Re-enable Genshi treewalker.

gsnedders · gsnedders · commit 6bc4fc4ecb43 · 2013-04-13T15:31:12.000+01:00
The Genshi treewalker was never fully updated when support for
namespaces was added to treewalkers. This finally fixes this, allowing
all tests to once again pass.

This removes the ignoring of children of void elements as the commonly
used treewalkers don't support this and there was some subtle bug I
couldn't see in the code. Given most don't support this, this seems
like little loss. (You will, however, still get an error from the
treewalker if a void element has children.)
diff --git a/html5lib/tests/test_treewalkers.py b/html5lib/tests/test_treewalkers.py
@@ -127,15 +127,19 @@ def GenshiAdapter(tree):
                     name = "{%s}%s" % (token["namespace"], token["name"])
                 else:
                     name = token["name"]
-                yield (START,
-                       (QName(name),
-                        Attrs([(QName(attr),value) for attr,value in token["data"]])),
-                       (None, -1, -1))
+                attrs = Attrs([(QName("{%s}%s" % attr if attr[0] is not None else attr[1]), value)
+                               for attr, value in token["data"].items()])
+                yield (START, (QName(name), attrs), (None, -1, -1))
                 if type == "EmptyTag":
                     type = "EndTag"
 
             if type == "EndTag":
-                yield END, QName(token["name"]), (None, -1, -1)
+                if token["namespace"]:
+                    name = "{%s}%s" % (token["namespace"], token["name"])
+                else:
+                    name = token["name"]
+
+                yield END, QName(name), (None, -1, -1)
 
             elif type == "Comment":
                 yield COMMENT, token["data"], (None, -1, -1)
@@ -150,10 +154,10 @@ def GenshiAdapter(tree):
         if text is not None:
             yield TEXT, text, (None, -1, -1)
 
-    #treeTypes["genshi"] = \
-    #    {"builder": treebuilders.getTreeBuilder("simpletree"),
-    #     "adapter": GenshiAdapter,
-    #     "walker":  treewalkers.getTreeWalker("genshi")}
+    treeTypes["genshi"] = \
+       {"builder": treebuilders.getTreeBuilder("simpletree"),
+        "adapter": GenshiAdapter,
+        "walker":  treewalkers.getTreeWalker("genshi")}
 except ImportError:
     pass
 
diff --git a/html5lib/treewalkers/genshistream.py b/html5lib/treewalkers/genshistream.py
@@ -1,52 +1,49 @@
 from __future__ import absolute_import, division, unicode_literals
 
+from genshi.core import QName
 from genshi.core import START, END, XML_NAMESPACE, DOCTYPE, TEXT
 from genshi.core  import  START_NS, END_NS, START_CDATA, END_CDATA, PI, COMMENT
 from genshi.output import NamespaceFlattener
 
 from . import _base
 
-from html5lib.constants import voidElements
+from html5lib.constants import voidElements, namespaces
 
 class TreeWalker(_base.TreeWalker):
     def __iter__(self):
-        depth = 0
-        ignore_until = None
+        # Buffer the events so we can pass in the following one
         previous = None
         for event in self.tree:
             if previous is not None:
-                if previous[0] == START:
-                    depth += 1
-                if ignore_until <= depth:
-                    ignore_until = None
-                if ignore_until is None:
-                    for token in self.tokens(previous, event):
-                        yield token
-                        if token["type"] == "EmptyTag":
-                            ignore_until = depth
-                if previous[0] == END:
-                    depth -= 1
+                for token in self.tokens(previous, event):
+                    yield token
             previous = event
+
+        # Don't forget the final event!
         if previous is not None:
-            if ignore_until is None or ignore_until <= depth:
-                for token in self.tokens(previous, None):
-                    yield token
-            elif ignore_until is not None:
-                raise ValueError("Illformed DOM event stream: void element without END_ELEMENT")
+            for token in self.tokens(previous, None):
+                yield token
 
     def tokens(self, event, next):
         kind, data, pos = event
         if kind == START:
-            tag, attrib = data
+            tag, attribs = data
             name = tag.localname
             namespace = tag.namespace
-            if tag in voidElements:
-                for token in self.emptyTag(namespace, name, list(attrib),
+            converted_attribs = {}
+            for k, v in attribs:
+                if isinstance(k, QName):
+                    converted_attribs[(k.namespace, k.localname)] = v
+                else:
+                    converted_attribs[(None, k)] = v
+
+            if namespace == namespaces["html"] and name in voidElements:
+                for token in self.emptyTag(namespace, name, converted_attribs,
                                            not next or next[0] != END 
                                            or next[1] != tag):
                     yield token
             else:
-                yield self.startTag(namespace, name, list(attrib))
+                yield self.startTag(namespace, name, converted_attribs)
 
         elif kind == END:
             name = data.localname
