00:30
<Hixie>
does webidl define how callback return values get converted to idl values?
00:30
<Hixie>
i can't find it anywhere
02:25
<hober>
with 3/31 approaching, I suppose I should work on http://hober.jottit.com/ISSUE-95_Change_Proposal more
02:26
<hober>
any and all feedback greatly appreciated.
02:42
<boblet>
hixie: you there?
04:24
<Hixie>
borey: here
04:24
<Hixie>
er
04:25
<Hixie>
nevermind, boblet left
04:25
<Hixie>
sorry borey, you were next on the nick completion list :-)
07:00
<zcorpan>
Hixie: was EXAMPLE workers/shared/001/test.html supposed to be replaced in the generated spec?
07:27
<zcorpan>
"our html5 involvement goes back years and years and years" - http://ecn.channel9.msdn.com/o9/ch9/8/2/8/8/3/5/ch9livemix10dean_ch9.mp4
07:28
<Hixie>
zcorpan: was it not replaced? hm
07:28
<Hixie>
wonder what's going on with wikileaks
07:29
<JonathanNeal>
Hey zcorpan, Hixie
07:29
<Hixie>
oh, too many /s for the EXAMPLE
07:29
Hixie
fixes
07:31
<zcorpan>
"will ie9 support canvas?" "everything in ie9 is hardware accelerated!"
08:02
<JonathanNeal>
did you guys see the pangram?
08:05
<hsivonen>
jgraham, Hixie: I see the textarea resizer on Live DOM now. I guess the resizer appears by default but I was using the wrong build yesterday
08:06
<zcorpan>
Hixie: the second example seems to be 001/
08:11
<Hixie>
i suck
08:11
<Hixie>
fixing
08:11
<Hixie>
thanks
08:17
<nessy>
it's bad hair day
08:20
<JonathanNeal>
http://sandbox.thewikies.com/html5-pangram/ and http://sandbox.thewikies.com/html5-pangram/short.html
08:20
annevk
always has a bad hair morning; unless it's extremely short
08:22
<zcorpan>
JonathanNeal: you should put some ads on that page
08:23
<JonathanNeal>
So it can be even uglier?
08:23
<JonathanNeal>
I'm not done, the short will be how I write an html5 normalize stylesheet
08:23
<zcorpan>
no, to match expectations; i think you've posted the link about 20 times now :P
08:24
<JonathanNeal>
It first follows the whatwg's rendering recommendation, followed then by popular vendor styling.
08:24
<annevk>
JonathanNeal, most people in this channel read the backlog
08:24
<JonathanNeal>
Note taken.
08:24
<JonathanNeal>
:)
08:25
<JonathanNeal>
I didn't even know there was a backlog to be read. I thought the best strategy was to always stir up trouble.
08:25
<Hixie>
zcorpan: who implements the three-argument postMessage on Window?
08:26
<Hixie>
zcorpan: if nobody does, or if everyone is willing to change their implementations, i don't mind changing it
08:26
<zcorpan>
Hixie: opera does (not shipped yet, but is implemented), not sure about the others
08:26
<Hixie>
zcorpan: the reason postMessage()'s ports argument was added in the middle was to keep it closer to the message argument, where it makes more sense
08:26
<Hixie>
i.e. having "message, message, target" rather than "message, target, message"
08:27
<zcorpan>
Hixie: i think the common reaction to having it in the middle is "WTF?"
08:27
<Hixie>
if everyone is willing to change their implementations, i don't mind changing it
08:27
<Hixie>
no need to convince me on logical grounds :-)
08:28
zcorpan
can only speak for opera
08:33
<Hixie>
well if you can find out who else implements it i can go ask them
08:34
<annevk>
pretty sure WebKit must have it
08:34
<zcorpan>
Hixie: it seems chrome implements it
08:35
<Hixie>
k, i'll ask them about it tomorrow
08:35
<zcorpan>
but not firefox
08:36
<Hixie>
ooh, new freeman's mind
08:36
<Hixie>
bbiab!
