00:23
<GPHemsley>
Hixie: Hmm... not easily. I don't think we can do it from within the normal interface, but we can probably do it by hacking the template/theme somehow.
00:26
<GPHemsley>
Hixie: Hmm, actually, we might be able to do it with some custom CSS.
00:26
<GPHemsley>
annevk: Were there other pages that you didn't want to see the anonnotice on?
00:33
<annevk>
GPHemsley: on the wiki? All of them I think
00:33
<annevk>
GPHemsley: you solved a minor problem with something far too heavyweight in my opinion
00:34
<GPHemsley>
annevk: Did you have an alternative solution?
00:34
<annevk>
GPHemsley: not sure how many page views the thing gets from non logged in users though, but whenever I visit it on my phone I dislike it
00:35
<annevk>
GPHemsley: maybe include the account banner on the registry pages
00:35
<annevk>
GPHemsley: the bit about specification discussions does not seem warranted
00:35
<annevk>
GPHemsley: we can just remind individuals of that fact every now and then
00:35
<annevk>
GPHemsley: or put in the FAQ
00:39
<GPHemsley>
Hixie: Gone from the FAQ page now. If there are any other pages where you want it gone, edit MediaWiki:Common.css (and remember to clear the cache by clicking on the clock in the upper right corner of the page in question).
00:40
<GPHemsley>
annevk: Everything in the anonnotice was done in response to something.
00:40
<annevk>
GPHemsley: I believe you, but I'm saying that boilerplate is not the right solution
00:41
<annevk>
GPHemsley: and that the problem might not be as big as you make it out to be
00:41
<annevk>
GPHemsley: the wiki is more read than written
00:42
GPHemsley
shrugs
00:42
<GPHemsley>
I'll revisit it at some point.
01:54
<rniwa>
sicking: yt?
01:55
<sicking>
rniwa: yup
01:55
<rniwa>
sicking: I just replied on your response to custom elements reaching CR
01:55
<rniwa>
sicking: as I mentioned in the email, we share the same concern you expressed
01:56
<rniwa>
sicking: in particular, it seems strange to move the custom elements spec into CR without moving other parts of the web components at least into the last call
01:56
<sicking>
agreed
01:57
<rniwa>
sicking: it might make sense to restrict the shadow DOM to only custom elements for now
01:57
<rniwa>
sicking: so that we don't have to resolve all the issues associated with attaching shadow DOM onto existing builtin html elements
01:57
<rniwa>
such as form controls, img, etc... that has been tricky to spec
01:57
<sicking>
don't know, i haven't followed this in enough detail
01:57
<rniwa>
hayato_: ping
01:58
<rniwa>
sicking: i think the problem, as you correctly pointed out, is that other vendors haven't had enough time to review the specifications :(
01:58
<rniwa>
sicking: i sympathize with google in that they've been working on this stuff for years
01:58
<rniwa>
but it'll be really unfortunate if we end up not being able to implement what's been spec'ed at the end of the day
01:58
<hayato_>
rniwa: y
01:59
<sicking>
yup
01:59
<rniwa>
hayato_: am I correct in saying that the reason shadow DOM spec. is having a hard to move forward
01:59
<rniwa>
hayato_: is because we can't really spec what happens with replaced elements, etc...?
02:00
<sicking>
rniwa: Is Dave Hyatt involved in the shadow DOM stuff at all from your side? It would be great to have his input
02:00
<rniwa>
hayato_: as far as I know, the behavior of shadow DOM is very well defined if we restrict ourselves to custom elements. am I wrong on this?
02:00
<rniwa>
sicking: unfortunately not that much at this point :(
02:00
<rniwa>
sicking: but I'm working with Sam and Maciej
02:00
<hayato_>
rniwa: let me catch up. what topic are you discussing? shadow dom on built-in elements?
02:00
<rniwa>
sicking: and they're also familiar with XBL so hopefully I've got it covered
02:01
<rniwa>
hayato_: yeah
02:01
<rniwa>
hayato_: so there is a concern sicking raised
02:01
<rniwa>
hayato_: which is that other browser vendors haven't had a time to review custom elements
02:01
<rniwa>
hayato_: and we have a concern that moving custom elements spec. forward without moving the rest of web components forward might be risky
02:01
<rniwa>
hayato_: in that we might end up having to "fix" custom elements later
02:02
<hayato_>
rniwa: there is a lot of unclear issues on shadow dom for built-in elements.
02:02
<rniwa>
hayato_: yeah
02:02
<rniwa>
hayato_: so why can't we punt that for now and move with the rest of the spec?
02:02
<hayato_>
rniwa: and there is no strong demand for that.
02:02
<rniwa>
hayato_: as far as I understand it, there is a tremendous value in supporting shadow DOM for custom elements
02:02
<rniwa>
hayato_: right
02:03
<rniwa>
hayato_: and extending the spec to support attaching shadow DOM on bulletin elements is a natural extension we can add later
02:03
<rniwa>
hayato_: so we can spec. the shadow DOM for custom elements so that we wouldn't block custom elements forever
02:03
<rniwa>
hayato_: does that make sense?
