Fix Doctype tokens generated by tree walkers wrt publicId and systemId attributes (not all trees support these attributes)

tbroyer · tbroyer · commit fcf0d4a2a6b7 · 2007-07-09T18:41:08.000Z
--HG--
extra : convert_revision : svn%3Aacbfec75-9323-0410-a652-858a13e371e0/trunk%40881
diff --git a/src/html5lib/treewalkers/_base.py b/src/html5lib/treewalkers/_base.py
@@ -51,9 +51,10 @@ def text(self, data):
     def comment(self, data):
         return {"type": "Comment", "data": unicode(data)}
 
-    def doctype(self, name):
-        return {"type": "Doctype", "name": unicode(name),
-                "data": name and name.upper() == "HTML"}
+    def doctype(self, name, publicId=None, systemId=None, correct=True):
+        return {"type": "Doctype", "name": name and unicode(name) or None,
+                "publicId": publicId, "systemId": systemId,
+                "correct": correct}
 
     def unknown(self, nodeType):
         return self.error(_("Unknown node type: ") + nodeType)
diff --git a/src/html5lib/treewalkers/dom.py b/src/html5lib/treewalkers/dom.py
@@ -10,7 +10,7 @@
 class TreeWalker(_base.NonRecursiveTreeWalker):
     def getNodeDetails(self, node):
         if node.nodeType == Node.DOCUMENT_TYPE_NODE:
-            return _base.DOCTYPE, node.nodeName or ''
+            return _base.DOCTYPE, node.name, node.publicId, node.systemId
 
         elif node.nodeType in (Node.TEXT_NODE, Node.CDATA_SECTION_NODE):
             return _base.TEXT, node.nodeValue
diff --git a/src/html5lib/treewalkers/genshistream.py b/src/html5lib/treewalkers/genshistream.py
@@ -57,7 +57,7 @@ def tokens(self, event, next):
                 yield token
 
         elif kind == DOCTYPE:
-            yield self.doctype(data[0])
+            yield self.doctype(*data)
 
         elif kind in (XML_DECL, DOCTYPE, START_NS, END_NS, \
           START_CDATA, END_CDATA, PI):
diff --git a/src/html5lib/treewalkers/simpletree.py b/src/html5lib/treewalkers/simpletree.py
@@ -26,7 +26,7 @@ def getNodeDetails(self, node):
             return (_base.DOCUMENT,)
 
         elif node.type == 3: # DocumentType
-            return _base.DOCTYPE, node.name
+            return _base.DOCTYPE, node.name, node.publicId, node.systemId
 
         elif node.type == 4: # TextNode
             return _base.TEXT, node.value
