IRC log of pointerevents on 2021-03-31

Timestamps are in UTC.

14:52:48 [RRSAgent]
RRSAgent has joined #pointerevents
14:52:48 [RRSAgent]
logging to https://www.w3.org/2021/03/31-pointerevents-irc
14:53:02 [Patrick_H_Lauke]
Meeting: PEWG
14:53:06 [Patrick_H_Lauke]
Chair: Patrick H. Lauke
14:53:10 [Patrick_H_Lauke]
Agenda: https://www.w3.org/events/meetings/9718517d-0e08-4377-bb7c-07332948233b/20210331T110000
14:53:21 [Patrick_H_Lauke]
Scribe: Patrick H. Lauke
15:02:22 [Patrick_H_Lauke]
present+ plh
15:02:27 [Patrick_H_Lauke]
present+ smaug
15:03:13 [mustaq]
present+ mustaq
15:04:31 [plh]
plh has joined #pointerevents
15:04:46 [liviu]
liviu has joined #pointerevents
15:05:08 [Patrick_H_Lauke]
TOPIC: Reword/expand touch-action definition #349 Think we're close to the finish on this one - would like to see if we can finalise this on the call.
15:05:14 [Patrick_H_Lauke]
https://github.com/w3c/pointerevents/pull/349
15:09:55 [Patrick_H_Lauke]
Patrick: [discussion around Olli's comment on "none" being unclear] wondering if we do need to be more explicit there in the definition, since straight after the definition list we have the paragraph about what touch-action does or doesn't cover
15:10:15 [Patrick_H_Lauke]
Patrick: but we could move that para *before* the definition list, less likely to be missed by a casual reader
15:10:24 [Patrick_H_Lauke]
Mustaq: yes that sounds like a good idea
15:10:53 [Patrick_H_Lauke]
https://pr-preview.s3.amazonaws.com/w3c/pointerevents/pull/349.html#details-of-touch-action-values
15:12:11 [Patrick_H_Lauke]
present+ liviu
15:14:56 [Patrick_H_Lauke]
Olli: yes, happy for the paragraph to be moved and to keep definition of none short as currently proposed
15:15:16 [Patrick_H_Lauke]
Patrick: i had further thoughts on this section as well https://github.com/w3c/pointerevents/pull/349#issuecomment-811092269
15:16:14 [Patrick_H_Lauke]
at the start of section 9 https://pr-preview.s3.amazonaws.com/w3c/pointerevents/pull/349.html#declaring-candidate-regions-for-default-touch-behaviors
15:20:07 [Patrick_H_Lauke]
[discussion on the first part about unclear intro to section 9]
15:20:28 [Patrick_H_Lauke]
Mustaq: I can see what you mean now, you mentioned this before. Let's land this PR first and make this a separate PR
15:20:33 [Patrick_H_Lauke]
Patrick: agreed
15:23:10 [Patrick_H_Lauke]
Patrick: second concern is whether the second note in https://pr-preview.s3.amazonaws.com/w3c/pointerevents/pull/349.html#declaring-candidate-regions-for-default-touch-behaviors which talks about "it's 'touch' but not just touch"
15:23:18 [Patrick_H_Lauke]
Mustaq: this is more related to your other PR
15:23:31 [Patrick_H_Lauke]
Patrick: yes, just wondering generally if this should be made normative rather than an informative note
15:23:43 [Patrick_H_Lauke]
Olli: does your other PR (#350) do this already?
15:24:32 [Patrick_H_Lauke]
Patrick: probably makes sense to look at this thought as part of the other PR
15:27:23 [Patrick_H_Lauke]
ACTION: Patrick to move paragraph explaining touch-action does not affect text selection/highlighting, link activation, etc *before* the definition list in section 9.3
15:27:39 [Patrick_H_Lauke]
TOPIC: Major refactoring: refer to "direct manipulation" rather than "touch" #350
15:27:49 [Patrick_H_Lauke]
https://github.com/w3c/pointerevents/pull/350
15:28:09 [Patrick_H_Lauke]
Patrick: not had a chance to look at this in much depth since last time, but one comment
15:28:15 [Patrick_H_Lauke]
https://github.com/w3c/pointerevents/pull/350#issuecomment-811094420
15:36:08 [Patrick_H_Lauke]
[discussion on using "direct manipulation *for* panning / zooming" to cover the concept better and bridge the gap between the two different views of what "direct manipulation" means (touchscreen-style panning/zooming/touching screen versus the "any GUI with a mouse" view of the term)]
15:36:39 [Patrick_H_Lauke]
[general agreement that the term like that is better, but concern that the spec itself may get too heavy-handed and difficult to read/understand]
15:37:24 [Patrick_H_Lauke]
Patrick: what i propose is that for next time, i'll make either a modification to PR 350, or a fresh PR, using this term and keeping it clear. moving "direct manipulation for panning/zooming" into a definition and try to avoid having to define it everywhere
15:37:53 [Patrick_H_Lauke]
ACTION: Patrick to either change PR 350 or supersede with a new one, incorporating the "direct manipulation *for* panning/zooming" concept
15:38:07 [Patrick_H_Lauke]
TOPIC: Click event while a pointer event is captured #75
15:38:16 [Patrick_H_Lauke]
https://github.com/w3c/pointerevents/issues/75
15:38:21 [Patrick_H_Lauke]
Patrick: just wanted to get a feel
15:38:52 [Patrick_H_Lauke]
Olli: surprised by Safari's behavior. was ready to make Gecko behave like Chrome, but now not sure why Safari is doing it differently
15:42:22 [Patrick_H_Lauke]
Patrick: Safari for scenario 2 behaves like Firefox, so both behave differently from Chrome
15:45:31 [Patrick_H_Lauke]
