00:00
<Hixie>
SimonSapin: yes, very asynchronously. but it shouldn't give you that error, because it has a separate idea of what version is current, iirc...
00:03
<Hixie>
woah
00:03
<Hixie>
weird
00:03
<Hixie>
what's going on here
00:05
<Hixie>
oh aaaaanneeee
00:07
<Hixie>
annevk: so uh
00:07
<Hixie>
annevk: for some reason the multipage splitter is splitting an old version?
00:23
<annevk>
hmm
00:23
<annevk>
maybe stuff got bogus again
00:25
<annevk>
svn: Checksum mismatch while updating 'source'; expected: '7d276e7c114c0dff2072062753a37225', actual: '296d8957f3f9fce92805f2bd67155ca0'
00:25
<annevk>
svn: Delta source ended unexpectedly
00:29
<annevk>
Hixie: I also get
00:29
<Hixie>
annevk: it might be worth putting in some defensive code in case i call you twice in parallel? maybe that broke it?
00:29
<annevk>
warning: can't find target for #CSSGC
00:29
<annevk>
warning: can't find target for #CSSWM
00:29
<Hixie>
those are fixed on my side
00:34
<annevk>
Fixed using http://stackoverflow.com/questions/10352934/svn-checksum-mismatch-while-updating
00:34
<annevk>
I think it's all good now
00:34
<Hixie>
cool, thanks
01:18
<a-ja>
Hixie: ping
02:15
<Hixie>
a-ja: pong
04:52
<MikeSmith>
Hi Hixie
04:52
<MikeSmith>
will take a look at the e-mail about the registry stuff today
05:57
<MikeSmith>
Hixie: haven't read the message about the registry yet but as far as the complexity, from the valiator side at least we there's no complexity needed as far as the set of data
05:57
<MikeSmith>
because the only thing the validator needs to consume is the list of valid names, and that's it
05:59
<MikeSmith>
whatever other data that people people add as part of registering a name, the validator doesn't need to see or do anything with
06:01
<MikeSmith>
so as far as I can see all we need exposed to us for the validator side is a space-separated list of the names
06:03
<MikeSmith>
no point in even having them as a JSON document even, because it would just be overkill since all the validator code would need to do is take the character stream with the names and split it up on the spaces
07:09
<foolip>
Hixie: yeah, the usage is shockingly low, and the intent to remove seems to have its required LGTMs, so here I go...
08:56
<Ms2ger>
foolip, what are you removing?
09:38
<jgraham>
SimonSapin: The point would be to make libraries in the virtualenv importable in random scripts, not just to make things in the venv/bin directory run (which I don't think they would without the venv being active anyway)
09:50
<Ms2ger>
jgraham, feel like reviewing? :)
11:19
<Ms2ger>
jgraham, r? https://github.com/w3c/testtwf-website/pull/211
11:46
<MikeSmith>
protip the help manage critic e-mail filtering: go into https://critic.hoppipolla.co.uk/config and add a "[critic]" or something prefix to all the email.subjectLine.* params
11:49
Ms2ger
waves at zcorpan
11:50
zcorpan
is not here
11:50
<Ms2ger>
Didn't mean to imply you were
11:51
<Ms2ger>
Will you be around for testtwf?
11:55
<zcorpan>
no
12:47
Ms2ger
really needs a victim for FileAPI reviews
13:10
<MikeSmith>
"As usual, I agree with everything Simon said."
13:58
<Ms2ger>
Which Simon?
14:03
<MikeSmith>
the best Simon
14:04
<annevk-cloud>
Hixie: registry proposal looks good to me
14:04
<MikeSmith>
https://groups.google.com/a/chromium.org/forum/#!msg/blink-dev/5sWUMC_d8Tg/gnxjyGTenvwJ
14:05
<annevk-cloud>
MikeSmith: would you not want JSON or some such for the data associated with meta names?
14:05
<MikeSmith>
annevk-cloud: because I will only use the names
14:05
<MikeSmith>
that's all we are using now
14:06
<MikeSmith>
and I have not plans to implement any of the other checking and reporting that Hixie is proposing
14:06
<annevk-cloud>
Hixie had some interesting suggestions as to how you could do more
14:06
<annevk-cloud>
I see
14:06
<MikeSmith>
right
14:07
<MikeSmith>
yeah I would be happy to add any of that if somebody else wants to provide a patch
14:08
<MikeSmith>
but I'm not convinced there's any problem that needs solving here, other than the very simple case of helping people catch unregistered/misspelled names
14:09
<MikeSmith>
and the problem of the latency between when people register new names and when Henri or me manage to get them added to the validator
14:09
<MikeSmith>
the only problem I've really had is I don't want to have to write code to scrape the wiki
14:10
<MikeSmith>
so I sure don't want to instead write even 10 times more code on the validator side to consume the rest of the registration data
14:11
<MikeSmith>
so all I want is a very simple API for retrieving just the list of registered names
14:13
<MikeSmith>
and as far as I can see the majority of the registered names are just junk that no (or almost no) existing applications actually consume
14:14
<MikeSmith>
so wI really have to wonder who are we really helping if we add yet more baroqueness around validator support for this feature
15:17
<Ms2ger>
If the element is mutable, the element's activation behavior is to do nothing.
15:17
<Ms2ger>
If the element is not mutable, it has no activation behavior.
15:17
Ms2ger
wonders what the difference is
15:20
<Ms2ger>
jgraham, want to sign off my changes in https://critic.hoppipolla.co.uk/r/642?
16:55
<MikeSmith>
Ms2ger: are the scripts that generate the canvas tests now working as expected again? can we close https://github.com/w3c/web-platform-tests/issues/215 ?
