| 07:08 | <annevk> | Does anyone here know what is happening with uievents and pointerevents? I see a bunch of issues moving? |
| 07:24 | <sideshowbarker> | Yeah, we’re moving whole chunks of the UI Events spec into the Pointer Events spec |
| 07:24 | <sideshowbarker> | Where the specific “we” actually doing the work of it is plh |
| 07:27 | <sideshowbarker> | but the broader “we” is, a group of implementors and editors — plus me and Xiaoqian and plh — who have been meeting to do stuff to get the UI Events to be actually implementable |
| 07:29 | <sideshowbarker> | https://docs.google.com/document/d/1iRFgqtReyomoCwZFEdRUdoW1ManD5t656CVUGJm7Xos |
| 07:35 | <sideshowbarker> | You’re welcome to join the meeting 😀 Next one is Feb 10, at 4pm CET |
| 13:37 | <Dominic Farolino> | When A.com requests a B.com subresource, only A.com service workers can intercept. Are there SOP/security concerns for that, or is it more about predictability or other concerns? I ask because foreign fetch was before my time and I can't tell how much of its dismissal was security-based, and also intuitively, B.com's server ultimately gets to respond to its subresources, so it doesn't seem wrong security-wise if its service worker were first involved. |
| 13:41 | <annevk> | Dominic Farolino: with double/triple/quadruple-keyed storage it becomes very heavy-weight. |
| 13:41 | <annevk> | Dominic Farolino: https://github.com/w3c/ServiceWorker/issues/1188 has some breadcrumbs from a quick search. |
| 15:35 | <julienw> | Hey folks! by chance, would you know why event.target is null after dispatching the event to an element inside a shadow dom, when it has bubbles: false, composed: false (esp composed: false I believe)? While it has a target when bubbles: true, composed: true(this is in chrome and firefox so I believe this is spec-ed and not a bug) I use this to know if an event has ever been dispatched before -- but maybe there's a better way? |
| 16:26 | <Dominic Farolino> | So the complexity is based on how we'd partition the service worker in a double+ keyed world? E.g., if two different sites both make requests for destination.com, we'd need two destination.com SWs running, each keyed under their respective initiators? |
| 16:31 | <annevk> | julienw: with bubbling we can retarget. Without bubbling there's nothing to retarget too and we can't reveal shadow tree nodes. I think we end up resetting all state at the end of dispatch, except for targets when clearTargets is false. |
| 16:32 | <annevk> | Dominic Farolino: that sounds right. There's the additional privacy and security concern that now destination.com gets to execute script with the blame going to whatever was in the address bar, despite the address bar party thinking it only embedded a harmless image. |
| 16:59 | <cwilso> | Doh. Github is currently experiencing issues, which may make our WHATNOT meeting challenging. |
| 17:51 | <Noam Rosenthal> | Lea Verou: the explainer for out of order streaming is here: https://github.com/WICG/declarative-partial-updates/blob/main/patching-explainer.md But we haven't updated it with the latest proposal (which is best shown in this example: https://gist.github.com/noamr/b50b56ca75dfdddfb243507078758618) |
| 19:34 | <Dominic Farolino> | For A.com requesting B.com/img as a subresource, I'd expect https://fetch.spec.whatwg.org/#determine-the-same-site-mode to compute a same-site mode of "unset-or-less", but I don't see any step that would yield that. Where am I wrong? I'm getting "strict-or-less" which has to be wrong. |