc, *names)
 
     def putfile(self, filename, flags, data):
         if 'l' in flags:
             self.wopener.symlink(data, filename)
         else:
             try:
                 if os.path.islink(self.wjoin(filename)):
                     os.unlink(filename)
             except OSError:
                 pass
             self.wopener(filename, 'w').write(data)
 
             if self.is_exec:
                 was_exec = self.is_exec(self.wjoin(filename))
             else:
                 # On filesystems not supporting execute-bit, there is no way
                 # to know if it is set but asking subversion. Setting it
                 # systematically is just as expensive and much simpler.
                 was_exec = 'x' not in flags
 
             util.set_flags(self.wjoin(filename), False, 'x' in flags)
             if was_exec:
                 if 'x' not in flags:
                     self.delexec.append(filename)
             else:
                 if 'x' in flags:
                     self.setexec.append(filename)
 
     def _copyfile(self, source, dest):
         # SVN's copy command pukes if the destination file exists, but
         # our copyfile method expects to record a copy that has
         # already occurred.  Cross the semantic gap.
         wdest = self.wjoin(dest)
         exists = os.path.lexists(wdest)
         if exists:
             fd, tempname = tempfile.mkstemp(
                 prefix='hg-copy-', dir=os.path.dirname(wdest))
             os.close(fd)
             os.unlink(tempname)
             os.rename(wdest, tempname)
         try:
             self.run0('copy', source, dest)
         finally:
             if exists:
                 try:
                     os.unlink(wdest)
                 except OSError:
                     pass
                 os.rename(tempname, wdest)
 
     def dirs_of(self, files):
         dirs = set()
         for f in files:
             if os.path.isdir(self.wjoin(f)):
                 dirs.add(f)
             for i in strutil.rfindall(f, '/'):
                 dirs.add(f[:i])
         return dirs
 
     def add_dirs(self, files):
         add_dirs = [d for d in sorted(self.dirs_of(files))
                     if not os.path.exists(self.wjoin(d, '.svn', 'entries'))]
         if add_dirs:
             self.xargs(add_dirs, 'add', non_recursive=True, quiet=True)
         return add_dirs
 
     def add_files(self, files):
         if files:
             self.xargs(files, 'add', quiet=True)
         return files
 
     def tidy_dirs(self, names):
         deleted = []
         for d in sorted(self.dirs_of(names), reverse=True):
             wd = self.wjoin(d)
             if os.listdir(wd) == '.svn':
                 self.run0('delete', d)
                 deleted.append(d)
         return deleted
 
     def addchild(self, parent, child):
         self.childmap[parent] = child
 
     def revid(self, rev):
         return u"svn:%s@%s" % (self.uuid, rev)
 
     def putcommit(self, files, copies, parents, commit, source, revmap):
         # Apply changes to working copy
         for f, v in files:
             try:
                 data, mode = source.getfile(f, v)
             except IOError:
                 self.delete.append(f)
             else:
                 self.putfile(f, mode, data)
                 if f in copies:
                     self.copies.append([copies[f], f])
         files = [f[0] for f in files]
 
         for parent in parents:
             try:
                 return self.revid(self.childmap[parent])
             except KeyError:
                 pass
         entries = set(self.delete)
         files = frozenset(files)
         entries.update(self.add_dirs(files.difference(entries)))
         if self.copies:
             for s, d in self.copies:
                 self._copyfile(s, d)
             self.copies = []
         if self.delete:
             self.xargs(self.delete, 'delete')
             self.delete = []
         entries.update(self.add_files(files.difference(entries)))
         entries.update(self.tidy_dirs(entries))
         if self.delexec:
             self.xargs(self.delexec, 'propdel', 'svn:executable')
             self.delexec = []
         if self.setexec:
             self.xargs(self.setexec, 'propset', 'svn:executable', '*')
             self.setexec = []
 
         fd, messagefile = tempfile.mkstemp(prefix='hg-convert-')
         fp = os.fdopen(fd, 'w')
         fp.write(commit.desc)
         fp.close()
         try:
             output = self.run0('commit',
                                username=util.shortuser(commit.author),
                                file=messagefile,
                                encoding='utf-8')
             try:
                 rev = self.commit_re.search(output).group(1)
             except AttributeError:
                 if not files:
                     return parents[0]
                 self.ui.warn(_('unexpected svn output:\n'))
                 self.ui.warn(output)
                 raise util.Abort(_('unable to cope with svn output'))
             if commit.rev:
                 self.run('propset', 'hg:convert-rev', commit.rev,
                          revprop=True, revision=rev)
             if commit.branch and commit.branch != 'default':
                 self.run('propset', 'hg:convert-branch', commit.branch,
                          revprop=True, revision=rev)
             for parent in parents:
                 self.addchild(parent, rev)
             return self.revid(rev)
         finally:
             os.unlink(messagefile)
 
     def puttags(self, tags):
         self.ui.warn(_('writing Subversion tags is not yet implemented\n'))
         return None, None

mercurial/dispatch.py

@@ -1,642 +1,642
 # dispatch.py - command dispatching for mercurial
 #
 # Copyright 2005-2007 Matt Mackall <[e-mail]>
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 from i18n import _
 import os, sys, atexit, signal, pdb, socket, errno, shlex, time, traceback, re
 import util, commands, hg, fancyopts, extensions, hook, error
 import cmdutil, encoding
 import ui as uimod
 
 def run():
     "run the command in sys.argv"
     sys.exit(dispatch(sys.argv[1:]))
 
 def dispatch(args):
     "run the command specified in args"
     try:
         u = uimod.ui()
         if '--traceback' in args:
             u.setconfig('ui', 'traceback', 'on')
     except util.Abort, inst:
         sys.stderr.write(_("abort: %s\n") % inst)
         if inst.hint:
             sys.stderr.write("(%s)\n" % inst.hint)
         return -1
     except error.ParseError, inst:
         if len(inst.args) > 1:
             sys.stderr.write(_("hg: parse error at %s: %s\n") %
                              (inst.args[1], inst.args[0]))
         else:
             sys.stderr.write(_("hg: parse error: %s\n") % inst.args[0])
         return -1
     return _runcatch(u, args)
 
 def _runcatch(ui, args):
     def catchterm(*args):
         raise error.SignalInterrupt
 
     try:
         for name in 'SIGBREAK', 'SIGHUP', 'SIGTERM':
             num = getattr(signal, name, None)
             if num:
                 signal.signal(num, catchterm)
     except ValueError:
         pass # happens if called in a thread
 
     try:
         try:
             # enter the debugger before command execution
             if '--debugger' in args:
                 ui.warn(_("entering debugger - "
                         "type c to continue starting hg or h for help\n"))
                 pdb.set_trace()
             try:
                 return _dispatch(ui, args)
             finally:
                 ui.flush()
         except:
             # enter the debugger when we hit an exception
             if '--debugger' in args:
                 traceback.print_exc()
                 pdb.post_mortem(sys.exc_info()[2])
             ui.traceback()
             raise
 
