| 01:25 | <MikeSmith> | what does "QoI testing" mean? |
| 01:25 | <MikeSmith> | http://www.w3.org/2012/06/26-coremob-minutes.html#item06 |
| 01:25 | <MikeSmith> | quality of implementation? |
| 01:37 | <zewt> | (what does "core mobile web platform" even mean? sounds roughly opposite the design of the platform) |
| 02:35 | <Hixie> | has the CSS 'cursor' property gone out of favour? |
| 02:35 | <Hixie> | i can't get it to use a png as a cursor image on either chrome or firefox |
| 02:38 | <zewt> | seems to work with .ico and .png files, but not .jpg |
| 02:39 | <zewt> | (not sure why you'd want to use a format without alpha as a cursor, but it's pretty weird that it doesn't work at all) |
| 02:39 | <Hixie> | maybe it's the size that matters... |
| 02:40 | <Hixie> | aha, yes |
| 02:40 | <Hixie> | browsers just don't seem to resize cursors down |
| 02:40 | <zewt> | https://zewt.org/~glenn/temp.html |
| 02:40 | <zewt> | do they fall back if you give multiple size cursors and earlier ones are too big? |
| 02:40 | <gsnedders> | (Opera doesn't support cursor: url(…) at all) |
| 02:41 | <Hixie> | zewt: looks like yes |
| 02:41 | <zewt> | looks like no on my quick test |
| 02:41 | <zewt> | in ff |
| 02:42 | <Hixie> | yes in chrome |
| 02:42 | <zewt> | yeah |
| 02:42 | <Hixie> | also looks like both accept cursors up to 128x128 on mac |
| 02:42 | <gsnedders> | OS limit? |
| 02:42 | <zewt> | ff seems to be loading both cursors, then ignoring both when the first doesn't work and jumping to the fallback cursor |
| 02:43 | <zewt> | gsnedders: yeah, i think the cursor size is an OS thing (but the fallback issue is a browser one) |
| 02:43 | <zewt> | oh, chrome also falls back properly |
| 02:43 | <zewt> | so ff just doesn't handle them so well |
| 02:43 | <zewt> | afk a few |
| 08:20 | <zcorpan> | Yuhong: i've reworded the intro a bit, lemme know what you think |
| 09:17 | <matjas> | ES6 Annex B is normative for web browsers (yay!), but not for other ES implementations ಠ_ಠ |
| 09:26 | <odinho> | matjas: What update did you get on the Opera bug, if any (from JIRA)? |
| 09:27 | <matjas> | odinho: none at all :( |
| 09:27 | <matjas> | waiting for hallvors to contact me with the NDA details |
| 09:31 | <odinho> | matjas: Okay, it has a patch and is reviewed, but it has to bubble up to integration, hopefully it'll be accepted. Haven't run it through regression tests though. |
| 09:31 | <matjas> | I did get notified of your comment to the ECMAScript bug, though :) |
| 10:08 | <Ms2ger> | Animations and transitions looks like they're going to be unprefixed in tomorrow's Nightly |
| 10:22 | <annevk> | what happened here: https://www.w3.org/Bugs/Public/show_bug.cgi?id=17426 ? |
| 10:23 | <Ms2ger> | Something strange |
| 10:29 | <Ms2ger> | Given an interface declared without [Constructor], should its interface object have a 'length' property? |
| 10:37 | <annevk> | Ms2ger: in WebKit and Opera, yup |
| 10:39 | <annevk> | Ms2ger: IE/Gecko, no |
| 10:39 | <annevk> | Ms2ger: for Node |
| 10:52 | <Ms2ger> | annevk, and the value? zero? |
| 10:55 | <annevk> | Ms2ger: yeah |
| 10:55 | <Ms2ger> | That matches Gecko's new bindings, let's go with that |
| 10:56 | <annevk> | poor IE |
| 10:56 | <annevk> | does it match IDL? |
| 10:56 | <Ms2ger> | I'm not entirely sure |
| 10:57 | <Ms2ger> | IDL says that the interface object is a Function object |
| 10:57 | <Ms2ger> | And ES seems to suggest that Functions get a length property |
| 10:57 | <annevk> | okay |
| 10:58 | <annevk> | would be nice if IDL had some examples of how objects would be represented in ES |
| 10:58 | <Ms2ger> | OTOH, WebIDL says "Interface objects for interfaces declared with a [Constructor] extended attribute MUST have a property named �length��" |
| 10:59 | <Ms2ger> | And by WebKit you mean V8, I guess. |
| 10:59 | <Ms2ger> | s/./?/ |
| 11:28 | <annevk> | Ms2ger: yes |
| 11:28 | <annevk> | Ms2ger: but I can test Safari |
| 11:28 | <annevk> | Safari says undefined |
| 11:28 | <Ms2ger> | So can heycam, I guess :) |
| 11:29 | <annevk> | (Version 5.1.7 (7534.57.2)) |
| 12:06 | <zcorpan> | Ms2ger: is it intentional that using xspec xref doesn't add the referenced spec to the References section? |
| 12:07 | <Ms2ger> | Sorta |
| 12:08 | <zcorpan> | is it fine that it's not in References, or am i expected to place a reference somewhere? |
| 12:26 | <Ms2ger> | I've used something like Some of the terms used in this specification are defined in Web IDL, XML and Namespaces in XML. [WEBIDL] [XML] [XMLNS] |
| 12:41 | <zcorpan> | ok |
