Pointer Events WG -- 15 Jul 2014

<ArtB> Scribe: Doug

<ArtB> ScribeNick: shepazu

<rbyers_> I can scribe, but will want to hand off to someone else if we get deep into discussion blink, touch-action etc...

<smaug> I think

<rbyers> scribe: rbyers

<scribe> scribeNick: rbyers

<scribe> Agenda: http://lists.w3.org/Archives/Public/public-pointer-events/2014JulSep/0007.html

tweak agenda

OP: Touch-action and touch-events

Test Assertion "holes"

is the TestAssertion data up-to-date; where are the holes? Who can commit to create tests to fill the holes (and by when)? <https://www.w3.org/wiki/PointerEvents/TestAssertions>

AB: Jacob, Matt, Cathy - any idea what coverage we have?
... and is the data we have here accurate and up-to-date?

Asir: I don't know if it's fully up-to-date, but I can take an action to ensure there's a link for each assertion from the wiki to test case

AB: That would be excellent

Asir: 66 in the wiki, the recent PR covers 54, leaving 12. Jacob has been pushing some changes to cover 2-3.
... Scott has the remaining ones, has an action to create PRs for them
... Scott's should cover about 9 of the remaining ones

SG: TWF submission covers 3.1, 3.2, 12.1, and some additional ones not listed on the wiki
... 12.1 is covered in Microsoft's submission PR1074, so I won't merge mine in - just verify we have the coverage
... for the others we should create new test assertions on the wiki, right?

Asir: how many do we need to create?

SG: Not sure. Tests for gotpointercapture/lostpointercapture firing async.
... and more extensive pointerleave tests with deeply nested elements, but not sure the complexity is necessary

<scott_gonzalez> https://github.com/dmethvin/pointerevents-test/blob/master/pointerenterleave-continuous.html

SG: see diagrams on line 175

RB: some nested element case for enter/leave would be good to test - an easy thing to break

SG: So I can add assertions to the wiki and put together a PR for whatever doesn't overlap

<scott_gonzalez> https://github.com/dmethvin/pointerevents-test/blob/master/pointerenterleave-continuous.html#L175

SG: I can try to get this done for next week

AB: Next call will likely be next tuesday in August unless someone else wants to chair

DS: I would probably be available to chair

AB: Ok, let Doug know if you want to have a meeting next week

AV: Implementations are blocked on testing, so any help to get tests landed quickly would be helpful
... let us know if anyone needs help

AB: Scott let us know if you need help getting this done for next week
... Anything else on this topic?
... Also covered the next: Test Assertion cleanup: if a TA is covered by a merged tests, the data in "Test Status" column should be a link to the merged/approved test case <http://w3c-test.org/pointerevents>. Who can commit to helping with this?

status of PR 1074

<mbrubeck> https://github.com/w3c/web-platform-tests/pull/1074

AV: If there are no comments we should merge

CC: I had some questions in github

JR: There were two questions. 1) is it true that IE fails some of the tests (constructor test and pointerleave after pointercancel test). Yes, we're looking at what we need to do. Tests are correct.
... 2) a minor thing - title element on constructor test should go on head. If Artim hasn't corrected then I will.

CC: Also one with an incomplete assertion
... a quick fix, but needs to be fixed

AV: Can we make those quick fixes and merge it in?

JR: Cathy do you agree it's fairly obvious what needs to happen to address these? Or do you want to review?

CC: They're trivial, fine to fix and land

SG: Also one for script changes to be relative

AB: Sounds like we have agreement on what needs to change. Jacob can merge once those fixes land.

<jrossi> ACTION: jrossi to make final corrections and merge PR 1074 [recorded in http://www.w3.org/2014/07/15-pointerevents-minutes.html#action01]

<trackbot> Created ACTION-114 - Make final corrections and merge pr 1074 [on Jacob Rossi - due 2014-07-22].

AV: I believe we have one or two more PRs coming in to plug remaining holes
... it would be great if we could get these in within a weeks timeframe - would really help implementers

AB: Anything else on testing for today?

CR implementation: status and updates

AB: Main open question is Firefox metro - is that 100% implemented now?

MB: It's very close, out of 72 test cases (including merged and submmitted) it's passing all but 1.

<mbrubeck> https://bugzilla.mozilla.org/show_bug.cgi?id=1036985

MB: 1 is a recent regression we're actively working on fixing.

OP: And there's a bug open to handle the legacy mouse events properly

AB: Anybody have anything else about the FF implementation?
... Any other new news regarding CR implementation?

touch-action for touch events

<scribe> scribenick: mbrubeck

OP: the amount of new spec required is rather small
... it's mainly about what events are required and when

RB: We talked about this when we were first landing touch-action in Blink
... including whether touch-action should be in its own spec.
... We talked about two things: The effect touch-action has on actions in the browser, and the effect on events.
... Aside from the stuff directly around events (like pointercancel), everything else is about browser actions.
... In terms of what happens to the events, there's no change. Blink has made some changes, but it's all within the (not very precise) wording of the Touch Events spec.
... implementations differ here already, regardless of touch-action.
... I think you could formulate those interop questions without bringing up touch-action.
... e.g. the Touch Events spec would specify whether touchcancel is dispatched when scrolling starts.
... That's independent of the touch-action question.
... I agree with Olli that it would be nice to have this written down somewhere. But I'm not sure how/where to do that.

Zakim: who is talking

JR: I'm not sure if this needs an additional spec.

RB: Maybe it doesn't; I'm mostly just concerned about the explicit mention of pointercancel in the touch-action section of the spec.
... Would it make sense to just move that one line to a different section of the PE spec?
... It could be a hook, so any other event system that wanted to support touch-action could specify its equivalent behavior.

JR: Yes, I think we could move that to 5.2.8.

RB: Another reason that could make sense is that consumers of pointer events might not use touch-action anywhere. But that behavior is still relevant even if the property is not used.

JR: I think you're right. I think we can just lift that paragraph into the section that defines pointercancel.
... I'd be okay with that if it makes it easier to write a lightweight spec referencing just 9.1.

RB: We still have the problem of all the details around how Touch Events respond to scrolling.
... We've talked about that in the touch-events CG and documented existing browser behavior.

OP: Sounds good to me.
... I just wanted to make sure we agree it should be documented somewhere.

AB: So Jacob can take an action to make that one change.
... Any objections?

[None]

<rbyers> scribenick: rbyers

AOB

AB: I'll be away for the next 2 weeks, if anyone wants to have a meeting they should notify Doug and group Monday early morning boston time

AV: I wonder if we should meet next week to finalize the test case

AB: If scott completes his action by Monday then meeting on Tuesday sounds good to me
... Sounds like we have a plan to meet next Tuesday, topic: reviewing the tests - blocked on Scott's action

- DRAFT -

Pointer Events WG

15 Jul 2014

Attendees

Contents

tweak agenda

Test Assertion "holes"

status of PR 1074

CR implementation: status and updates

touch-action for touch events

AOB

Summary of Action Items

Scribe.perl diagnostic output