     # Global exception handling, alphabetically
     # Mercurial-specific first, followed by built-in and library exceptions
     except error.AmbiguousCommand, inst:
         ui.warn(_("hg: command '%s' is ambiguous:\n    %s\n") %
                 (inst.args[0], " ".join(inst.args[1])))
     except error.ParseError, inst:
         if len(inst.args) > 1:
             ui.warn(_("hg: parse error at %s: %s\n") %
                              (inst.args[1], inst.args[0]))
         else:
             ui.warn(_("hg: parse error: %s\n") % inst.args[0])
         return -1
     except error.LockHeld, inst:
         if inst.errno == errno.ETIMEDOUT:
             reason = _('timed out waiting for lock held by %s') % inst.locker
         else:
             reason = _('lock held by %s') % inst.locker
         ui.warn(_("abort: %s: %s\n") % (inst.desc or inst.filename, reason))
     except error.LockUnavailable, inst:
         ui.warn(_("abort: could not lock %s: %s\n") %
                (inst.desc or inst.filename, inst.strerror))
     except error.CommandError, inst:
         if inst.args[0]:
             ui.warn(_("hg %s: %s\n") % (inst.args[0], inst.args[1]))
             commands.help_(ui, inst.args[0])
         else:
             ui.warn(_("hg: %s\n") % inst.args[1])
             commands.help_(ui, 'shortlist')
     except error.RepoError, inst:
         ui.warn(_("abort: %s!\n") % inst)
     except error.ResponseError, inst:
         ui.warn(_("abort: %s") % inst.args[0])
         if not isinstance(inst.args[1], basestring):
             ui.warn(" %r\n" % (inst.args[1],))
         elif not inst.args[1]:
             ui.warn(_(" empty string\n"))
         else:
             ui.warn("\n%r\n" % util.ellipsis(inst.args[1]))
     except error.RevlogError, inst:
         ui.warn(_("abort: %s!\n") % inst)
     except error.SignalInterrupt:
         ui.warn(_("killed!\n"))
     except error.UnknownCommand, inst:
         ui.warn(_("hg: unknown command '%s'\n") % inst.args[0])
         try:
             # check if the command is in a disabled extension
             # (but don't check for extensions themselves)
             commands.help_(ui, inst.args[0], unknowncmd=True)
         except error.UnknownCommand:
             commands.help_(ui, 'shortlist')
     except util.Abort, inst:
         ui.warn(_("abort: %s\n") % inst)
         if inst.hint:
             ui.warn(_("(%s)\n") % inst.hint)
     except ImportError, inst:
         ui.warn(_("abort: %s!\n") % inst)
         m = str(inst).split()[-1]
         if m in "mpatch bdiff".split():
             ui.warn(_("(did you forget to compile extensions?)\n"))
         elif m in "zlib".split():
             ui.warn(_("(is your Python install correct?)\n"))
     except IOError, inst:
         if hasattr(inst, "code"):
             ui.warn(_("abort: %s\n") % inst)
         elif hasattr(inst, "reason"):
             try: # usually it is in the form (errno, strerror)
                 reason = inst.reason.args[1]
             except: # it might be anything, for example a string
                 reason = inst.reason
             ui.warn(_("abort: error: %s\n") % reason)
         elif hasattr(inst, "args") and inst.args[0] == errno.EPIPE:
             if ui.debugflag:
                 ui.warn(_("broken pipe\n"))
         elif getattr(inst, "strerror", None):
             if getattr(inst, "filename", None):
                 ui.warn(_("abort: %s: %s\n") % (inst.strerror, inst.filename))
             else:
                 ui.warn(_("abort: %s\n") % inst.strerror)
         else:
             raise
     except OSError, inst:
         if getattr(inst, "filename", None):
             ui.warn(_("abort: %s: %s\n") % (inst.strerror, inst.filename))
         else:
             ui.warn(_("abort: %s\n") % inst.strerror)
     except KeyboardInterrupt:
         try:
             ui.warn(_("interrupted!\n"))
         except IOError, inst:
             if inst.errno == errno.EPIPE:
                 if ui.debugflag:
                     ui.warn(_("\nbroken pipe\n"))
             else:
                 raise
     except MemoryError:
         ui.warn(_("abort: out of memory\n"))
     except SystemExit, inst:
         # Commands shouldn't sys.exit directly, but give a return code.
         # Just in case catch this and and pass exit code to caller.
         return inst.code
     except socket.error, inst:
         ui.warn(_("abort: %s\n") % inst.args[-1])
     except:
         ui.warn(_("** unknown exception encountered, details follow\n"))
         ui.warn(_("** report bug details to "
                  "http://mercurial.selenic.com/bts/\n"))
         ui.warn(_("** or [e-mail]\n"))
         ui.warn(_("** Python %s\n") % sys.version.replace('\n', ''))
         ui.warn(_("** Mercurial Distributed SCM (version %s)\n")
                % util.version())
         ui.warn(_("** Extensions loaded: %s\n")
                % ", ".join([x[0] for x in extensions.extensions()]))
         raise
 
     return -1
 
 def aliasargs(fn):
     if hasattr(fn, 'args'):
         return fn.args
     return []
 
 class cmdalias(object):
     def __init__(self, name, definition, cmdtable):
         self.name = self.cmd = name
         self.cmdname = ''
         self.definition = definition
         self.args = []
         self.opts = []
         self.help = ''
         self.norepo = True
         self.badalias = False
 
         try:
             aliases, entry = cmdutil.findcmd(self.name, cmdtable)
             for alias, e in cmdtable.iteritems():
                 if e is entry:
                     self.cmd = alias
                     break
             self.shadows = True
         except error.UnknownCommand:
             self.shadows = False
 
         if not self.definition:
             def fn(ui, *args):
                 ui.warn(_("no definition for alias '%s'\n") % self.name)
                 return 1
             self.fn = fn
             self.badalias = True
 
             return
 
         if self.definition.startswith('!'):
             self.shell = True
             def fn(ui, *args):
                 env = {'HG_ARGS': ' '.join((self.name,) + args)}
                 def _checkvar(m):
                     if int(m.groups()[0]) <= len(args):
                         return m.group()
                     else:
                         return ''
                 cmd = re.sub(r'\$(\d+)', _checkvar, self.definition[1:])
                 replace = dict((str(i + 1), arg) for i, arg in enumerate(args))
                 replace['0'] = self.name
                 replace['@'] = ' '.join(args)
                 cmd = util.interpolate(r'\$', replace, cmd)
                 return util.system(cmd, environ=env)
             self.fn = fn
             return
 
         args = shlex.split(self.definition)
         self.cmdname = cmd = args.pop(0)
         args = map(util.expandpath, args)
 
         for invalidarg in ("--cwd", "-R", "--repository", "--repo"):
             if _earlygetopt([invalidarg], args):
                 def fn(ui, *args):
                     ui.warn(_("error in definition for alias '%s': %s may only "
                               "be given on the command line\n")
                             % (self.name, invalidarg))
                     return 1
 
                 self.fn = fn
                 self.badalias = True
                 return
 
         try:
             tableentry = cmdutil.findcmd(cmd, cmdtable, False)[1]
             if len(tableentry) > 2:
                 self.fn, self.opts, self.help = tableentry
             else:
                 self.fn, self.opts = tableentry
 
             self.args = aliasargs(self.fn) + args
             if cmd not in commands.norepo.split(' '):
                 self.norepo = False
             if self.help.startswith("hg " + cmd):
                 # drop prefix in old-style help lines so hg shows the alias
                 self.help = self.help[4 + len(cmd):]
             self.__doc__ = self.fn.__doc__
 
         except error.UnknownCommand:
             def fn(ui, *args):
                 ui.warn(_("alias '%s' resolves to unknown command '%s'\n") \
                             % (self.name, cmd))
                 try:
                     # check if the command is in a disabled extension
                     commands.help_(ui, cmd, unknowncmd=True)
                 except error.UnknownCommand:
                     pass
                 return 1
             self.fn = fn
             self.badalias = True
         except error.AmbiguousCommand:
             def fn(ui, *args):
                 ui.warn(_("alias '%s' resolves to ambiguous command '%s'\n") \
                             % (self.name, cmd))
                 return 1
             self.fn = fn
             self.badalias = True
 
     def __call__(self, ui, *args, **opts):
         if self.shadows:
             ui.debug("alias '%s' shadows command '%s'\n" %
                      (self.name, self.cmdname))
 
         if self.definition.startswith('!'):
             return self.fn(ui, *args, **opts)
         else:
             try:
                 util.checksignature(self.fn)(ui, *args, **opts)
             except error.SignatureError:
                 args = ' '.join([self.cmdname] + self.args)
                 ui.debug("alias '%s' expands to '%s'\n" % (self.name, args))
                 raise
 
 def addaliases(ui, cmdtable):
     # aliases are processed after extensions have been loaded, so they
     # may use extension commands. Aliases can also use other alias definitions,
     # but only if they have been defined prior to the current definition.
     for alias, definition in ui.configitems('alias'):
         aliasdef = cmdalias(alias, definition, cmdtable)
         cmdtable[aliasdef.cmd] = (aliasdef, aliasdef.opts, aliasdef.help)
         if aliasdef.norepo:
             commands.norepo += ' %s' % alias
 
 def _parse(ui, args):
     options = {}
     cmdoptions = {}
 
     try:
         args = fancyopts.fancyopts(args, commands.globalopts, options)
     except fancyopts.getopt.GetoptError, inst:
         raise error.CommandError(None, inst)
 
     if args:
         cmd, args = args[0], args[1:]
         aliases, entry = cmdutil.findcmd(cmd, commands.table,
                                      ui.config("ui", "strict"))
         cmd = aliases[0]
         args = aliasargs(entry[0]) + args
         defaults = ui.config("defaults", cmd)
         if defaults:
             args = map(util.expandpath, shlex.split(defaults)) + args
         c = list(entry[1])
     else:
         cmd = None
         c = []
 
     # combine global options into local
     for o in commands.globalopts:
         c.append((o[0], o[1], options[o[1]], o[3]))
 
     try:
         args = fancyopts.fancyopts(args, c, cmdoptions, True)
     except fancyopts.getopt.GetoptError, inst:
         raise error.CommandError(cmd, inst)
 
     # separate global options back out
     for o in commands.globalopts:
         n = o[1]
         options[n] = cmdoptions[n]
         del cmdoptions[n]
 
     return (cmd, cmd and entry[0] or None, args, options, cmdoptions)
 
 def _parseconfig(ui, config):
     """parse the --config options from the command line"""
     for cfg in config:
         try:
             name, value = cfg.split('=', 1)
             section, name = name.split('.', 1)
             if not section or not name:
                 raise IndexError
             ui.setconfig(section, name, value)
         except (IndexError, ValueError):
             raise util.Abort(_('malformed --config option: %r '
                                '(use --config section.name=value)') % cfg)
 
 def _earlygetopt(aliases, args):
     """Return list of values for an option (or aliases).
 