Liviu: just tested 2 in Chrome, and grey gets the click
15:46:16 [Patrick_H_Lauke]
Patrick: I propose that I spend some time documenting current state of Chrome, Firefox, Safari for all 4 scenarios, so we get a better picture of what the current state of play is / where the delta is. This issue was opened 2016, so the whole convo is a bit unwieldy to follow
15:46:43 [Patrick_H_Lauke]
ACTION: Patrick to document what current browser behaviors for the 4 tests actually is, to ease further discussion on #75
15:46:52 [Patrick_H_Lauke]
plh: do we have a test in WPT
15:47:27 [Patrick_H_Lauke]
mustaq: i think we have a test https://github.com/w3c/pointerevents/pull/174
15:47:41 [mustaq]
https://github.com/web-platform-tests/wpt/issues/3213
15:48:40 [liviu]
https://github.com/web-platform-tests/wpt/pull/4763/commits/960dbfabe0ed8b551b9c9e6f39c0e08a18b4c85d
15:48:59 [Patrick_H_Lauke]
plh: we should have test results, but maybe this is from a time before we had web drivers
15:49:09 [Patrick_H_Lauke]
s/web drivers/web driver support
15:49:18 [Patrick_H_Lauke]
mustaq: the test is manual, not an automated test
15:49:30 [mustaq]
WPT result: https://wpt.fyi/results/pointerevents/pointerevent_click_during_capture.html?label=experimental&label=master&aligned
15:49:36 [Patrick_H_Lauke]
mustaq: actually, scratch that, test is automated
15:50:10 [Patrick_H_Lauke]
mustaq: so yes we have some issues there
15:51:04 [Patrick_H_Lauke]
Patrick: so either chromium needs to change (and we redefine what is correct), or gecko/webkit need to change
15:51:20 [Patrick_H_Lauke]
Olli: also differences in how capture is lost
15:51:32 [Patrick_H_Lauke]
mustaq: probably best to separate out those two concerns, easier to discuss
15:51:58 [mustaq]
I suggest covering "click event target" through this issue...
15:52:08 [mustaq]
and file a new issue for ordering of lostpointercapture.
15:52:57 [Patrick_H_Lauke]
Patrick: suggest closing this issue (from 2016 onwards) as it's a bit unwieldy. supersede it with a fresh issue. Olli do you mind doing that?
15:52:59 [Patrick_H_Lauke]
Olli: sure
15:53:35 [Patrick_H_Lauke]
ACTION: Olli to open new fresh issue for click event target issue once Patrick has documented table of current behaviors
15:54:12 [Patrick_H_Lauke]
Patrick: AOB?
15:54:34 [Patrick_H_Lauke]
Liviu: i opened a new issue https://github.com/w3c/pointerevents/issues/355
15:56:43 [Patrick_H_Lauke]
Mustaq: [explains at high level the issue of what happens/what's the order of events between frames. DOM/UI events spec may not handle this either]
15:57:04 [mustaq]
https://dom.spec.whatwg.org/#dispatching-events
15:57:24 [Patrick_H_Lauke]
Mustaq: don't think we can solve this in meeting today, but worth considering
16:01:05 [smaug]
mustaq: ah, I see, release on move
16:01:35 [Patrick_H_Lauke]
Patrick: worth iterating over this in github, let's see what where we land for next meeting
16:02:06 [Patrick_H_Lauke]
Patrick: in meantime, I'll be working/reworking #350, and merge #349 after moving the paragraph before the definition list as discussed
16:02:20 [Patrick_H_Lauke]
rrsagent, set logs world-visible
16:02:42 [Patrick_H_Lauke]
rrsagent, create minutes
16:02:42 [RRSAgent]
I have made the request to generate https://www.w3.org/2021/03/31-pointerevents-minutes.html Patrick_H_Lauke
16:06:20 [smaug]
oh, and the spec is nicely inconsistent with lostpointercapture. "click and lostpointercapture events are fired, click SHOULD be fired before lostpointercapture." vs "Immediately after firing the pointerup or pointercancel events, a user agent MUST clear the pending pointer capture target override for the pointerId of the pointerup or pointercancel event that was just dispatched, and then run Process Pending Pointer Capture steps to
16:06:20 [smaug]
fire lostpointercapture if necessary."
16:07:12 [Patrick_H_Lauke]
sigh
16:21:17 [Patrick_H_Lauke]
rrsagent, bye
16:21:17 [RRSAgent]
I see 4 open action items saved in https://www.w3.org/2021/03/31-pointerevents-actions.rdf :
16:21:17 [RRSAgent]
ACTION: Patrick to move paragraph explaining touch-action does not affect text selection/highlighting, link activation, etc *before* the definition list in section 9.3 [1]
16:21:17 [RRSAgent]
recorded in https://www.w3.org/2021/03/31-pointerevents-irc#T15-27-23
16:21:17 [RRSAgent]
ACTION: Patrick to either change PR 350 or supersede with a new one, incorporating the "direct manipulation *for* panning/zooming" concept [2]
16:21:17 [RRSAgent]
recorded in https://www.w3.org/2021/03/31-pointerevents-irc#T15-37-53
16:21:17 [RRSAgent]
ACTION: Patrick to document what current browser behaviors for the 4 tests actually is, to ease further discussion on #75 [3]
16:21:17 [RRSAgent]
recorded in https://www.w3.org/2021/03/31-pointerevents-irc#T15-46-43
16:21:17 [RRSAgent]
ACTION: Olli to open new fresh issue for click event target issue once Patrick has documented table of current behaviors [4]
16:21:17 [RRSAgent]
recorded in https://www.w3.org/2021/03/31-pointerevents-irc#T15-53-35