04:48
<MikeSmith>
hsivonen: three of the W3C validator.nu instances crashed today at almost the same time with segfaults
04:52
<MikeSmith>
hsivonen: looking at the log output, I have no idea why. The last stack trace in the logs is for a SAXStreamer.fatal error from TreeBuilder.startTag(TreeBuilder.java:2859) > TreeBuilder.pushHeadPointerOntoStack(TreeBuilder.java:4574) in the htmlparser code
04:54
<MikeSmith>
but that's a noraml error that appears to occur pretty often in the logs anyway, without any segfault happening
05:10
<MikeSmith>
hsivonen: hmm all of them show that somebody was trying to use the /html5 UI for some reason
05:11
<MikeSmith>
which is not linked to from the main UI nor anywhere, so dunno how they'd even know it was there. I guess I should just disable that but I don't know how
07:15
<stevefaulkner>
MikeSmith: email to webkit dev list from James Craig http://lists.webkit.org/pipermail/webkit-dev/2013-January/023371.html :-)
08:02
<hsivonen>
MikeSmith: seems weird for SAXStreamer to be used in ways that’d crash
08:43
<MikeSmith>
Stevef: yep, saw that. Will get my test fixed and update the patch
08:45
<MikeSmith>
hsivonen: yeah, pretty sure that had nothing to do with it
08:45
<MikeSmith>
nothing else at all in the logs that would provide a clue about what the cause was, though
09:31
<Stevef>
TabAtkins: didn't think I was putting words in your mouth, I qualified the statement with "appears to think"
09:33
<Stevef>
TabAtknins: on mobile/tablets not an issue, but on desktop...
09:34
<odinho>
Hmm. Whatever happened to the tests-on-GitHub thing?
09:35
<MikeSmith>
odinho: going as planned, far as I know
09:35
<MikeSmith>
some reason to think otherwise?
09:35
<odinho>
MikeSmith: Okay, I thought it was done already? :P
09:36
<odinho>
MikeSmith: I can't find it on github.com/w3c
09:36
<MikeSmith>
eh?
09:36
<MikeSmith>
html-testsuite, right?
09:36
<odinho>
Hmmmm. But but, I'm interested in webapps and webappsec :)
09:36
<MikeSmith>
aah
09:36
<odinho>
Yeah, -- I remember asking about that.
09:36
<MikeSmith>
well
09:37
<odinho>
And "blabla decision needed". So guess I have to do some actual admin work there :P
09:37
<MikeSmith>
odinho: bug Art
09:37
<odinho>
MikeSmith: Okay :]
09:38
<MikeSmith>
Art is great on getting decisions made and stuff done
09:39
<odinho>
MikeSmith: I've seen that :-) So I'll just wait till the end of the day when I guess he'll come on.
09:40
<MikeSmith>
yeah
09:42
<MikeSmith>
Stevef: in the bug, Niwa-san is asking that I put my name in the name field in the changelogs for the patch. So I'll do that but put "Authored by Steve Faulkner." at the top of the body of the changelog.
09:42
<MikeSmith>
if that's OK with you
09:43
<Stevef>
MikeSmith: yeah no problem, the importnat thing is it gets implemented :-)
09:43
<Stevef>
and you did all the hard work getting the tests done, thanks!
09:44
<MikeSmith>
ain't done yet
09:44
<MikeSmith>
well mostly done
09:44
<MikeSmith>
but then everything's always mostly done
09:45
<Stevef>
well the patch itself took a few hours and that is only because I am a noob
09:45
<Stevef>
a few lines of code
12:50
<MikeSmith>
Stevef: question just out of curiosity, one of the existing tests I'm looking at has a comment, "When a header and footer are inside an article, they should not be landmarks."
12:50
<MikeSmith>
and the test has an assertion for checking that
12:50
<MikeSmith>
why is that the expected behavior?