     The values are listed in the order they appear in args.
     The options and values are removed from args.
     """
     try:
         argcount = args.index("--")
     except ValueError:
         argcount = len(args)
     shortopts = [opt for opt in aliases if len(opt) == 2]
     values = []
     pos = 0
     while pos < argcount:
         if args[pos] in aliases:
             if pos + 1 >= argcount:
                 # ignore and let getopt report an error if there is no value
                 break
             del args[pos]
             values.append(args.pop(pos))
             argcount -= 2
         elif args[pos][:2] in shortopts:
             # short option can have no following space, e.g. hg log -Rfoo
             values.append(args.pop(pos)[2:])
             argcount -= 1
         else:
             pos += 1
     return values
 
 def runcommand(lui, repo, cmd, fullargs, ui, options, d, cmdpats, cmdoptions):
     # run pre-hook, and abort if it fails
     ret = hook.hook(lui, repo, "pre-%s" % cmd, False, args=" ".join(fullargs),
                     pats=cmdpats, opts=cmdoptions)
     if ret:
         return ret
     ret = _runcommand(ui, options, cmd, d)
     # run post-hook, passing command result
     hook.hook(lui, repo, "post-%s" % cmd, False, args=" ".join(fullargs),
               result=ret, pats=cmdpats, opts=cmdoptions)
     return ret
 
 def _getlocal(ui, rpath):
     """Return (path, local ui object) for the given target path.
-    
+
     Takes paths in [cwd]/.hg/hgrc into account."
     """
     try:
         wd = os.getcwd()
     except OSError, e:
         raise util.Abort(_("error getting current working directory: %s") %
                          e.strerror)
     path = cmdutil.findrepo(wd) or ""
     if not path:
         lui = ui
     else:
         lui = ui.copy()
         lui.readconfig(os.path.join(path, ".hg", "hgrc"), path)
 
     if rpath:
         path = lui.expandpath(rpath[-1])
         lui = ui.copy()
         lui.readconfig(os.path.join(path, ".hg", "hgrc"), path)
 
     return path, lui
 
 def _checkshellalias(ui, args):
     cwd = os.getcwd()
     norepo = commands.norepo
     options = {}
 
     try:
         args = fancyopts.fancyopts(args, commands.globalopts, options)
     except fancyopts.getopt.GetoptError:
         return
 
     if not args:
         return
 
     _parseconfig(ui, options['config'])
     if options['cwd']:
         os.chdir(options['cwd'])
 
     path, lui = _getlocal(ui, [options['repository']])
 
     cmdtable = commands.table.copy()
     addaliases(lui, cmdtable)
 
     cmd = args[0]
     try:
         aliases, entry = cmdutil.findcmd(cmd, cmdtable, lui.config("ui", "strict"))
     except error.UnknownCommand:
         commands.norepo = norepo
         os.chdir(cwd)
         return
 
     cmd = aliases[0]
     fn = entry[0]
 
     if cmd and hasattr(fn, 'shell'):
         d = lambda: fn(ui, *args[1:])
         return lambda: runcommand(lui, None, cmd, args[:1], ui, options, d, [], {})
 
     commands.norepo = norepo
     os.chdir(cwd)
 
 _loaded = set()
 def _dispatch(ui, args):
     shellaliasfn = _checkshellalias(ui, args)
     if shellaliasfn:
         return shellaliasfn()
 
     # read --config before doing anything else
     # (e.g. to change trust settings for reading .hg/hgrc)
     _parseconfig(ui, _earlygetopt(['--config'], args))
 
     # check for cwd
     cwd = _earlygetopt(['--cwd'], args)
     if cwd:
         os.chdir(cwd[-1])
 
     rpath = _earlygetopt(["-R", "--repository", "--repo"], args)
     path, lui = _getlocal(ui, rpath)
 
     # Configure extensions in phases: uisetup, extsetup, cmdtable, and
     # reposetup. Programs like TortoiseHg will call _dispatch several
     # times so we keep track of configured extensions in _loaded.
     extensions.loadall(lui)
     exts = [ext for ext in extensions.extensions() if ext[0] not in _loaded]
     # Propagate any changes to lui.__class__ by extensions
     ui.__class__ = lui.__class__
 
     # (uisetup and extsetup are handled in extensions.loadall)
 
     for name, module in exts:
         cmdtable = getattr(module, 'cmdtable', {})
         overrides = [cmd for cmd in cmdtable if cmd in commands.table]
         if overrides:
             ui.warn(_("extension '%s' overrides commands: %s\n")
                     % (name, " ".join(overrides)))
         commands.table.update(cmdtable)
         _loaded.add(name)
 
     # (reposetup is handled in hg.repository)
 
     addaliases(lui, commands.table)
 
     # check for fallback encoding
     fallback = lui.config('ui', 'fallbackencoding')
     if fallback:
         encoding.fallbackencoding = fallback
 
     fullargs = args
     cmd, func, args, options, cmdoptions = _parse(lui, args)
 
     if options["config"]:
         raise util.Abort(_("option --config may not be abbreviated!"))
     if options["cwd"]:
         raise util.Abort(_("option --cwd may not be abbreviated!"))
     if options["repository"]:
         raise util.Abort(_(
             "Option -R has to be separated from other options (e.g. not -qR) "
             "and --repository may only be abbreviated as --repo!"))
 
     if options["encoding"]:
         encoding.encoding = options["encoding"]
     if options["encodingmode"]:
         encoding.encodingmode = options["encodingmode"]
     if options["time"]:
         def get_times():
             t = os.times()
             if t[4] == 0.0: # Windows leaves this as zero, so use time.clock()
                 t = (t[0], t[1], t[2], t[3], time.clock())
             return t
         s = get_times()
         def print_time():
             t = get_times()
             ui.warn(_("Time: real %.3f secs (user %.3f+%.3f sys %.3f+%.3f)\n") %
                 (t[4]-s[4], t[0]-s[0], t[2]-s[2], t[1]-s[1], t[3]-s[3]))
         atexit.register(print_time)
 
     if options['verbose'] or options['debug'] or options['quiet']:
         ui.setconfig('ui', 'verbose', str(bool(options['verbose'])))
         ui.setconfig('ui', 'debug', str(bool(options['debug'])))
         ui.setconfig('ui', 'quiet', str(bool(options['quiet'])))
     if options['traceback']:
         ui.setconfig('ui', 'traceback', 'on')
     if options['noninteractive']:
         ui.setconfig('ui', 'interactive', 'off')
 
     if options['help']:
         return commands.help_(ui, cmd, options['version'])
     elif options['version']:
         return commands.version_(ui)
     elif not cmd:
         return commands.help_(ui, 'shortlist')
 
     repo = None
     cmdpats = args[:]
     if cmd not in commands.norepo.split():
         try:
             repo = hg.repository(ui, path=path)
             ui = repo.ui
             if not repo.local():
                 raise util.Abort(_("repository '%s' is not local") % path)
             ui.setconfig("bundle", "mainreporoot", repo.root)
         except error.RepoError:
             if cmd not in commands.optionalrepo.split():
                 if args and not path: # try to infer -R from command args
                     repos = map(cmdutil.findrepo, args)
                     guess = repos[0]
                     if guess and repos.count(guess) == len(repos):
                         return _dispatch(ui, ['--repository', guess] + fullargs)
                 if not path:
                     raise error.RepoError(_("There is no Mercurial repository"
                                       " here (.hg not found)"))
                 raise
         args.insert(0, repo)
     elif rpath:
         ui.warn(_("warning: --repository ignored\n"))
 
     msg = ' '.join(' ' in a and repr(a) or a for a in fullargs)
     ui.log("command", msg + "\n")
     d = lambda: util.checksignature(func)(ui, *args, **cmdoptions)
     return runcommand(lui, repo, cmd, fullargs, ui, options, d,
                       cmdpats, cmdoptions)
 
 def _runcommand(ui, options, cmd, cmdfunc):
     def checkargs():
         try:
             return cmdfunc()
         except error.SignatureError:
             raise error.CommandError(cmd, _("invalid arguments"))
 
     if options['profile']:
         format = ui.config('profiling', 'format', default='text')
 
         if not format in ['text', 'kcachegrind']:
             ui.warn(_("unrecognized profiling format '%s'"
                         " - Ignored\n") % format)
             format = 'text'
 
         output = ui.config('profiling', 'output')
 
         if output:
             path = ui.expandpath(output)
             ostream = open(path, 'wb')
         else:
             ostream = sys.stderr
 
         try:
             from mercurial import lsprof
         except ImportError:
             raise util.Abort(_(
                 'lsprof not available - install from '
                 'http://codespeak.net/svn/user/arigo/hack/misc/lsprof/'))
         p = lsprof.Profiler()
         p.enable(subcalls=True)
         try:
             return checkargs()
         finally:
             p.disable()
 
             if format == 'kcachegrind':
                 import lsprofcalltree
                 calltree = lsprofcalltree.KCacheGrind(p)
                 calltree.output(ostream)
             else:
                 # format == 'text'
                 stats = lsprof.Stats(p.getstats())
                 stats.sort()
                 stats.pprint(top=10, file=ostream, climit=5)
 
             if output:
                 ostream.close()
     else:
         return checkargs()

mercurial/encoding.py

@@ -1,97 +1,97
 # encoding.py - character transcoding support for Mercurial
 #
 #  Copyright 2005-2009 Matt Mackall <[e-mail]> and others
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 import error
 import unicodedata, locale, os
 
 def _getpreferredencoding():
     '''
     On darwin, getpreferredencoding ignores the locale environment and
     always returns mac-roman. http://bugs.python.org/issue6202 fixes this
     for Python 2.7 and up. This is the same corrected code for earlier
     Python versions.
 
