change parse.py to not import html5lib; fix a bug in endTagP handling; add a regression testcase for that; make a comment in _base.py clearer; add some more parseError messages

annevk · annevk · commit 8abc31b7b37e · 2007-03-01T10:37:35.000Z
--HG--
extra : convert_revision : svn%3Aacbfec75-9323-0410-a652-858a13e371e0/trunk%40511
diff --git a/parse.py b/parse.py
@@ -12,7 +12,7 @@
 from src import html5parser, liberalxmlparser
 #END RELEASE
 #RELEASE add
-from html5lib import html5parser, liberalxmlparser
+#from html5lib import html5parser, liberalxmlparser
 #END RELEASE
 
 def convertTreeDump(treedump):
diff --git a/src/html5parser.py b/src/html5parser.py
@@ -840,7 +840,8 @@ def startTagOther(self, name, attributes):
         self.tree.insertElement(name, attributes)
 
     def endTagP(self, name):
-        self.tree.generateImpliedEndTags("p")
+        if self.tree.elementInScope("p"):
+            self.tree.generateImpliedEndTags("p")
         if self.tree.openElements[-1].name != "p":
             self.parser.parseError("Unexpected end tag (p).")
         while self.tree.elementInScope("p"):
@@ -1150,7 +1151,8 @@ def startTagImplyTbody(self, name, attributes):
         self.parser.phase.processStartTag(name, attributes)
 
     def startTagTable(self, name, attributes):
-        self.parser.parseError()
+        self.parser.parseError(_(u"Unexpected start tag (table) in table "
+          u"phase. Implies end tag (table)."))
         self.parser.phase.processEndTag("table")
         if not self.parser.innerHTML:
             self.parser.phase.processStartTag(name, attributes)
@@ -1168,14 +1170,16 @@ def endTagTable(self, name):
         if self.tree.elementInScope("table", True):
             self.tree.generateImpliedEndTags()
             if self.tree.openElements[-1].name != "table":
-                self.parser.parseError()
+                self.parser.parseError(_(u"Unexpected end tag (table). "
+                  u"Expected end tag (" + self.tree.openElements[-1].name +\
+                  u")."))
             while self.tree.openElements[-1].name != "table":
                 self.tree.openElements.pop()
             self.tree.openElements.pop()
             self.parser.resetInsertionMode()
         else:
-            self.parser.parseError()
             # innerHTML case
+            self.parser.parseError()
 
     def endTagIgnore(self, name):
         self.parser.parseError(_("Unexpected end tag (" + name +\
diff --git a/src/treebuilders/_base.py b/src/treebuilders/_base.py
@@ -303,9 +303,8 @@ def generateImpliedEndTags(self, exclude=None):
         if (name in frozenset(("dd", "dt", "li", "p", "td", "th", "tr"))
             and name != exclude):
             self.openElements.pop()
-            # XXX Until someone has broven that the above breaks stuff I think
-            # we should keep it in.
-            # self.processEndTag(name)
+            # XXX This is not entirely what the specification says. We should
+            # investigate it more closely.
             self.generateImpliedEndTags(exclude)
 
     def getDocument(self):
