09:06
<zcorpan>
if old gecko looked for "mathml" in the SI, and this is what the mathml spec uses, why is a new FPI suggested instead of doing what old gecko did?
09:06
<zcorpan>
the new FPI isn't going to fix the mathml spec
09:07
<zcorpan>
if the mathml spec is going to be changed (to use the new FPI), it could be changed to use one of the already supported doctypes or to avoid entities
09:50
<odinho>
So, webapps (particulary) and webappsec joining web-platform-tests will create a folder explosion because of all the different specs.
09:50
<odinho>
I am right in that we'll have toplevel folders for every spec, right?
09:50
<odinho>
(because I just said so on the webappsec list :P heh)
09:51
<Ms2ger>
Yes
09:51
<Ms2ger>
Webapps has ~15 at this point
09:51
<odinho>
Okay, goodie, I didn't lie.
10:03
<zcorpan>
Hixie: is it possible to allow https for resources.whatwg.org ?
10:03
<Ms2ger`>
Ah, darobin_
10:04
<zcorpan>
or should we mirror e.g. file-bug.js somewhere on dvcs.w3.org or so? (how?)
10:05
<zcorpan>
darobin_: ^
10:05
<darobin>
zcorpan: what do you want to do with file-bug.js?
10:06
<Ms2ger`>
darobin, I want you to fix https://w3c-test.org/webapps/DOMCore/tests/approved/interfaces.html
10:06
<zcorpan>
darobin: include it in e.g. https://dvcs.w3.org/hg/csswg/raw-file/tip/cssom-view/Overview.html
10:07
<zcorpan>
darobin: http: no longer works and resources.whatwg.org doesn't support https
10:07
<darobin>
zcorpan: ok, but I'm missing some context :) does it need to be on the same origin?
10:07
<darobin>
ah, gotcha
10:08
<darobin>
I reckon you could mirror it yes, or I could set that up
10:08
<darobin>
I could even mirror it to w3c-test.org which does support https
10:08
<zcorpan>
darobin: if you could set up a mirror somewhere, that'd be splendid :-)
10:09
<darobin>
zcorpan: is the source on github somewhere?
10:09
<darobin>
Ms2ger`: it's looking like someone clobbered the webidl2.js clone there, investigating
10:10
<Ms2ger`>
darobin, more to the point, there's no WebIDLParser.js
10:10
<zcorpan>
yes. https://github.com/whatwg/resources.whatwg.org
10:10
<darobin>
zcorpan: ta, looking at that now
10:10
<darobin>
Ms2ger`: hmmmm, yeah, that's... from webidl.js v1 — it doesn't exist anymore
10:11
<Ms2ger`>
darobin, why not just call it that too?
10:11
<darobin>
call what what?
10:11
<Ms2ger`>
The new parser .js file
10:11
<darobin>
it doesn't have the same interface
10:12
<darobin>
that DOM test is referring to a dirty hack from ancient history
10:12
<Ms2ger`>
And so do all other idlharness.js tests
10:12
<darobin>
Ms2ger`: oooh, fun!
10:12
<darobin>
so, when I announced an update, I'm guessing no one took notice?
10:13
<Ms2ger`>
Oh, I did test the update
10:13
<Ms2ger`>
Only at that point, it didn't require changing the test
10:13
<darobin>
I didn't remove the WebIDLParser.js
10:13
<darobin>
I have no idea who did
10:13
<darobin>
ok, let me apply a quick fix
10:15
<darobin>
oh right, the quick fix won't work
10:15
darobin
injects some coffee
10:15
<darobin>
Ms2ger`: may I propose a way forward?
10:15
<darobin>
how about you move those tests to GH
10:16
<darobin>
once they're there, I fix all the idlharness madness?
10:16
<Ms2ger`>
How?
10:17
<Ms2ger`>
darobin, I mean, the tests need to include a <script> that points to the parser, right?
10:19
<darobin>
zcorpan: http://w3c-test.org/resources.whatwg.org/
10:19
<tobie>
Wouldn't it make sense to compile those tests server side anyway?
10:19
<zcorpan>
darobin: thanks!
10:19
<Ms2ger`>
I probably object to any compilation step
10:19
<tobie>
Ms2ger`: why?
10:20
<Ms2ger`>
Because that makes it harder to set up
10:20
<darobin>
Ms2ger`: the right way with the new idlharness is: http://berjon.com/tmp/geotest/
10:20
<tobie>
makes what harder to setup?
10:20
<darobin>
so yes you need to change the <script>
10:20
<darobin>
Ms2ger`: what I was offering was to do that part, but in exchange for that you move the files to GH
10:21
<darobin>
that way, I don't get to do any hg, and you don't get to do any JS
10:21
<darobin>
seems like a good plan for all involved
10:21
<Ms2ger`>
darobin, why? What's wrong with changing the file name of the script to what it was before, instead of changing all the references?
