hgext/convert/subversion.py

43075:57875cf423c9

    import svn.core
    import svn.ra
    import svn.delta
    from . import transport
    import warnings
    warnings.filterwarnings('ignore',
            module='svn.core',
            category=DeprecationWarning)
    svn.core.SubversionException # trigger import to catch error

except ImportError:
    svn = None

class SvnPathNotFound(Exception):
    pass

def revsplit(rev):
    """Parse a revision string and return (uuid, path, revnum).
    >>> revsplit(b'svn:a2147622-4a9f-4db4-a8d3-13562ff547b2'
    ...          b'/proj%20B/mytrunk/mytrunk@1')

    if len(parts) > 1 and parts[0].startswith('svn:'):
        uuid = parts[0][4:]
        mod = '/' + parts[1]
    return uuid, mod, revnum

def quote(s):
    # As of svn 1.7, many svn calls expect "canonical" paths. In
    # theory, we should call svn.core.*canonicalize() on all paths
    # before passing them to the API.  Instead, we assume the base url
    # is canonical and copy the behaviour of svn URL encoding function
    # so we can extend it safely with new components. The "safe"
    # characters were taken from the "svn_uri__char_validity" table in
    # libsvn_subr/path.c.
    return urlreq.quote(s, "!$&'()*+,-./:=@_~")

def geturl(path):
    try:
        return svn.client.url_from_path(svn.core.svn_path_canonicalize(path))
    except svn.core.SubversionException:

        # by svn API, which is UTF-8.
        path = encoding.tolocal(path)
        path = 'file://%s' % quote(path)
    return svn.core.svn_path_canonicalize(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):
        paths = {}
        if orig_paths is not None:
            for k, v in orig_paths.iteritems():
                paths[k] = changedpath(v)
        pickle.dump((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 IOError:
        # Caller may interrupt the iteration
        pickle.dump(None, fp, protocol)
    except Exception as inst:
        pickle.dump(stringutil.forcebytestr(inst), fp, protocol)

    # 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.
    """
    if svn is None:
        raise error.Abort(_('debugsvnlog could not load Subversion python '
                           'bindings'))

    args = decodeargs(ui.fin.read())
    get_log_child(ui.fout, *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 error.Abort(_('Mercurial failed to run itself, check'
                                   ' hg executable is in PATH'))
            try:
                orig_paths, revnum, author, date, message = entry
            except (TypeError, ValueError):
                if entry is None:
                    break

    def close(self):
        if self._stdout:
            self._stdout.close()
            self._stdout = None

class directlogstream(list):
    """Direct revision log iterator.
    This can be used for debugging and development but it will probably leak
    memory and is not suitable for real conversions."""
    def __init__(self, url, paths, start, end, limit=0,
                  discover_changed_paths=True, strict_node_history=False):

        def receiver(orig_paths, revnum, author, date, message, pool):
            paths = {}
            if orig_paths is not None:
                for k, v in orig_paths.iteritems():
                    paths[k] = changedpath(v)
            self.append((paths, revnum, author, date, message))

        # 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)

    def close(self):
        pass

# 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):

        rsp = opener.open('%s://%s/!svn/ver/0/.svn' % (proto, path), 'rb')
        data = rsp.read()
    except urlerr.httperror as 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 Exception:
        # Could be urlerr.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':
            if (pycompat.iswindows and path[:1] == '/'
                  and path[1:2].isalpha() and path[2:6].lower() == '%3a/'):
                path = path[:2] + ':/' + path[6:]
            path = urlreq.url2pathname(path)
    except ValueError:
        proto = 'file'
        path = os.path.abspath(url)

    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,

#
class svn_source(converter_source):
    def __init__(self, ui, repotype, url, revs=None):
        super(svn_source, self).__init__(ui, repotype, url, revs=revs)

        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)
        if svn is None:
            raise MissingTool(_('could not load Subversion python bindings'))

