IRC log of pointerevents on 2023-12-20

Timestamps are in UTC.

15:39:30 [RRSAgent]
RRSAgent has joined #pointerevents
15:39:35 [RRSAgent]
logging to https://www.w3.org/2023/12/20-pointerevents-irc
15:39:47 [Patrick_H_Lauke]
Meeting: PEWG
15:39:51 [Patrick_H_Lauke]
Chair: Patrick H. Lauke
15:39:59 [Patrick_H_Lauke]
Agenda: https://www.w3.org/events/meetings/6246bc85-4dae-43a8-a50c-9bc5a0829585/20231220T110000/
15:40:03 [Patrick_H_Lauke]
Scribe: Patrick H. Lauke
15:40:09 [Patrick_H_Lauke]
ScribeNick: Patrick_H_Lauke
15:40:29 [Patrick_H_Lauke]
rrsagent, make logs world-visible
16:00:25 [flackr]
flackr has joined #pointerevents
16:00:43 [Patrick_H_Lauke]
present+ flackr
16:01:52 [Patrick_H_Lauke]
present+ smaug
16:05:24 [Patrick_H_Lauke]
present+ mustaq
16:07:19 [flackr]
https://bugs.chromium.org/p/chromium/issues/detail?id=269917#c49
16:07:55 [Patrick_H_Lauke]
[group briefly discussed an interesting interop issue they came across - some browsers "soft-capturing" mouse events to a particular window/frame on mousedown, so can't drag/drop between frames; firefox and safari do it, chrome doesn't. differences across platforms. possibly cross-origin differences. might not be pointer events specific, more generally UI events, as this is undefined behaviour]
16:12:51 [mustaq]
The drag out of window causing reversed typing: https://crbug.com/1402816
16:13:13 [Patrick_H_Lauke]
[decided group may tackle this aspect to spec in future, even though this may need to then seep into UI events]
16:17:42 [Patrick_H_Lauke]
TOPIC: Clarify mousedown event target if the preceding pointerdown event listener removes the target #492 https://github.com/w3c/pointerevents/issues/492
16:20:35 [Patrick_H_Lauke]
Mustaq: was hoping to write a PR for this week, but will happen in new year
16:20:58 [Patrick_H_Lauke]
Patrick: after that, we will then submit for wide review, once all normative changes are in the spec and it's "complete" for v3
16:21:26 [Patrick_H_Lauke]
TOPIC: Meta-issue: update WPT to cover Pointer Events Level 3 #445 https://github.com/w3c/pointerevents/issues/445
16:21:31 [Patrick_H_Lauke]
Olli: we have 6
16:21:51 [Patrick_H_Lauke]
Olli: wonder if we can go through these in the new year and just settle them
16:22:26 [Patrick_H_Lauke]
Mustaq: just need some time. when are we next meeting
16:22:54 [Patrick_H_Lauke]
Patrick: we next meet on 17 January (I have next two weeks off)
16:23:02 [Patrick_H_Lauke]
Patrick: will send reminder earlier
16:23:41 [Patrick_H_Lauke]
Rob: we also have weirdness with DOM modification (e.g. lostPointerCapture etc when node is removed). don't have concrete proposal, but our naive assumption that we resolve on next pointer event doesn't quite work on that case
16:24:19 [Patrick_H_Lauke]
Rob: not sure it needs spec change, but need to verify that what we expect to happen happens correctly ?
16:24:54 [Patrick_H_Lauke]
Rob: if during lostPointerCapture of the previous element, or gotPointerCapture of the new element, and then the element is removed...
16:25:27 [Patrick_H_Lauke]
Rob: we don't like sending events to targets that are removed...but we should talk about it some more
16:25:46 [Patrick_H_Lauke]
Rob: if the event is not going to the new capture target, where IS it going
16:25:55 [Patrick_H_Lauke]
Rob: don't want to drop the event, but need to decide who gets it
16:26:25 [Patrick_H_Lauke]
Rob: i landed a WPT that tests what i think is the expected behaviour. will file an issue to see if there's agreement
16:26:30 [mustaq]
The WPT Rob is landing now: https://chromium-review.googlesource.com/c/chromium/src/+/5053186
16:26:45 [Patrick_H_Lauke]
Olli: might be one of the cases where none of the browsers are consistent/do what we'd expect
16:27:21 [Patrick_H_Lauke]
Rob: there's also other oddities, like boundary events and elements are removed, we still send boundary events. may need to say we snapshot
16:27:26 [Patrick_H_Lauke]
Mustaq: snapshot but don't send the event...
16:27:51 [Patrick_H_Lauke]
Rob: ...every browser sends it to the full chain, including the removed nodes
16:28:32 [Patrick_H_Lauke]
Rob: this doesn't seem at all PE specific, just a bunch of edge cases that should have been covered in UI events...
16:29:01 [Patrick_H_Lauke]
Rob: capture is weird because it gives you a chance to change the DOM before the events come in, so you potentially send events to a removed node
16:29:24 [Patrick_H_Lauke]
Olli: hit testing should be specified...
16:29:48 [Patrick_H_Lauke]
Rob: none of this is contingent on hit testing. assuming there IS a consistent-enough hit testing, THEN what should we do...
16:33:17 [Patrick_H_Lauke]
Patrick: thank you all, as ever. Once we have the last issue for v3 resolved, we'll finally be able to draw a line under PE3, and then regroup to see where we want to take this after that (e.g. living standard). in the meantime, have a good break, and we'll reconvene on the 17 January. I'll send an earlier reminder in the new year about outstanding work, so it doesn't creep up on people the night before of the meeting.
16:33:23 [Patrick_H_Lauke]
rrsagent, set logs world-visible
16:33:27 [Patrick_H_Lauke]
rrsagent, generate minutes
16:33:28 [RRSAgent]
I have made the request to generate https://www.w3.org/2023/12/20-pointerevents-minutes.html Patrick_H_Lauke
16:33:49 [Patrick_H_Lauke]
rrsagent, bye
16:33:49 [RRSAgent]
I see no action items