10:21
<darobin>
hmmmm
10:22
<darobin>
I could add a rewrite for sure, but I'm wondering if it's enough
10:22
<Ms2ger`>
tobie, first, I have no idea what kind of compilation you're thinking about, and second, I want to be able to checkout a repository and not need to do any compilation step before I can run my tests
10:22
<Ms2ger`>
darobin, let's try that first, then :)
10:22
<darobin>
sorry, I got up at 5 to go get my daughter at the train station, this "logic" you insist on using is not making it's way through
10:22
darobin
needs sugar
10:23
Ms2ger`
wonders what darobin's daughter was doing at the train station at 5am
10:24
<darobin>
Ms2ger`: coming back from her ski holiday :)
10:32
<tobie>
Ms2ger`: what's your use case for running those tests?
10:32
<tobie>
I mean: what are you aiming to do when running them like that?
10:33
<Ms2ger`>
tobie, running tests while offline, adding tests and running them, integrating them in a continuous integration system
10:34
<tobie>
Ms2ger`: there's a bunch of tests that have build steps: e.g.: all ref tests, i18n tests, some other ones I can't remember right now.
10:34
<zcorpan>
Ms2ger`: i guess you have write access to anolis, right?
10:35
<Ms2ger`>
zcorpan, I do, yes
10:35
<Ms2ger`>
tobie, my ref tests run fine without any build steps
10:35
<zcorpan>
Ms2ger`: could you change the <!--logo--> substitution to load the image over https?
10:35
<tobie>
oh. on FF.
10:35
<zcorpan>
Ms2ger`: sub.py, search for "logo"
10:35
<tobie>
Sure.
10:35
<Ms2ger`>
Sure
10:35
<Ms2ger`>
zcorpan, you could also send me a pull request ;)
10:35
<zcorpan>
thanks
10:36
<tobie>
You have a custom build of the browser however. :)
10:36
<zcorpan>
Ms2ger`: oh it's on github now?
10:36
<Ms2ger`>
zcorpan, bitbucket
10:37
<Ms2ger`>
tobie, no, I've also got a test runner that allows running reftests
10:37
<Ms2ger`>
tobie, based on manual inspection, of course
10:37
<tobie>
oh.
10:37
<tobie>
yeah.
10:38
<tobie>
I'm wondering if there's be value in providing build of the test suite.
10:38
zcorpan
101 switching trains
10:38
<Ms2ger`>
I still don't know what you mean by "build"
10:38
<tobie>
damned. I thought switching terminology in the middle of my argument would trick you.
10:40
<tobie>
Ms2ger`: for example, i18n tests are all contained in a php file and generated on the fly into testharness tests.
10:41
<tobie>
idlharness tests are also turned into testharness tests, but client-side.
10:41
<Ms2ger`>
I'm not quite convinced that's a reasonable thing to do, but sure
10:41
<tobie>
I'm not convinced either.
10:41
<tobie>
I'd be doing it if I was.
10:42
<tobie>
Not asking about it.
10:42
<tobie>
:)
10:42
<Ms2ger`>
Anyway, I'm off for lunch and class after that
10:45
<Ms2ger`>
And tell zcorpan that I pushed his change
11:04
<zcorpan>
logo-ED is still http: but maybe that was pointed out already? is someone fixing that?
11:12
<zcorpan>
darobin: ^
11:13
<darobin>
damn, he left
12:00
<MikeSmith>
Ms2ger: https://w3c-test.org/resources/WebIDLParser.js seems to be back
12:01
<Ms2ger>
https://w3c-test.org/webapps/DOMCore/tests/approved/interfaces.html is still broken
12:01
Ms2ger
pokes darobin
12:01
<darobin>
MikeSmith: it's not the right one
12:01
<darobin>
I'll fix this after lunch
12:02
<MikeSmith>
ok
12:02
<darobin>
go back to vacation :)
12:03
<MikeSmith>
:-)
12:32
<zcorpan_>
darobin: here now
12:33
<darobin>
zcorpan_: I think that bug's been reported before but I can't find a trace of it so I reported it again
12:34
<zcorpan_>
darobin: ok, thanks
12:46
<zcorpan_>
would be fun to see how join() behaves with host objects as this
12:48
<darobin>
zcorpan_: I approve of your idea of fun
12:49
<zcorpan_>
http://software.hixie.ch/utilities/js/live-dom-viewer/saved/2165
12:51
<zcorpan_>
that one had the same behavior in opera/chrome/firefox. this one is different in firefox: http://software.hixie.ch/utilities/js/live-dom-viewer/saved/2166
12:52
<Ms2ger>
zcorpan_, yeah, that's probably getting fixed as we move form to our new bindings
13:07
<zcorpan_>
for how long are we going to have "experimental-webgl" as the context? forever?
