03:26
<alix88>
Are there any WebVTT people in here?
03:30
<alix88>
This room is sure lively.
03:30
<alix88>
I'm looking for anybody who might be familiar with cue settings in the WebVTT spec
03:38
<alix88_>
..
03:44
<alix88>
ERMEGERD
03:46
<celtboy>
WERB STERNDERDS
03:46
<celtboy>
Was someone wanting to discuss VTT?
03:46
<alix88>
That was me. Know much about it?
03:50
<alix88>
Guess this isn't the place to discuss w3 specs :) Goodnight, room full of silent people.
08:14
<zcorpan>
wow alix88 can't have heard of this thing called time zones
08:14
<zcorpan>
but maybe i should have popped in here when my daughter waked me up...
08:16
<zcorpan>
alix88: if you're reading logs, then, next time, just ask you question, and maybe somebody can answer it later
08:21
<krijnh>
Tee hee
09:42
<smaug____>
hsivonen: welcome back
09:44
<annevk>
so the new Maps.app is terrible, but the new Phone.app lets me call again
09:44
<annevk>
working Phone.app prolly beats Maps.app
09:45
<annevk>
krijn: teehee indeed
09:51
<jgraham>
annevk: My 200SEK Nokia thing lets me make calls, so it seems to me that if you are paying 6000+SEK then that's 5800SEK's for broken mapping
09:52
<jgraham>
s/SEK's/SEK/
09:52
<annevk>
jgraham: if it only were that simple :)
09:53
<zcorpan>
5800 for broken Maps.app
09:53
<zcorpan>
sounds like a fair deal :-)
10:22
<hsivonen>
smaug____: thanks
11:30
<annevk>
Does http://url.spec.whatwg.org/#concept-url-parse look comprehensible enough?
11:31
<annevk>
It's quite a bit of work to write it all out, so I rather not complete it only to hear it has to be redone
11:40
<annevk>
Ah, no feedbac?
11:46
<odinho>
annevk: Well, it takes time to read and understand, -- but that's just because that's how it is. I understand what's happening if that was the question.
11:47
<annevk>
sounds good enough to me :)
11:47
<annevk>
and yeah, URLs are annoying :)
11:49
<Ms2ger>
Maybe a list of the possible states would be useful?
11:59
<annevk>
yeah maybe, lets first write the whole thing down
12:00
<annevk>
then fix a few small things and then maybe start raising some questions as to how people want various interop issues resolved
12:10
<zcorpan>
annevk: do you generate the spec from the javascript impl?
12:10
<annevk>
nah
12:11
<annevk>
it's only a couple of hours to write it out, figuring out the conversion code would be more and it would not be good or readable anyway
12:12
<annevk>
but it's written based on the JavaScript impl
12:13
<zcorpan>
annevk: would be good if it said at the top what the algorithm can return (or what it does)
12:14
<zcorpan>
annevk: e.g. i expected it to return an error instead of just aborting in step 3 in scheme start
12:15
<annevk>
you can only get in step 3 if you set .protocol
12:15
<annevk>
override stuff is like innerHTML madness in the HTML parser
12:16
<annevk>
but yeah, some notes should be added in due course
12:16
<annevk>
forks encouraged :)
12:19
<zcorpan>
yay https://bugzilla.mozilla.org/show_bug.cgi?id=744907
12:22
<zcorpan>
annevk: should it be possible to pass in a URL object as baseUrl?
12:23
<annevk>
I'd rather not do that
12:24
<zcorpan>
why not?
12:24
<zcorpan>
what happens if the base url is invalid?
12:24
<annevk>
we can make it work, but URL objects can be manipulated in weird ways and the parser would then have to account for that
12:24
Ms2ger
wonders why this Zynga guy sends email in euc-kr
12:24
<annevk>
good question
12:24
<annevk>
zcorpan: should maybe throw for that case
12:25
<annevk>
zcorpan: also if the base URL hierarchical flag is not set, both are kinda useless
12:25
<annevk>
zcorpan: I'll add an XXX
12:25
<zcorpan>
if i have a URL and want to resolve something against it, it seems silly to have to serialize it for it to work
12:26
<marcosc_>
Anyone here interested in the responsive images stuff?
