| 00:02 | <virtuelv> | Hixie: are you asking about what should happen, or what is happening with the onbeforeunloads? |
| 00:03 | <virtuelv> | (in other words, I think I am advocating the Useneticism "Just ask the question instead of asking about asking" |
| 00:08 | <Hixie> | i was hoping someone might tell me what teh spec should say :-) |
| 00:09 | <Hixie> | seems that onbeforeunload doesn't fire in nested iframes if you navigate the parent |
| 00:09 | <Hixie> | eeenteresting |
| 00:09 | <nessy> | Hixie, you are a stirrer - I had a good laugh at your email reply to Paul :-) |
| 00:09 | <Hixie> | wasn't meant to be funny -- if he really does mean it, then we might be able to fix web storage |
| 00:10 | <virtuelv> | crikey, my DNS lookups are taking 20 seconds |
| 00:10 | Hixie | hands virtuelv 8.8.8.8 |
| 00:11 | <virtuelv> | Hixie: I know |
| 00:11 | <virtuelv> | I however don't know if I can be arsed to reconfigure my entire network |
| 00:12 | Hixie | just had to change the dhcp server |
| 00:13 | <Hixie> | hm, gecko does run the nested onbeforeunloads |
| 00:13 | Hixie | fires up virtualbox to see what IE does, since that now becomes the tie breaker |
| 00:14 | <Hixie> | aww, IE and gecko agree, nested iframes fire onbeforeunload |
| 00:16 | <Hixie> | and they agree that it is breadth first, interesting |
| 00:18 | <Hixie> | wait, no |
| 00:18 | <Hixie> | IE is inconssitent |
| 00:18 | <Hixie> | gah, Gecko is consistent with the inconsistent |
| 00:18 | <Hixie> | wtf |
| 00:19 | <Hixie> | ok maybe i just misinterpreted the first result |
| 00:19 | <Hixie> | ok depth first |
| 00:19 | <Hixie> | i can do that |
| 01:00 | <Hixie> | ok so what happens if the page has an onbeforeunload that adds a child browsing context with an onbeforeunload... |
| 01:08 | <Hixie> | http://www.hixie.ch/tests/adhoc/html/navigation/beforeunload/003.html |
| 01:08 | <Hixie> | http://www.hixie.ch/tests/adhoc/html/navigation/beforeunload/004.html |
| 01:08 | <Hixie> | what's the right behaviour? |
| 01:10 | <Hixie> | (the question being, is onbeforeunload fired for browsing contexts created during onbeforeunload? how about if they are created and inserted before browsing contexts that have already fired onbeforeunload?) |
| 01:19 | <ap> | Hixie: fwiw, https://bugs.webkit.org/show_bug.cgi?id=19418 |
| 01:24 | <Hixie> | thanks ap |
| 02:17 | <Dashiva> | Firefox could this, firefox could that. We _could_ all jump off a cliff too... |
| 02:27 | <Hixie> | hey, a tab |
| 02:33 | <cardona507> | Hixie - are the WHATWG spec and the W3C spec still textually the same? or is <device> etc the differece now? I am wondering because I like to link people to the whatwg spec and I am wondering how different it is than the W3C spec. |
| 02:35 | <Hixie> | the whatwg equivalent of the w3c htmlwg spec is http://www.whatwg.org/specs/web-apps/html5/ |
| 02:35 | <cardona507> | awesome - thanks |
| 02:35 | <Hixie> | np |
| 02:36 | <cardona507> | <device> is pretty exciting by the way |
| 02:36 | <Hixie> | the faq has a (worryingly long) explanation of the various versions of the spec, btw |
| 02:36 | <Hixie> | in case you're interested |
| 02:43 | <cardona507> | Hixie - is the idea with <device> etc and current work (including HTML5) to finally shift out of versions (4, 5, etc...) and into a fluid HTMLCurrent? |
| 02:43 | <Hixie> | yeah |
| 02:43 | <cardona507> | ncie |
| 02:43 | <cardona507> | *nice |
| 02:43 | <Hixie> | see http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2009-December/024477.html |
| 02:46 | <cardona507> | wow - thats a pretty big deal |
| 02:47 | <Hixie> | few people noticed :-) |
| 02:47 | <Hixie> | the few that did complained that i was trying to take over html from the w3c or something |
| 02:47 | <cardona507> | :) classic |
| 02:47 | <Hixie> | (not sure how that works since all the text is also being done in public-html and public-device-apis at the w3c, but anyway) |
| 02:52 | <Hixie> | well i'm never gonna hit 100 bugs today at this rate |
| 02:52 | <Hixie> | i just spent 5 hours on one bug |
| 02:52 | <Hixie> | it was a !@#%^$#&%$* hard bug, but that's besides the point... |
