Meeting minutes
‘Logical’ values for the ‘touch-action’ property #505
Patrick: as discussed in last meeting, i left comment on that issue about our decision to defer to Level 4
Add persistentDeviceId to PointerEvent #495
Patrick: as discussed in our last meeting, I've now merged this into Level 4 / gh-pages branch
Triage unlabelled issues https://github.com/w3c/pointerevents/issues
Patrick: want to see if anything is urgent and needs to be added to Level 3, or if we can defer
Rob: the issues asking for clarification may be things we want to do for v3, but need time to review
ACTION: review any unlabelled issues and mark them as v3 or future
Meta-issue: update WPT to cover Pointer Events Level 3 #445 w3c/pointerevents#445
https://
Rob: I have something started
<flackr> https://
Rob: discovered that Chrome at least doesn't seem to match spec behaviour. matches in one direction but not other
Rob: spec seems to imply that you can never capture a pointer if it's not in the active document. In chrome the capture works in one direction but not the other
Rob: haven't checked whether Firefox or Safari do the "correct" thing yet. depending on the outcome, we can come back and debate if the spec needs to allow capturing in one direction but not other
Rob: if i remember correctly, the issue I found was that if you try to set capture into an inner frame, Chrome allows it, but by spec we shouldn't
Rob: i'll post a repro example to the issue, then we can debate (and get results from Safari/Firefox to inform discussion)
ACTION: futher exploration needed
Level 3 and Level 4
Patrick: I've gone ahead and followed PLH suggestion to set up a Level 3 branch, and treat gh-pages as Level 4
ACTION: Patrick to catch up with PLH to make sure the new branch structure doesn't auto-publish to wrong place
<flackr> In https://
<flackr> > If the pointer is not in the active buttons state or the element's node document is not the active document of the pointer, then terminate these steps.
Patrick: so we want to make a change to spec anyway to clarify / don't keep failure silent and instead throw error
Patrick: I can see how you'd possibly want to allow capturing down (from parent document into frame) but not other way around, as it could be a third party script that should not be allowed...
Rob: keep in mind it can't be cross-origin
[Rob gave a very brief additional context/discussion on the thinking behind w3c/
RSSAgent, set logs world-visible