| 03:14 | <MikeSmith> | anybody else having problems with GitHub issues right now? |
| 03:14 | <Domenic> | yep, 500s |
| 03:14 | <MikeSmith> | getting 500s |
| 03:14 | <MikeSmith> | ah OK |
| 03:15 | <MikeSmith> | well rhwn I guess I need to find something else to do for a while |
| 03:28 | <MikeSmith> | kinda weird that affects only some issues/PRs but not others |
| 08:00 | <tobie> | MikeSmith: the app is probably running on multiple servers in parallel |
| 08:14 | <tobie> | marcosc: re permission spec, have you considered opening up the API to ask for permission passing an object, not a string, e.g.: |
| 08:14 | <tobie> | let gyro = new Gyroscope(); permissions.request(gyro).then... |
| 08:15 | <tobie> | Developers wouldn't need to know what permissions they would need for an object, just pass the object instead to be able to use it. |
| 09:10 | <tobie> | mounir: ^ |
| 10:02 | <mounir> | tobie: the permission request takes an objec |
| 10:02 | <mounir> | tobie: but it's not a random object |
| 10:09 | <tobie> | Mounir: I know. I'm asking if you've considered a mechanism by which you'd ask for permission not by using a string or permDesc object that you had to know about as a developer, but just create the object you need and pass it to the permission API to get permission to use it. |
| 10:10 | <mounir> | tobie: no, it wasn't |
| 10:11 | <tobie> | mounir: would you, though? |
| 10:12 | <mounir> | tobie: I no longer actively work on this so my opinion probably doesn't matter |
| 10:13 | <tobie> | mounir: who does? |
| 10:13 | <mounir> | tobie: but I have two concernS: you will have to have a way thot doesn't take an object and the object solution sounds really hard to spec |
| 10:13 | <mounir> | or it would have to be spec'd in the client spec which might be fine |
| 10:13 | <mounir> | tobie: jeffrey |
| 10:14 | <mounir> | but I think he is on paternity leave so might be hard to get a hold of him |
| 10:14 | <tobie> | mounir: indeed |
| 10:15 | <tobie> | The objects could expose a special property or method. |
| 10:15 | <tobie> | That would hold a perm descriptor. |
| 11:56 | <zcorpan> | MikeSmith: can you rebuild the checker so the messages use the current spec's content models in the error messages? |
| 11:57 | <MikeSmith> | yeah |
| 11:57 | <MikeSmith> | lemme do that right now |
| 11:57 | <zcorpan> | 👍 |
| 12:04 | <MikeSmith> | zcorpan: ok, now updated |
| 12:05 | <zcorpan> | MikeSmith: thanks |
| 12:06 | <MikeSmith> | thanks for the heads-up about it |
| 12:06 | <MikeSmith> | the excerpts were out of date in other places as well |
| 12:07 | <zcorpan> | MikeSmith: what are the steps to fix this locally? i tried `python ./build/build.py all` but it uses the old messages |
| 12:08 | <MikeSmith> | yeah it does not download the spec copies automatically at startup as it does for the wiki things |
| 12:08 | <MikeSmith> | you need to do this: |
| 12:08 | <MikeSmith> | cd resources/spec && make update && make |
| 12:10 | <zcorpan> | ok. would it be a good idea to have that happen as part of build.py, so it is kept up-to-date? |
| 12:27 | <MikeSmith> | zcorpan: yes it would be right now it would then add a new requirement that anybody running the python build script would also need to have make make(1) installed |
| 12:28 | <MikeSmith> | so I think the right way to do it would be to move it into the python script |
| 12:28 | <zcorpan> | MikeSmith: ah ok, yeah |
| 12:28 | <MikeSmith> | anyway I think that is not much work |
| 12:28 | <MikeSmith> | will take a look right now |
| 12:31 | <MikeSmith> | OK yeah it should be relatively easy |
| 12:36 | <MikeSmith> | zcorpan: https://github.com/validator/validator/issues/393 |
| 13:49 | <annevk> | tobie: I'd like to make some progress on https://www.w3.org/Bugs/Public/show_bug.cgi?id=27301 |
| 13:50 | <annevk> | tobie: if we were to expose "this object", what would we call it and how would we expose it (argument vs ambient value)? |
| 13:50 | <annevk> | tobie: the "context object" thing from DOM is still spreading and that's not particularly great for a hack |
| 13:53 | <tobie> | I'm busy with the sensor work atm (and also admin stuff), was planning to spend Thursday + Friday on WebIDL this week. |
