14:53
<innovati>
What's the story about HTML's support for <![CDATA[cdata sections]]> and <?processsing-instruction?>?
14:53
<innovati>
In Chrome they parse as <!--[CDATA[cdata sections]]--> and <!--?processsing-instruction?-->
14:54
<innovati>
Is their parsing in HTML to allow parsing them inside embedded XML documents only?
15:04
<Domenic>
innovati: https://html.spec.whatwg.org/multipage/syntax.html#cdata-sections might help.
15:04
<Domenic>
Outside of foreign content they are "bogus comments", I believe. I.e. they are similar to <!anything else>
15:06
<innovati>
interesting it seems <!a> and <?a> parse as <!--a--> and <!--?a-->
15:07
<innovati>
does the spec talk about bogus comments, or is this just something browsers do?
15:07
<innovati>
(I'm writing XML, and HTML templating functions, and functions that process DOM nodes so I have to write tests and I'm realizing gaps in my own understanding :D )
15:14
<andreubotella>
The parsing algorithm describes in full detail all aspects of parsing, including such "bogus comments".
15:16
<andreubotella>
though in many particular occasions, an HTML parser is allowed to give up and reject the input when it's invalid
15:17
<andreubotella>
not sure if in the case of bogus comments
15:18
<andreubotella>
although browsers are not exempted from giving up like that, no major browser does
15:30
<innovati>
ah found this https://html.spec.whatwg.org/multipage/parsing.html#parse-error-unexpected-question-mark-instead-of-tag-name
15:30
<innovati>
and in this section it seems to cover the <!-- and <! style comments, I think that's what you were just talking about? https://html.spec.whatwg.org/multipage/parsing.html#comment-start-state
15:31
<innovati>
Thanks Domenic && andreubotella <3
15:31
<andreubotella>
that's right
15:31
<andreubotella>
np
15:32
<andreubotella>
the CDATA and <!a> cases are covered in the markup declaration state
15:32
<andreubotella>
*markup declaration open
15:34
<innovati>
it's pretty beautiful in practice how resilient HTML is
15:36
<Domenic>
annevk: edgarchen: Web IDL reviews this week? :)
16:16
<annevk>
Domenic: it's very annoying that you cannot accept commit suggestions on whatwg/html; I guess the file is too big or some such?
16:17
<Domenic>
Ah, yeah, that makes sense
17:00
<annevk>
Domenic: I did one and pinged Edgar for the other
17:20
<Domenic>
annevk: so should we merge all the extended attribute renames, you think?
17:25
<Domenic>
annevk: I'm gonna do it
17:29
<annevk>
Domenic: the one remaining concern I had was tooling/impl interest; I can say Fx will get to it eventually
17:30
<Domenic>
Hmm fair
17:30
<Domenic>
I checked the tooling and they're pretty much OK
17:37
<Domenic>
Ugh the Bikeshed service is broken? https://github.com/tabatkins/bikeshed/issues/1649
18:28
<bkardell>
Domenic: can you help me figure out how I can move https://github.com/whatwg/html/pull/5248 properly? To recap: We suggested changing the interface name first because all of the things follow, but that didn't fly because it wasn't observeable directly by itself until mathml does something in implementations - it does in 2 browsers now mostly agree to what those things are... you suggested the special link/tabindex bit,
18:28
<bkardell>
we took some time and I thought sorted that out, wrote the tests and stuff but then webkit doesn't want to do that in this pull. It seems to me that another pull would depend on a lesser pull first else there isn't even a tabIndex... Unsure how to move this forward
18:29
<Domenic>
bkardell: I don't see why a tabindex PR would depend on that PR
18:29
<bkardell>
ok then :)
18:30
<bkardell>
I mean, I assumed because nothing specified it should even have a tabindex until you route through all the things, but ok.. .I'm good with that, will send a separate one
18:32
<Domenic>
We might delay merging until they can be merged at the same time, but it seems worthwhile having a separate PR for that separate change.
19:10
<annevk>
This situation is a bit confusing also with ARIA adding to this potentially
19:10
<annevk>
No good suggestions though; I support the high bar
19:11
<annevk>
Or consistent bar, if you will
20:31
<Domenic>
annevk or others: do you know if Gecko is working on out of process iframes?
20:34
<jgraham>
Yes
20:34
<jgraham>
https://wiki.mozilla.org/Project_Fission
20:35
<jgraham>
Unless you mean out of process same origin, in which case I don't know
20:44
<Domenic>
Nah, cross-origin
20:45
<Domenic>
Thanks; I thought Fission was just for tabs, not iframes