        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 = urlreq.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 svn.core.SubversionException:
            ui.traceback()
            svnversion = '%d.%d.%d' % (svn.core.SVN_VER_MAJOR,
                                       svn.core.SVN_VER_MINOR,
                                       svn.core.SVN_VER_MICRO)
            raise NoRepo(_("%s does not look like a Subversion repository "
                           "to libsvn version %s")
                         % (self.url, svnversion))

        if revs:
            if len(revs) > 1:
                raise error.Abort(_('subversion source does not support '
                                   'specifying multiple revisions'))
            try:
                latest = int(revs[0])
            except ValueError:
                raise error.Abort(_('svn: revision %s is not an integer') %
                                 revs[0])

        trunkcfg = self.ui.config('convert', 'svn.trunk')
        if trunkcfg is None:
            trunkcfg = 'trunk'
        self.trunkname = trunkcfg.strip('/')

        try:
            self.startrev = int(self.startrev)
            if self.startrev < 0:
                self.startrev = 0
        except ValueError:
            raise error.Abort(_('svn: start revision %s is not an integer')
                             % self.startrev)

        try:
            self.head = self.latest(self.module, latest)
        except SvnPathNotFound:
            self.head = None
        if not self.head:
            raise error.Abort(_('no revision found in module %s')
                             % self.module)
        self.last_changed = self.revnum(self.head)

        self._changescache = (None, None)

        if os.path.exists(os.path.join(url, '.svn/entries')):

                lastrevs[module] = revnum
        self.lastrevs = lastrevs

    def exists(self, path, optrev):
        try:
            svn.client.ls(self.url.rstrip('/') + '/' + quote(path),
                                 optrev, False, self.ctx)
            return True
        except svn.core.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):

            if not self.exists(path, rev):
                if self.module.endswith(path) and name == 'trunk':
                    # we are converting from inside this directory
                    return None
                if cfgpath:
                    raise error.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)

        if trunk:
            oldmodule = self.module or ''
            self.module += '/' + trunk
            self.head = self.latest(self.module, self.last_changed)
            if not self.head:
                raise error.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 + '/' + quote(branches),
                                        rev, False, self.ctx)
            for branch in sorted(branchnames):
                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 error.Abort(_('svn: start revision is not supported '
                                   'with more than one branch'))
            revnum = self.revnum(self.heads[0])
            if revnum < self.startrev:
                raise error.Abort(
                    _('svn: no revision found after start revision %d')
                                 % self.startrev)

        return self.heads

    def _getchanges(self, rev, full):
        (paths, parents) = self.paths[rev]

        if parents:
            files, self.removed, copies = self.expandpaths(rev, paths, parents)
        if full or not parents:
            # Perform a full checkout on roots
            uuid, module, revnum = revsplit(rev)
            entries = svn.client.ls(self.baseurl + quote(module),
                                    optrev(revnum), True, self.ctx)
            files = [n for n, e in entries.iteritems()
                     if e.kind == svn.core.svn_node_file]
            self.removed = set()

        files.sort()
        files = zip(files, [rev] * len(files))
        return (files, copies)

        del self.commits[rev]
        return revcommit

    def checkrevformat(self, revstr, mapname='splicemap'):
        """ fails if revision format does not match the correct format"""
        if not re.match(r'svn:[0-9a-f]{8,8}-[0-9a-f]{4,4}-'
                              r'[0-9a-f]{4,4}-[0-9a-f]{4,4}-[0-9a-f]'
                              r'{12,12}(.*)\@[0-9]+$',revstr):
            raise error.Abort(_('%s entry %s is not a valid revision'
                               ' identifier') % (mapname, revstr))

    def numcommits(self):
        return int(self.head.rsplit('@', 1)[1]) - self.startrev

    def gettags(self):

        try:
            for entry in stream:
                origpaths, revnum, author, date, message = entry
                if not origpaths:
                    origpaths = []
                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:

                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])

                # /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
                renamings = []
                for source, sourcerev, dest in pendings:
                    tagname = dest.split('/')[-1]