08:43
<Hixie>
ok where was i
08:43
<zcorpan>
you were about to fix all my bugs
08:48
<Hixie>
oh yes
09:11
<zcorpan>
it's weird that people's reaction to "feature x and y seem like they should do what i want, but are buggy in browsers A and B" is "how about introducing feature z?" instead of "how about fixing x and y in A and B?"
09:14
<Hixie>
yeah
09:14
<Hixie>
i see that all the time also
09:20
<Hixie>
anyone else having difficulty using the bug filing thing in the spec?
09:27
<Hixie>
ok bed time
09:27
<Hixie>
nn
09:31
<zcorpan>
Hixie: nn
09:34
<gsnedders>
Anyone with IE9 around?
09:35
<Philip`>
Yes
09:35
<gsnedders>
Philip`: http://stuff.gsnedders.com/es/getterparsing.html
09:35
<gsnedders>
Philip`: What does that give?
09:38
<Philip`>
FAIL (threwEvalError: Expected ':')
09:38
<Philip`>
6 times
09:39
<gsnedders>
Huh
09:39
<gsnedders>
I thought IE supported it...
09:40
<lazni>
is opera's treatment of page-break-* in @media projection okay in standards?
09:41
<virtuelv>
lazni: to answer that question, you would have to first explain what it is you think Opera does with it
09:41
<annevk>
lazni, yes it is
09:41
<lazni>
it split advancement by pagedown by the pages
09:41
<annevk>
lazni, projection is paged-based media
09:42
<lazni>
okay
09:42
lazni
is writing a CSS+fragment slideshow
10:17
<hsivonen>
sigh. I just had a "Someone is wrong on the Internet" moment about http://robertnyman.com/2010/03/23/what-will-happen-to-open-video-on-the-web/
10:18
<hsivonen>
(I left a comment)
10:24
<annevk>
hmm, from that post (via #fronteers): "My friend spoke to a Mozilla guy at Libre Planet this last weekend who said Mozilla has a “big announcement” regarding video coming this week or next."
10:35
<jgraham>
Hmm, "my friend said someone said something was going to happen" doesn't exactly sound concrete
10:35
<Dashiva>
hsivonen: It's disappointing to see people still think using Vorbis has any weight at all on how safe Theora is
10:37
<hsivonen>
Dashiva: I didn't even read that far carefully before I went into the "Someone is wrong on the Internet" mode.
10:39
<zcorpan>
Dashiva: leave a comment?
10:40
<Dashiva>
People will still be wrong
10:41
<zcorpan>
but you know you get a warm fuzzy feeling when pointing it out
10:43
<Philip`>
If the Ogg people said Vorbis was perfectly safe and patent-free and encouraged everyone to use it, and then a few years later everybody got sued for infringing patents, then that would cast significant doubt over similar claims about Theora
10:43
<Philip`>
but everyone uses Vorbis and nobody has been sued yet, so that particular source of doubt has not been cast
10:43
<Philip`>
so it does seem to have some weight on the safety of Theora
10:48
<Dashiva>
In the same way having built a shed does give some weight to the guy who wants to build your skyscraper
10:51
<Philip`>
Sure
10:51
<jgraham>
For large values of "shed"
12:58
<lazni>
is there any CSS way I can make the nav#toc xx-large when there's no fragment id, but small when the fragment id is not "toc" in http://hdh.dyn-o-saur.com/public-domain/doc/sd1/goal-setting.html ?
13:02
<Rik`>
:target ?
13:12
<lazni>
:not(:target) makes it large when another id is in focus too
13:13
<AryehGregor>
No, pretty sure there's no way to do that. Use JS.
13:18
<hsivonen>
aaargh. I've broken validator.nu buildability with javac
13:19
<hsivonen>
ecj still works
13:52
<zcorpan>
hmm, maybe i should look at the entity data
13:54
<MikeSmith>
hsivonen: how'd you manage to break buildability with javac?
13:58
<zcorpan>
Philip`: have you tried http://krijnhoetmer.nl/irc-logs/whatwg/20100311#l-738 ?
