| 02:12 | <dekiss> | annevk-cloud |
| 02:12 | <dekiss> | here? |
| 02:12 | <dekiss> | hwo can I help making the html ? ^^ |
| 02:12 | <dekiss> | hehe |
| 02:12 | <dekiss> | suggestions on mailing list are only considered to be implemented? |
| 04:36 | <JonathanNeal> | Is there ARIA role or similar attribute that identifies a link as being to the current page, like one would find in a navigation menu? |
| 04:47 | <Hixie> | JonathanNeal: the href="" pointing to the local page seems sufficient to indicate that, no? |
| 04:49 | <JonathanNeal> | Perhaps. Still, I was wondering if there was anything more blatantly assistive. |
| 04:50 | <Hixie> | i don't understand |
| 04:50 | <Hixie> | assistive how? |
| 04:50 | <Hixie> | the AT has access to the link... |
| 04:51 | <JonathanNeal> | Hixie: some ATs are not as A as others. Is rel="self" valid? |
| 04:52 | <Hixie> | in general i would strongly recommend against us making anything valid that is redundant with existing data |
| 04:52 | <Hixie> | since that's just asking for it to be inconsistent |
| 04:52 | <JonathanNeal> | Hixie: responding to "i don't understand", what I mean is: in theory, navigations with self pointing links have been around for ages so I would hope this has been resolved by ATs without the need for assistive roles or aria-*. Still, I'm asking because I do not know if there is, in fact, a way to be explicit. |
| 04:53 | <Hixie> | how is href="" not explicit? |
| 04:53 | <Hixie> | i don't understand what you mean by explicit |
| 04:55 | <JonathanNeal> | To be clear, you mean an href with an empty value? |
| 04:56 | <Hixie> | no i mean the href="" attribute |
| 04:58 | <JonathanNeal> | You mean an href pointing to the page itself should be explicit enough for any AT, and anything additional is harmfully redundant? |
| 04:58 | <dekiss> | Hixie can I be editor of HTML ? :) |
| 04:59 | <Hixie> | JonathanNeal: "harmfully redundant" is tautological, but yeah |
| 04:59 | <Hixie> | dekiss: sure |
| 04:59 | <dekiss> | ^^ NICE :) |
| 04:59 | <dekiss> | is there some official application I should send? |
| 04:59 | <Hixie> | yeah |
| 04:59 | <JonathanNeal> | Hixie: not entirely so, as is the case for roles and html5 attributes. |
| 05:00 | <JonathanNeal> | err, html5 elements |
| 05:00 | <Hixie> | dekiss: attach a patch to bug https://www.w3.org/Bugs/Public/show_bug.cgi?id=23659 |
| 05:00 | <Hixie> | JonathanNeal: it's just as redundant for those |
| 05:00 | <JonathanNeal> | <nav role="navigation"> redundant but not necessarily harmful. |
| 05:00 | <dekiss> | Hixie THANKS! |
| 05:00 | <Hixie> | JonathanNeal: it's both redundant and harmful |
| 05:00 | <Hixie> | dekiss: (the patch should actually fix the bug, obviously) |
| 05:00 | <Hixie> | dekiss: (not just any random patch) |
| 05:01 | <dekiss> | ok :) |
| 05:02 | <JonathanNeal> | well, I don't believe that redundancy for compatibility is as harmful as redundancy for, say, uniformity or verbosity. |
| 05:02 | <JonathanNeal> | If you don't sympathize with that statement, I'd love to read your thoughts on why. |
| 05:03 | <Hixie> | redundancy in HTML leads to copy-pasta which leads to ATs being given bogus data where they would instead work fine if there was no redundancy |
| 05:07 | <JonathanNeal> | Thanks. |
| 05:08 | <dekiss> | Hixie you think Webgl will be used much? |
| 05:12 | <JonathanNeal> | Hixie, speaking of redundancy, do you think the <head> element has any real value? |
| 05:13 | <Hixie> | dekiss: not my department |
| 05:13 | <Hixie> | JonathanNeal: never really thought about it, since we can't drop it |
| 05:13 | <dekiss> | ok |
| 05:16 | <JonathanNeal> | That's interesting and curious. So, if you did think about it, and found it to be the meta equivalent of <font>, why couldn't you drop it? |
| 05:19 | <Hixie> | JonathanNeal: we can't drop it just like we can't drop <font> |
| 05:19 | <Hixie> | JonathanNeal: browsers implement it, pages depend on it |
| 05:19 | <Hixie> | JonathanNeal: we couldn't even make it non-conforming (which is what HTML4 did with <font>) since the parser implies it |
| 05:19 | <JonathanNeal> | I just learned something. I thought <font> was deprecated in HTML4 and removed in HTML5. |
| 05:20 | <Hixie> | it's obsolete, but browsers still have to support it |
| 05:23 | <dekiss> | Chrome has some bug with antialising on Helvetica Neur |
| 05:23 | <dekiss> | also it had with css transitions on fixed elements |
| 05:24 | <dekiss> | where can I send those bugs? |
| 05:28 | <GPHemsley> | Hixie: Is the HTML UA stylesheet available in a single file or block of text? |
