Emit a parse error if no DOCTYPE was seen (when going from initial to the root element phase; if the DOCTYPE has an error emit an error as well (should we?); modify Phase.processEOF to match the spec more accurate; this all should fix some testcases

annevk · annevk · commit be631552c376 · 2006-12-30T20:36:02.000Z
--HG--
extra : convert_revision : svn%3Aacbfec75-9323-0410-a652-858a13e371e0/trunk%40384
diff --git a/src/parser.py b/src/parser.py
@@ -174,10 +174,11 @@ def __init__(self, parser, tree):
 
     def processEOF(self):
         self.tree.generateImpliedEndTags()
-        if self.parser.innerHTML == False\
-          or len(self.tree.openElements) > 1:
+        if self.parser.innerHTML == True and len(self.tree.openElements) > 1:
             # XXX No need to check for "body" because our EOF handling is not
             # per specification. (Specification needs an update.)
+            #
+            # XXX Need to check this more carefully in the future.
             self.parser.parseError()
         # Stop parsing
 
@@ -212,29 +213,34 @@ class InitialPhase(Phase):
     # "quirks mode". It is expected that a future version of HTML5 will defin
     # this.
     def processEOF(self):
+        self.parser.parseError("No DOCTYPE seen.")
         self.parser.phase = self.parser.phases["rootElement"]
         self.parser.phase.processEOF()
 
     def processComment(self, data):
         self.tree.insertComment(data, self.tree.document)
 
     def processDoctype(self, name, error):
+        if error:
+            self.parser.parseError("DOCTYPE is in error.")
         self.tree.insertDoctype(name)
         self.parser.phase = self.parser.phases["rootElement"]
 
     def processSpaceCharacters(self, data):
         self.tree.insertText(data, self.tree.document)
 
     def processCharacters(self, data):
-        self.parser.parseError()
+        self.parser.parseError("No DOCTYPE seen.")
         self.parser.phase = self.parser.phases["rootElement"]
         self.parser.phase.processCharacters(data)
 
     def processStartTag(self, name, attributes):
+        self.parser.parseError("No DOCTYPE seen.")
         self.parser.phase = self.parser.phases["rootElement"]
         self.parser.phase.processStartTag(name, attributes)
 
     def processEndTag(self, name):
+        self.parser.parseError("No DOCTYPE seen.")
         self.parser.phase = self.parser.phases["rootElement"]
         self.parser.phase.processEndTag(name)
 