16:55
<MikeSmith>
Ms2ger: nm
16:55
<MikeSmith>
plh was asking me about the same thing
16:56
<MikeSmith>
I see you replied already
16:56
<Ms2ger>
He was?
16:57
<MikeSmith>
oh wait
16:57
<MikeSmith>
different issue
16:57
<MikeSmith>
I was thinking he asked you on #testing but I see now that was about #105
16:57
<MikeSmith>
ok so can we close https://github.com/w3c/web-platform-tests/issues/215 ?
16:58
<Ms2ger>
Yep
16:58
Ms2ger
does that
17:02
<MikeSmith>
thanks
17:18
<Ms2ger>
annevk-cloud, any reason not to put the TextEncoder/Decoder tests into w-p-t rather than https://code.google.com/p/stringencoding?
17:40
<Hixie>
MikeSmith: the proposal isn't gonna work if you only implement one tenth of it :-)
17:59
<MikeSmith>
Hixie: I'll reply to the message when I have more time but basically the only existing problem that I can so far agree is in need of solving is the part about having an actual API to expose the name values so that we can have the validator use that instead of us needing to hardcode the names in the validator sources
18:00
<MikeSmith>
I'm not opposed to the other parts of the proposal if somebody were to care enough to spend time implementing them
18:06
<SpiritDrift>
join #webgl
18:11
<a-ja>
MikeSmith: newly-allowed ARIA values being added to validator?
18:12
<a-ja>
(8409 and 8412 in tracker)
18:12
<MikeSmith>
a-ja: I think we already have them in the validator
18:12
<MikeSmith>
if not please files bugs
18:13
<a-ja>
oh cool
18:16
<Hixie>
MikeSmith: makes sense
18:17
<Hixie>
MikeSmith: if the only thing we do is have a formal way to update validators, though, without e.g. the registration system, or the advice paragraphs on obsolete but known values, then we can build a much simpler system
18:18
<a-ja>
Hixie: see my message 4 or 5 days ago re tracker 8409?....bad? data-x values for group
18:19
<zewt>
grr @ python and its silly unhashable dicts
18:19
<Hixie>
a-ja: message where?
18:20
<a-ja>
<code data-x="attr-aria-role-directory">group</code>,
18:20
<a-ja>
2 occurrences
18:20
<Hixie>
isn't that fixed already?
18:20
<Hixie>
i think i fixed it when you first mentioned it
18:21
<a-ja>
maybe locally?
18:21
<a-ja>
just hadn't seen another checkin changing it
18:21
<Hixie>
i find no occurrences of /role-directory">group/ in the spec
18:21
<Hixie>
i probably fixed it in some other checkin
18:21
<a-ja>
k...cool
18:21
<Hixie>
i don't bother with dedicated checkins for minor stuff like that, i usually fix it in passing while i'm doing a bigger thing
18:22
<Hixie>
(thanks for noticing, btw)
18:22
<Hixie>
ok, gotta go, bbiab
18:22
<Hixie>
(noticing the typo, i mean.)
18:22
<Ms2ger>
http://html5.org/tools/web-apps-tracker?from=8409&to=8410
18:22
<a-ja>
btw...just saw un-prefixed border-box land in moz inbound
18:23
<a-ja>
(10yo bug)
18:23
<a-ja>
Ms2ger: that'd be rework of your patch
18:24
<Ms2ger>
Yep
18:24
<Ms2ger>
It even landed with my name on it :)
18:24
<a-ja>
noticed that :)
18:25
<a-ja>
Ms2ger: now uplift that sucker!
18:25
<a-ja>
if possible
18:26
<Ms2ger>
It's taken ten years, it can wait 6 more weeks :)
18:26
a-ja
know it's likely too late in cycle
18:27
<a-ja>
i'm just tired of it polluting my stylesheets
18:28
<Ms2ger>
Yeah, uplift is Monday after next, and I wouldn't be surprised by some breakage
18:40
<Ms2ger>
foolip, how about Window.captureEvents?
18:42
<a-ja>
Ms2ger: fyi, just submitted feedback to caniuse.com
18:42
<Ms2ger>
Thanks :)
19:05
<satazor>
Hello guys, I'm a developer of the recently released demo of http://baboom.com which is a new music service. We are using the new WebAudio API to provide FLAC playback, though we are having a few issues with it, and hopefully some of you guys can enlighten me
19:07
<satazor>
we are decoding FLAC files and using createScriptProcessor to fill raw pcm audio data, but the fact that the refill callback is on the UI thread has a major issue, which is that when browser is doing some heavy stuff, the audio data does not called in time, and that causes audio glitches
19:07
<MikeSmith>
satazor: that API isn't discussed here so much. I'm not sure who here is very familiar with it
19:07
<MikeSmith>
maybe roc
19:08
<satazor>
roc isn't here, or is under other username?
19:09
<MikeSmith>
satazor: no he'd be using that nick
19:10
<satazor>
I see, anyone else that could give me an hand?
19:12
<MikeSmith>
satazor: not that I can think of
19:12
<MikeSmith>
other options are posting to http://lists.w3.org/Archives/Public/public-audio/
19:12
<MikeSmith>
or to https://github.com/WebAudio/web-audio-api/issues
19:13
<MikeSmith>
which I realize aren't real-time discussion but I don't think the audio people are around on IRC much
19:17
<MikeSmith>
satazor: you could try #audio on irc.w3.org
19:17
<satazor>
alright thanks for all these resources
20:48
<annevk-cloud>
Ms2ger: ask jsbell
20:48
<Ms2ger>
!summon jsbell