Meeting minutes
Update on recharter and new PE3 snapshot
Patrick: from last time, i removed the unidirectional touch-action from PE3 w3c/
Patrick: based on that philippe requested a new dated snapshot w3c/
Patrick: from our last meeting, I updated our charter w3c/
Patrick: adding mention of gestures, and clarifying that we're not going to try to define the black box magic of how UA decides what is / isn't a gesture
Patrick: but we just want to have an API for content to then consume gestures once recognised
Patrick: based on that, Philippe requested the rechartering w3c/
Robert: does charter need to say anything more about mouse?
Patrick: didn't need to do anything more I think, as we have https://
Web Platform Tests https://wpt.live/pointerevents/ / https://wpt.fyi/results/pointerevents?label=experimental&label=master&aligned
Patrick: I believe you did something on this since last time, Mustaq
Mustaq: yes, I've looked at tweaking how failures are surfaced (?)
<mustaq> w3c/
Mustaq: that first list isn't completely precise, but goes in the right direction
Rob: wonder if we can filter by tests that relate to PE level *3* to weed out some things from 4
Mustaq: don't think tests are tagged in a granular-enough way / make a distinction between versions of PE
Rob: don't need to be super granular, but removing things that are explicitly in 4 would be good to exclude those
Mustaq: good idea to label. wonder if we can reuse existing labels like "optional" or "tentative"
Patrick: "tentative" sounds appropriate to me as an outsider...
Rob: "tentative" is usually reserved for things where there's not even a draft, so not quite the right one
Patrick: is it a problem creating a new label, do we HAVE to reuse an existing label?
Mustaq: no, not a problem
Rob: want to make sure we also remove things that aren't REALLY pointer events issues per se, but more pointerlock or similar
Mustaq: CSS specs go through stages ... wonder how they handle it
Rob: they'd have same problem, as tests aren't split by levels
Rob: having meta tags, or lists of tags...
Gaston: tests can be labelled for interop ... maybe the mechanism that's used for THAT can be used for our purpose
Mustaq: need to be careful with renaming filenames, as it will break lots of external references
Mustaq: so any method we choose should not touch them
Patrick: should we do an action for this? or we can also discuss this in person next week at TPAC...
Rob: there will be other spec people there too, so that might be good
ACTION: discuss labelling/thinslicing/filtering for different levels of PE for the WPT tests at TPAC
Any particular old or new issues that we want to start thinking about/tackling? https://github.com/w3c/pointerevents/issues
Mustaq: not issue related, but about a WPT that fails in all browsers
Mustaq: the test is wrong, but working on it
Mustaq: specific question for Olli - seeing Safari+Chrome fail similarly, but Gecko's behaviour is different
Mustaq: (context is failing test for targets removed from slot ?)
Mustaq: A is host, B inside shadow DOM. you get mouse down/up on B. in Safari/Chrome, the host A sees an actual (fake) event on A, with target A. In Firefox we see it bubbling to A, so target is B
Olli: i think it should behave like Safari/Chrome...
Mustaq: I will send you the test, we can look at it. Question is: are events meant to be "replicated" on host?
Rob: yes we change the target of the event when it goes to the light DOM
Olli: if you have a minimal test case
Olli: even if it's manual
Mustaq: will ping you in IRC
Gaston: is there a specific agenda already set for our meeting at TPAC next week?
Patrick: not crystallised anything yet ... was hoping to play it by ear, more of the same as our regular meetings, just easier to then discuss things while sitting around the same laptop etc. but if anything comes up, I'll send an email/set a proper agenda
Patrick: thank you all, see you physically/virtually next week at TPAC