WHATWG on 2023-02-15

09:17	<annevk>	Domenic: not using "return" in "in parallel" to exit is still a thing, right? (Context: https://github.com/whatwg/fs/pull/95.)
10:54	<annevk>	krosylight: there's "<li><p>Run the following steps [=in parallel=]:" too; I guess with regards to CI maybe we need to wait a little
10:55	<annevk>	krosylight: at least https://github.com/whatwg/fetch/pull/1605 suggests there's some nonsense going on
11:05	<krosylight>	Seems it's https://github.com/speced/bikeshed/issues/2471
12:36	<sideshowbarker>	annevk: Is there anything blocking https://github.com/whatwg/fetch/pull/1442 (ORB)?
12:37	<sideshowbarker>	ah, now I see the checklist — “Define "validate a partial response" (and Content-Range header parser)”
12:45	<annevk>	Yeah, there's a couple minor things still. I'd like to get back to it at some point. Or find someone who wants to take over.
12:45	<sideshowbarker>	OK
12:47	<sideshowbarker>	Chrome is still shipping CORB I guess? I think it’s a continuing point of confusion to developers — both that they get this behavior only in Chrome, and also now that there’s no spec which defines the behavior they’re seeing
12:48	<sideshowbarker>	Developers posting to Stack Overflow almost always assume it’s some part of CORS
12:49	<annevk>	They ship a variant of the original CORB and I think are still planning to converge on ORB
12:49	<annevk>	I mean, they're not entirely wrong; it's all part of the SOP
14:47	<Yagiz Nizipli>	It's harder to visualize this, than to implement it :)
15:30	<annevk>	:ouch:
15:38	<Domenic>	Domenic: not using "return" in "in parallel" to exit is still a thing, right? (Context: https://github.com/whatwg/fs/pull/95.) Still a thing. Not sure where we've landed is great, but current style is to use "abort these steps" for nested algorithms and reserve "return" for top-level ones.
15:52	<annevk>	Domenic: so callbacks can never return anything? In my mind the divide was main thread / in parallel
15:54	<annevk>	I guess it might need some more experience
18:06	<annevk>	Domenic: should we consider a dedicated Maintainers section in standards for https://github.com/whatwg/sg/issues/210 as well? Maybe just something to keep in mind if it keeps coming up
23:55	<Domenic>	Domenic: so callbacks can never return anything? In my mind the divide was main thread / in parallel So, again, none of this is very principled or even good. But the style I've been using is that inline callbacks can't return things; if it has a real return value you need to pull it out into a separate named algorithm.
23:55	<Domenic>	The more I describe it, the worse it sounds, so maybe we should have a real discussion about coming up with something better and more principled.
23:56	<Domenic>	You mean like, separate from the Acknowledgments?