| 05:29 | <Hixie> | dekiss: crbug.com |
| 05:29 | <Hixie> | GPHemsley: it's available in a single file, but not separate from interleaving other text |
| 05:30 | <Hixie> | GPHemsley: in part because it's not entirely expressible in CSS |
| 05:30 | <GPHemsley> | Hixie: :P |
| 05:30 | <GPHemsley> | fair enough, I suppose |
| 05:31 | GPHemsley | uses CSS 2.1's default stylesheet for HTML 4 instead. |
| 05:32 | <Hixie> | GPHemsley: what's your purpose? |
| 05:32 | <GPHemsley> | just looking for a sample CSS file |
| 05:32 | <Hixie> | ah |
| 05:34 | GPHemsley | wonders how easy it is to find a sample ECMAScript file that's not contaminated by JavaScript extensions... |
| 05:36 | <JonathanNeal> | GPHemsley: last updated 2012 http://www.iecss.com/whatwg.css |
| 05:38 | <dekiss> | thanks |
| 05:39 | <GPHemsley> | JonathanNeal: Thanks; need an ECMAScript file now. |
| 05:41 | <JonathanNeal> | Are there (and what are the) standard ways to communicate the language of a document outside of using <html> with lang=""? |
| 05:44 | <GPHemsley> | JonathanNeal: The Content-Language HTTP header |
| 05:47 | <Hixie> | The Content-Language HTTP header actually sets the intended languages understood by the reader, not the language of the document. |
| 05:47 | <Hixie> | iirc |
| 05:48 | <GPHemsley> | Hixie: That's Accept-Language |
| 05:49 | <GPHemsley> | JonathanNeal: http://www.w3.org/International/questions/qa-http-and-lang |
| 05:49 | <GPHemsley> | JonathanNeal: http://www.w3.org/International/techniques/authoring-html#language |
| 05:50 | <GPHemsley> | JonathanNeal: http://www.w3.org/International/tutorials/language-decl/ |
| 05:57 | <Hixie> | GPHemsley: no, i mean content-language |
| 05:57 | <Hixie> | GPHemsley: accept-language is what the client sends |
| 05:58 | <GPHemsley> | Hixie: Content-Language is the language of the document. Not sure what you expect the server to know about what languages the reader understands... |
| 05:58 | <Hixie> | "The Content-Language entity-header field describes the natural language(s) of the intended audience for the enclosed entity." |
| 06:00 | <GPHemsley> | == "the language of the document" |
| 06:01 | <Hixie> | not necessarily |
| 06:02 | <Hixie> | consider for example a document written in french but intended to teach english speakers french |
| 06:07 | <JonathanNeal> | i understand the arguments, but i didn't catch which header the server should send, Content-Language? |
| 06:08 | <Hixie> | lang="". |
| 06:09 | <Hixie> | if you want to set the language of the document, that's the best way to do it. |
| 06:10 | <GPHemsley> | Hixie: In that case, the server would probably send `Content-Language: fr, en` |
| 06:11 | <Hixie> | not much point a french language speaker looking at it :-) |
| 06:11 | <JonathanNeal> | Like I originally asked, the server solution was the one I was looking for. "Content-Language", got it. |
| 06:13 | <GPHemsley> | Hixie: Despite it being easy to get that interpretation in the letter of the spec, I'm not sure the spirit of the spec meant for that to be the case. |
| 06:13 | <GPHemsley> | JonathanNeal: Yeah, we're just quibbling about specs at this point. Content-Language is what you want. :) |
| 06:16 | <Hixie> | JonathanNeal: content-language is almost certainly not what you want unless you're writing some sort of intellectual exercise |
| 06:16 | <Hixie> | GPHemsley: specs mean what they say |
| 06:16 | <Hixie> | GPHemsley: you can't just say "well the spec says X, but it obviously means Y", and then interpret it as saying something else |
| 06:17 | <JonathanNeal> | Is there a header that defines the language of the document? |
| 06:17 | <Hixie> | no |
| 06:17 | <Hixie> | (but even if there was, you wouldn't want to use it) |
| 06:18 | <JonathanNeal> | Well, if I intend something for a French audience, can I use "Content-Language"? |
| 06:18 | <JonathanNeal> | I would surely be following the spec and not "obviously" interpretting. |
| 06:19 | <Hixie> | unless you're writing an intellectual exercise, you almost certainly don't want to bother with HTTP headers. |
| 06:19 | <Hixie> | almost nothing does anything useful with them. |
| 06:19 | <GPHemsley> | that's true |
| 06:20 | <GPHemsley> | Hixie: I do believe French is a "natural language of the intended audience" if the intended audience is trying to learn French. |
| 06:20 | <JonathanNeal> | Hixie: with all HTTP headers or just language ones? |
| 06:21 | <Hixie> | JonathanNeal: certainly all the metadata ones. things like caching headers have more implementations. |
| 06:21 | <JonathanNeal> | Cause I would disagree with all HTTP headers being useless in general. UTF-8 is super useful. So is X-UA. |
| 06:21 | <Hixie> | JonathanNeal: security-related headers tend to be useful, too |
| 06:21 | <Hixie> | UTF-8 isn't a header? :-) |
| 06:21 | <Hixie> | X-UA is way bogus. |
| 06:22 | <JonathanNeal> | UTF-8 comes with Content-Type |
| 06:22 | <Hixie> | content-type is mostly bogus too |
| 06:22 | <JonathanNeal> | And the HTTP spec is goofy and follows ISO. |
| 06:22 | <GPHemsley> | Hixie: Nearly by definition, the natural language(s) your intended audience must include whatever language the document is in. |
| 06:22 | <GPHemsley> | +of |
| 06:23 | <JonathanNeal> | But speaking of intellectual exercises, how many browsers require <!doctype html> to render the page in non quirks mode. |
| 06:23 | <JonathanNeal> | ? |
| 06:23 | <GPHemsley> | Put another way, why would you send a document in a particular language to your intended audience if it was not one of their languages? |
| 06:24 | <GPHemsley> | 42 |
| 06:24 | GPHemsley | should go to bed... |
| 06:26 | <Hixie> | GPHemsley: to teach them the language, as above |
| 06:27 | <GPHemsley> | Hixie: But then it is one of their languages |
| 06:27 | <Hixie> | JonathanNeal: all of them, i think |
| 06:27 | <GPHemsley> | Hixie: It seems to all come down to what "of" means in this context. |
| 06:27 | <Hixie> | GPHemsley: i read zero hebrew, but one could make a page to teach me hebrew that was exclusive in hebrew (and lots of pictures, probably) |
| 06:28 | <GPHemsley> | Hixie: I still think Hebrew would be a "natural language of the intended audience" in that case. |
| 06:29 | <GPHemsley> | Also, "describe" and "specify" mean different things, IIUC. |
| 06:29 | <GPHemsley> | "Describe" is much more broad and hand-wavy |
| 06:29 | <Hixie> | hebrew isn't, by any definition i can see, a natural language of mine. |
| 06:30 | <JonathanNeal> | Hixie: actually, IE8+, Chrome, Safari. |
| 06:30 | <Hixie> | any definition that would put hebrew into the category of "hixie's natural languages" would put all natural languages into that category. |
| 06:30 | <Hixie> | JonathanNeal: and firefox, and opera |
| 06:30 | <GPHemsley> | Hixie: Yes, this use of "natural" is meant to distinguish from "artificial" or "constructed" or "programming" |
| 06:30 | <JonathanNeal> | Yea, I didn't realize that. |
| 06:31 | <GPHemsley> | Hixie: https://en.wikipedia.org/wiki/Natural_language |
| 06:31 | <Hixie> | GPHemsley: i'm not sure what you're arguing against |
| 06:31 | <GPHemsley> | Hixie: Don't confuse "natural" or "native" |
| 06:31 | <GPHemsley> | s/or/with/ |
| 06:32 | <GPHemsley> | I don't either |
| 06:32 | <GPHemsley> | Which is why I should go to bed |
| 06:32 | <GPHemsley> | g'night everybody |
| 06:32 | <Hixie> | nn |
| 06:34 | <JonathanNeal> | That was fun. I learned a lot. |
| 07:05 | <gjsrivastava> | moo |
| 07:14 | <SteveF> | JonathanNeal: ATs recognise and expose in page links, there is no ARIA property to indicate them |
| 07:14 | <JonathanNeal> | Thanks, SteveF! |
| 07:19 | <SteveF> | JonathanNeal: <nav role=navigation> is not yet redundant as <nav> is not exposed correctly via accessibility APIs in some browsers see https://rawgithub.com/stevefaulkner/HTML5accessibility/master/index.html, so suggest if you want AT users to recognise <nav> then addition of role=navigation is useful, but in general see note http://www.w3.org/html/wg/drafts/html/master/dom.html#aria-usage-note |
| 07:20 | <JonathanNeal> | Yes, there was some discussion about this. Ultimately, the redundancy will likely create some headaches later, once ATs have caught up. |
| 07:20 | <JonathanNeal> | So, my take is: for now, they're useful, but keep an eye on them. |
| 07:20 | <JonathanNeal> | You also wrote a blog about this that I consulted, SteveF. |
| 07:20 | <SteveF> | JonathanNeal: "Ultimately, the redundancy will likely create some headaches later, once ATs have caught up." how so? |
| 07:22 | <JonathanNeal> | SteveF: IMO, people will misunderstand the redundancy later and use the roles inappropriately during copy and pasting. That may or may not be similar to what Hixie was getting at when he mentioned copy-pasta. But that's my take. |
| 07:22 | <SteveF> | Jonathanneal: using <nav> for something other than a <nav> will always create a headache regardless of presence of role=navigation |
| 07:23 | <JonathanNeal> | I'm guilty of this too. Until today, I didn't know the difference between "menu(-item)" and "list(-item)", as I saw them used interchangeably in stackoverflow posts. |