12:26
<Ms2ger>
No
12:26
<marcosc_>
:)
12:26
<annevk>
zcorpan: e.g. consider if you have a URL and then set its protocol to data
12:27
<annevk>
zcorpan: it's still a "hierarchical" URL because browsers do weird stuff
12:27
<annevk>
not sure what result you want out of that
12:27
<annevk>
I mean if you come with a proposal that solves all the weird shit, mkay
12:32
<zcorpan>
annevk: can you give a list of weird states i need to cover?
12:33
<annevk>
prolly once I have looked at it in more detail
12:33
<annevk>
what we could do btw
12:33
<annevk>
is that if you pass a URL object in
12:33
<annevk>
we just parse its URL.href again
12:33
<annevk>
and if that fails, you fucked up
12:33
<zcorpan>
annevk: tentatively i'd just check for the weird states and throw INVALID_STATE_ERR
12:34
<zcorpan>
or have a flag "weird state" that gets set when the url is mutated in weird ways, or whatever
12:36
<zcorpan>
i guess parsing URL.href could be black-box equivalent
12:36
<annevk>
it better be :)
13:11
<annevk>
Anyone know if it's normal for GitHub to lag behind?
13:11
<annevk>
Something I committed has not shown up on the site thus far...
13:12
<Ms2ger>
Hmm, https://github.com/whatwg suddenly looks rather different
13:13
<annevk>
today's redesign I guess
13:13
<annevk>
but I'm more worried about the syncing
13:14
<jgraham>
Doesn't sound that normal
13:14
<jgraham>
Did you check if it is actually in the repo?
13:18
<annevk>
it seems to be in now
13:19
<annevk>
but that was at least a delay of several minutes
13:40
<annevk>
and again
13:40
<annevk>
committed 3:40
13:40
<annevk>
lets see when it shows up
13:56
<annevk>
still not
14:03
<annevk>
now it is
14:03
<annevk>
so that's about 20 minutes
14:03
<annevk>
zcorpan: added the URL argument thingie: http://url.spec.whatwg.org/#api
14:03
<odinho>
maybe they're going freemium on the time.
14:03
<annevk>
zcorpan: haven't actually fixed the bugs it might cause
14:04
<annevk>
I wanted to expose invalid URLs
14:04
<annevk>
should I make that URLUtils.isInvalid?
14:04
<odinho>
Hmm. Just did a small private OTW. Opera is the only one having spaces in its Accept header...
14:10
<annevk>
they haven't fixed that yet?
14:10
<annevk>
ideally Accept is just nuked
14:10
<annevk>
but that's a compat issue
14:13
<odinho>
annevk: I might just do a troll patch for it. Saw the bug now. Really need those spaces out. I just had to fix a compat issue with a website because we have spaces.
14:14
<annevk>
ideally you also write that down somewhere
14:15
<annevk>
odinho: add it to http://wiki.whatwg.org/wiki/HTTP_Issues please
14:15
<annevk>
should maybe rename that page to HTTP to make it easier to find
14:16
<odinho>
annevk: I haz no user.
14:16
<annevk>
email?
14:16
<odinho>
annevk: the obvious one I guess :]
14:21
<zcorpan>
ilovespicegirls⊙hc ?
14:21
<Ms2ger>
I thought that was you, zcorpan?
14:21
<zcorpan>
wait, maybe that was lachy
14:21
<annevk>
odinho: moved it to HTTP
14:22
<annevk>
wait, I've not been chatting with a girl all this time?
14:22
<Ms2ger>
That's what I thought when I heard "Anne"
14:23
<annevk>
Belgians
14:23
<zcorpan>
Ms2ger though is clearly a girl
14:23
<annevk>
or an unproductive robot
14:23
<Ms2ger>
You saw through the "Ms."?
