06:43
<asmodai>
interesting WebGL shader experiment http://glsl.heroku.com/e#4039.0
07:46
<jgraham>
Hixie: Yeah, my opinion on those is "just pick something sensible"
07:46
<jgraham>
I suppose there is some danger that a too-rigid heirachy stops people from writing that kind of test at all
07:47
<jgraham>
But it doesn't seem that likely
10:36
<asmodai>
gsnedders: just saw your code review tweet arrive via acquaintances, nifty!
13:03
<smaug____>
where is anne
13:07
<hsivonen>
there he is
13:07
<hober>
he was with jeff earlier
13:07
<hsivonen>
cunning timing
13:17
<annevk>
I talked with W3C's CEO, we could not reach some kind of agreement for now where the W3C would publish the WHATWG standards either with a permissive license or with me editing the WHATWG standards as is and publishing them under a restrictive license at the W3C.
13:17
<annevk>
So the situation will be that the W3C looks for editors to fork the WHATWG work under their license, which comes down to the same thing, but is more agreeable apparently.
13:18
<hober>
:(
13:19
<hsivonen>
:-(
13:21
<annevk>
He made it quite clear that he is not the person that can make this change. Member companies, including those representing browsers, would have to feel strongly about this for change to happen.
13:21
<hsivonen>
is this about different rules for Invited Experts vs. Members? there is Member precedent for this scnario, right?
13:22
<hsivonen>
specifically, Hixie specs
13:22
<hsivonen>
so the AC makes things bad again?
13:23
<hsivonen>
why does the CEO need to make this an AC issue?
13:23
<hober>
IIRC the relevant precedent is the non-exclusive copyright that a/o/m granted: http://lists.w3.org/Archives/Public/public-html/2007Apr/0429.html
13:23
<hsivonen>
why not just not ask the AC?
13:23
<annevk>
No, I think the rules are roughly the same there. They do not see that as a precedent they want to repeat even though it is repeating essentially anyway (with different editors). Thinking about it more makes it harder for me to understand their position.
13:24
<annevk>
They did ask the AC and a vocal minority said no (including some browser representatives I'm led to believe) and a vocal majority declined to comment.
13:25
<annevk>
*non-vocal majority, doh
13:25
<annevk>
and maybe not declined, but just did not comment
13:26
<hsivonen>
what's the upside of having your work published by the W3C? just the PP or do the threaten to ban normative refs to your stuff?
13:27
<othermaciej>
it seems like the biggest benefit would be to the w3c
13:27
<annevk>
The W3C believes their DNA is to be the standards body of the web (minus I suppose the little misdirection with regard to XHTML and such). From the perspective of being the standards body of the web, it makes sense to publish the standards that comprise it.
13:27
hsivonen
thinks it was a strategic error to ask the AC instead of relxing stuff by executive order
13:28
<annevk>
Agreed, this was something timbl should have just told the world about how it is.
13:29
<annevk>
s/was/is/
13:34
<zcorpan>
annevk: the browsers i have seem to fail http://simon.html5.org/test/css/css-charset/page-utf16-css-no-decl.html :-(
13:36
<GPHemsley>
Ms2ger: The anolis readme has not been updated to link to bitbucket instead of gsnedders
13:40
<zcorpan>
annevk: http://simon.html5.org/test/css/css-charset/page-utf16-css-no-decl-ascii-only.html - still fails
13:46
karlcow
is not sure how hsivonen makes executive order compatible with no member companies. It seems strange. I mean on the logical level.
13:46
<zcorpan>
annevk: http://simon.html5.org/test/css/css-charset/page-windows-1251-css-at-charset-utf16.html is kinda what hsivonen asked for. except the stylesheet ruleset is ascii-compatible
13:48
<karlcow>
The only way I see it changing is by convincing companies that it is in their own interests to do the work. Somehow it seems partly more a fight of some company employees with their own companies. :/ difficult issue.
13:49
<karlcow>
s/to do the work/to relax the license/ brain fork
13:50
<annevk>
zcorpan: so if you make an ASCII-only CSS file of that too, we can find out if browsers decode it as utf-16 or the page encoding
13:50
<annevk>
zcorpan: now the fail is ambiguous
13:50
<zcorpan>
annevk: http://simon.html5.org/test/css/css-charset/page-windows-1251-css-at-charset-utf16-ascii-only.html
13:51
<annevk>
whoa, both Opera and Gecko fail
13:51
<annevk>
CSS 2.1 requires them to fail, but still that is some pretty crazy requirement to begin with...
13:52
<hsivonen>
from code inspection, gecko decodes as utf-16 if @charset claims so
13:53
<zcorpan>
opera even decodes http://simon.html5.org/test/css/css-charset/at-charset-utf16-ascii-only.css as utf-16 when loading in a top-level browsing context
13:53
<zcorpan>
(might be a cache thing)
13:53
<hsivonen>
in gecko, the code path is totally different for that
13:54
<hsivonen>
top level context uses text/plain code
13:54
<hsivonen>
i.e. HTML with meta disabled
13:54
<zcorpan>
i thought we use text/plain behavior as well, but it's likely that we remember the encoding we used for decoding the file as a stylesheet
13:54
<zcorpan>
there are issues like that
13:55
<hsivonen>
we should too but I think I accidentally broke that
13:56
<zcorpan>
are there other cases i should test?
13:56
<zcorpan>
should we change the inherit-utf-16 case?