02:03
<rniwa>
sicking: ^
02:04
<sicking>
like I said, i haven't looked at the shadow DOM spec enough to have an opinion on this
02:04
<rniwa>
sicking: okay.
02:04
<rniwa>
sicking: will you be fine with spec'ing shadow DOM just for custom elements as v1 though?
02:04
<hayato_>
spec is here: http://w3c.github.io/webcomponents/spec/shadow/#html-elements-and-their-shadow-trees
02:04
<sicking>
i'll defer to William Chen and mrbkap
02:04
<rniwa>
sicking: or do you think attaching shadow DOM on builtin shtml elements is an important use case?
02:04
<rniwa>
mrbkap: ping?
02:04
<hayato_>
rniwa: But this is outdated.
02:05
<sicking>
i definitely think that attaching shadow DOM to form controls is an important use case. It might be part of one solution for styling form controls
02:05
<sicking>
which is a sorely needed feature
02:05
<hayato_>
tkent-san has been working on supporting shadow dom for built-in element, such as input elements in blink.
02:06
<rniwa>
sicking: isn't that decorator though
02:08
<hayato_>
we thought that attaching Shadow DOM to built-in elements is important use case, but it turned out there is no much demands from developers.
02:08
<esprehn>
We can't restrict it to custom elements without making shadow dom and custom elements tightly coupled
02:08
<rniwa>
esprehn, hayato_: we can add some optional argument in document.register to create a shadow DOM, right?
02:09
<esprehn>
For example a browser may implement support for one before the other, shipping shadow dom shouldn't be blocked on custom elements
02:09
<rniwa>
esprehn: but nobody can ship shadow DOM because there is no spec.
02:09
<esprehn>
It had a spec, what do you mean?
02:09
<rniwa>
esprehn: attaching shadow DOM on builtin elements is undefined as is
02:10
<esprehn>
It is not undefined
02:10
<esprehn>
No more than attaching it to a div or a span, you mean specific things like input
02:10
<hayato_>
rniwa: http://w3c.github.io/webcomponents/spec/shadow/#html-elements-and-their-shadow-trees is trying to address that.
02:10
<rniwa>
esprehn: input element is a builtin element
02:11
<esprehn>
So is div, but shadow dom works fine with it
02:11
<rniwa>
esprehn: for the spec to have a defined behavior for builtin elements, it needs to define behaviors of attaching shadow DOM on every single element we have
02:11
<rniwa>
esprehn: including SVG and MathML elements
02:11
<esprehn>
Or it could allow it only only div, span, td, etc for now and we could expand it later
02:11
<rniwa>
esprehn: that'll be fine
02:12
<rniwa>
esprehn: I think the problem here is that we can't move the spec. forward because there are too many edge cases to consider
02:12
<esprehn>
Restricting to only custom elements is too restrictive
02:12
<rniwa>
esprehn: maybe.
02:12
<rniwa>
esprehn: but what are use cases of using shadow DOM for other elements than custom elements?
02:13
<esprehn>
Wanting something special on a <td>
02:13
<rniwa>
esprehn: why don't you just put an element inside td then?
02:13
<esprehn>
I guess if you could <td is=foo> as custom
02:14
<esprehn>
That argument does not generalize
02:14
<esprehn>
We could eliminate all features :)
02:14
<rniwa>
esprehn: as I understand it, we've punted declarative syntax for custom elements for v2
02:14
<esprehn>
Yes
02:14
<rniwa>
esprehn: then I don't see why we want to address that use case in shadow DOM v1
02:14
<esprehn>
Huh?
02:14
<rniwa>
esprehn: <td is=foo>
02:15
<sicking>
rniwa: decorators would just be a way of attaching a shadow DOM. The actual rendering should still be done using shadow DOM
02:15
<rniwa>
sicking: sure
02:15
<esprehn>
Declarative custom elements and type extensions are totally unrelated
02:15
<rniwa>
sicking: but as I understand it, we haven't spec'ed that anywhere, right?
02:15
<rniwa>
sicking: so it seems natural for attaching shadow DOM to builtin elements and decorator to be spec'ed at the same time
02:16
<sicking>
rniwa: and in the meantime, being able to attach a shadow DOM to a form control, or a header/table/whatever is a great way of working around the lack of decorators
02:16
<esprehn>
Not having declarative syntax has no bearing on td is=foo
02:16
<sicking>
rniwa: maybe
02:16
<sicking>
rniwa: (the "maybe" being a reply to your latest comment)
02:17
<rniwa>
sicking: that's a good point.
02:17
<rniwa>
sicking: I'm not convinced that shadow DOM is the right way to address the problem of styling form controls though
02:17
<esprehn>
If we're concerned with complex widgets I'm fine with a whitelist
02:17
<rniwa>
sicking: it seems like we want something like what webkit exposes for form controls spec'ed
02:18
<sicking>
rniwa: what's that?