14:23
<zcorpan>
yep
14:23
<jgraham>
Oooh, like Ms Pacman?
14:24
<jgraham>
Ms2ger is actually an 80s video game gone sentinent?
14:24
<Ms2ger>
Beep
14:24
Ms2ger
eats the ghost of longdesc
14:24
<jgraham>
*sentient
14:27
<jgraham>
Actually longdesc is quite like the pacman ghosts
14:27
<Ms2ger>
It always comes back?
14:27
<jgraham>
Yeah :(
14:27
<Ms2ger>
http://w3cmemes.tumblr.com/post/31865121758/the-joker-shares-his-approach-on-css2-1-issues also applies to HTML5, I guess
14:27
<Ms2ger>
"W3C Invites Implementations of The WebSocket API"
14:28
<Ms2ger>
Oh, why thank you, I'll start implementing right away
14:28
<jgraham>
The W3C would be rubbish as party planners
14:28
<zcorpan>
yeah we've been waiting for the w3c invitation!
14:28
<jgraham>
We'd be half way through trashing their house before they sent out the invites
14:29
<Ms2ger>
Of only that was all they were rubbish at
14:29
<Ms2ger>
If, even
14:58
<annevk>
TC39 has some complicated way to define percent escaping: http://es5.github.com/#Encode
14:59
<annevk>
I was hoping for a couple of lines :)
15:44
<annevk>
mine are a bit shorter
15:44
<annevk>
http://url.spec.whatwg.org/#percent-escaped
16:05
marcosc
notes that the hour glass emoji don't worky in Chrome
16:39
SamB_MacG5
wonders why he's supposed to leave his sense of logic at the door
16:39
SamB_MacG5
wonders "What door?"
16:40
<SamB_MacG5>
So, I was wondering why XBL 2.0 was scrapped. Is this a good place to find out, or do you know a better place?
16:40
<annevk>
nobody implemented it
16:41
<SamB_MacG5>
I wish W3C docs would/could mention stuff like that ...
16:42
<annevk>
you mean like http://www.w3.org/TR/xbl/#status-of-this-document
16:44
<SamB_MacG5>
Pretty sure I read that, but it wasn't clear that the lack of plans to implement was the reason for abandonment, rather than the other way around
16:44
<annevk>
anyway
16:44
<annevk>
search for "web components"
16:44
<annevk>
or bug dglazkov
16:45
<annevk>
that's the new hot thing
16:45
<annevk>
though like XBL vendor interest is scarce
16:45
<annevk>
with /\
16:45
<SamB_MacG5>
/\ ?
16:46
<Ms2ger>
An arrow
16:46
<SamB_MacG5>
ah
16:46
SamB_MacG5
is used to ^ for that, possibly with dashes on one side or the other
16:46
<Ms2ger>
~^~
16:47
<annevk>
mustache?
16:48
<SamB_MacG5>
anyway, who can I bug about clarifying http://www.w3.org/TR/xbl/#status-of-this-document ?
16:53
<marcosc>
SamB_MacG5: there should be a mailing list mentioned in the status of this document
16:53
<SamB_MacG5>
true
16:54
<SamB_MacG5>
there are actually *two* of them
16:55
<SamB_MacG5>
I guess public-webapps⊙wo is probably more appropriate for such administrative feedback?
16:55
<marcosc>
Yeah, hit those guys up... but I would recommend some text to be added... otherwise, I doubt anyone will do anything about it
16:56
<SamB_MacG5>
you mean, I should suggest a specific piece of text to insert?
16:57
<SamB_MacG5>
and possibly replacements for parts I think are unclear?
18:26
<annevk>
Ms2ger: so should we move DOM Parsing into DOM?