    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'),
                                  'ab')
        self.convertfp.write(util.tonativeeol('%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)

        revision. 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.
        """
        def findchanges(path, start, stop=None):
            stream = self._getlog([path], start, stop or 1)
            try:
                for entry in stream:
                    paths, revnum, author, date, message = entry

                        break
                    if revnum <= stop:
                        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
                if not paths:
                    revnum = None
                return revnum, path

            dirent = svn.ra.stat(self.ra, path.strip('/'), stop)
            self.reparent(prevmodule)
        except svn.core.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.
        revnum, realpath = findchanges(path, stop, dirent.created_rev)

        new_module, revnum = revsplit(rev)[1:]
        if new_module != self.module:
            self.module = new_module
            self.reparent(self.module)

        progress = self.ui.makeprogress(_('scanning paths'), unit=_('paths'),
                                        total=len(paths))
        for i, (path, ent) in enumerate(paths):
            progress.update(i, item=path)
            entrypath = self.getrelpath(path)

            kind = self._checkpath(entrypath, revnum)

                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 = revsplit(parents[0])[1:]
                parentpath = pmodule + "/" + entrypath
                fromkind = self._checkpath(entrypath, prevnum, pmodule)

                        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:
                if ent.action == 'M':
                    # If the directory just had a prop change,
                    # then we shouldn't need to look for its children.
                    continue

                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)

        progress.complete()
        changed.update(removed)

        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

            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 = 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

                if branch == self.trunkname:
                    branch = None
            except IndexError:
                branch = None

            cset = commit(author=author,
                          date=dateutil.datestr(date, '%Y-%m-%d %H:%M:%S %1%2'),
                          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)

                        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:

                except SvnPathNotFound:
                    pass
        except svn.core.SubversionException as xxx_todo_changeme:
            (inst, num) = xxx_todo_changeme.args
            if num == svn.core.SVN_ERR_FS_NO_SUCH_REVISION:
                raise error.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:

            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 svn.core.SubversionException as 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
                return None, None
            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 = svn.core.Pool()
        rpath = '/'.join([self.baseurl, quote(path)]).strip('/')
        entries = svn.client.ls(rpath, optrev(revnum), True, self.ctx, pool)
        if path:
            path += '/'
        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

        # extract the "entry" portion (a relative path) from what
        # svn log --xml says, i.e.
        #   "/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

            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]
        # developer config: convert.svn.debugsvnlog
        if not self.ui.configbool('convert', 'svn.debugsvnlog'):
            return directlogstream(*args)
        arg = encodeargs(args)
        hgexe = procutil.hgexecutable()

        stdin, stdout = procutil.popen2(procutil.quotecommand(cmd))
        stdin.write(arg)
        try:
            stdin.close()
        except IOError:
            raise error.Abort(_('Mercurial failed to run itself, check'
                               ' hg executable is in PATH'))
        return logstream(stdout)

pre_revprop_change = b'''#!/bin/sh

REPOS="$1"
REV="$2"

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(br'Committed revision (\d+).', re.M)
    uuid_re = re.compile(br'Repository UUID:\s*(\S+)', re.M)

        else:
            if not re.search(br'^(file|http|https|svn|svn\+ssh)\://', path):
                path = os.path.realpath(path)
                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 '%s'\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

            wcpath = os.path.join(encoding.getcwd(), os.path.basename(path) +
                                '-wc')
            ui.status(_("initializing svn working copy '%s'\n")
                      % os.path.basename(wcpath))
            self.run0('checkout', path, wcpath)

            self.wc = wcpath
        self.opener = vfsmod.vfs(self.wc)
        self.wopener = vfsmod.vfs(self.wc)

        doc = xml.dom.minidom.parseString(output)
        for e in doc.getElementsByTagName(r'entry'):
            for n in e.childNodes:
                if n.nodeType != n.ELEMENT_NODE or n.tagName != r'name':
                    continue
                name = r''.join(c.data for c in n.childNodes
                                if c.nodeType == c.TEXT_NODE)
                # Entries are compared with names coming from
                # mercurial, so bytes with undefined encoding. Our
                # best bet is to assume they are in local
                # encoding. They will be passed to command line calls
                # later anyway, so they better be.

        # already occurred.  Cross the semantic gap.
        wdest = self.wjoin(dest)
        exists = os.path.lexists(wdest)
        if exists:
            fd, tempname = pycompat.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)

            for i in iter(lambda: f.rfind('/', 0, i), -1):
                dirs.add(f[:i])
        return dirs

    def add_dirs(self, files):
        add_dirs = [d for d in sorted(self.dirs_of(files))
                    if d not in self.manifest]
        if add_dirs:
            self.manifest.update(add_dirs)
            self.xargs(add_dirs, 'add', non_recursive=True, quiet=True)
        return add_dirs

        self.childmap[parent] = child

    def revid(self, rev):
        return "svn:%s@%s" % (self.uuid, rev)

    def putcommit(self, files, copies, parents, commit, source, revmap, full,
                  cleanp2):
        for parent in parents:
            try:
                return self.revid(self.childmap[parent])
            except KeyError:
                pass

        fd, messagefile = pycompat.mkstemp(prefix='hg-convert-')
        fp = os.fdopen(fd, r'wb')
        fp.write(util.tonativeeol(commit.desc))
        fp.close()
        try:
            output = self.run0('commit',
                               username=stringutil.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] if parents else 'None'
                self.ui.warn(_('unexpected svn output:\n'))
                self.ui.warn(output)
                raise error.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 hascommitforsplicemap(self, rev):
        # This is not correct as one can convert to an existing subversion
        # repository and childmap would not list all revisions. Too bad.
        if rev in self.childmap:
            return True
        raise error.Abort(_('splice map revision %s not found in subversion '
                           'child map (revision lookups are not implemented)')
                         % rev)

43076:2372284d9457

    import svn.core
    import svn.ra
    import svn.delta
    from . import transport
    import warnings

    warnings.filterwarnings(
        'ignore', module='svn.core', category=DeprecationWarning
    )
    svn.core.SubversionException  # trigger import to catch error

except ImportError:
    svn = None


class SvnPathNotFound(Exception):
    pass


def revsplit(rev):
    """Parse a revision string and return (uuid, path, revnum).
    >>> revsplit(b'svn:a2147622-4a9f-4db4-a8d3-13562ff547b2'
    ...          b'/proj%20B/mytrunk/mytrunk@1')

    if len(parts) > 1 and parts[0].startswith('svn:'):
        uuid = parts[0][4:]
        mod = '/' + parts[1]
    return uuid, mod, revnum


def quote(s):
    # As of svn 1.7, many svn calls expect "canonical" paths. In
    # theory, we should call svn.core.*canonicalize() on all paths
    # before passing them to the API.  Instead, we assume the base url
    # is canonical and copy the behaviour of svn URL encoding function
    # so we can extend it safely with new components. The "safe"
    # characters were taken from the "svn_uri__char_validity" table in
    # libsvn_subr/path.c.
    return urlreq.quote(s, "!$&'()*+,-./:=@_~")


def geturl(path):
    try:
        return svn.client.url_from_path(svn.core.svn_path_canonicalize(path))
    except svn.core.SubversionException:

        # by svn API, which is UTF-8.
        path = encoding.tolocal(path)
        path = 'file://%s' % quote(path)
    return svn.core.svn_path_canonicalize(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):
        paths = {}
        if orig_paths is not None:
            for k, v in orig_paths.iteritems():
                paths[k] = changedpath(v)
        pickle.dump((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 IOError:
        # Caller may interrupt the iteration
        pickle.dump(None, fp, protocol)
    except Exception as inst:
        pickle.dump(stringutil.forcebytestr(inst), fp, protocol)