14:01
<Philip`>
zcorpan: I think I forgot
14:07
<hsivonen>
MikeSmith: I changed NamedCharacters.java to use byte[][] for names
14:07
<hsivonen>
MikeSmith: ecj optimized the static initializer to under 2^16 bytes
14:07
<hsivonen>
MikeSmith: javac didn't
14:07
<MikeSmith>
ah
14:08
<hsivonen>
so javac failed to produce a legal .class file
14:14
<othermaciej>
howdy everyone
14:15
<annevk>
public-html :/
14:15
<annevk>
besides that pretty good
14:15
<annevk>
just had my first beer in the open, weather is changing for the better :)
14:16
<Philip`>
zcorpan: That regexp looks scary so maybe http://philip.html5.org/data/meta-content-charset-no-http-equiv-content-type.txt is better
14:16
<Philip`>
zcorpan: That's from egrep -i 'content="[^;"]*;\s*charset='|grep -iv 'http-equiv="content-type"'
14:16
jgraham
wonders who annevk is talking to
14:17
<Philip`>
on a file containing <meta uri http-equiv name content> extracted from parsed pages
14:17
<annevk>
jgraham, othermaciej?
14:17
hsivonen
wonders if WebKit and Opera have set themselves up for Microsoft moments like https://bugzilla.mozilla.org/show_bug.cgi?id=546857#c22
14:18
hsivonen
has no idea what kind of maintenance burdens Dashboard and Opera Widgets are
14:19
<jgraham>
annevk: Oh. I'm not really sure that follows from "Howdy everyone"
14:19
<annevk>
jgraham, it just did...
14:20
<othermaciej>
heh :-)
14:20
<annevk>
hsivonen, I suppose they're somewhat closer to HTML/CSS than XUL, other than that I have no idea
14:21
<zcorpan>
Philip`: nice, thanks
14:22
<AryehGregor>
"CSSStyleDeclaration" hurts my eyes. Too many S's.
14:23
<annevk>
should have been CSSDeclarationBlock but it's too late for that now
14:23
<AryehGregor>
Acronyms in names should be capitalized like normal words. That's the only sane way to do it.
14:23
<annevk>
much stuff in the CSSOM has weird names
14:23
<AryehGregor>
The convention that gives you CSSStyleDeclaration would also give you XMLHTTPRequest, which is just evil.
14:23
<AryehGregor>
CssStyleDeclaration would be so much more sensible.
14:23
<annevk>
euhm, now it is XMLHttpRequest
14:24
<annevk>
which is also weird
14:24
<AryehGregor>
Well, that's even more insane.
14:24
<zcorpan>
Philip`: there were no other attributes, like httpequiv? or http-equiv=contenttype?
14:24
<annevk>
yeah, not sure what you're saying
14:24
<hsivonen>
AryehGregor: more of a .NET than Java naming fan, eh?
14:24
<AryehGregor>
I don't touch either.
14:25
<AryehGregor>
annevk, I would do things like XmlHttpRequest.
14:25
<AryehGregor>
That way you can tell where words begin and end.
14:26
<annevk>
not really interesting to argue about such conventions now
14:26
<AryehGregor>
No, not at all, I was just randomly spamming.
14:26
<hsivonen>
I've used the .NET style in Java to improve readability
14:26
<hsivonen>
one of the main benefits .NET has over Java :-)
14:26
<AryehGregor>
Well, I guess it could be remembered for the future, if anyone gets to set new conventions for new languages or whatnot.
14:26
<AryehGregor>
hsivonen, so what *are* the .NET and Java styles?
14:27
<hsivonen>
CssStyleDeclaration would be .NET and CSSStyleDeclaration Java
14:27
<Philip`>
zcorpan: This data is just <meta uri http-equiv name content>
14:27
<Philip`>
i.e. it ignores any other attributes that may exist
14:27
<zcorpan>
Philip`: ok
14:28
<AryehGregor>
Then I'm a .NET convention fan. One thing Microsoft did right. :P
14:28
<zcorpan>
Philip`: i was kinda also interested in any other attributes that may exist :)
14:29
<Philip`>
zcorpan: I think I'd have to modify the Java code to do that properly
14:29
<Philip`>
which is possible
14:29
<Philip`>
(later, if I don't forget)
14:30
<Philip`>
The regexp wouldn't match anything across multiple lines so that's probably not so good
14:43
<hsivonen>
grr. maybe I should just ship a .class file compiled using Eclipse and stop supporting compilation using javac
14:45
<hsivonen>
alternatively, I could split NamedCharacter.java so that each member is in a class alone
14:46
<othermaciej>
is the file too big to compile with javac?