18:26
<Ms2ger>
Mm, I dunno
18:26
<Ms2ger>
It's icky markup ;)
18:27
<annevk>
you mean the source code or you mean you don't want markup features to spoil the DOM?
18:27
<Ms2ger>
The latter
18:28
<annevk>
okay
18:32
<annevk>
only a few days left and apart from IPv6 I will be 80% done
18:32
<annevk>
then of course comes the next 80%
18:33
<Ms2ger>
Heh
18:35
<annevk>
it's not too bad for a couple of weeks of research, although I guess it was closer to one or two months, don't quite remember when I started
18:35
<annevk>
I wonder what Hixie estimated back in the days for writing the URL part
18:36
<annevk>
I wonder if he included writing the Encoding Standard so http://url.spec.whatwg.org/#query-percent-escaped would be defined :)
18:42
<matjas>
Ms2ger: you magnificent bastard, you https://www.w3.org/Bugs/Public/show_bug.cgi?id=11204#c46
18:50
<annevk>
hey matjas
18:50
<annevk>
matjas: did you get my feedback the other day?
18:51
<matjas>
annevk: i didn’t get anything
18:52
<matjas>
annevk: was it here in #whatwg?
18:53
<annevk>
matjas: http://logbot.glob.com.au/?c=freenode%23whatwg&s=18%20Sep%202012&e=18%20Sep%202012#c722062
18:53
<annevk>
matjas: though in the end I guess all that was valid was the typo, as what should really be done is not entirely clear yet
18:54
<matjas>
thanks
18:56
<annevk>
matjas: are you using that Punycode in the wild btw?
18:56
<annevk>
matjas: I was wondering whether we should expose such a thing as native API
18:57
<matjas>
annevk: Punycode.js ships with Node.js, and some other open-source projects depend on it, so yes I think it’s being used
18:58
<annevk>
ah yeah, outside of browsers
18:58
<matjas>
annevk: (Node.js uses it internally for its URL stuff)
18:58
<annevk>
yeah, noticed that when looking at their URL impl
18:58
<matjas>
IMHO it would be a useful API to have built-in
18:58
<matjas>
in browsers
18:59
<matjas>
especially since stuff like <input type=email> disallows IDNA
18:59
<annevk>
maybe we could have static methods on URL
18:59
<annevk>
URL.toPunycode() / URL.fromPunycode() or some such
19:02
matjas
donates Punycode.js unit tests to browser vendors
19:08
<matjas>
annevk: just noticed your last three blog posts too, yay!
19:16
<matjas>
“There are other characters that are used as "full stops" that one could consider mapping as label separators, but their use as such has not been investigated thoroughly.” Thanks, RFC 5895
19:34
<humph_>
I had a question about the best way to write conformance tests for WebVTT, specifically how and where to capture metadata about the different good/bad .vtt files
19:35
<humph_>
should we put them in the .vtt file's comment block, in a separate file per test, in a wiki, in one file keyed by test file, etc.
19:35
<humph_>
what's the right approach?
19:35
<jgraham>
zcorpan: ^
19:36
<jgraham>
If it were me and I was writing a parser, I might develop a simple text format to embed the VTT and a desciption in
19:36
<zcorpan>
humph_: webvtt doesn't currently support comments, so your options are in the cue text or in a separate file
19:36
<jgraham>
Like we did for the HTML parser tests
19:36
<jgraham>
Then have a script to spit out the actual browser-readable tests
19:36
<humph_>
jgraham: so a file that you'd split the .vtt out of at test time?
