00:41
<Hixie>
man, do we still not have an answer to people who want a set of push buttons that act like a radio button set?
00:44
<annevk>
Hixie: aren't forms your area?
00:44
<Hixie>
this is rendering
00:44
annevk
wonders what a push button is
00:44
<annevk>
Hixie: ah, your question is whether the CSS WG has finally acknowledged they need to address form control styling?
00:44
<Hixie>
well either CSS or web components, but yeah
00:44
<Hixie>
basically :-)
00:45
<Hixie>
i mean like a version of <select> that shows adjacent buttons instead of a drop-down
00:45
<Hixie>
like: [ Arr ] [ Dep ]
00:45
<Hixie>
instead of ( ) Arr (o) Dep
00:45
<Hixie>
or [ Arr |v]
00:46
<annevk>
I think components wants to at least explain form controls, but I'm not sure how far that is
00:46
<annevk>
But I don't think there's anyone trying to solve this issue
00:46
<Hixie>
this is probably the second most important thing to work on, imho
00:47
<Hixie>
in css
00:47
<Hixie>
the first being mobile-related UI stuff
00:47
<Hixie>
though there might not be much of that needed in the CSS world
00:48
<annevk>
So offline is my #1 thing this year and fixing various issues with all my specs. Forms and UI events were things I was thinking about tackling once I got tired with those
00:48
<annevk>
But I need to consider that more carefully since they are sinkholes
00:48
<Hixie>
yeah
00:49
<Hixie>
i've tried to fix it twice so far
00:49
<Hixie>
web controls, and xbl
00:49
<Hixie>
someone else's turn now :-)
01:05
<Domenic_>
I was also considering form styling
01:05
<Domenic_>
But it's pretty hard to find time given this is a hobby, so things like streams and even promsies seem likely to occupy my free time for quite a while.
01:06
<annevk>
Someone needs to hire you :/
01:06
<annevk>
Yeah, streams are important too
01:07
<annevk>
But I guess form styling is more overdue
03:22
<Domenic_>
What did we do before promises? So far i have IDBRequest and File API's successCallback, errorCallback. Were there more?
03:23
<Domenic_>
i guess fullscreen API having you listen to separate events to see if your method call succeeded is another example
06:03
<annevk-cloud>
Domenic_: fullscreen is different, events are fired in other documents too
06:05
<annevk-cloud>
Domenic_: also see WebRTC, toBlob on canvas, geolocation (also has more complicated API promises do not work for), etc.
06:36
<MikeSmith>
zcorpan_: thanks for the heads-up about http://html5.org/tools/web-apps-tracker?from=8432&to=8433 (making caption end tag optional)
06:37
<zcorpan_>
np
06:45
<MikeSmith>
zcorpan_: also about the validator doctype-sniffing option in the REST API yeah, the change is, no sniffing except for opt-in in web service api
06:45
<MikeSmith>
but I don't plan to remove the option to sniff later
06:45
<zcorpan_>
MikeSmith: ok, sounds good
06:46
<MikeSmith>
ok thanks
06:46
<bhanu>
hi, I have few queries on recent changes being made on range methods, could anybody help me understand?
06:46
<zcorpan_>
bhanu: don't ask to ask just ask :-)
06:49
<bhanu>
ok, here are the queries : first being in DOM2 for setStartBefore/setStartAfter/setEndBefore/setEndAfter the specification says InvalidNodeTypeError should be thrown when root container of refNode is not an Attr, Document, or DocumentFragment node
06:50
<bhanu>
in DOM4 which is a work in draft its mentioned if parent of refNode is null, then throw InvalidNodeTypeError exception
06:52
<bhanu>
again there is no mention about what type of refNode would lead to the InvalidNodeTypeError exception
06:54
<zcorpan_>
bhanu: you're looking at http://dom.spec.whatwg.org/ yes?
06:55
<bhanu>
zcorpan : yes
06:56
<zcorpan_>
i find http://dom.spec.whatwg.org/#concept-range-bp-set
06:58
<zcorpan_>
so if parent is null or a doctype, you get InvalidNodeTypeError, but doctypes can't have children so it's just if parent is null for those methods
06:59
<bhanu>
parent specification is fine
07:01
<bhanu>
If I am not wrong the root node requirement is eradicated right which was there is in DOM2?
07:01
<zcorpan_>
oh so dom2 throws if the root is e.g. an element?
07:02
<bhanu>
yes
07:02
<bhanu>
I believe that is why this change is brought in
07:02
<zcorpan_>
yeah i don't see that in the dom spec. what do browsers do?
07:03
<bhanu>
I tried in chrome and mozilla, it doesn't throw any exception
07:03
<bhanu>
but its a old bug in webkit
07:06
<bhanu>
again there was restriction on the refNode in DOM2 i.e it shouldn't be any of Document, DocumentFragment, Attr, Entity, or Notation node
07:06
<bhanu>
which i don't see in DOM4
07:09
<zcorpan_>
i'd recommend ignoring what dom2 says and go by what the new spec says and what browsers do (in case all browsers disagree with the spec, maybe the spec should be changed)
07:10
<bhanu>
ok, thanks
08:37
<gsnedders>
Is there any way to make a table split into columns using CSS?