| 13:53 | <tobie> | annevk: would you be available to chat about this on either one of these days? |
| 13:54 | <tobie> | annevk: it would be useful to get me up to pace. |
| 13:54 | <annevk> | tobie: sounds good |
| 13:54 | <tobie> | annevk: cool |
| 14:10 | <tobie> | annevk: fwiw, I don't really like ambient values, I think it makes everything harder to reason about and adds magic. |
| 14:10 | <annevk> | tobie: fair, but note that the alternative will be a ton of work |
| 14:10 | <annevk> | tobie: and even more refactoring of all the things |
| 14:11 | <tobie> | annevk: oh--well in that case... :D |
| 14:11 | <daleharvey> | Anyone happen to know if it is ok to do <img srcset="source.png any, small.png 100w" ? |
| 14:11 | <daleharvey> | (trying to figure out how to convert manifest icons to a srcset) |
| 14:14 | <Domenic> | daleharvey: I think you're supposed to do src="source.png" srcset="small.png 100w" |
| 14:15 | <Domenic> | annevk: tobie: I think it should be either "this value" (like ES), "this" in some special font/bold (nice and short, and helps with internal slots like this.[[foo]]; streams does this); or maybe "this X object" (very precise, what we're doing in HTML these days). |
| 14:15 | <Domenic> | I guess I should leave that comment on the bug |
| 14:16 | <daleharvey> | Domenic: ok cool cheers |
| 14:17 | <annevk> | I'd like just "this" |
| 15:36 | <zcorpan> | MikeSmith: FYI I failed to push changes to jdm's web-platform-test branch using checkout-pr; https://github.com/w3c/web-platform-tests/pull/1427#issuecomment-257593016 |
| 15:51 | <MikeSmith> | zcorpan: what failed? does jdm have the branch writable by you? |
| 15:51 | <zcorpan> | MikeSmith: he said he checked the checkbox in the PR |
| 15:51 | <zcorpan> | MikeSmith: git push failed |
| 15:51 | <MikeSmith> | OK that sounds like a github bug |
| 15:52 | <MikeSmith> | checkout-pr is not doing anything special in that regard as far as I can see |
| 15:52 | <zcorpan> | yeah |
| 15:53 | <MikeSmith> | you using the https URL for the remote, not the ssh one? |
| 15:53 | <MikeSmith> | if so I wonder if that may affect it |
| 15:53 | <MikeSmith> | I have never tested myself with https URLs for PRs |
| 16:18 | <MikeSmith> | zcorpan: I also get “ERROR: Permission to jdm/web-platform-tests.git denied to sideshowbarker.” when using the SSH URL for that remote |
| 16:18 | <MikeSmith> | when trying to push |
| 16:18 | <zcorpan> | using https, yes |
| 16:20 | <MikeSmith> | yeah so it seems like if jdm has the box checked, tehre must be some github bug causing it to not actually be making the branch writeable by upstream committers |
| 16:58 | <tobie> | Is there a best practice to write an async abstract operation that returns a value? |
| 17:18 | <Domenic> | tobie: not really yet but I have done e.g. https://html.spec.whatwg.org/multipage/webappapis.html#fetch-a-classic-worker-script |
| 17:19 | <tobie> | Domenic: thanks. Will have a look. |
| 19:15 | <TabAtkins> | Domenic, daleharvey: Domenic's suggestion (putting the "plain" one in src) is preferred, but just for back-compat. The two are otherwise equivalent. |
| 20:21 | <zcorpan_> | tobie: returns a value sync (early) or when the async part is finished? |
| 20:22 | <tobie> | zcorpan_: the latter |
| 20:22 | <tobie> | I need algo prose for Promise.all([...]) |
| 20:23 | <TabAtkins> | That exists in the Promises Guide |
| 20:23 | <Domenic> | <a>waiting for all</a> |
| 20:23 | <Domenic> | It's not defined super-great but it's what we've got for now |
| 20:24 | <TabAtkins> | https://www.w3.org/2001/tag/doc/promises-guide/#aggregating-promises |
| 20:24 | <TabAtkins> | Whoops, beat me. |
| 20:25 | <tobie> | TabAtkins, Domenic: except its for abstract operations which don't return a promise (the resolved value isn't exposed, e.g. getting a permission or figuring out if a sensor is present) |
| 20:26 | <TabAtkins> | Ah, so you've got some async abstract ops, not user-exposed, that you want to do *the equivalent of* Promise.all on? |
| 20:27 | <Domenic> | Oh dang yeah |
| 20:27 | <Domenic> | I think I did that for modules too |
