| 00:05 | <zcorpan_> | kinetik: "When a media element is removed from a Document, if the media element's networkState attribute has a value other than NETWORK_EMPTY then the user agent must act as if the pause() method had been invoked." |
| 00:05 | <zcorpan_> | http://www.whatwg.org/specs/web-apps/current-work/multipage/video.html#video |
| 00:06 | <kinetik> | yes, i mentioned that later |
| 00:06 | <zcorpan_> | oh, indeed |
| 00:06 | <kinetik> | i was confused originally |
| 00:10 | <zcorpan_> | maybe i should write a tutorial or something for html5 video and other things i've doing qa for |
| 00:11 | <zcorpan_> | however i'm a bit disconnected to real-world use cases |
| 00:12 | <zcorpan_> | i guess i could look at flash sites or desktop apps for inspiration |
| 00:22 | <AryehGregor> | That's the problem, isn't it? I mean, the people who can write good tutorials are the ones who spend their time actually building websites and/or writing tutorials, while the people who know WTF they're talking about are the ones who spend their time writing specs. The two sets have a fairly small intersection. |
| 00:30 | <AryehGregor> | Tsk, tsk, hacks.mozilla.org suggesting people use multiple="true". http://hacks.mozilla.org/2009/12/w3c-fileapi-in-firefox-3-6/ |
| 00:34 | <AryehGregor> | It also forces Courier for form input instead of respecting my default monospace font. |
| 02:00 | AryehGregor | wonders why people misspell his name when it's five letters and easily copy-pastable |
| 02:00 | <AryehGregor> | I can understand mispronouncing it, but . . . |
| 02:07 | AryehGregor | doesn't care too much, once let a professor call him "Awry" for a semester without correcting him |
| 02:09 | <deltab> | that page also defines a function getByteAt that returns "the Unicode value of the character at the given index" |
| 02:12 | <deltab> | oh, the API does seem to work that way |
| 02:12 | <TabAtkins> | What a wonderfully named function. |
| 02:14 | <TabAtkins> | Ah, nice. Turns out Manu has made his argument for removing Microdata much, much weaker. |
| 02:15 | <deltab> | hmm, so there's no way to get bytes directly from a Blob? |
| 02:16 | <Dashiva> | TabAtkins: Doesn't that imply someone else (or he) could just restate the previous, stronger arguments? |
| 02:17 | <TabAtkins> | Dashiva: The 'strong' arguments were based on fairness. The remove-Microdata crowd has rebuked that very line of argument pretty thoroughly. |
| 02:18 | <Dashiva> | Do they have actual authority to do that, though? |
| 02:18 | <TabAtkins> | Debunking themselves? I don't see why not. |
| 02:18 | <TabAtkins> | It would look pretty silly to come back and say they really meant it was a good argument. |
| 02:19 | <TabAtkins> | (It wasn't, but it was better than what's there now.) |
| 02:19 | <Dashiva> | No, someone else I mean. "Those people hate microdata, but there are some of us who really believe in fairness and such..." |
| 02:20 | <TabAtkins> | Sure. It's still a crap argument. Technologies don't succeed or fail based on fairness. |
| 02:20 | <Dashiva> | Fair enough (pun noticed) |
| 02:22 | TabAtkins | didn't intend to be punny. ?_? |
| 02:22 | <Dashiva> | Oh, the pun was that fairness being irrelevant was "fair enough" |
| 02:25 | <TabAtkins> | omg |
| 02:26 | <MikeSmith> | AryehGregor: I guess you probably saw bugmail a while back related to issues you raised regarding spurious errors for wikipedia.org validation |
| 02:26 | <MikeSmith> | I think everything you reported is now fixed in the source |
| 02:27 | <MikeSmith> | you will still get a bunch of errors reported, but they are real errors |
| 02:29 | <MikeSmith> | like "Bad value bat-smg for attribute lang" |
| 02:30 | <MikeSmith> | and about 5 other instances of language tags of the form "foo-bar" |
| 02:30 | <TabAtkins> | Microdata has an experimental FF impl, doesn't it? |
| 02:31 | <TabAtkins> | Of the DOM api, I mean. |
| 02:32 | TabAtkins | doesn't want to read back through the permathread to find out if he's remembering correctly. |
| 02:42 | MikeSmith | finds http://meta.wikimedia.org/wiki/Language_code#Subdomains_that_do_not_conform_valid_ISO_639_language_code |
| 02:50 | <MikeSmith> | TabAtkins: what's the status on your CSS gradient syntax proposal? |
| 02:50 | <TabAtkins> | MikeSmith: I've got a minor change I need to make so that it interacts sanely with background-size (right now you have to guess to come up with implementable behavior). |
| 02:51 | <TabAtkins> | And then shepazu and I have been seriously discussing aligning CSS and SVG's notions of gradients, so it's possible that further changes may arrive in the near future. |
| 02:51 | <TabAtkins> | (We should see something about that on the www-style list soon.) |
| 02:51 | <MikeSmith> | cool |
| 02:57 | MikeSmith | finds CSS WG issue 179 (Add SVG equivalents to gradients proposal.) |
| 02:57 | <othermaciej> | hi MikeSmith! |
| 02:57 | <MikeSmith> | othermaciej: hey man |
| 02:58 | <othermaciej> | how are things? |
| 02:58 | <MikeSmith> | good good |
| 02:58 | <MikeSmith> | one thing I wanted to ask you about is if SVG filters have been flipped on in nightlies yet |
| 03:00 | <MikeSmith> | plus I found a crasher - https://bugs.webkit.org/show_bug.cgi?id=32357 |
| 03:02 | <othermaciej> | MikeSmith: filters are on in trunk - I don't know if they have made it to a nightly yet |
| 03:02 | <othermaciej> | thanks for reporting the crasher |
| 03:10 | <TabAtkins> | MikeSmith: Yeah, issue 179 is actually impossible! Turns out that several of the radial gradients can't be reproduced in SVG. |
