00:13
<heycam>
jamesr_, you can consider abarth as "internal affairs" then :)
00:13
<heycam>
where internal affairs isn't aware of Management's nefarious plot
00:14
<Hixie>
aw man
00:14
<Hixie>
how did i end up the bad guy here!
00:14
<Hixie>
i was just following orders!
00:15
heycam
thinks perhaps we shouldn't talk of this lest tom's hardware write a scathing expose
00:45
<gsnedders>
http://dev.ckeditor.com/ticket/6666 — yay!
07:55
<hsivonen>
hober, AryehGregor, zcorpan: Script execution post fixed. thanks
07:55
<hsivonen>
zcorpan: missing favicon acknowledged. thanks. (not fixed yet)
09:51
<zcorpan>
so for <img src=404>, in ie .complete is false while in gecko/webkit/opera it's true
09:51
<zcorpan>
afaict the spec sides with ie for that case
09:52
<zcorpan>
do we want to change the spec?
09:53
<annevk>
prolly
09:56
<zcorpan>
it makes .complete less useful though
09:57
<Ms2ger>
Even less useful?
09:57
<zcorpan>
yes
09:58
<zcorpan>
i was using it in my media testsuite and expected it to be false for 404, but now we had to add a workaround :)
09:59
<annevk>
should it be false for all error status codes?
10:00
<annevk>
what about a 301 without a Location header?
10:00
<zcorpan>
"no data could be obtained" http://www.whatwg.org/specs/web-apps/current-work/complete/embedded-content-1.html#img-error
10:01
<annevk>
oh, so in that case it ought to just work
10:01
<annevk>
if the 404 is an image
10:01
<zcorpan>
actually the status code is not so relevant for <img>, the relevant part is if the body is a valid image
10:01
<zcorpan>
yeah
10:01
<annevk>
I was already wondering when Hixie would have changed his mind on this
10:02
<annevk>
because "we" decided on this behavior long ago
10:02
<zcorpan>
so an HTML 404 is the same as a "broken" image
10:03
<annevk>
it's the same as a HTML 200
10:41
<annevk>
there's a typo in the novel
10:42
<annevk>
"He cackled to himself as his committed" s/his/he/ right?
10:42
<annevk>
or is using "his" some kind of old English I am not familiar with?
10:45
<jgraham>
No, it's a typo
10:46
<jgraham>
Did anyone ever mirror the webapps repo in some sensible format?
10:46
<jgraham>
i.e. git/mercurial
10:47
<jgraham>
(and keep it up to date)?
10:47
<MikeSmith>
jgraham: not that I know of
10:47
<MikeSmith>
if you think it's useful I could set it up at github
10:49
<jgraham>
MikeSmith: I would find it useful right about now :)
11:24
<MikeSmith>
jgraham: importing now
11:24
<MikeSmith>
it's taking a while
11:25
<jgraham>
MikeSmith: I actually started doing it myself as well. I guess Hixie loves us…
11:25
<jgraham>
(not to github though, which indeed seems more useful)
11:31
<MikeSmith>
jgraham: wiil be at https://github.com/sideshowbarker/html5-spec when it's done
11:31
<MikeSmith>
git⊙gc:sideshowbarker/html5-spec.git
11:31
<gsnedders>
A git-svn clone for the spec?
11:31
gsnedders
has one of those locally :P
11:32
<jgraham>
gsnedders: If you had said that like an hour ago (and uploaded it somewhere sensible) you coudl have saved everyone lots of heartache
11:33
gsnedders
was writing a five page essay worth 2% of the course…
11:36
<MikeSmith>
use really small paper
11:42
<jgraham>
gsnedders: It is likely worth 0% of your final degree. The point of the excerise is to *learn*
11:42
<jgraham>
Not to accumulate points
11:43
<gsnedders>
jgraham: I've learnt I've totally forgotten everything I knew about writing essays from school :)
11:43
<jgraham>
(although the points are useful)
11:45
<MikeSmith>
the point it to learn how to parrot your instructors' words in such a way that reinforces their egos so they give you higher marks
11:46
<MikeSmith>
make them see that you understand well how clever they are
11:47
<Philip`>
Be sure to get a section on pathetic fallacy in there too
11:47
<Philip`>
That's the main thing I remember from writing essays
14:12
<jgraham>
Is it me or does the spec allow literal U+0000 through in the foster parenting case?
14:15
jgraham
files a bug
14:21
<gsnedders>
I thought U+0000 was handled within the tokenizer so where it goes within the tree constructor is surely irrelevent?
14:21
gsnedders
hasn't read the spec closely since it was changed for null-handling
14:25
<jgraham>
No, now the handling is split over the treebuilder and the tokenizer in a slightly ugly way
14:30
<annevk>
parser changes seem to confuse everyone and satisfy nobody :/
14:31
<annevk>
oh, so Apple will keep Java support on Lion?