02:18
<esprehn>
Although the current idea (and what we implement) is that adding a ShadowRoot makes the native behavior go away
02:18
<sicking>
rniwa: see also my latest email to WHATWG
02:18
<rniwa>
sicking: these: http://trac.webkit.org/wiki/Styling%20Form%20Controls
02:18
<esprehn>
It turns into a <div> effectively
02:21
<sicking>
rniwa: yeah, something in that general direction.
02:21
<rniwa>
sicking: I think spec'ing these will be better than having authors replace builtin form controls
02:22
<rniwa>
sicking: and then can provide some form serialization mechanism
02:22
<esprehn>
Authors already replace the builtins
02:22
<esprehn>
The pseudo elements are not flexible enough
02:23
<sicking>
rniwa: authors need to be able to fully replace the styling though. So that you can turn a checkbox into a switch, or a <select> into a map
02:23
<rniwa>
sicking: why don't authors just write a custom element at that point?
02:24
<sicking>
rniwa: accessibility
02:24
<sicking>
rniwa: and you could get better platform integration
02:24
<rniwa>
sicking: but if authors are replacing it with a shadow DOM or whatever, then accessibility is somewhat broken unless they've explicitly set tabindex, etc... right?
02:24
<sicking>
rniwa: which is kind'a the same thing
02:24
<esprehn>
Or because they want the HTMLSelectElement API or search engines
02:25
<sicking>
rniwa: and having to re-implement the full form API, and keep it up-to-date as the web grows, seems annoying
02:25
<sicking>
rniwa: no. accessibility should still work with a different rendering
02:25
<rniwa>
esprehn: API on HTMLSelectElement wouldn't work because map isn't a select element.
02:26
<esprehn>
Yes it is
02:26
<sicking>
rniwa: i.e. a screenreader would still be able to tell the user that a checkbox is a checkbox and that it can be flipped, and let the user use voice commands to flip it, even if it's rendered as a switch
02:26
<rniwa>
sicking: how so? AT wouldn't know how to interact with the new UI.
02:26
<esprehn>
<select is=x-map> is indeed a select
02:26
<rniwa>
sicking: that is true although you can just use ARIA role for that.
02:26
<sicking>
rniwa: AT interacts with the DOM
02:26
<rniwa>
sicking: not really in webkit.
02:27
<sicking>
rniwa: AT should interact with the HTMLSelectElement even if there's a shadow DOM attached.
02:27
<esprehn>
Shadow dom just changes the presentation
02:28
<sicking>
yup
02:28
<sicking>
anyhow, time to head home
02:28
<rniwa>
sicking: night!
02:28
<sicking>
g'night!
02:29
<esprehn>
You create a renderblock instead of a renderselect, the select element API should continue working much like if the select was display none it continues working
02:29
<sicking>
rniwa: i think we agree on the central issue though. It seems too early to put these parts into CR
02:29
<rniwa>
sicking: yeah.
02:30
<rniwa>
esprehn: only if the implementation of map respected the changes made by that API and emulated the behavior
02:30
<rniwa>
esprehn: even then there are certain APIs that doesn't make any sense on maps
02:31
<rniwa>
e.g. what would changing "size" attribute mean on a map?
02:31
<rniwa>
or "multiple"
02:31
<rniwa>
would map respect "disabled" attribute?
02:31
<rniwa>
what if we added new attributes in the future like "disabled"?
05:21
<esprehn>
rniwa: what does changing size mean on a multi select on the iphone?
05:22
<esprehn>
rniwa: IIRC iOS shows only one option and shows a spinner in some native UI. The size attribute does nothing.
07:21
<marcosc>
Hixie: any chance you could comment on this thread? http://lists.w3.org/Archives/Public/public-webapps/2013OctDec/0725.html
07:22
<marcosc>
if we have to use JSON for application metadata... should it be in a <script> or <meta>/<link>?
07:23
<marcosc>
putting JSON into an attribute just feels really wrong to me
07:24
<marcosc>
if anyone else has an opinion, would really like to hear it
07:25
<marcosc>
Ms2ger: you're always full of kind things to say... http://lists.w3.org/Archives/Public/public-webapps/2013OctDec/0725.html
07:25
<Ms2ger>
Meh, manifests
07:27
<marcosc>
Ms2ger: you know you love it
07:28
<marcosc>
hsivonen: if possible, would really appreciate any feedback on http://lists.w3.org/Archives/Public/public-webapps/2013OctDec/0725.html
07:43
<hsivonen>
Ms2ger: we have whines about late metas. we might not have whines about double metas
07:43
<hsivonen>
marcosc: I'll take a look
07:44
<hsivonen>
for those who are participants in the HTML WG but don't actually read the mail: there's now a poll where, contrary to the usual mode of operation, the Chairs are counting votes: https://www.w3.org/2002/09/wbs/40318/polyglot-status-preference-poll/
07:45
<marcosc>
hsivonen: thanks! really appreciate any feedback
07:47
<Ms2ger>
Oh wait
07:47
<Ms2ger>
It looks like this might actually be sent as a/xhtml+mxl
07:47
<Ms2ger>
xml, even
07:58
<hsivonen>
marcosc: is there an explanation somewhere why the new manifest doesn't need to come as early in the page load as the old (Hixie-AppCache) manifest?
