| 01:18 | <Domenic> | Yet again implementers are misled by a non-normative note -_- |
| 03:46 | <MikeSmith> | Domenic: which note? |
| 03:46 | <Domenic> | https://github.com/whatwg/html/issues/1852 |
| 03:52 | MikeSmith | looks |
| 04:16 | <MikeSmith> | posted a comment just now |
| 04:16 | <MikeSmith> | with link to https://github.com/whatwg/html/commit/65216fce862018e2052a7edd24485e41839dae6d |
| 04:16 | <MikeSmith> | which is not any more enlightening |
| 07:15 | <annevk> | MikeSmith: https://github.com/whatwg/html/pull/1847 is waiting for you |
| 07:45 | <MikeSmith> | annevk: yeah will take a look at it shortly |
| 11:02 | <zcorpan> | annevk: https://github.com/whatwg/html/pull/1854/files#diff-36cd38f49b9afa08222c0dc9ebfe35ebR78676 have we stopped with "return null and abort these steps" in HTML? |
| 11:07 | <jgraham> | zcorpan: https://twitter.com/zcorpan/status/783212065850875905 good point. Someone should write that as a followup |
| 11:07 | <jgraham> | idk if someone should be me :) |
| 11:08 | <zcorpan> | 🙂 |
| 11:45 | <JakeA> | annevk: I'm speccing navigationPreload.getState() to return an object like {enabled: false, headerValue: "blah"} - would the returned object have its own interface in webidl?? |
| 11:45 | <JakeA> | hah, too many ? |
| 11:45 | <JakeA> | makes it sounds stressy |
| 11:49 | <annevk> | JakeA: return a dictionary |
| 11:49 | <JakeA> | ta! |
| 12:08 | <jgraham> | https://hackernoon.com/how-it-feels-to-learn-javascript-in-2016-d3a717dd577f#.bte6fur6n |
| 12:31 | <smaug> | heh |
| 12:31 | <smaug> | and I guess rather true |
| 12:51 | <JakeA> | annevk: I'm struggling to find a spec that returns a dictionary, they only seem to be used for arguments. Know anywhere I can cargo-cult from? |
| 13:03 | <annevk> | JakeA: maybe RTC somewhere? It exists for sure |
| 13:03 | <JakeA> | annevk: will keep digging, cheers! |
| 13:17 | <annevk> | JakeA: as for language, just create an IDL dictionary with the right members in prose and return that |
| 13:58 | <annevk> | smaug: could you check https://github.com/whatwg/dom/pull/336#issuecomment-251067202 please? |
| 13:59 | <annevk> | JakeA: might be better to create a WIP PR next time if you need feedback |
| 13:59 | <smaug> | hrm, spec issue being discussed in a pull request |
| 14:00 | <JakeA> | annevk: sure. I'll create one when I fix those issues & add the preload response promise |
| 14:00 | <JakeA> | annevk: thanks for the feedback |
| 14:00 | <annevk> | smaug: seems to just happen |
| 14:01 | <annevk> | smaug: recently one happened on a commit, but I steered it towards an issue |
| 14:01 | <annevk> | smaug: maybe it's something where if we remind folks enough it doesn't happen anymore at some point? |
| 17:12 | <smaug> | are there still some cases when there is nested event loop spinning happening in the spec in such way that there might be JS on stack? |
| 18:45 | <tobie> | Would it be helpful to anybody else than me to link to things like ES [[Get]] from the WebIDL spec? |
| 18:50 | <annevk> | smaug: not sure, don't think so |
| 18:50 | <annevk> | smaug: but bz now and then mentions alert() does not match reality |
| 18:51 | <annevk> | smaug: and UAs still mess up sync xhr |
| 18:51 | <smaug> | annevk: oh, I know what Gecko does |
| 18:51 | <smaug> | but those are special cases |
| 18:51 | <smaug> | showModalDialog is/was a bit different |
| 18:52 | <smaug> | since the underlying page was still kind of running |
| 18:53 | <smaug> | it is just that the spec prevents entering to microtask loop if we're there already |
| 18:53 | <annevk> | Hmm okay |
| 18:53 | <smaug> | my guess is that it never worked with showModalDialog and it got just fixed by removing showModalDialog |
| 18:53 | <TabAtkins> | That matches what I remember, yes. |
| 18:54 | <annevk> | I've never really understood the full model |
| 18:54 | <annevk> | Close, but there was always something |
| 19:45 | <smaug> | is there any way to review spec changes using the generated HTML files? Some tool which shows then in the new page what was added and perhaps in previous version what was removed? |
| 19:51 | <Domenic> | tobie: it would be helpful, however, part of the issue is that most places that link to [[Get]] should probably instead be linking to Get() |
| 19:52 | <Domenic> | as part of the ES5 -> ES6+ update |
| 19:52 | <Domenic> | smaug: there is https://html.spec.whatwg.org/commit-snapshots/ which is currently kind of broken pending https://github.com/whatwg/html/pull/1857 |
| 19:57 | <tobie> | Domenic: https://tc39.github.io/ecma262/#sec-ordinary-object-internal-methods-and-internal-slots-get-p-receiver ? |
| 19:57 | <Domenic> | tobie: the problem is that's only the definition for ordinary objects |
| 19:57 | <Domenic> | it's not applicable to any web platform objects for example |
| 19:58 | <tobie> | Domenic: does that mean there's something else applicable in that case, or does that mean we're missing bits here? |
| 19:59 | <jyasskin> | tobie: Yes, please link from WebIDL to ES. https://tc39.github.io/ecma262/#sec-get-o-p in this case. |
| 19:59 | <Domenic> | So for example, in the modern era, "value is the result of calling the [[Get]] internal method of V passing key and V as arguments." should become "Let value be ? Get (V, key)" |
| 19:59 | <Domenic> | With the link going to the thing jyasskin said |
| 20:00 | <Domenic> | (for Get(), not for [[Get]]) |
| 20:04 | <Domenic> | Oh poop, there's no way of getting DataTransfer objects |
| 20:10 | <tobie> | Domenic: where is [[Get]] defined for platform objects, tho? |
| 20:11 | <Domenic> | tobie: indeed it looks like [[Get]] is one of the few things platform objects don't overwrite; my bad. But it's still incorrect to link to ordinary object [[Get]] in most cases, since e.g. in that dictionary case, we could be handed a proxy. |
| 20:26 | <tobie> | Domenic: any idea who we talk to to get better links in the ES spec? |
| 20:26 | <Domenic> | tobie: bterlson, preferably by filing an issue. |
| 21:35 | <jsbell> | Whelp, rewrite of Blink's UTF-16 decoder landed. What could go wrong... |
| 22:58 | <TabAtkins> | annevk I guess?: Yo I stole your caniuse styles for Bikeshed. |