| 00:05 | <domfarolino> | JakeA: you around? (Unsure if you're in CA given I/O this week) |
| 03:20 | <devsnek> | gg chrome |
| 03:20 | <devsnek> | didn't firefox already implement this? |
| 03:20 | <devsnek> | like without breaking anything |
| 03:47 | <domfarolino> | devsnek: are you referring to the message about preloaded audio object? |
| 03:47 | <devsnek> | yea |
| 03:51 | <domfarolino> | Ok. I have no info on that, was just curious |
| 03:59 | <devsnek> | chrome really locks developers into developing for chrome |
| 04:00 | <devsnek> | they had that "made for chrome" badge |
| 05:53 | <domfarolino> | devsnek: There's lots of examples for that kinda stuff: https://twitter.com/patrickkettner/status/993897844867186689, https://twitter.com/jaffathecake/status/986341782395863040 for example |
| 05:53 | <domfarolino> | wow, didn't think that would take up my _entire_ screen when expanded in IRCCloud |
| 08:14 | <annevk> | hsivonen: I'll work on a URL-side fix for that query issue |
| 08:15 | <annevk> | justJanne: isn't that proving the point though? |
| 08:35 | <hsivonen> | annevk: thanks |
| 09:11 | <annevk> | There has to a better way than asciitable.com to lookup hex digits |
| 09:12 | <annevk> | Another embarrassing thing I do: data:text/html,— for an em dash (similar for ellipsis) |
| 09:14 | <hsivonen> | I use https://r12a.github.io/app-conversion/ , https://www.binaryhexconverter.com/ and the search "ascii table" on DDG |
| 09:15 | <annevk> | I guess I can find solace in that other people reach for URLs too |
| 09:16 | <MikeSmith> | annevk: on MacOS you can type an em dash with Shift+Option+- |
| 09:16 | <MikeSmith> | hyphen |
| 12:23 | <JakeA> | domfarolino: I'm in CA yeah |
| 12:43 | <domfarolino> | JakeA: cool, wasn’t sure if you were up since it was late in UK. Do you know if there’s an issue open for FF about <script> </script> being needed after in-body <link>s to render earlier content? |
| 12:43 | <domfarolino> | Context: https://github.com/whatwg/html/issues/3672#issuecomment-387414275 |
| 12:57 | <JakeA> | domfarolino: not that I'm aware of, but the behaviour of link-in-body is undefined so it isn't really a bug |
| 12:58 | <domfarolino> | Ok, just wondering. Yeah getting this standardized would be a big win |
| 13:16 | <MikeSmith> | Domenic: as far as the WebIDL-less-ness of the Async Local Storage spec, curious if you’ve gotten any feedback from implementors about that |
| 13:16 | <Domenic> | MikeSmith: yeah, bholley from Mozilla seemed pretty happy since it means they wouldn't need a Web IDL-to-JS generator. And in Chrome it works better for us for the same reason. |
| 13:18 | <MikeSmith> | cool |
| 13:27 | <wanderview> | I liked that the layered APIs effectively gives the API a namespace under its import "url"... so in theory something that is broken could be completely rewritten with a new API under a new URL without breaking anything using the old one |
| 13:34 | <Domenic> | +1. Not sure it quite makes it worth the other costs of versioning (e.g. dev confusion) but it at least makes it feasible. And you could even no longer ship the old version if it gets low enough usage, but instead lazy-load it from some CDN or something. |
| 13:35 | <wanderview> | Domenic: you could even do (perhaps unwise) things like ship 5 versions of the API and see which one people actually use and prefer |
| 13:35 | <wanderview> | in the experimentation phase |
| 13:35 | <wanderview> | for example, if you're unsure where to draw the line between features/perf in a virtual-list API |
| 13:36 | <wanderview> | although I guess that could happen without being in the browser as normal modules on the web |
| 13:45 | <Domenic> | Heh yeah fun idea though, might be worth trying |
| 13:46 | <domfarolino> | There's not a way to observe the attributes of a script element in which code is executing right? |
| 13:47 | <domfarolino> | (from the code that is executing that is) |
| 13:51 | <wanderview> | domfarolino: in a window with classic scripts you can use document.currentScript I think |
| 13:51 | <wanderview> | no idea how that works with modules |
| 13:51 | <wanderview> | and I don't think workers have anything like that |
| 13:52 | wanderview | reads about non-standards document.onafterscriptexecute... |
| 13:56 | <domfarolino> | wanderview: oh wow, didn't even know document.currentScript was a thing |
| 13:59 | <wanderview> | domfarolino: thats the great thing about the web... there is always more stuff you would not expect to find there |
| 13:59 | <domfarolino> | lol, always true |
| 14:21 | <domfarolino> | Domenic: Hate to bother you but have you had a chance to look at https://github.com/w3c/web-platform-tests/pull/10722? Looks like we're getting this merged in Chrome very shortly and I'm fixing node's implementation too |
| 14:21 | <domfarolino> | s/this/countReset |
| 14:34 | <Domenic> | domfarolino: no problem, I thought one review from pwnall would be enough, but happy to also weigh in |
| 14:34 | <domfarolino> | Domenic: One review sounds good to me, I just don't have merge access or else I'd prob push it in |
| 14:35 | <Domenic> | Oh, we should get you that |
| 14:35 | <domfarolino> | That would be awesome :) Does someone just have to extend that access to me? |
| 14:38 | <Domenic> | Yeah, although they might delay because they're currently moving to a dedicated web-platform-tests org and are overhauling the permissions structure while doing so |