-    However, we can't use a version check for this method, as some distributions 
+    However, we can't use a version check for this method, as some distributions
     patch Python to fix this. Instead, we use it as a 'fixer' for the mac-roman
     encoding, as it is unlikely that this encoding is the actually expected.
     '''
     try:
         locale.CODESET
     except AttributeError:
         # Fall back to parsing environment variables :-(
         return locale.getdefaultlocale()[1]
 
     oldloc = locale.setlocale(locale.LC_CTYPE)
     locale.setlocale(locale.LC_CTYPE, "")
     result = locale.nl_langinfo(locale.CODESET)
     locale.setlocale(locale.LC_CTYPE, oldloc)
 
     return result
 
 _encodingfixers = {
     '646': lambda: 'ascii',
     'ANSI_X3.4-1968': lambda: 'ascii',
     'mac-roman': _getpreferredencoding
 }
 
 try:
     encoding = os.environ.get("HGENCODING")
     if not encoding:
         encoding = locale.getpreferredencoding() or 'ascii'
         encoding = _encodingfixers.get(encoding, lambda: encoding)()
 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 error.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 error.Abort("decoding near '%s': %s!" % (sub, inst))
     except LookupError, k:
         raise error.Abort("%s, please check your locale settings" % k)
 
 def colwidth(s):
     "Find the column width of a UTF-8 string for display"
     d = s.decode(encoding, 'replace')
     if hasattr(unicodedata, 'east_asian_width'):
         w = unicodedata.east_asian_width
         return sum([w(c) in 'WFA' and 2 or 1 for c in d])
     return len(d)
 

mercurial/subrepo.py

@@ -1,580 +1,580
 # subrepo.py - sub-repository handling for Mercurial
 #
 # Copyright 2009-2010 Matt Mackall <[e-mail]>
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 import errno, os, re, xml.dom.minidom, shutil, urlparse, posixpath
 from i18n import _
 import config, util, node, error, cmdutil
 hg = None
 
 nullstate = ('', '', 'empty')
 
 def state(ctx, ui):
     """return a state dict, mapping subrepo paths configured in .hgsub
     to tuple: (source from .hgsub, revision from .hgsubstate, kind
     (key in types dict))
     """
     p = config.config()
     def read(f, sections=None, remap=None):
         if f in ctx:
             p.parse(f, ctx[f].data(), sections, remap, read)
         else:
             raise util.Abort(_("subrepo spec file %s not found") % f)
 
     if '.hgsub' in ctx:
         read('.hgsub')
 
     for path, src in ui.configitems('subpaths'):
         p.set('subpaths', path, src, ui.configsource('subpaths', path))
 
     rev = {}
     if '.hgsubstate' in ctx:
         try:
             for l in ctx['.hgsubstate'].data().splitlines():
                 revision, path = l.split(" ", 1)
                 rev[path] = revision
         except IOError, err:
             if err.errno != errno.ENOENT:
                 raise
 
     state = {}
     for path, src in p[''].items():
         kind = 'hg'
         if src.startswith('['):
             if ']' not in src:
                 raise util.Abort(_('missing ] in subrepo source'))
             kind, src = src.split(']', 1)
             kind = kind[1:]
 
         for pattern, repl in p.items('subpaths'):
             # Turn r'C:\foo\bar' into r'C:\\foo\\bar' since re.sub
             # does a string decode.
             repl = repl.encode('string-escape')
             # However, we still want to allow back references to go
             # through unharmed, so we turn r'\\1' into r'\1'. Again,
             # extra escapes are needed because re.sub string decodes.
             repl = re.sub(r'\\\\([0-9]+)', r'\\\1', repl)
             try:
                 src = re.sub(pattern, repl, src, 1)
             except re.error, e:
                 raise util.Abort(_("bad subrepository pattern in %s: %s")
                                  % (p.source('subpaths', pattern), e))
 
         state[path] = (src.strip(), rev.get(path, ''), kind)
 
     return state
 
 def writestate(repo, state):
     """rewrite .hgsubstate in (outer) repo with these subrepo states"""
     repo.wwrite('.hgsubstate',
                 ''.join(['%s %s\n' % (state[s][1], s)
                          for s in sorted(state)]), '')
 
 def submerge(repo, wctx, mctx, actx):
     """delegated from merge.applyupdates: merging of .hgsubstate file
     in working context, merging context and ancestor context"""
     if mctx == actx: # backwards?
         actx = wctx.p1()
     s1 = wctx.substate
     s2 = mctx.substate
     sa = actx.substate
     sm = {}
 
     repo.ui.debug("subrepo merge %s %s %s\n" % (wctx, mctx, actx))
 
     def debug(s, msg, r=""):
         if r:
             r = "%s:%s:%s" % r
         repo.ui.debug("  subrepo %s: %s %s\n" % (s, msg, r))
 
     for s, l in s1.items():
         a = sa.get(s, nullstate)
         ld = l # local state with possible dirty flag for compares
         if wctx.sub(s).dirty():
             ld = (l[0], l[1] + "+")
         if wctx == actx: # overwrite
             a = ld
 
         if s in s2:
             r = s2[s]
             if ld == r or r == a: # no change or local is newer
                 sm[s] = l
                 continue
             elif ld == a: # other side changed
                 debug(s, "other changed, get", r)
                 wctx.sub(s).get(r)
                 sm[s] = r
             elif ld[0] != r[0]: # sources differ
                 if repo.ui.promptchoice(
                     _(' subrepository sources for %s differ\n'
                       'use (l)ocal source (%s) or (r)emote source (%s)?')
                       % (s, l[0], r[0]),
                       (_('&Local'), _('&Remote')), 0):
                     debug(s, "prompt changed, get", r)
                     wctx.sub(s).get(r)
                     sm[s] = r
             elif ld[1] == a[1]: # local side is unchanged
                 debug(s, "other side changed, get", r)
                 wctx.sub(s).get(r)
                 sm[s] = r
             else:
                 debug(s, "both sides changed, merge with", r)
                 wctx.sub(s).merge(r)
                 sm[s] = l
         elif ld == a: # remote removed, local unchanged
             debug(s, "remote removed, remove")
             wctx.sub(s).remove()
         else:
             if repo.ui.promptchoice(
                 _(' local changed subrepository %s which remote removed\n'
                   'use (c)hanged version or (d)elete?') % s,
                 (_('&Changed'), _('&Delete')), 0):
                 debug(s, "prompt remove")
                 wctx.sub(s).remove()
 
     for s, r in s2.items():
         if s in s1:
             continue
         elif s not in sa:
             debug(s, "remote added, get", r)
             mctx.sub(s).get(r)
             sm[s] = r
         elif r != sa[s]:
             if repo.ui.promptchoice(
                 _(' remote changed subrepository %s which local removed\n'
                   'use (c)hanged version or (d)elete?') % s,
                 (_('&Changed'), _('&Delete')), 0) == 0:
                 debug(s, "prompt recreate", r)
                 wctx.sub(s).get(r)
                 sm[s] = r
 
     # record merged .hgsubstate
     writestate(repo, sm)
 
 def reporelpath(repo):
     """return path to this (sub)repo as seen from outermost repo"""
     parent = repo
     while hasattr(parent, '_subparent'):
         parent = parent._subparent
     return repo.root[len(parent.root)+1:]
 
 def subrelpath(sub):
     """return path to this subrepo as seen from outermost repo"""
     if not hasattr(sub, '_repo'):
         return sub._path
     return reporelpath(sub._repo)
 
 def _abssource(repo, push=False, abort=True):
     """return pull/push path of repo - either based on parent repo .hgsub info
     or on the top repo config. Abort or return None if no source found."""
     if hasattr(repo, '_subparent'):
         source = repo._subsource
         if source.startswith('/') or '://' in source:
             return source
         parent = _abssource(repo._subparent, push, abort=False)
         if parent:
             if '://' in parent:
                 if parent[-1] == '/':
                     parent = parent[:-1]
                 r = urlparse.urlparse(parent + '/' + source)
                 r = urlparse.urlunparse((r[0], r[1],
                                          posixpath.normpath(r[2]),
                                          r[3], r[4], r[5]))
                 return r
             else: # plain file system path
                 return posixpath.normpath(os.path.join(parent, repo._subsource))
     else: # recursion reached top repo
         if push and repo.ui.config('paths', 'default-push'):
             return repo.ui.config('paths', 'default-push')
         if repo.ui.config('paths', 'default'):
             return repo.ui.config('paths', 'default')
     if abort:
-        raise util.Abort(_("default path for subrepository %s not found") % 
+        raise util.Abort(_("default path for subrepository %s not found") %
             reporelpath(repo))
 
 def itersubrepos(ctx1, ctx2):
     """find subrepos in ctx1 or ctx2"""
     # Create a (subpath, ctx) mapping where we prefer subpaths from
     # ctx1. The subpaths from ctx2 are important when the .hgsub file
     # has been modified (in ctx2) but not yet committed (in ctx1).
     subpaths = dict.fromkeys(ctx2.substate, ctx2)
     subpaths.update(dict.fromkeys(ctx1.substate, ctx1))
     for subpath, ctx in sorted(subpaths.iteritems()):
         yield subpath, ctx.sub(subpath)
 
 def subrepo(ctx, path):
     """return instance of the right subrepo class for subrepo in path"""
     # subrepo inherently violates our import layering rules
     # because it wants to make repo objects from deep inside the stack
     # so we manually delay the circular imports to not break
     # scripts that don't use our demand-loading
     global hg
     import hg as h
     hg = h
 
     util.path_auditor(ctx._repo.root)(path)
     state = ctx.substate.get(path, nullstate)
     if state[2] not in types:
         raise util.Abort(_('unknown subrepo type %s') % state[2])
     return types[state[2]](ctx, path, state[:2])
 