14:46
<hsivonen>
othermaciej: the output from javac is too big
14:46
<hsivonen>
to fit within the spec for .class files
14:47
<hsivonen>
the corresponding C++ file has been a problem with GCC and MSVC, too
14:48
<annevk>
hmm, CSS Value API is somewhat hard
14:49
<Philip`>
Making it harder to edit and recompile the code with standard tools doesn't sound like a good idea from a third-party contributor perspective
14:51
<hsivonen>
Philip`: Eclipse is the standard tool, isn't it. :-)
14:51
<hsivonen>
but yeah, I guess I need to split this file
14:51
<hsivonen>
and do the same split on the C++ side
14:53
<hsivonen>
this isn't the first time that javac produces more verbose bytecode than ecj
14:54
<Philip`>
Does javac attempt any kind of optimisation at all?
14:54
<hsivonen>
dunno
14:54
Philip`
briefly looked at its output once and it didn't seem to do any of the obvious things that would save space and time
14:55
<Philip`>
but I may have been mistaken
14:56
<JonathanNeal>
word
14:57
<Philip`>
JonathanNeal: excel
14:57
<JonathanNeal>
:)
14:57
<Philip`>
Wasn't that the pattern?
14:58
<JonathanNeal>
After one entry? Very presumptuous Philip`
15:00
<Philip`>
You get more points for spotting the pattern sooner
15:00
<Dashiva>
More like assuming
15:00
<Philip`>
I was hoping "sharepoint workspace" would come next
15:04
<JonathanNeal>
Unfortunately, the next entry was to be "sentence"
15:04
<zcorpan>
hmm, opera doesn't do incremental rendering of giant text nodes in xml
15:05
<jgraham>
Microsoft Sentence?
15:05
annevk
asks for thoughts on www-style
15:15
<MikeSmith>
hsivonen: fwiw, I don't think adding a dependency on ecj would be too unreasonable
15:15
<MikeSmith>
if that's what you were to decide to do
15:15
<MikeSmith>
it's cross-platform and easy to install
15:17
<zcorpan>
http://www.w3.org/Bugs/Public/show_bug.cgi?id=9207 - anyone feeling like analyzing the individual characters to see if they'd benefit from changing the parser?
16:16
<JonathanNeal>
Excell
16:20
<Philip`>
Where did the ent go?
16:20
<jgraham>
That sounds like the start of a Tolkien joke
16:48
<JonathanNeal>
Philip`, I don't know.
17:18
<CrazyFoam>
Hello, I have a question regarding XML Namespaces
17:18
<CrazyFoam>
say I have <a xmlns="urn:a" xmlns:b="urn:b"> <b:b> <c/> </b:b> </a>
17:18
<CrazyFoam>
element <c/> is assumed to be in namespace urn:a
17:19
<CrazyFoam>
but if I have <a:a xmlns:a="urn:a" xmlns:b="urn:b"> <b:b> <c/> </b:b> </a:a>
17:19
<CrazyFoam>
is element <c/> invalid or assumed to be in namespace urn:b?
17:20
<Philip`>
In the second case, the <c/> is in no namespace
17:20
<Philip`>
(which is a perfectly legitimate thing to do)
17:20
<CrazyFoam>
ah that's what I thought, but how is it legitimate
17:20
<CrazyFoam>
say the doc has no default namespace
17:20
<Philip`>
XML doesn't require namespaces
17:20
<CrazyFoam>
whatever processing the doc may not be able to understand <c/>
17:21
<Philip`>
You could write <a><b><c/></b></a> and the <c/> would be in no namespace, just as in your example
17:21
<Philip`>
Namespaces don't cause understanding
17:22
<annevk>
they just cause confusion
17:22
<CrazyFoam>
haha
17:22
<Philip`>
I could write <a:axghrlktxjhlkj xmlns:a="urn:a"/> and whatever's processing the document would have just as much chance of understanding it as of understanding <c/>
17:22
<Dashiva>
You could use the namespace URI as input to a RDF-based knowledge network that would provide all the data you need!