| 03:10 | <othermaciej> | interesting |
| 03:11 | <TabAtkins> | Yeah, which is why we're planning on aligning our abstractions so that we can express the same things using roughly the same concepts. |
| 03:11 | <MikeSmith> | TabAtkins: well, kind of a bummer to hear that |
| 03:11 | <MikeSmith> | (I meant bummer to hear that they can't be reproduced) |
| 03:12 | <TabAtkins> | Eh, it just means that we need to lift SVG a bit, because I think the cases that it's not hitting are useful. |
| 03:15 | <shepazu> | yeah, I think there is a pretty easy way to make SVG versions of most of the gradients, but some of them are not really possible in the general case... should be easy to add them to SVG... I am not sure I like some of the CSS syntax, but that's relatively unimportant |
| 03:16 | <shepazu> | specifically, SVG doesn't allow the author to explicitly declare whether the radial gradient is elliptical or circular |
| 03:17 | <shepazu> | (though it's possible to get both results) |
| 03:21 | <TabAtkins> | That latter part is due to SVG having a slightly crazy notion of image servers that isn't sufficiently developed and general, while CSS doesn't have such an idea at all and has so far relied on url() images all working basically the same. |
| 03:22 | <TabAtkins> | I want to both generalize SVG's image-server concept, and upgrade CSS to also use it. |
| 03:25 | <TabAtkins> | Also: omg I'm stopping reading all the issue-76 threads. They are entirely uninteresting. >_< |
| 03:25 | <TabAtkins> | Also: now I need to go write a change proposal for that one dt/dd issue. Not a counter-proposal, at least. |
| 03:31 | <othermaciej> | TabAtkins: thanks for updating your Change Proposal |
| 03:32 | <TabAtkins> | np. There are several points that Manu specifically rebutted in my change proposal, and while I'd like to re-rebut them, I don't believe they're actually relevant to the proposal, so I left the mout. |
| 05:40 | <Lachy> | more calls for draconian error handling in HTML, but still no new or valid arguments to back it up. http://rebuildingtheweb.com/en/error-messages/ |
| 06:23 | <JonathanNeal> | Hey guys! |
| 06:25 | <JonathanNeal> | Just the other day, I was asked to downgrade our entire website to HTML4 because FCKEditor was trashing HTML5 elements. Have you heard about that? I got around it by "protecting" the specific tags themselves, which means they don't show up in the WYSIWYG but they do show up in the code. |
| 06:26 | <MikeSmith> | I though the FCKEditor guy was following the HTML5 work |
| 06:27 | <JonathanNeal> | 1. FCKEditor has since become CKEditor. 2 After several pages of google + searches in their forums will guide you to 0 results of FCKEditor modified to use HTML5. |
| 06:27 | <JonathanNeal> | Also, the demo of CKEditor trashes HTML5 elements as well, fwiw. |
| 06:28 | <boblet> | http://dev.fckeditor.net/ticket/4556 |
| 06:28 | <JonathanNeal> | Try "<article><h1>Foo</h1><p>Bar</p></article>" in http://ckeditor.com/demo |
| 06:28 | <boblet> | “type changed from Bug to New Feature” |
| 06:29 | <JonathanNeal> | new feature as in upcoming feature? |
| 06:29 | <boblet> | not sure if that is an attempt at comedy or an indication that it’s in the works |
| 06:29 | <MikeSmith> | I think it's simply an indication that it is a low priority |
| 06:29 | <MikeSmith> | this is not specific to new HTML5 elements, right? |
| 06:30 | <boblet> | JonathanNeal: you could try the old <div class="new-html5-element-name"> in the meantime |
| 06:30 | <MikeSmith> | I mean, it does this for any elements it doesn't know, right? |
| 06:30 | <JonathanNeal> | I tried messing with the source, checking for every place they mentioned "div" or "address" and cloned those properties for "article" and "section" but it still didn't work. |
| 06:30 | <MikeSmith> | an editor that wraps unknown elements in <p> elements would seem to be broken |
| 06:30 | <JonathanNeal> | MikeSmith, yes you're right. And FCKEditor has a "protected tags" function, but it does not actually work. |
| 06:31 | boblet | thought that FCKEditor was more … apt :| |
| 06:31 | <MikeSmith> | good times |
| 06:31 | <boblet> | as a name |
| 06:31 | <MikeSmith> | why in the world does it have this <p>-wrapping misfeature to begin with? |
| 06:31 | <JonathanNeal> | But hey, despite it all, I protected the source of <(all new html 5 elements and /)> and we have http://vm-20.liferay.com/ |
| 06:32 | <JonathanNeal> | It's still in progress, but man the thing sings to the outliner sometimes. |
| 06:32 | <boblet> | MikeSmith: because it’s FCKEd…itor, of course |
| 06:34 | <JonathanNeal> | boblet, lol yes, I tried to get help with the bug by making this demonstration in pictures @ http://sandbox.thewikies.com/fck-that/ |
| 06:35 | <MikeSmith> | WTFEditor |
| 06:36 | <MikeSmith> | marketing slogan: It's full of surprises! |
| 06:37 | <boblet> | the state of WYSIWYG is just sad. I really hope someone will do one that only supports modern browsers and doesn’t use MSWord as interface inspiration |
| 06:37 | <boblet> | but I’m not holding my breath |
| 06:40 | <JonathanNeal> | I enjoy http://en.wikipedia.org/wiki/Fckeditor#Name_change |
| 06:42 | <JonathanNeal> | But seriously, it doesn't look like there is a solution right now, so the fact that we have it working and playing nicely is a dream come true, or at least a waking dream for now. |
| 08:45 | <erlehmann> | I dont get why JonathanNeal didn't just change the source code. |
