00:59
<wirepair>
hsivonen : you about?
04:17
<falken>
Hixie: if possible, could you see https://www.w3.org/Bugs/Public/show_bug.cgi?id=20418 (related to the objections about <dialog> being magical)
06:07
<Hixie>
falken: yeah, saw that one earlier. wasn't sure what the implications were.
06:07
<Hixie>
falken: treating modal and non-modal dialogs the same doesn't seem like an obviously right thing to do
06:09
<Hixie>
falken: (abs pos on the "top" layer should hopefully work via nesting? not sure exactly.)
06:13
<falken>
Hixie: not sure what you mean by working via nesting. if it's in the top layer, abs pos will just work because cb is ICB
06:14
<Hixie>
i mean abs posed boxes in the boxes that are themeselves in the top layer
06:14
<Hixie>
(isn't z-index independent of containing block?)
06:14
<Hixie>
(i though z-index only depended on stacking contexts)
06:15
<falken>
right i think in theory z-index only depends on stacking context, but in practice we contruct the layers by traversing the render tree, and the easiest way to make something's cb the icb is by moving its renderer to be a child of the root
06:18
<falken>
I'm not sure what you mean though. Abs pos box of a box in the top layer just works. But if the abs pos box itself is in the top layer, it will be positioned relative to ICB, regardless of whether it's a child of something in the top layer
06:29
<MikeSmith>
hey good afternoon falken
06:29
<falken>
hi MikeSmith
06:31
<MikeSmith>
falken: I know you are busy with <dialog> at the moment but I wanted to ask you if you been thinking about Streams, or following the discussions about it
06:32
<falken>
MikeSmith: not so much... what's up?
06:35
<tyoshino_>
MikeSmith: hi Mike
06:35
<MikeSmith>
hey hey tyoshino_
06:35
<MikeSmith>
didn't know you were on here
06:36
<tyoshino_>
sorry. i wasn't watching
06:36
<tyoshino_>
Streams API?
06:37
<MikeSmith>
falken: we have a bit of a split between tyoshino_ Streams API draft and one the Domenic_ is working on
06:37
<MikeSmith>
so tyoshino_ it would be nice if we can try to get agreement on a unified spec this month or next, say
06:38
<tyoshino_>
yes. we're also talking with Marcos of Mozilla about that
06:38
<MikeSmith>
cool
06:39
<MikeSmith>
woudl be great to chat with Domenic_ directly too
06:39
<MikeSmith>
here's on this channel most every day at US time
06:39
<MikeSmith>
tyoshino_: you planning a visit to US any time soon?
06:40
<tyoshino_>
yeah. i had some chat with him and smurg regarding Promise vs. event in December
06:41
<tyoshino_>
no for now. but i can be online in US day time if needed
06:43
<Hixie>
falken: i guess i just don't understand the problem you're trying to solve in the bug in sufficient detail.
06:43
<Hixie>
falken: gotta go to bed though, and i'm ~afk until monday.
06:44
<falken>
Hixie: OK thanks I'll try to add more detail in the meantime
06:51
<MikeSmith>
tyoshino_: maybe you could stay on some evening and ping Domenic_ to chat here real time
06:51
<MikeSmith>
tyoshino_: alternatively, start a discussion on the whatwg list or public-webapps
06:51
<MikeSmith>
discussion about resolving the spec difference
06:51
<MikeSmith>
*differences
06:53
<tyoshino_>
thanks for suggestions. we've already asked Marcos for arrangement of such a place. so, i'd like to wait for his response first.
07:00
<MikeSmith>
tyoshino_: sounds good
08:27
<hsivonen>
wirepair: pong
08:29
<hsivonen>
MikeSmith: what's my bad analogy? The software patents one? That TimBL countered with the ice cream truck analogy?
08:30
<wirepair>
hey hsivonen
08:30
<wirepair>
so i want to use your html5 parser to basically hook each tokenizer/location transition
08:30
<wirepair>
was wondering if there were any examples anywhere
08:32
<MikeSmith>
hsivonen: yeah, it was your software-patents analogy (which to be clear I didn't think was a bad analogy at all)
08:32
<hsivonen>
wirepair: there might be a Java example in NetBeans, but I haven't looked
08:32
<hsivonen>
wirepair: there's a C++ example in Gecko's View Source highlighting
08:33
<wirepair>
hmm, any hints as to which objects i should be looking at?