14:31
<hsivonen>
annevk: that's because Hixie didn't spec *exactly* the conceptual model I described :-)
14:31
<hsivonen>
annevk: URL about Java?
14:33
<hsivonen>
in other parser news, we got another <foo </bar> bug
14:33
<hsivonen>
(sent over to evang, so nothing special about it. just noting that those bugs do show up)
14:35
<annevk>
hsivonen, http://www.apple.com/pr/library/2010/11/12openjdk.html
14:35
<Philip`>
Is there any way the spec could distinguish between cases where content relies on it closing the foo, and where content relies on it not closing the foo?
14:35
<Philip`>
(like the exact tokeniser state in which the '<' is encountered)
14:36
<hsivonen>
annevk: cool. Apple contributing their previously proprietary code to OpenJDK is probably the best outcome for all stake holders
14:37
<hsivonen>
Philip`: I think it would be even more confusing to vary < by state
14:38
<Philip`>
hsivonen: Maybe increasing compatibility is more important than decreasing confusion
14:38
<Philip`>
(I don't know if it's really possible to increase compatibility by complexifying the logic, though)
14:39
<hsivonen>
It would be awesome if we could just make things more complex to make them more compatible :-)
14:40
<hsivonen>
(that is, making them more complex in whatever random ways :-)
14:42
<Philip`>
Hmm, random complexity seems unlikely to work, unless we could develop some automatic measure of compatibility and then use a genetic algorithm to evolve the parser algorithm to optimise compatibility
14:43
<zcorpan>
we could change the parser every week based on the output of the compat measurement
14:44
<jgraham>
The problem would be measuring compatibility
14:45
<hsivonen>
I wonder how common it is that sites respond to beta cycle evang by saying that they'll fix the site once the browser is final (as in no longer beta)
14:46
<hsivonen>
as opposed to fixing the site right away
14:46
<Philip`>
Maybe they don't want to make changes they can't test, and don't want to install beta browsers for testing since it may disturb their current browser installations
14:47
<Philip`>
Or maybe they think that if they hold out long enough then you'll revert the browser changes and they won't have to do anything
14:49
<jgraham>
I wonder what will happen to foreignContent mode (again)
14:49
<hsivonen>
jgraham: which bug?
14:49
<jgraham>
hsivonen: Dunno if there was a bug
14:50
<Philip`>
Re http://www.apple.com/pr/library/2010/11/12openjdk.html : "Apple [...] has recently introduced its magical iPad" - they actually call it magical?
14:50
<jgraham>
I should have filed one on the changes to end tag handling in the last updates, but I don't recall if I did
14:53
<hsivonen>
jgraham: would that one get fixed if Hixie adopted the model I've been proposing?
14:53
<jgraham>
Yeah, I think so
14:54
<hsivonen>
Philip`: you should watch one of those Stevenote remixes with only the adjectives
14:54
<hsivonen>
jgraham: let's just do it then
14:54
<jgraham>
I was vaugely contemplating making an experimental spec change to the model you proposed
14:55
<jgraham>
Because it seems like the big barrier at this point is that it is a big change and Hixie only has the bandwidth for bugfixes
15:14
hsivonen
is a bit surprised that https://bugzilla.mozilla.org/show_bug.cgi?id=587452 hasn't been fixed at Google's end yet
15:15
<annevk>
smaug____, I made a small update to the XHR tests
15:15
<annevk>
smaug____, they now include a reference to testharnessreport.js
15:15
<annevk>
smaug____, that also made http://tc.labs.opera.com/apis/XMLHttpRequest/testrunner.htm possible
15:15
<annevk>
smaug____, although the way that works is not entirely how I'd like it
15:16
<smaug____>
annevk: ok, will look at that soon
16:01
<annevk>
hsivonen, so this OpenJDK is not so open that Google can use it for Android in a slightly different way?
16:17
<tabatkins>
hsivonen: Do you know if anyone's bugged us about it yet?
16:37
<annevk>
Ms2ger, html5.org is updated now
16:37
<Ms2ger>
Thanks
16:37
<annevk>
Ms2ger, kind of strange that bitbucket.org does not propagate things directly
16:37
<Ms2ger>
Yeah
16:39
<annevk>
so the parsing spec should prolly say when <script> is executed and when it isn't
16:40
<Ms2ger>
I'm waiting for Hixie to give me a flag
16:40
<Ms2ger>
Currently, it's conditional on being a fragment parser
19:10
<AryehGregor>
hsivonen, "cross-broser-compatible" http://hsivonen.iki.fi/script-execution/
19:11
<Ms2ger>
hsivonen, did all this land for b7?
