00:09
<smaug____>
is whatwg.org down?
00:11
<smaug____>
ah, works now
00:39
<Jasper>
Down again?
00:46
<zewt>
... who would think that browsers can't generate a securely random token? heh
00:47
<jamesr>
4
00:47
<zewt>
4 people?
00:47
<jamesr>
no 4 - chosen by a fair dice roll
00:47
<jamesr>
more seriously, what about https://developer.mozilla.org/en-US/docs/DOM/window.crypto.getRandomValues ?
00:48
<jamesr>
chrome 11+, firefox 21+
00:48
<jamesr>
safari ????
00:48
<zewt>
i mean, sicking guessed that there was a concern that browsers couldn't generate securely-random blob urls, i'm just wondering who would actually think that, heh
00:49
<jamesr>
oh heh. well they ship an API that's supposed to generate securely-random values to JS, so they'd better have a way to do it
00:49
<zewt>
on a technical level if you don't have a secure prng, you probably can't even do tls
00:49
<jamesr>
maybe you could point him at that API
01:22
<Hixie_>
Jasper: yeah, see dreamhoststatus.com for updates
01:22
<Jasper>
Aha, thanks.
01:22
<zewt>
http://www.downforeveryoneorjustme.com/whatwg.org
11:22
<annevk>
beverloo: too pedantic for Twitter: they could have named it in a camelcase way ;-)
11:22
<beverloo>
perhaps, there's an inconsistent mix between lowercase and camelcase right now
11:23
<darobin>
that would make a great tagline
11:23
<darobin>
"AnneVK — Too Pedantic For Twitter"
11:50
<jgraham>
I thought that was Ms2ger
13:40
<zcorpan>
anyone with IE9+ around?
13:41
<jgraham>
Yeah, kind of, if I switch computers
13:43
<zcorpan>
i'm wondering what this logs: http://software.hixie.ch/utilities/js/live-dom-viewer/saved/2184
13:44
<jgraham>
600px
13:47
<zcorpan>
jgraham: thanks
13:47
<zcorpan>
so opera implements the spec, everyone else doesn't
13:50
<jgraham>
unpossible
13:51
<zcorpan>
inorite
13:58
zcorpan
installs jumpcut
14:06
<darobin>
does anyone recall something about it being bad for a random third party to mint their own values for @role?
14:07
<darobin>
ISTR a discussion in which it was frowned upon, but I can't seem to find it again
14:10
<jgraham>
Doesn't that have the same problem as minting random proprietary things in general?
14:11
<darobin>
slightlyoff: I take it you're aware of http://webdevdata.org/ and http://www.w3.org/community/webdevdata/?
14:12
<darobin>
jgraham: well, a) extension points are always evil, and b) there are varying degrees of proprietariness
14:12
<darobin>
in this case, the epub people are looking at transitioning from XHTML to proper HTML
14:12
<darobin>
they have idpf:type attributes the values of which would be a good fit for role
14:12
<darobin>
and dropping the namespacing would help
14:13
<slightlyoff>
darobin: nope. It looks to just be scraped via a crawler?
14:13
<darobin>
but at the same time, I don't think that importing all of the IDPF roles into mainline HTML is necessarily great either
14:13
<darobin>
slightlyoff: the data is from a crawler yes, but the CG is the author and I reckon you'd find like-minded souls in there to help port your extension and such
14:13
<wilhelm_>
zcorpan: browserstack.com is your friend.
14:14
<slightlyoff>
darobin: true story
14:14
<slightlyoff>
darobin: was entirely unaware of the effort
14:14
<slightlyoff>
darobin: thanks for the pointer!
14:14
<darobin>
slightlyoff: you're most welcome :)
14:14
<zcorpan>
wilhelm_: thanks
14:14
<slightlyoff>
...and now I get to wait on my AC rep to approve it = )
14:15
<darobin>
yeah, I love how CGs are trivial for anyone *unless* you're a member :)
14:19
<slightlyoff>
also, this looks to be an incredibly shallow crawl
14:19
<slightlyoff>
taht's not bad...but yeah
14:20
<darobin>
slightlyoff: I haven't been tracking the effort but I suspect they did a first page crawl based on Alexa, similar to the Paciello dataset
14:20
<darobin>
so yes, shallow
14:20
<darobin>
marcosc would know, but he's on some remote Greek island this week
14:21
<slightlyoff>
darobin: is there a graph analysis framework on top of this?
14:21
<slightlyoff>
darobin: or is it just an occasional crawl for now?
14:22
<darobin>
slightlyoff: I have no idea, but it's a fairly recent endeavour so I would assume it has goodwill but not necessarily much yet
14:22
<slightlyoff>
who is doing the engineering?
