RE: [minutes] 2012-01-25 Web Performance WG Teleconference #59

I also meant to comment on the document readiness test: http://w3c-test.org/webperf/tests/approved/navigation-timing/html5/test_document_readiness_exist.html

It looks like the test has removed the frame which is a good starting point.  My comment last week about the readyState phases wasn’t that IE didn’t implement them, but instead that we do not notify all events for iframes when accessed from the root document.  By the time that the root document has access to the frame, it is already loaded.  Thus, only the "complete" state will exist.

IE does have all of the states for the root document.  Changing the test to use addEventListener shows all of the states, with the exception of "uninitialized" which we don't test for, but seems that the only way to get into this state would be to do object injection.  We probably don't care about this corner case as that state should be well before navigationStart.

Also, line 21 is not correct.  It is an assignment, not a comparison:
    if (document.readyState = "complete")

I think the AI is already closed, though can we reopen and make these changes?

Thanks,
Karen

From: Karen Anderson (IE) [mailto:Karen.Anderson@microsoft.com]
Sent: Wednesday, January 25, 2012 2:39 PM
To: Jatinder Mann; public-web-perf@w3.org
Subject: RE: [minutes] 2012-01-25 Web Performance WG Teleconference #59

Sorry I was not able to attend our call this morning.  I am a little confused talking point 1.b in the notes:

The test test_timing_attributes_order.htm regarding using sleeps in the unload handlers of the previous document are disallowed for Chrome as James pointed out on the mailing list.  It looks like the decision in today’s call was to just to verify the order and not the time of the unload events.  However, there was a past thread (attached) on the desire to have a test to ensure that the time between unloadEventStart/End only captured the onunload and not the onbeforeunload event.  What changed our minds?

Regardless of whether we want to separate the time between onbeforeunload and onunload, we still need a method of generating a delay within the unload sequence.  How do we feel about using an XHR request to a PHP script that performs this logic on the server instead?  Any other ideas?

Thanks,
Karen

From: Jatinder Mann [mailto:jmann@microsoft.com]
Sent: Wednesday, January 25, 2012 10:50 AM
To: public-web-perf@w3.org
Subject: [minutes] 2012-01-25 Web Performance WG Teleconference #59


Meeting Summary:



1.       Navigation Timing to PR

a.       Closed ACTION 77 - Update the test_navigation_type_reload.html to fix the compare issue

The Working Group reviewed this test and agree to move the updated test to the submission folder and close this work item.



b.       Test test_timing_attributes_order.html needs update

James has reviewed the test_timing_attributes_order.html test and has verified that it isn’t a valid test for Chrome as sleeps are not allowed in loops in the unload events. The Working Group has agreed to just test the order and not worry about the duration. Karen to make updates.



c.        Navigation Timing to PR

The last remaining change prior to Navigation Timing going to PR is to update the test_timing_attributes_order.html to remove the sleeps and duration test. Considering this is a simple change, we are targeting getting this done by end of the week. As soon as this test change has been made, we are good to move this spec to PR.



2.       Resource Timing to CR

a.       ACTION-55 needs to be closed

Jatinder is working on closing ACTION-55 prior to moving this spec to CR.



3.       Performance Timeline to CR

a.       ACTION-63 needs to be closed

Jatinder is working on closing ACTION-63 prior to moving this spec to CR.



b.       High Resolution Time Spec

Jatinder is working on a High Resolution Time spec that would replace the simple definition of the time base currently found in the Performance Timeline spec.



4.       Page Visibility to CR

a.       Working Group has agreed to move this specification to CR

The Working Group has agreed that all open issues with this spec have been closed and would like to move this spec to the CR phase. Philippe is working on moving this spec to CR.



b.       Page Visibility Test Suite

Karen is working on submitting a complete Page Visibility test suite in the next two weeks.



5.       requestAnimationFrame to CR

a.       Move spec to LC

Considering there are no remaining issues open on this specification, the Working Group would like to know if this specification can be moved to the Last Call phase. We will ping the editors to see if they feel there are no remaining issues.



b.       requestAnimationFrame Test Suite

Microsoft can submit a requestAnimationFrame test suite in the near future.



6.       HAR Format Specification and Rechartering WG

We had last discussed rechartering the Web Perf WG to include the HAR format specification. Prior to the rechartering effort, we were hoping to see a spec draft. Arvind, what is the status on the HAR format spec draft?




Detailed Notes:



Web Perf Teleconference #59 1/25/2012



IRC log: http://www.w3.org/2012/01/25-webperf-irc


Meeting Minutes: http://www.w3.org/2012/01/25-webperf-minutes.html



Attendees

Present for Navigation Timing, Resource Timing and User Timing (4-5PM EST/1-2PM PST)

Philippe Le Hegaret, Jatinder Mann, Arvind Jain, Karen Anderson, Zhiheng Wang, James Simonsen, Ganesh Rao


Present for Page Visibility, Efficient Script Yielding, Display Paint Notifications (4-5PM EST/2-3PM PST)

Meeting cancelled.



Scribe

Jatinder Mann



Contents

Agenda

1.       Discuss all specification status



--------------------------------------------------------------------------------



http://w3c-test.org/webperf/tests/approved/navigation-timing/html5/test_navigation_type_reload.html

Jatinder: Per Action 77, James updated navigation_type_reload.html. It works in IE9 and IE10.

plh: Are we good with pushing this to approved?

JatindeR: Yes.

Tony: Yes.

close action-77

<trackbot> ACTION-77 Update the test_navigation_type_reload.html to fix the compare issue closed

<plh> http://w3c-test.org/webperf/tests/submission/Google/NavigationTiming/test_document_readiness_exist.html

<plh> http://w3c-test.org/webperf/tests/submission/Google/NavigationTiming/test_performance_attributes_exist_in_object.html

Zhiheng: Can we move these tests to approved?
... They work in IE and Chrome.

Jatinder: And Karen responded saying document readiness doesn't work as expected on iframe. This shouldn't impact Navigation timing.

Plh: Yes, let's move them to approved.

<plh> CR issues list<http://www.w3.org/2012/01/navigation-cr-issues.html>

Tony: James provided feedback on the test_timing_attributes_order.html on the mailing list. We should just test the order.

Jatinder: I will work with Karen to get that test updated. This will be the last update needed for PR to move forward.

Plh: Considering Page Visibility is stable, we should move it forward to CR. Do we have any other feedback on that spec?

Tony: No, it appears all feedback has been met. We should be able to move it forward.

Jatinder: Yes, let's move to CR. I will work with Karen to have the Page Visibility test suite available in two weeks.

plh: We should consider moving requestAnimationFrame to last call. We should ping James/Cameron to see if there are any other issues.

JatindeR: I can send mail asking James/Cameron if there are any other issues. If not, we can move this one to LC and then to CR shortly thereafter.

plh: Is anyone working on a test suite for requestAnimationFrame?

Jatinder: I can work on getting a test suite for requestAnimationFrame together.

plh: What about the HAR format? Any updates?

Jatinder: I believe we wanted a draft spec prior to re-chartering. I will ask Arvind the question in the meeting minutes.