img.complete test has bogus loop thing

[bzbarsky]

The test at html/semantics/embedded-content/the-img-element/img.complete.html does this:

async_test(function(t) {
  assert_false(imageInstance.complete, "imageInstance.complete should be false");
  var startTime = Date.now();
  while (true) {
    if (Date.now() - startTime > 2000)
      assert_unreached('.complete didn\'t change to true');
    if (imageInstance.complete === true) break;
  }
  t.done();

This can't possibly pass, because the value of "complete" cannot change while in a busy-loop in JS. I'm not sure what this test is trying to test, but it's bogus.

[bzbarsky]

Looks like @deniak committed this as part of "move TTWF tests". No useful info on where it was moved from and why the test is trying to do this... :(

[jdm]

The spec claims that complete should reflect whether a particular task has been queued, which is how this is intended to pass: https://html.spec.whatwg.org/multipage/embedded-content.html#dom-img-complete

[bzbarsky]

That's ... not OK. That information is not generally available and shouldn't be. Filed whatwg/html#2038

[foolip]

Spec bug is now whatwg/html#1055.

[annevk]

Fixed by dc2e245.