| 08:46 | <erlehmann> | Of the editor. |
| 09:57 | annevk2 | likes the idea of just using <input> for video et al |
| 10:11 | <nessy> | annevk2: how and what for? |
| 10:14 | <annevk2> | as a way of dragging video input streams and other input streams to |
| 10:14 | <annevk2> | or selecting them |
| 10:14 | <annevk2> | rather than an api that gives some kind of dialog for the task |
| 10:15 | <csarven> | If any of the terms 'external', 'internal', 'inline' are defined in the HTML4 or 5 spec, where are they? |
| 10:16 | <annevk2> | if they are, within a <dfn> :) |
| 10:18 | <csarven> | http://www.w3.org/Terms.html#database mentions the difference between external and internal wrt to the document |
| 10:21 | <zcorpan_> | "1992/updated Apr-95," |
| 10:23 | <csarven> | =) |
| 11:25 | <mitnavn> | Hey, I'm messing with the HTML 5 drag and drop. Does anyone know how to cancel the original action of the drop, so only my own function will run? return false; doesn't do it. |
| 11:26 | <zcorpan_> | e.preventDefault() ? |
| 11:26 | <mitnavn> | Shouldn't return false do it then? Anyway, let me try. |
| 11:27 | <mitnavn> | Yeah it doesn't do it. I'm dragging an image to a textarea, I want to prevent the source URL of being added to the textarea. |
| 11:32 | <zcorpan_> | what happens if you DELETE the green player? |
| 11:32 | <zcorpan_> | (sorry, public-html reference) |
| 11:33 | <annevk2> | he no longer plays |
| 11:38 | jgraham | is surprised that he doesn't get the reference |
| 11:38 | <jgraham> | Must be a thread I am not reading or something |
| 11:39 | jgraham | wonders if a whole change proposal is needed to say that <fltcap> is a stupid element name |
| 11:51 | <Dashiva> | "If a technical argument favors microdata, it is also a political argument." |
| 11:52 | <Hixie> | mitnavn: try canceling some of the other events... might be ondrop that's doing it or something |
| 11:52 | <Hixie> | jgraham: actually a www-archive reference |
| 11:53 | <zcorpan_> | oh, right |
| 11:53 | zcorpan_ | doesn't pay attention to which list he's reading |
| 11:57 | <Dashiva> | What is fltcap supposed to mean? |
| 11:58 | <Dashiva> | figure-like-thing caption? |
| 11:58 | <jgraham> | 10 points to anyone who guesses without reading the proposal |
| 11:58 | <jgraham> | (and we all know what points mean...) |
| 12:03 | <Hixie> | figure... < caption? |
| 12:03 | <Hixie> | i give up |
| 12:03 | <Philip`> | It's nearly an anagram of flatpack, which might be a clue |
| 12:05 | <Philip`> | It's no harder to guess the meaning of fltcap than of dl/dt/dd anyway |
| 12:05 | <Hixie> | dt and dd don't stand for anything in <figure> |
| 12:05 | <Hixie> | they just happen to already work |
| 12:06 | <Hixie> | it'll be interesting to see what happens with <figure> |
| 12:06 | <Hixie> | because personally that's one of the things that i dislike the most about html5 right now |
| 12:06 | <Hixie> | i much preferred <legend> |
| 12:07 | <Dashiva> | Does anyone know the answer? |
| 12:07 | Hixie | doesn't remember |
| 12:07 | <Hixie> | it's been a while since i read the proposal |
| 12:12 | <daedb> | I think it was "floating caption" or something like that |
| 12:14 | <MikeSmith> | why is it again that <legend> isn't the right solution for the long term? |
| 12:14 | <Hixie> | because people care more about the short term, sadly |
| 12:15 | <MikeSmith> | that doesn't seem like a terrifically compelling reason |
| 12:16 | Philip` | searches the mailing list and can't actually find any explanation of what fltcap means |
| 12:16 | <MikeSmith> | I would people 10 or 20 years from now are not going to be especially happy about that rationale |
| 12:16 | <MikeSmith> | *would think |
| 12:16 | <jgraham> | daedb wins although I guess he looked at the proposal |
| 12:16 | <daedb> | I liked <legend> just because it's a single element that can actually be pronounced. |
| 12:17 | <jgraham> | MikeSmith: But "don't worry about the short term" thinking leads to XHTML 2 |
| 12:17 | <daedb> | jgraham: I actually did remember it, no cheating :p |
| 12:17 | <jgraham> | daedb: But had you previously looked? |
| 12:17 | <Philip`> | daedb: It could be pronounced wrongly and cause people to think it's about feet |
| 12:17 | <MikeSmith> | jgraham: sterling logic there |
| 12:17 | <daedb> | jgraham: Yeah, weeks ago |
| 12:18 | <jgraham> | Philip`: Too much I'm Sorry I Haven't A Clue? |
| 12:18 | <Philip`> | I thought there was a Design Principle about designing features so that they work as well as possible in legacy browsers or something along those lines |
| 12:19 | <Hixie> | jgraham: you can't ignore the longterm either |
| 12:19 | <Hixie> | jgraham: one has to find a balance |
| 12:19 | <jgraham> | MikeSmith: The point is that doing stuff that people can use now is important because otherwise they get frustrated and ignore the solution. And in 10 years time everyone is so used to the workaround that no one bothers with the "official" solution anymore |
| 12:19 | <daedb> | Philip`: It keeps reminding me of felt-tip pens for some reason. |
| 12:20 | <jgraham> | or they cargo-cult the idea that the element is broekn long after the actual brokenness has vanished |
| 12:20 | <MikeSmith> | that too |
| 12:20 | Hixie | thinks we should either use <legend> or drop the elements for now and add them later when the parsers are updated |
| 12:20 | <Hixie> | which shouldn't take more than a few years, i think |
| 12:20 | <Hixie> | 5-10 at the most |
| 12:21 | <MikeSmith> | in the short term people can use whatever they're using now for marking up figures |