13:58
<annevk>
zcorpan: can you do "IndexOptions DescriptionWidth=*"
13:58
<annevk>
see http://httpd.apache.org/docs/2.2/mod/mod_autoindex.html for other options
13:59
<annevk>
sorry
13:59
<annevk>
NameWidth=* is the one
13:59
<zcorpan>
done
14:00
<annevk>
zcorpan: xml-stylesheet test? hsivonen said Gecko supports charset= there
14:00
<zcorpan>
i guess i can test with bogus decl in more places and see that it falls back to a recognized label at different places
14:00
<zcorpan>
ah yeah
14:01
<hsivonen>
do we know yet what exactly @charset "utf-16"; does in WebKit and IE
14:02
<hsivonen>
what Gecko does makes no sense
14:04
<asmodai>
w3.org dead due to Sandy?
14:04
<hober>
asmodai: or tpac
14:04
<annevk>
zcorpan: do you know what step 3 does in IE? The utf-16 -> utf-8 thing for @charset
14:04
<zcorpan>
no. don't have ie
14:04
<annevk>
ah, hsivonen asked the same
14:04
<asmodai>
hober: :\
14:04
<annevk>
yeah me neither :/
14:05
<annevk>
oh wait
14:05
<karlcow>
asmodai: working for me
14:05
<asmodai>
karlcow: Ah, it's back now
14:05
<asmodai>
couldn't reach it in the past 5-10 minutes or so
14:05
<asmodai>
Still has problems loading chrome, heh
14:08
<hsivonen>
sad when the position is "we need need to ask lawyers"
14:09
<zcorpan>
http://simon.html5.org/test/css/css-charset/xml-stylesheet-page-windows-1251-charset-attribute-windows-1250.xhtml hmm. elm is null. is there something about <div id="&#x010C;"></div> not doing what it looks like it does in XML?
14:11
<annevk>
it returns Č afaict
14:49
<zcorpan>
hsivonen: http://simon.html5.org/test/css/css-charset/page-windows-1251-css-at-charset-utf16.html - webkit uses utf-8 for @charset "utf-16" afaict
14:50
<hsivonen>
zcorpan: good. so the rule is not inventing something new
14:50
<zcorpan>
i solved the xml issue by setting .id in script
15:09
<zcorpan>
hsivonen: firefox seems to fail (all?) the "bogus" tests
15:09
<zcorpan>
e.g. http://simon.html5.org/test/css/css-charset/page-windows-1251-css-http-bogus-at-charset-windows-1250.html
15:11
<annevk>
zcorpan: you mean they apply the "get an encoding" step at the end?
15:11
<zcorpan>
i haven't figured out what gecko does exactly
15:59
<dglazkov>
good morning, Whatwg!
16:50
<zcorpan>
TabAtkins: any idea where css tests should be submitted? http://hg.csswg.org/test ? http://dvcs.w3.org/hg/csswg/ ?
17:09
<zcorpan>
do non-gecko browsers on windows do the file://c:/something -> file:///c:/something thing?
17:12
<GPHemsley>
annevk: I wonder if we could abstract the decoder algorithm in encoding out to that generic definition document? I seem to be writing almost the same thing for mimesniff, at least in terms of reading bytes and keeping track of the pointer, etc.
17:17
GPHemsley
wonders why you can't highlight a line in the logs as important after 24 hours.
19:25
<Hixie>
TabAtkins: src="url(foo)" behaviour is already pretty much fixed in stone, fwiw (it's the same as src="url%28foo%29", more or less)
19:26
<Hixie>
(i onl briefly glanced at the thread so please disregard if my statement is redundant :-) )
20:41
<bencc>
websocket is version 13
20:42
<bencc>
is it possible that in the future there will be version 14, 15... ?
20:46
<Ms2ger>
GPHemsley, the reason you can't highlight the logs from the past is that they need to be cached aggressively, because krijn serves the logs from his toaster (or equivalent hardware)
20:46
<krijn>
:D
20:46
<GPHemsley>
ah :)
20:46
<krijn>
Yeah, sorry about that
20:46
<Ms2ger>
Thanks for providing them anyway ;)
20:47
<krijn>
I think the highlighting of lines is misused most of the times anyway
20:47
<krijn>
What the
20:47
<krijn>
Speaking of the devil
20:48
<Ms2ger>
As troll bait, you mean?
20:48
<Ms2ger>
That's how I use it, at least
20:49
<krijn>
That's correct usage :)
20:51
<GPHemsley>
Ms2ger: Oh, subtle!
20:51
<Ms2ger>
Hmm? :)
20:57
<Ms2ger>
Anything interesting happened in the webapps meeting, btw?
20:59
Ms2ger
confirms nor denies he wrote multicol and flexbox tests
21:05
<pablof>
TabAtkins: is window.CSS in some spec?
21:07
<Ms2ger>
Not yet, I don't think
21:11
<isaacd>
Running validator.nu code locally is now giving me an exception. (Re-checking-out-and-building didn't help. Error log: http://pastebin.com/r2i3LrX2 )
21:16
GPHemsley
enjoys tagging random elements with 'class=XXX' to see how badly the styling breaks.
22:10
<roc>
what's Philip Jagenstadt's IRC nick again?
22:11
<Ms2ger>
foolip
22:11
<roc>
thank you sir
22:11
<Ms2ger>
Np
22:11
<roc>
foolip: Web Audio once again requires the attention of someone who groks Web standards
22:12
<Ms2ger>
Hixie, pointers to callbacks that do propagate exceptions?
22:18
<Hixie>
Ms2ger: i dunno
23:48
<Hixie>
where's ms2ger?
23:59
<gsnedders>
Bed, I guess.
23:59
<gsnedders>
Some of us sleep, sometimes.
23:59
<Hixie>
that can't be right