07:59
<hsivonen>
marcosc: my immediate feedback is: why aren't inline manifests, <link> and <meta> all too late?
07:59
<marcosc>
hsivonen: because you only need that stuff async... if you ever need it at all
07:59
<marcosc>
most people don't go to a website and bookmark it straight away
08:00
<hsivonen>
marcosc: so no part of the first page load ever goes through the NavigationController?
08:00
<marcosc>
hsivonen: no. there is no relationship between manifest and navcon
08:01
<marcosc>
not yet, anyway
08:01
<hsivonen>
I guess I need to read more about what this thing is before I comment further
08:01
<marcosc>
http://manifest.sysapps.org/
08:02
<hsivonen>
same as http://w3c.github.io/manifest/ apparently
08:02
<marcosc>
yep, same thing
08:02
<marcosc>
hsivonen: the long version is here http://w3c-webmob.github.io/installable-webapps/
08:03
<marcosc>
(not done yet... but tell the whole story)
08:03
<marcosc>
hsivonen: section 3 is basically what we are trying to enable: http://w3c-webmob.github.io/installable-webapps/#add-to-homescreen
08:25
<Ms2ger>
Lovely: https://code.google.com/p/chromium/issues/detail?id=325330
08:35
<rniwa>
Ms2ger: hi Ms2ger!
08:35
<rniwa>
Ms2ger: lol
08:35
<Ms2ger>
I see you cc'd me on a bug :)
08:35
<rniwa>
Ms2ger: yeah
08:36
<rniwa>
Ms2ger: XML fragment parsing bug :(
08:36
<rniwa>
Ms2ger: it's such an edge case that I initially thought it's a bug in the test
08:36
<rniwa>
Ms2ger: but then Simon convinced me that it wasn't
08:36
<Ms2ger>
Dammit Simon :)
08:36
<rniwa>
Ms2ger: it's crazy I had to spend 2+ hours reading various parts of HTML5, DOM, & Parsing/Serialization specifications to figure it out
08:37
rniwa
is back
08:37
<Ms2ger>
!
08:37
<rniwa>
Ms2ger: anyway, I was wondering if you could approve https://github.com/w3c/web-platform-tests/pull/456 ?
08:37
<rniwa>
MikeSmith: ^
08:37
<rniwa>
it's a pretty straight forward fix for xhtml files
08:38
<Ms2ger>
Sure
08:38
<Ms2ger>
Ugh, CRLF
08:39
<rniwa>
Ms2ger: are we supposed to always use LF there?
08:39
<rniwa>
Ms2ger: I was almost going to fix it but I wasn't sure if it was policy.
08:39
<Ms2ger>
Vague policy :)
08:39
<Ms2ger>
Anyway, merged
08:40
<rniwa>
Ms2ger: I need to start merging tests for your parsing & serialization spec
08:40
<rniwa>
so that we can give a meaningful feedback
08:40
<rniwa>
s/merging/importing/
08:40
<Ms2ger>
Yeah, I should probably put them into wpt
08:40
<Ms2ger>
Do you have a th.js test for that bug, btw?
08:41
<rniwa>
Ms2ger: what do you mean by th.js test?
08:41
<Ms2ger>
One that uses testharness.js
08:41
<rniwa>
Ms2ger: oh no.
08:41
<rniwa>
Ms2ger: I had assumed it exists somewhere
08:41
<rniwa>
Ms2ger: and I was hoping that you could tell me where it might be :(
08:41
<rniwa>
Ms2ger: but I could write one up if you want
08:42
<Ms2ger>
There's some tests in the spec repo
08:42
<Ms2ger>
But not at all exhaustive :)
08:42
<rniwa>
Ms2ger: could you move it to https://github.com/w3c/web-platform-tests/ ?
08:42
rniwa
can merge such a PR
08:42
<Ms2ger>
Yeah
08:42
<rniwa>
Ms2ger: that's why...
08:42
<rniwa>
Ms2ger: I was looking for those tests in that repo
08:42
<rniwa>
but couldn't find it :(
08:42
<Ms2ger>
I'll get to it :)
08:42
<rniwa>
Ms2ger: great. thanks.
08:43
<rniwa>
now I can announce on webkit-dev that we're the first major browser engine to pass the W3C tests for template elements :D
08:45
<Ms2ger>
Are they all wrong? ;)
08:45
<rniwa>
Ms2ger: they're all good :D
08:45
<rniwa>
Ms2ger: in fact, i've been fixing all those tests because they had a bunch of bugs like the one you just merged
08:45
<Ms2ger>
That seems implausible :)
08:46
<Ms2ger>
Unless you fixed them up first
08:46
<rniwa>
Ms2ger: I had to manually check each test case to make sure they're doing sensible things
08:46
<rniwa>
Ms2ger: you can check what i've done
08:46
<Ms2ger>
As always :)
08:46
<Ms2ger>
I'd rather not be sucked into that stuff :)
08:46
<rniwa>
Ms2ger: yeah
08:47
<rniwa>
Ms2ger: you'd better spend that time writing more spec, test, or doing other more useful things
08:47
<Ms2ger>
Indeed
08:50
<MikeSmith>
rniwa: you still needs me to do anything?
