I object to this. I had a reason for using on*. The spec said the ontimeout attribute had to exist and map to a timeout event listener. This patch removes the test for ontimeout working.
Ok, then we need to test that differently. Per spec onfoo properties don't accept objects with handleEvent methods, only functions.
Yeah, kyle is right. I should have noticed the problem in the xhr.timeout tests.
I agree - the code should've caught that... but I think that should be a separate bug. I'd feel better if the test file itself were cloned, and the clone used the addEventListener approach.
Created attachment 610603 [details] [diff] [review] v2
To be clear, when we land bug 740069 this test will stop working if we don't do anything to it. Peterv is going to post a new patch that addresses your comments.