12:51
<annevk>
landmarks are document/application global
12:52
<annevk>
iirc
12:53
<hsivonen>
hmm. so much wrong with the pro-Polyglot rationale
12:53
<hsivonen>
probably not worth the 386, though
12:55
<MikeSmith>
annevk: I see
12:55
hsivonen
is slightly surprised to see “simple majority wins” as opposed to “chairs decide” polls at the HTML WG
12:56
<MikeSmith>
hsivonen: me too actually
12:57
<MikeSmith>
they seem to be making a distinction between process issues and technical issues
12:57
<hsivonen>
yeah
12:59
<Stevef>
MikeSmith: thats how webkit have implemented header/footer as their use in HTML is not the same as in ARIA, having lots of headers and footers in pages makes noise if they are all landmarks
13:00
<MikeSmith>
Stevef: you mean that behavior is not required by the ARIA spec?
13:01
<Stevef>
Its not required by HTML5 there is not a one to one mapping between definition and recommended use of header/footer and definition and recommended use of banner/contentinfo
13:01
<Stevef>
as anne said "landmarks are document/application global"
13:02
<Stevef>
they are not meant to be overused otherwise their utiltiy is lost
13:02
<Stevef>
landmark are typically big chunks a macro division of content
13:04
<MikeSmith>
if that's the case then the behavior that WebKit has adopted should be made a requirement in the spec, I think
13:04
<MikeSmith>
Firefox handles them differently?
13:05
<Stevef>
Firefox at the moment does am working on getting them to adopt webkit behaviour, at the moment header isn't exposed as a landmark at all
13:06
<MikeSmith>
oh
13:06
<Stevef>
note: they still get a role in webkit, a more generic 'group' role
13:06
<MikeSmith>
ok
13:07
<MikeSmith>
so to be clear, for the case where they are children of <main> the should still be landmarks, though, right?
13:07
<Stevef>
yes
13:07
<MikeSmith>
if so I can have the test check for that
13:07
<MikeSmith>
ok
13:07
<Stevef>
its only article/section
13:08
<MikeSmith>
hai
13:09
<Stevef>
will need to think on it a bit re: children of main
13:11
<Stevef>
typically header and footer are used as would be expected of banner/contentinfo in the wild, but when they are not they tend to be many instances on a page
13:12
<Stevef>
expected as in a single instance of each
13:12
<Stevef>
278 HTML5 pages using Header and Footer elements http://www.html5accessibility.com/HTML5data/header-footer/index-all.html
13:13
<Stevef>
usually header at top and footer at bottom of page unsurprisingly
13:36
<MikeSmith>
Stevef: oh strike what I said before. It seems like header and footer should not be landmarks if they're childen of main
13:37
<MikeSmith>
because people should not be using them as children of main to begin with, right?
13:43
<hsivonen>
is anyone maintaining a service that diffs W3C HTML and WHATWG HTML and visualizes the differences?
13:44
<MikeSmith>
hsivonen: not that I know of
13:45
<hsivonen>
ok :-(
13:47
<annevk>
MikeSmith: was it announced who got two year and who got one year on the TAG?
13:47
<MikeSmith>
hsivonen: one issue is, I think HTML diffing tools choke on the spec in its current size
13:47
<annevk>
I just noticed again that difference exists
13:48
<MikeSmith>
annevk: it was announced to the AC
13:48
<MikeSmith>
annevk: yeah you got the short straw
13:48
<MikeSmith>
oh
13:48
<MikeSmith>
hmm
13:48
<annevk>
that sounds positive
13:49
<annevk>
http://lists.w3.org/Archives/Public/www-tag/2013Jan/0016.html could be seen either as order of elected or on order of last name with my name incorrectly ordered as would be typical in the US
13:49
<MikeSmith>
I guess I shouldn't have said it here if it's not been announced
13:49
<annevk>
well http://www.w3.org/2001/tag/ has it as public
13:50
<MikeSmith>
annevk: yeah I've no insight into the order he chose
13:50
<annevk>
thanks for looking
13:52
<hsivonen>
annevk: I don’t see that page saying anything about the newly-elected members
13:52
<annevk>
hsivonen: I meant it'll be public soon enough
13:52
<hsivonen>
ah
13:54
<MikeSmith>
Stevef: if header and footer really are not meant to be used a direct children of main then you could make that an explicit requirement in the spec, in the Content Model part ("Flow content, with no header or or footer children.")
