Re: [Navigation Timing] test_document_readiness_exist.htm

Hi, Karen,

   The test was add per one of the AIs from TPAC 2011:
http://www.w3.org/2010/webperf/track/actions/68  It was added to make sure
the corresponding readiness states exist as Navigation Timing has
dependency on them. Unfortunately the existing test might not be
the best way to ensure that for the reasons you mentioned...

cheers,
Zhiheng

On Wed, Apr 25, 2012 at 11:22 AM, Karen Anderson (IE) <
Karen.Anderson@microsoft.com> wrote:

>  Hi All,
>
>
>
> IE has failures in the document readiness test (
> http://w3c-test.org/webperf/tests/approved/navigation-timing/html5/test_document_readiness_exist.html)
> and I've been looking into how the browsers are handling readiness events
> and it isn't consistent.  This was actually brought up a long time ago by
> Tony:
> http://lists.w3.org/Archives/Public/public-web-perf/2010Jul/0016.html.
> My additional homework has found that some of the inconsistencies is also
> around how the different browsers handle the onreadystatechange event for
> subdocuments.
>
>
>
> My concern with the current test is that most likely by the time the code
> is running on the page, we have already missed the loading phase.  If we
> change the test to use an iframe, then we run into the discrepancies of how
> the different browsers fire the onreadystatechange event on subdocuments.
>
>
>
> Recalling history, adding the dom* events to the performance object was to
> give a consistent story on these events when one didn't already exist.
> Creating a test that calls out this discrepancy seems odd to me.  I think
> given that, I think we should remove the test from the suite.  The timing
> order test covers that the dom* events include timestamps and in the
> correct location in the timeline, so we are covered from that perspective.
>
>
>
> What are your thoughts?
>
>
>
> Thanks,
>
> Karen
>