08:34
<wirepair>
or if you can point me to the C++ example i might be able to figure it out
08:35
<hsivonen>
wirepair: in Java, TransitionHandler and ErrorReportingTokenizer
08:36
<hsivonen>
wirepair: in C++, follow identifiers from https://mxr.mozilla.org/mozilla-central/source/parser/html/nsHtml5Highlighter.cpp to the tokenizer
08:37
<wirepair>
ok cool, i'll try to figure it out
08:37
<wirepair>
thanks!
08:42
<hsivonen>
hmm. http://krijnhoetmer.nl/irc-logs/whatwg/20140109 shows more context than /lastlog. I should read the full log instead of my /lastlog
08:43
<hsivonen>
Domenic_: is https://hsivonen.fi/eme/ at all like what you are looking for?
08:46
hsivonen
wonders if anyone has registered domain names with Hong Kong-only ideographs in them and how those domain names are working out in terms of users finding the sites
09:00
<MikeSmith>
hsivonen: hmm yeah people should read that writeup about EME
09:06
hsivonen
sees a TAG thread about a new HTTP response code without any of the reformers posting to the thread
09:06
<hsivonen>
I don't understand the use case for the new code
09:13
<zcorpan_>
who needs use cases
09:17
<zcorpan_>
hsivonen: sounds like it's a redirect but without another RTT
09:24
<hsivonen>
who is "dka" and who is "twirl" in TAG minutes?
09:25
<hsivonen>
Is dka Daniel K. Appelquist?
09:25
<hsivonen>
is twirl Sergey?
09:35
<MikeSmith>
dka ia Dan
09:36
<MikeSmith>
I don't know who twirl is
09:49
<hsivonen>
zcorpan_: it seems to me the motivating use case for the new response code has more to do with SemWeb than with avoiding one RTT
09:50
hsivonen
decides not to care about the new response code
09:54
hsivonen
sees SPARQL mentioned in the TAG Push API discussion. Uh oh.
10:30
<darobin>
annevk: so, got drunk all night to celebrate your liberation?
10:30
<annevk>
darobin: three more weeks no?
10:30
<annevk>
darobin: also, I kind of agreed to come one last time in April
10:30
<darobin>
oh, you're probably right
10:31
<darobin>
that's outrageous
10:31
darobin
starts a FREE ANNE campaign
10:31
<hsivonen>
darobin: how's your freedom at the HTML WG?
10:31
<darobin>
hsivonen: once a week MikeSmith brings me a nice sandwich down in the mines
10:31
<annevk>
But after that I'll have at least a month worth of extra time each year, which seems pretty good.
10:32
<annevk>
Haha, shit would fall apart if not for MikeSmith
10:32
<hsivonen>
how to get an extra month each year: run for the TAG, get elected, don't run again
10:32
<darobin>
I certainly would
10:32
<annevk>
hsivonen: are you trying to get quoted by @w3cmemes?
10:33
<darobin>
MikeSmith is the Care Bear from the Pit of Despair
10:33
<hsivonen>
annevk: not particularly
10:36
hsivonen
is still trying to form an opinion on whether to implement the Encoding Standard for GB18030 in Gecko or to get cold feet and suggest retaining gbk as a distinct encoding in order to submit NCRs as before
10:55
<annevk>
hsivonen: the question is what kind of gbk would you implement, would you have it be a different two-byte table from gb18030?
10:55
<annevk>
hsivonen: that would simplify the least
11:04
<Ms2ger>
hsivonen, hmm, clearly I should run for the TAG, then...