13:09
<Ms2ger>
Probably
13:10
<zcorpan_>
splendid
13:14
<zcorpan_>
Ms2ger: reviewed http://critic.hoppipolla.co.uk/r/8
13:15
<Ms2ger>
Yep, thanks
13:15
<Ms2ger>
And good that that one is notreached indeed :)
13:23
<Ms2ger>
So is it expected that any time I click a critic link, I get redirected to /home instead?
13:30
<Ms2ger>
zcorpan_, so about the spec bug for typed arrays, apparently TC39 is going to spec them and the editors don't care about the Khronos spec anymore
13:30
<zcorpan_>
Ms2ger: is it reported to tc39?
13:31
<Ms2ger>
zcorpan_, not by me, but dherman said they'd make sure to match implementations
13:32
<zcorpan_>
Ms2ger: i got redirected to home once as well. weird. jgraham, you know what's up?
13:33
<odinho>
Man, that was GitHub integration. It asked me to authorize critic when I clicked that.
13:33
<odinho>
I pressed OK, got to home screen, -- but this next time I got to the review. :-)
13:33
<Ms2ger>
I've put a note on my todo list to review the tc39 spec once it happens
13:35
<zcorpan_>
Ms2ger: ok
13:35
<webben>
Hmm. Is there a sane way to read the HTTP bis work in HTML format?
13:36
<webben>
or is going and picking through https://wiki.tools.ietf.org/wg/httpbis/ to get the slightly crufty "HTML" the best currently possible?
13:36
<hsivonen>
From the "are CDMs plug-ins or not" dept: https://plus.google.com/100314989943113365362/posts/ULA2Wvb2MNE
13:39
<zcorpan_>
Ms2ger: jgraham: https://github.com/jensl/critic/issues/33
13:39
<Ms2ger>
webben, things like http://wiki.tools.ietf.org/id/draft-ietf-httpbis-authscheme-registrations-06.html are the best you can get
13:42
<Ms2ger>
odinho, fwiw, I tend to keep my manifests sorted on the filename
13:44
<odinho>
Ms2ger: But they don't even seem alphabetically sorted atm :-)
13:44
<Ms2ger>
odinho, they should be if you ignore the "support" part
13:46
<odinho>
Oh, that's too advanced for me to understand, okay :-) No fan, but I see you have a system.
13:47
<odinho>
I think I would've done -> manual a, support a, support b, a, b, c
13:48
<Ms2ger>
That would also work
13:48
Ms2ger
doesn't think it matters too much
13:48
<odinho>
But haven't written in those extra things. -- Nah, it doesn't. Hence note :D
13:49
<webben>
Ms2ger: that's an improvement over what I'd found cheers
13:50
<Ms2ger>
np
13:50
<darobin>
Ms2ger: check https://w3c-test.org/webapps/DOMCore/tests/approved/interfaces.html again sweetie
13:50
<Ms2ger>
Thanks, my love
13:56
<jgraham>
Ms2ger, odinho: Yeah, the critic GH integration doens't pass the url you were actually tryimng to access through the OAuth redirection hoops, so you always end up on the homepage after you log in
13:56
<jgraham>
It's a bug
13:56
<Ms2ger>
Okay
13:57
<odinho>
jgraham: Still I'm impressed, quite nice work there :-)
14:39
<Ms2ger>
Hmm, does critic automatically catch addressed issues?
14:54
<darobin>
zcorpan_: I'm told the logo bug should be gone
14:55
<zcorpan_>
darobin: verified. thanks!
14:56
<darobin>
zcorpan_: I only sent an email, the almighty systeam actually took care of it :)
14:57
<zcorpan_>
Ms2ger: if something is changed on the line(s) of the issue, critic assumes the issue is addressed, i think
14:57
<Ms2ger>
Interesting
15:18
<Ms2ger>
So, anyone who still wants to have a look at the webgl tests?
15:45
<Ms2ger>
If so, feel free to do that at some later point
15:49
<zcorpan_>
jgraham: hmm. i'm trying to use the return value of a function that i run with step()
15:49
<zcorpan_>
jgraham: but step() seems to throw away the return value
15:50
<Ms2ger>
What would you expect it to return?
15:50
<Ms2ger>
step() runs immediately
15:52
<zcorpan_>
var rv = this.step(function() { return 'Foo'; }); // rv === 'Foo'
15:53
<Ms2ger>
Mm
15:53
<zcorpan_>
currently rv === undefined
15:53
<dglazkov>
good morning, Whatwg!
