00:01
<TabAtkins>
Both done.
04:15
<annevk>
benjamingr__: I think legacycaller is there for plugins, if that is what you meant
04:53
<Domenic>
annevk: yay, thanks for the review. Will merge tomorrow morning!!
04:53
<annevk>
\o/
04:54
annevk
woke up early and figured he should ignore all distraction until that was done
04:54
<annevk>
I mostly ended up adding "then"
04:54
<annevk>
I also removed the comment under the HTMLUnknownElement list since it was no longer accurate and I renamed "re-throw" to "rethrow" once to match established conventions
05:00
<Domenic>
perfect <3
05:31
<laughinghan>
question that's narrowly about history.scrollRestoration but more broadly about when breaking changes are allowed to the spec
05:31
<laughinghan>
if something has been implemented in two major browsers, compatibly, but hasn't been implemented in the others yet (like at all, not even experimentally)
05:32
<laughinghan>
is a breaking change out of the question? Especially if it makes more sense
05:51
<MikeSmith>
laughinghan: not out of the question and there have been non-hypothetical cases of that
05:51
<laughinghan>
really? awesome
05:51
<laughinghan>
I'm working on an illustration right now
05:51
<laughinghan>
(of the change I'm gonna propose)
05:52
<MikeSmith>
cool
05:52
<laughinghan>
MikeSmith: out of curiosity, know off the top of your head
05:52
<laughinghan>
of such cases?
05:53
<MikeSmith>
not off the top of my head but annevk or Domenic might have some at hand
05:55
<MikeSmith>
one thing I do remember, there is something that everybody changed that affected Acid2 rendering
05:55
<MikeSmith>
http://acid2.acidtests.org/#top
05:55
<MikeSmith>
that band across the eyes
05:55
<laughinghan>
what about it?
05:56
<MikeSmith>
that is an artifact of a subsequent change in implementations
05:56
<MikeSmith>
originally all browsers did not have the in their renderings
05:57
<MikeSmith>
instead it looked just like http://acid2.acidtests.org/reference.html
05:57
<annevk>
I think for new features that is rare, but who knows
06:00
<laughinghan>
ohhh
06:00
<laughinghan>
interesting
06:00
<laughinghan>
wait but that got enshrined as official behavior right?
06:00
<laughinghan>
so the opposite of what I want?
06:01
<laughinghan>
Not that I'm complaining, that's still interesting
06:01
<MikeSmith>
it was a breaking change for anything that relied on the behavior that acid2 required previously
06:01
<MikeSmith>
maybe there wasn’t much that relied on it, I dunno
06:02
<MikeSmith>
that test tests some crazy-ass corner-case stuff
06:03
<laughinghan>
huh, interesting
06:04
<MikeSmith>
anyway you never know til you try
06:04
<MikeSmith>
stranger things have happened
06:06
<laughinghan>
:)
06:48
<laughinghan>
okay here it is: http://output.jsbin.com/zapoyu
06:48
<laughinghan>
I'm also gonna open a GitHub ticket (unless you say I shouldn't)
07:01
<MikeSmith>
haha
07:01
<MikeSmith>
massive points for incorporation of the dada engine
07:02
<MikeSmith>
btw you may enjoy https://logopoeia.com/wisdom/
07:02
<MikeSmith>
ah he’s gone
07:02
<MikeSmith>
well I hope he does open an issue
07:45
<laughinghan>
okay submitted issue: https://github.com/whatwg/html/issues/1042
07:46
<laughinghan>
also created a test case for the desired browser behavior (automated & manual)
07:46
<laughinghan>
good night
10:59
<zcorpan>
disappointed that github's squash and merge isn't smart about fixup commits
11:01
<jgraham>
Well yes
11:01
<jgraham>
GitHub are great at half-assing features
11:15
<annevk>
zcorpan: you mean that you need to determine yourself whether to keep the commit message?
11:16
<zcorpan>
annevk: git commit --fixup hash means there's no message, and it gets automatically squashed into the provided commit when rebasing
11:16
<zcorpan>
annevk: but github doesn't seem to care and i still need to do cleanup in the textarea
11:16
<annevk>
I see
13:18
<Ms2ger>
Anyone know what spec defines a dictionary constructor?