    # 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.
    """
    if svn is None:
        raise error.Abort(
            _('debugsvnlog could not load Subversion python ' 'bindings')
        )

    args = decodeargs(ui.fin.read())
    get_log_child(ui.fout, *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 error.Abort(
                    _(
                        'Mercurial failed to run itself, check'
                        ' hg executable is in PATH'
                    )
                )
            try:
                orig_paths, revnum, author, date, message = entry
            except (TypeError, ValueError):
                if entry is None:
                    break

    def close(self):
        if self._stdout:
            self._stdout.close()
            self._stdout = None


class directlogstream(list):
    """Direct revision log iterator.
    This can be used for debugging and development but it will probably leak
    memory and is not suitable for real conversions."""

    def __init__(
        self,
        url,
        paths,
        start,
        end,
        limit=0,
        discover_changed_paths=True,
        strict_node_history=False,
    ):
        def receiver(orig_paths, revnum, author, date, message, pool):
            paths = {}
            if orig_paths is not None:
                for k, v in orig_paths.iteritems():
                    paths[k] = changedpath(v)
            self.append((paths, revnum, author, date, message))

        # 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,
        )

    def close(self):
        pass


# 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):

        rsp = opener.open('%s://%s/!svn/ver/0/.svn' % (proto, path), 'rb')
        data = rsp.read()
    except urlerr.httperror as 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 Exception:
        # Could be urlerr.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':
            if (
                pycompat.iswindows
                and path[:1] == '/'
                and path[1:2].isalpha()
                and path[2:6].lower() == '%3a/'
            ):
                path = path[:2] + ':/' + path[6:]
            path = urlreq.url2pathname(path)
    except ValueError:
        proto = 'file'
        path = os.path.abspath(url)

    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,

#
class svn_source(converter_source):
    def __init__(self, ui, repotype, url, revs=None):
        super(svn_source, self).__init__(ui, repotype, url, revs=revs)

        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
            )
        if svn is None:
            raise MissingTool(_('could not load Subversion python bindings'))

        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 = urlreq.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 svn.core.SubversionException:
            ui.traceback()
            svnversion = '%d.%d.%d' % (
                svn.core.SVN_VER_MAJOR,
                svn.core.SVN_VER_MINOR,
                svn.core.SVN_VER_MICRO,
            )
            raise NoRepo(
                _(
                    "%s does not look like a Subversion repository "
                    "to libsvn version %s"
                )
                % (self.url, svnversion)
            )

        if revs:
            if len(revs) > 1:
                raise error.Abort(
                    _(
                        'subversion source does not support '
                        'specifying multiple revisions'
                    )
                )
            try:
                latest = int(revs[0])
            except ValueError:
                raise error.Abort(
                    _('svn: revision %s is not an integer') % revs[0]
                )

        trunkcfg = self.ui.config('convert', 'svn.trunk')
        if trunkcfg is None:
            trunkcfg = 'trunk'
        self.trunkname = trunkcfg.strip('/')

        try:
            self.startrev = int(self.startrev)
            if self.startrev < 0:
                self.startrev = 0
        except ValueError:
            raise error.Abort(
                _('svn: start revision %s is not an integer') % self.startrev
            )

        try:
            self.head = self.latest(self.module, latest)
        except SvnPathNotFound:
            self.head = None
        if not self.head:
            raise error.Abort(_('no revision found in module %s') % self.module)
        self.last_changed = self.revnum(self.head)

        self._changescache = (None, None)

        if os.path.exists(os.path.join(url, '.svn/entries')):

                lastrevs[module] = revnum
        self.lastrevs = lastrevs

    def exists(self, path, optrev):
        try:
            svn.client.ls(
                self.url.rstrip('/') + '/' + quote(path),
                optrev,
                False,
                self.ctx,
            )
            return True
        except svn.core.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):

            if not self.exists(path, rev):
                if self.module.endswith(path) and name == 'trunk':
                    # we are converting from inside this directory
                    return None
                if cfgpath:
                    raise error.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)