| 07:24 | <JonathanNeal> | But someone is more likely to accidentally <nav><ul role="navigation"> if role is left in prominent examples on the web long enough. |
| 07:24 | <SteveF> | JonathanNeal: you may also find Using WAI-ARIA in HTML http://rawgithub.com/w3c/aria-in-html/master/index.html useful |
| 07:25 | <zcorpan> | i wonder why the AT situation with <nav> is still worse than role=navigation. the two were invented about at the same time |
| 07:25 | <SteveF> | JonathanNeal: <nav><ul role="navigation"> is a conformance error |
| 07:25 | <zcorpan> | and it was almost 10 years ago now |
| 07:26 | <SteveF> | zcorpan: because (some) browsers have not implemented the mappings |
| 07:27 | <SteveF> | zcorpan: bugs have been filed on implementers in most cases |
| 07:29 | <zcorpan> | yeah, so from that i conclude that accessibility is low priority |
| 07:32 | <SteveF> | zcorpan: depends on the browser |
| 07:33 | <zcorpan> | indeed |
| 08:28 | <zcorpan> | TabAtkins: did anything happen with <http://www.w3.org/mid/CAAWBYDCodt_vi7ajvnWLYMNLwN-8rjYqyMHW8jb40HXHOguVog⊙mgc>; ? |
| 08:34 | <JonathanNeal> | SteveF: i'm having trouble finding something in the spec, is the first or root <header> implicitly role="banner" or *must* it be declared? |
| 08:37 | <Ms2ger> | Default: no role; |
| 08:37 | <Ms2ger> | If specified, role must be banner |
| 08:37 | <Ms2ger> | http://www.whatwg.org/specs/web-apps/current-work/multipage/elements.html#wai-aria |
| 08:39 | <JonathanNeal> | Thanks! |
| 09:17 | <SteveF> | JonathanNeal: am progressively adding such info into the 'head' section of each element definition , for example http://www.w3.org/html/wg/drafts/html/master/sections.html#the-header-element, the default role for <header>header element that is not a descendant of an article or section element. is banner, otherwise no role |
| 09:18 | <SteveF> | (as implemented) in webkit/firefox |
| 09:23 | <JonathanNeal> | oh, that's different. |
| 09:25 | <SteveF> | JonathanNeal: suggest filing a bug against the whatwg spec as the w3c spec reflects implementations |
| 09:26 | <SteveF> | and implementer consensus |
| 10:03 | <xDevz> | If you need any help with development, full Website coding, Script modifications, small fixes and others.. msg me |
| 10:09 | <wilhelm> | No. |
| 10:13 | <odinho> | Agreed. |
| 10:13 | <annevk> | I'm somewhat against making private email public, but what the fuck: http://lists.w3.org/Archives/Public/public-restrictedmedia/2014Jan/0074.html |
| 10:17 | <Ms2ger> | wat |
| 10:47 | <annevk> | I don't like what Gecko and IE do for hostname |
| 10:47 | <annevk> | and Chrome does not do it |
| 10:47 | <annevk> | and Safari does not do it |
| 11:04 | <smaug____> | annevk: and what does Gecko do? |
| 11:05 | <annevk> | smaug____: https://www.w3.org/Bugs/Public/show_bug.cgi?id=11587 |
| 11:05 | smaug____ | doesn't see anything too interesting in the code |
| 11:05 | <annevk> | obj.hostname = "2001::1" works |
| 11:05 | <smaug____> | ah, that is low level stuff |
| 11:05 | <smaug____> | Necko stuff |
| 11:05 | <annevk> | even though 2001::1 needs to be [2001::1] in order to parse |
| 11:06 | <annevk> | so there is some check somewhere to see if there's a : in the input |
| 11:06 | <annevk> | same when serializing |
| 11:06 | <annevk> | it's somewhat ugly |
| 11:10 | <Ms2ger> | Dammit IPv6 |
| 11:11 | <smaug____> | http://mxr.mozilla.org/mozilla-central/source/netwerk/base/src/nsStandardURL.h#352 https://bugzilla.mozilla.org/show_bug.cgi?id=134793 |
| 11:11 | <smaug____> | er, even older |
| 11:12 | <smaug____> | https://bugzilla.mozilla.org/show_bug.cgi?id=124042 |
| 11:12 | <smaug____> | ancient stuff |
| 11:13 | <smaug____> | hmm, or is there something when we parse the url |
| 11:19 | <annevk> | Ms2ger: reviewed https://critic.hoppipolla.co.uk/r/539 for you |
| 11:19 | <Ms2ger> | Thanks! |
| 11:21 | <annevk> | smaug____: so that StandardURL.h thing is for returning right? |
| 11:22 | <annevk> | well yeah must be |
| 11:23 | <annevk> | http://mxr.mozilla.org/mozilla-central/source/netwerk/base/src/nsStandardURL.cpp#1447 seems to be for setting |
| 11:25 | <annevk> | http://mxr.mozilla.org/mozilla-central/source/netwerk/base/src/nsStandardURL.cpp#387 |
| 11:26 | <annevk> | it has the exact weird ass magic you'd expect |
| 11:26 | <smaug____> | yup |
| 11:26 | <annevk> | and it's been there since the CVS era |
| 11:27 | <smaug____> | https://bugzilla.mozilla.org/show_bug.cgi?id=103916 |
| 11:28 | <annevk> | that bug doesn't mention IPv6 |
| 11:28 | <smaug____> | but look at the patch |
| 11:29 | <annevk> | whoa, Alec Flett is also from back in the days |