08:51
<rniwa>
MikeSmith: nope. Ms2ger took care of it for me :D
08:51
<MikeSmith>
k
08:55
rniwa
is going to enable template element by default on all webkit ports
08:56
<Ms2ger>
rniwa, so I'll try to look at the bug tonight
08:58
<hsivonen>
marcosc: for the "what exists" overview, it might be worthwhile to investigate IE, too, since, IIRC, IE comes the closest to using the metadata elements from the HTML spec
08:58
<marcosc>
hsivonen: yes, I've done a little bit of research
08:59
<marcosc>
I'll make sure it gets added
08:59
<marcosc>
there is a lot of use of <meta name="application-name">
08:59
<marcosc>
I was impressed
08:59
<marcosc>
it's more popular than the apple equivalent
09:06
<hsivonen>
marcosc: any new recent exploration of installing apps to the homescreen but keeping the URL field available to the user somehow? (e.g. for copying tweet URLs)
09:07
<marcosc>
hsivonen: yeah, that's come up
09:07
<marcosc>
it might be some kind of option... "allow deep linking" or something
09:08
<marcosc>
but it might be that links might just work as normal, so you can long hold and copy a link
09:08
<marcosc>
hsivonen: there is also talk of enabling the user to access the address bar
09:10
<hsivonen>
oh so you'd long-hold the link on the view *previous* to the one you want to link to?
09:12
<hsivonen>
why did Apple create apple-mobile-web-app-title *after* application-name was in the spec?
10:02
<zcorpan>
annevk-cloud: what replaced "CORS-same-origin"?
10:03
<zcorpan>
annevk-cloud: see http://dev.w3.org/csswg/cssom/#requirements-on-user-agents-implementing-the-xml-stylesheet-processing-instruction
11:02
<zcorpan>
Hixie: is http://software.hixie.ch/utilities/js/live-dom-viewer/delayed-image supposed to have an empty body?
11:56
<hsivonen>
who operates web-sniffer.net and how does one report bugs to that person?
12:02
<annevk>
hsivonen: use http://redbot.org/ instead?
12:38
<hsivonen>
hmm. when reporting a WebKit security bug, most bugmail goes to @chromium.org
13:31
<philipj>
jgraham, ping
14:00
<jgraham>
philipj: Pong
14:01
<Ms2ger>
Did we hire philipj too?
14:02
<jgraham>
AFAIK no, he has just picked up bad irc habits from somewhere
14:04
<Ms2ger>
Oh, this isn't #ateam
14:05
<jgraham>
No, it isn't
14:18
<zcorpan>
annevk: what replaced "CORS-same-origin"?
14:20
<hsivonen>
marcosc: I sent an email to the manifest thread
14:20
<annevk>
zcorpan: the way it works now is that you get a response back and if you don't look behind the curtain (as some APIs such as <img> need to do) you can't really distinguish
14:21
<annevk>
zcorpan: I guess I should still have some kind of generic flag exposed so you can more easily determine whether to look behind the curtain
14:22
<zcorpan>
annevk: it's not obvious to me what i should say in cssom
14:23
<annevk>
zcorpan: I need more context
14:23
<zcorpan>
annevk: search for "cors-same-origin" in http://dev.w3.org/csswg/cssom/
14:25
<annevk>
zcorpan: the first instance looks like a bug
14:25
<annevk>
zcorpan: and yeah, for the second instance we should probably have that kind of convenience flag
14:26
<zcorpan>
annevk: why is it a bug?
14:27
<annevk>
zcorpan: I think you only want to do that kind of thing for actual same-origin
14:27
<annevk>
zcorpan: not allow some kind of CORS opt-in to a quirk
14:30
<zcorpan>
annevk: ok. how do i do that?
14:33
<annevk>
zcorpan: I guess you'd compare document's origin with request's origin
14:34
<annevk>
zcorpan: would prolly be cleaner if response had a URL and origin associated with it too
14:34
<zcorpan>
annevk: what if it redirects?
14:34
<annevk>
zcorpan: request's origin is updated
14:34
<zcorpan>
ah
14:34
<annevk>
zcorpan: so you'd compare after the fetch finalizes
14:34
<annevk>
zcorpan: it's not entirely clear to me how stable all of that is though
14:35
<annevk>
nobody has really given feedback yet whether this is the right abstraction for what's going on in a browser :/
14:35
<zcorpan>
annevk: ok. i'm not that familiar with the code in this area
14:37
<zcorpan>
annevk: thanks
14:41
<tj_vantoll>
Hi all, is it possible to respond to a mailing list thread from the digest email? I'm assuming that it's not but I wanted to check.
14:43
<tj_vantoll>
I'm basically wondering if there's a way to respond to a thread if I did not receive the actual email.