19:37
<jgraham>
Yeah
19:37
<zcorpan>
humph_: for some of my parser tests i've put json in cue data with expected settings/timings; and a separate json file with expected number of cues or "error"
19:37
<humph_>
I was afraid of putting it in the cue text and making the tests unnecessarily large
19:38
<zcorpan>
for my other parser tests i used html5lib style like jgraham suggested
19:38
<zcorpan>
(i.e. for testing cue text parsing)
19:38
<zcorpan>
i haven't released these tests but i intend to
19:38
<humph_>
cool. I'm implementing webvtt in firefox with my class this term, just wanted to figure out the best way to start
19:39
<humph_>
they are beginning with these tests
19:39
<zcorpan>
ok
19:39
<jgraham>
Nice project :)
19:39
<jgraham>
And testing++
19:39
<humph_>
http://vocamus.net/dave/?p=1525
19:39
<humph_>
yeah, should be fun
19:40
<jgraham>
I hope you will release the results to W3C for inclusion in the official testsuite
19:40
<humph_>
yeah, sylvia suggested the same
19:40
<humph_>
I think that's a good plan
19:40
<jgraham>
sylvia++
19:40
<humph_>
indeed
19:40
<humph_>
alright, thanks for the ideas
19:41
<jgraham>
If you need any help with test harness wrangling or getting the tests submitted feel free to ask me
19:41
<humph_>
k
19:41
<humph_>
I started by doing a node module to wrap the js parser
19:41
<jgraham>
Or there is #testing on the W3C server
19:41
<humph_>
https://github.com/humphd/node-webvtt
19:41
<humph_>
so they could at least validate in a test harness
19:42
<zcorpan>
i used testharness.js for testing the parser and the dom api, and reftests for rendering
19:48
<annevk>
matjas: nice find
19:48
<annevk>
matjas: so yeah, what we end up with here kinda depends on what browsers are going to implement :(
19:48
<annevk>
matjas: Opera has some kind of mixture, Gecko is aiming to adopt http://jprs.co.jp/idn/index-e.html
19:48
<annevk>
matjas: WebKit has no plans
19:49
<annevk>
dunno about Internet Explorer
19:49
<Hixie>
annevk: any chance i can get "nodes are inserted" to be defined so as to give me one hook per node that is inserted, in tree order?
19:49
<Hixie>
annevk: (if not no worries, i can do it on my end)
19:50
<jgraham>
Hixie: Doing the "which order should scripts run when several are inserted" bug?
19:50
<Hixie>
good guess
19:50
<Hixie>
ues
19:50
<Hixie>
yes, even
19:50
<jgraham>
:)
19:51
<annevk>
Hixie: yeah we could rephrase it like that
19:52
<annevk>
Hixie: file a bug, can fix it tomorrow prolly
19:52
<annevk>
I think HTML is the only one that makes use of this right now (SVG should prolly too) so whatever you want works
19:53
<Hixie>
annevk: k, thanks
19:55
<Hixie>
annevk: https://www.w3.org/Bugs/Public/show_bug.cgi?id=18941
19:55
<smaug____>
annevk: where should I file DOM4 bugs nowadays
19:56
<smaug____>
is w3c bugzilla still ok?
19:56
<annevk>
smaug____: yeah works for me
19:56
<annevk>
I should maybe get a component inside the WHATWG product but so far nobody in the WebApps WG has complained
20:00
<zcorpan>
[tm]: ^
21:35
<Hixie>
who's in charge of cssom these days, do they have a bugzilla account?
21:38
<jgraham>
Supposedly Glenn Adams
21:40
<Hixie>
k
21:41
<jgraham>
http://dvcs.w3.org/hg/csswg/log/911a391cb16c/cssom/cssom-source suggests that it might even be slightly true
21:50
<Hixie>
anne know if apple is on board with Opus as an audio codec?
21:52
<Hixie>
hober: ^
22:00
jgraham
has no idea how to get <input type=file multiple> to expose paths
22:00
<jgraham>
What am I missing?
22:01
<Hixie>
it doesn't expose paths
22:01
<jgraham>
*names* are rather easy
22:01
<Hixie>
(i suppose a UA could include the path in the name, if so inclined)
22:01
<Hixie>
oh you mean how _could_ we do it?