17:23
<Philip`>
You'd want some kind of schema if you want to restrict the document to known elements
17:23
<CrazyFoam>
knowledge network haha
17:24
<CrazyFoam>
okay well thanks for the info, really what i was looking for is that <c/> in the last case defeintiely shouldn't get interptreted as belonging to either urn:a or urn:b
17:25
<Philip`>
Yep - if it has no prefix then it's in the default namespace, and the default namespace is only affected by xmlns="..."
17:27
<CrazyFoam>
on that note, what do you call it when a doc changes default namespace all over the place, e.g. <doc xmlns="urn:doc"> <a xmlns="urn:a"/> <b xmlns="urn:b"/> </doc>, still valid right? can't think of a term to search for specifications - "default namespace switching"?
17:36
<Philip`>
CrazyFoam: You mean when it changes the default namespace in non-root elements?
17:37
<Philip`>
I guess it's neurotic by the definitions in http://lists.xml.org/archives/xml-dev/200204/msg00170.html
17:40
<CrazyFoam>
hmmm... is that post discussing maintaining prefix names through serialization?
17:41
<Philip`>
That's part of it
17:41
<CrazyFoam>
ah got to the neruotic part
17:41
<CrazyFoam>
what is a GI?
17:43
<Philip`>
Tag name
17:43
<Philip`>
('Generic identifier')
17:47
<lazni>
http://www.mikematas.com/ is rather empty in opera
17:49
<brucel>
Is there a webkit guru in the house?
17:49
<Philip`>
There might be more in #webkit
17:49
<jgraham>
brucel: I suggest you just ask the question and someone will probably answer
17:50
<jgraham>
lazni: Which version of OPera, and what exactly do you see?
17:51
<lazni>
10.51, I see only the 3 menu items on the left
17:52
<jgraham>
lazni: Curious. It works for me, although it is a bit slow to load (and has an obnoxious alert telling me to get a modern browser "such as Apple Safari")
17:52
<brucel>
Am trying to build up video demo that, if Ogg video isn;t supported (but video *is*, eg Safari) it will fall back to the contents of the video element. But can't get it working - onerror event seems not to fire. http://people.opera.com/brucel/dev/html5-video-fallback.html Have a brain full of cold so am trying to understand where I'm going wrong
17:53
<lazni>
I restarted opera and the page loads now
17:53
<jgraham>
lazni: Could be cache bug or something I guess
17:53
<lazni>
but it doesn't pop up about safari
17:54
<jgraham>
If you work out how to reproduce then file a bug
17:54
<annevk>
brucel, did you check that error is dispatched at all?
17:54
<lazni>
okay
17:55
<brucel>
annevk, seems never to fire
17:55
<jgraham>
lazni: You're not missing much. There's nothing that says "retard" quite like being a supposed UI designer who uses a modal alert to tell people that a site might not work in a browser that it actually works fine in and that they should buy a Mac instead
17:56
<lazni>
and a totallly JS-driven site
17:57
<annevk>
brucel, fwiw, it should be dispatched on the video element
17:57
<annevk>
brucel, also, <video> does not have a type attribute
17:58
<brucel>
annevk whoops, will whisk that away and see how it goes
17:59
<jgraham>
lazni: Well yeah it would be nice if the site fell back to something sensible. I have to admit that it looks pretty cool though
17:59
<annevk>
(if it then works that would also be a bug, but I don't think it'll help)
17:59
<annevk>
brucel, if you add autoplay does it then fire maybe?
18:01
<annevk>
brucel, you can maybe hack around it by checking the associated error object on some kind of timeout loop
18:01
<annevk>
maybe they do properly update that
18:03
<brucel>
annevk didn't work when I removed type from video element and added autoplay
18:03
<brucel>
annevk hack, thanks, will try. But this *should* work, right?