| 12:21 | <daedb> | Dropping is not an option imo |
| 12:21 | <jgraham> | I think that they are really useful elements today <details> in particular |
| 12:21 | <Philip`> | jgraham: I thought ISIHAC involved making up homonyms, not mispronunciations |
| 12:21 | <MikeSmith> | yeah, definitely significant demand for <details> |
| 12:22 | daedb | cares more about <figure> than <details> |
| 12:22 | <MikeSmith> | on another topic, does anybody have a clue where to report GTK+ bugs? |
| 12:22 | <Philip`> | so they'd find some joke about ends of ledges |
| 12:22 | <jgraham> | Philip`: Yeah I guess |
| 12:23 | <jgraham> | but this is IRC so obviously you have to adapt the rules |
| 12:23 | <Philip`> | MikeSmith: http://www.gtk.org/development.html |
| 12:23 | <daedb> | Whatever happened to the <p caption> idea? I kinda like that as an alternative. |
| 12:23 | jgraham | notes that he also thinks that <figure> is worthy and should go in HTML5 |
| 12:23 | <Philip`> | jgraham: You'd have to adapt the title too, since on the internet people don't apologise for being clueless |
| 12:24 | <jgraham> | We need I'm Sorry I Haven't a Change Proposal That Meets All The Formal Criteria |
| 12:25 | <jgraham> | (But Will Work On It Promptly) |
| 12:36 | <zcorpan_> | can't we make hixie go "hmm that's not a bad idea actually. ok done." for <p caption>? :) |
| 12:37 | <Hixie> | captions can have multiple paragraphs and lists and stuff |
| 12:37 | <zcorpan_> | <div caption> |
| 12:37 | <Hixie> | the whole point of <div> is it has no semantics |
| 12:37 | <jgraham> | It feels like an ugly hack |
| 12:38 | jgraham | just wishes that English had another word that meant "caption" |
| 12:38 | <annevk> | <legend> |
| 12:38 | <zcorpan_> | can we use words from another language? |
| 12:38 | <Philip`> | <subtitle> |
| 12:39 | <Hixie> | <legend> is a fine word |
| 12:39 | <Philip`> | <caption2> |
| 12:39 | <zcorpan_> | c4p710n |
| 12:39 | <daedb> | <captionator> |
| 12:40 | <Dashiva> | <caption2:captionharder> |
| 12:40 | <jgraham> | Well I think the first problem is the assumption that <figure> and <details> should have the same solution |
| 12:41 | <Hixie> | i also think <table> and <fieldset> should have the same solution |
| 12:42 | <Hixie> | (or at least three of the four, since two of them can't ever be made the same) |
| 12:42 | <Hixie> | if we had <di>, i'd want it to be the same too |
| 12:42 | <csarven> | <caption> could trigger a table layout in legacy browsers |
| 12:42 | <csarven> | <legend> is better |
| 12:42 | <jgraham> | I don't understand why you want them all to be the same |
| 12:42 | <Hixie> | that was pretty much my reasoning, yes |
| 12:43 | <jgraham> | Well I understand why you would want <table><caption> to be the same as <figure><caption> but that isn't possible |
| 12:43 | <jgraham> | But <details> seems rather different to <figure> |
| 12:44 | <Hixie> | jgraham: because i think that having parallel constructs be similar leads to better usability (untested hypothesis) |
| 12:45 | <daedb> | Does <label> have any bad issues that prevent usage in <figure>/<detail>? |
| 12:45 | <jgraham> | I suspect having identical constructs for dissimmilar things leads to poorer usability (and I think these things are somewhat dissimmilar) |
| 12:45 | <Hixie> | daedb: yeah, <label> has all kinds of issues |
| 12:45 | <jgraham> | daedb: You can't use <label> in <figure> because it prevents you putting form controls inside the caption |
| 12:45 | <jgraham> | and has legacy issues |
| 12:45 | <daedb> | sigh, I suspected as much... |
| 12:46 | <Hixie> | jgraham: i don't think that having them be the same is as strong a requirement as our solution not having ugly names like "fltcap" |
| 12:46 | <Hixie> | jgraham: but i certainly wouldn't want to have them all be <x> <xcaption/> ... </x> for every value of x |
| 12:47 | <Hixie> | jgraham: if there are names that are significantly superior to <legend> but that only work for one of them, i wouldn't be opposed to exploring that |
| 12:47 | <jgraham> | Yes, I agree that fltcap is a non-starter |
| 12:48 | <jgraham> | And I agree that having n names is probably too many for the relevant n |
| 12:48 | <csarven> | <details> seems to be a superset of <figure> |
| 12:48 | <jgraham> | csarven: Huh? |
| 12:48 | jgraham | doesn't see that at all |
| 12:48 | <Hixie> | i think they're pretty distinct |
| 12:49 | <csarven> | <details> is way too generic |
| 12:49 | <Hixie> | i think <details> has more in common with <fieldset> than <figure> |
| 12:49 | <jgraham> | I agree with that |
| 12:50 | <jgraham> | (It's a pity that fieldset is so screwed up or we could just have <fieldset details> and not invent anything new) |
| 12:50 | <zcorpan_> | <fieldset closed> |
| 12:50 | <annevk> | what's up with the <something semantic> construct? |
| 12:50 | <daedb> | <key>? |
| 12:51 | <annevk> | did it start with <time>? |
| 12:51 | daedb | is looking at legend synonyms on wiktionary... |
| 12:52 | <zcorpan_> | <img usemap> was before <time> |
| 12:52 | <jgraham> | annevk: Well for fieldset it makes sense since <details> is a special kind of <fieldset> |
| 12:52 | <zcorpan_> | and ismap |
| 12:52 | <annevk> | zcorpan_, usemap was a pointer |
| 12:52 | <annevk> | but yeah... |
| 12:53 | <zcorpan_> | can we make all <fieldset>s closeable? |
| 12:53 | <zcorpan_> | i.e. say that <fieldset> is our <details> |
| 12:54 | <zcorpan_> | but only if it has a legend |
