03:21
<howdoi>
oninput event must work for contenteditable=true right?
08:56
<tobie>
heycam: any idea what this HTML comment is doing here? https://github.com/heycam/webidl/blob/1d575a6/index.xml#L461
08:57
<heycam>
tobie: that was from me half-arsing the raises()/setraises()/getraises() annotation removal
08:57
<heycam>
it can be removed
08:57
<heycam>
same with any of the <!--JAVA ... --> sections
08:58
<tobie>
yes, kicked those out already
08:58
<heycam>
cool :)
08:58
heycam
must go to dinner now
08:58
<tobie>
bon appetit
13:33
<Ms2ger>
My nexus 5x has that
13:34
<Ms2ger>
Did the MS people end up publishing a table layout spec?
13:35
<nox>
There is no headphone jack on iPhone 6?
13:36
<annevk>
nox: there's a rumor about the next iPhone
13:37
<nox>
I don't understand the parallel with floppy drives. I wouldn't use floppies when the drives got removed. I've never had a headset that wouldn't use a jack.
13:38
<Ms2ger>
Moz shipped me a usb headset
13:40
<nox>
Ms2ger: I ordered a Bose headset.
13:40
<nox>
No USB thingie on it.
14:16
<zcorpan>
MikeSmith: nice. confirmed those two cases. is there a diff for this latest change?
14:19
<MikeSmith>
zcorpan: there is not, because I squashed the commits and force-pushed to the fork of the parser repo
14:20
<zcorpan>
ah ok
14:21
<zcorpan>
MikeSmith: now i'm curious what the flub was :-)
14:30
<MikeSmith>
zcorpan: <!--<!---> or <!--<!----> or <!--<!-----> etc caused the parser to never end the comment and return to the data state :)
14:31
<zcorpan>
nice
20:46
<smaug____>
Domenic: curious, are the currently any APIs which need to refer to relevant realm ?
20:46
<Domenic>
smaug____: yeah, we have an example of navigator.getBattery in https://html.spec.whatwg.org/multipage/webappapis.html#realms-settings-objects-global-objects
20:47
<smaug____>
that is an example
20:47
<Domenic>
smaug____: it is not a hypothetical example
20:48
<Domenic>
smaug____: customElements does a similar thing; window1.customElements.define.call(window2.customElements, ...) will define the element in window2
20:49
<smaug____>
and the spec actually uses relevant realm? I was just surprised to see 4th global type :)
20:49
<smaug____>
thanks for adding the example to the spec
20:49
<smaug____>
https://html.spec.whatwg.org/multipage/webappapis.html#realms-settings-objects-global-objects:the-iframe-element
20:50
<smaug____>
"type", 4th way to refer to a global
20:50
<Domenic>
yeah, it's been kind of implicit; "relevant" was the most-often-used term I believe, so we settled on that
20:51
<Domenic>
And yeah I figured after I spent all the time sucking this information from bz's mind I was going to make sure it was at least somewhat more accessible for the next person
20:57
<tobie>
TabAtkins: Some of the WebIDL grammar gets mixed up for macros by Bikeshed (e.g. -?([1-9][0-9]*|0[Xx][0-9A-Fa-f]+|0[0-7]*)). Is there an undocumented way to escape these in groups (e.g. insert them in a <pre> element or something) or am I out of luck and will need to escape all of them separately (e.g.: -?(\[1-9]\[0-9]*|0[Xx][0-9A-Fa-f]+|0\[0-7]*))?
21:00
<Domenic>
tobie: <pre> can be used
21:00
<Domenic>
tobie: you can also use a custom element and put it in the "opaque elements" list
21:00
<Domenic>
tobie: https://github.com/whatwg/streams/blob/master/index.bs#L26
21:00
<TabAtkins>
tobie: Unfortunately macro replacement happens at the text level (pre-parsing) because you can use macros in attribute values, or insert markup with them. So I don't have any implicit escaping. That said, I already skip over macro-looking things that only contain digits, and I can skip ones that also contain a dash.