| 13:58 | <matjas> | should CORS apply when requesting a data URL from an http:// resource? I thought `data` was a scheme as per RFC 2397 (in which case it would be a cross-origin request) |
| 13:59 | <matjas> | i’m asking because browser implementations seem to differ |
| 16:16 | <hool> | hello whatwg |
| 16:18 | <Ms2ger> | Good evening |
| 16:19 | <zcorpan> | TabAtkins_: "[, ], (, ), [, ]." in css3-syntax, i guess you meant to have {, } in there |
| 16:36 | <TabAtkins_> | Yes, I'll fix. |
| 16:41 | <JonathanNeal> | Can I access the prototype or property descriptors of className? |
| 16:41 | <TabAtkins_> | className is a string, no? |
| 16:42 | <JonathanNeal> | yes, but I can't seem to find it on Element.prototype or HTMLElement.prototype, I can override the getter / setter on an element, but I can't even find it's property descriptor. |
| 16:43 | <TabAtkins_> | Oh, I see what you're saying. |
| 16:43 | <TabAtkins_> | I *think* it should be defined on HTMLElement.prototype, but it wouldn't surprise me if it's magic. |
| 16:45 | <hool> | has anyone here worked with frame accuracy in html5 video? |
| 16:45 | <hool> | specifically with 29.97fps? |
| 16:45 | <hool> | (drop frame) |
| 16:45 | <JonathanNeal> | Yea, defineProperty(HTMLElement.prototype, 'className') will not hijack className. I can hijack an individual element via defineProperty(myNode, 'className'), but I can't getOwnPropertyDescriptor for it. |
| 16:46 | <JonathanNeal> | Which defeats the purpose of using defineProperty to hijack a getter or setter without destroying the actual functionality. I have never seen this for any other property. |
| 16:47 | <TabAtkins_> | Nah, you've got similar magic with event listeners. |
| 16:47 | <TabAtkins_> | And it's a bug that we're trying to fix. |
| 16:53 | <zcorpan> | JonathanNeal: iirc webidl says that idl attributes are supposed to be visible on the prototype, but opera and webkit don't do that (yet) |
| 17:00 | <JonathanNeal> | Well, I found a way around it. I defineProperty the individual node, then on set I use delete to remove the setter, set the value, and then restore the alternate setter. |
| 17:12 | <annevk> | matjas: data URLs are considered same-origin under certain circumstances |
| 17:12 | <annevk> | matjas: see HTML |
| 17:13 | <annevk> | matjas: hmm maybe it needs some additional wording to make that happen |
| 17:13 | <annevk> | matjas: I kind of forgot how that works |
| 17:21 | <JonathanNeal> | Thanks for the heads up on magic, TabAtkins_. I would have kept trying to do something that didn't work. |
| 17:22 | <JonathanNeal> | But now I have a classList polyfill that handles errors, enumerables, className that works in IE8+ https://gist.github.com/2956234 |
| 17:22 | <TabAtkins_> | Heh, cool. |
| 17:38 | <annevk> | hmm |
| 17:38 | <annevk> | CSSOM still not updated |
| 17:38 | <TabAtkins_> | Poke Glenn and Shane in w3c-css-wg? |
| 17:38 | <TabAtkins_> | Even if they're working on it, they should be updating live. |
| 17:39 | <annevk> | just observing what happens to stuff I've abandoned |
| 17:39 | <TabAtkins_> | Okay, so you're not actually interested in it. ^_^ |
| 17:40 | <annevk> | well I'm kind of interested in having the IDL updated |
| 17:40 | <Ms2ger> | I know complaints to the chairs have been made |
| 17:40 | <Ms2ger> | They claimed the editors were "new" |
| 17:40 | <Ms2ger> | I prefer "incompetent" |
| 17:41 | <annevk> | like the chairs? |
| 17:41 | <Ms2ger> | No comment |
| 17:41 | <annevk> | oh my, I better go back vacationing |
| 17:41 | <Ms2ger> | Or, well, start editing CSSOM again outside the WG |
| 17:47 | <annevk> | there's some minor fixing that could be done, but the best long term approach is writing out the CSS model (i.e. rewriting CSS 2.1 a la "HTML5") and then having the API map on that |
| 17:48 | <annevk> | that's the main thing I learned writing the CSSOM and I have tried to convey that at various points but the CSS WG is just not wired to think about specifications that way |
| 17:48 | <annevk> | it always turned quickly into W3C Process discussions and splitting drafts, etc. etc. |
| 17:50 | <annevk> | prescient https://twitter.com/hober/status/5511711 |
| 17:55 | <hober> | annevk: :) |
| 17:56 | <annevk> | hober: or just maybe this keeps following you around because you really are :p |
| 17:57 | <hober> | annevk: i think this is the first tweet mentioning 'html5', at least according to the kellan.io/oldtweets search: https://twitter.com/BenWard/status/882773/ |
| 17:57 | <hober> | annevk: "hober: claiming to not be the bottleneck he really is since 2007." |