18:08
<annevk>
yeah
18:08
<annevk>
you might wanna file a bug: https://bugs.webkit.org/
18:14
<brucel>
annevk thx
19:41
<mr_danie1>
everyon is so excited about html5, its great new features; also many times html5 is seen as a flash-killer, making hopefully anothe propritery technology obsolete, and in many aspect I think html5 will make the race with widely accepted standards
19:42
<mr_danie1>
but what is about p2p capabilites? for example chatroulette.com is completely build on flash, because it uses the p2p technology from flash to make direct end-user communication possible
19:43
<annevk>
it's work in progress
19:43
<mr_danie1>
so I am asking myself: is there a way to build p2p webapplications without flash?
19:43
<mr_danie1>
is there a working draft document available annevk ?
19:44
<annevk>
http://www.whatwg.org/specs/web-apps/current-work/multipage/commands.html#peer-to-peer-connections
20:19
<zcorpan>
hmm, why doesn't the first video autoplay in opera? http://people.opera.com/brucel/dev/html5-video-fallback.html
20:23
<zcorpan>
annevk: "<annevk> brucel, fwiw, it should be dispatched on the video element" - no, it should fire on <source>
20:25
<zcorpan>
(at least when using <source>)
20:32
zcorpan
sees a your momma joke about html5 on twitter
20:39
<erlehmann>
this onerror thing is nice
20:39
<erlehmann>
zcorpan, link ?
20:39
<zcorpan>
http://twitter.com/iamkory/statuses/11051962698
20:44
<annevk>
zcorpan, ah yeah, I see now, but it'll bubble anyway so that shouldn't matter
20:45
gsnedders
opens up laptop and finds a listing of x86 jump instructions
20:45
<annevk>
the event summary makes this somewhat confusing
20:45
<gsnedders>
Oh dear.
20:46
<zcorpan>
annevk: error doesn't bubble
20:47
<annevk>
meh
20:47
<annevk>
seems cumbersome to put onerror on each source element
20:47
<annevk>
oh well
20:47
<zcorpan>
why would you put it on each source element?
20:48
<annevk>
debugging :)
20:48
<zcorpan>
you could use a capturing listener for debugging purposes :)
20:50
<annevk>
if i put the script in the right place, yeah...
20:51
<Hixie>
zcorpan: i think a "your momma so fat" joke would work better with the html spec :-)
20:51
<Hixie>
"your momma so fat, she makes the html spec look small"
20:52
<Dashiva>
More like "she got the HTML5 spec as a tattoo"
20:52
<zcorpan>
lol
20:53
<Hixie>
lol
20:54
<annevk>
hehe
21:09
<mr_danie1>
back, thanks for the link annevk
21:20
<zcorpan>
content-aware fill looks nice
21:25
<paul_irish>
apparently its been available in GIMP for years as a plugin. looks almost better, actually: http://o3.tumblr.com/post/470608946/photoshops-caf-content-aware-fill-unbelievable
21:26
<Hixie>
"almost better" so "not quite as good"? :-)
21:27
<paul_irish>
heh. i think it's a little nicer, in fact. I perfered the resynthesizer's desert landscape job vs adobe's.
21:29
zcorpan
wasn't aware of GIMP-resynthesize
21:29
<zcorpan>
paul_irish: yes, the desert looks better
21:30
<zcorpan>
the sky in the first picture looks worse, but might be easily fixable
22:21
<Dashiva>
So RDFa 1.1 keeps xmlns unchanged, but also adds two additional ways to declare prefixes
22:22
<Dashiva>
And this is supposed to make it less complicated?
22:24
<Philip`>
Perhaps you could then define a profile of RDFa 1.1 that only includes one or two of the prefix declaration methods, to keep things simple
22:26
<gsnedders>
Anyone have any clue about ARM based computers? Like, what is decent and not overly expensive?
22:32
<jgraham>
gsnedders: Huh? Are there many ARM-based general purpose computers for sale?
22:32
<jgraham>
(or: what is your use case?)
22:33
<gsnedders>
jgraham: There are a few
22:33
<Philip`>
Does an iPhone count?
23:08
<Bolkonskij>
hi
23:52
<zcorpan>
reading the opera desktop team comments, it seems people don't understand what RC means
23:57
<gsnedders>
zcorpan: or "pre-pre-alpha" for that matter.