Teach lint & treewalkers that elements are only void in HTML ns

Author: gsnedders

html5lib/filters/lint.py

@@ -3,7 +3,7 @@
 from six import text_type
 
 from . import _base
-from ..constants import voidElements
+from ..constants import namespaces, voidElements
 
 from ..constants import spaceCharacters
 spaceCharacters = "".join(spaceCharacters)
@@ -19,17 +19,22 @@ def __iter__(self):
         for token in _base.Filter.__iter__(self):
             type = token["type"]
             if type in ("StartTag", "EmptyTag"):
+                namespace = token["namespace"]
                 name = token["name"]
+                if namespace is not None and not isinstance(namespace, text_type):
+                    raise LintError("Tag namespace is not a string or None: %(name)r" % {"name": namespace})
+                if namespace == "":
+                    raise LintError("Empty tag namespace")
                 if not isinstance(name, text_type):
                     raise LintError("Tag name is not a string: %(tag)r" % {"tag": name})
                 if not name:
                     raise LintError("Empty tag name")
-                if type == "StartTag" and name in voidElements:
+                if type == "StartTag" and (not namespace or namespace == namespaces["html"]) and name in voidElements:
                     raise LintError("Void element reported as StartTag token: %(tag)s" % {"tag": name})
-                elif type == "EmptyTag" and name not in voidElements:
+                elif type == "EmptyTag" and (not namespace or namespace == namespaces["html"]) and name not in voidElements:
                     raise LintError("Non-void element reported as EmptyTag token: %(tag)s" % {"tag": token["name"]})
                 if type == "StartTag":
-                    open_elements.append(name)
+                    open_elements.append((namespace, name))
                 for (namespace, localname), value in token["data"].items():
                     if namespace is not None and not isinstance(namespace, text_type):
                         raise LintError("Attribute namespace is not a string or None: %(name)r" % {"name": namespace})
@@ -43,15 +48,20 @@ def __iter__(self):
                         raise LintError("Attribute value is not a string: %(value)r" % {"value": value})
 
             elif type == "EndTag":
+                namespace = token["namespace"]
                 name = token["name"]
+                if namespace is not None and not isinstance(namespace, text_type):
+                    raise LintError("Tag namespace is not a string or None: %(name)r" % {"name": namespace})
+                if namespace == "":
+                    raise LintError("Empty tag namespace")
                 if not isinstance(name, text_type):
                     raise LintError("Tag name is not a string: %(tag)r" % {"tag": name})
                 if not name:
                     raise LintError("Empty tag name")
-                if name in voidElements:
+                if (not namespace or namespace == namespaces["html"]) and name in voidElements:
                     raise LintError("Void element reported as EndTag token: %(tag)s" % {"tag": name})
                 start_name = open_elements.pop()
-                if start_name != name:
+                if start_name != (namespace, name):
                     raise LintError("EndTag (%(end)s) does not match StartTag (%(start)s)" % {"end": name, "start": start_name})
 
             elif type == "Comment":

html5lib/treewalkers/_base.py

@@ -2,7 +2,7 @@
 from six import text_type, string_types
 
 from xml.dom import Node
-from ..constants import voidElements, spaceCharacters
+from ..constants import namespaces, voidElements, spaceCharacters
 
 __all__ = ["DOCUMENT", "DOCTYPE", "TEXT", "ELEMENT", "COMMENT", "ENTITY", "UNKNOWN",
            "TreeWalker", "NonRecursiveTreeWalker"]
@@ -154,7 +154,7 @@ def __iter__(self):
 
             elif type == ELEMENT:
                 namespace, name, attributes, hasChildren = details
-                if name in voidElements:
+                if (not namespace or namespace == namespaces["html"]) and name in voidElements:
                     for token in self.emptyTag(namespace, name, attributes,
                                                hasChildren):
                         yield token
@@ -187,7 +187,7 @@ def __iter__(self):
                     type, details = details[0], details[1:]
                     if type == ELEMENT:
                         namespace, name, attributes, hasChildren = details
-                        if name not in voidElements:
+                        if (namespace and namespace != namespaces["html"]) or name not in voidElements:
                             yield self.endTag(namespace, name)
                     if self.tree is currentNode:
                         currentNode = None

html5lib/treewalkers/genshistream.py

@@ -48,7 +48,7 @@ def tokens(self, event, next):
         elif kind == END:
             name = data.localname
             namespace = data.namespace
-            if name not in voidElements:
+            if namespace != namespaces["html"] or name not in voidElements:
                 yield self.endTag(namespace, name)
 
         elif kind == COMMENT: