| 02:12 | <tusf> | tusf: it's a tree. Not aware of articles, but in the early days linked list vs array for representing children was a topic. |
| 06:55 | <annevk> | Benjamin Gruenbaum: FWIW, composing AbortSignal objects is still something we should do, but we need someone to do the work essentially :-) |
| 06:56 | <annevk> | Benjamin Gruenbaum: there's also another tricky issue about making AbortSignal work well with timeouts, for which you really want to expose a cancel reason (or use a different exception, as XHR does) |
| 08:42 | <annevk> | Jake Archibald: for the issue you just filed you haven't explained what an "isolated browsing context group" is |
| 08:43 | <Jake Archibald> | annevk: I thought it'd be obvious it's https://html.spec.whatwg.org/multipage/browsers.html#cross-origin-isolation-mode, or do we have another form of isolation? |
| 08:44 | <annevk> | Okay, I was thinking it would be that, but I wasn't sure. |
| 08:44 | <Jake Archibald> | I'll add a note |
| 08:47 | <annevk> | Jake Archibald: also is it intentional that it's first about bcg and later about bc? |
| 08:51 | <Jake Archibald> | annevk: made some edits. Did I fix the bit you mean? |
| 08:55 | <annevk> | Jake Archibald: not the second question; the post seems to be about bcgs but the questions are about bcs |
| 08:56 | <Jake Archibald> | annevk: I thought it was impossible for a top level to change browsing context without also changing group? |
| 08:56 | <Jake Archibald> | And a browsing context cannot exist in two groups |
| 08:56 | <annevk> | Jake Archibald: yeah that is correct, maybe I'm being too pedantic |
| 08:57 | <Jake Archibald> | I'll add those points in as asserts 😀 |
| 08:58 | <Jake Archibald> | Added some clarifying text to question 1 |
| 09:05 | <annevk> | Jake Archibald: so the main reason to care about bcg/bc identity is because of popups, right? Is there anything else you are concerned with here? |
| 09:07 | <annevk> | Jake Archibald: question 2 is rather complex; I would not expect identity to be preserved if they are not visited in order (e.g., if you visit something else in between). As in, the traversal is what causes the swap and if you then traverse to non-isolated and then to isolated again, two more swaps. |
| 12:12 | <annevk> | Glad we migrated away from Travis: https://twitter.com/peter_szilagyi/status/1437646118700175360 |
| 18:13 | <sideshowbarker> | MDN PSA: as of a few moments ago, the sources for all the docs in the https://developer.mozilla.org/en-US/docs/Web/API tree are now in Markdown rather than HTML |
| 18:17 | <sideshowbarker> | And the all the https://developer.mozilla.org/en-US/docs/Web/JavaScript and https://developer.mozilla.org/en-US/docs/Web/CSS sources also got converted to Markdown recently, so that means the core of MDN reference docs is now all a lot more pleasant to work on and contribute to than it was before |
| 18:19 | <sideshowbarker> | The https://developer.mozilla.org/en-US/docs/Web/HTTP and https://developer.mozilla.org/en-US/docs/Web/HTML trees and other remaining trees will soon be following |
| 18:25 | <bakkot> | is there a list somewhere of the shipped APIs that WHATWG/browsers would like to deprecate? I'm considering shipping something in the script I build for work which will touch a whole bunch of APIs, which would mess up any use counters (this script runs on a nontrivial fraction of all page loads across the web). if there's a list of APIs which are being considered for deprecation (or which have use counters for any other reason) I can try to avoid those. |
| 18:34 | <sideshowbarker> | bakkot: FWIW all data in https://github.com/mdn/browser-compat-data/tree/main/api has a true/false deprecated field that you could filter on to programmatically generate such a list |
| 18:44 | <bakkot> | sideshowbarker: that's helpful, thanks! but I'm also thinking of things like alert, which I don't think is formally deprecated but I know there is a desire to deprecate it |
| 18:44 | <bakkot> | (though alert itself is not one I would touch either way) |
| 21:21 | <Domenic> | bakkot: the other big ones I'm aware of that will be multi-year efforts and as such haven't gotten much of a callout yet are document.domain and unload/beforeunload handlers |
| 21:42 | <bakkot> | Domenic: sweet, thanks, I will stay away from those |
| 21:53 | <Domenic> | bakkot: asking internally I also got window.webkitStorageInfo and navigator.webkitTemporaryStorage/webkitPersistentStorage |