        if trunk:
            oldmodule = self.module or ''
            self.module += '/' + trunk
            self.head = self.latest(self.module, self.last_changed)
            if not self.head:
                raise error.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 + '/' + quote(branches), rev, False, self.ctx
            )
            for branch in sorted(branchnames):
                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 error.Abort(
                    _(
                        'svn: start revision is not supported '
                        'with more than one branch'
                    )
                )
            revnum = self.revnum(self.heads[0])
            if revnum < self.startrev:
                raise error.Abort(
                    _('svn: no revision found after start revision %d')
                    % self.startrev
                )

        return self.heads

    def _getchanges(self, rev, full):
        (paths, parents) = self.paths[rev]

        if parents:
            files, self.removed, copies = self.expandpaths(rev, paths, parents)
        if full or not parents:
            # Perform a full checkout on roots
            uuid, module, revnum = revsplit(rev)
            entries = svn.client.ls(
                self.baseurl + quote(module), optrev(revnum), True, self.ctx
            )
            files = [
                n
                for n, e in entries.iteritems()
                if e.kind == svn.core.svn_node_file
            ]
            self.removed = set()

        files.sort()
        files = zip(files, [rev] * len(files))
        return (files, copies)

        del self.commits[rev]
        return revcommit

    def checkrevformat(self, revstr, mapname='splicemap'):
        """ fails if revision format does not match the correct format"""
        if not re.match(
            r'svn:[0-9a-f]{8,8}-[0-9a-f]{4,4}-'
            r'[0-9a-f]{4,4}-[0-9a-f]{4,4}-[0-9a-f]'
            r'{12,12}(.*)\@[0-9]+$',
            revstr,
        ):
            raise error.Abort(
                _('%s entry %s is not a valid revision' ' identifier')
                % (mapname, revstr)
            )

    def numcommits(self):
        return int(self.head.rsplit('@', 1)[1]) - self.startrev

    def gettags(self):

        try:
            for entry in stream:
                origpaths, revnum, author, date, message = entry
                if not origpaths:
                    origpaths = []
                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:

                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])

                # /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
                renamings = []
                for source, sourcerev, dest in pendings:
                    tagname = dest.split('/')[-1]

    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'), 'ab'
            )
        self.convertfp.write(
            util.tonativeeol('%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)

        revision. 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.
        """

        def findchanges(path, start, stop=None):
            stream = self._getlog([path], start, stop or 1)
            try:
                for entry in stream:
                    paths, revnum, author, date, message = entry

                        break
                    if revnum <= stop:
                        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
                if not paths:
                    revnum = None
                return revnum, path

            dirent = svn.ra.stat(self.ra, path.strip('/'), stop)
            self.reparent(prevmodule)
        except svn.core.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.
        revnum, realpath = findchanges(path, stop, dirent.created_rev)

        new_module, revnum = revsplit(rev)[1:]
        if new_module != self.module:
            self.module = new_module
            self.reparent(self.module)

        progress = self.ui.makeprogress(
            _('scanning paths'), unit=_('paths'), total=len(paths)
        )
        for i, (path, ent) in enumerate(paths):
            progress.update(i, item=path)
            entrypath = self.getrelpath(path)

            kind = self._checkpath(entrypath, revnum)

                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 = revsplit(parents[0])[1:]
                parentpath = pmodule + "/" + entrypath
                fromkind = self._checkpath(entrypath, prevnum, pmodule)

                        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:
                if ent.action == 'M':
                    # If the directory just had a prop change,
                    # then we shouldn't need to look for its children.
                    continue

                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)

        progress.complete()
        changed.update(removed)

        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

            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 = 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

                if branch == self.trunkname:
                    branch = None
            except IndexError:
                branch = None

            cset = commit(
                author=author,
                date=dateutil.datestr(date, '%Y-%m-%d %H:%M:%S %1%2'),
                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)

                        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:

                except SvnPathNotFound:
                    pass
        except svn.core.SubversionException as xxx_todo_changeme:
            (inst, num) = xxx_todo_changeme.args
            if num == svn.core.SVN_ERR_FS_NO_SUCH_REVISION:
                raise error.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:

