[webkit-reviews] review denied: [Bug 16798] <button> does not have type=submit (Acid3 bug) : [Attachment 18433] Proposed patch

[bugzilla-daemon at webkit.org]

Darin Adler <darin at apple.com> has denied Andrew Wellington
<proton at wiretapped.net>'s request for review:
Bug 16798: <button> does not have type=submit (Acid3 bug)
http://bugs.webkit.org/show_bug.cgi?id=16798

Attachment 18433: Proposed patch
http://bugs.webkit.org/attachment.cgi?id=18433&action=edit

------- Additional Comments from Darin Adler <darin at apple.com>
+void HTMLButtonElement::attributeChanged(Attribute* attr, bool preserveDecls)
+{
+    if (attr->name() == typeAttr && !attr->value())
+	 m_type = SUBMIT;
+    HTMLGenericFormElement::attributeChanged(attr, preserveDecls);
+}

We should not add this override. Instead the code for this should just go into
parseMappedAttribute. It should just be rewritten so it sets m_type to SUBMIT
any time the value is neither RESET nor BUTTON.

I'd like to confirm that HTMLButtonElement.type really returns a different
value than getAttribute("type") in the other browsers. Accordingly, the test
case should check both type and getAttribute("type").

+debug('<span>Start test 1</span>');

I don't think these are great. We normally don't include markup in the debug
statements, and I think it's better not to. And I don't think it adds much to
the output to just number tests here. Maybe if the messages said things like
"removing type attribute" it would be better.