02:57
<annevk>
Hixie: nah man, rename all the things!
03:37
<Hixie>
jesus
03:37
<Hixie>
MPEG-4 is a seriously messed up format
03:46
<rniwa>
Hixie: any spec. to do with video and audio are pretty much messed up IMO.
03:46
<Hixie>
the spec is impenetrable
03:46
<Hixie>
as far as i can tell
03:47
<Hixie>
(if anyone here is familiar with MPEG-4 let me know, you could really help me out)
03:57
<MikeSmith>
Hixie: track stuff?
04:16
<rniwa>
Hixie: does HTML5 require named getter on document object?
04:17
<rniwa>
Hixie: i.e. named access; http://www.whatwg.org/specs/web-apps/current-work/multipage/browsers.html#named-access-on-the-window-object
04:17
<rniwa>
Hixie: it appears that webkit has some code to allow this on document but is not working.
04:17
<rniwa>
Hixie: it'll be nice if we can just get rid of altogether if we don't need them.
04:17
<rniwa>
get rid of them*
04:26
<Hixie>
MikeSmith: yeah
04:26
<MikeSmith>
Hixie: yeah saw the bug now
04:26
<Hixie>
rniwa: dunno, look at the spec :-)
04:27
<rniwa>
Hixie: i've looked at the spec but couldn't find it.
04:27
Hixie
looks
04:27
<Hixie>
yes
04:27
<Hixie>
it's right there in the idl
04:27
<Hixie>
getter object (DOMString name);
04:35
<rniwa>
Hixie: can I have a pointer?
04:35
<Hixie>
search for what i just pasted :-)
04:35
<Hixie>
it's in the Document idl block
04:35
<rniwa>
Hixie: ah, now i see it.
04:36
<rniwa>
Hixie: :( it's called named properties here
04:36
<Hixie>
?
04:36
<rniwa>
that's why it wasn't hitting my search for "getters"
04:36
<Hixie>
how do you mean?
04:36
<rniwa>
or rather "accessors"
04:37
<rniwa>
Hixie: i was looking for named getters & named accessors
04:37
<Hixie>
ah
04:37
<rniwa>
Hixie: it seems like there are a lot of synonymous for this thing
04:37
<Hixie>
well this is just webidl syntax
04:38
<zcorpan>
Hixie: wait, we can't use RESOLVED FIXED for fixed bugs now?
04:39
<Hixie>
trying to get clarity on this from the chairs
04:39
<Hixie>
i expect we'll clone the bugs
04:39
<Hixie>
so that they don't get lost
04:40
<zcorpan>
is the w3c spec frozen?
04:43
<Hixie>
not frozen, but it doesn't currently have an editor
04:44
<zcorpan>
you don't push updates to it anymore?
04:47
<Hixie>
no
04:47
<Hixie>
the chairs asked me to stop a while back
04:47
<zcorpan>
ok
04:47
<Hixie>
(i finally got around to updating my script recently, after going through the bulk of bug report e-mails)
04:48
<Hixie>
(so that i didn't have to find a way to get the e-mails into bug form)
04:49
<zcorpan>
i wonder how long the chairs will take to decide on the bugs thing, having fixed ones stay open seems annoying for me
04:50
<Hixie>
sounded like they were going to come to a decision pretty quickly
04:50
<zcorpan>
good
04:52
<Hixie>
oh sweet lord
04:52
<Hixie>
MPEG2 is EVEN WORSE
04:53
<Hixie>
also wtf is with these standards organisations that DON'T PUT THEIR STANDARDS ON THE WEB
04:53
<Hixie>
wtf is the point of a standard YOU CAN'T READ
04:54
<zcorpan>
is mpeg2 something we need to support?
04:54
<Hixie>
mpeg2 is what tv stations use, i believe
04:54
<Hixie>
so it's likely that most uses of the feature i'm speccing will use mpeg2
04:54
<Hixie>
note that i'm not convinced anyone will actually ever use this
04:56
<roc>
as I'm sure you know, selling copies of standards was (is) a common business model for standards bodies to fund themselves
04:56
<Hixie>
such standards bodies are idiots.
04:57
<Hixie>
or at least, seriously misguided
04:57
<Hixie>
and somewhat lacking in understanding as to how standards work
04:58
<Hixie>
(then again, i don't understand why standards bodies need funding. all you need is to pay for a web host and that's pretty cheap these days.)