20:13
<cying>
CSS nerdy question:
20:13
<cying>
how do i style the first child of a container's first line in all caps?
20:13
<tabatkins>
Lines don't have children.
20:14
<tabatkins>
(At least, not in the element tree, which is where Selectors operate.)
20:14
<tabatkins>
Or... wait. I think I misparsed your sentence.
20:14
<tabatkins>
#container > :first-child::first-line should work
20:15
<tabatkins>
That'll select the ((first child of a container)'s first line), but not the (first child of (a container's first line)).
20:16
<tabatkins>
And then, of course, text-transform: uppercase;.
20:17
<tabatkins>
cying: ^^^
20:17
<tabatkins>
cying: Also, your request is a heisen-sentence. It can be read either way, and swapping the order of the "first *" clauses is equally ambiguous.
20:18
<cying>
hehe
20:18
<cying>
tabatkins: i mean the former not the latter
20:18
<cying>
((first child of a container)'s first line)
20:18
<tabatkins>
K, then my selector will work.
20:18
<cying>
sadly not working in mobile safari
20:18
<tabatkins>
That'll be a bug, then.
20:20
<tabatkins>
What's the failure? It doesn't work at all? It selects more than you think? Less than you think?
20:33
<cying>
tabatkins: it doesn't work at all
20:33
<cying>
tabatkins: it selects nothing
20:34
<annevk>
is the child a block?
20:36
<cpojer>
hey guys
20:36
<cpojer>
I have a question about window.onpopstate
20:37
<cpojer>
I am not clear about what the expected behavior is
20:37
<cpojer>
in chrome/firefox4 popstates fires for the initial page load
20:37
<cpojer>
in safari it doesn't
20:37
<cpojer>
I read this: http://lists.w3.org/Archives/Public/public-html-bugzilla/2010Nov/0078.html
20:38
<cpojer>
but I am not sure I understand it :D
20:38
<cpojer>
if anyone cares to enlighten me that would be great
20:38
<Ms2ger>
cpojer, so
20:40
<cpojer>
so.
20:40
<Ms2ger>
Say, you click a link
20:40
<Ms2ger>
http://www.whatwg.org/specs/web-apps/current-work/multipage/text-level-semantics.html#the-a-element
20:40
<Ms2ger>
Finally, the user agent must follow the hyperlink
20:40
<Ms2ger>
http://www.whatwg.org/specs/web-apps/current-work/multipage/links.html#following-hyperlinks
20:41
<cpojer>
wait what?
20:41
<Ms2ger>
... must navigate a browsing context to the resulting absolute URL.
20:41
<Ms2ger>
http://www.whatwg.org/specs/web-apps/current-work/multipage/history.html#navigate
20:42
<cpojer>
wait sorry, I don't really want to read through all the specs if thats ok
20:42
<cpojer>
just want to know what the expected behavior for popstate is
20:42
<cpojer>
should it fire on the initial page load or not
20:42
<cpojer>
that is, if I navigate to example.com from an empty tab
20:42
<cpojer>
should it fire popstate or not?
20:42
<cpojer>
chrome/ff4 do, safari does not
20:42
<cpojer>
any maybe because its friday night I am not fully able to understand that response in the bug tracker
20:48
<Ms2ger>
Yes, should fire
20:48
<cpojer>
ok thanks
21:52
<zcorpan>
http://www.d.umn.edu/~lcarlson/html5bugchart/20101002/ld/chart1.html is indeed a pretty long description
22:08
<Hixie>
zcorpan: i wonder why she didn't just put that on the main page, it's more useful than the pie charts!
22:08
<tabatkins>
Dammit, Hixie, I was just about to say that.
22:09
<tabatkins>
The information about trends and such is not obvious from the infographic. It's much easier to understand from the description.
22:14
<annevk>
I can't follow those tables, why is there no summary?
22:17
<Philip`>
The first table seems to have a typo - INVALID is 6% for 1, LATER is 2% for 5
22:17
<zcorpan>
does the typo in the table propagate to the graph?
22:18
<Philip`>
No
22:18
<karlcow>
http://okfn.org/projects/annotator/ - Open-Source Annotation Toolkit for Inline, Online Web Annotation
22:26
<cying>
annevk: it sure is
22:26
<cying>
annevk: (re: :first-child::first-line)
22:26
<tabatkins>
data:text/html,<style>#container > :first-child::first-line { color: red; }</style><div id=container>foo foo<p>foo bar<br>bar baz</p>
22:27
<tabatkins>
cying: Given ^^^, which works on desktop, it's clear that if Mobile Safari isn't doing it, it has a bug.
22:27
<cying>
ahhhh
22:27
<cying>
tabatkins: thanks