W3C

– DRAFT –
PEWG

01 February 2023

Attendees

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

Meeting minutes

<smaug> Patrick_H_Lauke: we're still using webex?

<mustaq> I am being asked for webex login!

Coordinates of a pointercancel event w3c/pointerevents#463

Olli: didn't have time to write a test

Mustaq: do firefox and chrome behave differently with touchcancel as well?

Olli: need to test, also good to check what Safari does

Rob: what do pointerenter/pointerleave do?

Olli: those are just hacks on top of over/out

Rob: but even then they still refer to the document...

Rob: would be good to make sure it gives coordinates of last known position inside the document?

Mustaq: i think Navid was looking at this...

Mustaq: for cancel it's tricky because it may have been cancelled for a variety of reasons

Olli: do we want to expose coords for situations when coords WERE known?

Rob: we could always report last known good coords. would be good to see what the developer use case would be.

Olli: wasn't there a chrome bug about this? what was the complaint

Mustaq: the bug reported complained that they got 0/0, don't think they understood the situation that may lead to a cancel

Rob: [speculates about situations where relying on getting last known good x/y on cancel may be useful]

Rob: it won't give devs anything more useful or wrong than listening to all the move events and looking at the last one that came in

Olli: might be good to see how touchcancel is handled, for possible consistency

Olli: for pointercancel just before a dragstart we send the last known good coordinates in gecko

Olli: per what the drag'n'drop spec demands

Patrick: conceptually i like the idea of "send the last known good, otherwise 0/0" or should it be null/null?

Patrick: we would always have a coordinate, as otherwise pointer would not exist in the first place

Rob: we'll still have the challenge of pressure, like we had to do for up event, but we have precedent

Patrick: we do have a slightly different situation though as cancel could happen while pressure is full, etc. so maybe copy/transplant last known good values for all of those

ACTION: propose some additional wording to clarify values of properties on cancel

ACTION: Olli to write test for touchcancel, for consistency check

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

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

<flackr> https://w3c.github.io/touch-events/#dom-touchevent-changedtouches

Rob: touch events DO say what happens with coordinates for touchcancel

Rob: so the proposal here would be consistent

Patrick: might have a look to see if TE spec says anything about pressure etc

Patrick: so for the WPT one we do now have the list of things lacking tests https://github.com/w3c/pointerevents/issues?q=is%3Aclosed+label%3Aneeds-wpt+ ... now how do we go about it/split this up

Mustaq: some of those tests would be trivial, others not so much. expect to look at them sometime this month, but no progress

Olli: I should try to find some time. Ask Edgar, also

Mustaq: how do we coordinate?

Olli: add a comment to the issue w3c/pointerevents#445

Olli: sorry, comment on the relevant PR

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

Olli: I did ask Anne if he could find somebody from Apple to comment on it, but still waiting

<mustaq> w3c/pointerevents#462

raw trackpad data issue

<mustaq> We discussed this here before: w3c/pointerevents#228 (comment)

Patrick: and we also have w3c/pointerevents#206

[short discussion on the topic]

Patrick: fundamentally, this feels too low-level and not quite in line with what PE do. has shades of sensor api (access sensor for trackpad directly) and even pointer lock (as if i'm, say, doing a signature on a trackpad, i don't want my mouse to move around the screen).

Patrick: I'll file a comment on all the above issues pointing to https://discourse.wicg.io/ where interested devs can suggest this as a new spec and see if it gains traction

ACTION: Patrick to comment on the three related issues/feature requests about raw trackpad data

w3c/touch-events#134

[further brief discussion on the similarities between the ask of raw trackpad and pointer lock API]

Summary of action items

  1. propose some additional wording to clarify values of properties on cancel
  2. Olli to write test for touchcancel, for consistency check
  3. Patrick to comment on the three related issues/feature requests about raw trackpad data
Minutes manually created (not a transcript), formatted by scribe.perl version 210 (Wed Jan 11 19:21:32 2023 UTC).

Diagnostics

Maybe present: Olli, Patrick, Rob

All speakers: Mustaq, Olli, Patrick, Rob

Active on IRC: flackr, mustaq, Patrick_H_Lauke, smaug