| 12:54 | <annevk> | maybe <fieldset controls> |
| 12:54 | Philip` | has made a site using lots of fieldsets that have checkboxes in their legends, that cause their content to expand/collapse |
| 12:55 | <Philip`> | (and it looked like it worked reasonably nicely) |
| 15:30 | <AryehGregor> | MikeSmith, yeah, I noticed that, thanks. Changing the subdomains will be a pain, and probably won't happen too soon, but we can change the lang="" pretty easily. I'll bug someone about it. Is validator.nu up-to-date? It looks like there are still a couple of warnings raised as errors. |
| 16:04 | <AryehGregor> | Apparently there are some Wikipedias in languages so obscure that the English Wikipedia doesn't have an article on them. |
| 16:04 | <AryehGregor> | http://cbk-zam.wikipedia.org/wiki/El_Primero_Pagina |
| 16:34 | <JonathanNeal> | Hello! |
| 17:01 | <MikeSmithX> | AryehGregor: I looked at a few of those invalid language tags case-by-case |
| 17:02 | <MikeSmithX> | e.g., for the "bat-smg" case, they should be using just "sgs" |
| 17:02 | <MikeSmithX> | which is not in the IANA registry yet |
| 17:02 | <MikeSmithX> | but http://www.sil.org/iso639-3/cr_files/2009-050_sgs.pdf seems to indicate there's already been a decision about it |
| 17:02 | <AryehGregor> | I concluded "bat" or "lit" would be best for now, for that one. |
| 17:03 | <AryehGregor> | Is there a code for "no code assigned yet"? |
| 17:03 | <MikeSmithX> | I dunno |
| 17:03 | <AryehGregor> | Or extensions for languages with no assigned code? |
| 17:03 | <MikeSmithX> | yeah there is that, sorta |
| 17:03 | <MikeSmithX> | can do this: "bat-x-smg" |
| 17:04 | <MikeSmithX> | the "x" means it's for "private use" |
| 17:04 | <MikeSmithX> | which is not exactly the right semantic for this case |
| 17:04 | <MikeSmithX> | but close enough |
| 17:05 | <MikeSmithX> | anyway, Richard Ishida says that any new tags for the form "aaa-bbb" are always wrong |
| 17:05 | <MikeSmithX> | if anything, they should be "aaa-bbbbb" |
| 17:05 | <MikeSmithX> | that is, the subtag after hyper should be at least 5 characters long |
| 17:05 | <MikeSmithX> | if it is indicating a dialect |
| 17:06 | <MikeSmithX> | that form indicates it's a "variant" |
| 17:06 | <AryehGregor> | Interesting. |
| 17:07 | <MikeSmithX> | s/hyper/hyphen/ |
| 17:12 | <MikeSmithX> | AryehGregor: anyway, in general I would suggest that they should go through the registration procedure for registering actual new tags for all those cases |
| 17:13 | <MikeSmithX> | http://www.inter-locale.com/ID/rfc5646.html#registrationProc |
| 17:13 | <AryehGregor> | They = Wikipedia? |
| 17:13 | <MikeSmithX> | yeah |
| 17:13 | <AryehGregor> | The language approval committee is a few random volunteers, I guess someone should suggest it to them. |
| 17:13 | <AryehGregor> | Probably a non-starter if it requires significant work. |
| 17:13 | <AryehGregor> | I doubt Wikipedia's standards for a new wiki are as high as the IETF or ISO or whoever for a new language code. |
| 17:14 | <MikeSmithX> | well, the fact that a Wikipedia exists for the language/dialect is a good data point that it's important enough to a not-insignificant number of people |
| 17:16 | <MikeSmithX> | anyway, I would suggest that in the mean time at the language approval committee adopt a policy of using "foo-x-bar" tags for any languages that don't have standard tags |
| 17:16 | <AryehGregor> | No it's not, one person can approve a new wiki if no one actually objects. |
| 17:16 | <MikeSmithX> | oh, OK |
| 17:16 | <AryehGregor> | That person being some random dude who's doing it because no one else cares. |
| 17:16 | <MikeSmithX> | well then, they really should be doing the "foo-x-bar" thing for those cases |
| 17:17 | <AryehGregor> | I think it's being done in some cases. |
| 17:17 | <MikeSmithX> | OK |
| 17:17 | <AryehGregor> | Most of these wikis were set up quite a while ago. I don't really know so much about language approval, I just ran into this because of the validation errors. |
| 17:18 | <MikeSmithX> | well, if there's anything more I can do to help with it, lemme know. it would be nice to have the wikipedia.org home page actually validate |
| 17:18 | <AryehGregor> | So could we just change abc-def to abc-x-def and that would be legitimate in all cases? |
| 17:18 | <AryehGregor> | Yes, I agree. |
| 17:18 | <AryehGregor> | I need to 1) figure out what to change the language codes to, 2) ask someone to do it. |
| 17:18 | <AryehGregor> | (2) is probably trivial. |
| 17:19 | <AryehGregor> | My conclusions after quickly looking up some stuff on Wikipedia were: bat-smg -> bat or lit; roa-rup -> rup; map-bms -> jav; roa-nrm -> ???; fiu-vro -> vro; cbk-zam -> ??? (doesn't even have Wikipedia article); eml -> egl or rgn; mo -> ro (?). |
| 17:19 | <MikeSmithX> | AryehGregor: yeah, I think if you change to abc-x-defgh, that would be legitimate in all cases |
| 17:19 | <MikeSmithX> | AryehGregor: yeah, that summary looks like what I had found too for the ones I looked at |
| 17:20 | <MikeSmithX> | and I suggest defgh because of the fact that real variant subtags must be at least 5 letters long |
| 17:20 | <MikeSmithX> | http://www.inter-locale.com/ID/rfc5646.html#variant |
| 17:21 | <Hixie> | wow, you know you're doing well when someone in a w3c mailing list balks at you taking several WEEKS to do something |
| 17:21 | <AryehGregor> | MikeSmithX, is eml still supposed to be an error, or only a warning? |
| 17:22 | <AryehGregor> | Apparently it's been split into egl and rgn, and I have no idea which is correct here, if either. |
