02:39
<Hixie>
jamesr_: no idea. i think hallvord was the last custodian, try asking him.
03:10
<MikeSmith>
Hixie: which spec was that?
04:01
<MikeSmith>
"An expert is a man who has stopped thinking because he knows."
04:27
<Hixie>
MikeSmith: hit testing in IE
04:28
<MikeSmith>
ah
05:53
<annevk>
Latest on hit testing was this http://lists.w3.org/Archives/Public/www-style/2010Aug/0407.html
05:53
<annevk>
Then tantek was going to add it to CSS3 UI
05:54
<annevk>
Then that did not happen and it was going to be CSS4 UI
05:54
<annevk>
I kind of lost hope. They should just have let Leif do it
06:04
<annevk>
jamesr_: Hixie: ^^
06:26
<rniwa>
annevk: hit testing is hard :\
06:26
<rniwa>
annevk: i think one thing we need to do is decide the scope of the spec.
06:27
<rniwa>
annevk: if you try to spec everything including caret position, selection, etc... then it becomes an impossible task
06:27
<rniwa>
annevk: if, however, we restrict ourselves to be focusing on things like just elements, then there might be some hope :)
06:28
<annevk>
rniwa: I've mostly been thinking about mouse events and :hover, but editing is indeed a problem too
06:28
<rniwa>
annevk: yeah, just spec'ing mouse events and :hover will be good for most use cases.
06:28
<rniwa>
annevk: i don't think we can spec hit testing for editing because of different platform conventions
06:28
<annevk>
e.g. http://dev.w3.org/csswg/cssom-view/#dom-document-caretpositionfrompoint is not defined much at all
06:29
<rniwa>
annevk: i mean can you imagine having to specify 5-6 different conventions for different caret behaviors around bidi embedding boundaries?
06:29
<annevk>
rniwa: I can imagine it would be painful :)
06:29
<rniwa>
annevk: i know at least windows, mac, and linux all have different behaviors when you click around bidi boundary.
06:30
<rniwa>
annevk: not to mention i think android & chrome os also use different conventions from all the rest :(
06:30
<rniwa>
annevk: i'm sure there are plenty more on this.
06:30
<annevk>
yeah maybe once you get to that point it would be okay to let it be platform-dependent with some rough guidelines
06:30
<rniwa>
yup.
06:30
<rniwa>
it's not realistic to spec everything at some point.
06:30
<annevk>
but which text you hit depending on which element it is in would be a great start
06:30
<rniwa>
annevk: btw, we should really fix click event
06:31
<rniwa>
annevk: it's insane that we don't fire click event whenever cursor moves across an element boundary :\
06:31
<rniwa>
annevk: (the spec currently says we fire click only if mousedown & mouseup happen on the same target)
06:31
<rniwa>
annevk: (which is very unlikely scenario if you have tiny spans, etc...)
06:32
<annevk>
is that actually what is implemented?
06:32
<rniwa>
annevk: yeah, in webkit :(
06:32
<rniwa>
annevk: i really want to fix it.
06:32
<rniwa>
annevk: because our current behavior is insanely bad from UX stand point of view.
06:32
<rniwa>
annevk: i think almost all other browsers do something different, which i don't quite understand.
06:32
<annevk>
UI Events spec is kind of a mess
06:32
<rniwa>
annevk: e.g. IE always fires click event
06:33
<rniwa>
annevk: indeed.
06:33
<rniwa>
annevk: but that's the most important kind of events!
06:33
<rniwa>
annevk: i mean... who cares about all those exotic events. UI events are the one developers use the most.
06:34
<rniwa>
but it's kind of understandable the way it is now. spec'ing UI events is hard :\
06:34
<annevk>
I guess the problem is that they're hard and that properly defining them is not something many people can do
06:34
<annevk>
:)
06:34
<rniwa>
annevk: yup.
06:35
<rniwa>
annevk: i've seen enough code in webkit to know that much.
06:35
<annevk>
e.g. focus events are a mess too
06:36
<rniwa>
annevk: yeah.
06:36
<rniwa>
annevk: oh my... don't get me started on focusin, focusout, etc...
06:36
<rniwa>
they're giving me nightmares.
06:36
<annevk>
better not talk about them now then :p
06:36
<rniwa>
you can talk about it.
06:37
<rniwa>
but i might just rant for an hour :P
06:37
<rniwa>
annevk: i think the biggest headache for me right now is the relationship between selection & focus
06:38
<rniwa>
annevk: in webkit, focus follows selection
06:38
<rniwa>
on all other browsers, selection follows focus
06:38
<rniwa>
meaning that if you type something, the character appears at where selection is in webkit
06:38
<rniwa>
but at where focus is on all other browsers.
