Minutes from Pointer Events WG call 23 June 2021

Dear all,

minutes from today's call will be at 
https://www.w3.org/2021/06/23-pointerevents-minutes.html (though 
RRSAgent has been having issues) and copied below (apologies for the 
rough and ready manual formatting)

Meeting: PEWG
Chair: Patrick H. Lauke
Agenda: 
https://www.w3.org/events/meetings/9718517d-0e08-4377-bb7c-07332948233b/20210623T110000
Scribe: Patrick H. Lauke

Present:

Olli Pettay
Robert Flack
Mustaq Ahmed
Philippe Le Hégaret
Patrick H. Lauke

TOPIC: Review 'Expand explanation for non-coalesced events' 
https://github.com/w3c/pointerevents/pull/379 (in particular, use of 
"measurable" in the definition)
Patrick: [explains the conundrum we had with continuous/discrete / 
measurable/countable, etc]
Olli: concern about including/listing UI-Events ... are there more 
properties?
Rob: reason why i didn't want to list all properties, what if things 
change or other specs add more properties
Olli: think PR is quite close though
Rob: will see if there are suggestions to make on this
Mustaq: would also remove shiftkey/altkey/metakey as they don't fire 
pointer events at all
present+ mustaq
Rob: maybe we can even give the example about button not being 
continuous as the value just changes, you don't get the "in-between" values
Patrick: might be getting too deep into weeds of buttons which is not 
even our spec
ACTION: review this PR for next meeting and suggest changes by then
***RRSAgent records action 1

TOPIC: Review 'task queue is not reliable HTML concept. Use event loop 
instead' https://github.com/w3c/pointerevents/pull/386
PLH: we are getting better at referencing, but other specs can tell when 
WE do bad referencing. task queue not reliable concept, so should use 
event loop
Patrick: defer to other people who are more knowledgeable
Olli: it's ... ok
Olli: technically you put a task into the queue...
Mustaq: the spec does handwave saying that "queue is not a queue"
PLH: we rely on DOM to fire event, so we don't have to say anything 
about when/how to fire
Patrick: so we happy to merge this?
[no concerns]
ACTION: merge and close
***RRSAgent records action 2

TOPIC: Review 'Simplify/clarify coalesced and predicted events' 
https://github.com/w3c/pointerevents/pull/377
Rob: I should have more time soon, so can look at this for next meeting
Patrick: if anybody else also has any ideas, jump in. the PR is mostly 
ok, the last sticking point was the need to go into the 
targeting/retargeting (my PR was a bit too brutal in removing it altogether)
ACTION: Rob to review this PR and add suggested wording for the 
targeting issue
***RRSAgent records action 3

TOPIC: Unclear note about PointerEvent initialization of attributes to 
reflect coalesced events https://github.com/w3c/pointerevents/issues/374
Rob: high-level purpose here is that developers don't need to care about 
the individual (non-coalesced) events, but just look at the fired events
Patrick: but is this going into the weeds of sometihng we don't want to 
define/that they shouldn't care?
Rob: this is just to reinforce/give an example
Rob: implementation shouldn't need to do this, but this is only an example
Patrick: any value in me having a stab at generalising this, without 
having to give a specific example?
Rob: movement is a good example though - it's a delta from the last change
Mustaq: we could say this without mentioning pointerlock
Rob: yes this is true even without pointerlock
Patrick: I might give a try proposing a slightly more generalised 
wording, without mentioning pointerlock, that achieves same
ACTION: Patrick to propose more generalised wording for the note
***RRSAgent records action 4

TOPIC: The behavior of getCoalescedEvent in 
pointerevent_constructor.html is inconsistent with the spec 
https://github.com/w3c/pointerevents/issues/229
Patrick: is this going to be resolved by the work Rob is going to do on 
the PR?
Rob: we were going to say this is true for trusted events
Olli: will review this and the other related issues
"This API always returns at least one coalesced event for pointermove 
events and an empty list for other types of PointerEvents." 
https://github.com/w3c/pointerevents/issues/224
How is pointer event ctor supposed to work when coalescedEvents is 
passed using the PointerEventInit 
https://github.com/w3c/pointerevents/issues/223
Olli: I can try to write some PR
ACTION: Olli to review/propose PRs for these
***RRSAgent records action 5

TOPIC: HTML monkeypatching: initiate the drag-and-drop operation 
definition https://github.com/w3c/pointerevents/issues/384 and HTML 
monkeypatching: animation frame callbacks 
https://github.com/w3c/pointerevents/issues/385
Patrick: Starting with the drag and drop one #384
Olli: we should do this, but think it's still not possible to write 
platform test for d'n'd. at least wasn't last time
PLH: you agree we should add the spec the requirement to fire 
pointercancel etc?
Olli: that or add a hook
PLH: they declined to add a hook. asked for it to be exported
PLH: should i make a stab at this?
Olli: i don't have time for this right now
Mustaq: I can take a look
Patrick: you're it mustaq, assigned you. next up animation frame 
monkeypatching
Olli: ... it's about scheduling of tasks, and that's not defined 
anywhere. HTML spec lets you schedule tasks any way you want. maybe we 
should clarify somehow in PE spec, not sure how
PLH: domenic suggested adding a new function to fire pointer event
Olli: it's not even about pointer events...
Rob: it also doesn't have any well defined time WHEN you want to fire, 
you may want to do it before or after, depending on performance etc
Olli: implementations will likely change over time
Rob: we could say it should happen at some point between two points - 
producing a frame
Olli: even that could be not true / browser may choose to paint more and 
listen to events less, depending on priority. would prefer having plenty 
of flexibility
Patrick: i think it's a "soft" *may* we use in that sentence in the spec
https://w3c.github.io/pointerevents/#the-pointermove-event
Rob: only thing we can guarantee is that events will be dispatched in order
we would flush coalesced events before next fired event
PLH: we could remove the sentence altogether as we already say things in 
the specific coalesced events section
PLH: suggest either removing the sentence, or moving it to the section 
for coalesced. and we don't mention animation callbacks
mustaq: https://w3c.github.io/pointerevents/#the-pointerrawupdate-event
Mustaq: there's also mention of animation (without link) in the pointerraw
Patrick: could we just change "which might be aligned to animation 
callbacks" to "which might be delayed"
Rob: we just need to say that raw events will not be delayed
ACTION: Patrick to write PR that removes the referenced animation frame 
callback sentence altogether and removes mention in pointerraw
***RRSAgent records action 6
Patrick: thank you all, see you in two weeks' time

P
-- 
Patrick H. Lauke

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

Received on Wednesday, 23 June 2021 16:12:59 UTC