 # subrepo classes need to implement the following abstract class:
 
 class abstractsubrepo(object):
 
     def dirty(self):
         """returns true if the dirstate of the subrepo does not match
         current stored state
         """
         raise NotImplementedError
 
     def checknested(self, path):
         """check if path is a subrepository within this repository"""
         return False
 
     def commit(self, text, user, date):
         """commit the current changes to the subrepo with the given
         log message. Use given user and date if possible. Return the
         new state of the subrepo.
         """
         raise NotImplementedError
 
     def remove(self):
         """remove the subrepo
 
         (should verify the dirstate is not dirty first)
         """
         raise NotImplementedError
 
     def get(self, state):
         """run whatever commands are needed to put the subrepo into
         this state
         """
         raise NotImplementedError
 
     def merge(self, state):
         """merge currently-saved state with the new state."""
         raise NotImplementedError
 
     def push(self, force):
         """perform whatever action is analogous to 'hg push'
 
         This may be a no-op on some systems.
         """
         raise NotImplementedError
 
     def add(self, ui, match, dryrun, prefix):
         return []
 
     def status(self, rev2, **opts):
         return [], [], [], [], [], [], []
 
     def diff(self, diffopts, node2, match, prefix, **opts):
         pass
 
     def outgoing(self, ui, dest, opts):
         return 1
 
     def incoming(self, ui, source, opts):
         return 1
 
     def files(self):
         """return filename iterator"""
         raise NotImplementedError
 
     def filedata(self, name):
         """return file data"""
         raise NotImplementedError
 
     def fileflags(self, name):
         """return file flags"""
         return ''
 
     def archive(self, archiver, prefix):
         for name in self.files():
             flags = self.fileflags(name)
             mode = 'x' in flags and 0755 or 0644
             symlink = 'l' in flags
             archiver.addfile(os.path.join(prefix, self._path, name),
                              mode, symlink, self.filedata(name))
 
 
 class hgsubrepo(abstractsubrepo):
     def __init__(self, ctx, path, state):
         self._path = path
         self._state = state
         r = ctx._repo
         root = r.wjoin(path)
         create = False
         if not os.path.exists(os.path.join(root, '.hg')):
             create = True
             util.makedirs(root)
         self._repo = hg.repository(r.ui, root, create=create)
         self._repo._subparent = r
         self._repo._subsource = state[0]
 
         if create:
             fp = self._repo.opener("hgrc", "w", text=True)
             fp.write('[paths]\n')
 
             def addpathconfig(key, value):
                 if value:
                     fp.write('%s = %s\n' % (key, value))
                     self._repo.ui.setconfig('paths', key, value)
 
             defpath = _abssource(self._repo, abort=False)
             defpushpath = _abssource(self._repo, True, abort=False)
             addpathconfig('default', defpath)
             if defpath != defpushpath:
                 addpathconfig('default-push', defpushpath)
             fp.close()
 
     def add(self, ui, match, dryrun, prefix):
         return cmdutil.add(ui, self._repo, match, dryrun, True,
                            os.path.join(prefix, self._path))
 
     def status(self, rev2, **opts):
         try:
             rev1 = self._state[1]
             ctx1 = self._repo[rev1]
             ctx2 = self._repo[rev2]
             return self._repo.status(ctx1, ctx2, **opts)
         except error.RepoLookupError, inst:
             self._repo.ui.warn(_('warning: error "%s" in subrepository "%s"\n')
                                % (inst, subrelpath(self)))
             return [], [], [], [], [], [], []
 
     def diff(self, diffopts, node2, match, prefix, **opts):
         try:
             node1 = node.bin(self._state[1])
             # We currently expect node2 to come from substate and be
             # in hex format
             if node2 is not None:
                 node2 = node.bin(node2)
             cmdutil.diffordiffstat(self._repo.ui, self._repo, diffopts,
                                    node1, node2, match,
                                    prefix=os.path.join(prefix, self._path),
                                    listsubrepos=True, **opts)
         except error.RepoLookupError, inst:
             self._repo.ui.warn(_('warning: error "%s" in subrepository "%s"\n')
                                % (inst, subrelpath(self)))
 
     def archive(self, archiver, prefix):
         abstractsubrepo.archive(self, archiver, prefix)
 
         rev = self._state[1]
         ctx = self._repo[rev]
         for subpath in ctx.substate:
             s = subrepo(ctx, subpath)
             s.archive(archiver, os.path.join(prefix, self._path))
 
     def dirty(self):
         r = self._state[1]
         if r == '':
             return True
         w = self._repo[None]
         if w.p1() != self._repo[r]: # version checked out change
             return True
         return w.dirty() # working directory changed
 
     def checknested(self, path):
         return self._repo._checknested(self._repo.wjoin(path))
 
     def commit(self, text, user, date):
         self._repo.ui.debug("committing subrepo %s\n" % subrelpath(self))
         n = self._repo.commit(text, user, date)
         if not n:
             return self._repo['.'].hex() # different version checked out
         return node.hex(n)
 
     def remove(self):
         # we can't fully delete the repository as it may contain
         # local-only history
         self._repo.ui.note(_('removing subrepo %s\n') % subrelpath(self))
         hg.clean(self._repo, node.nullid, False)
 
     def _get(self, state):
         source, revision, kind = state
         try:
             self._repo.lookup(revision)
         except error.RepoError:
             self._repo._subsource = source
             srcurl = _abssource(self._repo)
             self._repo.ui.status(_('pulling subrepo %s from %s\n')
                                  % (subrelpath(self), srcurl))
             other = hg.repository(self._repo.ui, srcurl)
             self._repo.pull(other)
 
     def get(self, state):
         self._get(state)
         source, revision, kind = state
         self._repo.ui.debug("getting subrepo %s\n" % self._path)
         hg.clean(self._repo, revision, False)
 
     def merge(self, state):
         self._get(state)
         cur = self._repo['.']
         dst = self._repo[state[1]]
         anc = dst.ancestor(cur)
         if anc == cur:
             self._repo.ui.debug("updating subrepo %s\n" % subrelpath(self))
             hg.update(self._repo, state[1])
         elif anc == dst:
             self._repo.ui.debug("skipping subrepo %s\n" % subrelpath(self))
         else:
             self._repo.ui.debug("merging subrepo %s\n" % subrelpath(self))
             hg.merge(self._repo, state[1], remind=False)
 
     def push(self, force):
         # push subrepos depth-first for coherent ordering
         c = self._repo['']
         subs = c.substate # only repos that are committed
         for s in sorted(subs):
             if not c.sub(s).push(force):
                 return False
 
         dsturl = _abssource(self._repo, True)
         self._repo.ui.status(_('pushing subrepo %s to %s\n') %
             (subrelpath(self), dsturl))
         other = hg.repository(self._repo.ui, dsturl)
         return self._repo.push(other, force)
 
     def outgoing(self, ui, dest, opts):
         return hg.outgoing(ui, self._repo, _abssource(self._repo, True), opts)
 
     def incoming(self, ui, source, opts):
         return hg.incoming(ui, self._repo, _abssource(self._repo, False), opts)
 
     def files(self):
         rev = self._state[1]
         ctx = self._repo[rev]
         return ctx.manifest()
 
     def filedata(self, name):
         rev = self._state[1]
         return self._repo[rev][name].data()
 
     def fileflags(self, name):
         rev = self._state[1]
         ctx = self._repo[rev]
         return ctx.flags(name)
 
 
 class svnsubrepo(abstractsubrepo):
     def __init__(self, ctx, path, state):
         self._path = path
         self._state = state
         self._ctx = ctx
         self._ui = ctx._repo.ui
 
     def _svncommand(self, commands, filename=''):
         path = os.path.join(self._ctx._repo.origroot, self._path, filename)
         cmd = ['svn'] + commands + [path]
         cmd = [util.shellquote(arg) for arg in cmd]
         cmd = util.quotecommand(' '.join(cmd))
         env = dict(os.environ)
         # Avoid localized output, preserve current locale for everything else.
         env['LC_MESSAGES'] = 'C'
         write, read, err = util.popen3(cmd, env=env, newlines=True)
         retdata = read.read()
         err = err.read().strip()
         if err:
             raise util.Abort(err)
         return retdata
 