| 02:52 | <cardona507> | Another 495 hours and you can be done for today. :) what is the context of <device> - and who will likely implement it? could it be for a skype like service on chrome os?Are there any nightly builds that support it? |
| 03:20 | <Hixie> | cardona507: it's exploratory for now, being discussed on public-device-apis⊙wo |
| 03:30 | <smq-monstros1ty> | hello there |
| 03:30 | <smq-monstros1ty> | in regards to html5 coming from xhtml |
| 03:31 | <smq-monstros1ty> | i read in xhtml they put a trailing slash becuase of compatibility issues with some browsers |
| 03:31 | <smq-monstros1ty> | and i read that in html5 you dont need it |
| 03:31 | <Hixie> | like in <br> or <img>? |
| 03:31 | <smq-monstros1ty> | but it wouldnt be an issue if br hr meta still use it? |
| 03:31 | <smq-monstros1ty> | yeah <br /> <hr /> yeah <img /> |
| 03:31 | <Hixie> | html5 lets you do it either like html4 or like xhtml, so you can do <br> or <br/>, doesn't matter |
| 03:32 | <Hixie> | (the / is ignored) |
| 03:32 | <smq-monstros1ty> | cause i just converted www.truecarnage.org |
| 03:32 | <smq-monstros1ty> | it wasnt that much of a change but i kinda want to cater to both |
| 03:32 | <smq-monstros1ty> | is that possible? |
| 03:32 | <Hixie> | both what? |
| 03:32 | <smq-monstros1ty> | to do html5 and yet keep to standards im used to in xhtml? |
| 03:33 | <Hixie> | yes, you can do that, HTML5 allows you to include the / |
| 03:33 | <Hixie> | in HTML5 you can do <br/> and it is valid |
| 03:33 | <smq-monstros1ty> | how much of html5 is usable now? |
| 03:33 | <smq-monstros1ty> | browser compatibility? |
| 03:34 | <Hixie> | pretty much all the stuff that was already in html4, as well as many other parts |
| 03:34 | <Hixie> | there are guides on the web that discuss that in more detail |
| 03:34 | <smq-monstros1ty> | you guys deprecating <dv>? |
| 03:34 | <smq-monstros1ty> | <div> |
| 03:35 | <smq-monstros1ty> | <section> <aside> i read some stuff on that I will definitely read more on it |
| 03:35 | <smq-monstros1ty> | do you guys have icons for passing validation? |
| 03:36 | <Hixie> | <div> isn't deprecated, it should be used when you need something to hook some styles on |
| 03:36 | <smq-monstros1ty> | why did xhtml force transitional for <iframe> and target="_blank" |
| 03:36 | <Hixie> | no icons for passing validation, but we do have a validator at html5.validator.nu |
| 03:36 | <Hixie> | dunno, that was another working group |
| 03:36 | <Hixie> | html5 doesn't have a transitional anymore |
| 03:36 | <Hixie> | and it allows <iframe> and target="" |
| 03:36 | <smq-monstros1ty> | nice cause i like to use _blank and it would fail me on strict |
| 03:37 | <smq-monstros1ty> | i read w3c has accepted html5 |
| 03:37 | <smq-monstros1ty> | and its still ok to specify the types of script and styles? |
| 03:38 | <smq-monstros1ty> | <script type="text/javascript">? |
| 03:38 | <Hixie> | you can, don't need to though |
| 03:38 | <Hixie> | <script> is fine |
| 03:38 | <cardona507> | Hixie - will being a member of the HTMLwg help me get into the device api wg? |
| 03:39 | <Hixie> | cardona507: as far as i know it's just a mailing list, unless you want to call into the meetings (but i don't bother with that) |
| 03:39 | <smq-monstros1ty> | html5 is separate so there won't be a worry about bias towards browsers? |
| 03:40 | <smq-monstros1ty> | that would be the browser issue? |
| 03:40 | <Hixie> | how do you mean? |
| 03:40 | <smq-monstros1ty> | thats one huge reason i used xhtml |
| 03:40 | <smq-monstros1ty> | is that it tried to be ocmpatible |
| 03:40 | <smq-monstros1ty> | across browsers and devices |
| 03:40 | <smq-monstros1ty> | i wouldnt want to work with a language that favors IE or Safari or chrome |
| 03:40 | <smq-monstros1ty> | over another |
| 03:40 | <Hixie> | html5 is significantly more compatible with browsers and devices than either XHTML or HTML4 ever were |
| 03:41 | <smq-monstros1ty> | cool so it wont be driven by browser its its own |
| 03:42 | <smq-monstros1ty> | one thing i notice in internet explorer 8, many sites have javascript errors |
| 03:42 | <smq-monstros1ty> | big sites |
| 03:43 | <smq-monstros1ty> | espn, even id's quakelive, fancast |
