00:12
<gsnedders>
mpilgrim: I guess the best answer is to see if that still exists once I finish and commit and push the changes to bring html5lib up-to-date with the spec
00:25
<Dashiva>
Rik`_: That's a failure of the outliner, technically
00:25
<Rik`_>
Dashiva: oh ?
00:26
<Dashiva>
The spec recommends useful defaults for stuff like <nav>
00:26
<Dashiva>
Since the element already specifies the semantics of the content
00:27
<gsnedders>
Dashiva: I don't want to implement any heuristics beyond what the spec requires in the outliner though
00:28
<Dashiva>
gsnedders: "User agents should provide default headings for sections that do not have explicit section headings."
00:28
<Dashiva>
And then goes on to used "Navigation" in the example
00:28
<Dashiva>
-d
00:28
<zcorpan>
"Untitled section" is a default heading
00:29
<gsnedders>
Dashiva: I'm well aware of what the spec says :)
00:29
<gsnedders>
Dashiva: And as zcorpan says, there is a default :)
00:29
<Dashiva>
gsnedders: Well, you're going ... the opposite of beyond ... what the spec does :P
00:31
<Dashiva>
It's not like it's more than a element name -> section name dict
00:31
<cardona507>
TabAtkins - did you have any slides or a url from your css gradient talk?
00:34
<Hixie>
ok between lachy yesterday and anne choking himself just now, y'all are just _trying_ to take yourselves out
00:34
<Hixie>
wtf
00:35
<gsnedders>
I'm still alive!
00:35
annevk2
was reading http://twitter.com/sgalineau/status/5493563264
00:35
gsnedders
facepalms
00:35
<cardona507>
haha
00:35
<gsnedders>
(But has a SMIL come to his face too)
00:36
Hixie
puts his drink down before reading that msg
00:36
<Hixie>
heh
01:09
<Dashiva>
gsnedders: Do you know enough Norwegian to construct that pun on your own?
01:10
<gsnedders>
But maNo.
01:11
<gsnedders>
*no
03:10
<gantrixx>
I'm curious to know the latest on the HTML5 <video> tag
03:10
<gantrixx>
this is desperately needed
03:11
<gantrixx>
I'm tired of trying to get video to work using conflicting plugins
04:03
TabAtkins
is on the plane home from TPAC.
10:55
zcorpan_
wonders what triple-equality checking is
10:59
<zcorpan_>
i understand that anne means ===, but first it sounded like checking == three times
11:04
<Philip`>
It's possible a cosmic ray will invert the output of ==
11:04
<Philip`>
so you need to check it several times if you want to be really really sure
11:32
gsnedders
realies he left headphones in the office. Gah.
12:00
zcorpan_
looks at a bug for <table><form> and sees that it has a number of dups
12:00
<zcorpan_>
clearly what opera does is not good enough for web compat
12:07
gsnedders
remembers looking at those bugs
15:30
<Dashiva>
http://www.flickr.com/photos/25541021%40N00/3706760751/in/set-72157620803945238/
15:30
<Dashiva>
That's a lot of empty accounts
15:47
<erlehmann>
ohmigawd
15:47
<erlehmann>
the german parliamentary tv guy wrote back
15:47
<erlehmann>
ant asked stuff like „can the video element do rtmp“
15:48
<erlehmann>
and attributed HTML5 solely to the W3C
15:48
<erlehmann>
*double facepalm*
15:54
<jgraham>
Dashiva: Looking at that twitter stream, I am impressed by the ratio of design goodness to insightfulness
15:54
<jgraham>
Because there is a lot of the former and almost none of the latter
15:55
<jgraham>
e.g. almost the flight fatality data he has deals with absolute numbers rather than per journey
15:55
<jgraham>
Which makes it useless for understanding pretty much anything interesting
16:03
<Philip`>
Dashiva: Is there really a precise 50% male/female split on Twitter, or did he just make that up?
16:08
<Dashiva>
:iiam:
19:44
gsnedders
opens up text editor to try and fight html5lib into matching the spec
19:49
gsnedders
reaches "in body" and sighs
19:58
<jgraham>
gsnedders: Up. Tommorrow. Yes
19:58
<gsnedders>
jgraham: So num += 2?
19:58
<gsnedders>
or only 1?
19:59
<jgraham>
yes
19:59
<jgraham>
2
20:22
<Dashiva>
That canvas focus box thing made me think about flash
20:22
<Dashiva>
In many cases it's possible to use tab to focus the active areas
20:23
<Dashiva>
But in some cases that actually ruins the flash, games where you aren't supposed to know where to click.
