convert/svn: stop returning unicode revision identifiers

hgext/convert/subversion.py

@@ -10,7 +10,7 b' import cPickle as pickle'
 import tempfile
 import urllib
 
-from mercurial import strutil, util
+from mercurial import strutil, util, encoding
 from mercurial.i18n import _
 
 # Subversion stuff. Works best with very recent Python SVN bindings
@@ -50,6 +50,9 b' def geturl(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
 
@@ -226,7 +229,7 b' class svn_source(converter_source):'
             self.rootmodule = self.module
             self.commits = {}
             self.paths = {}
-            self.uuid = svn.ra.get_uuid(self.ra).decode(self.encoding)
+            self.uuid = svn.ra.get_uuid(self.ra)
         except SubversionException:
             ui.traceback()
             raise NoRepo("%s does not look like a Subversion repo" % self.url)
@@ -253,8 +256,8 b' class svn_source(converter_source):'
 
         self.head = self.latest(self.module, latest)
         if not self.head:
-            raise util.Abort(_('no revision found in module %s') %
-                             self.module.encode(self.encoding))
+            raise util.Abort(_('no revision found in module %s')
+                             % self.module)
         self.last_changed = self.revnum(self.head)
 
         self._changescache = None
@@ -314,8 +317,8 b' class svn_source(converter_source):'
             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.encode(self.encoding))
+                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]
@@ -333,8 +336,7 b' class svn_source(converter_source):'
                     continue
                 brevid = self.latest(module, self.last_changed)
                 if not brevid:
-                    self.ui.note(_('ignoring empty branch %s\n') %
-                                   branch.encode(self.encoding))
+                    self.ui.note(_('ignoring empty branch %s\n') % branch)
                     continue
                 self.ui.note(_('found branch %s at %d\n') %
                              (branch, self.revnum(brevid)))
@@ -511,16 +513,13 b' class svn_source(converter_source):'
         self.convertfp.flush()
 
     def revid(self, revnum, module=None):
-        if not module:
-            module = self.module
-        return u"svn:%s%s@%s" % (self.uuid, module.decode(self.encoding),
-                                 revnum)
+        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.encode(self.encoding).rsplit('@', 1)
+        url, revnum = rev.rsplit('@', 1)
         revnum = int(revnum)
         parts = url.split('/', 1)
         uuid = parts.pop(0)[4:]
@@ -786,7 +785,7 b' class svn_source(converter_source):'
                           desc=log,
                           parents=parents,
                           branch=branch,
-                          rev=rev.encode('utf-8'))
+                          rev=rev)
 
             self.commits[rev] = cset
             # The parents list is *shared* among self.paths and the

tests/svn/encoding.svndump

@@ -1,6 +1,6 b''
 SVN-fs-dump-format-version: 2
 
-UUID: 64cf20ed-79ff-4f10-8aaf-041d1993b42e
+UUID: 7ec0009b-ac71-43a6-9416-7fc295ee5ca1
 
 Revision-number: 0
 Prop-content-length: 56
@@ -9,7 +9,7 b' Content-length: 56'
 K 8
 svn:date
 V 27
-2009-06-21T10:05:28.031033Z
[phone]T15:50:22.535630Z
 PROPS-END
 
 Revision-number: 1
@@ -27,7 +27,7 b' pmezard'
 K 8
 svn:date
 V 27
-2009-06-21T10:05:28.108227Z
[phone]T15:50:22.584588Z
 PROPS-END
 
 Node-path: branches
@@ -72,7 +72,7 b' pmezard'
 K 8
 svn:date
 V 27
-2009-06-21T10:05:29.109026Z
[phone]T15:50:23.143007Z
 PROPS-END
 
 Node-path: trunk/à
@@ -125,7 +125,7 b' pmezard'
 K 8
 svn:date
 V 27
-2009-06-21T10:05:32.061964Z
[phone]T15:50:26.088415Z
 PROPS-END
 
 Node-path: trunk/è
@@ -167,7 +167,7 b' pmezard'
 K 8
 svn:date
 V 27
-2009-06-21T10:05:33.096353Z
[phone]T15:50:27.162784Z
 PROPS-END
 
 Node-path: trunk/è
@@ -178,3 +178,53 b' Node-path: trunk/\xc3\xb9'
 Node-action: delete
 
 
+Revision-number: 5
+Prop-content-length: 120
+Content-length: 120
+
+K 7
+svn:log
+V 18
+branch to branché
+K 10
+svn:author
+V 7
+pmezard
+K 8
+svn:date
+V 27
[phone]T15:50:30.079728Z
+PROPS-END
+
+Node-path: branches/branché
+Node-kind: dir
+Node-action: add
+Node-copyfrom-rev: 4
+Node-copyfrom-path: trunk
+
+
+Revision-number: 6
+Prop-content-length: 121
+Content-length: 121
+
+K 7
+svn:log
+V 19
+branch to branchée
+K 10
+svn:author
+V 7
+pmezard
+K 8
+svn:date
+V 27
[phone]T15:50:33.072456Z
+PROPS-END
+
+Node-path: branches/branchée
+Node-kind: dir
+Node-action: add
+Node-copyfrom-rev: 5
+Node-copyfrom-path: branches/branché
+
+

tests/svn/svndump-encoding.sh

@@ -35,6 +35,16 b' svn ci -m "copy files"'
 svn rm trunk/è
 svn rm trunk/ù
 svn ci -m 'remove files'
+
+# Create branches with and from weird names
+svn up
+svn cp trunk branches/branché
+echo a > branches/branché/a
+svn ci -m 'branch to branché'
+svn up
+svn cp branches/branché branches/branchée
+echo a >> branches/branché/a
+svn ci -m 'branch to branchée'
 cd ..
 
 svnadmin dump svn-repo > ../encoding.svndump