04:58
<roc>
bureaucracy
04:58
<roc>
it costs money
04:58
<Hixie>
bureaucracy is not required to write a standard.
04:59
<Hixie>
in fact, it only detracts from it.
04:59
<roc>
I think you're preaching to the choir here
04:59
<Hixie>
yes. :-)
05:00
<Hixie>
just venting
05:09
<Hixie>
zcorpan: (oh, and canvas stopped updating even longer ago)
05:09
<MikeSmith>
Hixie: http://mumudvb.braice.net/mumudrupal/sites/default/files/iso13818-1.pdf
05:10
<Hixie>
ooh
05:10
<MikeSmith>
that's just part 1, I think
05:11
<MikeSmith>
apparently there are 7 or more parts to it
05:11
<Hixie>
apparently i need the "Systems" chapter
05:11
<Hixie>
(?)
05:12
<Hixie>
which that seems to be
05:12
<Hixie>
let me see if it has what i need
05:13
<Hixie>
oh i think i read this spec once
05:20
<zcorpan>
Hixie: but fixed canvas bugs are not left open?
05:20
<Hixie>
i don't think i've fixed any since then have i?
05:21
<zcorpan>
dunno
05:21
<Hixie>
it's always possible i've accidentally closed some, especially if it was back when i still thought we'd resolve the canvas issue before we forked
05:21
<Hixie>
i kinda assume they're going to take all the changes
05:22
<Hixie>
if they're not, they'll probably have bigger problems
05:22
<Hixie>
than a few missed bugs
05:29
<zcorpan>
WONTFIX resolutions are OK?
05:31
<Hixie>
probably not, but i told the chairs about that one
05:32
<Hixie>
it seemed pretty uncontroversial, other than you disagreeing with me :-)
05:33
<zcorpan>
yeah imma spend days ISSUEing and CPing over it
05:33
<Hixie>
:-)
05:35
<MikeSmith>
zcorpan: If you need any secret backroom deals later to get an extension, lemme know
05:35
<Hixie>
so THAT's how hober and the a11y people are getting all these extensions
05:36
<zcorpan>
MikeSmith: shhh!
05:36
<MikeSmith>
whoops yeah I guess it's not secret any more
05:36
<MikeSmith>
damn
05:37
<MikeSmith>
can always count on dude to provide the lulz
05:37
<MikeSmith>
at his own expense
05:40
<MikeSmith>
Hixie: oh I see you did manage to find the MPEG-2 info you needed
05:42
<MikeSmith>
hahaha "I have no idea what I'm doing here."
05:42
<MikeSmith>
somebody's going to quote you on that man
05:42
<MikeSmith>
out of context of course
05:42
<MikeSmith>
well, I will at least
05:44
<zcorpan>
i already tweeted it
05:49
<MikeSmith>
I'm saving it for a future CP
06:42
<MikeSmith>
am I missing some context for http://lists.w3.org/Archives/Public/public-html-a11y/2012Jun/0119.html ?
06:45
<webben>
MikeSmith: I'm guessing "lightweight JSON objects" are the Canvas regions that have ARIA roles but not associated elements.
06:46
<webben>
MikeSmith: http://www.whatwg.org/specs/web-apps/current-work/multipage/the-canvas-element.html#hit-region's-unbacked-region-description
06:47
<MikeSmith>
hey it's webben
06:47
<webben>
hey :)
06:47
<MikeSmith>
webben: thanks for the pointer
06:48
<webben>
yw
07:07
<Stevef>
mikesmith: it's the half baked part of the canvas additions where there is no agreement or information about how it will work and whether the regions will be focusable or not
07:12
<MikeSmith>
Stevef: I don't see what it has to do with JSON but OK
07:15
<Stevef>
MikeSmith: mentioned here at F2F http://krijnhoetmer.nl/irc-logs/html-wg/20120503#l-1083
07:16
<webben>
Yes. Strictly speaking they are effectively DOM properties of the <canvas> element, not "JSON".