20:23
<Dashiva>
I wonder if it's possible to play such a game using AT
21:32
<Philip`>
"assuming the people of the future are at least halfway sane"
21:32
<Philip`>
That's a dangerous assumption
21:32
<Philip`>
since I think many of the people of the past and present aren't
21:36
<gsnedders>
Such as Philip`
21:48
<jgraham>
Philip`: Maybe the futre here means "a far future in which genetic engineering has boosted average intelligence to remarkable levels"
22:02
<gsnedders>
Weee, 1500 test failures in html5lib
22:02
<gsnedders>
I guess that's what I get for having a whole insertion mode not implemented at all
22:05
<Trede>
Do i interpret the draft correctly that a websocket server can return multiple "websocket-protocol" header names only when a client did not specify any protocol ?.
22:11
<gsnedders>
Why on earth do we implement the scripting enabled case in html5lib anyway?
22:14
<Trede>
its quite sad that a draft is allowed to be written so poorly such as the websocket one by ian is. its hardly on the level you expect from something on ietf.org
22:17
<gsnedders>
zcorpan_: How kind do you feel at the moment?
22:22
<zcorpan_>
gsnedders: hmm?
22:29
<gsnedders>
zcorpan_: You know how you wrote all those TCs for <script>? Could you go through all the other test inputs in the tree construction tests that contain <script and update them? :)
22:29
<gsnedders>
I'll buy you lunch if you do. :)
22:33
<Philip`>
jgraham: I'm not sure intelligence has much link to sanity
22:34
<Philip`>
Otherwise there wouldn't be so many mad professors
22:35
<Philip`>
and TV has taught me that there's a lot of them
22:36
<gsnedders>
<!doctype html></html> <head> should result in a text node " " under body, right?
22:38
<zcorpan_>
gsnedders: wasn't the spec changed to ignore end tags there?
22:39
<gsnedders>
zcorpan_: If I read the spec right (unlikely) then no
22:40
<zcorpan_>
hmm bash says cvs is not a command :(
22:40
<gsnedders>
'An end tag whose tag name is one of: "head", "body", "html", "br"' takes you through all the steps to create the doc, then you end up in after-after-body
22:41
zcorpan_
tries smartcvs
22:41
<zcorpan_>
but now i don't remember what my passphrase is
22:43
gsnedders
is amazed at the number of bugs in his tokenizer update code
22:44
<TabAtkins>
Philip`: It's a better assumption than that the future people will be completely sane.
22:45
<TabAtkins>
And, I think, is also a better assumption than that the future people will be stupider than us.
22:45
<gsnedders>
jgraham: 15:15 just inside the cinema tomorrow.
22:46
<Trede>
like humanity ever improves /
22:47
<gsnedders>
zcorpan_: Want to see Up in English tomorrow?
22:48
<zcorpan_>
gsnedders: maybe
22:49
<gsnedders>
zcorpan_: Just tell me yes/no sometime before I go and buy the tickets tomorrow morning :)
22:53
Philip`
thought Up was surprisingly sad
22:55
<zcorpan_>
oh now i remembered my passphrase
23:47
Philip`
never knew that commenting on a spec could be so fraught with difficulties
23:48
<Philip`>
Looks like there's been more comments commenting on the process of commenting on the MathML spec, than there were comments on the spec
23:51
<AryehGregor>
Hmm, it seems like someone has checked in RDFa support to MediaWiki.
23:51
<Philip`>
What does "support" mean?
23:52
<AryehGregor>
RDFa attributes on elements seem to be whitelisted in user input. Although I haven't looked very hard: http://www.mediawiki.org/wiki/Special:Code/MediaWiki/58712
23:52
<AryehGregor>
I'm not sure this actually allows valid RDFa to be output, since it's not namespaced or anything?
23:53
<gsnedders>
Valid RDFa? It relies upon namespaces, how can it be valid when it relies upon a DTD? :)
23:53
<AryehGregor>
Eh?
23:53
<AryehGregor>
Conformant RDFa, if you like. I didn't mean valid in the XML sense.
23:54
<gsnedders>
I know what you mean, but I'm just being silly about the whole problem with validity and RDFa :)
23:55
<Philip`>
"RDFa properties allow URIs" - um, no they don't
23:56
<Philip`>
(except about and resource)
23:56
<Philip`>
(which don't allow only URIs)
23:58
<Philip`>
Also: using a regep to detect 'unsafe' attribute values that might run scripts?
23:58
<Philip`>
*regexp
23:58
<AryehGregor>
Yes, I commented on that separately.
23:59
<gsnedders>
html5lib's sanitizer does that too
23:59
<AryehGregor>
MediaWiki already has a whitelist of safe URL protocols, that should be used if anything is.
23:59
<AryehGregor>
Regex is fine, just it needs to be whitelist-based, not blacklist.
23:59
<AryehGregor>
That's assuming any of these URLs can actually be activated by the user, though, which I doubt.