W3C

– DRAFT –
PEWG

21 June 2023

Attendees

Present
flackr, mustaq, Patrick_H_Lauke, smaug
Regrets
-
Chair
Patrick H. Lauke
Scribe
Patrick H. Lauke, Patrick_H_Lauke

Meeting minutes

ScribeNick Patrick_H_Lauke

Meta-issue: update WPT to cover Pointer Events Level 3 w3c/pointerevents#445

Patrick: where are we up to? seeing some activity...

<mustaq> Reopened w3c/pointerevents#463 when adding a test

Patrick: https://github.com/w3c/pointerevents/issues?q=is%3Aclosed+label%3Aneeds-wpt+

<mustaq> Sent a PR to fix.

Patrick: 6 still open. do you think we can get these done, do you need extra help?

Mustaq: should be ok, going a bit slow

Olli: there will be holidays soon - will be away in July likely

Mustaq: should be able to split between us

ACTION: continue working on these

Heartbeat: Clarify what the target of the click event should be after capturing pointer events w3c/pointerevents#356

Patrick: made a changed codepen, but didn't collate all results. should i still do it?

Rob: I did testing myself, and mousedown target is always consistent with the pointerup target, even when the capture was established during pointer event which happened "first"

Olli: and even if you release in pointerup?

Rob: yes

Rob: Mustaq and I had offline discussion last week...what user may be doing, what dev expectation may be

Rob: we talked about boundary events previously, and they're not consistent with click event already at this point

Mustaq: makes sense to maintain special status of click event...

Patrick: sorry, my connection is really bad, missing every other word at the moment...

Patrick: did we get to an agreement then?

[Rob and Olli confirm]

Mustaq: compat side...

Mustaq: [is low risk?] because it's different in all browsers already anyway (?)

Mustaq: in case something still breaks, we can reconsider

Olli: we can say that this is resolved, we need a PR

Mustaq: I can take care of that

Olli: that's fine, I can review

Olli: and then there's the issue with lostpointercapture w3c/pointerevents#357

Mustaq: second comment...

Rob: it's already agreed that the click's target is based on down and up, so don't need to ... (?)

Olli: so lostpointercapture can happen first

Mustaq: because click is a special high-level event

Olli: i'm fine with that

Olli: if you explicitly or implicitly release, it's the same

Rob: yes because the top event is captured

Olli: if pointercaputre released before pointerup, then click is not handled in any special way

Rob: conceptually calculates target at some point where we get pointerup

Olli: adding comment to #357 - fire lostpointercapture before click

Patrick: does this need change/more explicit wording in spec?

Mustaq: i think so, will try to tackle with the PR for both issues

Olli: because currently spec seems to say both behaviours are fine which is contradictory

Mustaq: question - does this thinking apply to auxclick and contextmenu etc as well?

Rob: shouldn't be differrent...

Mustaq: UI event spec doesn't mention common ancestor at all for those events

Rob: we're using the targeting based on the events that led to the click...

Rob: same idea could be applied to auxclick etc

Rob: e.g. if auxclick is result of a single event, then its targeting depends on targeting of that single event (?)

Rob: this is platform dependent

Rob: contextmenu shows up on the down, and sometimes on the up. will vary by platform, but consistent with our handling of click. when based on down event, use targeting from down event, etc

Rob: if not using common ancestor, just go by what the up event is

Olli: but auxclick is different

Mustaq: don't even know if auxclick has common ancestor concept

Olli: might be a gap/shortcoming of UI event spec, not necessarily intentional that it's not mentioned

Olli: ...there is also double-click

Mustaq: inclusive ancestor wording in UI spec is very handwavy. so think we need to add auxclick there as well

<mustaq> https://w3c.github.io/uievents/#events-mouseevent-event-order

<mustaq> auxclick should be there too, right?

<mustaq> By "there" I mean the text around "nearest common inclusive ancestor"

Rob: looks to me auxclick should use same inclusive ancestor targeting

Rob: auxclick definition says it's always a press and a release

<flackr> https://w3c.github.io/uievents/#event-type-auxclick

Rob: should use same targeting behaviour as click

Olli: Firefox does it

ACTION: mustaq to draft PR to cover #356 and #357, Olli to review

TPAC

Patrick: confirming i'll be physically at TPAC

[Olli and Mustaq confirm they will also be physically there, Rob will join virtually]

Patrick: additionally, worked with Philippe for the charter extension, and said that we're roughly aiming at getting last blocking issues addressed before TPAC

Patrick: then to move forward with spec, we'll need to do the implementation report etc. Even if not fully implemented (though ideal), we need at least commitment/signals from different browser engines that they plan to implement the new parts, so shouldn't be a problem

Patrick: thank you all. As said, will send reminders well before next meeting.

Summary of action items

  1. continue working on these
  2. mustaq to draft PR to cover #356 and #357, Olli to review
Minutes manually created (not a transcript), formatted by scribe.perl version 210 (Wed Jan 11 19:21:32 2023 UTC).

Diagnostics

No scribenick or scribe found. Guessed: Patrick_H_Lauke

Maybe present: Olli, Patrick, Rob

All speakers: Mustaq, Olli, Patrick, Rob

Active on IRC: flackr, mustaq, Patrick_H_Lauke