00:02
<hober>
alphabetical except for DOMContentLoaded?
00:02
<annevk>
smola: sorry it's all a bit slow, I'm in the midst of meetings (and should really be asleep now)
00:03
<annevk>
Hixie: cssom-view should prolly define scroll, dunno about resize
00:03
<Hixie>
html now says "some spec one day will" or something
00:03
<annevk>
Hixie: CSS in general really ought to own the whole scrolling concept
00:03
<jwalden>
hmm; http://www.w3.org/ "Learn how to program Web applications with W3C&#8217;s Mobile Web 2 online course"
00:04
<Hixie>
annevk: i'd hope
02:11
<MikeSmith>
Hixie: yay for http://www.whatwg.org/specs/web-apps/current-work/#events-0
02:15
<MikeSmith>
Hixie: hmm how come it's not in the multipage version? just didn't regenerate it yet?
07:24
<zcorpan>
Hixie: http://dev.w3.org/csswg/cssom-view/#events
07:25
<zcorpan>
smola: ?
07:26
<smola>
zcorpan: morning; I just wondered where should I report problems with web-platforms-tests
07:26
<smola>
GitHub Issues or somewhere else
07:26
<Hixie>
zcorpan: excellent, cool
07:26
<Hixie>
MikeSmith: done
07:26
<Hixie>
MikeSmith: dunno why it didn't take before
07:26
<MikeSmith>
thanks
07:27
<zcorpan>
smola: github issues
07:27
<smola>
zcorpan: alright, thanks
07:29
<zcorpan>
Hixie: help http://lists.w3.org/Archives/Public/public-web-perf/2014Jan/0014.html
07:31
<Hixie>
i'm too tired to work out the answer. ask me again tomorrow? :-)
07:31
<Hixie>
or copy some place in HTML that does it already... :-)
07:32
<zcorpan>
ok
07:42
<zcorpan>
i think we want incumbent-settings-object although that's different from xhr
08:30
<annevk-cloud>
New APIs should not copy from XHR
08:32
<darobin>
it would be fun
08:34
<darobin>
element.onmousestatechange = function () { if (element.mouseState !== 17) return; ... }
08:34
<darobin>
actually, that'd make for a nice April Fool's spec
09:57
<zcorpan>
annevk-cloud: you may want to review the Beacon spec, i think it has copied much of its stuff from xhr
11:04
<jgraham>
Wow, even the innerHTML getter seems to be full of interop fail
11:22
<annevk>
Reliable way to find out about why a feature could not be implemented. Search for the feature and "Boris Zbarsky"
11:22
<annevk>
jamesr__: so that email is about Event.systemTime
11:23
<annevk>
jamesr__: if you are just talking about the typedef, I think that could be directly in IDL
11:23
<annevk>
jamesr__: DOMTimeStamp is defined there too
11:35
<jgraham>
annevk: heh
11:55
<SimonSapin>
darobin: careful of Poe's law…
12:02
<darobin>
SimonSapin: Poe's law is part of what makes it fun :)
13:25
<Ms2ger>
jgraham, fun
13:25
<Ms2ger>
darobin, you realize that the CSSWG decided to reference an April Fools RFC?
13:26
<darobin>
Ms2ger: I know, RFC 6919
13:26
<darobin>
honestly, I don't understand why referencing that isn't a requirement for all documents
14:04
<SimonSapin>
april fools or not, the definition of "may wish to" is perfect for this case
16:06
<Hixie>
SimonSapin: we shouldn't be publishing specs that contain "behaviour which is regarded as ridiculous", and we certainly shouldn't be making compromises to expedite going through W3C process. Though I understand that that is a liability of working with the w3c.
16:09
<SimonSapin>
Hixie: this was not about W3C process, it was about two working group members that could not agree after endless discussion on a topic that is so minor that is was judged not worth blocking the rest of the spec.
16:10
<SimonSapin>
IIRC it only affects a handful of code points, and only with broken fonts
16:10
<jgraham>
There are non-W3C processes where two members couldn't hold the group hostage like that
16:11
<jgraham>
e.g. anything that is not consensus based
16:11
<Hixie>
that couldn't happen in the whatwg. the editor gets final say on what the spec says.
16:11
<Hixie>
it's only because of the w3c's unhealthy obsession with "consensus" that it could happen
16:11
<Hixie>
what jgraham said.
16:12
<SimonSapin>
well, the two in this case are co-editors
16:13
<Hixie>
ah, well. there are whatwg specs with co-editors. but i think that's a mistake.
17:25
tantek
dials into HTMLWG telcon, hears discussion of longdesc. What year is it?
17:25
<Ms2ger>
You're dialing into an HTMLWG telcon
17:25
<Ms2ger>
2001?