| 17:58 | <annevk> | nice find |
| 18:00 | <Lachy> | Given a WebIDL dictionary defined as: dictionary D { DOMString foo = "" }, I need some clarification about what happens when I pass the value as {foo: null} |
| 18:00 | <Lachy> | I initially thought it would stringify as "null", but reading WebIDL, just made me a little confused. |
| 18:01 | <Lachy> | oh, no. nevermind. I was misreading the algorithm. |
| 18:10 | <Ms2ger> | Yes, "null" |
| 18:39 | <JonathanNeal> | length, item, contains, add, remove, and toggle are enumerables on a DOMTokenList. Is this intentional? |
| 18:41 | <Ms2ger> | Yes, see http://dev.w3.org/2006/webapi/WebIDL/#es-attributes and http://dev.w3.org/2006/webapi/WebIDL/#es-operations |
| 18:44 | <JonathanNeal> | Ms2ger: I couldn't actually understand where the logic was being presented. I thought they would not be enumerable so that for each would only go over the indexes of the list. |
| 18:45 | <Ms2ger> | That's never been true in browsers |
| 18:45 | <Ms2ger> | for (var a of arr) will do that, though |
| 18:46 | <kennyluck> | Ms2ger, just curious who have made the complaints to the chairs (re. CSSOM)? |
| 18:46 | <Ms2ger> | I'm not sure I should tell you that |
| 18:47 | <kennyluck> | Okay. |
| 18:49 | <JonathanNeal> | Ms2ger: yea, I always liked how it did that on Array. To be clear, the spec does not prefer that? If you're able to chime in why, it would satisfy my curiosity, but I wouldn't be hurt if you didn't. |
| 18:50 | <Ms2ger> | Probable web-compat issues, like usual :) |
| 18:51 | <annevk> | there's no compat issues for DOMTokenList though |
| 18:52 | <JonathanNeal> | annevk: do you know why that's the preference then? |
| 18:53 | <annevk> | probably just not considered |
| 18:54 | <kennyluck> | It's indeed quite annoying to see Glenn post to almost every mailing list and join HTML WG calls but not edit the spec for which he is an editor. |
| 18:54 | <annevk> | JonathanNeal: could you file a bug on DOM? |
| 18:55 | <JonathanNeal> | if it's a bug, I will, yes. |
| 18:55 | <Ms2ger> | It would be rather silly to make DOMTokenList inconsistent with all the other lists, though |
| 18:55 | <annevk> | JonathanNeal: well it's something we can consider changing |
| 19:02 | <JonathanNeal> | Ms2ger: what other lists are there? |
| 19:02 | <JonathanNeal> | Besides Array, of course. |
| 19:03 | <annevk> | JonathanNeal: NodeList, StyleSheetList, HTMLCollection, etc. |
| 19:03 | <Ms2ger> | NodeList, PropertyNodeList, HTMLCollection, HTMLOptionsCollection, HTMLPropertiesCollection, DOMTokenList, DOMSettableTokenList, ClientRectList, PaintRequestList, TouchList, FileList, SVGLengthList, SVGNumberList, SVGPathSegList, SVGPointList, SVGTransformList, |
| 19:03 | <Ms2ger> | Are just a few |
| 19:04 | <JonathanNeal> | Right, duh. My bad. |
| 19:05 | <JonathanNeal> | length, and item seem to be enumerable on NodeList. What's the link to file a ticket? |
| 19:05 | <BenoitRen> | Hi there. |
| 19:05 | <annevk> | JonathanNeal: see http://dvcs.w3.org/hg/domcore/raw-file/tip/Overview.html |
| 19:06 | <BenoitRen> | Does anyone know where I can find resources on how to mark up a message board (commonly called forums) semantically? I've been drawing a blank for a long time and googling isn't helping. |
| 19:11 | <annevk> | BenoitRen: what problem are you running into? |
| 19:14 | <annevk> | friend of mine once did that exercise but it seems it's no longer online |
| 19:15 | <BenoitRen> | annevk: It's not a specific problem, really. I don't even know where to start. |
| 19:16 | <annevk> | BenoitRen: seems like it would just be a bunch of <article>'s with some content inside them really |
| 19:16 | <annevk> | BenoitRen: with topic overview lists as lists, etc. |
| 19:16 | <annevk> | BenoitRen: kind of depends on the style you're going for too |
| 19:17 | <BenoitRen> | annevk: But the format is always: "poster X with this avatar and this location and amount of posts says <article> and here's some a bunch of buttons to quote it oh and here's a PM button". |
| 19:19 | <BenoitRen> | annevk: That's the main problem I'm stuck on. How do I mark something like that up semantically? |
| 19:19 | <annevk> | BenoitRen: that's just some <p>s with classes |
| 19:19 | <annevk> | BenoitRen: maybe <aside> if it's not too relevant |
| 19:19 | <annevk> | BenoitRen: and I would put that inside the <article> |
| 19:21 | <TabAtkins_> | <article class=thread><article class=post><address>Urist McDwarf <img src></address> <p>Blah blah blah</article>...</article> |
| 19:21 | <BenoitRen> | annevk: It's that simple? I find that hard to accept... How would that read? |
| 19:22 | <annevk> | BenoitRen: simple is good :) |
| 19:22 | <annevk> | Ms2ger: https://www.w3.org/Bugs/Public/show_bug.cgi?id=17318 I'm not sure I follow entirely |
| 19:23 | <BenoitRen> | annevk: Sure, but a HTML document like that seems weird. One paragraph is just a poster name, then the avatar, then a paragraph with just the post count, and then suddenly actual text? |
| 19:23 | <annevk> | Ms2ger: oh I do now |
| 19:23 | <Ms2ger> | Order :) |
| 19:23 | <TabAtkins_> | BenoitRen: Yeah, that's about right. |
| 19:24 | <annevk> | BenoitRen: I guess you would do <article> <aside> poster info </aside> post </article> |
| 19:24 | <annevk> | BenoitRen: maybe <article> <header> poster info </header> post </article> |
| 19:25 | <JonathanNeal> | annevk: I went to overview but i couldn't find a link to anything that looked like it was the right place to file a ticket, is there another place? |
| 19:25 | <annevk> | JonathanNeal: :( |
| 19:26 | <annevk> | JonathanNeal: it's right under "Participate" |
| 19:26 | <annevk> | JonathanNeal: but it's kind of bad you didn't find that, I wonder how we can make it more obvious |
| 19:26 | <annevk> | Ms2ger: do you have a fix in mind? |
| 19:28 | <BenoitRen> | How about... <article><h2>MyAwesomeUserName</h2><aside>My Cool Title</aside><aside>Posts: 9001</aside><p>My awesome post with knowledge.</p></article> |
| 19:28 | <annevk> | the username is not the header |
| 19:28 | <annevk> | heading* |
| 19:28 | <zcorpan> | JonathanNeal: which spec? |
| 19:29 | <BenoitRen> | Why not? |
| 19:29 | <annevk> | zcorpan: DOM |
| 19:30 | <zcorpan> | annevk: maybe bug should be a line on its own? |
| 19:30 | <zcorpan> | JonathanNeal: where did you look? :-) |
| 19:30 | <Ms2ger> | annevk, for? |
| 19:31 | <annevk> | Ms2ger: the replaceChild thing |
| 19:31 | <Ms2ger> | Reordering the first few steps of the algorithm, I think |
| 19:31 | <annevk> | so sometimes Gecko does throw NotFoundError? |
| 19:33 | <annevk> | hmm confusing |
| 19:34 | <Ms2ger> | Yeah |
| 19:34 | <BenoitRen> | TabAtkins_: I just read the spec on <address>. How you propose using it is valid, but I don't understand why it wouldn't be a header. |
| 19:35 | <annevk> | Ms2ger: I was hoping it would be an easy fix, postponed now :) |
| 19:35 | <Ms2ger> | Aww :) |
| 19:35 | <annevk> | Ms2ger: those algorithms are starting to become scary :/ |
| 19:36 | <Ms2ger> | The code is worse :/ |
| 19:36 | <odinho> | Opening for cleanup!!11 |
| 19:36 | <odinho> | :D |
| 19:36 | <Ms2ger> | Why are you using that name this time of the day? |
| 19:36 | <odinho> | Yeah, I really need to go home... :-/ |
| 19:37 | <odinho> | *leaving |
| 19:37 | <smaug____> | oh, odinho is odinho only in the office |
| 19:37 | <odinho> | smaug____: There's a system to it ;-) |
| 19:39 | <Ms2ger> | A method, if you like |
| 19:45 | <Hixie> | (updated my e4h proposal) |
| 19:46 | <Ms2ger> | Now, your h spec ;) |
| 19:51 | BenoitRen | feels ignored |
| 19:51 | <TabAtkins_> | BenoitRen: Sorry, was doing other things. |
| 19:52 | <TabAtkins_> | BenoitRen: The poster's name isn't a *heading*. Headings are title for their section. |
| 19:52 | <TabAtkins_> | Someone's name isn't a title for their post. |
| 19:52 | <TabAtkins_> | <header> is a different beast, it's structual markup. It's fine to put it the poster's name in a <header> if you want. |
| 19:53 | <BenoitRen> | I thought <header> was a way to cram two headings in one semantically. Like Title: Subtitle. |
| 19:53 | <TabAtkins_> | No, that's <hgroup>. |
| 19:53 | <TabAtkins_> | English is confusing. |
| 19:53 | <BenoitRen> | Oh. |
| 19:54 | <BenoitRen> | So I put the poster's name in an address along with their avatar (with empty alt attribute), then the rest of their info in an aside? |
| 19:55 | <TabAtkins_> | Sure, that sounds reasonable. |
| 19:55 | <Hixie> | BenoitRen: <header> was renamed <hgroup> a few years ago when we added what is now <header>, fwiw |
| 19:55 | <zcorpan> | Hixie: it needs xlink: support as well, right? i'm not sure the prefix story really works for writing svg and mathml |
| 19:56 | <BenoitRen> | Hixie: Ah, I see. Thanks. By the way, don't forget my W3 bug. ;) |
| 19:56 | <Hixie> | zcorpan: why xlink? i thought the svg guys were going away from xlink |
| 19:56 | <TabAtkins_> | We are. Fuck xlink. |
| 19:57 | <Hixie> | BenoitRen: i have 506 open bugs on the pile, but am prioritising the 1245 open e-mails first, sorry :-) |
| 19:57 | <BenoitRen> | Hixie: B-but I've been waiting for over 6 months! :'( |
| 19:58 | <TabAtkins_> | Everyone else has been waiting longer. ^_^ |
| 19:58 | <Hixie> | BenoitRen: oldest e-mail on my pile is from June 2009. :-) |
| 19:59 | <Hixie> | of the things i haven't intentionally deferred, i'm currently up to January 2012 or so |
| 19:59 | <Hixie> | so if your bug is older than that, let me know |
| 19:59 | <Hixie> | i might have missed it |
| 19:59 | <zcorpan> | TabAtkins_: it wasn't clear to me whether that was going to be reality or just fiction |
| 19:59 | <Hixie> | (and if it's truly urgent, e.g. a browser vendor is blocked on it, let me know and i'll prioritise it) |
| 19:59 | <TabAtkins_> | zcorpan: It's one of the less controversial breaks, so it's reality afaik. |
| 20:00 | <Hixie> | zcorpan: if it turns out xlink is important, i don't mind adding it later, but that seems like a minor issue overall |
| 20:00 | <zcorpan> | yeah sure |
| 20:07 | <BenoitRen> | TabAtkins_: What about the buttons? Things like the PM buttons is simple: just put them in a <p> below the message. But the quote button is often found at the top right of the message... :S |
| 20:08 | <Hixie> | TabAtkins_: btw, re <style scoped> (since you're the csswg member who posted on the relevant thread) -- my intent is to leave the HTML spec as is until someone can take over on the CSS side and define the cascade, @global, et al properly, and then I'll coordinate with them to have the rights hooks in the specs |
| 20:12 | <zcorpan> | BenoitRen: it's position visually doesn't mean it can't go in a <p> below the message |
| 20:13 | <BenoitRen> | zcorpan: It makes it very hard to style, though. :( |
| 20:13 | <TabAtkins_> | Hixie: fantasai and I plan to do that later this year. |
| 20:14 | <Hixie> | cool |
| 20:14 | <TabAtkins_> | BenoitRen: Yeah, that's the rub right now. It'll get fixed as Flexbox and Grid propagate out. Do what you need to do for now, and styling will catch up with the proper semantics later. |
| 20:22 | <JonathanNeal> | Woot, got it in, thanks annevk, I don't know how I missed it, but I really did. https://www.w3.org/Bugs/Public/show_bug.cgi?id=17726 |
| 20:31 | <zcorpan> | JonathanNeal: a link should also appear when you select some text in the spec, btw |
| 20:37 | <annevk> | Hixie: what's new? |
| 20:38 | <Hixie> | new where? |
| 20:38 | <Hixie> | what? |
| 20:38 | Hixie | blinks like a deer in headlights |
| 20:38 | <annevk> | a Google Plus |
| 20:38 | <annevk> | ah* |
| 20:38 | <Hixie> | oh on e4h? |
| 20:39 | <Hixie> | attributes="can have inline {substitutions}" and the new checked?={booleanexpression} or title?={title} syntax for optional attributes |
| 21:12 | <annevk> | neat |
| 21:12 | <annevk> | and there's @<test>{substitutions}</test> too right? |
| 21:14 | <annevk> | confirmed |
| 21:14 | <Hixie> | yeah |
| 21:14 | <TabAtkins_> | What does @ do, repeat as a list? |
| 21:15 | <Hixie> | nothing really, it just prevents XSS attacks by being invalid XML |
| 21:15 | <TabAtkins_> | Explain? |
| 21:17 | <Hixie> | you at your cube? |
| 21:18 | <TabAtkins_> | yeah |
| 21:18 | <Hixie> | it's hard to explain over irc. let me just come over. |
| 21:18 | <Velmont> | oh man |
| 21:18 | Velmont | was following along from home ;] |
| 21:19 | <Velmont> | But go do that high bandwidth thing. |
| 21:22 | <annevk> | Velmont: http://code.google.com/p/doctype-mirror/wiki/ArticleE4XSecurity |
| 21:22 | <gsnedders> | jwalden: I got __proto__ into Carakan again! :) I got __proto__ reverted in Carakan again! :( |
| 21:23 | <jwalden> | \o/ /o\ |
| 21:23 | <jwalden> | (I read that left to right and reacted exactly that way :-D ) |
| 21:23 | <gsnedders> | Well, at least we know how the bidi algorithm applies here! |
| 21:23 | <jwalden> | I am still busy shaving some yaks, but I can almost see something pretty underneath it, so I'm making good progress |
| 21:24 | <jwalden> | I leave for a month-long vacation probably next Wednesday, so I need to push here to get this done now |
| 21:24 | <gsnedders> | :( |
| 21:25 | <annevk> | https://twitter.com/eeppa/status/221661734824443904 :) |
| 21:26 | <Velmont> | annevk: Trying to get us to start using RDF? :P |
| 21:26 | <annevk> | depends, are you a hipster? |
| 21:27 | <Velmont> | Are we not all? :] |
| 21:29 | <gsnedders> | jwalden: I'm just hoping I still beat you. (Though jl seems to be doing most of the follow-up work now…) :P |
| 21:29 | <annevk> | relevant for whatwg⊙wo: http://i.qkme.me/3q13ei.jpg |
| 21:29 | <hober> | that should be revise *an* rfc :) |
| 21:30 | jwalden | wasn't a hipster before it was cool not to be a hipster |
| 21:30 | <jwalden> | hober++ |
| 21:31 | <hober> | annevk: also, should you post that one or should i? :) |
| 21:31 | <annevk> | hober: jwalden: have it your way: http://i.qkme.me/3q13fu.jpg |
| 21:31 | <TabAtkins_> | Depends on your pronunciation, I suppose. |
| 21:31 | <annevk> | hober: you are at work, so you can do it :) |
| 21:31 | <jwalden> | \o/ |
| 21:31 | <TabAtkins_> | "reff-seh" |
| 21:31 | <Hixie> | Velmont: basically, look up the e4x security problems; the @ thing just stops those being viable in e4h |
| 21:31 | <hober> | annevk: hahahaha. :) |
| 21:32 | <TabAtkins_> | Velmont: The confusing bit for me was that I didn't realize the @ came at the *start* of the document. That's why an e4h literal is invalid XML - it has character data before the root element. |
| 21:32 | <gsnedders> | TabAtkins_: No, it shouldn't, as it'll be the same phonemes. But, well, whether you treat it as an acronym or an initialism would change it. |
| 21:32 | <TabAtkins_> | gsnedders: That's what I'm saying, yeah. |
| 21:33 | <BenoitRen> | Well, I'm off. Thanks for the help, all! |
| 21:33 | <TabAtkins_> | hober: Got some questions about image-set() impl if you have a few. |
| 21:33 | <hober> | annevk: posted. |
| 21:33 | <hober> | TabAtkins_: ok, shoot |
| 21:34 | <TabAtkins_> | Does the Nx dimension in its syntax work the same as the Nx component in @srcset, where it adjusts the used resolution accordingly? |
| 21:35 | <hober> | it does the same intrinsic dimension calculation, if that's what you mean |
| 21:35 | <TabAtkins_> | Yeah, that's what I mean. kk. |
| 21:35 | <hober> | word. |
| 21:36 | <TabAtkins_> | To addess the "Indicating it's optional..." thread, I'm thinking of allowing someone to specify a fallback color as well, for use if the UA can't download any of them, or just doesn't want to download anything. Sound okay? |
| 21:36 | <annevk> | hober: haha, nice line to go with it :p |
| 21:36 | <TabAtkins_> | Well, really for the latter, since the former can be done by nesting it in an image(). |
| 21:36 | <hober> | TabAtkins_: isn't that what image() is for? |
| 21:37 | <TabAtkins_> | Basically, "use this color if you're so constrained you'd prefer not to download the images at all". |
| 21:38 | <jwalden> | gsnedders: yeah, might or might not happen; these yaks in the way need the work regardless, and it would be painful to hack around them |
| 21:38 | <hober> | TabAtkins_: hmm. i doubt there's much of a use case there. also, that breaks the homogeneity of the comma-separated bits |
| 21:39 | <TabAtkins_> | Eh, not really. It's like the other arguments, except with a resolution indicator of 0x. ^_^ |
| 21:39 | <hober> | well, wait a sec. couldn't you do that with image-set(url(foo.png) 1x, url(bar.ping) 2x, color(black)) or some such? |
| 21:39 | <hober> | yeah, also the arguments aren't ordered |
| 21:40 | <hober> | i assume the fallback color would come last |
| 21:40 | <TabAtkins_> | Yeah, the ordering is irrelevant here - you can distinguish between a color and an image. |
| 21:40 | <Velmont> | TabAtkins_: Yeah, I started by looking at the parser. -- I needed the opposite thing, knowing why it could be vulnerable at all :P |
| 21:40 | <hober> | which breaks the unorderedness |
| 21:40 | <TabAtkins_> | No need to put it last. |
| 21:41 | <hober> | I think ultimately YAGNI |
| 21:42 | <TabAtkins_> | I disagree. ^^; Avoiding the roundtrip can be even better than reducing the transfer size. |
| 21:43 | <Velmont> | or even put a small dataurl there :P |
| 21:43 | <TabAtkins_> | But, hm. Using image(black) to generate an image, and setting it with an extremely low x multiplier, might be sufficient. |
| 21:43 | <TabAtkins_> | Perhaps special-case 0x, or require 1x for normal images and special-case a lack of an x multiplier. |
| 21:44 | <hober> | lack-of-x-multiplier should be equivalent to 1x |
| 21:44 | <TabAtkins_> | That's what I thought. Then we'd need either 0x or some other indicator in place of an x multiplier. |
| 21:44 | <hober> | i think even if you really want such a fallback, doing it with image(black) 0.1x is better: it's weird looking, and what you're doing is weird. also, it keeps the feature straightforward |
| 21:45 | <hober> | the problem with 0x or <1x with an image that has no intrinsic size is, well. |
| 21:45 | <Hixie> | ojan: did you ever decide (when implementing this for webkit) what the intrinsic width of a seamless iframe should be when it's floating? |
| 21:46 | <TabAtkins_> | hober: The problem is identical for vector images without intrinsic sizes. |
| 21:46 | <TabAtkins_> | hober: In that case the value of the x doesn't actually do anything, and the intrinsic size is determined normally for dimensionless images, as specified in Image V" |
| 21:46 | <TabAtkins_> | Values. |
| 21:47 | <TabAtkins_> | hober: Is image-set() exposed to the web anywhere yet? |
| 21:48 | <hober> | -webkit-image-set() is enabled in ToT WebKit |
| 21:48 | <TabAtkins_> | Okay, it wasn't in m21 so I was wondering. |
| 21:48 | <hober> | m21? |
| 21:48 | <TabAtkins_> | Chrome 21. |
| 21:49 | <TabAtkins_> | Current dev channel. |
| 21:49 | <hober> | ahh. i dunno if chrome enables it |
| 21:50 | <TabAtkins_> | It would be nice if we didn't do so yet. ^_^ |
| 21:53 | <Hixie> | anyone know if any non-chrome browsers implement seamless="" yet? |
| 21:53 | <Hixie> | TabAtkins_: any suggestions on what i should make the intrinsic width of shrinkwrapping seamless iframes? |
| 21:53 | <TabAtkins_> | I assume you dont' want to expose the min-content width of the contained document. |
| 21:53 | <TabAtkins_> | So... probably 300px. |
| 21:54 | <Hixie> | i think the ideal width to expose would be the canvas shrink-wrap width, but that seems impossible to define in a way that isn't sometimes impossible to calculate |
| 21:55 | <TabAtkins_> | The canvas doesn't have a shrink-wrap width, unless you mean the document's shrink-wrap width. That's always possible to calculate, but it will expose data about the document. |
| 21:56 | <Hixie> | exposing data is fine, these are always same-origin |
| 21:56 | <Hixie> | we're already exposing the height |
| 21:56 | <TabAtkins_> | Oh, okay. Then yeah, just get the document's own intrinsic width. That's always well-defined. |
| 21:57 | <Hixie> | fair enough |
| 21:57 | <Hixie> | (what should i reference?) |
| 21:57 | <TabAtkins_> | CSS 2.1. |
| 21:59 | <Hixie> | i don't see anything in css2.1 about documents having intrinsic dimensions, unless i'm missing something. |
| 21:59 | <Hixie> | there's stuff about the root element, but that depends on the viewport |
| 21:59 | <Hixie> | and it's the viewport i'm trying to size |
| 21:59 | <TabAtkins_> | You want the fit-content size of the root element. |
| 22:00 | <TabAtkins_> | fit-content depends on min-content, max-content, and the available space, none of which are dependent on the viewport size. |
| 22:00 | <TabAtkins_> | Though, hm, actually the available space would depend on the viewport size naively. |
| 22:01 | <TabAtkins_> | So instead, just use the definition directly - max(min-content, min(max-content, available space for the iframe)) |
| 22:02 | <Hixie> | "available space for the iframe" is undefined in this case, i think (we're talking about out-of-flow or inline-blocks iframes) |
| 22:02 | <TabAtkins_> | No, it's well-defined. It's the width of the iframe's containing block. |
| 22:02 | <Hixie> | (and fit-content doesn't seem to be in css2.1) |
| 22:02 | <TabAtkins_> | Which, in the worst case, is the width of the iframe's viewport. |
| 22:03 | <TabAtkins_> | No, it's called "shrinkwrap width" or something in 2.1. fit-content is defined in Writing Modes currently. |
| 22:03 | <Hixie> | oh. css2.1 doesn't define the shrink-wrap width. :-) |
| 22:03 | <Hixie> | that's why i was asking in the first place :-) |
| 22:03 | <TabAtkins_> | (By "iframe's viewport" I mean the viewport of the document the iframe is in, not the viewport it defines for its contained document.) |
| 22:04 | <Hixie> | (sure) |
| 22:04 | <TabAtkins_> | Okay, then refer to Writing Modes, say that it's the fit-content width of the document's root element, except with the available space of the iframe used in place of the available space in the algorithm. |
| 22:04 | <Hixie> | even if i do use the css2.1 underdefined "shrink-to-fit" width, that still won't really work, i think. |
| 22:04 | <Hixie> | because if the document just consists of positioned elements, the shrink-to-fit width will be 0 |
| 22:05 | <TabAtkins_> | Yes. |
| 22:05 | <TabAtkins_> | That's a problem? |
| 22:05 | <Hixie> | well for the height i avoid the problem by fixing the width and using the canvas boundary extents or some such |
| 22:06 | <TabAtkins_> | Why are you going to that much trouble? |
| 22:06 | <TabAtkins_> | Just let it fall down to 0x0. |
| 22:06 | <Hixie> | that seems... less than ideal |
| 22:06 | <TabAtkins_> | They can set width/height on either the iframe, the document's root, or something else in there if they want. |
| 22:06 | <TabAtkins_> | You're trying to be smarter than necessary. ^_^ |
| 22:07 | <Hixie> | story of my life :-) |
| 22:08 | <Hixie> | i'm thinking of just punting and making it 300x150 if they're floating/positioning the iframe |
| 22:08 | <Hixie> | or 300xresulting-height |
| 22:08 | <TabAtkins_> | <iframe seamless> is supposed to act like you just transplanted the document's root into the outer document. |
| 22:08 | <TabAtkins_> | So just do that. |
| 22:08 | <Hixie> | yeah, that's fair enough |
| 22:08 | <TabAtkins_> | If the inner document is all positioned, shrug, its root will be 0x0, just like a container element in the same circumstances. |
| 22:08 | <Hixie> | ok |
| 22:10 | <Velmont> | Hehe. Seems easy enough to relate to. |
| 22:15 | <ojan> | Hixie: i agree with TabAtkins_. i'm pretty sure that's roughly what webkit does |
| 22:16 | <ojan> | Hixie: although i just reviewed the code. eseidel (on #webkit) wrote the code...he might know better. |
| 22:16 | <ojan> | making float do the wrong thing because of the edge case of a document only containing positioned elements would be a shame |
| 22:17 | <Hixie> | webkit seems to do weird things |
| 22:18 | <Hixie> | e.g. http://software.hixie.ch/utilities/js/live-dom-viewer/saved/1655 |
| 22:18 | <Hixie> | why scrollbars? |
| 22:18 | <Velmont> | film at 11? |
| 22:19 | <Hixie> | (with overflow:hidden: http://software.hixie.ch/utilities/js/live-dom-viewer/saved/1656 ) |
| 22:19 | Hixie | joins a CG that has published some FSA specs to see what the UI looks like when you have published FSA specs |
| 22:19 | <Hixie> | (prepare for the linked data community to think i've gone over to their side...) |
| 22:20 | <ojan> | Hixie: lol...it's a bug...there's some crazy hackery w scrollbars and seamless iframes |
| 22:20 | <Hixie> | ojan: k |
| 22:20 | <Hixie> | aw, man, i can't join the damn group without going through my AC rep |
| 22:20 | <ojan> | Hixie: i'll file these. hopefully eseidel will fix them. :) |
| 22:21 | <TabAtkins_> | Hixie: I know, it's so annoying. >_< |
| 22:21 | <ojan> | Hixie: anyways...the 0x0 behavior is what a regular element would do...so i don't see why it's problematic for a seamless iframe to do that |
| 22:21 | <Hixie> | ojan: k |
| 22:21 | <hober> | Hixie: I'm sure tvr would be thrilled that you've developed an interest in linked data :) |
| 22:21 | <Hixie> | yeah but not if i wanted to leave the group 5 minutes later :-) |
| 22:21 | <hober> | heh |
| 22:22 | <Hixie> | is anyone here either a member of a CG that has published an FSA or an individual member who isn't a member company employee who could do me a favour? |
| 22:22 | <JonathanNeal> | Hixie: do I qualify? |
| 22:22 | <Hixie> | JonathanNeal: possibly :-) |
| 22:23 | <Hixie> | JonathanNeal: i'm trying to get a screenshot of the page that lets you sign on to the FSA deed |
| 22:23 | <Hixie> | JonathanNeal: once you're a member of teh group, it should be either on this page: http://www.w3.org/community/json-ld/ |
| 22:23 | <Hixie> | or this page: http://www.w3.org/community/json-ld/spec/22/commitments |
| 22:24 | <Hixie> | oh i guess i could just create a bogus account for testing purposes |
| 22:25 | <Hixie> | except i bet w3c has a human in that loop |
| 22:25 | <Hixie> | so that wouldn't work either |
| 22:25 | <JonathanNeal> | Sorry. |
| 22:25 | <JonathanNeal> | I saw the "Join this group" button, but if I can't see the agreement without a human approving something, meh. |
| 22:26 | <Hixie> | oh i don't think you need a human to approve joining |
| 22:26 | <Hixie> | unless you don't have an account |
| 22:26 | <Hixie> | in which case don't worry |
| 22:27 | <JonathanNeal> | I definitely have some kind of account with this site, I used it to join a css group. |
| 23:21 | <TabAtkins_> | Hixie: For <iframe seamless>, I wonder if it the inner document should use the outer document's information for viewport-related media queries. |
| 23:22 | <Velmont> | Opera has something in the speeddails there. Never looked at it though. |
| 23:25 | <Hixie> | TabAtkins_: possibly, yeah. file a bug? |
| 23:26 | <TabAtkins_> | Related: Does inline SVG create a sub-document with its own viewport size? |
| 23:26 | TabAtkins_ | doesn't want to hunt for the relevant details to try and answer the question himself. |
| 23:38 | <Hixie> | TabAtkins_: svg always kinda confuses me on this front |
| 23:38 | <Hixie> | it used to be very poorly defined |
| 23:38 | <Hixie> | now it's defined, but i'm not 100% sure where |
| 23:51 | <TabAtkins_> | Ah, cool, dbaron already filed the bug about seamless and MQ. |
