– DRAFT –
PEWG

Web Platform Tests https://wpt.live/pointerevents/ / https://wpt.fyi/results/pointerevents?label=experimental&label=master&aligned

Patrick: We discussed last time, Gaston you were going to look over the WPT harness problems. did you have time?

gaston: i looked over the various failing tests. seems there's a lot of assertions throwing in firefox due to lack of pen support in the harness. also a lot of timeouts in safari, so tests are hanging up. there were a few tests failing with assertions throwing rather than failing. in chromium, there was a directory for pointerlock tests, and we have an exception with the browser because the root of the document was not va[CUT]

ointerlock

Olli: think there was proposal for interop to improve touch event testing. but we have overall pointer and mouse interop issues, but we've not been looking at all those other issues for touch and pen

Olli: wonder if we should have that as interop 2026? i think touch is more important than pen

mustaq: at least TPAC there were rumours about improve touch testing. don't know more

gaston: if you want, i can send writeup to the irc

<gastonr> - In Firefox, pen events are not supported: error: Action action_sequence failed: Error: Unimplemented pointerMove for pointerType pen (https://wpt.fyi/results/pointerevents?label=master&label=experimental&aligned&q=firefox%3Aerror)

<gastonr> - A lot of the tests that time out in Firefox and Safari have either touch-action:none or pointer-events: none. It seems like they aren't handling that case well.

<gastonr> - A lot of the Safari ERROR results are due to a crash in the python script with webdriver.error.MoveTargetOutOfBoundsException: move target out of bounds (500) (https://pastebin.com/CeEhKDeK) (most of the pointerevent_touch-action-xyz tests do this) safari:error wpt.fyi query

<gastonr> (https://wpt.fyi/results/pointerevents?label=master&label=experimental&aligned&q=safari%3Aerror)

<gastonr> - For safari, pointerevent_attributes.html?touch throws an error because of the unhandled promise of a rejected assert_equals

<gastonr> - For Safari, pointerevent_pointerrawupdate_changes_pointer_capture.https.html and pointerevent_capture_mouse_and_release_and_capture_again.html throw the exception: JavaScript exception occured: window.__wptrunner_process_next_event is not a function.. I think it's interesting that the other browsers don't have that issue, but it seems that for

<gastonr> one pointerevent test in chromium we did see this problem: pointerevent_pointerleave_after_pointercancel_touch-expected.txt

<gastonr> -In Chromium, pointerlock/ tests fail because the browser throws the exception: "The root document of this element is not valid for pointer lock" (https://wpt.fyi/results/pointerevents/pointerlock?label=master&label=experimental&aligned)

Patrick: if we know what the shortcomings are, is there a path forward for fixes to WPT? is it down to us to fix, or is there somebody we need to harrangue?

Rob: i think we agree that touch is more important than pen, and there seem to be efforts to fix these. once touch is sorted, we should be able to get to pen ones

Rob: out of the list that gaston posted, we can ignore pen support for now. touch event stuff seem like they should be working, so we should give those tests some attention

Patrick: so should we do action to look at this?

Olli: interop 2026 opening proposals soon, should submit a proposal for this

Mustaq: is Firefox going to propose it?

Olli: not discussed yet, but think Masayuki proposed for last time?

ACTION: Olli to raise possibility of touch test fixes for interop 2026 from Mozilla side

Recharter w3c/strategy#515

<smaug> Gecko has internal https://searchfox.org/firefox-main/source/dom/webidl/SimpleGestureEvent.webidl

[the group discusses idea of potentially adding gesture handling to IN SCOPE for the group's charter]

Philippe: are there any existing APIs that you would want to point to and say "this is what we want to standardise?"

Rob: we have internal gesture recognisers in browser code, but no existing API that's already there

Philippe: want to avoid having to start from scratch, as otherwise this needs to be incubated separate first

Patrick: the old API in webkit https://developer.mozilla.org/en-US/docs/Web/API/GestureEvent

Patrick: hammer.js https://hammerjs.github.io/

Patrick: not married to any particular syntax/approach - whether one single event (which then has an attribute for the type of gesture), or separate events for each gesture like hammer

Olli: would this then all be in our single pointer events spec, or would it be in a separate spec? and then we have UI events still without an editor...

Patrick: good question here... Philippe, would we as a group be able to separate out the gesture stuff in a separate document, or are we only chartered to work on one spec document?

Philippe: we could put it in charter as a separate document, but if we want to change our mind later we can integrate it

ACTION: Patrick to review/add any more info to the Recharter w3c/strategy#515 issue before Philippe takes it forward

ACTION: Patrick to make PR against charter for #515

any interesting issues/PRs to discuss?

Mustaq: I have one interop 2025 test

<mustaq> web-platform-tests/wpt#53685

Mustaq: tried to tweak before, until i realised Firefox failed. then tweaked again, and it should work

[mustaq and olli talk over the PR]

Patrick: if there's nothing else, thank you all as ever. We'll readjurn in two weeks' time