07:16
<MikeSmith>
Stevef: ah, OK
07:17
<Stevef>
mikesmith: there is other discussion can't find at the mo'
07:18
<MikeSmith>
Stevef: I understand the context now
07:19
<MikeSmith>
just, using that inaccurate terminology to describe it doesn't help
07:34
<Stevef>
mikesmith: json idea originally floated by philip http://krijnhoetmer.nl/irc-logs/whatwg/20120223#l-117
07:40
<MikeSmith>
Stevef: I see
08:24
<wirepair>
man, CORS sure has made it easy to execute XSS against people who think passing parts of the URI to XHR is a Good Idea(tm)
08:25
<wirepair>
s/people/sites
08:28
<annevk>
passing parts of the URI directly is never a good idea
08:30
<wirepair>
of course not but CORS makes it much easier to get third party script code executed via an XHR call than without CORS support
08:31
<annevk>
zcorpan: retweeted from @whatwg
08:31
<annevk>
wirepair: yeah and new HTML features make it easier to bypass blacklist-based filtering
08:32
wirepair
shrugs
08:32
<wirepair>
if you're blacklisting you're doing it wrong :>
08:33
<annevk>
if you're reusing parts of the URL without checking you're doing it wrong too
08:33
<wirepair>
in that we agree as well :>
08:35
<zcorpan>
XDomainRequest doesn't have that problem
08:35
<zcorpan>
(at least so long as XHR didn't support CORS in ie)
08:36
<wirepair>
yeah prob is people forget that xhr can now load from 3rd party sites
08:36
<wirepair>
so all taht old code is suddenly vulnerable
08:39
<zcorpan>
yes, and when i first saw this problem being discussed, the proposed fix was *still* vulnerable because the guy who proposed the fix didn't know that urls like "//example.org" work
08:39
<zcorpan>
(the fix being "just prepend a slash and you'll be fine")
08:42
<wirepair>
fantastic. :)
08:42
<annevk>
wirepair: hmm yeah, I guess it's too late now :(
08:45
<wirepair>
well yeah, i mean it's how CORS works so not really sure what you could even do.
08:47
<annevk>
well for XHR we could have had a different way to get access to CORS
10:32
<zcorpan>
anyone have advice for this guy? (redirect and appcache) http://forums.whatwg.org/bb3/viewtopic.php?f=10&t=5022
11:15
<niloy>
zcorpan, move resources to cdn maybe?
11:18
<zcorpan>
niloy: feel free to reply in the forum :-)
11:19
<niloy>
zcorpan, sure :)
11:19
<zcorpan>
thanks
13:01
<gsnedders>
jgraham: given you were asking yesterday, https://groups.google.com/d/msg/mozilla.dev.tech.js-engine/yYQyMCcMf-0/c1kHxnhOUlYJ
13:03
<gsnedders>
(E4X in SM)
15:51
<Stevef>
no meeting this week
15:52
<Stevef>
sorry wrong window
15:52
<Ms2ger>
No meeting here either :):
15:52
Ms2ger
wonders why he has eyes on his chin
15:54
<jgraham>
Practising canibalism?
15:54
<Ms2ger>
Not usually
16:04
<dglazkov>
good morning, Whatwg!
17:54
<Ms2ger>
"When rendering li elements, user agents are expected to use the ordinal value of the li element to render the counter in the list item marker."
17:54
<Ms2ger>
"If the parent element is an ol element, then the li element has an ordinal value."
17:55
<Ms2ger>
If the parent element is *not* an ol element, what is rendered?
18:01
<Hixie>
if the parente lement is not an ol element, there's no counter in the list item marker
18:03
<Ms2ger>
<ol><div><li>A</div></ol>
18:05
<Ms2ger>
Should that have a number anywhere, and if so, where?
18:05
<TabAtkins>
Yes, before the A.
18:05
<Ms2ger>
Which number?
18:05
<TabAtkins>
The <li> generates a marker, and inherits a numeric type from <ol>
18:05
<TabAtkins>
1.
18:05
<Ms2ger>
Now, <ol><div><li>A</div> <li>B</ol>
18:05
<TabAtkins>
The counter from <ol> isn't distrubed by having the <div> come between it and the <li>.
18:06
<Ms2ger>
You say 1. A / 2. B?
18:06
<TabAtkins>
If my mental model of the parser is right, sure. If you give me the DOM produced, I can tell you for sure.