13:58
<annevk>
http://digbysblog.blogspot.ca/2013/01/the-poor-poor-rich-of-wall-street.html trololol
14:12
<MikeSmith>
annevk: yeah in that graphic the distressed looks on the faces of all the people was the best touch
14:16
<MikeSmith>
people struggling on a meager $200,000 salary plus only $25,000 a year on top of that in investment income
14:16
<MikeSmith>
you really have to empathize with their hardship
14:17
<MikeSmith>
but then the kicker is you find out out their taxes are going to increase by $2900
14:18
<MikeSmith>
and then at the injustice of that heinous increase you can only feel outrage
14:19
<MikeSmith>
masterful
14:20
<Stevef>
MikeSmith: that will involve chnages to authroing advice for footer/header is spec
14:20
<MikeSmith>
Stevef: my point is it should not be mere advice
14:21
<MikeSmith>
it should be a requirement
14:21
<MikeSmith>
and yeah I realize it would need a change to the spec
14:21
<Stevef>
MikeSmith: well that means significant changes...
14:21
<MikeSmith>
I'm not asking about it right now for the purposes of tests for the patch
14:22
<MikeSmith>
Stevef: yeah, sure
14:22
<Stevef>
MikeSmith: realise that, need to look at data and use cases then...
14:22
<Stevef>
see what shakes out
14:23
<MikeSmith>
ok
15:06
<Somatt_wrk>
any idea what happens when loading new html5 semantic elements via xhr in IE8 once we've made it discover them ?
15:07
<Somatt_wrk>
can they be parsed properly ?
15:29
<MikeSmith>
Stevef: OK I just posted and updated patch to the bug, with separate ax tests for Chrome and Safari
15:30
<Stevef>
yeah saw that thanks, hopefully will pass
15:30
<MikeSmith>
fingers crossed that it will actually pass when the mac-ews bot runs it
15:30
<MikeSmith>
cause I am not able to actually run it locally
15:34
<MikeSmith>
Stevef: it takes quite a while for the bots to run, but if you check later and all the platform indicators are green for the latest patch, then it's happy times
15:35
<Stevef>
cool will check
15:35
<MikeSmith>
I will now take the opportunity to drop off for a while and drink some shochu
15:38
<Stevef>
have one for me you deserve it
16:47
<annevk>
So is tag⊙wo subscribed to w3c-ac-forum⊙wo and am I therefore getting Member-only email again even though I cannot access Member-only archives?
16:47
<annevk>
And also, nobody told me I was added to tag⊙wo
17:06
<hober>
annevk: you're probably the first person elected to the tag who didn't simultaneously also have Member access
17:07
<annevk>
Does marcosc have Member access? I guess he might because of WebApps.
18:38
<dglazkov>
good morning, Whatwg!
18:41
<astearns>
good morning, dglazkov - is the 'is' attribute still immutable in the latest proposal?
18:41
<dglazkov>
astearns: yup
18:41
<astearns>
ok, thanks
20:20
<Hixie>
dglazkov: by "immutable" you mean "changing it does nothing", right? (not that the attribute itself doesn't change)
20:25
<dglazkov>
Hixie: right
20:27
<Hixie>
good good
23:23
<Hixie>
i stuck the html spec intro into http://splasho.nfshost.com/upgoer5/
23:23
<Hixie>
it said: UH OH! YOU HAVE USED NON-PERMITTED WORDS (SPECIFICATION, DEFINES, WEB, PLATFORM, DETAIL, WEB, PLATFORM, SPECIFICATION, STACK, RELATIVE, SPECIFICATIONS, SUMMED).
23:23
<Hixie>
well bummer
23:23
<Hixie>
that's like half the words of the intro
23:52
<astearns>
hixie: This paper shows how a big part of the world-wide tied-up things works, in lots and lots of ways. It goes with other papers that show how to make things work, shown in this picture: