| 06:42 | <annevk> | bkardell: I think you only have to scroll to the right? |
| 08:02 | <annevk> | Noam Rosenthal: are you joining WHATNOT? |
| 08:05 | <keithamus> | I'm trying to join WHATNOT but its saying someone needs to let me in |
| 08:07 | <keithamus> | annevk: are you in? |
| 08:09 | <keithamus> | Panos Astithas: ^ |
| 08:23 | <zcorpan> | keithamus: I invited you, try again? |
| 08:23 | <Noam Rosenthal> | Noam Rosenthal: are you joining WHATNOT? |
| 08:24 | <keithamus> | Perhaps because I'm using a different email address? I'm using username @ gmail |
| 08:24 | <keithamus> | I can switch to my github one though |
| 08:25 | <zcorpan> | keithamus: I invited gmail address now instead |
| 08:25 | <keithamus> | Thanks! |
| 09:05 | <annevk> | keithamus: in https://github.com/whatwg/html/pull/9834 you still need to flip the order of the first two columns |
| 09:05 | <keithamus> | doing so right now π |
| 10:16 | <keithamus> | keithamus: in https://github.com/whatwg/html/pull/9834 you still need to flip the order of the first two columns |
| 12:45 | <sideshowbarker> | annevk: About https://bugs.webkit.org/show_bug.cgi?id=262635 if you think the change in https://github.com/WebKit/WebKit/pull/18689 should actually be made β so that we can have interop on this β then I can add some WPT tests to the PR (and export them after the PR is reviewed). And I guess I could could also write a spec patch for https://fetch.spec.whatwg.org/#dom-body-formdata to explicitly state that the Content-Disposition and Content-Type headers names must be parsed case-insensitively |
| 13:21 | <annevk> | sideshowbarker: I guess that makes sense for a multipart/form-data payload |
| 13:21 | <annevk> | sideshowbarker: there's https://github.com/web-platform-tests/wpt/issues/40946 |
| 13:22 | <annevk> | sideshowbarker: https://andreubotella.github.io/multipart-form-data/#multipart-form-data-parser has a WIP parser |
| 13:22 | <Andreu Botella> | very much WIP |
| 13:22 | <annevk> | It does seem to handle case-insensitivity of headers correctly, so there's that |
| 13:27 | <annevk> | sideshowbarker: also did a non-reviewer review of your WebKit PR |
| 14:00 | <sideshowbarker> | very much WIP |
| 14:31 | <annevk> | jarhar: so Chromium might be alone in having something close to what the spec suggests for type=email, accepting IDN and converting to Punycode, I agree with you that preserving that seems good |
| 14:32 | <annevk> | jarhar: in terms of supporting IP addresses things are somewhat bad, I guess mainly due to the impl being in terms of a regexp |
| 14:33 | <annevk> | jarhar: like yes, some IPv4 stuff works, but IPv6 does not at all; and invalid IPv4 works too and is considered valid |
| 14:39 | <annevk> | jarhar: although there's a weird aspect in that the normalization doesn't happen when the value is set through script |
| 15:16 | <Jeremy Roman> | Data URLs are weird. In data:text/html,Hello?world, the pathname is text/html,Hello and the search is ?world, but the response body is Hello?world |
| 15:17 | <Jeremy Roman> | (There may well be no way to avoid the hilarity that is how this interacts with URL patterns.) |
| 15:17 | <annevk> | Jeremy Roman: Chromium used to make #blah part of the body too |
| 15:18 | <annevk> | And that should still happen for javascript: URLs I think |
| 15:23 | <Jeffrey Yasskin> | Many non-hierarchical URL schemes are going to have weird semantics. If their semantics were non-weird, they'd be hierarchical. Definitely don't look at the semantics of https://w3c-ccg.github.io/did-resolution/#dereferencing-algorithm if you want to avoid weirdness. π |
| 16:21 | <jarhar> | jarhar: in terms of supporting IP addresses things are somewhat bad, I guess mainly due to the impl being in terms of a regexp |
| 16:26 | <annevk> | jarhar: judging on the latest in the thread, I don't think the way we do IP addresses in <input type=email> actually work in practice? |
| 16:27 | <annevk> | jarhar: as in, we don't reject s@123.123.123.123 currently, but if you actually wanted that you'd have to write s@[123.123.123.123] which we do reject |