18:06
<gavinc>
Sometimes validator.nu is a bit... amusing The encoding "utf-8" is not the preferred name of the character encoding in use. The preferred name is utf-8
18:07
<Ms2ger>
Yeah, nothing special going on in the parser here
18:07
<Ms2ger>
Now, <ol reversed><div><li>A</div> <li>B</ol>
18:07
<TabAtkins>
The <ol> genreates a fresh counter scope. Each <li> increments it.
18:07
<TabAtkins>
Ah, fun fun. Depends on how you implement @reversed.
18:08
<Ms2ger>
Not with CSS counters, because I'm told those can't reverse
18:08
<TabAtkins>
They can, but you have to have a starting value.
18:08
<TabAtkins>
You can say "start at 5, and each counter adds -1".
18:09
<Ms2ger>
You can?
18:09
<Ms2ger>
Hi fantasai
18:09
<TabAtkins>
Yeah, the counter-increment property takes any integer.
18:09
<fantasai>
hi Ms2ger :)
18:10
<fantasai>
TabAtkins: where are you?
18:10
<TabAtkins>
In the google office.
18:10
<Ms2ger>
<TabAtkins_> fantasai: I'm in the google office. Where's the moz office, and how do I get in?
18:10
<fantasai>
TabAtkins: is our phone dead again :/
18:11
<fantasai>
s/our/your/
18:11
<TabAtkins>
fantasai: Nearly, so I've turned it off.
18:11
fantasai
sent a text with instructions half an hour ago
18:11
<fantasai>
~_~
18:11
<Ms2ger>
Hmm, I didn't know we were right on the waterfront
18:12
<TabAtkins>
fantasai: I was talking with Ojan and Tony when I got in anyway.
18:12
fantasai
didn't know they were in sf
18:12
<TabAtkins>
Yup!
18:12
<fantasai>
anyway!
18:12
<TabAtkins>
Do I need security or something to get up to floor 7?
18:12
<TabAtkins>
Or do you just want to come down and get me?
18:13
<Ms2ger>
TabAtkins, however, ol isn't specced in terms of CSS counters
18:13
<fantasai>
No, and we don't ask you to sign an NDA that you're legally incapable of signing, either :P
18:13
<TabAtkins>
kk, I'll be up in a sec then.
18:13
<fantasai>
Ms2ger: if it's specced using display: list-item, then it will be in L3
18:13
<fantasai>
Ms2ger: using an implied counter
18:14
<fantasai>
right, so I was going to ask a question ...
18:14
<fantasai>
hsivonen: Is it intentional that </ins> gets ignored when </p> is implicit?
18:14
<fantasai>
hsivonen: http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%0A%3Cins%3E%3Cp%3Einserted%3C%2Fins%3E%0A%0A%3Cp%3Enot%20inserted
18:15
<Ms2ger>
Well, it has display: list-item, but with an override for the counter
18:15
<[tm]>
gavinc: file a bug. our maybe there is actually an open bug on that one already
18:18
<fantasai>
Ms2ger: for 'start' and 'value', you mean?
18:18
<Ms2ger>
And reversed
18:18
<fantasai>
ah, right
18:18
<fantasai>
that one, we can't do in CSS yet
18:18
<Ms2ger>
Hah
18:18
<Ms2ger>
TabAtkins just claimed he can
18:19
<Ms2ger>
But if bz and fantasai can't do it... :)
18:19
<fantasai>
you can count backwards :)
18:19
<fantasai>
that's easy
18:19
<fantasai>
but you need to know what number to start at
18:19
<fantasai>
and that's not possible afaik
18:19
<TabAtkins>
fantasai: I'm in the lobby.
18:20
<Ms2ger>
TabAtkins (~tabatkins@nat/mozilla
18:20
<Ms2ger>
Welcome to the bright side, sir
18:21
<Ms2ger>
fantasai, well, you can count the
18:21
<Ms2ger>
Wrong button
18:21
<Ms2ger>
Another case
18:21
<Ms2ger>
<ol><div style=display:list-item>A</div><li>B</ol>
18:22
fantasai
notes that Ms2ger got both of us to lol
18:22
<Ms2ger>
Always glad to do that, m'lady
18:28
<Hixie>
Ms2ger: sorry, was afk
18:28
<Hixie>
Ms2ger: looks like the spec should say the ordinal value is 1 if there's no parent ol.
18:29
<Hixie>
Ms2ger: note that if the author uses css counters to override the default rendering, the numbers won't match, since css doesn't provide a way to get the ordinal value at the moment
18:29
<jgraham>
I have been told that webkit does something weird for @reversed like making the start value CSS dependant
18:30
<Ms2ger>
So, <ol><div><li>A</div><div><li>B</div></ol>
18:30
<Hixie>
it's not just reversed="" that causes problems, also display:none and so on
18:31
<Hixie>
bbiab
18:35
<pablof>
has there been any additional work on appcache's prefer-online since http://html5.org/tools/web-apps-tracker?from=7135&to=7136 ?
18:35
<pablof>
can't seem to find most of that text in the spec
18:35
<pablof>
(tangentially, is there a place to do a blame on the svn files somewhere online?)
18:37
<jgraham>
You want a web service that does something like 'def get_blame:return "Hixie"'?
18:38
<jgraham>
(yeah, I know svn blame does more than that :p)
18:39
<pablof>
i'm after the commit/diff, not the author, we all know who to blame :P
18:40
<jgraham>
I doubt anyone provides that
18:41
<jgraham>
I imagine it is slow to compute
18:41
<jgraham>
No one even provides a convenient git-svn version of the spec :(
18:43
<Ms2ger>
Hixie has per-1000-commit blames cached somewhere
18:43
<Ms2ger>
But I always forget where
20:03
<Hixie>
Ms2ger, pablof: http://www.whatwg.org/specs/web-apps/current-work/blame.r5000 is available, for various values of 5000
20:05
<Hixie>
blame.r0005 blame.r0954 blame.r2200 blame.r2800 blame.r3000 blame.r4667 blame.r5372 blame.r7000 blame.r0050 blame.r1000 blame.r2216 blame.r2905 blame.r3075 blame.r5000 blame.r5600 blame.r6400 blame.r0500 blame.r2000 blame.r2500 blame.r2947 blame.r4000 blame.r5287 blame.r6000 blame.r6666
20:05
<Hixie>
er let me sort that and try again
20:05
<Hixie>
blame.r0005 blame.r0050 blame.r0500 blame.r0954 blame.r1000 blame.r2000 blame.r2200 blame.r2216 blame.r2500 blame.r2800 blame.r2905 blame.r2947
20:05
<Hixie>
blame.r3000 blame.r3075 blame.r4000 blame.r4667 blame.r5000 blame.r5287 blame.r5372 blame.r5600 blame.r6000 blame.r6400 blame.r6666 blame.r7000
20:06
<Hixie>
(some of the more weird numbers like 2947 and 5287 correspond to diffs where a lot changed for editorial reasons so they come up often in archeology)
20:07
<Hixie>
feel free to document this on the wiki if you want btw
20:09
<pablof>
thanks
20:11
<pablof>
Hixie: were there any changes in the spec recently on appcache's prefer-online?
20:16
<Hixie>
pablof: well it was added to the spec...
20:16
<Hixie>
pablof: how recently do you mean?
20:17
<pablof>
Hixie: i mean that i can't seem to find the text added in http://html5.org/tools/web-apps-tracker?from=7135&to=7136
20:18
<pablof>
let me re-check in the one-page one, in case i'm dumb
20:19
<pablof>
Hixie: yeah, never mind, the parts i was looking for weren't in the application cache section, sorry about that
20:19
<pablof>
doh
20:37
<Hixie>
pablof: ah ok
21:30
<Hixie>
AryehGregor: you around? got some questions about drag and drop and contenteditable (in case you are and want to pretend you're not ;-) )
21:42
jgraham
would totally always pretend to be away rather than answer questions about contenteditable or drag and drop
21:44
<Hixie>
yeah me too
21:44
<Hixie>
hence me letting him know :-)
21:44
<jgraham>
Also, I wonder what would happen if web developers saw that the HTML spec specifically thanks the inventors of the drag and drop API. I expect it would involve pitch forks and buring torches
21:45
<jgraham>
Or failing that some mild sarcasm on twitter
21:45
<jgraham>
21st century revolution style
21:47
<Hixie>
lol
21:47
<Hixie>
dnd has been dramatically improved
21:47
<Hixie>
with draggable and dragzone
21:48
<Hixie>
it might even be possible to use it now
23:42
<TabAtkins>
I wonder if Chrome has added draggable and dragzone yet? I've got some dnd code that I wouldn't mind simplifying.