| 11:30 | <annevk> | smaug____: but given how much code it adds I guess it's even older |
| 11:30 | <annevk> | smaug____: this seems more like a rewrite |
| 11:31 | <annevk> | hmm maybe not |
| 11:31 | <jgraham> | annevk: How many digits in the bug number do you need to make you happy? :p |
| 11:35 | <annevk> | So it's a nice legacy thing that Alec and Darin forgot to port to Chrome :-) |
| 11:36 | <annevk> | It seems pretty trivial to remove from Gecko, but compat... Of course the compat thing applies to other engines too. |
| 11:41 | annevk | filed https://bugzilla.mozilla.org/show_bug.cgi?id=960014 |
| 11:46 | <smaug____> | annevk: I'd guess it was added because IE did it |
| 11:47 | <smaug____> | but don't have old IE to test |
| 13:44 | <zcorpan> | MikeSmith: heycam|away: please can http://dev.w3.org/2006/webapi/WebIDL/ be a proper redirect? |
| 13:45 | <MikeSmith> | zcorpan: to where? |
| 13:45 | <MikeSmith> | oh that's ugly now |
| 13:45 | <MikeSmith> | I dint make that |
| 13:46 | <MikeSmith> | so heycam|away must have |
| 13:46 | <zcorpan> | likely |
| 13:46 | <zcorpan> | i'm only bugging you because heycam|away is away and you might be able to fix it :-) |
| 13:47 | <darobin> | zcorpan: I'm sure you have a dev CVS account up your sleeve somewhere :) |
| 13:48 | <annevk> | Why is João asking that question on www-dom rather than WHATWG? |
| 13:49 | <zcorpan> | true. but i haven't used cvs on this laptop so i'd have to spend some time to set things up |
| 13:50 | <MikeSmith> | zcorpan: lame :) |
| 13:50 | <jgraham> | annevk: I'm not sure João believes in WHATWG |
| 13:51 | <zcorpan> | MikeSmith: guilty :-) |
| 13:51 | <MikeSmith> | anyway I'll do it before I forget because otherwise .. I'll forget. And I do really love CVS |
| 13:53 | <MikeSmith> | Ms2ger: re forgetting, is w3c-test.org w-p-t getting updated again after pushes? |
| 13:54 | <Ms2ger> | Doesn't seem to be |
| 13:54 | <MikeSmith> | ちぇ |
| 13:55 | <MikeSmith> | OK |
| 13:57 | <MikeSmith> | "fatal: unable to access 'https://github.com/w3c/testharness.js/';: transfer closed with outstanding read data remaining" |
| 13:57 | <MikeSmith> | darobin: ↑ |
| 13:57 | <MikeSmith> | nm |
| 13:57 | <darobin> | MikeSmith: is that the error you're getting from the sync script? |
| 13:58 | <MikeSmith> | cron job |
| 13:58 | <MikeSmith> | from teh sync script yeah I guess |
| 13:58 | MikeSmith | looks at the cron job |
| 13:59 | <darobin> | MikeSmith: that looks like a submodule is unhappy |
| 13:59 | <darobin> | perhaps trying to run the script by hand would work? |
| 13:59 | <MikeSmith> | yeah will try |
| 13:59 | <darobin> | most of the time when it breaks it's a permissions problem |
| 14:01 | <MikeSmith> | "fatal: remote error: Repository not found." |
| 14:01 | <MikeSmith> | that might be a clue possibly |
| 14:02 | <MikeSmith> | darobin: url = git://github.com/w3c/html-testsuite.git |
| 14:03 | <jgraham> | Does GH still support git:// urls? |
| 14:03 | <jgraham> | I don't think you can get them from the UI at least |
| 14:04 | <darobin> | I'm pretty sure it does |
| 14:04 | <MikeSmith> | ah that's for the old path on the filesystem anyway |
| 14:04 | <darobin> | oh |
| 14:04 | <darobin> | WTF |
| 14:04 | <darobin> | that repo has not existed in ages |
| 14:05 | <darobin> | MikeSmith: you want to make that github.com/w3c/web-platform-tests |
| 14:05 | <darobin> | well, with the gits at both ends |
| 14:05 | <darobin> | something went wrong with the restore? |
| 14:05 | <MikeSmith> | darobin: makes me wonder how it's been getting update recently if now from that cron job |
| 14:05 | <MikeSmith> | *not from |
| 14:06 | <darobin> | MikeSmith: yeah, that's... bizarre |
| 14:07 | <MikeSmith> | d'ogh |
| 14:07 | <MikeSmith> | I was checking the www-data user, not the github user |
| 14:07 | <darobin> | aha! |
| 14:07 | <MikeSmith> | so not bizarre, just pilot error |
| 14:10 | <MikeSmith> | ah figured out the cause of the problem. again my fault. I had made local changes to config.json and serve.py and it's refusing to clobber them |
| 14:14 | <annevk> | zcorpan: did you test https://www.w3.org/Bugs/Public/show_bug.cgi?id=14832 as well? |
| 14:15 | <zcorpan> | annevk: nope |
| 14:17 | <zcorpan> | maybe it's relatively easy to add though. not for all possible places but <a href> (and maybe <form action>) could be enough |
| 14:18 | <annevk> | if you have tests I'd be interested in hearing adhoc results |
| 14:22 | <zcorpan> | actually i wouldn't know how to test mailto: in an automated fashion, but i recall mailto: having to be utf-8 in presto to work correctly. but that's my memory so not so reliable :-) |
| 14:25 | <MikeSmith> | zcorpan heycam|away I set up the redirect for the WebIDL spec. Should be working now |
| 14:26 | <zcorpan> | annevk: data:text/html;charset=windows-1251,<!DOCTYPE html><a href="mailto:foo@bar?subject=å">foo shows literal "å" in the status bar in (new) opera while same for http: form-escapes it |
| 14:26 | <annevk> | cool |
| 14:26 | <zcorpan> | annevk: and clicking the link makes the å round-trip successfully to opera mail in both opera and firefox |
| 14:27 | <zcorpan> | annevk: but firefox falls back to utf-8 for unencodeable characters so that's not telling much |
| 14:27 | <zcorpan> | don't have ie right now |
| 14:28 | <zcorpan> | i guess javascript: and data: can be automated |
| 14:29 | <zcorpan> | and mailto: for getting .href, too |
| 14:31 | <zcorpan> | data:text/html;charset=windows-1252,<!DOCTYPE html><a href="mailto:foo@bar?subject=å">foo<script>alert(document.links[0].href)</script> gives mailto:foo@bar?subject=%C3%A5 in both opera and firefox. and now the character is encode-able so looks like firefox uses utf-8 for mailto: |
| 14:33 | <zcorpan> | data:text/html;charset=windows-1252,<!DOCTYPE html><a href="data:text/plain;charset=windows-1252,?å">foo<script>alert(document.links[0].href)</script> -> utf-8 in opera/firefox |
| 14:34 | <zcorpan> | data:text/html;charset=windows-1252,<!DOCTYPE html><a href="javascript:'?å'">foo<script>alert(document.links[0].href)</script> -> utf-8 in opera/firefox |
| 14:37 | <zcorpan> | ftp: %E5, ws: %E5 (boo!), bogus: utf-8 |
| 14:41 | <zcorpan> | ftps: utf-8 (is it supported?) |
| 14:47 | <annevk> | no |
| 14:47 | <annevk> | ftp is |
| 16:02 | <annevk> | Anyone suggestions for how to define "domain label"? |
| 16:14 | <annevk> | Where are all the opinions? Maybe I should go back to arguing about <cite> or some such... |
| 16:14 | <Ms2ger> | Ships, dammit! |
| 16:14 | <Ms2ger> | And cats! |
| 16:17 | <jgraham> | annevk: Too busy working out whether I have already gone past the point of "if you touch it you own it", meaning that I may as well just rewrite all this code |
| 16:17 | <Ms2ger> | You own it |
| 16:17 | <Ms2ger> | Following the "you ask about it, you own it" rule |
| 16:18 | <jgraham> | Ms2ger: Sadly I think that is probably the truth |
| 16:18 | <annevk> | I feel like I fixed enough URL bugs for a while. Maybe I can now try to fix the major Encoding bug... |
| 16:18 | Ms2ger | does not ask about the context |
| 16:18 | <annevk> | Unless someone has any suggestions for how to define "domain label" that is... |
| 16:44 | <Ms2ger> | Hixie, ping |
| 17:15 | <JonathanNeal> | would anyone confirm that quirksmode behaviors are not winning @ http://sandbox.thewikies.com/notype.php ? |
| 17:18 | <jsbell> | JonathanNeal: FWIW, they look identical to me in Chrome 31 |
| 17:36 | <GPHemsley> | They look identical to me in Aurora 28, too |
| 17:39 | <GPHemsley> | annevk-cloud: "A *domain label* is a string of characters identifying a level of a _domain_." |
| 17:39 | <GPHemsley> | annevk-cloud: How's that? |
| 17:41 | GPHemsley | wonders why 'host', 'domain', and 'domain label' are "defined" in section 4.0 when their actual definitions are in section 4.2... |
| 17:42 | <GPHemsley> | actually, s/'host',// |
| 17:43 | <GPHemsley> | actually, more usefully, s/'host'/'IPv6 address'/ |
| 17:45 | <GPHemsley> | annevk-cloud: I also recommend s/Otherwise, $host is a domain, return its _labels_ separated from each other by U+002E./Otherwise, $host is a domain; return its _domain labels_ separated from each other by U+002E./ |
| 17:45 | <GPHemsley> | (Step 3 of *host serializer*) |
| 17:46 | <GPHemsley> | Because presumably "it" refers to $host not _domain_ |
| 17:46 | <GPHemsley> | (forgot the markup for _domain_ in my suggestion) |
| 17:48 | <MikeSmith> | annevk-cloud: A domain label is one or more URL code points other than the code points U+002E, U+3002, U+FF0E, and U+FF61. |
| 18:43 | <dglazkov> | good morning, Whatwg! |
| 19:59 | <Domenic_> | Anyone remember the title of marcosc_'s bookmarking spec? |
| 20:04 | <Domenic_> | found it http://w3c-webmob.github.io/installable-webapps/ |
| 20:13 | <Domenic_> | are MS's GestureEvents specced anywhere? |
| 20:23 | <bterlson> | Domenic_: http://msdn.microsoft.com/en-us/library/ie/dn433243(v=vs.85).aspx might be helpful |
| 20:24 | <Domenic_> | bterlson: yeah, I kind of meant proposed for other browsers to implement, i guess. |
| 20:24 | <Domenic_> | was trying to counter a coworker's claim that you needed a phonegap-like shell to get a platform's native gesture events, by saying the open web has that covered |
| 20:24 | <Domenic_> | but i guess it doesn't yet |
| 20:33 | <Hixie> | Ms2ger: pong |
| 20:34 | <Ms2ger> | Hixie, can you fix https://www.w3.org/Bugs/Public/show_bug.cgi?id=24301 soon? Gecko wants to implement |
| 20:34 | <Hixie> | just started on it 5 minutes ago |
| 20:35 | <Ms2ger> | Great, thanks |
| 20:36 | <Hixie> | wonder why the w3c version has both width and height |
| 20:36 | <Hixie> | that doesn't make much sense |
| 20:39 | <Hixie> | wow, the w3c spec text for this really makes no sense |
| 20:39 | Hixie | closes that tab and disregards it |
| 20:40 | <marcosc_> | Domenic_: that's the use cases doc... the spec is manifest.sysapps.org |
| 20:40 | <Domenic_> | marcosc_: right, yeah, misspoke |
| 20:40 | <Domenic_> | the use cases doc is more fun to read :) |
| 20:40 | <Domenic_> | especially now that dont-share-cookies-and-stuff is gone :( |
| 20:41 | <marcosc_> | heh, need to rename more properties |
| 20:41 | <marcosc_> | "and-what-do-you-think-you-are-doing": |
| 20:41 | <Hixie> | btw i fixed a few things in html that were needed by that |
| 20:42 | <Hixie> | i think the only thing html doesn't handle itself now is the orientation thing |
| 20:42 | <marcosc_> | Hixie: yep |
| 20:42 | <marcosc_> | awesome, thanks for doing that |
| 20:42 | <marcosc_> | still trying to work out orientation with mounir |
| 20:43 | <Hixie> | yeah i looked at orientation but it wasn't clear that html was the right place for it really, nor that i had the requisite expertise |
| 20:43 | <Hixie> | so i punted on that |
| 20:43 | <Hixie> | let me know if anything else comes up |
| 20:44 | <marcosc_> | do you have any pointers on what you changed? |
| 20:44 | <marcosc_> | (been away last 6 weeks) |
| 20:44 | <Hixie> | not at hand |
| 20:45 | <marcosc_> | I'll check the change logs |
| 20:45 | <Hixie> | at least r8348 and r8349, looks like |
| 20:46 | <Ms2ger> | marcosc_, slacker :) |
| 20:48 | <TabAtkins> | zcorpan: No, unfortunately. In fact, I still have a tab open for that as a reminder. I've never gotten around to getting a working Blink install again to add this myself. |
| 20:50 | <marcosc_> | Hixie: the i18n support for application-name there looks good. Icon stuff looks good too. |
| 20:51 | <marcosc_> | I've been testing icon support and it's fairly good in IE and Chrome, but fairly sad in Gecko |
| 21:04 | <Hixie> | hmmmm |
| 21:04 | <Hixie> | Ms2ger: if you create an ImageBitmap from an ImageData object, what density should the ImageBitmap have? |
| 21:04 | <Hixie> | if we're using constructors, there's no way to forward the density from the canvas |
| 21:04 | <Hixie> | since there's no canvas.. |
| 21:04 | <Ms2ger> | E_DONTSUPPORTDENSITY |
| 21:05 | <Hixie> | ? |
| 21:06 | <Hixie> | oh like make drawImage() always assume 1:1 source data to backing store pixel? |
| 21:06 | <Hixie> | hmm |
| 21:07 | <Hixie> | i guess that could work |
| 21:07 | <Hixie> | whatever replaces srcset="" will have to expose the actual density, too |
| 21:08 | <Hixie> | so that you can know what to do with it |
| 21:10 | <TabAtkins> | Yeah, I was wondering if I should add a property to <img> to expose the density of the chosen resource. |
| 21:15 | <Hixie> | oh man, i painted myself into a corner with these createImageData() methods |
| 21:15 | <jory> | TabAtkins: I'd find that really useful. |
| 21:16 | <Hixie> | i have a createImageData(imagedata) method that returns a new blank imagedata of the same dimensions |
| 21:16 | <Hixie> | but an ImageData(imagedata) constructor should obviously be a copy constructor |
| 21:16 | <Hixie> | and having both do different things would be highly confusing |
| 21:16 | <Hixie> | what a mess |
| 21:24 | <Hixie> | Ms2ger: should this constructor copy the data? or reference the underlying ArrayBuffer object? |
| 21:25 | <Ms2ger> | No opinion |
| 21:25 | <Ms2ger> | Maybe bz has one |
| 21:52 | <heycam> | MikeSmith, thanks |
| 22:10 | <Hixie> | jgraham: ping (you're the one maintaining hoppipolla.co.uk right?) |
| 22:11 | <Hixie> | jgraham: bugs update thing seems to be broken |
| 22:12 | <Ms2ger> | He is |
| 22:15 | <jgraham> | Hixie: What thing now? |
| 22:16 | <Hixie> | jgraham: looks like the bugs update thingy that you have hasn't updated in fa ew months |
| 22:16 | <Hixie> | in a few |
| 22:19 | <Hixie> | https://www.w3.org/Bugs/Public/show_bug.cgi?id=24298 |
| 22:20 | <Hixie> | the whatwg vs w3c thing isn't even causing as much confusion as the w3c vs w3c thing |
| 22:20 | <Hixie> | sigh |
| 22:38 | <arunranga> | I think it's just confusion with workflow tools and bad labeling ^Hixie. |