09:21
<zcorpan_>
gsnedders: doesn't the html spec do that for the entity table?
09:22
<zcorpan_>
Hixie: https://www.w3.org/Bugs/Public/show_bug.cgi?id=24411 seems like a bug for html, right?
09:26
<zcorpan_>
TabAtkins: do you plan to fix https://github.com/ResponsiveImagesCG/picture-element/issues/91 soon? it seems imperative to actually implementing the spec and writing tests
10:24
<gsnedders>
zcorpan_: I couldn't get CSS Multicol to work :(
10:27
<zcorpan_>
gsnedders: the spec's entity table multicols for me in presto and blink
10:29
<gsnedders>
zcorpan_: I dunno. I was trying really quickly just before printing stuff out.
10:30
<zcorpan_>
hsivonen: WDYT about https://www.w3.org/Bugs/Public/show_bug.cgi?id=24439 ?
10:31
<zcorpan_>
dom consistency was an argument in 2006 but maybe it's not so important today?
11:15
<jgraham>
zcorpan_: see https://tbpl.mozilla.org/php/getParsedLog.php?id=33794007&tree=Cedar the tests seem to time out in gecko? Could be expected I suppose
11:19
<zcorpan_>
jgraham: do you have https://github.com/w3c/web-platform-tests/pull/534 in?
11:19
<zcorpan_>
jgraham: the css tests don't time out in gecko for me
11:20
<hsivonen>
zcorpan_: commented on the bug
11:20
<zcorpan_>
jgraham: only 'loading webvtt <track>' which is expected (it's a gc bug i think)
11:20
<jgraham>
zcorpan_: Not sure, my sync is a bit out of date, but I can fix that easilly enough
11:21
<jgraham>
zcorpan_: OK. The weird thing is that those tests supposedly didn't time out on my machine. But that doesn't seem to be reproducable.
11:22
<zcorpan_>
hsivonen: thx
11:35
<zcorpan_>
jgraham: hmm. i'm confused. i want to move the structured clone tests from Workers to workers, but in my working copy git has them at both locations. :-S
11:37
<zcorpan_>
jgraham: could you try moving them?
11:39
<jgraham>
zcorpan_: I see files in Workers/semantics/structured-clone/ but nothing else under Workers. They don't seem to be under workers/
11:40
<zcorpan_>
yeah that matches what the web interface says on github, but in my checkout i have all the tests in both locations. dunno if something gets confused by the names differing only by case or something
11:41
<jgraham>
zcorpan_: OK, I moved the files that were in Workers
11:42
<zcorpan_>
thanks!
11:42
<jgraham>
So now that shouldn't exist at all
12:17
hsivonen
wonders if howcome has an opinion on Client Hints
12:18
hsivonen
is late to Regions Considered Harmful
13:42
<MikeSmith>
hsivonen: I can't recall what his views are on conneg
13:43
<MikeSmith>
but one thing that's disturbing about CH is its potential giving of new life to conneg
13:50
<yoav>
MikeSmith: Do you hold negative views RE conneg even if it's server-initiated opt-in conneg? Because that's what CH is
13:56
<MikeSmith>
yoav: I am keeping an open mind but I think it's pretty clear that conneg on the Web so far has not been a big success
13:56
<MikeSmith>
maybe CH will change that
13:58
<yoav>
On the one hand I agree (Content-Encoding being the best example). On the other hand, I believe 'key' and TLS may change that.
13:58
<yoav>
And opt-in conneg has very limited risk
13:58
<MikeSmith>
yeah relatively
13:59
<yoav>
If it doesn't work, we can remove it. No risk of polluting requests forever.
13:59
<MikeSmith>
that's definitely a nice aspect
14:00
<MikeSmith>
I'm curious why hsivonen thought Haakon might have a particular opinion on CH
14:00
<yoav>
Now I am too :)
14:25
<zcorpan_>
yoav: if we implement CH and sites start to rely on it being supported, it can't be removed
14:28
jgraham
thinks that CH is probably not Switzerland, but hasn't yet figured out what it is
14:28
<Ms2ger>
* hsivonen wonders if howcome has an opinion on Client Hints
14:29
<jgraham>
Yes, but what does CH stand for? ;)
14:39
<ondras>
Cyanide & Happiness ?
15:11
<whinzou>
hello
15:42
GPHemsley
wonders what a Client Hint is
15:45
<MikeSmith>
Internet Draft
15:45
<MikeSmith>
GPHemsley: ↑
15:46
<MikeSmith>
from Ilya at Google
15:50
<jory>
So in theory Client Hints would allow mobile devices to specify different DPRs depending on available bandwidth...
17:19
<dglazkov>
good morning, Whatwg!
17:21
<jcgregorio>
good morning dglazkov
18:04
<annevk>
Domenic_: new delay() looks better
18:05
<annevk>
Domenic_: seems about the same as saying to return and run the remaining steps async
18:05
<Domenic_>
yeah
18:05
<Domenic_>
as i said might abandon that attempt to make things make more sense to me
19:10
<annevk>
Hixie: http://www.unicode.org/mail-arch/unicode-ml/y2014-m01/0053.html explains the situation
19:11
<Hixie>
that e-mail... makes no sense
19:12
<annevk>
:-)
19:12
<Ms2ger>
Did you look at the mailing list it was sent to first?
