| 06:23 | <annevk> | domfarolino: I suspect wpt.fyi is outdated |
| 06:24 | <domfarolino> | That's my guess too |
| 06:26 | <annevk> | It might also use stable browsers or some such |
| 08:16 | <zcorpan> | jgraham: dunno if you keep up with github notifications, but I assigned this to you. https://github.com/w3c/web-platform-tests/issues/10799#issuecomment-390902187 |
| 08:16 | <zcorpan> | Let me know if I should bug someone else about enabling tests in gecko |
| 08:42 | <jgraham> | zcorpan: I guess I'm the right person to ask :/ |
| 08:42 | <jgraham> | And I saw that notification at least :) |
| 08:42 | <zcorpan> | Ok☺ |
| 08:53 | <zcorpan> | jgraham: do you prefer github issue or bugzilla? |
| 09:31 | <jgraham> | zcorpan: Well I'm less likely to forget about a needinfo on bugzilla |
| 09:32 | <jgraham> | GitHub really needs to copy that model |
| 11:09 | <yoav> | annevk: hey! I'm looking at https://github.com/whatwg/fetch/issues/726 and trying to see what the best place would be to define the Accept-CH-Lifetime cache |
| 11:09 | <yoav> | is there a previous notion of a persistent cache in Fetch? |
| 11:10 | <annevk> | yoav: CORS cache I guess |
| 11:10 | <annevk> | yoav: but that's very old so might not be that great an inspiration |
| 11:11 | <yoav> | annevk: who owns that cache? Or is it defined as a singleton? |
| 11:11 | <annevk> | yoav: that's a thing that's not defined atm, it should probably say that the user agent holds it |
| 11:12 | <yoav> | ok |
| 11:14 | <smaug____> | anyone want to compare Chrome to Safari on https://jsfiddle.net/wcvncxma/ (I don't have Safari at hand) |
| 11:15 | <yoav> | smaug____: the number outputs look the same for me (stable chrome vs. stable safari) |
| 11:15 | <smaug____> | ok, thanks |
| 11:23 | <annevk> | tobie: how do you generate the HTML Standard for pr-preview? |
| 11:30 | <tobie> | annevk: I just POST to https://build.whatwg.org/wattsi |
| 11:30 | <tobie> | annevk: see: https://github.com/tobie/pr-preview/blob/2bc932997ebe66aebccb0f089f11d0f9d1840281/lib/wattsi-client.js#L75-L113 |
| 12:01 | <annevk> | tobie: and then run the html-build scripts somehow? |
| 12:01 | <annevk> | tobie: it seems you'd miss a bunch of details otherwise |
| 12:01 | <tobie> | annevk: no, the wattsi server returns a zip, I just unpack it. |
| 12:03 | <annevk> | tobie: okay, that'll mostly work but get some things wrong |
| 12:14 | <JakeA> | If I want to write a WPT that tests a fetch to another origin, how do I do that? The linter is telling me off for using www.web-platform.test |
| 12:17 | <annevk> | JakeA: you could use /common/get-host-info.sub.js |
| 12:18 | <JakeA> | ahhh, cheers |
| 12:19 | <annevk> | JakeA: or https://wptserve.readthedocs.io/en/latest/pipes.html?highlight=domains#built-in-pipes |
| 12:35 | <yoav> | annevk: can you take a look at the (still rough) PR at https://github.com/whatwg/fetch/pull/729 and let me know what you think? There are several definitions still missing and some things there are a bit hand-wavy, but wanted to get your opinion on this early draft |
| 12:41 | <tobie> | annevk: I only build the multipage version, btw |
| 12:41 | <tobie> | annevk: s/build/use/ |
| 12:41 | <tobie> | annevk: what things would this get wrong? |
| 12:43 | <annevk> | tobie: scripts at https://github.com/whatwg/html-build/blob/master/build.sh#L343 are not applied |
| 12:44 | <tobie> | annevk: on the server? why not? |
| 12:46 | <annevk> | tobie: because the build server is only a replacement for running wattsi itself |
| 12:46 | <tobie> | annevk: OK, I get it. |
| 12:48 | <annevk> | yoav: what kind of feedback are you looking for? |
| 12:50 | <yoav> | annevk: Is this PR a step in the right direction from your perspective? Should the client-hints list hang off the active document or its env settings object? |
| 12:50 | <yoav> | anything else you see there where I'm Holding It Wrong™? |
