| 06:29 | <annevk> | Domenic: I think WebKit folks were interested in implementing, but wanted a real spec first |
| 06:29 | <Domenic> | From what I understand the draft PR is such a spec? |
| 06:31 | <annevk> | Well yeah, but also tentative |
| 08:26 | <sideshowbarker> | PSA: Comments at https://github.com/orgs/mdn/discussions/264#discussioncomment-3903551 welcome β specifically on what the URLs in MDN for Web/DOM API static methods and instance methods should look like (a separate question from what the actual article titles should look like) |
| 08:27 | <sideshowbarker> | Domenic: β |
| 08:28 | <annevk> | I was gonna look at Bikeshed for inspiration, but it seems it might still put these all in the same namespace and the reason it didn't cause a problem is because the instance json() is on the Body interface. |
| 08:28 | <sideshowbarker> | ah |
| 08:29 | <Andreu Botella> | it sounds like that might cause issues the moment that Bikeshed updates to have different namespaces |
| 08:29 | <Andreu Botella> | with builds using an earlier version of Bikeshed but the new data |
| 08:30 | <annevk> | Yeah, we'd get a bunch of broken IDs. Probably have to hardcode all the existing statics and coordinate with TabAtkins on the rollout |
| 08:33 | <Luca Casonato> | We should have just called the static method fromJSON π |
| 08:39 | <sideshowbarker> | As far as MDN goes, I am thankful that for the JavaScript docs at least, the titles/names and URLs for methods were done right from the beginning β so thereβs no problem to fix there That said, thereβs also been some suggestion that when (if) the title/name/URL changes get made to the API pages, then the JavaScript pages should also be changed to match what ends up being done for the API pagesβ¦ |
| 08:39 | <annevk> | Luca Casonato: maybe, but also, specs don't matter when it comes to design decisions: https://www.w3.org/TR/html-design-principles/#priority-of-constituencies |
| 08:47 | <Domenic> | PSA: Comments at https://github.com/orgs/mdn/discussions/264#discussioncomment-3903551 welcome β specifically on what the URLs in MDN for Web/DOM API static methods and instance methods should look like (a separate question from what the actual article titles should look like) |
| 08:47 | <Domenic> | I will try not to introduce any instance methods named foo_static ;) |
| 09:04 | <annevk> | Domenic: also so Chrome is no longer interested in ShadowRealm? |
| 09:05 | <Domenic> | I don't believe anyone is currently working on ShadowRealm at Google, but as with all TC39 features that have reached stage 3, we decided not to object to it in plenary. |
| 09:05 | <annevk> | We already did all the Exposed=* work, which should have been blocked on the HTML PR, but I forgot that was still outstanding. |
| 09:05 | <Domenic> | I think there might be community contributors, but IIUC they have done V8 work, not Chrome work |
| 09:07 | <annevk> | yulia: if you're around do you happen to know if Gecko is doing any of the non-JS work on ShadowRealm? |
| 09:07 | <yulia> | mgaudet: might be a good person to answer that |
| 09:13 | <Domenic> | Friggin javascript: URLs are a fractal of badness... |
| 09:15 | <Domenic> | The new model makes obvious questions like "do they interrupt ongoing navigations?" "when do they snapshot the previous page's URL?" "what state does the new document carry over from the previous document?" "what referrer is used for creating the new document?"... and now I need to test everything, I guess. Or at least put a bunch of XXX boxes to investigate. |
| 09:29 | <annevk> | I think you can celebrate it as a win that it reveals these kind of questions |
| 09:29 | <annevk> | I think that's something that typically happens (or maybe is supposed to happen) when you take a bad spec and formalize it better |
| 09:32 | <Andreu Botella> | so am I the only one working on non-JS-engine work on ShadowRealm (for Deno)? |
| 11:59 | <Ms2ger πππ> | I think WebKit has implemented the things that are in the HTML PR |
| 12:20 | <annevk> | Ms2ger πππ: right you are. https://bugs.webkit.org/show_bug.cgi?id=241448. Would still be good to hear from mgaudet for Gecko and also from Chromium, but it does seem like there's an implementation. And it's disabled because the HTML PR is still outstanding. Domenic thoughts? |
| 12:21 | <Domenic> | Yeah, that does help bump the priority for review, thanks for the detective work |
| 12:22 | <Ms2ger πππ> | Thanks! |
| 13:32 | <mgaudet> | So we've also done a high level sketch implementation of the HTML integration of ShadowRealms: https://bugzilla.mozilla.org/show_bug.cgi?id=1747059 Having a final spec to work against would certainly be helpful to us shipping; it's not my only blocker to shipping, but other things mostly just need effort. |