| 17:22 | <MikeSmithX> | AryehGregor: I need to talk with i18n folks more about that one |
| 17:22 | <AryehGregor> | Also not sure about mo. |
| 17:22 | <MikeSmithX> | eml is "retired" |
| 17:22 | <AryehGregor> | Well, it only says deprecated for mo, so that's not an error, right? |
| 17:22 | <AryehGregor> | I take it validator.nu isn't running trunk of whatever code this is? |
| 17:22 | <MikeSmithX> | AryehGregor: yeah, that one is just a warning now, in the source |
| 17:23 | <AryehGregor> | How long until that goes live at validator.nu? |
| 17:23 | <MikeSmithX> | AryehGregor: probably a couple more weeks. Henri has some parser changes to check in |
| 17:23 | <AryehGregor> | k. |
| 17:23 | <hsivonen> | Sorry about the silence. I've been first sick and now traveling. |
| 17:24 | <hsivonen> | Past experience suggests I shouldn't try to redeploy V.nu while traveling or just before. |
| 17:24 | <AryehGregor> | :) |
| 17:24 | <MikeSmithX> | hsivonen: I got some other patches I want to bug you about anyway |
| 17:24 | <hsivonen> | MikeSmithX: ok |
| 17:25 | <hsivonen> | MikeSmithX: do you have them on pastebin? |
| 17:25 | <MikeSmithX> | AryehGregor: I have no idea why they picked "bat-" for that anyway, given that it's clearly a dialect of Lithuanian |
| 17:25 | <AryehGregor> | Heck if I know. |
| 17:25 | <MikeSmithX> | hsivonen: not yet anywhere except in my workspace |
| 17:26 | <AryehGregor> | This isn't a very professional operation, for all I know they picked it out of a hat. |
| 17:26 | <MikeSmithX> | I will try to get them to you in the next couple days |
| 17:27 | <hsivonen> | MikeSmithX: I've pushed the major parser changes. The ones that are still queued up are C++ translation-specific except for one astral NCR fix |
| 17:27 | <MikeSmithX> | OK |
| 17:28 | <MikeSmithX> | actually, stuff that I have queued up does not need to delay deployment |
| 17:28 | <MikeSmithX> | it can wait til after |
| 17:28 | <hsivonen> | (yay for test cases for the NCR thing. I did some bit masking that was right for a BMP code unit but wrong for a full code point) |
| 17:28 | <AryehGregor> | Okay, changes to the main page should be live within an hour. |
| 17:28 | <hsivonen> | my current plan is to redeploy on Monday-ish |
| 17:28 | <AryehGregor> | Only ones left are eml, mo, and the compatibility character thing. |
| 17:29 | <MikeSmithX> | hsivonen: OK, I will look through what I've got and e-mail you the diffs |
| 17:29 | <AryehGregor> | Is there a simple fix for that, like the exact same string without compatibility characters? |
| 17:29 | <hsivonen> | MikeSmithX: ok |
| 17:29 | <AryehGregor> | My Unicode-fu is weak. |
| 17:30 | <MikeSmithX> | mostly simple stuff but want to get your review. e.g., making the PrudentHttpEntityResolves connection timeout and socket timeout configurable through system properties |
| 17:31 | <AryehGregor> | There we go: http://validator.nu/?doc=http://www.wikipedia.org |
| 17:31 | <AryehGregor> | So eml will remain an error for now, the other two will be demoted to warnings on redeployment? |
| 17:32 | <MikeSmithX> | eml will become a warning |
| 17:32 | <AryehGregor> | Ah, okay. |
| 17:32 | <AryehGregor> | So it will validate. Great. |
| 17:32 | <MikeSmithX> | oh no |
| 17:32 | <MikeSmithX> | sorry, no, it will remain an error |
| 17:32 | <MikeSmithX> | because it's not actually in the registry |
| 17:33 | <AryehGregor> | Right. |
| 17:33 | <hsivonen> | AryehGregor: is eml an actual ISO code? |
| 17:33 | <AryehGregor> | It's apparently "retired". |
| 17:34 | <MikeSmithX> | but it should be. I need to talk to the ltru people about getting it added. as far as I understand it, it is "retired" but still valid. so it should be listed in the registry as "grandfathered", I think, with a "deprecated" field and a "preferred-value" field. but the problem there is that the "preferred-value" field for that case is actually two tags, not one |
| 17:35 | <MikeSmithX> | s/should be/should be in the registry/ |
| 17:35 | <AryehGregor> | Yes, that's why I didn't convert it. |
| 17:35 | <AryehGregor> | Not sure which is right here, if either. |
| 17:35 | <MikeSmithX> | actually, one of the patches I have here is the same one I e-mailed George about |
| 17:35 | <MikeSmithX> | which is a change to jing |
| 17:35 | <MikeSmithX> | not to v.nu itself |
| 17:36 | <MikeSmithX> | change to enabled the jing schematron backend to emit warnings |
| 17:36 | <MikeSmithX> | by recognizing assert/@role=warning |
| 17:37 | <MikeSmithX> | and that change wouldn't have any effect on v.nu anyway |
| 17:37 | <MikeSmithX> | hmm, or I guess it would for the bring-your-own-schema case |
| 17:38 | <zcorpan_> | MikeSmithX: do you fancy fixing bug 14? :) |
| 17:39 | MikeSmithX | looks |
| 17:40 | <MikeSmithX> | that would seem to be a parser change |
| 17:40 | <zcorpan_> | why? |
| 17:41 | <zcorpan_> | xml-stylesheet processing is layered on top of the infoset |
| 17:41 | <MikeSmithX> | hmm, yeah.. just that the schema-validation code is not doing any PI checking now |
| 17:41 | <MikeSmithX> | I guess it could |
| 17:44 | <hsivonen> | I think the implementation for bug 14 should be a SAX consumer that appears as a URI-identified pseudo-schema to the UI |
| 17:47 | <zcorpan_> | maybe you could create a new document for each potential xml-stylesheet pi, and feed the document "<xml-stylesheet "+pi.content+"/>" |
| 17:47 | <zcorpan_> | and have a schema for that |
