Changeset 143830 in webkit

Timestamp:

Feb 22, 2013 7:28:41 PM (11 years ago)

Author:

tonyg@chromium.org

Message:

Threaded HTML parser fails resources/plain-text-unsafe.dat
​https://bugs.webkit.org/show_bug.cgi?id=110538

Reviewed by Eric Seidel.

With this patch, the background HTML parser passes all tests in plain-text-unsafe.dat.

No new tests because covered by existing tests.

• html/parser/BackgroundHTMLParser.cpp:

(WebCore::tokenExitsSVG):
(WebCore):
(WebCore::tokenExitsMath):
(WebCore::BackgroundHTMLParser::simulateTreeBuilder):

Location:

trunk/Source/WebCore

Files:

• ChangeLog (modified) (1 diff)
• html/parser/BackgroundHTMLParser.cpp (modified) (5 diffs)

trunk/Source/WebCore/ChangeLog

@@ +1 @@
+2013-02-22  Tony Gentilcore  <tonyg@chromium.org>
+
+        Threaded HTML parser fails resources/plain-text-unsafe.dat
+        https://bugs.webkit.org/show_bug.cgi?id=110538
+
+        Reviewed by Eric Seidel.
+
+        With this patch, the background HTML parser passes all tests in plain-text-unsafe.dat.
+
+        No new tests because covered by existing tests.
+
+        * html/parser/BackgroundHTMLParser.cpp:
+        (WebCore::tokenExitsSVG):
+        (WebCore):
+        (WebCore::tokenExitsMath):
+        (WebCore::BackgroundHTMLParser::simulateTreeBuilder):
+
 2013-02-22  Tony Gentilcore  <tonyg@chromium.org>
 

trunk/Source/WebCore/html/parser/BackgroundHTMLParser.cpp

@@ -62 +62 @@
 #endif
 
-static inline bool tokenExitsForeignContent(const CompactHTMLToken& token)
+static bool tokenExitsForeignContent(const CompactHTMLToken& token)
 {
     // FIXME: This is copied from HTMLTreeBuilder::processTokenInForeignContent and changed to use threadSafeMatch.
@@ -113 +113 @@
 }
 
+static bool tokenExitsSVG(const CompactHTMLToken& token)
+{
+    const String& tagName = token.data();
+    return equalIgnoringCase(tagName, SVGNames::foreignObjectTag.localName());
+}
+
+static bool tokenExitsMath(const CompactHTMLToken& token)
+{
+    // FIXME: This is copied from HTMLElementStack::isMathMLTextIntegrationPoint and changed to use threadSafeMatch.
+    const String& tagName = token.data();
+    return threadSafeMatch(tagName, MathMLNames::miTag)
+        || threadSafeMatch(tagName, MathMLNames::moTag)
+        || threadSafeMatch(tagName, MathMLNames::mnTag)
+        || threadSafeMatch(tagName, MathMLNames::msTag)
+        || threadSafeMatch(tagName, MathMLNames::mtextTag);
+}
+
 static const size_t pendingTokenLimit = 1000;
 
@@ -180 +197 @@
         if (inForeignContent() && tokenExitsForeignContent(token))
             m_namespaceStack.removeLast();
-        // FIXME: Support tags that exit MathML.
-        if (m_namespaceStack.last() == SVG && equalIgnoringCase(tagName, SVGNames::foreignObjectTag.localName()))
+        if ((m_namespaceStack.last() == SVG && tokenExitsSVG(token))
+            || (m_namespaceStack.last() == MathML && tokenExitsMath(token)))
             m_namespaceStack.append(HTML);
         if (!inForeignContent()) {
@@ -203 +220 @@
     if (token.type() == HTMLToken::EndTag) {
         const String& tagName = token.data();
-        // FIXME: Support tags that exit MathML.
         if ((m_namespaceStack.last() == SVG && threadSafeMatch(tagName, SVGNames::svgTag))
             || (m_namespaceStack.last() == MathML && threadSafeMatch(tagName, MathMLNames::mathTag))
-            || (m_namespaceStack.contains(SVG) && m_namespaceStack.last() == HTML && equalIgnoringCase(tagName, SVGNames::foreignObjectTag.localName())))
+            || (m_namespaceStack.contains(SVG) && m_namespaceStack.last() == HTML && tokenExitsSVG(token))
+            || (m_namespaceStack.contains(MathML) && m_namespaceStack.last() == HTML && tokenExitsMath(token)))
             m_namespaceStack.removeLast();
         if (threadSafeMatch(tagName, scriptTag)) {
@@ -215 +232 @@
     }
 
-    // FIXME: Need to set setForceNullCharacterReplacement based on m_inForeignContent as well.
+    // FIXME: Also setForceNullCharacterReplacement when in text mode.
+    m_tokenizer->setForceNullCharacterReplacement(inForeignContent());
     m_tokenizer->setShouldAllowCDATA(inForeignContent());
     return true;