00:03
<heycam>
w3.org down?
02:30
<paul_irish>
hober: can you help me out with a whatwg wiki account? username "Paul Irish"
02:50
<MikeSmith>
paul_irish: will try to see if I have perms
02:51
<paul_irish>
you do
02:51
<MikeSmith>
heycam|away: entire MIT network is dropping off due to a routing problem
02:51
<MikeSmith>
paul_irish: ok gimme a minute
02:51
<paul_irish>
no rush! :)
02:55
<MikeSmith>
paul_irish: ok, you want the account created?
04:14
<paul_irish>
MikeSmith: yeah
04:19
<MikeSmith>
paul_irish: pusshed on the button on it a while back
04:19
<MikeSmith>
should be mail in your inbox
04:19
<MikeSmith>
sent to your gmail account
04:24
<paul_irish>
thx dude
09:01
<annevk>
If an interface has a constructor, does a derived interface necessarily need a constructor?
10:12
<annevk>
I had not realized UI Events was just some patching on top of DOM3 Events :/
10:27
<annevk>
smaug____: we could also do new EventTarget(eventParent)
10:27
<annevk>
smaug____: I think
10:28
<annevk>
smaug____: we could even consider always exposing the eventParent as a readonly attribute on EventTarget, but I'm not sure that's worth it
10:28
<hsivonen>
from the department of entirely unsurprising news: the IE9 mode in IE10 does not actually match IE9. For example, IE9 mode in IE10 accepts unprefixed 2D transforms but real IE9 does not (requires -ms-).
10:30
<annevk>
"House of cards"
10:30
<smaug____>
I think we need some kind of parent
10:31
<smaug____>
but having it in EventTarget might be a bit odd
10:31
<smaug____>
though...
10:31
<annevk>
smaug____: Why do we need to expose it explicitly?
10:31
<smaug____>
because it makes sense ?
10:32
<smaug____>
:)
10:32
<annevk>
smaug____: The best kind of argument
10:32
<smaug____>
I wonder if IDB might need eventParent
10:32
<smaug____>
need to ask sicking about that
10:33
<annevk>
Well I'm going to rewrite concept-event-dispatch to use an "event parent" concept at least to calculate the object chain before dispatching
10:33
<annevk>
So IDB and others can use that
10:34
<annevk>
But it does not need to be exposed. E.g. that for event purposes Window is the parent of Document is not exposed in such a way either at the moment
10:35
<smaug____>
it is
10:35
<smaug____>
.defaultView
10:36
<annevk>
That's not really convincing and was not introduced for that purpose at all...
11:13
<SimonSapin>
Hi annevk_. A few questions about http://encoding.spec.whatwg.org/#utf-8-decoder
11:14
<annevk>
SimonSapin: what about it?
11:14
<SimonSapin>
the algorithm "returns" one code point/EOF/error and is called repeatedly, right?
11:15
<SimonSapin>
when it says "emit" something, the steps stop there for this iteration
11:16
<SimonSapin>
"Let byte be the value at byte pointer." is that a copy, or does byte (potentially) change when byte pointer does?
11:24
<SimonSapin>
annevk: and I think that "Increase the byte pointer by one." is not at the right place, but I’m not sure what would fix it
11:25
<SimonSapin>
and it might be possible to avoid "Decrease the byte pointer by one." by moving the range checking earlier
11:26
<annevk>
SimonSapin: /byte/ is either a copy or a pointer, but if byte pointer changes byte does not
11:28
<annevk>
SimonSapin: I believe the algorithm is correct though
11:28
<annevk>
SimonSapin: but there's many ways to implement a utf-8 decoder
11:29
<SimonSapin>
annevk: it’s probably correct, but I don’t find it very clear
11:29
<SimonSapin>
it’s not obvious that you should not continue the steps after "emit"
11:30
<annevk>
SimonSapin: did you read section 4?
11:31
<annevk>
but yeah, I'm not entirely satisfied with how that works at the moment
11:31
<SimonSapin>
section 4 might be what I was missing
11:31
<annevk>
I did it mainly this way to reduce the overhead as it would be the same everywhere, but maybe self-contained algorithms would be better
11:32
<SimonSapin>
maybe just refer to section 4 every time
11:33
<SimonSapin>
annevk: anyway I started https://github.com/SimonSapin/rust-webencodings
11:33
<SimonSapin>
and learning rust along the way
11:33
<annevk>
oh cool
11:33
<annevk>
you should tell the Mozilla guys
11:34
<annevk>
they might wanna use it ;)
11:35
<SimonSapin>
I will, once it does a bit more
11:37
<annevk>
SimonSapin: fwiw, seems they already link to section 4
11:38
<annevk>
"The utf-8 decoder (decoder for utf-8) is:" the second "decoder" is a link back to what decoder actually means
11:38
<SimonSapin>
oh, missed that
11:47
<SimonSapin>
annevk: do you have encoding tests?
11:48
<annevk>
SimonSapin: I have http://dump.testsuite.org/encoding/single-byte-test.html
11:48
<annevk>
I've been meaning to convert that one to testharness.js to make Ms2ger happy
11:49
<annevk>
I'd need some information on whether using PHP is fine...