| 17:47 | <MikeSmithX> | OK, I can take a crack at that one after I get some of these other changes in |
| 17:48 | <MikeSmithX> | is <?xml-stylesheet href="<"?> not simply an XML well-formedness error? |
| 17:48 | <zcorpan_> | no |
| 17:49 | <Hixie> | why would it be a well-formedness error? |
| 17:49 | <AryehGregor> | < isn't allowed in XML attributes unescaped, AFAIK. |
| 17:49 | <AryehGregor> | I have no idea why, but . . . |
| 17:49 | <gsnedders> | That's not an attribute. |
| 17:49 | <Hixie> | PIs don't have attributes |
| 17:49 | <AryehGregor> | Well, okay, then. |
| 17:49 | <Hixie> | :-) |
| 17:49 | <gsnedders> | That's a PI, which is a random bunch of characters between <? and ?> |
| 17:49 | <AryehGregor> | Great. |
| 17:50 | <gsnedders> | AryehGregor: The reason for < being disallowed in SGML compatibility |
| 17:50 | <AryehGregor> | Wasn't XML supposed to be reasonably simple? I guess it is compared to SGML . . . |
| 17:50 | <gsnedders> | (As XML is meant to be an SGML subset) |
| 17:50 | <hsivonen> | AryehGregor: I think the DPH got forgotten pretty early |
| 17:50 | <Hixie> | SGML certainly puts XML in perspective, in terms of complexity |
| 17:51 | <Philip`> | HTML does too |
| 17:51 | <AryehGregor> | So does JSON. :) |
| 17:51 | <Hixie> | Philip`: indeed |
| 18:03 | <JonathanNeal> | So, I have multiple navigations on a page, should I title some of them differently? "Navigation" then "Sub navigation" ? |
| 18:05 | <JonathanNeal> | Another option is to have the name of the page in the heading of the navigation, like "About Us Navigation", or "Our Leadership Navigation". The other option I can think of is to just give them their title, like "Our Leadership". |
| 18:12 | <TabAtkins> | The latter is what I do for the section-nav, as differentiated from the main-nav. |
| 18:13 | <Hixie> | how convenient for shelley to suddenly decide that she's too good to reply to my e-mail, when if i did that she'd accuse me of ignoring her |
| 18:13 | <TabAtkins> | Hixie: She said she's leaving the email threads alone. Don't question it. |
| 18:15 | <Hixie> | she said that _in an e-mail thread_ |
| 18:16 | <Hixie> | it's not like she's actually leaving e-mail threads alone, just the ones that would inconveniently expose her inconsistencies |
| 19:29 | <AryehGregor> | whatwg.org is down? |
| 19:29 | <Hixie> | well that ain't good |
| 19:29 | <Hixie> | up again |
| 19:30 | <AryehGregor> | Up again? |
| 19:30 | <Hixie> | it's up |
| 19:30 | <AryehGregor> | Yeah. |
| 19:30 | <AryehGregor> | Hixie, what's supposed to happen if you have an infinite loop of iframes? |
| 19:30 | <Hixie> | the UA uses an infinite amount of RAM, i guess |
| 19:30 | <AryehGregor> | That's not what browsers do, so shall I file a bug? |
| 19:31 | <Hixie> | the browsers are probably applying the "hardware limitations" clause |
| 19:31 | <AryehGregor> | They stop after one or two nesting levels, though. |
| 19:31 | <AryehGregor> | So no hardware limitations in sight. |
| 19:32 | <AryehGregor> | Worth speccing, or better to just forget about it? |
| 19:32 | <Hixie> | the hardware limitations clause allows you to protect yourself from hitting the limits |
| 19:32 | <Hixie> | doesn't say you have to be near them |
| 19:33 | <AryehGregor> | That's pretty generous of it. Okay, then, I won't bother filing a bug. Probably not important to interop. |
| 19:35 | <zcorpan_> | is a UA that does nothing a conforming UA if it claims to do nothing as a protection against hardware limitations? |
| 19:37 | zcorpan_ | has a kitchen table that is a conforming HTML5 UA running with very limited hardware |
| 19:38 | <Hixie> | zcorpan_: the exact line is "User agents may impose implementation-specific limits on otherwise unconstrained inputs, e.g. to prevent denial of service attacks, to guard against running out of memory, or to work around platform-specific limitations." |
| 19:38 | <Hixie> | i suppose you could say your table is constrained to 0-byte documents |
| 19:38 | <Hixie> | and 0-byte URLs |
| 19:39 | <zcorpan_> | yep |
| 19:39 | <zcorpan_> | i call it the Zero Table |
| 19:39 | <gsnedders> | Can you put anything on it? Or eat off it? |
| 19:39 | <zcorpan_> | sure |
| 19:42 | <Lachy> | good morning everyone |
| 19:43 | <gsnedders> | Good evening Australia. |
| 19:43 | <gsnedders> | zcorpan_: But that means putting something on nothing, and as zero is the additive identity (for real numbers) then it is no longer zero! |
| 19:44 | <zcorpan_> | gsnedders: i didn't say the table is nothing |
| 19:45 | <Lachy> | zcorpan_, does the table have a network connection, or does it only accept input printed on cards or paper? |
| 19:47 | <zcorpan_> | Lachy: it has a network connection, but it has a limitation of 0 connections at a time |
| 20:20 | <zcorpan_> | Hixie: what's the problem with defining it in terms of the infoset? |
| 20:29 | <Hixie> | zcorpan_: ? |
| 20:30 | <zcorpan_> | Hixie: what's the problem with defining xml-stylesheet in terms of the infoset as opposed to the dom? |
| 20:30 | <Hixie> | i meant CSSOM DOM stuff |
| 20:31 | <Hixie> | not DOM DOM stuff |
| 20:31 | <Hixie> | sorry for the confusion |
| 20:31 | <zcorpan_> | ah |
| 20:31 | <zcorpan_> | ok |
| 20:35 | <zcorpan_> | "The xml-stylesheet processor will say nothing about CSSOM integration." is the wg consensus |
| 20:55 | <Hixie> | lovely |
| 21:34 | <hsivonen> | can anyone remember why the "already executed" flag on scripts was changed to "already started"? |
| 21:34 | <hsivonen> | what problem was addressed? |
| 21:35 | zcorpan_ | wasn't aware of the change |