13:22
<smaug____>
Ms2ger: webidl?
13:25
<smaug____>
Ms2ger: like http://heycam.github.io/webidl/#es-dictionary-constructors
13:26
<Ms2ger>
Yeah
13:26
<smaug____>
can't recall why it was added though
13:26
<smaug____>
I guess there must have been some use case
13:27
Ms2ger
looks at blame
13:28
<Ms2ger>
> heycam committed on 26 Mar 2013
13:29
<smaug____>
I hope there was some message why it was committed
13:29
<Ms2ger>
" Allow [Constructor] on dictionaries. "
13:30
smaug____
kicks w3c spec stylesheet
13:32
<smaug____>
good that webidl still uses the old, good one, but some other specs ...
14:26
<zcorpan>
can SVG elements be customized built-in elements?
14:29
<annevk>
zcorpan: don't think so
14:37
<zcorpan>
how come we allow any attribute on customized built-in elements? isn't that as problematic for future extensions to the language as any attribute on non-custom standard elements?
15:33
<Domenic>
yeah, maybe we should leave them disallowed for now and discuss in a separate issue
20:53
<zcorpan>
https://www.peterkrautzberger.org/0186/?utm_source=html5weekly&utm_medium=email "MathML is a failed web standard"
20:55
<laughinghan>
zcorpan: https://www.w3.org/community/mathonwebpages/2016/04/13/call-for-participation-in-getting-math-onto-web-pages-community-group/
20:55
<Domenic>
zcorpan: oh, you're back. review custom elements ^_^
20:56
<laughinghan>
the charter for the Math WG (in charge of MathML) has lapsed; the author of that blogpost launched the new CG I just linked to
20:56
<Domenic>
MathML has 2 impls, right? WebKit and Gecko? So not a straightforward thing to remove :-/
20:56
<laughinghan>
also is that being posted here 'cos we hate MathML here? :D
20:57
<laughinghan>
Domenic: according to pkra, at least, both of those impls were written and primarily maintained by volunteers from his community, not browser vendor employees
20:57
<Domenic>
meh, seems fine, just not implemented and doesn't seem to be going places.
20:58
<Domenic>
yeah, still
20:58
<Domenic>
Gecko had a debate on removing mathml a while ago that didn't seem to go anywhere
20:59
<laughinghan>
the dev.platform.mozilla or whatever mailing list thread? Yeah I have that open a browser tab, read halfway through haha
20:59
<Domenic>
yeah
20:59
<zcorpan>
laughinghan: i don't think it's hated here, i posted it because i found it interesting and thought it would be interesting for people here also
20:59
<laughinghan>
damn haha
21:00
<laughinghan>
but yeah, def a good article
21:02
<zcorpan>
in practical terms there being two partial implementations shouldn't prevent better solutions from being developed, but maybe mentally people expect MathML to eventually magically become good
21:04
<TabAtkins>
annevk: <3 for getting some useful Shadow DOM terms into DOM; reffing them now in the CSS Scoping spec instead of redefining them.
21:08
<zcorpan>
Domenic: r+
21:08
<Domenic>
\o/
21:10
<zcorpan>
nice work 👏
21:16
<gsnedders>
Domenic: Presto supported the CSS subset of MathML
21:16
<gsnedders>
Domenic: not that that is relevant any more
21:28
<jgraham>
Presto invented the CSS subset of MathML
22:07
<jyasskin>
TabAtkins: Might it be the case that Shepherd isn't indexing enum-value dfns? My {{"prompt"}} isn't finding the definition in Permissions.
22:17
<TabAtkins>
jyasskin: Hmmmmmmmmm. It shows up in the db just fine, so it's being indexed. `bikeshed refs --text='"prompt"'` returns it. It's just not working in the {{"prompt"}} shorthand. Investigating.
22:18
<TabAtkins>
Oh, I see the problem.
22:18
<TabAtkins>
It's indexed as type=value for some reason. Your markup is right, so yeah, looks like a Shepherd bug. Bugging plinss now.
22:48
<TabAtkins>
jyasskin: plinss is fixing the issue now. Should show up in a little while.
22:48
<jyasskin>
Great, thanks!