11:22
<darobin>
Ms2ger: I reckon you're more the AB type
11:22
<gsnedders>
SimonSapin: Seems I don't know how function comparison works :)
11:22
<darobin>
there's an election later this year, I'll start rallying around your candidacy
11:22
<SimonSapin>
gsnedders: it’s not about comparaison, it’s that instance.method is a wrapper (bound method) over class.method, in order to add the "self" parameter
11:23
<SimonSapin>
so it’s not the same object
11:23
<gsnedders>
SimonSapin: Oh, right
11:25
<SimonSapin>
(see "descriptors" for how this all works)
11:46
<smola>
annevk: sorry for the bug-noise
11:46
<annevk>
no worries
11:47
<annevk>
the review is great
12:59
<smaug____>
is DOMStringList deprecated
13:00
<smaug____>
DOM3 has it
13:00
<smaug____>
but DOM doesn't say anything about it
13:03
<annevk>
yes we killed it
13:03
<annevk>
I guess I should put it on the obsolete list indeed
13:03
<annevk>
forgot to do that I think
13:04
<smaug____>
annevk: IDB still refers to DOMStringList
13:05
<smaug____>
annevk: what should be used in APIs there days then?
13:05
<annevk>
smaug____: https://github.com/whatwg/dom/commit/8f73159d4a9e2e1a002f4eb1dfd4df3f645e9575
13:06
<annevk>
An Array
13:06
smaug____
tries to recall what was decided about [] and sequence
13:06
<smaug____>
annevk: DOMString[] ?
13:07
<annevk>
smaug____: well, as bz will tell you we're not quite clear [] is the right thing
13:07
<smaug____>
I thought that got resolved
13:08
<annevk>
I think there are some proposals, not sure if anything stuck yet
13:08
<smaug____>
https://bugzilla.mozilla.org/show_bug.cgi?id=942631
13:12
<annevk>
cool
13:19
<hsivonen>
annevk: is the Encoding Standard already compatible with the HK govt standard?
13:20
hsivonen
was too lazy to compare the Encoding Standard with the HK govt character list
13:20
<annevk>
hsivonen: I'd have to check; foolip did most of the work on big5
13:20
<annevk>
hsivonen: I can go through that bug in an hour or so
13:20
<hsivonen>
annevk: cool
13:56
<Ms2ger>
annevk, do you get critic email?
13:56
<annevk>
no
13:59
<jgraham>
annevk: Boo
13:59
<Ms2ger>
Then please look at https://critic.hoppipolla.co.uk/r/29 manually :)
14:00
<hsivonen>
Hixie: are you planning on adopting this change to the parsing algorithm: https://github.com/w3c/html/commit/e2ddb663fd04803d2be7f16026e2117ced167c01#diff-36cd38f49b9afa08222c0dc9ebfe35ebL98664 ?
14:02
<hsivonen>
darobin: I'm pretty worried abou the precedent of the W3C forking the parsing algorithm. The HTML WG is supposed to bikeshed stuff that doesn't affect implementations.
14:04
<darobin>
hsivonen: if implementers back out of it then it'll get yanked of course
14:04
<Ms2ger>
Of course?
14:04
<darobin>
Ms2ger: well, so long as I have commit rights :)
14:05
<darobin>
hsivonen: I'm all for sticking to unifying bikeshedding and all, but I have yet to find a user who doesn't agree that's the way to do this
14:05
<darobin>
also, it matches usage in content
14:06
<darobin>
(that I could get my hands on at least)
14:07
<Ms2ger>
Did you manage to convince Hixie you're right?
14:09
<darobin>
Ms2ger: no, as far as I know he just called it "overengineered"
14:10
<darobin>
but at this point it's Hixie on one side and pretty much everyone else who has an opinion on the other side
14:10
<jgraham>
Oh sweet jesus please don't fork the parser
14:10
<darobin>
jgraham: do not panic, there is plenty of time to fix this
14:10
<darobin>
it just takes everyone to chill :)
14:16
Ms2ger
assumes Hixie is right, then
14:23
<MikeSmith>
I'm chill already
14:23
<MikeSmith>
so basically everybody should just be more like me
14:24
<Ms2ger>
MikeSmith, it's true
14:24
<Ms2ger>
Every night before I go to bed, I reflect on how I can be more like MikeSmith
14:26
<Ms2ger>
zco... well dammit
14:41
<MikeSmith>
but anyway I agree that when somebody else actually takes time to work with others to understand a problem and use cases and then actually write a spec that addresses the problem and use cases he's learned about, it's right that we should all gang up together and beat that person down
14:42
<MikeSmith>
we must exercise zero tolerance for that kind of factionalist reactionary behavior in our party
14:44
<hsivonen>
(I'm not saying that the parsing algorithm change is a bad idea. I'm saying that the WHATWG and W3C parsing specs diverging is.)