            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 svn.core.SubversionException as 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
                return None, None
            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 = svn.core.Pool()
        rpath = '/'.join([self.baseurl, quote(path)]).strip('/')
        entries = svn.client.ls(rpath, optrev(revnum), True, self.ctx, pool)
        if path:
            path += '/'
        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

        # extract the "entry" portion (a relative path) from what
        # svn log --xml says, i.e.
        #   "/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

            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,
        ]
        # developer config: convert.svn.debugsvnlog
        if not self.ui.configbool('convert', 'svn.debugsvnlog'):
            return directlogstream(*args)
        arg = encodeargs(args)
        hgexe = procutil.hgexecutable()

        stdin, stdout = procutil.popen2(procutil.quotecommand(cmd))
        stdin.write(arg)
        try:
            stdin.close()
        except IOError:
            raise error.Abort(
                _(
                    'Mercurial failed to run itself, check'
                    ' hg executable is in PATH'
                )
            )
        return logstream(stdout)


pre_revprop_change = b'''#!/bin/sh

REPOS="$1"
REV="$2"

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(br'Committed revision (\d+).', re.M)
    uuid_re = re.compile(br'Repository UUID:\s*(\S+)', re.M)

        else:
            if not re.search(br'^(file|http|https|svn|svn\+ssh)\://', path):
                path = os.path.realpath(path)
                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 '%s'\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

            wcpath = os.path.join(
                encoding.getcwd(), os.path.basename(path) + '-wc'
            )
            ui.status(
                _("initializing svn working copy '%s'\n")
                % os.path.basename(wcpath)
            )
            self.run0('checkout', path, wcpath)

            self.wc = wcpath
        self.opener = vfsmod.vfs(self.wc)
        self.wopener = vfsmod.vfs(self.wc)

        doc = xml.dom.minidom.parseString(output)
        for e in doc.getElementsByTagName(r'entry'):
            for n in e.childNodes:
                if n.nodeType != n.ELEMENT_NODE or n.tagName != r'name':
                    continue
                name = r''.join(
                    c.data for c in n.childNodes if c.nodeType == c.TEXT_NODE
                )
                # Entries are compared with names coming from
                # mercurial, so bytes with undefined encoding. Our
                # best bet is to assume they are in local
                # encoding. They will be passed to command line calls
                # later anyway, so they better be.

        # already occurred.  Cross the semantic gap.
        wdest = self.wjoin(dest)
        exists = os.path.lexists(wdest)
        if exists:
            fd, tempname = pycompat.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)

            for i in iter(lambda: f.rfind('/', 0, i), -1):
                dirs.add(f[:i])
        return dirs

    def add_dirs(self, files):
        add_dirs = [
            d for d in sorted(self.dirs_of(files)) if d not in self.manifest
        ]
        if add_dirs:
            self.manifest.update(add_dirs)
            self.xargs(add_dirs, 'add', non_recursive=True, quiet=True)
        return add_dirs

        self.childmap[parent] = child

    def revid(self, rev):
        return "svn:%s@%s" % (self.uuid, rev)

    def putcommit(
        self, files, copies, parents, commit, source, revmap, full, cleanp2
    ):
        for parent in parents:
            try:
                return self.revid(self.childmap[parent])
            except KeyError:
                pass

        fd, messagefile = pycompat.mkstemp(prefix='hg-convert-')
        fp = os.fdopen(fd, r'wb')
        fp.write(util.tonativeeol(commit.desc))
        fp.close()
        try:
            output = self.run0(
                'commit',
                username=stringutil.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] if parents else 'None'
                self.ui.warn(_('unexpected svn output:\n'))
                self.ui.warn(output)
                raise error.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 hascommitforsplicemap(self, rev):
        # This is not correct as one can convert to an existing subversion
        # repository and childmap would not list all revisions. Too bad.
        if rev in self.childmap:
            return True
        raise error.Abort(
            _(
                'splice map revision %s not found in subversion '
                'child map (revision lookups are not implemented)'
            )
            % rev
        )