| 22:44 | <JonathanNeal> | ah, whatwg versus whichw3c |
| 22:45 | <Hixie> | arunranga: it's the natural outcome of the w3c process, sadly. |
| 22:45 | <Hixie> | arunranga: even with labeling, people end up reading random old TR/ drafts |
| 22:47 | <arunranga> | Maybe. |
| 22:47 | <arunranga> | Hixie, do you have an opinion on https://www.w3.org/Bugs/Public/show_bug.cgi?id=17765 |
| 22:47 | <Hixie> | no maybe about it, people are confused about that all the time |
| 22:47 | <Hixie> | looking... |
| 22:48 | <Hixie> | arunranga: i have many opinions on it, but nothing really helpful, unfortunately |
| 22:48 | <arunranga> | Hixie: The problem is implementations shake out bugs with /TR specs VERY fast these days. |
| 22:49 | <Hixie> | no, the problem is that TR/ specs exist at all |
| 22:49 | <Hixie> | they serve no useful purpose except for patent lawyers and government contract lawyers |
| 22:49 | <Hixie> | and should be treated accordingly |
| 22:50 | <arunranga> | Hixie, re: bug 17765, MY opinion is we don't necessarily need a URL parse bit. Fetch alone should solve this for mediastream, blob, etc. |
| 22:50 | <Hixie> | (e.g. having a snapshot directory that just automatically archives versions every 6 months with a boring style sheet and a clear heading like "January 2014 Snapshot of the Foo Bar Standard for the Purposes of Patent Lawyers and Government Officials") |
| 22:51 | <arunranga> | Hixie: unfortunately, /TR exists directly to serve the patent issues. But even within /TR, it's now become commonplace to cross-refer to WHATWG specs. |
| 22:51 | <Hixie> | TR/ doesn't exist to serve the patent issues |
| 22:51 | <Hixie> | it exists because the w3c thinks their process is good |
| 22:51 | <arunranga> | So /TR specs have normative WHATWG refs. |
| 22:51 | <Hixie> | if it was just patent issues, the process would be very different |
| 22:51 | <arunranga> | Hixie: umm, process is an upshot of patent circumvention IHMO |
| 22:51 | <arunranga> | s/IHMO/IMHO |
| 22:53 | <Hixie> | arunranga: no, come on, that's not even close to true. the process existed in much the same as its current form long before the patent stuff was added. |
| 22:53 | <Hixie> | arunranga: the process predates the w3c, for one |
| 22:53 | <Hixie> | arunranga: it's basically the same as the IETF's |
| 22:53 | <Hixie> | arunranga: if all the process was for was patents, it would be much more light-weight and would just consist of automatic snapshots and a review period for those snapshots |
| 22:54 | <Hixie> | arunranga: it wouldn't have any of the WD/LC/CR/PR/REC crap |
| 22:54 | <Ms2ger> | Mmm, more process arguments |
| 22:55 | <arunranga> | Hixie, Ms2ger, I'm not actually *arguing.* |
| 22:55 | <Ms2ger> | Mmm, more process talk |
| 22:57 | <arunranga> | Hixie, File/FileSystem are cases in point about /TR being overridden, so no *real* arguments. However, till ALL browser vendors bail uniformly on process, we're in it. |
| 22:58 | arunranga | supposes he has to wait for annevk (annevk-cloud) in order to fix https://www.w3.org/Bugs/Public/show_bug.cgi?id=17765 properly |
| 22:58 | <Hixie> | i'm not sure what you mean by "till ALL browser vendors bail uniformly on process, we're in it" |
| 22:58 | <Hixie> | all the browsers _have_ bailed |
| 22:58 | <Hixie> | nobody references TR/ specs for what to implement except by accident |
| 22:59 | <Hixie> | i mean, do you see any browsers who have skipped out on the improvements to HTML since 1999? |
| 23:00 | <Domenic_> | I feel fortunate to have gotten the message "see TR in URL = outdated, click on latest ED link asap" early in my spec-reading career |
| 23:00 | <arunranga> | Hixie: *sigh NO, but some of the post-99 improvements were done under W3C's watch, at least as far as IE team's external claims |
| 23:01 | <Hixie> | arunranga: if it was only the IE team left, they'd quickly move on. |
| 23:02 | <arunranga> | Hixie, ANYway, I still have no real love for /TR (and less for LCWD --> RC). |
| 23:14 | <Hixie> | arunranga: feel free to not use /TR :-) |
| 23:14 | <Hixie> | nobody's forcing you :-) |
| 23:33 | <TabAtkins> | miketaylr: Yo, you have a corpus of JS to run regex on, right? |
| 23:35 | <miketaylr> | TabAtkins: yeah, just a folder on my HD right now, hopefully something more structured in the near future |
| 23:36 | <TabAtkins> | miketaylr: kk. I'm just wondering about checking for any usage of CSSCharsetRule, CHARSET_RULE, or ".encoding". |
| 23:37 | <miketaylr> | TabAtkins: sure thing, will grep for those later tonight and ping you (about to peace out for dinner) |
| 23:37 | <TabAtkins> | Thanks! |