14:46
<MikeSmith>
yeah understood
14:49
<MikeSmith>
hsivonen: and I'm not saying that the parsing algorithm change is a good idea. I'm not convinced myself that the ruby spec needs to change. But a lot of other people in Japan who've been considering this seems to believe different.
14:52
gsnedders
doesn't want the parser forked because it means we can get into arguments about whether html5lib supports HTML5 or HTML.
15:07
<odinho>
Ms2ger: Testing that FormData worked, -- you could just send it to some python echo script and read that back, no?
15:13
<hsivonen>
annevk: can you refresh my memory about what GB2312 means in IE?
15:13
<hsivonen>
annevk: is it a label of GBK? a distinct encoding that works identically to GBK? a different encoding?
15:19
<hsivonen>
decodes like GBK at least in IE
15:21
<annevk>
hsivonen: will test now
15:23
<hsivonen>
encodes as GBK, too
15:23
<annevk>
Ms2ger: ask again Monday
15:23
<hsivonen>
http://hsivonen.com/test/moz/gb2312.htm
15:24
<hsivonen>
now I'm less sure about aliasing all these to GB18030 again :-(
15:24
<annevk>
hsivonen: gbk becomes gb2312
15:24
<hsivonen>
annevk: as a label?
15:25
<annevk>
yeah
15:25
<annevk>
per http://dump.testsuite.org/encoding/label-test.html
15:25
<annevk>
I like how IE uppercases gb18030 but not gb2312
15:25
<annevk>
whereas other browsers lowercase the former
15:39
<Ms2ger>
odinho, I was asking easy :)
15:44
<annevk>
I wonder where foolip is
16:03
<odinho>
annevk: Vacation. Might be back tuesday, might not. :)
16:14
<jgraham>
odinho: Yes, I think logic agrees with you there
16:20
<annevk>
Ms2ger: answered critic questions
16:21
<jgraham>
Is anyone else not getting critic mail?
16:22
<annevk>
I configured it too, fwiw
16:22
<jgraham>
That's not really what I meant
16:23
<jgraham>
Mail delivery might be having problems
16:24
<Ms2ger>
jgraham, I just got an email for a review I did yesterday
16:24
<jgraham>
I just restarted the email service
16:31
<jgraham>
Oh, it didn't appreciate the 33Mb emails that removing 32MB.txt generated
16:37
<annevk>
hsivonen: is there anything more you need me to look into?
16:38
<annevk>
hsivonen: replied to the gbk / big5 bugs
16:38
<annevk>
Fixing some Fullscreen stuff today, URL & more Encoding next week I guess
16:39
<annevk>
And at some point I'll define xhr.responseURL
16:39
<Hixie>
hsivonen: given that there are zero use cases for that complexity (literally the only purpose for adding it is that it briefly aids doing some obscure ruby things while not all browsers support it), that it was developed without bothering to engage in the relevant bugs, and that it causes parser churn, no, i've no intention of doing anything with it.
16:39
<annevk>
As well as a URL member for response objects
16:39
<Hixie>
hsivonen: see https://bugzilla.mozilla.org/show_bug.cgi?id=33339
16:39
<annevk>
morning Hixie
16:41
<Hixie>
hsivonen: it's the perfect example of what happens when a hyper-focused group who thinks their feature is very important gets to make decisions without considering the overall picture.
16:41
<qrk>
like EME?
16:41
<Hixie>
not quite as bad as EME, but yeah
16:42
<qrk>
EME makes me cry.