| 20:28 | <Domenic> | tobie: https://html.spec.whatwg.org/#fetch-the-descendants-of-a-module-script step 5 |
| 20:29 | <tobie> | TabAtkins: yes |
| 20:31 | <tobie> | Domenic: this (well except I was hoping for something less verbose) |
| 20:31 | <tobie> | Thanks folks. Much appreciated. |
| 20:34 | <gsnedders_web> | How does the AC vote at PR work in W3C land? Is it merely dissent is publicly noted? |
| 20:35 | <tobie> | TabAtkins: think you could give https://github.com/heycam/webidl/pull/218 a quick scan? |
| 20:36 | <tobie> | TabAtkins: fixing the webidl brokenness following the latest bikeshed update |
| 20:37 | <tobie> | TabAtkins: It did not ask me for "for=/". |
| 20:38 | <TabAtkins> | What do you mean by that last bit? |
| 20:40 | <tobie> | TabAtkins: that I followed the link to you explanation and tried reading it while soothing a 1yo. :-/ |
| 20:40 | <TabAtkins> | Hmm, this is firing more than I expected. It looks like I'm not excluding alternative versions of the current spec when checking if any other spec defines a for-less version of the text. |
| 20:41 | <TabAtkins> | In particular, webidl-1 is the thing causing you problems. Gimme a sec and I can probably wipe out most or all fo these errors for you. |
| 20:41 | <gsnedders_web> | https://gist.github.com/gsnedders/a20f06ebf3d32554e61333f93fbc66c0 --- does this seem like a not stupid thing to say about the W3C? |
| 20:42 | <gsnedders_web> | (feel free to respond when I soon vanish, gsnedders will still be here :)) |
| 20:42 | <tobie> | TabAtkins: Well they should be all fixed in that PR, no? |
| 20:43 | <TabAtkins> | Yeah, the PR should fix everything, but there's no good reason to clutter up your spec with a lot of "for" values just to disambiguate it from a previous version of itself, which it won't even link to unless you explicitly tell it to. ^_^ |
| 20:44 | <tobie> | TabAtkins: I can't disagree. |
| 20:48 | <tobie> | gsnedders: what's the context? |
| 20:51 | <smaug> | zcorpan_: do you happen to know if there are tests for cssom-view event.x/.y |
| 20:54 | <TabAtkins> | tobie: Just pushed - every single one of the errors you were getting were spurious, so you can throw away the PR. ^_^ |
| 20:54 | <zcorpan_> | smaug: i'm not aware of such tests |
| 20:54 | <smaug> | zcorpan_: ok. someone will then write such. |
| 20:54 | <zcorpan_> | smaug: nice! |
| 20:55 | <smaug> | zcorpan_: ok to put them to https://github.com/w3c/web-platform-tests/tree/master/cssom-view ? |
| 20:55 | <zcorpan_> | sure |
| 20:56 | <zcorpan_> | currently there are tests for cssom and cssom-view in both wpt and csswg-test, but the two will be merged soon anyway |
| 20:56 | <zcorpan_> | right gsnedders :-) |
| 20:57 | <smaug> | ah, right, rather messy right now |
| 20:57 | <smaug> | zcorpan_: I assume csswg will merge to wpt |
| 21:01 | zcorpan_ | is reminded by f.lux that he is tired, so will try to not stay up |
| 21:18 | <tobie> | Domenic: I've updated gh-pages with the latest bikeshed |
| 21:18 | <Domenic> | tobie: \o/ |
| 21:18 | <tobie> | Domenic: do you want me to rebase your PR and render it? |
| 21:18 | <Domenic> | tobie: that would be great. I think I have two outstanding PRs now? |
| 21:42 | <tobie> | TabAtkins: thanks for the bs fixes, btw. |
| 21:43 | <TabAtkins> | I will *never* get tired of the Bikeshed file extension being used that way. |
| 21:47 | <gsnedders> | tobie: relevance of W3C following 5.1 becoming a REC |
| 21:48 | <tobie> | gsnedders: oh. that's a thing? |
| 21:49 | <gsnedders> | tobie: yes |
| 21:49 | <tobie> | gsnedders: and so why are you writing about it? |
| 21:54 | jgraham | hopes the answer isn't "someone's wrong on the internet" |
| 21:55 | <gsnedders> | tobie: trying to explain things better to people on the internet (nobody is actually wrong!) |
| 21:57 | <tobie> | gsnedders: think the context you provided here would be useful as intro there |
| 21:59 | <jgraham> | Nobody is wrong on the internet? |
| 22:02 | <KiChjang> | that just means everybody is wrong |
| 22:02 | <KiChjang> | and somehow the internet is still working |
| 22:03 | <jgraham> | I think it just means that gsnedders, and an undetermined number of other people, are wrong |
| 22:59 | <tobie> | Domenic: done |