     def _wcrev(self):
         output = self._svncommand(['info', '--xml'])
         doc = xml.dom.minidom.parseString(output)
         entries = doc.getElementsByTagName('entry')
         if not entries:
             return 0
         return int(entries[0].getAttribute('revision') or 0)
 
     def _wcchanged(self):
         """Return (changes, extchanges) where changes is True
         if the working directory was changed, and extchanges is
         True if any of these changes concern an external entry.
         """
         output = self._svncommand(['status', '--xml'])
         externals, changes = [], []
         doc = xml.dom.minidom.parseString(output)
         for e in doc.getElementsByTagName('entry'):
             s = e.getElementsByTagName('wc-status')
             if not s:
                 continue
             item = s[0].getAttribute('item')
             props = s[0].getAttribute('props')
             path = e.getAttribute('path')
             if item == 'external':
                 externals.append(path)
             if (item not in ('', 'normal', 'unversioned', 'external')
                 or props not in ('', 'none')):
                 changes.append(path)
         for path in changes:
             for ext in externals:
                 if path == ext or path.startswith(ext + os.sep):
                     return True, True
         return bool(changes), False
 
     def dirty(self):
         if self._wcrev() == self._state[1] and not self._wcchanged()[0]:
             return False
         return True
 
     def commit(self, text, user, date):
         # user and date are out of our hands since svn is centralized
         changed, extchanged = self._wcchanged()
         if not changed:
             return self._wcrev()
         if extchanged:
             # Do not try to commit externals
             raise util.Abort(_('cannot commit svn externals'))
         commitinfo = self._svncommand(['commit', '-m', text])
         self._ui.status(commitinfo)
         newrev = re.search('Committed revision ([0-9]+).', commitinfo)
         if not newrev:
             raise util.Abort(commitinfo.splitlines()[-1])
         newrev = newrev.groups()[0]
         self._ui.status(self._svncommand(['update', '-r', newrev]))
         return newrev
 
     def remove(self):
         if self.dirty():
             self._ui.warn(_('not removing repo %s because '
                             'it has changes.\n' % self._path))
             return
         self._ui.note(_('removing subrepo %s\n') % self._path)
         shutil.rmtree(self._ctx.repo.join(self._path))
 
     def get(self, state):
         status = self._svncommand(['checkout', state[0], '--revision', state[1]])
         if not re.search('Checked out revision [0-9]+.', status):
             raise util.Abort(status.splitlines()[-1])
         self._ui.status(status)
 
     def merge(self, state):
         old = int(self._state[1])
         new = int(state[1])
         if new > old:
             self.get(state)
 
     def push(self, force):
         # push is a no-op for SVN
         return True
 
     def files(self):
         output = self._svncommand(['list'])
         # This works because svn forbids \n in filenames.
         return output.splitlines()
 
     def filedata(self, name):
         return self._svncommand(['cat'], name)
 
 
 types = {
     'hg': hgsubrepo,
     'svn': svnsubrepo,
     }

mercurial/url.py

@@ -1,698 +1,698
 # url.py - HTTP handling for mercurial
 #
 # Copyright 2005, 2006, 2007, 2008 Matt Mackall <[e-mail]>
 # Copyright 2006, 2007 Alexis S. L. Carvalho <[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 version 2 or any later version.
 
 import urllib, urllib2, urlparse, httplib, os, re, socket, cStringIO
 import __builtin__
 from i18n import _
 import keepalive, util
 
 def _urlunparse(scheme, netloc, path, params, query, fragment, url):
     '''Handle cases where urlunparse(urlparse(x://)) doesn't preserve the "//"'''
     result = urlparse.urlunparse((scheme, netloc, path, params, query, fragment))
     if (scheme and
         result.startswith(scheme + ':') and
         not result.startswith(scheme + '://') and
         url.startswith(scheme + '://')
        ):
         result = scheme + '://' + result[len(scheme + ':'):]
     return result
 
 def hidepassword(url):
     '''hide user credential in a url string'''
     scheme, netloc, path, params, query, fragment = urlparse.urlparse(url)
     netloc = re.sub('([^:]*):([^@]*)@(.*)', r'\1:***@\3', netloc)
     return _urlunparse(scheme, netloc, path, params, query, fragment, url)
 
 def removeauth(url):
     '''remove all authentication information from a url string'''
     scheme, netloc, path, params, query, fragment = urlparse.urlparse(url)
     netloc = netloc[netloc.find('@')+1:]
     return _urlunparse(scheme, netloc, path, params, query, fragment, url)
 
 def netlocsplit(netloc):
     '''split [user[:passwd]@]host[:port] into 4-tuple.'''
 
     a = netloc.find('@')
     if a == -1:
         user, passwd = None, None
     else:
         userpass, netloc = netloc[:a], netloc[a + 1:]
         c = userpass.find(':')
         if c == -1:
             user, passwd = urllib.unquote(userpass), None
         else:
             user = urllib.unquote(userpass[:c])
             passwd = urllib.unquote(userpass[c + 1:])
     c = netloc.find(':')
     if c == -1:
         host, port = netloc, None
     else:
         host, port = netloc[:c], netloc[c + 1:]
     return host, port, user, passwd
 
 def netlocunsplit(host, port, user=None, passwd=None):
     '''turn host, port, user, passwd into [user[:passwd]@]host[:port].'''
     if port:
         hostport = host + ':' + port
     else:
         hostport = host
     if user:
         quote = lambda s: urllib.quote(s, safe='')
         if passwd:
             userpass = quote(user) + ':' + quote(passwd)
         else:
             userpass = quote(user)
         return userpass + '@' + hostport
     return hostport
 
 _safe = ('abcdefghijklmnopqrstuvwxyz'
          'ABCDEFGHIJKLMNOPQRSTUVWXYZ'
          '0123456789' '_.-/')
 _safeset = None
 _hex = None
 def quotepath(path):
     '''quote the path part of a URL
 
     This is similar to urllib.quote, but it also tries to avoid
     quoting things twice (inspired by wget):
 
     >>> quotepath('abc def')
     'abc%20def'
     >>> quotepath('abc%20def')
     'abc%20def'
     >>> quotepath('abc%20 def')
     'abc%20%20def'
     >>> quotepath('abc def%20')
     'abc%20def%20'
     >>> quotepath('abc def%2')
     'abc%20def%252'
     >>> quotepath('abc def%')
     'abc%20def%25'
     '''
     global _safeset, _hex
     if _safeset is None:
         _safeset = set(_safe)
         _hex = set('abcdefABCDEF0123456789')
     l = list(path)
     for i in xrange(len(l)):
         c = l[i]
         if (c == '%' and i + 2 < len(l) and
             l[i + 1] in _hex and l[i + 2] in _hex):
             pass
         elif c not in _safeset:
             l[i] = '%%%02X' % ord(c)
     return ''.join(l)
 
 class passwordmgr(urllib2.HTTPPasswordMgrWithDefaultRealm):
     def __init__(self, ui):
         urllib2.HTTPPasswordMgrWithDefaultRealm.__init__(self)
         self.ui = ui
 
     def find_user_password(self, realm, authuri):
         authinfo = urllib2.HTTPPasswordMgrWithDefaultRealm.find_user_password(
             self, realm, authuri)
         user, passwd = authinfo
         if user and passwd:
             self._writedebug(user, passwd)
             return (user, passwd)
 
         if not user:
             auth = self.readauthtoken(authuri)
             if auth:
                 user, passwd = auth.get('username'), auth.get('password')
         if not user or not passwd:
             if not self.ui.interactive():
                 raise util.Abort(_('http authorization required'))
 
             self.ui.write(_("http authorization required\n"))
             self.ui.status(_("realm: %s\n") % realm)
             if user:
                 self.ui.status(_("user: %s\n") % user)
             else:
                 user = self.ui.prompt(_("user:"), default=None)
 
             if not passwd:
                 passwd = self.ui.getpass()
 
         self.add_password(realm, authuri, user, passwd)
         self._writedebug(user, passwd)
         return (user, passwd)
 
     def _writedebug(self, user, passwd):
         msg = _('http auth: user %s, password %s\n')
         self.ui.debug(msg % (user, passwd and '*' * len(passwd) or 'not set'))
 
     def readauthtoken(self, uri):
         # Read configuration
         config = dict()
         for key, val in self.ui.configitems('auth'):
             if '.' not in key:
                 self.ui.warn(_("ignoring invalid [auth] key '%s'\n") % key)
                 continue
             group, setting = key.split('.', 1)
             gdict = config.setdefault(group, dict())
             if setting in ('username', 'cert', 'key'):
                 val = util.expandpath(val)
             gdict[setting] = val
 