15:53
<Ms2ger>
Pull request? ;)
15:54
<zcorpan_>
sure. i think it's just prepend 'return ' to this line: func.apply(this_obj, Array.prototype.slice.call(arguments, 2));
16:08
<zcorpan_>
https://github.com/w3c/testharness.js/pull/14
16:28
<zcorpan_>
are there specific Date values that are particularly interesting to test in structured clone?
16:30
zcorpan_
notices that new Date(-0) has two possible implementations per spec
16:46
<jwalden>
it might not take too much pushing to give that only one implementation, if browser engines agree on what should happen with that, but I don't know the actual reason both are permitted -- choice-of-representation seems pretty weak without different implementations actually requiring it
16:54
<zcorpan_>
http://software.hixie.ch/utilities/js/live-dom-viewer/saved/2167 - opera/chrome have different representations, firefox doesn't
16:55
<zcorpan_>
maybe it depends on the hardware, too
16:56
<zcorpan_>
i'll just test that, whatever the representation, it's consistent between window and worker and it doesn't get mangled from structured clone
18:39
<annevk>
MikeSmith: http://platform.html5.org/history/ needs an entry for fonts
18:39
<annevk>
MikeSmith: but I'm not sure where
20:26
<Hixie>
man, browsers don't seem to have caught up with the dnd model
20:26
<Ms2ger>
Hah
20:28
<tantek>
Hixie, maybe it's a bad model?
20:29
<Ms2ger>
Maybe it's such horrible code that nobody wants to touch it with a foot-long pole?
20:45
<tantek>
Ms2ger, why not both?
20:47
<Ms2ger>
Possible
22:01
<Hixie>
tantek: it's the improvements to the model they're not picking up, not the original model
22:09
<TabAtkins>
Yeah, the original terrible model is reasonably usable everywhere, last I checked.
22:09
<TabAtkins>
But not the good stuff with @dropzone and such.
22:20
<tantek>
maybe there's insufficient web developer demand for the improvements?
22:20
<tantek>
I don't see many web sites with dnd in general
22:20
<tantek>
how well does dnd work with mobile / touch interfaces?
22:20
<tantek>
(wild guess as to why people are not worrying/caring about it)
22:21
<Hixie>
yeah, i'm sure it's not a high priority
22:26
<gsnedders>
tantek: FB uses DND on desktop, at least.
22:31
<tantek>
gsnedders, as does G+ with the whole goofy circles thing
22:32
<tantek>
DnD people from one circle to another etc.
22:32
<tantek>
but of course none of that is available on mobile / touch / iOS etc.
22:32
<tantek>
so it's … meh
22:32
<Hixie>
i don't understand why we couldn't make dnd work on mobile
22:32
<Hixie>
but yeah
22:33
<tantek>
Hixie - ask the UI/UX/designer people
22:34
<tantek>
and I agree - there's something odd how mobile touch UIs are all about "swipe this and swipe that" but no drag & drop
22:34
<tantek>
which seem like almost the same physical gesture
22:34
<tantek>
is it the difference of having a specific hit-target to start swiping from?
22:35
<tantek>
and don't mobile games use swiping from specific hit-targets?
22:35
<tantek>
not sure why it hasn't made its way into web UIs
22:35
<tantek>
just observing that it hasn't
22:37
<gsnedders>
tantek: Does FB not serve the desktop version to tablets? On the other hand, you can't drag files around on iOS/Android…
22:43
<Hixie>
it took me about 3 minutes to work out that "R" in this bug means "are". https://www.w3.org/Bugs/Public/show_bug.cgi?id=9482
22:43
<Hixie>
and the guy probably saved all of 200ms.
22:43
<Hixie>
NOT A WIN.
22:43
<Hixie>
tantek: there's drag-and-drop in some mobile stuff, e.g. bad piggies uses it.
22:44
<gsnedders>
I feel like I should give a witty response to that in txtspk, but I don't think my knowledge of the language is good enough.
22:47
<tantek>
and here I thought R was a programming language http://en.wikipedia.org/wiki/R_%28programming_language%29
22:54
<jwalden>
somebody tell zcorpan if they see him before I do that implementation of new Date(-0) has two representations *only* so that implementations can store dates, internally, as either an int64_t or a double; SpiderMonkey stores internally as double but adds 0 in TimeClip so -0 doesn't have weird behavior; for any implementation using double representation it's easy to look like int64_t, but...
22:54
<jwalden>
...probably far harder the other way; thus mandating +0 behavior is probably the thing the most people'd be compatible with; I can't think of a good reason to support this difference at the hardware level -- anyone using double representation can do +0 to pretend to be int64_t-represented, and implementations have to have double capabilities to represent math generally, so I don't see why...
22:54
<jwalden>
...mandating as-if-by-int64_t would be prevented by hardware issues