| 03:43 | <Hixie> | that might have more to do with IE8 having bugs rather than the sites |
| 03:43 | <smq-monstros1ty> | i was wondering |
| 03:43 | <Hixie> | IE8 doesn't do javascript (or rather, the DOM) in a very standards-compliant way |
| 03:43 | <Hixie> | see e.g. how it does on acid3, which mostly tests DOM/JS stuff |
| 03:44 | <smq-monstros1ty> | how far are we from css3? |
| 03:44 | <webben> | Is it correct to say there is no subset of valid HTML5 documents that would also validate as HTML 4.01 or XHTML 1.0? |
| 03:44 | <Hixie> | no |
| 03:44 | <smq-monstros1ty> | www.truecarnage.org does |
| 03:44 | <Hixie> | webben: there are two HTML4 and XHTML1 DOCTYPEs that will validate as HTML5 (with a warning) |
| 03:45 | <webben> | ah okay |
| 03:45 | <Hixie> | smq-monstros1ty: css3 is a large number of specs, and i don't know how they are progressing, i'm out of touch with the css stuff these days |
| 03:46 | <smq-monstros1ty> | thats cool apreciate your prompt replies to the questions :) |
| 03:46 | <smq-monstros1ty> | i just worry about compatibility and proprietary stuff |
| 03:46 | <smq-monstros1ty> | since i dont make money and am not a business hehee |
| 03:46 | <smq-monstros1ty> | yet wanting to have viewable content on all browsers |
| 03:47 | <smq-monstros1ty> | I will read more i read a bit a while ago on <sectin> <aside> elements and the dictionary lists |
| 03:47 | <Hixie> | cool |
| 03:47 | <smq-monstros1ty> | im happy <iframe> are still in heheh i like to use iframe and name attribute so people can click a link and have it work in the inline frame |
| 03:49 | <smq-monstros1ty> | need to find a way to unify servers on irc heheh |
| 03:49 | <smq-monstros1ty> | thats a diff topic though i guess but |
| 03:49 | <smq-monstros1ty> | i often have the issue of being on another network on irc hehehe |
| 03:51 | <Hixie> | gotta go, bbiab |
| 03:52 | <cardona507> | this device api stuff is pretty cool |
| 04:37 | <Hixie> | annevk: i added a hook to html5 for xhr to close connections during unload, let me know if it needs tweaking |
| 04:52 | <Hixie> | anyone want to make a screenshot of a form control for me? |
| 04:52 | <Hixie> | i haven't quite worked out what the example will be yet |
| 04:53 | <cardona507> | Hixie - what do you have in mind? |
| 04:54 | <Hixie> | not sure yet, hold on |
| 04:54 | <Hixie> | something like a mail client's To: line with autocomplete coming from both a list="" and the native user contacts database |
| 04:55 | <Hixie> | and a second showing just using the native user contacts database |
| 04:58 | <cardona507> | I'm a little unclear - what do you mean 'autocomplete coming from both a list="" and the native user contacts database'? |
| 04:58 | <Hixie> | hold on, i'm writing up the example so it'll be clearer |
| 04:58 | <cardona507> | ok |
| 05:03 | <Hixie> | ok, regenning spec |
| 05:03 | <Hixie> | i made some ascii art versions to show what i meant |
| 05:04 | Hixie | twiddles thumbs waiting for it to finish |
| 05:04 | <Hixie> | http://www.whatwg.org/specs/web-apps/current-work/#attr-input-multiple |
| 05:08 | <cardona507> | I am firing up AI as we speak |
| 05:08 | <Hixie> | AI? |
| 05:09 | <cardona507> | adobe illustrator |
| 05:09 | <Hixie> | oooh |
| 05:15 | <smq-monstros1ty> | is there a display: still? |
| 05:16 | <Hixie> | in css? sure |
| 05:16 | <smq-monstros1ty> | display:inline-block; work? |
| 05:16 | <smq-monstros1ty> | looking through this reference---> http://www.w3schools.com/html5/tag_article.asp |
| 05:16 | <Hixie> | i wouldn't recommend using w3schools |
| 05:16 | <Hixie> | they have a lot of mistakes |
| 05:16 | <Hixie> | (i don't know if they do in that particular article) |
| 05:17 | <Hixie> | (i just mean in general) |
| 05:17 | <smq-monstros1ty> | is there an equivalent reference taglist for html5? |
| 05:17 | <Hixie> | there are some being made, or you can use the spec itself :-) |
| 05:17 | <smq-monstros1ty> | i often want to use an inline-block |
| 05:18 | <Hixie> | the spec has a radio button at the top that lets you turn off the implementation-specific stuff |
| 05:18 | <Hixie> | http://whatwg.org/html5 |
| 05:18 | <Hixie> | labeled "Hide UA text" |