11:49
<darobin>
annevk: using PHP is fine
11:49
<darobin>
it just needs to get reviewed before being imported to the test server
11:49
<darobin>
but since the repo isn't synced yet, that's not an immediate problem
11:49
<annevk>
darobin: so the other problem with that test seems to be that it triggers the slow script dialog in Gecko
11:50
<annevk>
darobin: I could probably make it 30 or so static files instead generated with Python
11:51
<darobin>
annevk: doesn't trigger that here, though it did trigger my mental "There's something wrong here" dialog :)
11:51
<annevk>
although hmmm might require some .htaccess
11:51
<annevk>
heh
11:51
<annevk>
it fetches the same thing about 50 times or so sync
11:51
<annevk>
each time with a different header
11:52
<darobin>
hmmm, synchronous XHR, adding methods to a core prototype — full of lovely :)
11:55
<annevk>
darobin: well all that can be rewritten, the test was mostly a quick hack to figure a few things out
11:55
<annevk>
just not sure what the desired features are of a largish test
11:57
<darobin>
annevk: I know, I was just poking fun
11:57
<darobin>
annevk: I wonder, could this be turned into a bunch of distinct tests?
11:57
<darobin>
with testharness and all
11:57
<annevk>
yeah, as I said, I could generate them from Python
11:57
<darobin>
just so that it doesn't trigger any slowness warnings
11:57
<annevk>
or some other language
11:58
<darobin>
ah, ok that's what you meant — sorry, wasn't clear on that bit
11:58
<SimonSapin>
annevk: testing single bytes encoding based on their index is easy enough. I’m more interested in corner cases like how many replacement characters do you get for a given invalid UTF-8 sequence
11:58
<annevk>
SimonSapin: http://code.google.com/p/stringencoding/ might have tests
11:58
<darobin>
annevk: I don't have a strong opinion, but at the very least if it's going to be big and slow it's better if there's some sort of progress indication
11:59
<darobin>
even if it's just a <pre> listing the parts being tested
11:59
<annevk>
SimonSapin: agreed that those are more desirable
12:00
<darobin>
I reckon that making it xhr-async will cause it to spin the event loop often enough to avoid the slow script dialog
14:03
<MikeSmith>
annevk: did you before have a thing somewhere like http://www.lookout.net/test/url/url-liveview.html ?
14:03
<MikeSmith>
Chris says in his blog post he thought you did but he can't remember the address
14:04
<annevk>
http://dump.testsuite.org/url/inspect.html
14:04
<annevk>
I commented on his G+ post
14:05
<annevk>
How did you get the correct link to his tool? :)
14:05
<MikeSmith>
ah ok
14:05
<MikeSmith>
annevk: through URL surgery
14:05
<annevk>
heh
14:05
<MikeSmith>
fitting :)
14:06
<hsivonen>
http://www.delorie.com:81/some/url.txt is down. Does anyone know a substitute?
14:06
<annevk>
hsivonen: something like http://software.hixie.ch/utilities/cgi/test-tools/echo maybe?
14:07
<hsivonen>
annevk: thanks. not quite as pure but works for my purposes today
14:07
<annevk>
yeah, the raw dump stuff was nice
14:10
<hsivonen>
can “Browser Modes” 7 through 9 (not “Document Modes”) in IE10 be activated by means other than the dev tools?
14:10
<hsivonen>
that is, can an intranet admin activate them for particular sites?
14:11
<hsivonen>
oh awesome. IE10 doesn’t have application/xml in Accept
14:11
<hsivonen>
we should drop that
14:12
<hsivonen>
oh. IE9 doesn’t either
14:12
<hsivonen>
I thought it did
14:20
<annevk>
Heh https://twitter.com/aaronsw/status/140810339771097088 hadn't seen that before
14:27
<MikeSmith>
heh
14:32
<jesusruiz>
one question, WHATWG specification differs from the W3C specification?.
14:33
<jesusruiz>
I speak from the time of the final specification, the W3C that seem set for 2014.
14:34
<jesusruiz>
But I do not see the WHATWG definite date.
14:35
<hsivonen>
jesusruiz: It’s a living spec. There’s no end date.
14:37
<jesusruiz>
hsivonen thanks for the reply ;)
14:38
<jesusruiz>
What I have not entirely clear is, after a bit of separation that has had the WHATG and W3C specification that must be followed.
14:39
<jesusruiz>
I like the freedom of WHATG.
14:39
<jesusruiz>
WHATWG
14:40
<jesusruiz>
but if the specs are going to be something different, and not as a developer which follow.
14:41
<hsivonen>
jesusruiz: follow the WHATWG spec
14:41
<hsivonen>
jesusruiz: the W3C spec will be out of date by construction, since the W3C publishes snapshots
14:42
<jesusruiz>
hsivonen Yes, I feel more identified with this group. So I think you're right.
14:43
<jesusruiz>
It is an important reason you mention.
15:02
<Ms2ger>
annevk, I'm so glad you want to make me happy :)
15:03
<Ms2ger>
GPHemsley, sounds like work...
15:42
<Stevef>
hsivonen: HTML 5.1 is only a little behind and in front in some ways
16:06
<MikeSmith>
http://lists.webkit.org/pipermail/webkit-dev/2013-January/023326.html
16:07
<MikeSmith>
(Sent from)
16:09
<Ms2ger>
Oh really
17:25
<dglazkov>
good morning, Whatwg!
17:41
<hober>
paul_irish: sorry, just seeing your msg now. thankfully MikeSmith got to you first. :)
17:51
<paul_irish>
hober: yup! np thx
18:47
<stevefaulkner>
MikeSmith: can you give me the URL for the numarkup dev validator?
19:04
<Ms2ger>
"To me, it looks longer."
19:16
<Ms2ger>
Hear, hear: https://www.w3.org/Bugs/Public/show_bug.cgi?id=20665
19:27
<GPHemsley>
Ms2ger: I'm just saying, if somebody comes to you and asks, you can help them. ;)
19:27
<Ms2ger>
GPHemsley, I wish I didn't know that :)
19:28
<GPHemsley>
Ms2ger: If it makes you feel any better, there are a bunch of people in this channel who have that power.
19:29
<GPHemsley>
http://wiki.whatwg.org/index.php?title=Special:ListUsers&group=autoconfirmed
19:29
<Ms2ger>
Good to know
19:31
<GPHemsley>
(And I'm happy to add more people to that list, if anyone wants to make suggestions.)
19:32
<jesusruiz>
GPHemsley Do you only keep the wiki?
19:33
<GPHemsley>
jesusruiz: I think I'm chief in charge, but I believe there are others. Why?
19:33
<jesusruiz>
documentation of the specification, not keep you?
19:33
<GPHemsley>
jesusruiz: Specifications are generally not maintained via the wiki.
19:34
<jesusruiz>
I know that is not normal, but you are missing a possible translation of the documentation.
19:34
<jesusruiz>
a platform that could help translate into some languages.
19:35
<jesusruiz>
but well, neither does the W3C. So, time to time :)
19:35
<jesusruiz>
GPHemsley thanks for the info ;)
19:36
Ms2ger
has never seen the point of translating specifications
19:36
<GPHemsley>
jesusruiz: Well, I'm not in charge of that. You'll have to talk to Hixie or others with more influence on that.
19:37
<jesusruiz>
Well, for those who know English no problem. In fact, English is very common to know, but some people do not know the language.
19:37
<Ms2ger>
Do those people implement browsers?
19:38
<jesusruiz>
for example php is translating the documentation. It was a simple idea.
19:38
<Ms2ger>
Specification != documentation
19:39
<Ms2ger>
Translating documentation makes a lot of sense, but we're not in the documentation business
19:41
<jesusruiz>
Ms2ger Yeah, right. Anyway, I guess the documentation can be translated personally by others. Always keeping course and asking permission rights.
19:43
<jesusruiz>
I ask on the mailing list :) Thanks to all
21:08
<jgraham>
https://www.frederik-braun.com/thesis/ (via miketaylr)
21:09
<jgraham>
Also, it looks like crap in pdf.js, if someone wants to report that
21:11
<jgraham>
Also, 73 pages? That's crazy-short (doesn't mean that it's not good, of course)
22:40
<eighty4>
gsnedders: You should know this :) What version/revision of html5 was target deprecated? http://dev.w3.org/html5/markup/a.html#a.attrs.target "The target attribute on the a element was deprecated in a previous version of HTML, but is no longer deprecated"
22:46
<tantek>
4
22:55
<eighty4>
gsnedders: right… some kind of brain**** nevermind.
23:22
<annevk>
Translation of specifications is most definitely useful. Not everyone knows English growing up (or for that matter later in life) and yet might want to know the inner workings of the web.
23:23
<annevk>
We lose out on massive amounts of feedback from the Japanese community alone due to the language barrier as far as I can tell.
23:32
<zewt>
massive amounts of useless feedback about mistranslations
23:43
<Hixie>
abarth: could you chat with Nasko and decide if Chrome wants or does not want cross-origin seamless iframes?
23:50
<jamesr>
Hixie, woah
23:52
<Hixie>
woah?
23:54
<annevk>
zewt: I think there would be some benefits to not having this huge Western bias, but it's hard to proof. Fortunately to some extent things are being translated (e.g. by myakura) and bits of feedback do come through.
23:55
<annevk>
Hixie: cross-origin <iframe seamless> without events is still a useful feature I think
23:55
<annevk>
Hixie: I thought it was pretty clear that introducing new synchronous cross-origin object access was a no-go
23:56
<Hixie>
i don't have this topic paged in right now
23:56
<Hixie>
so i've no idea what you're saying. :-)
23:57
<annevk>
Hixie: maybe buy some more RAM at Fry's?
23:57
<Hixie>
they don't sell ram compatible with my motherboard :-(
23:57
<annevk>
man, sad times