Bug 16799 – Acid3 expects HTMLObjectElement.data to canonicalize urls

• Component: DOM
• Status: RESOLVED FIXED

[Eric Seidel (no email)]

Test 64: FAIL (expected ./test.html, got: test.html - object elements didn't resolve URIs correctly)

    function () {
      // test 64: attributes of the <object> element
      var obj1 = document.createElement('object');
      obj1.setAttribute('data', 'test.html');
      var obj2 = document.createElement('object');
      obj2.setAttribute('data', './test.html');
      assertEquals(obj1.data, obj2.data, "object elements didn't resolve URIs correctly");
      assert(obj1.data.match(/^http:/), "object.data isn't absolute");
      obj1.appendChild(document.createElement('param'));
      assertEquals(obj1.getElementsByTagName('param').length, 1, "object is missing its only child");
      return 4;
    },

[Robert Blaut]

Created attachment 18735 [details]
Minimal test case

[Dave Hyatt]

Created attachment 18921 [details]
Make data absolute

Will land test case in bug.

[Mark Rowe (bdash)]

Comment on attachment 18921 [details]
Make data absolute

r=me

[Darin Adler]

Comment on attachment 18921 [details]
Make data absolute

+    if (dataVal.isNull())
+        return dataVal;

You don't need this. Document::completeURL already does this. It can just be a one-liner.

    return document()->completeURL(getAttribute(dataAttr));

[Darin Fisher (:fishd, Google)]

*** Bug 14448 has been marked as a duplicate of this bug. ***

[Darin Fisher (:fishd, Google)]

you might want to review my patch in bug 14448, which includes a test case, but more importantly it had to be backed out after it was found to regress some existing layout tests!

[Dave Hyatt]

I will just remove those layout tests, as they make a bogus assumption.

[Dave Hyatt]

Fixed in r30044.