14:43
<annevk>
tj_vantoll: I think it's possible, but it's not super desirable as it breaks threading
14:44
<annevk>
tj_vantoll: if this is whatwg.org email, you can browse to the message in http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/ and then hit the first link (next to the author's name); that should open up your email client with the appropriate headers set
14:44
<annevk>
tj_vantoll: if it's w3.org email, you can do the same thing, but different
14:45
<jgraham>
In theory you could fix the threading headers by hand ofc
14:45
<tj_vantoll>
annevk: It's a whatwg one. I had no idea that that link would do that.
14:46
<tj_vantoll>
I guess I'll give that a shot and see what happens. Thanks!
14:54
<sangwhan>
Anyone know which spec was the first to introduce a redundant named color "grey" when there was already "gray"?
14:55
<sangwhan>
SVG seems suspicious, but not sure
14:55
<annevk>
sangwhan: I think that's a CSS thing
14:59
<SimonSapin>
CSS blames SVG: http://www.w3.org/TR/css3-color/#svg-color
14:59
<hsivonen>
does Blink-Opera really not have a character encoding override menu?
15:00
<hsivonen>
hooray, if so!
15:00
<sangwhan>
hsivonen: not on the checkout/branch I have, but lemme check the nightly
15:01
<sangwhan>
SimonSapin: Yeah, and SVG doesn't quite blame anyone - but why would they do that is beyond me
15:02
<SimonSapin>
bikeshedding
15:02
<jgraham>
X11
15:03
<jgraham>
http://en.wikipedia.org/wiki/X11_color_names
15:04
<annevk>
Better add colour too
15:04
<sangwhan>
Copy the most impossible to use API on planet earth, well done :)
15:04
<SimonSapin>
Oh, we have a http://xkcd.com/978/
15:04
<sangwhan>
Now time to implement background-colour
15:04
<SimonSapin>
W3C spec refers to wikipedia, which talks about what the W3C does
15:04
<annevk>
Do gray and grey serialize back in the same way?
15:04
<annevk>
That might require more bits than you'd want to give out
15:04
<hsivonen>
sangwhan: I sure hope Opera doesn't add that misfeature later
15:05
<SimonSapin>
I think colors serialize to #hex or rgb()
15:05
<annevk>
SimonSapin: even for specified style?
15:05
<sangwhan>
annevk: don't colors serialize to non-named?
15:05
<annevk>
sangwhan: well that's what I wanted them to do
15:05
<sangwhan>
one way to find out
15:05
<annevk>
but glazou made a lot of noise
15:07
<zewt>
hsivonen: if not that's a bug, encoding override isn't an optional thing
15:07
<zewt>
at least for people who aren't in japan occasionally trying to view pages that are, heh
15:09
<annevk>
SimonSapin: it'll prolly take a while for me to fix those URL bugs
15:09
<SimonSapin>
annevk: I know you’re busy
15:10
<annevk>
SimonSapin: yeah, these games don't play themselves
15:11
<darobin>
IIRC the grey/gray thing comes from an antique Netscape
15:11
<darobin>
it was supposed to make HTML easier to use for all authors
15:11
darobin
glad they didn't push that localise in all languages...
15:11
odinho
want grå, now!
15:11
<darobin>
or some story like that, I don't recall
15:11
<darobin>
it's certainly stupid enough to be a Netscape feature
15:12
<sangwhan>
Presto keeps it as grey and exposes that (but internally seems to use the raw color) and Blink seems to go whammo rgb() the moment it goes in
15:12
<philipj>
jgraham, I was going to ask about how to use assert_throws for exceptions which aren't DOMExceptions, but gave up instead
15:12
<SimonSapin>
spec says to serialize <color> with rgb() or rgba() http://dev.w3.org/csswg/cssom/#serialize-a-css-component-value
15:12
<SimonSapin>
Gecko, Blink and Presto do keywords: data:text/html,<body style=color:grEen><script>document.write(document.body.style.color)</script>
15:12
<philipj>
jgraham, the exception object thrown for "not enough arguments" doesn't have a code attribute
15:13
<hsivonen>
zewt: what's the deal with this being more of a problem in Japan than in Russia even though there's a greater number of Cyrillic legacy encodings?
15:13
<hsivonen>
zewt: broken windows effect?
15:14
<annevk>
hsivonen: what about CESU-8 btw?
15:14
<sangwhan>
SimonSapin: my tests conflict with your results on blink, although TC is a bit different
15:14
<annevk>
hsivonen: include that too?
15:14
<SimonSapin>
sangwhan: what’s your test case?
15:14
<jgraham>
philipj: You should be able to pass in an instance of the exception type as the first argument
15:15
<hsivonen>
annevk: I couldn't figure out why CESU-8 would be dangerous
15:15
<jgraham>
e.g. assert_throws(new TypeError(), function() {throw new TypeError()}))
15:15
<hsivonen>
annevk: and for the BMP, it seems useful to get to read it as UTF-8
15:16
<sangwhan>
SimonSapin: document.querySelectorAll('body')[0].style.color='grey';alert(document.querySelectorAll('body')[0].style.color);
15:16
<annevk>
hsivonen: k
15:16
<annevk>
makes sense
15:16
<philipj>
jgraham, thanks!