| 14:39 | <domfarolino> | Yeah, I was wondering if it would even be worth it since everything seems to be moving over, which is exciting! |
| 14:40 | <domfarolino> | Thanks!! |
| 14:42 | <Domenic> | annevk: I don't feel terribly qualified to review https://github.com/w3c/web-platform-tests/pull/10915, as I (and jsdom/whatwg-url in general) have ignored non-UTF-8 URLs, but let me know if you think I'm the best bet for getting it landed and I'll try to figure it out... |
| 14:47 | <annevk> | Domenic: I think only you and r12a are somewhat familiar with most of the changed files there |
| 14:47 | <Domenic> | Supposedly jsbell too :) |
| 14:47 | <annevk> | It's unfortunate those tests were allowed to skip over a lot of the differences by normalizing inputs and outputs |
| 14:47 | <annevk> | Otherwise we would have uncovered this issue quite a while ago I think |
| 14:48 | <annevk> | There's probably still quite a bit lurking there, but that'll have to wait for another day |
| 14:48 | <Domenic> | Yeah I imagine there are more principled ways of writing these tests in general. I just saw a lot of outstanding PRs and wanted to get at least some coverage. |
| 14:48 | <Domenic> | But yeah there's a lot of code duplication and such |
| 14:48 | <Domenic> | annevk: also FYI I'm going to try to spend an hour today hacking on a new live-url-viewer that is auto-deployed as we update jsdom/whatwg-url. Feature requests welcome. I'll just steal your HTML and CSS? |
| 14:49 | <annevk> | More problematic than that is what I'm removing here to some extent, where it removes percent encoding papering over browser differences |
| 14:49 | <annevk> | Domenic: I'd love reliable copy-and-paste in the address bar |
| 14:49 | <annevk> | Domenic: I suspect base64 or some such is the way to go for that |
| 14:49 | <Domenic> | Hmm what do you mean? |
| 14:50 | <annevk> | Domenic: so you test a URL and then you want to copy that URL test into another browser to see its outputs |
| 14:50 | <Domenic> | Ahhh right OK that makes sense |
| 14:50 | <annevk> | Domenic: currently the fragment is mangled, but copy-and-pasting is not reliable due to different escaping techniques |
| 19:44 | <domfarolino> | Domenic: There's no reason for line 7 of https://html.spec.whatwg.org/multipage/webappapis.html#fetch-the-descendants-of-a-module-script to have a duplicate "module script's settings object" text right? |
| 19:44 | <domfarolino> | Trying to make sure I'm not going crazy here |
| 19:45 | <domfarolino> | s/line/step |
| 19:46 | <Domenic> | domfarolino: that step looks really broken actually, it's not passing the right arguments at all :( |
| 19:46 | <Domenic> | well, the wrong order |
| 19:46 | <Domenic> | there are two settings objects in the arguments to ISGMFP |
| 19:47 | <Domenic> | but the order is wrong |
| 19:50 | <domfarolino> | Domenic: Ok so we're attempting to pass the `module script's settings object` as both the `fetch client settings object` _and_ the `module map's settings object`, but the order is just messed up. |
| 19:50 | <Domenic> | yeah |
| 19:50 | <domfarolino> | I can take a stab at this in my referrer policy PR |
| 19:51 | <domfarolino> | seems minor enough |
| 19:54 | <Domenic> | Awesome, thanks for fixing. Separate PR also welcome. |
| 19:55 | <jyasskin> | Domenic: Is there a good example of another spec using Streams under the hood? I want to base https://github.com/WICG/webpackage/pull/98 (bundled exchanges) on them and could make something up, but it'd be easier to imitate something you endorse. :) |
| 19:56 | <domfarolino> | Yeah, I think I'll actually do a separate one. |
| 19:57 | <Domenic> | jyasskin: hmm, "under the hood" doesn't sound like a great match, since streams, like promises, are pretty tied to their public API. Or is this designing an API? |
| 19:58 | <jyasskin> | This isn't designing an API, so that's good information. |
| 19:59 | <jyasskin> | My idea was that there's a network stream feeding information to a parser, and then that parser exposes a collection of streams for the individual contained exchanges, so I should use those streams in the parser's definition. |
| 20:17 | <Domenic> | Hmm, yeah, we've talked about making the concept more disconnected from the API, but it's a pretty hard problem just because of how I wrote myself into a corner to start with. |
| 20:18 | smaug____ | wonders if any browser implements "If a listed form-associated element has a form attribute specified, then that attribute's value must be the ID of a form element in the element's tree." |
| 20:18 | <smaug____> | that element's tree part in general |
| 20:18 | <smaug____> | better to test |
| 20:18 | <Domenic> | That sounds like a conformance requirement not related to browser implementations |
| 20:22 | <smaug____> | oh, indeed |
| 20:22 | <smaug____> | "If element is listed, has a form content attribute, and is connected, then" is more interesting case |
| 20:23 | <smaug____> | and that makes this work sanely only with Document and Shadow DOM |
| 20:23 | <smaug____> | Domenic: thanks |
| 20:23 | <Domenic> | One day we'll try to make the separation clearer... |
| 21:40 | <Domenic> | TabAtkins: can I make Bikeshed require a Date for certain document statuses? |
| 21:42 | <TabAtkins> | If you add the status to the list of snapshotStatuses https://github.com/tabatkins/bikeshed/blob/master/bikeshed/config/status.py#L80 Date is "recommended", and displays a Warning if you're missing it. |
| 21:59 | <Domenic> | Perfect |
| 22:14 | <lgrahl> | annevk: Still awake? |