[minutes] Web Performance WG Teleconference #94 2013-01-23

Meeting Summary:



1.     Test Cases

The working group approved the User Timing test cases submitted by Microsoft, http://w3c-test.org/webperf/tests/submission/Microsoft/UserTiming/. These test cases will be moved to the approved folder.



I found issues with the User Timing test cases submitted by Intel, http://w3c-test.org/webperf/tests/submission/Intel/user-timing. Those test cases appear to be targeting an older version of the specs, as window.performance.mark() and window.performance.measure() had been removed from the spec prior to taking the spec to CR. I also wasn't able to find any additional test cases that weren't already covered by the Microsoft test cases. I will share that feedback with Intel.



Based on feedback, I will update the Resource Timing test cases submitted by Microsoft, http://w3c-test.org/webperf/tests/submission/Microsoft/ResourceTiming/, to not use sync XHRs. As the rest of the test cases have been reviewed, those test cases can be moved to the approved folder after this test bug fix. I will follow up on the mailing list with review feedback on the Resource Timing test cases submitted by Google, http://w3c-test.org/webperf/tests/submission/Google/resource-timing/html/.



I have fixed the bugs reported in the Page Visibility test cases submitted by Microsoft, http://w3c-test.org/webperf/tests/submission/Microsoft/PageVisibility/. As these test cases have been reviewed by Boris, we should be able to move these test cases to the approved folder. I will add one additional test case testing the timing of firing the visibilitychange event and the change of state of the document.hidden and document.visibilityState properties.





2.     Resource Timing Duplicate Resources

The working group has agreed to update the Resource Timing spec to clarify that for the case where two requests in different documents ask for the same resource, the time spent waiting for the resource must be shown in both timelines. E.g., if the DNS look occurred in one document, but not the other, this time will show up as 0.0ms. I will make the specification update.



3.     Navigation Timing L2 as FPWD and Timing control for script-based animations as CR

The working group has agreed to publish Navigation Timing L2 specification, https://dvcs.w3.org/hg/webperf/raw-file/tip/specs/NavigationTiming2/Overview.html, as a First Public Working Draft. The working group has also agreed to publish Timing control for script-based animations as CR next week. We will send a separate email to the mailing list to make sure there are no objections to publishing this document as CR.



4.     Charter Milestones

Philippe will be sending out an updated charter with proposed milestone dates for each of the specifications. If editors want to update those dates, they can propose revisions.




W3C Web Performance WG Teleconference #94 2013-01-23



IRC log: http://www.w3.org/2013/01/23-webperf-irc



Meeting Minutes: http://www.w3.org/2013/01/23-webperf-minutes.html



Attendees

Jatinder Mann, Arvind Jain, James Simonsen, Philippe Le Hegaret


Scribe

Jatinder Mann



Agenda

1.     Discuss Test Cases issues

2.     Discuss Open Spec issues

3.     New Specifications

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



Minutes:
Test Suite

Jatinder: I believe the Intel User Timing test cases, http://w3c-test.org/webperf/tests/submission/Intel/user-timing/, were targeting an older version of the spec, as most of the tests are hitting script errors as they are trying to call window.performance.mark() and window.performance.measure(). These functions were removed from the User Timing spec prior to taking the spec to CR.
... Without them fixing those script error issues, it's hard to see the rest of the tests run. From glancing over their scripts, I wasn't able to find anything new compared to the Microsoft submitted test cases. I'll take a closer look.
... I also noticed that they are testing mark("secureConnectionStart"), which is an optional parameter, meaning the test case should be marked optional or just not be tested.
... I've briefly reviewed the test cases that you had submitted at http://w3c-test.org/webperf/tests/submission/Google/resource-timing/
... Looks like there are a number of test bugs which are causing the test failures, which I will enumerate in a separate mail to the mailing list. The key ones seems to be that the tests don't recognize the "document" PerformanceEntry. Also, the tests seem to expect resources to be named by the URL from which they were originally loaded, not the complete resolved URL of the resource. I've also noticed that the tests uses the window.locati[CUT]
... which hasn't been standardized (http://www.w3.org/html/wg/drafts/html/master/browsers.html#the-location-interface). I'll send a more detailed email.
... On the Page Visibility front, I will upload the updated test_minimize.htm and test_tab_state_change.htm today. It's been updated to ensure that the visibilitychange event is registered on each subdocument.
Spec Issues

Jatinder: Resource Timing Duplicate Requests: http://lists.w3.org/Archives/Public/public-web-perf/2012Dec/0005.html
... I believe, based on that last mail thread, that we wanted to update the spec to clarify that for the case where two requests in different documents ask for the same resource, the time spent waiting for the resource should be shown in both timelines. If the DNS look occurred in one document, but not the other, this time.

James: I agree with that spec change.
... What if a resource is removed from the document and it was the first to make the request?

Jatinder: We're trying to show the network activity, so I would say that the resouce wouldn't be removed from the timeline. I can try to clarify the spec here.
Charter

Plh: The one piece of feedback on the Charter was that we should include the timeline for when we expect the specs to move to different milestones.

Arvind: I agree that we should put that in. Philippe, why don't you suggest dates and the editors can give feedback on them?

Plh: I'll send out the updated charter with milestone dates for review.

Arvind: What is our progress on our specs?

Jatinder: User Timing has no spec feedback. The test cases had been reviewed and approved by James.

Plh: Any objections to moving the Microsoft User Timing test cases to the approved folder?

James: No objections.

Plh: I'll move them.

Jatinder: I'll follow up with Intel on my feedback to their submitted test cases.
... Resource Timing has one spec update that we had just discussed earlier. James had reviewed the test cases, and there was on piece of feedback (async XHR instead of sync XHR). I'll have that submitted.
... Page Visibility has no remaing spec changes. I have the updated test cases per Boris' feedback. I'll upload them today. I believe this is the only remaining work left for this spec. We had planned to move this spec to PR in three weeks per our last discussion.

Plh: What about requestAnimationFrame?

Jatinder: There was only one issue raised, the animationStartTime issue I had raised, however, the editor has closed on that issue in his last mail thread. There are no more remaining issues for this spec.

Plh: Shall we move requestAnimationFrame to CR?

Jatinder: Yes.

Arvind: Yes.

Plh: I'll send mail to the mailing list next week to announce our intent. Jatinder, please make sure to add moving rAF to CR for next week's meeting agenda.

Jatinder: Definitely.

Plh: Are there any objections to moving Navigation Timing 2 to First Public Working Draft?

Jatinder: No objections.