00:03 | <smaug> | annevk: note that directionality is part of interop24 |
00:03 | <smaug> | (this network connection might be indeed bad) |
00:06 | <smaug> | FWIW https://wpt.fyi/results/html/dom/elements/global-attributes?label=master&label=experimental&aligned&view=interop&q=label%3Ainterop-2024-dir |
00:09 | <annevk> | smaug: ah thanks, sorry for not paying attention here. And yeah, that "Submit" concern is a thing that'd be tricky for certain implementations I'd expect. |
01:14 | <smaug> | Not sure how much OSes have that kind of data locally. Probably not for many languages |
08:24 | <jfkthame> | annevk: There are a bunch of :lang -matching tests in wpt/tests/css/css-selectors-4/lang-* , but I guess this exact case isn't covered afaik. |
08:26 | <jfkthame> | annevk: curious discrepancy between Safari and Chrome on this expanded testcase: https://software.hixie.ch/utilities/js/live-dom-viewer/?%3Cdiv%20lang%3D1%3Eone%3C%2Fdiv%3E%0A%3Cdiv%20lang%3D2%3Etwo%3C%2Fdiv%3E%0A%3Cdiv%20lang%3D3%3Ethree%3C%2Fdiv%3E%0A%3Cdiv%20lang%3D4%3Efour%3C%2Fdiv%3E%0A%3Cstyle%3E%0A%3Alang(1)%20%7B%20background%3Alime%20%7D%0A%3Alang(%222%22)%20%7B%20background%3Alime%20%7D%0A%3Alang(%5C33)%20%7B%20background%3Alime%20%7D%0A%3Alang(%22%5C34%22)%20%7B%20background%3Alime%20%7D%0A%3C%2Fstyle%3E |
08:27 | <jfkthame> | I'm seeing Safari match all three of lines 2-4, while Chrome matches only line 3. (And Firefox matches none of them, which I would argue is more correct.) |
12:04 | <annevk> | jfkthame: as an update, we discussed this with the i18n WG. Domenic was interested in pushing a JS Intl API to expose this exact algorithm to JavaScript, hopefully leading to increased interoperability (assuming implementations do the proper thing and share the logic). I don't really understand the results of your tests. The first not matching makes sense as 1 is not an IDENT. Oooh, maybe Chrome just doesn't support non-IDENT arguments. That would make sense then. |
13:27 | <jfkthame> | "maybe Chrome just doesn't support non-IDENT arguments" ... yes, that indeed seems to be the case |