| 21:38 | <hsivonen> | I wonder if I should freeze the presence of a for attribute at the same time src, async and defer get frozen... |
| 21:39 | <hsivonen> | hmm. problematic. |
| 21:39 | <hsivonen> | the already started flag doesn't get set for all non-defer, non-async parser-inserted scripts upon </script> |
| 21:40 | <hsivonen> | ouch |
| 21:40 | <hsivonen> | sigh |
| 21:41 | <hsivonen> | Hixie: I take it that it's intentional that scripts can cause a parser-inserted script to run later if the element didn't have non-whitespace content and didn't have the src attribute at the time the parser tried to run the script |
| 21:41 | <hsivonen> | Hixie: if that's the case, src, async and defer might need to be unfrozen at that point |
| 21:41 | hsivonen | tries to figure out what really needs to happen here |
| 21:42 | <zcorpan_> | hsivonen: opera seems to insert an empty text node for <script></script> |
| 21:43 | <hsivonen> | software.hixie.ch seems to be down :-( |
| 21:43 | <zcorpan_> | http://canvex.lazyilluminati.com/misc/dom-viewer/?%3C!DOCTYPE%20html%3E%0D%0A%3Cbody%3E%3Cscript%3E%3C%2Fscript%3E%3Cscript%3Edocument.body.firstChild.textContent%3D'w(1)'%3C%2Fscript%3E |
| 21:44 | <hsivonen> | zcorpan_: that's sad, but I can imagine the Web depending on it |
| 21:44 | <hsivonen> | we have a test case for it already in mochitest |
| 21:46 | <zcorpan_> | chrome seems to have weird behavior |
| 21:46 | <hsivonen> | how do I kill Opera on OS X when it goes crazy upon launch and keeps respawning itself with a new pid faster than I can kill it? |
| 21:47 | <zcorpan_> | sometimes the first script has no text node |
| 21:47 | <zcorpan_> | doesn't force quit work? |
| 21:48 | <hsivonen> | zcorpan_: by the time I can force quit, it has respawned itself under a different pid |
| 21:48 | <zcorpan_> | :/ |
| 21:49 | <hsivonen> | ok. now it managed to exhaust the process limit and killed my Gecko build, too |
| 21:49 | <zcorpan_> | that's pretty bad, haven't heard of such crazyness |
| 21:50 | <hsivonen> | this has happend to me multiple times with Opera |
| 21:50 | <hsivonen> | after Opera got a self-updater on Mac |
| 21:50 | <hsivonen> | anyway, I can't load the test case in Opera right now |
| 21:51 | <zcorpan_> | opera runs the script |
| 21:51 | <hsivonen> | ok |
| 21:51 | <hsivonen> | thanks |
| 21:51 | <hsivonen> | so I need to implement src/async/defer thawing in that case |
| 21:51 | <hsivonen> | I guess I should file a spec bug, too. |
| 21:53 | <hsivonen> | I wonder if I can reclaim the number of permitted processes without rebooting |
| 22:02 | <nessy> | killall ? |
| 22:02 | <hsivonen> | sigh. this script stuff became much more complicated than I expected |
| 22:02 | <hsivonen> | nessy: now the problem is that I can't start any more processes |
| 22:03 | hsivonen | reboots |
| 22:03 | <nessy> | I was going to say ... :-) |
| 22:22 | <JonathanNeal> | I notice that when I set the codec html5 video stopped working in safari. |
| 22:23 | <JonathanNeal> | Is there a way to detect the proper codec to set for the video? Does Safari not support the codec field set on the type attribute? |
| 22:24 | <zcorpan_> | do you mean the 'codecs' parameter? |
| 22:25 | <JonathanNeal> | yes |
| 22:26 | <zcorpan_> | i think it's supported but might have some bugs |
| 22:26 | <zcorpan_> | i'd point to the whatwg wiki but it's down atm |
| 22:27 | <JonathanNeal> | hancock, nice. |
| 22:27 | <JonathanNeal> | yea, i noticed that. |
| 22:36 | <JonathanNeal> | Would Content-Type: video/ogg;charset=UTF-8 throw off a browser? |
| 22:41 | <JonathanNeal> | You put in the "Please Update Your Code" song? |
| 22:45 | <hsivonen> | I hope my spec bug report was articulated well enough |
| 22:45 | <MindstormsKid> | Why is whatwg.org down? :/ |
| 22:47 | <Philip`> | MindstormsKid: Because it's broken |
| 22:48 | <Philip`> | Blame Hixie :-) |
| 22:48 | <MindstormsKid> | Is there a mirror? |
| 22:50 | <Philip`> | Of the spec? |
| 22:50 | <hsivonen> | MindstormsKid: http://dev.w3.org/html5/spec/Overview.html |
| 22:51 | <MindstormsKid> | hsivonen: ok, though I like the other one better :P |
| 22:51 | <TabAtkins> | Everything from the table of contents down is identical. |
| 22:53 | <Philip`> | No it's not |
| 22:53 | <Philip`> | It's all different colours |
| 22:53 | <MindstormsKid> | yes :P |
| 22:55 | <TabAtkins> | Identical content. >_< |
| 22:55 | <hsivonen> | colors of the spec are important |
| 22:59 | <jgraham> | hsivonen: Did you file a bug on the opera respawning thing? |
| 23:00 | <hsivonen> | jgraham: no. I whined about it here. |
| 23:00 | <hsivonen> | s/I whined/I just whined/ |
| 23:01 | <jgraham> | hsivonen: Would you mind filing a bug? Especially if you have some idea of when it reproduces e.g. after update |
| 23:01 | <hsivonen> | jgraham: I don't have an idea of the conditions required to reproduce |
| 23:01 | <hsivonen> | jgraham: which makes it hard to file a useful bug report |
| 23:03 | <jgraham> | hsivonen: Oh. Well perhaps you could file a useless one instead (at least with the version number and the OSX version) |
| 23:09 | <MikeSmith> | very nice to see that WebGL draft made publicly available, plus a public mailing list |
| 23:15 | <JonathanNeal> | I'm just leaving the codec out for now with mp4 video @ http://vm-20.liferay.com/documentation/5.2/installing-with-tomcat-6 |
| 23:24 | <aho> | https://www.khronos.org/webgl/ |
| 23:24 | <aho> | woo webgl specs :D |
| 23:24 | <aho> | (draft specs, mind you) :> |