reorder some methods to match the scheme set up in Phase and add a comment about a potential merging of the initial and root element phases

annevk · annevk · commit 94de3bd490e7 · 2006-12-29T23:05:03.000Z
--HG--
extra : convert_revision : svn%3Aacbfec75-9323-0410-a652-858a13e371e0/trunk%40375
diff --git a/src/parser.py b/src/parser.py
@@ -192,16 +192,24 @@ def processEndTag(self, name):
 
 
 class InitialPhase(Phase):
-    # XXX This phase deals with error handling as well which is currently not
-    # in the specification.
-
-    def processDoctype(self, name, error):
-        self.tree.insertDoctype(name)
+    # This phase deals with error handling as well which is currently not
+    # covered in the specification. The error handling is typically known as
+    # "quirks mode". It is expected that a future version of HTML5 will defin
+    # this.
+    #
+    # AT Given that. I think it should be possible to merge this phase with the
+    # root element phase in due course.
+    def processEOF(self):
         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):
+        self.tree.insertDoctype(name)
+        self.parser.phase = self.parser.phases["rootElement"]
+
     def processSpaceCharacters(self, data):
         self.tree.insertText(data, self.tree.document)
 
@@ -218,10 +226,6 @@ def processEndTag(self, name):
         self.parser.phase = self.parser.phases["rootElement"]
         self.parser.phase.processEndTag(name)
 
-    def processEOF(self):
-        self.parser.phase = self.parser.phases["rootElement"]
-        self.parser.phase.processEOF()
-
 
 class RootElementPhase(Phase):
     # helper methods
@@ -232,6 +236,10 @@ def insertHtmlElement(self):
         self.parser.phase = self.parser.phases["beforeHead"]
 
     # other
+    def processEOF(self):
+        self.insertHtmlElement()
+        self.parser.phase.processEOF()
+
     def processComment(self, data):
         self.tree.insertComment(data, self.tree.document)
 
@@ -250,10 +258,6 @@ def processEndTag(self, name):
         self.insertHtmlElement()
         self.parser.phase.processEndTag(name)
 
-    def processEOF(self):
-        self.insertHtmlElement()
-        self.parser.phase.processEOF()
-
 
 class BeforeHeadPhase(Phase):
     def __init__(self, parser, tree):
@@ -1585,12 +1589,12 @@ def endTagOther(self, name):
 
 
 class TrailingEndPhase(Phase):
-    def processComment(self, data):
-        self.parser.insertCommenr(data, self.tree.document)
-
     def processEOF(self):
         pass
 
+    def processComment(self, data):
+        self.parser.insertCommenr(data, self.tree.document)
+
     def processSpaceCharacters(self, data):
         self.parser.lastPhase.processCharacters(data)
 
diff --git a/src/tokenizer.py b/src/tokenizer.py
@@ -406,7 +406,6 @@ def closeTagOpenState(self):
             elif data == EOF:
                 self.tokenQueue.append({"type": "ParseError"})
                 self.tokenQueue.append({"type": "Characters", "data": u"</"})
-                self.stream.queue.append(data)
                 self.state = self.states["data"]
             else:
                 self.tokenQueue.append({"type": "ParseError"})
@@ -599,7 +598,7 @@ def commentState(self):
         if data == u"-":
             self.state = self.states["commentDash"]
         elif data == EOF:
-            self.emitCurrentTokenWithParseError(data)
+            self.emitCurrentTokenWithParseError()
         else:
             self.currentToken["data"] += data + self.stream.charsUntil(u"-")
         return True
@@ -609,7 +608,7 @@ def commentDashState(self):
         if data == u"-":
             self.state = self.states["commentEnd"]
         elif data == EOF:
-            self.emitCurrentTokenWithParseError(data)
+            self.emitCurrentTokenWithParseError()
         else:
             self.currentToken["data"] += u"-" + data +\
               self.stream.charsUntil(u"-")
@@ -627,7 +626,7 @@ def commentEndState(self):
             self.tokenQueue.append({"type": "ParseError"})
             self.currentToken["data"] += data
         elif data == EOF:
-            self.emitCurrentTokenWithParseError(data)
+            self.emitCurrentTokenWithParseError()
         else:
             self.tokenQueue.append({"type": "ParseError"})
             self.currentToken["data"] += u"--" + data