17:27
<annevk-cloud>
Sounds like someone is looking for a support forum
17:36
<Domenic_>
where should i direct web developers asking for background on "what is this DRM hubbub with the w3c"
17:36
Ms2ger
sticks his head in the sand
17:36
<Domenic_>
all i can think of is a few brendan eich posts which IIRC kind of assume you know what's going on already
17:38
<tantek>
Domenic_ - you could direct them to W3C blog posts on the subject
17:42
<MikeSmith>
Domenic_: http://longtermlaziness.wordpress.com/2013/10/08/the-w3c-is-a-restaurant/
17:47
<tantek>
good post MikeSmith
17:49
<MikeSmith>
tantek: thanks I paid David Bruant to write it and then he still managed to only get 80% of it right
17:50
<MikeSmith>
but I think I got good value for my money
17:50
<tantek>
it's a decent analogy
17:53
<Hixie>
it's a decent analogy except many restaurants would correctly turn away someone who came in waving a confederate flag or wearing a KKK hat
17:54
<Ms2ger>
Which reminds me that I noticed this morning that my university buys boxes of white chalk with "K" all along the outside
17:54
<Ms2ger>
And they are just wide enough to put three K's side by side
18:04
<jgraham>
It's a terrible analogy
18:04
<MikeSmith>
Ms2ger: don't bring that box into the restaurant that Hixie mentions. They might misunderstand. Especially if you stand on the box and wave your confederate flag.
18:04
<jgraham>
Argument by naology generally is
18:04
<jgraham>
*analogy
18:04
<Ms2ger>
MikeSmith, or they might laugh at me. It's not a large box :)
18:05
<jgraham>
At a resturant the resturant is in full control over the food + etc. The W3C isn't like that at all
18:07
<MikeSmith>
jgraham: put your money where your mouth is and pay David to write your side of the story
18:08
<jgraham>
Personally I see the W3C as like a standards organisation. One that has hired a lot of people who want to keep their jobs and so is likely to take decisions that will maintain its sources of income.
18:09
<MikeSmith>
well that sounds pretty unique
18:09
<MikeSmith>
there aren't much organizations like that
18:09
<MikeSmith>
sounds pretty crazy
18:09
<Ms2ger>
Do you like your job? ;)
18:10
<MikeSmith>
you want me to answer that question honestly?
18:10
<MikeSmith>
wait
18:10
<MikeSmith>
remind me what's my job
18:11
<Ms2ger>
Talking to monks
18:11
<jgraham>
MikeSmith: Fortunately the preponderance of other organisations that work in similar ways — not all of them standards organisations, even! — makes it easy to understand the W3C by comparison to the behaviour of those other organisations with which we are more familiar in my model. No such thing can be said of the "resturant" model.
18:13
<MikeSmith>
but I like David's restaurant
18:13
<MikeSmith>
you're right that all analogies suck
18:13
<MikeSmith>
some just suck less than others
18:14
<astearns>
a restaurant is not a for-profit organization?
18:15
<astearns>
arguments by analogy devolve pretty quickly to arguments over the analogy
18:15
<MikeSmith>
astearns: that's a feature
18:17
<MikeSmith>
jgraham: it's sometimes hard to find anybody who's made a point reasonably well who's not resorted to using an analogy
18:18
<SteveF>
MikeSmith: "remind me what's my job" MPAA fluffer
18:21
<jgraham>
MikeSmith: I dunno, it's usually just a case of waiting for hsivonen
18:24
<MikeSmith>
jgraham: oh but as I recall hsivonen recently got accused of using a bad analogy
18:25
<MikeSmith>
"emotive" even
18:26
<smola>
I feel like some cases in URL parsing in browsers were implemented just as an amusement...
18:26
<jgraham>
MikeSmith: Hmm, point
18:26
<jgraham>
I had forgotten that. Too much ice cream I guess.
18:26
<smola>
%ef%bc%85%ef%bc%94%ef%bc%91.com -> \uff05\uff14\uff11.com (%41 in full-width characters) -> %41.com -> a.com
18:26
<Domenic_>
annevk-cloud: is Mozilla at least implementing the URL spec's parsing algorithm?
18:32
<MikeSmith>
SteveF: applied for that job but was told another prominent WG member had already taken it. so oh well
18:33
<MikeSmith>
Ms2ger: btw I think that guy wasn't really a monk. Look at his shoes.
18:34
MikeSmith
goes back to fiddling with the w3c-test.org server to make CSS WG drafts disappear
18:36
<Hixie>
oh, bummer
18:36
<Hixie>
SimonSapin's bug about charset="" is next on my list
18:36
<Hixie>
quick, someone give me an excuse to do something else
18:38
<SimonSapin>
smola: amazing. Does it work across browsers?
