| 00:32 | <devsnek> | should localstorage instances called on windows between iframes still work? |
| 00:33 | <devsnek> | i found an interesting bit of js that works around a website deleting localStorage by creating an iframe and setting window.localStorage from iframe.contentWindow.localStorage |
| 00:54 | <Domenic> | That seems reasonable, we generally don't inspect callstacks in web specs at least |
| 02:57 | <devsnek> | yeah i just wasn't sure if it was some sort of like, branding thing that was broken. if it comes down to call stacks the behavior seems reasonable. |
| 06:37 | <annevk> | devsnek: web platform brands are cross-realm |
| 07:15 | <sideshowbarker> | what’s a “brand” in this context? Never heard the term used as such, and I don’t find it in the ES spec, so I guess it’s coined somewhere else? |
| 07:21 | sideshowbarker | finds “Passes various brand checks” in code comment in the WebIDL spec, but no reference anywhere to what a “brand” is |
| 07:22 | <Ms2ger 💉💉> | https://webidl.spec.whatwg.org/#implements , basically |
| 07:22 | sideshowbarker | looks |
| 07:23 | <sideshowbarker> | thanks — so I guess “brand” and “brand check” are some informal terms for that? |
| 07:26 | <sideshowbarker> | OK, finally found https://github.com/tc39/how-we-work/blob/main/terminology.md#brand-check |
| 07:27 | <sideshowbarker> |
|
| 07:48 | <annevk> | sideshowbarker: yeah, it's basically some internal slot that only implementers can set and inspect which prevents emulating the object in pure JS (it can often be observed through side effects of various features) |
| 08:46 | <zcorpan> | Do we care about unrelated drive-by fixes being in their own PRs? https://github.com/whatwg/html/pull/7829/files#diff-41cf6794ba4200b839c53531555f0f3998df4cbb01a4d5cb0b94e3ca5e23947dL75618 |
| 09:18 | <annevk> | zcorpan: it depends, wording nits and whitespace changes are fine, if it becomes a large part of the PR it's probably best split |
| 12:35 | <zcorpan> | annevk: ok, I guess this markup fix is small and therefore no need to split |
| 15:26 | <annevk> | Yeah, a markup fix should be fine |
| 15:53 | <Domenic> | Yeah my criteria is a mix of "Is the diff really confusing when you smash them into one PR" and "would people really want to see this in the GitHub commit log". So like, removing that one <td> line does not create a confusing diff and would be a bit silly to see while browsing the commit log. Whereas, if you reindent a big algorithm that is only slightly related to the part of the spec you're changing, that is worth a separate commit. |
| 16:52 | <smaug> | readonly attribute Uint8Array data; that isn't anything magical and can't be null or undefined, right? |
| 16:58 | <annevk> | smaug: yup |
| 16:58 | <smaug> | nm, found the existing bug about midi events 🙂 https://github.com/WebAudio/web-midi-api/issues/168 |
| 16:59 | <annevk> | "Basic" details of events are hard, film at 11 |
| 17:01 | <smaug> | Implementing what is spec'ed is hard. (turns out a crash in Gecko's midi impl happens because of this). If one would have tried to implement what is spec'ed, the spec would have been fixed long ago 🙂 |
| 17:03 | <annevk> | Browsers are hard; let's go shopping. See y'all tomorrow. |
| 20:53 | <Domenic> | annevk: smaug: do I recall Firefox trying to implement https://wicg.github.io/page-lifecycle/ ? Trying to figure out the maintenance story for that as apparently I am an editor... and in particular wondering if we should move it into HTML or not. |
| 21:34 | <smaug> | I'm not aware of any implementation plans for that |
| 21:35 | <smaug> | Would need to re-review it again now that Fission is shipping and bfcache handling is a tiny bit different because of that. |
| 21:35 | <smaug> | (page lifecycle API used to have issues with bfcache, but I think most of them were resolved ) |
| 22:59 | <Domenic> | OK, maybe I just remembered all of your reviews and pointing out those issues. |
| 23:00 | <Domenic> | Will plan on keeping it in WICG... and in particular put it on the list of things to rebase on https://github.com/whatwg/html/pull/6315 |