16:47
<SteveF>
hsivonen: "The HTML WG is supposed to bikeshed stuff that doesn't affect implementations. " primarily but not solely :-)
16:47
<Ms2ger>
jgraham, ha
17:45
<smola>
annevk: "scheme:// does not trigger hierarchical URI parsing" - I've dropped that from my implementation, after testing thoroughly... a lot of things get broken
17:46
<annevk>
smola: relative to?
17:47
<annevk>
ah you updated the bug
17:47
<smola>
yep, but notably, ed2k links would generate parsing errors
17:48
<annevk>
Time and again, it the "generic syntax" is invented and fails in practice.
17:49
<annevk>
Not sure if anyone implemented the <html:template> element yet in XML, but that does it in for XML.
17:50
<annevk>
I remember there being more examples, but they are not coming to mind.
17:52
<smola>
yes, it seems generic syntax failed so much that there are even funny contradictions such as an RFC saying this: "MUST follow the Uniform Resource Identifiers (RFC 3986) syntax [RFC3986]
17:52
<smola>
and then goes on and defines them in a way incompatible with RFC 3986
17:52
<Hixie>
generic syntaxes are silly, i keep telling people but they keep saying "you should base it on json!" or whatnot
17:54
<annevk>
JSON is pretty great, but a) it doesn't have comments and b) it doesn't stream.
18:23
<jgraham>
JSON lacking comments is pretty terrible
18:23
<jgraham>
Not sure that AppCache helped the reputation of "just make up a custom syntax" though
18:23
<jgraham>
(I am not claiming that most of the problems were syntax related)
18:35
<annevk>
I don't really understand what would be wrong with using Chrome and being against DRM.
18:36
<annevk>
I'm against the UK Internet censoring. Do I now "puzzle" people and should I move countries?
18:36
<annevk>
Seems rather contrived.
18:37
<jgraham>
Well switching browsers away from one whose makers have actively promoted DRM is much easier than switching countries
18:45
<Jasper>
jgraham, I tend to just have a "comment" key or something in a dict.
18:46
<Jasper>
jgraham, though what bites me is that I can't leave a comma at the end of a dict, like I can with JS/Python.
18:46
<jgraham>
JSON has the right data model but terrible syntax
18:47
<Jasper>
jgraham, well, you do know why Crockford removed comments from JSON, right?
18:49
<jgraham>
I don't know what reason he gave. And my personal opinion of why isn't suitable for a family channel (or even this one)
18:50
<jgraham>
Comments should never be used to transmit anything meaningful. That is
18:50
<jgraham>
what JSON is for.
18:50
<jgraham>
That reason?
20:05
<Ms2ger>
jgraham, so how do I get your server running?
22:51
<jgraham>
Ms2ger: Check it out and python serve.py
22:51
<jgraham>
It's even in the README
22:51
<Ms2ger>
ImportError: No module named wptserve
22:53
<jgraham>
In other news, am I reading the spec right to think that the DOM <span><textarea>\x0A</textarea></span> will give \x0A for the innerHTML of the textarea but <textarea>\x0a\x0a</textarea> for the innerHTML of the span or the outerHTML of the textarea?
22:54
<jgraham>
Ms2ger: git submodule update --recursive
22:54
<Hixie>
i hope not
22:54
<Hixie>
oh wait, of the DOM...
22:54
<Ms2ger>
Nope, doesn't help
22:54
<Hixie>
sounds plausible
22:55
<jgraham>
Ms2ger: Do you have something in [wpt root]/tools/wptserve/ ?
22:56
<Ms2ger>
Nope
22:56
<jgraham>
Hixie: It sounds plausible but also seems mildly insane :) I'm not sure I'm going to argue to change the spec though since I would have to change my implementation ;)
22:57
<jgraham>
Ms2ger: Did you git submodule init?
22:57
<Ms2ger>
Okay, that seems to help
22:58
<jgraham>
Ms2ger: I really suggest the README ;) It also tells you how to configure /etc/hosts
22:59
<Ms2ger>
I guess I should :)
22:59
<gsnedders>
We have to configure /etc/hosts? Ewwwww.
22:59
<Hixie>
jgraham: yeah, looks like we don't check what the parent is for that particular weirdness
23:00
<Hixie>
jgraham: what do browsers do?