22:01
<Hixie>
it'd be easy
22:01
<Hixie>
just make the filenames foo/bar/baz.quux
22:01
<jgraham>
No, no
22:01
<Hixie>
or C:\foo\bar\baz.quux
22:02
<jgraham>
"File names may contain partial paths, e.g. in the case that a user has selected an entire directory hierarchy. Path components should be separated from each other using U+005C REVERSE SOLIDUS character (\)."
22:02
<Hixie>
ah, there you go, the spec even says it already
22:02
<Hixie>
cool
22:02
<jgraham>
I don't want it to expose paths
22:02
<jgraham>
:)
22:02
<jgraham>
Darin claimed that it already does
22:02
<Hixie>
whether it does or not is entirely an implementation issue
22:03
<jgraham>
AFAICT this isn't actually the case
22:03
<Hixie>
what isn't
22:03
<jgraham>
That paths are exposed by this API
22:04
<Hixie>
they can be
22:04
<Hixie>
in theory
22:04
<Hixie>
whether implementations do so or not is up to them
22:06
<jgraham>
How would they expose it if they did?
22:06
<Hixie>
you just quoted the answer to your question
22:06
<Hixie>
i don't understand
22:06
<Hixie>
is it not clear?
22:08
<jgraham>
Not at all :)
22:09
<Hixie>
what's not clear
22:09
<Hixie>
i really don't understand the problem
22:09
<jgraham>
I don't see how the statement "file names may..." is related to the properties of the input element
22:09
<Hixie>
oh
22:10
<Hixie>
the file names it refers to are those that form part of hte /selected files/ list
22:10
<Hixie>
input.files returns a FileList of the /selected files/ list
22:10
<Hixie>
FileList is a list of File objects, which each have a name
22:10
<jgraham>
Right, but that explictly doesn't include a path
22:10
<Hixie>
they explicitly do, here
22:11
<Hixie>
well
22:11
<Hixie>
they don't include a path
22:11
<Hixie>
they just include a filename with backslashes
22:11
<jgraham>
What? The File API says they don't
22:11
<Hixie>
but you can interpret that as a path
22:11
<Hixie>
the HTML spec says they do
22:11
<jgraham>
"The name of the file; on getting, this must return the name of the file as a string. There are numerous file name variations on different systems; this is merely the name of the file, without path information"
22:11
<Hixie>
right
22:11
<Hixie>
hence why the HTML spec changes the name to include the path information
22:12
<jgraham>
Woah
22:12
<jgraham>
That is totally unclear
22:12
<jgraham>
Why is there a "may" level wilful violation of another spec that no UA actually violates?
22:12
<jgraham>
(afaict)
22:12
<Hixie>
so that people can expose paths
22:13
<jgraham>
But since no UA actually does that, it seems highly unclear and quite probably undesirable
22:13
<Hixie>
*shrug*
22:13
<Hixie>
i don't see why it's unclear
22:14
<Hixie>
nor undesireable
22:14
<jgraham>
It's unclear because usually when you are willfully violating other specs a) it says so explicitly and b) there is a compat. reqiurement
22:15
<Hixie>
it's not a wilful violation
22:16
<Hixie>
a wilful workaround maybe :-)
22:16
<jgraham>
Erm? If spec A says "you must not do X" and you say "you may do X" that sounds like a wilfull violation
22:17
<Hixie>
spec A doesn't say "you must not do X". It says "you must return A, and not return B" and HTML says "put B inside A"
22:17
<Hixie>
actually even that's not quite accurate
22:18
<jgraham>
Since "B inside A and return A" is a special case of "return B", that is just the same
22:18
<Hixie>
i read it differently but whatever
22:19
<Hixie>
i'm happy to change one or the other to avoid the appearance of disagreement
22:19
<jgraham>
OK. Well I should have been asleep some time ago. Seems like a good juncture to fix that.
23:13
<Hixie>
can anyone explain the purpose of this change? https://github.com/w3c/html/commit/37b3183567666a226d304c54d6ac5123ba564dcc
23:14
<Hixie>
hober?