18:38
<Hixie>
i guess i should jsut do all the CSSOM bugs at once
18:38
<smola>
SimonSapin: yes it does
18:38
<smola>
SimonSapin: Chrome and Firefox at least
18:39
<SimonSapin>
annevk-cloud: I don’t know if that’s in the URL spec ^
18:39
<SimonSapin>
not sure it should, either
18:39
<smola>
it isn't, I'm reporting it
18:41
<smola>
%ef%bc%85%ef%bc%94%ef%bc%91.com is even weirder
18:41
<smola>
SimonSapin: well, I can hardly imagine how can you get \uff05\uff14\uff11.com (manually entering %41.com instead of a.com in a CJK system)... but the second case,
18:41
<smola>
(swap the order of these sentences :p)
18:42
<smola>
I wonder if browsers work on any number of iterations of this encoding
18:43
<smola>
nope, if you repeat the process they do not work
18:54
<smaug____>
annevk-cloud: ping
18:56
<annevk-cloud>
Not today, sorry
19:05
<Hixie>
SimonSapin: ping https://www.w3.org/Bugs/Public/show_bug.cgi?id=24205 -- are we talking at cross-purposes? I don't understand your last comment.
19:07
<SimonSapin>
Hixie: as I understand the algorithm in the current spec, <meta http-equiv="Content-Language" content="fr and some random stuff"> will set the language to "fr"
19:07
<Hixie>
right
19:07
<Hixie>
that is as intended
19:07
<Hixie>
why would we not want that?
19:08
<SimonSapin>
I’d rather have it not set the language, like when there is a comma
19:08
<SimonSapin>
and I think browsers do that
19:09
<Hixie>
browsers seem to set it to the whole string
19:09
<Hixie>
as in "fr and some random stuff"
19:10
<Hixie>
the reason for aborting when there's a comma is that if there's a comma, it indicates the pragma is being used in some way like the HTTP header, and not like the pragma, and therefore doing anything with it is going to be bogus.
19:11
<SimonSapin>
ok, I got my testing wrong
19:11
<SimonSapin>
Chrome uses the whole thing, but not firefox
19:13
<SimonSapin>
well if that’s intentional, ok
19:13
<Hixie>
i can't tell if firefox is using content-language at all
19:14
<SimonSapin>
yes data:text/html,<meta http-equiv="content-language" content="fr"><style>body:lang(fr){background:green
19:15
<SimonSapin>
Hixie: I’m happy with just adding a note "Ignore the rest of the string" to show that this is intentional
19:16
<Hixie>
ah, typo in my test
19:16
<Hixie>
looks like firefox is setting the language to _seomthing_ when you have a bogus content-language
19:16
<Hixie>
but i can't work out what
19:16
<Hixie>
it's not "und", it's not leaving it set to the previous value
19:16
<SimonSapin>
"Ignore the rest of <var>input</var>"
19:16
<Hixie>
maybe "" but :lang() can't match that?
19:16
<Hixie>
k
19:16
<SimonSapin>
no, :lang() only takes idents, not strings
19:16
<gsnedders>
SimonSapin: As I just wrote on GitHub, uh, yeah, firstChunk is going to be horrible.
19:17
<gsnedders>
SimonSapin: Maybe we should just go for special-casing every single thing that's broken :(
19:17
<Hixie>
SimonSapin: right, but i'm trying to figure out how we could determine what the language was otherwise
19:17
<gsnedders>
SimonSapin: Could try avoid reading at all by doing isinstance(source, io.TextIO)/etc.
19:17
<Hixie>
not that it really matters
19:17
<Hixie>
but i presume it's internally the empty string
19:18
<gsnedders>
SimonSapin: But that doesn't work in general. But at least avoids hitting 200007 in Python 3.
19:18
<SimonSapin>
gsnedders: hum, do we really care about unicode streams? :)
19:19
<gsnedders>
SimonSapin: Yes. Consider TEXT fields in a database, for example.
19:21
<gsnedders>
SimonSapin: (Avoids hitting 200007 with urllib.response.HTTPResponse in Python 3, that is.)
19:23
<Hixie>
"When a ProcessingInstruction node node is inserted to a document, removed from a document, becomes part of the prolog, is no longer part of the prolog, ..."
19:23
<Hixie>
don't the first two there subsume the last two entirely?
19:28
<Ms2ger>
Hixie, can it become part of the prolog by removing a node from before it?
19:59
<Hixie>
Ms2ger: i guess you could remove the root element?
19:59
<Hixie>
Ms2ger: but it'd still be part of the document...
