| 01:01 | <sideshowbarker> | It looks like api.csswg.org may be having some issues. My CI runs are failing on the fetch spec: https://github.com/whatwg/fetch/runs/5472598698?check_suite_focus=true |
| 08:16 | <sideshowbarker> | Do we anywhere have in any spec a normative definition of the term “same-site”? |
| 08:20 | <sideshowbarker> | I thought https://w3c.github.io/webappsec-fetch-metadata/#sec-fetch-site-header used to have at least a clear normative definition for the same-site value of the Sec-Fetch-Site header — and used that when I wrote up https://stackoverflow.com/questions/66115532/what-does-the-sec-fetch-site-header-mean-why-is-the-origin-header-undefined/66228794#66228794 — but now even that spec no longer seems to make clear what it means by “same-site” (at least it’s not clear for developers reading the spec, in contrast to anybody who might be implementing it) |
| 08:20 | <Andreu Botella> | https://html.spec.whatwg.org/multipage/origin.html#same-site |
| 08:22 | <sideshowbarker> | Andreu Botella: OK thanks yeah I had seen that. So I guess that’s the best we’ve got — but I doubt it’s going to mean much to the average web developer reading it. |
| 08:22 | <Andreu Botella> | That seems to be the normative definition. I agree that it's not great for average devs |
| 08:22 | <sideshowbarker> | …it doesn’t even directly reference “registrable domain” |
| 08:23 | <sideshowbarker> | yeah |
| 08:24 | <sideshowbarker> | I think the concept of “registrable domain” is very clear and intuitive to everybody, so I’m glad that was coined and we have that defined in (the URL spec) at least — it’s been a big improvement over directly mentioning the public-suffix list and the “eTLD+1” term |
| 08:30 | <sideshowbarker> | by the way, I’m confused a bit by https://html.spec.whatwg.org/multipage/origin.html#sites, where we have the term “site” defined as this:
…while “obtain a site”, in the common case, reduces to being defined as this:
|
| 08:32 | <sideshowbarker> | I would have expected instead that the term “site” were defined as “A site is an opaque origin or a scheme-and-registrable-domain” |
| 08:59 | <annevk> | sideshowbarker: I think that's because an IP address is not a registrable domain |
| 08:59 | <annevk> | That is, I think I did it that way because of that, but as has been said above, it's all a blur 🙂 |
| 13:00 | <Yoav Weiss> | The infra spec no longer seem to have a definition for pair: https://infra.spec.whatwg.org/#pair |
| 13:00 | <Yoav Weiss> | Is there something I should use instead? Or should I simply unlink uses for pair in specs? |
| 13:00 | <Jake Archibald> | https://infra.spec.whatwg.org/#tuples ? |
| 13:02 | <Yoav Weiss> | yeah, that kinda works |
| 13:13 | <Yoav Weiss> | Hmm, is bikeshed broken on the bots for other folks as well? Or am I holding it wrong? https://github.com/w3c/largest-contentful-paint/runs/5480497271?check_suite_focus=true |
| 13:47 | <sideshowbarker> | To find out which other specs reference the term “same site” from the HTML spec, is there some way I can do that? Shepard or Webref? |
| 13:49 | <Ms2ger 💉💉> | sideshowbarker: https://dontcallmedom.github.io/webdex/s.html |
| 13:49 | <Ms2ger 💉💉> |
|
| 13:49 | sideshowbarker | looks |
| 13:50 | <Ms2ger 💉💉> | (How cool is that, btw.) |
| 13:50 | <sideshowbarker> | sideshowbarker: https://dontcallmedom.github.io/webdex/s.html |
| 13:50 | <sideshowbarker> | (How cool is that, btw.) |
| 13:52 | <Ms2ger 💉💉> | I've only known about it for two weeks myself: https://www.w3.org/mid/e6cfe633-a041-e5ba-c12d-9ed7fce59e51@w3.org :) |
| 13:54 | <Yoav Weiss> | The CSSOM links from HTML (e.g. https://drafts.csswg.org/cssom-view/#run-the-resize-steps) seem broken.. |
| 14:01 | <Ms2ger 💉💉> | https://github.com/w3c/csswg-drafts/issues/6725 |
| 14:01 | <Ms2ger 💉💉> | They're down mostly every day at this point |
| 14:02 | <Ms2ger 💉💉> | You can try https://andreubotella.com/csswg-auto-build/ |
| 14:08 | <annevk> | Yoav Weiss: I folded pair into tuple a while back as it had a syntax that people found confusing; sorry for the breakage |
| 14:08 | <annevk> | Yoav Weiss: context at https://github.com/whatwg/infra/pull/413 and linked issue |
| 14:09 | <Yoav Weiss> | No worries! tuple seems like a better fit for what LCP was doing anyway |
| 14:09 | <annevk> | TabAtkins: do you know about the Bikeshed breakage? |
| 14:10 | <annevk> | TabAtkins: ImportError: cannot import name 'Literal' from 'typing' (/usr/local/lib/python3.7/typing.py) |
| 14:21 | <Luca Casonato> | make local still works for me, but it seems CI runs are completely down now |
| 14:48 | <annevk> | Luca Casonato: did you get the latest Bikeshed though? |
| 14:48 | <annevk> | https://github.com/tabatkins/bikeshed/commit/c24e3ab95e18afe1ca9ee7e4a12269defbb5de0a is a pretty recent commit and seems related to the error above |
| 14:48 | <Luca Casonato> | updating now |
| 14:49 | <Luca Casonato> | yup, still works for me |
| 14:49 | <Luca Casonato> | I'm on Python 3.8.10. CI / build server looks to be on pyton 3.7.xx. maybe that makes a difference? |
| 16:01 | <Domenic> | OMG WebDex I've been looking for this for years. |
| 16:15 | <Fut Nada> | Hello. Is there anyone here who has access to a bookmaker site? I'm looking for one |
| 17:21 | <TabAtkins> | Yes, I knew about the issue; CI reported it to me right at the end of the day. Unfortunately the API server runs tip-of-tree Bikeshed (and is running 3.7, the version that was having problems), so it was broken until just now when I fixed it. |
| 17:24 | <TabAtkins> | (Anything running pip-installed Bikeshed should, to the best of my ability, always work great; at minimum, I verify it cycles green before I push new versions, and try to push versions soon after new fixes or features.) |
| 17:27 | <DerekNonGeneric> | Fut Nada: hello, this is not the channel for questions like that; this channel is for questions about specifications and has generally been about specification development and sometimes implementation |
| 18:00 | <annevk> | TabAtkins: thanks for the update! |
| 18:00 | <annevk> | It's so weird that GitHub doesn't jump to the latest attempt when you retrigger CI |