         # Find the best match
         scheme, hostpath = uri.split('://', 1)
         bestlen = 0
         bestauth = None
         for auth in config.itervalues():
             prefix = auth.get('prefix')
             if not prefix:
                 continue
             p = prefix.split('://', 1)
             if len(p) > 1:
                 schemes, prefix = [p[0]], p[1]
             else:
                 schemes = (auth.get('schemes') or 'https').split()
             if (prefix == '*' or hostpath.startswith(prefix)) and \
                 len(prefix) > bestlen and scheme in schemes:
                 bestlen = len(prefix)
                 bestauth = auth
         return bestauth
 
 class proxyhandler(urllib2.ProxyHandler):
     def __init__(self, ui):
         proxyurl = ui.config("http_proxy", "host") or os.getenv('http_proxy')
         # XXX proxyauthinfo = None
 
         if proxyurl:
             # proxy can be proper url or host[:port]
             if not (proxyurl.startswith('http:') or
                     proxyurl.startswith('https:')):
                 proxyurl = 'http://' + proxyurl + '/'
             snpqf = urlparse.urlsplit(proxyurl)
             proxyscheme, proxynetloc, proxypath, proxyquery, proxyfrag = snpqf
             hpup = netlocsplit(proxynetloc)
 
             proxyhost, proxyport, proxyuser, proxypasswd = hpup
             if not proxyuser:
                 proxyuser = ui.config("http_proxy", "user")
                 proxypasswd = ui.config("http_proxy", "passwd")
 
             # see if we should use a proxy for this url
             no_list = ["localhost", "127.0.0.1"]
             no_list.extend([p.lower() for
                             p in ui.configlist("http_proxy", "no")])
             no_list.extend([p.strip().lower() for
                             p in os.getenv("no_proxy", '').split(',')
                             if p.strip()])
             # "http_proxy.always" config is for running tests on localhost
             if ui.configbool("http_proxy", "always"):
                 self.no_list = []
             else:
                 self.no_list = no_list
 
             proxyurl = urlparse.urlunsplit((
                 proxyscheme, netlocunsplit(proxyhost, proxyport,
                                                 proxyuser, proxypasswd or ''),
                 proxypath, proxyquery, proxyfrag))
             proxies = {'http': proxyurl, 'https': proxyurl}
             ui.debug('proxying through http://%s:%s\n' %
                       (proxyhost, proxyport))
         else:
             proxies = {}
 
         # urllib2 takes proxy values from the environment and those
         # will take precedence if found, so drop them
         for env in ["HTTP_PROXY", "http_proxy", "no_proxy"]:
             try:
                 if env in os.environ:
                     del os.environ[env]
             except OSError:
                 pass
 
         urllib2.ProxyHandler.__init__(self, proxies)
         self.ui = ui
 
     def proxy_open(self, req, proxy, type_):
         host = req.get_host().split(':')[0]
         if host in self.no_list:
             return None
 
         # work around a bug in Python < 2.4.2
         # (it leaves a "\n" at the end of Proxy-authorization headers)
         baseclass = req.__class__
         class _request(baseclass):
             def add_header(self, key, val):
                 if key.lower() == 'proxy-authorization':
                     val = val.strip()
                 return baseclass.add_header(self, key, val)
         req.__class__ = _request
 
         return urllib2.ProxyHandler.proxy_open(self, req, proxy, type_)
 
 class httpsendfile(object):
     """This is a wrapper around the objects returned by python's "open".
 
     Its purpose is to send file-like objects via HTTP and, to do so, it
     defines a __len__ attribute to feed the Content-Length header.
     """
 
     def __init__(self, *args, **kwargs):
         # We can't just "self._data = open(*args, **kwargs)" here because there
         # is an "open" function defined in this module that shadows the global
         # one
         self._data = __builtin__.open(*args, **kwargs)
         self.read = self._data.read
         self.seek = self._data.seek
         self.close = self._data.close
         self.write = self._data.write
 
     def __len__(self):
         return os.fstat(self._data.fileno()).st_size
 
 def _gen_sendfile(connection):
     def _sendfile(self, data):
         # send a file
         if isinstance(data, httpsendfile):
             # if auth required, some data sent twice, so rewind here
             data.seek(0)
             for chunk in util.filechunkiter(data):
                 connection.send(self, chunk)
         else:
             connection.send(self, data)
     return _sendfile
 
 has_https = hasattr(urllib2, 'HTTPSHandler')
 if has_https:
     try:
         # avoid using deprecated/broken FakeSocket in python 2.6
         import ssl
         _ssl_wrap_socket = ssl.wrap_socket
         CERT_REQUIRED = ssl.CERT_REQUIRED
     except ImportError:
         CERT_REQUIRED = 2
 
         def _ssl_wrap_socket(sock, key_file, cert_file,
                              cert_reqs=CERT_REQUIRED, ca_certs=None):
             if ca_certs:
                 raise util.Abort(_(
                     'certificate checking requires Python 2.6'))
 
             ssl = socket.ssl(sock, key_file, cert_file)
             return httplib.FakeSocket(sock, ssl)
 
     try:
         _create_connection = socket.create_connection
     except AttributeError:
         _GLOBAL_DEFAULT_TIMEOUT = object()
 
         def _create_connection(address, timeout=_GLOBAL_DEFAULT_TIMEOUT,
                                source_address=None):
             # lifted from Python 2.6
 
             msg = "getaddrinfo returns an empty list"
             host, port = address
             for res in socket.getaddrinfo(host, port, 0, socket.SOCK_STREAM):
                 af, socktype, proto, canonname, sa = res
                 sock = None
                 try:
                     sock = socket.socket(af, socktype, proto)
                     if timeout is not _GLOBAL_DEFAULT_TIMEOUT:
                         sock.settimeout(timeout)
                     if source_address:
                         sock.bind(source_address)
                     sock.connect(sa)
                     return sock
 
                 except socket.error, msg:
                     if sock is not None:
                         sock.close()
 
             raise socket.error, msg
 
 class httpconnection(keepalive.HTTPConnection):
     # must be able to send big bundle as stream.
     send = _gen_sendfile(keepalive.HTTPConnection)
 
     def connect(self):
         if has_https and self.realhostport: # use CONNECT proxy
             self.sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
             self.sock.connect((self.host, self.port))
             if _generic_proxytunnel(self):
                 # we do not support client x509 certificates
                 self.sock = _ssl_wrap_socket(self.sock, None, None)
         else:
             keepalive.HTTPConnection.connect(self)
 
     def getresponse(self):
         proxyres = getattr(self, 'proxyres', None)
         if proxyres:
             if proxyres.will_close:
                 self.close()
             self.proxyres = None
             return proxyres
         return keepalive.HTTPConnection.getresponse(self)
 
 # general transaction handler to support different ways to handle
 # HTTPS proxying before and after Python 2.6.3.
 def _generic_start_transaction(handler, h, req):
     if hasattr(req, '_tunnel_host') and req._tunnel_host:
         tunnel_host = req._tunnel_host
         if tunnel_host[:7] not in ['http://', 'https:/']:
             tunnel_host = 'https://' + tunnel_host
         new_tunnel = True
     else:
         tunnel_host = req.get_selector()
         new_tunnel = False
 
     if new_tunnel or tunnel_host == req.get_full_url(): # has proxy
         urlparts = urlparse.urlparse(tunnel_host)
         if new_tunnel or urlparts[0] == 'https': # only use CONNECT for HTTPS
             realhostport = urlparts[1]
             if realhostport[-1] == ']' or ':' not in realhostport:
                 realhostport += ':443'
 
             h.realhostport = realhostport
             h.headers = req.headers.copy()
             h.headers.update(handler.parent.addheaders)
             return
 
     h.realhostport = None
     h.headers = None
 
 def _generic_proxytunnel(self):
     proxyheaders = dict(
             [(x, self.headers[x]) for x in self.headers
              if x.lower().startswith('proxy-')])
     self._set_hostport(self.host, self.port)
     self.send('CONNECT %s HTTP/1.0\r\n' % self.realhostport)
     for header in proxyheaders.iteritems():
         self.send('%s: %s\r\n' % header)
     self.send('\r\n')
 
     # majority of the following code is duplicated from
     # httplib.HTTPConnection as there are no adequate places to
     # override functions to provide the needed functionality
     res = self.response_class(self.sock,
                               strict=self.strict,
                               method=self._method)
 
     while True:
         version, status, reason = res._read_status()
         if status != httplib.CONTINUE:
             break
         while True:
             skip = res.fp.readline().strip()
             if not skip:
                 break
     res.status = status
     res.reason = reason.strip()
 
     if res.status == 200:
         while True:
             line = res.fp.readline()
             if line == '\r\n':
                 break
         return True
 
     if version == 'HTTP/1.0':
         res.version = 10
     elif version.startswith('HTTP/1.'):
         res.version = 11
     elif version == 'HTTP/0.9':
         res.version = 9
     else:
         raise httplib.UnknownProtocol(version)
 
     if res.version == 9:
         res.length = None
         res.chunked = 0
         res.will_close = 1
         res.msg = httplib.HTTPMessage(cStringIO.StringIO())
         return False
 
     res.msg = httplib.HTTPMessage(res.fp)
     res.msg.fp = None
 
     # are we using the chunked-style of transfer encoding?
     trenc = res.msg.getheader('transfer-encoding')
     if trenc and trenc.lower() == "chunked":
         res.chunked = 1
         res.chunk_left = None
     else:
         res.chunked = 0
 