20:00
<Ms2ger>
I dunno, what's the prolog? :)
20:00
<Ms2ger>
Except a sucky programming language
20:02
<jarek>
are there any plans to have standarized UI guidelines?
20:02
<jarek>
I mean something more advanced than CSS3 UI spec
20:03
<Ms2ger>
In general, browsers compete on UI
20:04
<jarek>
e.g. something akin to Human Interface Guidelines
20:04
<jarek>
by Apple
20:05
<jarek>
yeah, Mozilla works on pseudo-standard components in form of x-tags
20:05
<jarek>
and Google works on Polymer widgets
20:06
<jory>
Is there a spec re: viewport meta tag behaviour, or is it still just something that everyone kind of has implemented the same way?
20:06
<Ms2ger>
There's something in css
20:07
<Ms2ger>
http://dev.w3.org/csswg/css-device-adapt/
20:09
<jory>
Thanks Ms2ger
20:12
<jory>
One thing that I don't think is covered is the behaviour of the browser when the tag is added after the document has loaded.
20:13
<Ms2ger>
Quite possible
20:13
<jory>
Such edge cases. Many behaviours.
20:14
<Ms2ger>
Wow?
20:15
<jory>
Quite wow.
20:15
<jory>
heh
21:37
<zcorpan>
https://www.w3.org/Bugs/Public/show_bug.cgi?id=24257 is crazy
21:52
<zcorpan>
Hixie: being part of the document isn't enough to make xml-stylesheet "work". it needs to be in the prolog. hence the need to run the steps again if the root element is moved
21:53
<Hixie>
the steps run regardless of whether it's in the prolog
21:54
<Hixie>
and the steps don't mention the prolog
21:54
<Hixie>
(i filed a bug on this btw)
21:55
<zcorpan>
Hixie: it's mentioned in the xml-stylesheet spec
21:56
<Hixie>
xml-ss overrides cssom somehow?
21:56
<zcorpan>
no
21:56
<zcorpan>
cssom says "If node is not an xml-stylesheet processing instruction, terminate these steps."
21:56
<zcorpan>
-> http://www.w3.org/TR/xml-stylesheet/#dt-xml-stylesheet
21:57
<zcorpan>
-> http://www.w3.org/TR/xml-stylesheet/#dt-potential-xml-stylesheet
21:57
<zcorpan>
apologize for the infoset gibberish, but it's there
21:58
<zcorpan>
<root/><?xml-stylesheet?> isn't an "xml-stylesheet processing instruction"
21:58
<Ms2ger>
Sounds like we need an xmlss.spec.whatwg.org ;)
21:59
<zcorpan>
it would have been less painful than doing it in the xml core wg
22:18
<Hixie>
zcorpan: oh. then why the bit about the document?
22:20
<zcorpan>
Hixie: hmm. maybe it's not necessary
22:22
Ms2ger
wonders if jgraham feels like making a graph of the number of open w-p-t reviews
22:26
<zcorpan>
if i remove that, i can also remove the "part of the doctype" check (i guess it was no-op already since dom doesn't support that)
22:34
<zcorpan>
hmm. dvcs down :-(
22:34
zcorpan
tries again tomorrow
22:40
<Ms2ger>
MikeSmith, ^
22:51
<Hixie>
zcorpan: do we really want to refetch <link rel=stylesheet> if type="text/css" is added or removed?
22:51
<Hixie>
(given that the default is text/css?)
22:51
<Hixie>
also, are you handling _all_ rel=stylesheet links? including e.g. xslt?
22:52
<Hixie>
or should i only send you type=text/css ones?
22:53
<zcorpan>
Hixie: re first question, maybe not
22:54
<zcorpan>
Hixie: cssom only handles css. though xslt doesn't work with <link>
22:54
<Hixie>
i'm pretending that there might be some other style sheet language some day
22:55
<Ms2ger>
Why?
22:55
<Hixie>
i'll only send you the stuff if it's text/css. that means i'll keep the quirk, btw.
22:55
<Hixie>
why what?
22:56
<zcorpan>
why do you need to keep the quirk?
22:57
<Ms2ger>
<Hixie> i'm pretending that there might be some other style sheet language some day
22:57
<Hixie>
zcorpan: because i can't send it to you until i know it's css, and i don't know it's css until i've applied the quirk
22:57
<Hixie>
Ms2ger: seems like the responsible thing to do
22:58
<zcorpan>
Hixie: ok
23:02
<zcorpan>
Hixie: bedtime for me, but feel free to dump questions here, i'll check tomorrow
23:03
<Hixie>
zcorpan: roger
23:03
<Hixie>
zcorpan: nn
23:03
<zcorpan>
night
23:16
<Hixie>
is there a "destroy a css style sheet"? hmm...