15:17
<sangwhan>
hsivonen: nightlies don't have it either, but i don't know if the feature will make a comeback or not. not in the desktop team.
15:18
<SimonSapin>
sangwhan: wonderful. Blink uses rgb() for grey, but a keyword for grey
15:18
<SimonSapin>
a keyword for gray*
15:18
<annevk>
hsivonen: for what's worth, I am worried that there's pages out there that rely on the EBDIC label being ignored
15:19
<Ms2ger>
darobin, I thought that was x11 legacy
15:19
<hsivonen>
annevk: possible :-(
15:19
<Ms2ger>
I read an interesting post about the history of x11 at some point
15:19
<hsivonen>
annevk: getting telemetry for everything is so slow :-(
15:19
<darobin>
Ms2ger: IIRC it wasn't in Mosaic, but maybe that's named colours in general? — I don't know, it's been a long time
15:20
<hsivonen>
annevk: wouldn't such pages break in IE, then, though?
15:20
<Ms2ger>
"From notes it also seems some people was happy about this because they always forgot if it was grey or gray"
15:20
<sangwhan>
New Opera not having the encoding menu, not sure how that flies in Asia where encoding on every other website is broken beyond recognition though
15:20
<Ms2ger>
http://unix.stackexchange.com/questions/72919/what-are-the-origins-of-rgb-txt
15:20
<annevk>
hsivonen: fair point
15:20
<annevk>
hsivonen: if we only include the IE labels
15:20
<sangwhan>
SimonSapin: Hah, grey is different from gray? That's interesting
15:21
<SimonSapin>
that’s… I’m gonna go do something else now.
15:21
<hsivonen>
annevk: to serve the purpose of the mitigation, we should include all EBCDIC labels that ICU knows about
15:23
<sangwhan>
http://cvsweb.xfree86.org/cvsweb/*checkout*/xc/programs/rgb/rgb.txt?rev=1.1 has too many shades of gray or grey, thank god that wasn't adopted...
15:24
<Ms2ger>
sangwhan, 50?
15:28
<sangwhan>
Ms2ger: moar
15:28
<sangwhan>
Well, enough with this grey nonsense, I'll do a in memory swap to gray when I see grey from now on
15:29
<jgraham>
Ms2ger: 236 shades of gr[e|a]y
15:29
<Ms2ger>
Noooooooo
15:29
<annevk>
hsivonen: alright, I'll collect those and report the datasets in the bug
15:30
<hsivonen>
annevk: thanks!
15:37
<annevk>
hsivonen: oh man, ICU is a disaster
15:37
<annevk>
hsivonen: try data:text/html;charset=ibm-17593,<script>alert(document.characterSet)</script> in Chrome
15:38
<Ms2ger>
Good thing that TC39 forced everyone to use it
15:41
<annevk>
# This is a special version of ibm-1140 that the XML4C (Xerces) parser team
15:41
<annevk>
# requested in 2000.
15:41
<annevk>
# It maps both EBCDIC LF and NL controls to Unicode LF U+000A.
15:42
<karlcow>
annevk, I have asked the person who reported the bug 304/CORS to give a more explicit code example and http sequence to explain how and when it fails.
15:42
<annevk>
karlcow: seems pretty clear no?
15:42
<annevk>
karlcow: sicking should've looked up the status code...
15:44
<karlcow>
I was surprised by sicking's answer ;)
15:45
<annevk>
hsivonen: ICU also supports labels such as ISO_2022,locale=ko,version=0
15:45
<annevk>
hsivonen: if we were to take it into account we would have to do much more it seems
15:48
<hsivonen>
annevk: hmm. yeah, if that kind of labels get reflected to the Web, it's hopeless
15:49
<hsivonen>
annevk: I think that can't happen with ICU4J, though
15:49
<hsivonen>
annevk: I believe its label handling is what one would consider normal legacy
15:49
<annevk>
hsivonen: what is ICU4J and how do I get that data out of http://demo.icu-project.org/icu-bin/convexp ?
15:50
<annevk>
hsivonen: but see the Chrome example I gave earlier, it gives utf-8 for that
15:50
<annevk>
hsivonen: rather than iso-8859-1
15:50
<hsivonen>
annevk: yeah, that ibm-17593 thing is very, very sad
15:51
<hsivonen>
annevk: ICU4J is the Java version of ICU.
15:52
<annevk>
hsivonen: okay, so that exposes a lot less labels and encodings
15:53
<annevk>
hsivonen: but if that's safe :/
15:54
<hsivonen>
annevk: http://source.icu-project.org/repos/icu/icu4j/trunk/main/classes/charset/src/com/ibm/icu/charset/UConverterAliasDataReader.java gives a hint of what file to look for in the repo
15:54
<annevk>
hsivonen: the file is http://source.icu-project.org/repos/icu/icu/trunk/source/data/mappings/convrtrs.txt
15:54
<annevk>
hsivonen: which as you can see has many mappings for utf-8
15:55
<hsivonen>
ಠ_ಠ at Chrome just exposing ICU alias data to the Web
15:55
<annevk>
data:text/html;charset=windows-65001,<script>alert(document.characterSet)</script> is utf-8 too in Chrome
15:55
<gsnedders>
hsivonen: CESU-8 allows non-shortest-form null, no? So might break with some sanitizers…
15:55
<hsivonen>
not cool. bad Chrome.