14:23
<darobin>
it looks like it's all marcos, but you'd have to ask him really
15:49
<zcorpan>
jgraham: format_value (in testharness) doesn't play nice with arrays that have circular properties
15:50
<jgraham>
I wonder if that's my fault or not
15:50
<jgraham>
Should be easy enoguh to fix I think
15:51
<zcorpan>
Maximum recursion depth exceeded(stack: format_value([arguments not available])@https://w3c-test.org/resources/testharness.js:500 format_value([arguments not available])@https://w3c-test.org/resources/testharness.js:500 format_value([arguments not available])@https://w3c-test.org/resources/testharness.js:500 format_value([arguments not available])@https://w3c-test.org/resources/testharness.js:500 format_value([arguments no
15:51
<zcorpan>
t available])@https://w3c-test.org/resources/testharness.js:500 format_value([arguments not available])@https://w3c-test.org/resources/testharness.js:500 format_value([arguments not available])@https://w3c-test.org/resources/testharness.js:500 format_value([arguments not available])@https://w3c-test.org/resources/testharness.js:500 format_value([arguments not available])@https://w3c-test.org/resources/testharness.js:500 fo
15:51
<zcorpan>
rmat_value([arguments not available])@https://w3c-test.org/resources/testharness.js:500)
15:51
<zcorpan>
is the message when an assertion fails
15:51
<zcorpan>
for objects this problem doesn't surface, it just says "[object Object]"
15:52
<zcorpan>
unless there's a custom toString that iterates over the properties, i guess
15:52
<jgraham>
Yeah, it needs to record things that it has already seen and avoid recursing into those
15:53
<jgraham>
I can fix this evening
15:53
<zcorpan>
ok, thanks
16:04
<zcorpan>
apparently ie10 implements this requirement of structured clone but opera/webkit/gecko don't: "Let output be a newly constructed empty Array object whose length is equal to the length of input."
16:04
<zcorpan>
test: http://software.hixie.ch/utilities/js/live-dom-viewer/saved/2185
16:10
<darobin>
zcorpan: is there a test for that in the TS?
16:10
<darobin>
it's a requirement that seems to make sense to me
16:10
<zcorpan>
darobin: there will be when i submit the test i'm working on
16:11
<zcorpan>
which is "test everything you can think of in structured clone and transferable", which turns out to be quite a bit
16:11
<zcorpan>
i haven't even started with transferable
16:12
<darobin>
zcorpan: awesome
16:13
<zcorpan>
where "everything you can think of" is really more like "the least amount of tests to have a reasonable coverage of the spec", but still it's quite a bit
16:51
<gavinc>
URL spec question, http://a.example/%66oo-bar is not strictly equivalent to http://a.example/foo-bar but http://a.example/%E8oo-bar is equivalent to http://a.example/èoo-bar?
16:53
<SteveF>
slightlyoff: for background on how the webdevdata.org stuff was gathered http://blog.paciellogroup.com/2012/04/html5-accessibility-chops-data-for-the-masses/
16:53
<gavinc>
from, "Code points higher than U+009F will be converted to percent-encoded bytes by the URL parser."
16:53
<SteveF>
slightlyoff: i.e. extremly lo-fi but thats my style and the limit technically
16:54
<slightlyoff>
SteveF: I'm trying to join the CGs now (which i wasn't aware of)
16:54
<slightlyoff>
SteveF: apologies for duplication
16:55
<SteveF>
no duping I nearly creamed myself when I saw your stuff, its great!!!!
16:55
<SteveF>
will happily promote and help where I can
17:13
<jwalden>
gsnedders: making asm.js compilation parallel is coming along pretty quickly, I'd be slightly surprised if it doesn't land before the next aurora/beta/release uplift
17:37
<MikeSmith>
"Is Sam Ruby happy? If not, make him happy." https://github.com/rails/rails/blob/master/RELEASING_RAILS.rdoc#is-sam-ruby-happy--if-not-make-him-happy
18:26
<rillian>
MikeSmith: that was my question, yeah. :)
18:26
<rillian>
the spec says it "must, on getting, return the track element's text track's corresponding TextTrack object."
18:26
<rillian>
so if there's no corresponding object...
18:26
<rillian>
fwiw Chrome returns a default empty TextTrack
18:31
<Ms2ger`>
rillian, clearly there always is one
18:31
<rillian>
is that clear?
18:32
<Ms2ger`>
Well, otherwise the spec would handle that case
18:38
<rillian>
I don't get the sense that the spec thought about that code path
18:55
<smaug____>
Ms2ger`: :)
19:27
<jgraham>
Someone want to review http://critic.hoppipolla.co.uk/r/20 ?
19:27
<jgraham>
Ms2ger`: ^
19:28
<Ms2ger`>
I'll look in a bit
19:30
<jgraham>
Sure, no rush
20:46
<Krinkle>
paul_irish: The repaint improvements in MediaWiki went live yesterday on English Wikipedia (MediaWiki 1.21wmf12)
20:47
<paul_irish>
Krinkle: Thanks! that's great to hear.
20:51
<Krinkle>
paul_irish: Though we've both done some testing, would be nice to get some feedback from you (or people you work with) on how it works out in a real world scenario (like the scenario that brought us here)
20:53
<paul_irish>
Sure thing. I'll take a look later today
22:30
<Hixie_>
anyone here have a default-setup windows or mac install for either israel or egypt they can run some tests for me on?
23:38
<Hixie_>
hsivonen: ping https://www.w3.org/Bugs/Public/show_bug.cgi?id=17845
23:48
<zewt>
dear youtube: stop treating right-clicks as clicks. thanks -me