23:01
<jgraham>
Dunno, if only I had a testsuite
23:01
<jgraham>
Oh wait, I do!
23:01
<Ms2ger>
\o/
23:02
<Hixie>
O
23:02
<Hixie>
\
23:02
<Hixie>
d'oh
23:02
<Hixie>
my / was eaten
23:02
<Ms2ger>
Now I need to remember why I wanted to run the server
23:04
<jgraham>
Looks like Presto follows the spec but Gecko and Chrome don't add the extra newline in the case that you would expect them to
23:04
<jgraham>
I have no idea about IE
23:04
<gsnedders>
Damn it, jgraham. Making Presto break the web by following specs.
23:06
<Hixie>
(is there any reason to still be testing presto? i thought it had joined MacIE and Netscape...)
23:06
<jgraham>
Well I have it
23:06
<jgraham>
Which isn't true of the others
23:06
<Hixie>
k, just checking if i was wrong to ignore it :-)
23:06
<jgraham>
and is still the latest Opera for Linux
23:07
<jgraham>
And various still-shipping devices
23:07
<jgraham>
But basically no
23:07
<jgraham>
(but it occasionally makes me feel good about my life when Opera passes tests)
23:08
<Ms2ger>
jgraham, so do I need to click "prepare rebase"?
23:08
<jgraham>
(I mean not great, you understand, but slightly better than normal)
23:08
<jgraham>
Ms2ger: No
23:08
<jgraham>
For a tracking review that button should be hidden :)
23:09
<jgraham>
You do the rebase, then click "Rebase Review"
23:09
Ms2ger
tries, crosses fingers
23:09
<jgraham>
Then, when you fetch the branch, if there are more commits than you expected, you specify the parent by hand rather than using the default
23:12
<Ms2ger>
jgraham, that seemed to work, but now it's no longer tracking
23:12
<jgraham>
Ms2ger: Yup, you need to click enable tracking
23:12
<jgraham>
And then force refresh your browser for reasons that I don't understand
23:12
<Ms2ger>
Fascinating
23:13
<Ms2ger>
Okay, that's enough for tonight
23:13
<Ms2ger>
Thanks
23:13
<jgraham>
Well the force-refresh is only to update the display
23:13
<jgraham>
It reenables the tracking either way
23:13
<gsnedders>
So, what, the devices SDK seems to have just at CES officially moved to Chromium. So I guess there are going to be projects already underway, yet to ship, based on Presto.
23:14
<Ms2ger>
Yeah, I suspected as much
23:14
<Ms2ger>
"1 minutes ago"
23:14
<jgraham>
Hixie: So, if Chrome and Fx aren't following the spec here is it worth changing the spec?
23:15
<jgraham>
To just never add the extra newline
23:24
<odinho>
gsnedders: this is true.
23:24
<odinho>
We are shipping new products based on presto still. :)
23:24
<odinho>
Or, devicessdk-ppl are, that is.
23:25
<Hixie>
jgraham: never adding the newline leads to data loss.
23:25
<Hixie>
jgraham: so that seems like a bad idea.
23:25
<Hixie>
jgraham: but if they won't change...
23:25
<jgraham>
Well they haven't in N years
23:25
<gsnedders>
odinho: Including stuff beyond what shipped in dsk 12.1x or not?
23:26
<jgraham>
gsnedders: It would be very surprising if the configuration of TVSDK was just like that of desktop
23:26
<jgraham>
at least it never has been in the past
23:28
<gsnedders>
jgraham: This is true. But they were always wary of being the first to ship anything… as everyone was, really. And in this case they would be — and the last.
23:28
<odinho>
gsnedders: I think so. Desktop was lagging faaar behind core, as I'm sure you remember. And I think they are at newer core integrations.
23:28
<gsnedders>
odinho: I'm well aware.
23:29
<odinho>
I know that the eternal bugs-branch is the main master now, and it seems to be getting patches still.
23:29
<odinho>
So I think it's just stabilizing like crazy. Probably the highest quality it has ever had :P No churn.