15:56
<hsivonen>
gsnedders: oh. I didn't know that. I thought only Java "modified UTF-8" did that
15:56
<annevk>
where is jsbell
15:56
<gsnedders>
hsivonen: There was a reason for my "no?" at the end, I thought that was part of the CESU-8 too?
15:56
<gsnedders>
hsivonen: I could, of course, be wrong :)
15:57
<gsnedders>
hsivonen: I am wrong.
15:57
<hsivonen>
ok
15:58
<annevk>
gsnedders: "CESU-8 is similar to Java's Modified UTF-8 but does not have the special encoding of the NUL character (U+0000)."
15:58
<gsnedders>
I was just typing that.
15:58
<gsnedders>
(It is, notably, the only difference between the two)
16:05
<annevk>
hsivonen: Safari does the same for ICU
16:16
<annevk>
tj_vantoll: btw, if you have Gmail I recommend switching from digest and just using a filter
16:17
<annevk>
hsivonen: filed bugs on Chrome and Safari
16:17
<annevk>
hsivonen: not entirely sure what to do for the replacement encoding now
16:17
<tj_vantoll>
annevk: I just did thanks.
16:18
<Hixie>
GPHemsley: thanks
16:22
<JonathanNeal>
Mornin' (Pacific)
16:26
<Hixie>
zcorpan: fixed
16:26
<Hixie>
GPHemsley: btw, anne is right that we don't necessarily need to respond to everything. it's like the story of the new bar vs the old bar, where the old bar has a bunch of notices up for random stuff its customers have done over the years.
16:27
<Hixie>
GPHemsley: it may simply be better to live with the occasional person having trouble, and help them personally rather than with a notice
16:27
GPHemsley
is not familiar with said story
16:27
<Hixie>
marcosc: i'm unclear on what problem this is solving
16:28
<annevk>
GPHemsley: you go to the new bar because it has less bullshit bothering you
16:28
<Hixie>
GPHemsley: the old bar will have signs like "no backpacks on tables" because one day twenty years ago, some guy put his backpack on a bar and knocked a drink over, or whatever
16:29
<Hixie>
GPHemsley: which is a pointless sign really :-)
16:29
GPHemsley
doesn't go to bars
16:29
<Hixie>
well me either
16:30
<Hixie>
but the point is just that there's a cost to putting signs up for every last thing
16:30
<Hixie>
(and people dont read long signs anyway)
16:30
<GPHemsley>
ok
16:30
<GPHemsley>
like I said, I'll revisit it later
16:31
<Hixie>
lgtm
16:31
<Hixie>
i'm not particularly worried about the notices myself
16:31
<Hixie>
just wanted to make sure the faq didn't scare people off :-)
16:32
<GPHemsley>
I think the FAQ will scare people off with or without the notices :P
16:32
<GPHemsley>
it's humongous
16:32
<Hixie>
well, true
16:32
<Hixie>
but gotta give it a chance :-)
16:33
<GPHemsley>
were there any other pages that you wanted the notice removed from?
16:33
<GPHemsley>
(or, like I said, you can take care of it yourself)
16:33
<Hixie>
nah, i only send people to the faq :-)
16:34
<GPHemsley>
k
16:34
<GPHemsley>
annevk: In the meantime, you should now be able to set up a user style that removes the notice
16:35
<Hixie>
bbiab
16:41
<annevk>
GPHemsley: thanks, but I try not to change my default setup
16:41
<annevk>
GPHemsley: makes it harder to reproduce bugs
16:41
<GPHemsley>
very well
17:06
<Hixie>
annevk: just logging in removes them too :_)
17:09
<annevk>
Hixie: I'm in permanent private mode on my phone
17:09
<annevk>
Hixie: and also, I'm more concerned about other people than myself here, I can cope
17:11
<Hixie>
"permanent private mode"?
17:50
<dglazkov>
good morning, Whatwg!
19:29
<zcorpan>
TabAtkins: did you flesh out something with <picture>?
19:53
<zcorpan>
hmm, the web component thread became emotional. maybe it needs ice cream
21:13
<annevk>
zcorpan: actual ice cream or a metaphor?
21:16
<zcorpan>
🍨
21:16
<Ms2ger>
Ice cream vas
21:16
<Ms2ger>
vans
21:16
<annevk>
"Your search - 🍨 - did not match any documents."
21:24
<gsnedders>
Ms2ger: don't end up with people dead people of ice cream vans!
21:24
<gsnedders>
(Context: http://en.wikipedia.org/wiki/Glasgow_Ice_Cream_Wars)
21:50
<zcorpan>
annevk: https://duckduckgo.com/?q=🍨
23:43
<Hixie>
dglazkov: http://lists.w3.org/Archives/Public/public-webapps/2013JulSep/0454.html is hard to read because of all the cuteness :-/
23:59
<dglazkov>
Hixie: your argument is invalid