WHATWG on 2025-03-24

13:13	<Yagiz Nizipli>	https://blog.cloudflare.com/improving-web-standards-urlpattern
15:32	<gernknod>	hi all! i'm working with fetch keepalive and noticed the spec doesn't provide the user with a clear reason why the request returned a NetworkError. the spec at https://fetch.spec.whatwg.org/#http-network-or-cache-fetch describes if the 64kib limit is exceeded, a "NetworkError" should be returned, but that prevents any kind of reasonable error handling regarding the business code. the sending party could simply try again later, when the fetchGroup is empty again or there's enough space left, but since they don't know the reason, it might as well be a non-recoverable error and retrying won't bring any better results. does that make sense? is this the right place to discuss this? 😅
17:23	<Dominic Farolino>	Just FYI, I plan on landing a few customizable select HTML PRs towards the end of this week. They've been sitting for a while with addressed comments, 2 implementer support, stage 3, and they LGTM. I'm referring to: Content model: https://github.com/whatwg/html/pull/10586 selectedcontent: https://github.com/whatwg/html/pull/10633 UA stylesheet: https://github.com/whatwg/html/pull/10670 Probably the base appearance one too, assuming it also LGTM when I finish reviewing it: https://github.com/whatwg/html/pull/10629
17:27	<annevk>	Dominic Farolino: I'd like to review them before we land and I'm not sure I can get to it this week.
18:05	<Dominic Farolino>	Gotcha. I'll hold off on merging, let me know if you think you can review any this week or next.
19:31	<jmdyck>	https://html.spec.whatwg.org/#creating-a-source-set-from-attributes step 4's wording is a bit odd: "... source set does not contain an image source with a pixel density descriptor value of 1, and no image source with a width descriptor, ...". Since the second half (after the "and") doesn't have a verb, it would normally inherit the verb phrase from the first half, but that would be "does not contain no image source with a width descriptor". Seems like there are various ways that could be worded better.
22:03	<stelar7>	Hi all! Im looking at WEBIDL, spesifically converting a JS value to an IDL union type. Given a type like `(DOMString or sequence<DOMString>)`, and using the algorithm https://webidl.spec.whatwg.org/#js-union. If passed a value like `{a: 'hey', b: 2}`, we get to 11.2.2, fail the check, then to 15, and call toString on the object, returning a value like `[object Object]`. This seems wrong.. Is there something im missing here that would make this go to step 20 and throw, as I expect it to?
23:03	<akaster>	Does anyone have any insight on why the "magicification" of `<br>` and `<wbr>` styles has stalled in https://github.com/whatwg/html/pull/2298 ? I guess poking around it seems like it's tricky to write out in a reasonable manner what the behavior should be?