19:12
<Hixie>
what's wrong with "windows1252" as the name?
19:12
<Hixie>
windows-1252, even
19:12
<Ms2ger>
Already taken
19:12
<annevk>
I think the context is a Python library, but I'm not completely sure
19:15
<Hixie>
Ms2ger: isn't it already taken to mean exactly what this guy means? "what browsers do"?
19:15
<Ms2ger>
I dunno
19:27
<annevk>
Hixie: I think we convinced TC39 to take on structured cloning; might take a while for it to be integrated and such though
19:27
<Hixie>
when you say "take on structured cloning"
19:27
<Hixie>
you mean "provide hooks for structured cloning", right?
19:27
<Hixie>
i mean, i don't want us to change to non-DOM exceptions and all that just because it moves to a new forum
19:28
<annevk>
Hixie: yeah, and define how it behaves for new objects themselves
19:29
<annevk>
Hixie: DOMException is going to be just another Error subclass
19:29
<annevk>
Hixie: the plan is to move it to IDL and have IDL provide the hooks for throwing
19:29
<annevk>
Hixie: rather than DOM
19:30
<Hixie>
so long as it's not churning on authors for no good reason
19:31
<annevk>
I think apart from going to getter/setter to data properties it would remain identical and would make it clear that no new exceptions are to be minted outside IDL
19:41
<jory>
Hey I wanted to update a broken link on the wiki.
19:42
<jory>
http://wiki.whatwg.org/wiki/MetaExtensions
19:42
<jory>
The viewport-meta link is broken.
19:42
<annevk>
jory: go ahead
19:44
<jory>
annevk: It says I need to get access from one of the IRC regulars
19:45
<annevk>
jory: PM me your desired username and password
19:55
<jcgregorio>
Hixie: Canvas "Path" vs "Path2D" http://lists.w3.org/Archives/Public/public-whatwg-archive/2013Nov/0138.html
19:55
<jcgregorio>
is there agreement to update the spec?
19:56
<annevk>
Isn't it weird if Path is the only thing with a suffix?
19:57
<Hixie>
jcgregorio: i thought safari had shipped with it as Path
19:58
<jcgregorio>
Hixie, I'll have to check that, I was under the impression no one had implemented it yet
19:59
<Hixie>
there's been discussion of it on the whatwg list, but i can't find my last e-mail on the topic
20:00
<jcgregorio>
annevk: true, also for the sake of symmetry it's not like there's a lot of *3D objects in WebGL
20:00
<Hixie>
i guess http://lists.w3.org/Archives/Public/public-whatwg-archive/2013Nov/0125.html is the last thing from me on it
20:01
<Domenic_>
new goal in life: be the first to spec a 4D-suffixed class
20:03
<jcgregorio>
Hixie: sgtm, thanks
20:03
<Hixie>
jcgregorio: at the end of the day i'll spec whatever ships most widely, obviously
20:05
<annevk>
Safari has Path
20:05
<annevk>
We should not rename
20:06
<annevk>
Proof: http://software.hixie.ch/utilities/js/live-dom-viewer/saved/2788
20:06
<annevk>
jcgregorio: Hixie: ^^
20:07
<jcgregorio>
annevk: thanks!
20:09
<Hixie>
annevk: can you get it to do anything?
20:10
<Hixie>
annevk: e.g. http://goo.gl/iXggPJ
20:10
<Hixie>
annevk: doesn't do anything?
20:13
<annevk>
Hixie: hmm
20:14
<annevk>
hober: ^^
20:14
annevk
is fixing URLSearchParams
20:16
<MikeSmith>
annevk: did you see that othermaciej mentioned he's implementing the URL spec?
20:18
<annevk>
MikeSmith: yeah talked about it with him
20:18
<othermaciej>
I told annevk in person :-)
20:18
<othermaciej>
I will probably land the first patch today
20:18
<MikeSmith>
cool
20:18
<othermaciej>
but it is missing features and super buggy from the parsing and part replacement perspective
20:18
<othermaciej>
annevk: btw I hate 'URLUtils' as an interface name
20:18
<othermaciej>
annevk: it really should be something like 'URLParts' or 'URLDecomposition' or something
20:18
<annevk>
hah, I copied something from Hixie I think
20:19
<annevk>
that might be better
20:19
<annevk>
it's not exposed to script though
20:19
<othermaciej>
yeah I know
20:19
<othermaciej>
but I matched the interface name in our IDL anyway
20:19
<othermaciej>
(so people reading our code and the spec don't get confused)
21:57
<Hixie>
foolip: you around? you got any ideas for how to fix https://www.w3.org/Bugs/Public/show_bug.cgi?id=24350 ?
23:59
<annevk>
whoa
23:59
<annevk>
major shift in promises
23:59
<annevk>
for the better
23:59
<annevk>
it seems
23:59
<Domenic_>
:)