21:00
<TabAtkins>
Domenic: Unfortunately macros are in the set of things that this doesn't work for.
21:01
<Domenic>
oh hmm i guess i am unfamiliar with macros
21:02
<TabAtkins>
There are a few markup shorthands that have this problem too, if they'd interfere with HTML parsing, like the <<grammar-term>> one.
21:02
<TabAtkins>
(I should probably have switched macros to a much harder-to-collide syntax early on. Maybe I can still do this?)
21:02
<TabAtkins>
They're almost never used in the body of things, just in boilerplate.
21:05
<TabAtkins>
And I just pushed that, so the ones that are just [0-9] no longer need escaping.
21:06
<TabAtkins>
The ones for hex digits still do, tho.
21:06
<tobie>
TabAtkins: sorry, one of my daughters woke up. She's back to sleep now.
21:08
<tobie>
So yeah, they wouldn't really be macros if they happened on a constructed tree, right?
21:08
<tobie>
Thanks for fixing the [0-9] case.
21:08
<tobie>
I'll do the other ones by hand.
21:09
<TabAtkins>
tobie: Exactly. (Unfortunately.)
21:10
<tobie>
I think it would make sense to make the macro syntax a bit more out of the ordinary.
21:10
<TabAtkins>
I've idly contemplated writing an HTML parser just so I could hook these things into it and do *no* non-Markdown manipulations on the raw text.
21:10
<tobie>
But I can't really estimate the risk of breaking things.
21:10
<TabAtkins>
But that's quite a bit of work. :/
21:11
<TabAtkins>
(Writing my own HTML serializer was worth it, at least.)
21:11
<tobie>
You lost me at "*no* non-Markdown."
21:12
<TabAtkins>
Markdown-related things still have to happen at the textual level, I think. But everything else could move to post-parsing if not for it confusing the HTML parser.
21:12
<tobie>
OK. Read it three times and I think now I sort of get it.
21:14
<tobie>
I'm not familiar enough with the architecture of Bikeshed to comment on this, tbh.
21:14
<TabAtkins>
Yeah that's fine, I'm just thinking aloud. ^_^
21:16
<tobie>
Merged your changes and I no longer have macro issues at all.
21:16
<tobie>
:)
21:19
<TabAtkins>
Oh yeah! It requires the text to be all-uppercase, so all those hex character-classes aren't hitting it. ^_^
21:19
<tobie>
yup. :)
21:20
<tobie>
https://www.irccloud.com/pastebin/zuhGICmX/
21:20
<TabAtkins>
Hmmmm
21:20
<TabAtkins>
How... is that not hit before.
21:20
<tobie>
TabAtkins: now getting the above. The src is super, super messy, still, so that might be the cause.
21:23
<tobie>
TabAtkins: current src file: https://github.com/tobie/webidl/blob/bikeshed/index.bs
21:24
<TabAtkins>
Yeah, I spotted the problem, but it's in the "how the fuck did this ever work" category, so ???
21:25
<tobie>
:D
21:31
<tobie>
non-normative IDL is a rather new feature no?
21:33
<TabAtkins>
Yeah, brand new, but also I *tested* this.
21:34
<TabAtkins>
Or... maybe I didn't. Maybe I only tested the highlighting of normal IDL?
21:34
<TabAtkins>
That seems unlikely.
21:34
<TabAtkins>
I dunno.
21:36
<TabAtkins>
Okay, works now.
21:45
<tobie>
\o/
22:26
<tobie>
TabAtkins: Thoughts as to how best handle ES object internal properties (e.g.: [[DefineOwnProperty]])?
22:28
<TabAtkins>
Slots in an object are recognized as an IDL type, so you can do {{[[foo]]}}.
22:28
<tobie>
awesome. Thanks!
22:29
<TabAtkins>
Of course, your usage is generic, not referring to any actual existing slot, so that's not going to help you. ^_^
22:31
<TabAtkins>
But escaping it with \ works in that case.
22:32
<tobie>
Right.
22:32
<tobie>
ta