Minutes from Pointer Events WG call 30 August 2023

Dear all,

the minutes from today's meeting are at 
https://www.w3.org/2023/08/30-pointerevents-minutes.html and copied below:


PEWG
30 August 2023

Agenda: 
https://www.w3.org/events/meetings/6246bc85-4dae-43a8-a50c-9bc5a0829585/20230830T110000/

Attendees
flackr, mustaq, Patrick_H_Lauke, plh, smaug

Chair: Patrick H. Lauke
Scribe: Patrick H. Lauke, Patrick_H_Lauke


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

* consider to change the behavior of getCoalescedEvents on 
non-SecureContext w3c/pointerevents#472

* Clarify pointerleave and pointerout events when first pointer move 
after removing an element under the pointer w3c/pointerevents#477

* video for TPAC



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

Patrick: we have 5 open 
https://github.com/w3c/pointerevents/issues?q=is%3Aclosed+label%3Aneeds-wpt+

Olli: i did file the webdriver issue about coalesced/predicted events

Olli: but will still need to be manual for now, as it won't be supported 
anytime soon

w3c/webdriver#1759

Mustaq: fixed some of the test, hope to do 411 after TPAC



# consider to change the behavior of getCoalescedEvents on 
non-SecureContext w3c/pointerevents#472

Patrick: I made a naive minimal PR, can we check it? 
https://github.com/w3c/pointerevents/pull/480/files

Olli: before that, do we have data from the use counter?

Mustaq: we've not looked into this...

Rob: the use counter was added in juli

Mustaq: but since, we looked at the counter and there was no data, so 
might not have been used

<flackr> https://chromestatus.com/metrics/feature/timeline/popularity/4598

Rob: but it's relatively new counter, so may not be representative

Olli: if usage is very low, we could try to keep spec as it is right now...

Rob: i think this is low enough for us to consider NOT to have the API

Olli: so Patrick we don't need your PR at all

Patrick: so do we need to say anything in spec about it only being for 
secure context?

Rob: no, it already says so in the IDL

RESOLUTION: PR not needed

Patrick: can we close w3c/pointerevents#472 ?

Olli: one other issue already has needs-WPT, so don't need to keep this 
one open

<smaug> w3c/pointerevents#318 has needs-wpt, and I'll write a test



# Clarify pointerleave and pointerout events when first pointer move 
after removing an element under the pointer w3c/pointerevents#477

Rob: there's a discussion on the interop

<mustaq> web-platform-tests/interop#380

Olli: i was looking at it too, and your proposal seemed reasonable ... 
but there are issues: if you change slot attribute, you end up being 
slotted elsewhere, so it's almost like a removal...what should happen 
then, since ancestor chain will be different

Rob: we should track the nearest non-removed ancestor of the event path

Olli: not when you're changing node, but when you change slot

Rob: does that not fire change as well?

Rob: I think we should interpret it as: that entire slot was removed, so 
the node that the mouse was previously over that had enter fired to it...

Olli: nothing would be removed from the DOM, so we need to fire 
something else... the flatten tree is different because the slot is 
different ... need to think about this a bit more. but liked your idea

Mustaq: can we add note in issue?

Mustaq: added a few points in the interop issue

Mustaq: the test currently assumes that in and out events are 
interleaved on the parent node

Rob: we should test the case where the cursor is STILL over the same 
parent, so doesn't need to fire another enter

Rob: there are more cases that need to be tested

Mustaq: another question about event order (does over come before enter, 
etc). PE doesn't specify it...

Rob: we should match UI Events order

Mustaq: problem is UI Events doesn't define this either (?)

Mustaq: only WPTs that assert it

Rob: we should file an issue to specify which one fires first

Rob: suspect enter should fire before over, because we don't dispatch 
enter to ancestors, but over we do

Mustaq: because over bubbles, but enter doesn't

Olli: I think over is dispatched first...

[further discussion on current weird order in different implementations]

Rob: as developer, i'd expect enter - over and then out - leave

Olli: but as developer you wouldn't listen to both in most cases

Rob: we currently fire over then enter, and out then leave ... which is 
weird

Rob: test for mouse seems to pass in all browsers. think it's wrong, but 
not going to argue too strongly that we should change it

Mustaq: checking the mouse one ... over then enter, out then leave ... 
and that's been there forever

Mustaq: so keep this assumption for our own WPTs?

Rob: how hard would it be to change it (for mouse events as well?). 
might add a use counter

<mustaq> https://www.w3.org/TR/uievents/#events-mouseevent-event-order

Olli: this is defined in a sense in UI Events 5.3.3

Rob: don't like the order, but no reason why PE shouldn't do the same thing

<mustaq> The order specified by example there is (over, enter, out, leave).

<mustaq> Let's keep that in the current test then

Rob: not specified in PE, and maybe PE should say something about it...

Rob: should have the same event ordering as mouse events, explicitly

ACTION: consider if PE needs to clarify event order for enter/over 
leave/out etc

ACTION: Rob to consider if PE needs to clarify event order for 
enter/over and leave/out etc



# video for TPAC

Patrick: I've got the presentation ready with links to various demos 
that we want to show, have carved out time to record it tomorrow and 
send it, will then be on the w3c site before TPAC (will share link once 
I have it)


-- 
Patrick H. Lauke

https://www.splintered.co.uk/ | https://github.com/patrickhlauke
https://flickr.com/photos/redux/ | https://www.deviantart.com/redux
https://mastodon.social/@patrick_h_lauke | skype: patrick_h_lauke

Received on Wednesday, 30 August 2023 15:41:34 UTC