| 12:51 | <annevk> | yoav: I guess at a high-level more data types need to be explicit; e.g., is this cache a _list_ of entries? |
| 12:51 | <annevk> | yoav: active document seems wrong for workers |
| 12:51 | <annevk> | yoav: (and worklets) |
| 12:52 | <annevk> | yoav: does Accept-CH needs to be processed for each response or only by navigation responses? (and therefore needs some tie-in with HTML) |
| 12:54 | <annevk> | yoav: (the fact that you try to mutate the active document strongly suggests this should be in HTML instead; but it depends a bit on the earlier questions; if it should be in Fetch, modifying it should probably go via some task) |
| 12:55 | <annevk> | tobie: we could provide some endpoint for pr-preview that does both btw |
| 12:55 | <annevk> | tobie: at that point we could also make it so that it includes the commit snapshot warning |
| 12:56 | <annevk> | tobie: I'll file a tracking issue against html-build |
| 12:56 | <tobie> | annevk: that would be ideal. |
| 12:57 | <yoav> | annevk: which parts should be in HTML? the response header processing? the cache? |
| 12:58 | <yoav> | (I'm not sure if there are use-cases to setting Accept-CH{,-Lifetime} on subresource responses) |
| 12:58 | <yoav> | (but I suspect there aren't obvious ones) |
| 12:59 | <annevk> | tobie: filed https://github.com/whatwg/html-build/issues/156 |
| 13:00 | <annevk> | yoav: the processing; the cache depends a bit upon who accesses it |
| 13:02 | <yoav> | annevk: if processing moves to HTML, the cache will only be accessed by HTML (when setting the client-hint list, and when retrieving it, when initiating the client-hints list of a new env settings object) |
| 13:02 | <yoav> | so maybe worth while to move it there |
| 13:04 | <annevk> | yeah, if it's only for HTML that seems reasonable |
| 13:27 | <yoav> | annevk: any examples of response header processing in HTML? |
| 13:27 | <yoav> | https://html.spec.whatwg.org/#processing-link-headers seems the closest, but it's not very specific |
| 13:32 | <annevk> | yoav: we might not have much yet |
| 13:47 | <yoav> | annevk: Ok, any recommended place I should put the response processing bit in? |
| 13:53 | <tobie> | annevk: thanks. I've subscribed to it. Feel free to ping me directly on it so I can implement it. |
| 13:53 | <tobie> | annevk: (once it's ready.) |
| 14:21 | <annevk> | yoav: navigation I suspect |
| 17:55 | <Domenic> | annevk: filed https://github.com/whatwg/meta/issues/92 for review draft/snapshot coordination, maybe check if I missed anything |
| 17:56 | <Domenic> | It might be a bit tricky since we were planning on doing this earlyish Tuesday and TabAtkins might not be awake to merge the PR there... but we don't really want to merge it earlier if possible... |
| 17:57 | <TabAtkins> | ? |
| 17:57 | <annevk> | Domenic: we could do HTML first |
| 17:57 | <Domenic> | TabAtkins: basically what time will you be awake and able to merge a PR next Tuesday... |
| 17:57 | <Domenic> | annevk: the CSS refactoring affects both, but yeah as long as nobody builds a spec in the meantime... |
| 17:57 | <TabAtkins> | Next week? I wake up at 7am Pacific time |
| 17:58 | <Domenic> | Oh that should do the trick :) |
| 17:58 | <annevk> | That's half an hour into the scheduled meeting, that's pretty good :-) |
| 17:59 | <annevk> | Domenic: I suspect "Go back and add annoying warning to all previously-produced commit snapshots" may not be possible for whatpr drafts |
| 17:59 | <Domenic> | annevk: makes sense for whatpr, I'm more concerned about /commit-snapshots |
| 17:59 | <annevk> | Domenic: though maybe, if we really wanted to we could write some kind of script |
| 17:59 | <Domenic> | Right, that was my plan |
| 22:10 | <MikeSmith> | TabAtkins: friendly reminder about the syntax-highlighting thing (https://github.com/whatwg/wattsi/pull/63 is the PR for the change to the wattsi sources for consuming the output) |