     # will the connection close at the end of the response?
     res.will_close = res._check_close()
 
     # do we have a Content-Length?
     # NOTE: RFC 2616, S4.4, #3 says we ignore this if tr_enc is "chunked"
     length = res.msg.getheader('content-length')
     if length and not res.chunked:
         try:
             res.length = int(length)
         except ValueError:
             res.length = None
         else:
             if res.length < 0:  # ignore nonsensical negative lengths
                 res.length = None
     else:
         res.length = None
 
     # does the body have a fixed length? (of zero)
     if (status == httplib.NO_CONTENT or status == httplib.NOT_MODIFIED or
         100 <= status < 200 or # 1xx codes
         res._method == 'HEAD'):
         res.length = 0
 
     # if the connection remains open, and we aren't using chunked, and
     # a content-length was not provided, then assume that the connection
     # WILL close.
     if (not res.will_close and
        not res.chunked and
        res.length is None):
         res.will_close = 1
 
     self.proxyres = res
 
     return False
 
 class httphandler(keepalive.HTTPHandler):
     def http_open(self, req):
         return self.do_open(httpconnection, req)
 
     def _start_transaction(self, h, req):
         _generic_start_transaction(self, h, req)
         return keepalive.HTTPHandler._start_transaction(self, h, req)
 
 def _verifycert(cert, hostname):
     '''Verify that cert (in socket.getpeercert() format) matches hostname.
     CRLs and subjectAltName are not handled.
-    
+
     Returns error message if any problems are found and None on success.
     '''
     if not cert:
         return _('no certificate received')
     dnsname = hostname.lower()
     for s in cert.get('subject', []):
         key, value = s[0]
         if key == 'commonName':
             certname = value.lower()
             if (certname == dnsname or
                 '.' in dnsname and certname == '*.' + dnsname.split('.', 1)[1]):
                 return None
             return _('certificate is for %s') % certname
     return _('no commonName found in certificate')
 
 if has_https:
     class BetterHTTPS(httplib.HTTPSConnection):
         send = keepalive.safesend
 
         def connect(self):
             if hasattr(self, 'ui'):
                 cacerts = self.ui.config('web', 'cacerts')
             else:
                 cacerts = None
 
             if cacerts:
                 sock = _create_connection((self.host, self.port))
                 self.sock = _ssl_wrap_socket(sock, self.key_file,
                         self.cert_file, cert_reqs=CERT_REQUIRED,
                         ca_certs=cacerts)
                 msg = _verifycert(self.sock.getpeercert(), self.host)
                 if msg:
                     raise util.Abort(_('%s certificate error: %s') %
                                      (self.host, msg))
                 self.ui.debug('%s certificate successfully verified\n' %
                               self.host)
             else:
                 httplib.HTTPSConnection.connect(self)
 
     class httpsconnection(BetterHTTPS):
         response_class = keepalive.HTTPResponse
         # must be able to send big bundle as stream.
         send = _gen_sendfile(BetterHTTPS)
         getresponse = keepalive.wrapgetresponse(httplib.HTTPSConnection)
 
         def connect(self):
             if self.realhostport: # use CONNECT proxy
                 self.sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
                 self.sock.connect((self.host, self.port))
                 if _generic_proxytunnel(self):
                     self.sock = _ssl_wrap_socket(self.sock, self.cert_file,
                                                  self.key_file)
             else:
                 BetterHTTPS.connect(self)
 
     class httpshandler(keepalive.KeepAliveHandler, urllib2.HTTPSHandler):
         def __init__(self, ui):
             keepalive.KeepAliveHandler.__init__(self)
             urllib2.HTTPSHandler.__init__(self)
             self.ui = ui
             self.pwmgr = passwordmgr(self.ui)
 
         def _start_transaction(self, h, req):
             _generic_start_transaction(self, h, req)
             return keepalive.KeepAliveHandler._start_transaction(self, h, req)
 
         def https_open(self, req):
             self.auth = self.pwmgr.readauthtoken(req.get_full_url())
             return self.do_open(self._makeconnection, req)
 
         def _makeconnection(self, host, port=None, *args, **kwargs):
             keyfile = None
             certfile = None
 
             if len(args) >= 1: # key_file
                 keyfile = args[0]
             if len(args) >= 2: # cert_file
                 certfile = args[1]
             args = args[2:]
 
             # if the user has specified different key/cert files in
             # hgrc, we prefer these
             if self.auth and 'key' in self.auth and 'cert' in self.auth:
                 keyfile = self.auth['key']
                 certfile = self.auth['cert']
 
             conn = httpsconnection(host, port, keyfile, certfile, *args, **kwargs)
             conn.ui = self.ui
             return conn
 
 class httpdigestauthhandler(urllib2.HTTPDigestAuthHandler):
     def __init__(self, *args, **kwargs):
         urllib2.HTTPDigestAuthHandler.__init__(self, *args, **kwargs)
         self.retried_req = None
 
     def reset_retry_count(self):
         # Python 2.6.5 will call this on 401 or 407 errors and thus loop
         # forever. We disable reset_retry_count completely and reset in
         # http_error_auth_reqed instead.
         pass
 
     def http_error_auth_reqed(self, auth_header, host, req, headers):
         # Reset the retry counter once for each request.
         if req is not self.retried_req:
             self.retried_req = req
             self.retried = 0
         # In python < 2.5 AbstractDigestAuthHandler raises a ValueError if
         # it doesn't know about the auth type requested. This can happen if
         # somebody is using BasicAuth and types a bad password.
         try:
             return urllib2.HTTPDigestAuthHandler.http_error_auth_reqed(
                         self, auth_header, host, req, headers)
         except ValueError, inst:
             arg = inst.args[0]
             if arg.startswith("AbstractDigestAuthHandler doesn't know "):
                 return
             raise
 
 class httpbasicauthhandler(urllib2.HTTPBasicAuthHandler):
     def __init__(self, *args, **kwargs):
         urllib2.HTTPBasicAuthHandler.__init__(self, *args, **kwargs)
         self.retried_req = None
 
     def reset_retry_count(self):
         # Python 2.6.5 will call this on 401 or 407 errors and thus loop
         # forever. We disable reset_retry_count completely and reset in
         # http_error_auth_reqed instead.
         pass
 
     def http_error_auth_reqed(self, auth_header, host, req, headers):
         # Reset the retry counter once for each request.
         if req is not self.retried_req:
             self.retried_req = req
             self.retried = 0
         return urllib2.HTTPBasicAuthHandler.http_error_auth_reqed(
                         self, auth_header, host, req, headers)
 
 def getauthinfo(path):
     scheme, netloc, urlpath, query, frag = urlparse.urlsplit(path)
     if not urlpath:
         urlpath = '/'
     if scheme != 'file':
         # XXX: why are we quoting the path again with some smart
         # heuristic here? Anyway, it cannot be done with file://
         # urls since path encoding is os/fs dependent (see
         # urllib.pathname2url() for details).
         urlpath = quotepath(urlpath)
     host, port, user, passwd = netlocsplit(netloc)
 
     # urllib cannot handle URLs with embedded user or passwd
     url = urlparse.urlunsplit((scheme, netlocunsplit(host, port),
                               urlpath, query, frag))
     if user:
         netloc = host
         if port:
             netloc += ':' + port
         # Python < 2.4.3 uses only the netloc to search for a password
         authinfo = (None, (url, netloc), user, passwd or '')
     else:
         authinfo = None
     return url, authinfo
 
 handlerfuncs = []
 
 def opener(ui, authinfo=None):
     '''
     construct an opener suitable for urllib2
     authinfo will be added to the password manager
     '''
     handlers = [httphandler()]
     if has_https:
         handlers.append(httpshandler(ui))
 
     handlers.append(proxyhandler(ui))
 
     passmgr = passwordmgr(ui)
     if authinfo is not None:
         passmgr.add_password(*authinfo)
         user, passwd = authinfo[2:4]
         ui.debug('http auth: user %s, password %s\n' %
                  (user, passwd and '*' * len(passwd) or 'not set'))
 
     handlers.extend((httpbasicauthhandler(passmgr),
                      httpdigestauthhandler(passmgr)))
     handlers.extend([h(ui, passmgr) for h in handlerfuncs])
     opener = urllib2.build_opener(*handlers)
 
     # 1.0 here is the _protocol_ version
     opener.addheaders = [('User-agent', 'mercurial/proto-1.0')]
     opener.addheaders.append(('Accept', 'application/mercurial-0.1'))
     return opener
 
 scheme_re = re.compile(r'^([a-zA-Z0-9+-.]+)://')
 
 def open(ui, url, data=None):
     scheme = None
     m = scheme_re.search(url)
     if m:
         scheme = m.group(1).lower()
     if not scheme:
         path = util.normpath(os.path.abspath(url))
         url = 'file://' + urllib.pathname2url(path)
         authinfo = None
     else:
         url, authinfo = getauthinfo(url)
     return opener(ui, authinfo).open(url, data)