06:38
<rniwa>
apparently this is Mac-ism
06:40
<annevk>
how do you reproduce that?
06:42
<rniwa>
annevk: put focus & selection on different nodes
06:42
<rniwa>
and type a character
06:43
<annevk>
oh if you select something via an API focus is not changed?
06:44
<rniwa>
annevk: yeah something like that
06:44
<rniwa>
it's a little tricky to do.
06:46
<annevk>
I need to go get breakfast before they stop serving it
06:46
<rniwa>
annevk: we'll see if i can come up with an example by the time you come back ;)
06:47
<annevk>
I've kind of wanted to define UI events for a long time, but the last time I attempted with keyboard events I got stuck
06:52
<rniwa>
annevk: https://bug-56271-attachments.webkit.org/attachment.cgi?id=85623 is a fun demo
06:53
<rniwa>
(from https://bugs.webkit.org/show_bug.cgi?id=56271)
06:53
<rniwa>
Ms2ger: hi Ms2ger! how about some UI events discussion :D
06:54
<Ms2ger>
How about replaceChild? :)
06:54
<rniwa>
Ms2ger: LOL. I see what you did there.
07:00
<Velmont>
Oh what a beautiful moorning~
07:13
Ms2ger
writes some tests for WebKit to fail
07:48
<asmodai>
annevk: http://www.mediawiki.org/wiki/Extension:Math/MathJax_testing
08:48
<annevk>
Ms2ger: http://jamesmckay.net/2012/06/you-can-no-longer-afford-not-to-take-git-seriously/
08:52
<annevk>
uhm http://validator.nu/?doc=http%3A%2F%2Fhtml5boilerplate.com%2F
08:52
<annevk>
paul_irish: ^^
08:52
<Ms2ger>
<!doctype html public "✰"
08:52
<Ms2ger>
Oh really
08:53
<Ms2ger>
http://validator.nu/?doc=http%3A%2F%2Fwww.whatwg.org%2Fissues%2Fdata.html
08:53
<annevk>
HTML is hard
08:54
<Ms2ger>
Let's go shopping?
08:54
<annevk>
it's raining outside
08:55
<Ms2ger>
Sunny here
08:55
<Ms2ger>
... for now
08:57
<Velmont>
You say I'm leaving sunny southern europe to go up to rainy Norway? :-(
08:58
<annevk>
Velmont: you could stay :)
08:59
<Velmont>
annevk: And let y'all get the best places at the new office? Neh, I'll pass :]
08:59
<Ms2ger>
We have an office in Paris if you're interested in working further south? :)
09:00
<Velmont>
lol ;D
09:02
<Ms2ger>
(The same is true for other Operians, of course)
09:03
<annevk>
Ms2ger: are you actually employed these days or just helping out Mozilla recruiting on your personal time?
09:03
<Ms2ger>
The latter :)
09:09
<Velmont>
Why does everyone try to brain drain Opera all the time?
09:09
<Velmont>
What about stealing a bit from the others?
09:10
<Ms2ger>
We want smart people
09:33
<Velmont>
MikeSmith: I updated http://w3c-test.org/webappsec/tests/cors/submitted/opera/staging/resources/cors-makeheader.php -- makes it possible to send 2 allow-origin headers in response. For some extra tests for bugs in WebKit and Fx.
09:34
<Velmont>
MikeSmith: I might update it even more on monday, because I see I'm not checking the case where crossorigin.site.com -> site.com, all the tests are doing site.com -> crossorigin.site.com. Might be some bugs hiding there.
09:35
<Ms2ger>
Bugs in Gecko? Couldn't be
09:35
<annevk>
Opera found them first
09:35
<Ms2ger>
Well played
09:37
<Ms2ger>
And Opera had the weird throwing-nodes-as-exceptions first? :)
10:40
<Von_Davidicus>
Why is it that, in HTML5, the contents of a header attribute MUST refer to the ID of a <th> element in the same table, and what can I do to get around that headache?
10:46
<Ms2ger>
Don't use that attribute if you're going to misuse it
10:47
<Von_Davidicus>
Well, I gotta use SOMETHING to make the JS on my webpage work--I was using the headers attributes to refer to the IDs of <input> elements.
10:58
<Von_Davidicus>
I guess I could use the class attribute... *Misses axis, which was another useful info dump attribute*
11:56
<Ms2ger>
MikeSmith, http://w3c-test.org/framework/ is broken
17:55
<methuselahah>
where can I find the most detailed html5 javascript documentation for html5 video?