| 05:19 | <smq-monstros1ty> | thank you :) |
| 05:23 | <Hixie> | MikeSmith: waiting for www-data's lock in /sources/public/html5/spec a lot |
| 05:27 | <mpilgrim> | is it my imagination, or did maciej just tell john to sit in the timeout chair and calm his body? |
| 05:27 | <cardona507> | Hixie - feedback? http://img706.imageshack.us/img706/668/ex101.png |
| 05:27 | <cardona507> | I was a little unclear what the |v were in the top line |
| 05:28 | <Hixie> | wow that's nearly perfect except for the |v bit |
| 05:28 | <cardona507> | ooops - the a on the far right of the top line will be a v |
| 05:28 | <Hixie> | the |v bit was supposed to be a drop-down arrow like in a combo box |
| 05:28 | <cardona507> | heh - yeah fixin it now |
| 05:30 | <cardona507> | like a small circle with a downward arrow in it? |
| 05:31 | <Hixie> | sure |
| 05:31 | <Hixie> | it's not especially needed |
| 05:31 | <Hixie> | do whatever you think looks best :-) |
| 05:31 | <Hixie> | i guess some platforms wouldn't have anything at all there |
| 05:39 | <cardona507> | http://img98.imageshack.us/img98/668/ex101.png |
| 05:39 | <cardona507> | feedback? |
| 05:40 | <Hixie> | i'd remove the line next to the arrow, but otherwise that's perfect |
| 05:40 | <cardona507> | Hixie - and you would also like a screenshot of the example below that one with bob@example? - removing line next to arrow |
| 05:41 | <Hixie> | yes please, if you can! |
| 05:42 | <cardona507> | no problem - here is example #1 final - http://img121.imageshack.us/img121/7262/example1final01.png - should I email it or will you just grab it off of your screen? |
| 05:49 | <cardona507> | Hixie - and here is example #2 - http://img527.imageshack.us/img527/9046/ex201.png - let me know if you need any changes or if you need them emailed. |
| 05:54 | <Hixie> | cardona507: i'll grab it from the site, thanks so much! |
| 05:54 | <Hixie> | those are really sweet |
| 05:56 | <cardona507> | thanks man |
| 05:56 | <cardona507> | glad to help |
| 05:57 | <Hixie> | are you in the acknowledgements yet? |
| 05:58 | <cardona507> | I don't think so - |
| 05:58 | <Hixie> | dude, let me add you |
| 05:58 | <Hixie> | what's your full name? |
| 05:58 | <cardona507> | sweetness! :) Carlos Gabriel Cardona |
| 06:06 | <cardona507> | Thanks Ian - I really appreciate it |
| 06:10 | <Hixie> | cardona507: thank _you_! |
| 06:11 | <Dashiva> | It's a symbiotic relationship |
| 06:21 | <cardona507> | Hixie - I just noticed that in the examples - On the first on arthur is before adam and on the second example arthur is after adam - would you like me to fix that? or am I over thinking it? |
| 06:22 | <cardona507> | *on the first one arthur... |
| 07:44 | <Hixie> | waiting for www-data's lock in /sources/public/html5/spec again |
| 08:00 | <Hixie> | i need a sanity check |
| 08:00 | <cedricv> | hey guys! is there any current work going on wrt the "notifications API" that has been removed (recently?) ? |
| 08:00 | <Hixie> | when someone refers to http://example.com/#foo, where http://example.com/ has no script and no element with id=foo, clearly something is wrong |
| 08:00 | <Hixie> | but if there's a script, #foo could be an argument to that script representing state |
| 08:00 | <Hixie> | is there some way to distinguish the two in a way that's useful in a validator? |
| 08:00 | <cedricv> | and by that question, i'm asking where, if any work is actually going on about this :D |
| 08:00 | <Hixie> | cedricv: not that i know of, though i think the chrome guys are still working on it -- #chromium might be able to help |
| 08:01 | <cedricv> | thanks Hixie, i'll poke them |
| 08:01 | <Hixie> | np |
| 08:01 | <hsivonen> | Hixie: you don't know if the page is going to be loaded in an iframe that's going to get stuff injected to it by parent |
| 08:01 | <hsivonen> | stuff being script |
| 08:01 | <Hixie> | indeed |
| 08:02 | <Hixie> | would a non-machine-checkable conformance requirement be in any way useful here? i guess not |
| 08:02 | <hsivonen> | also, I'm not keen on making the validity of a document (that contains links) dependent on what the links deference to |
| 08:02 | <hsivonen> | so I'd prefer not having that as a conformance req |
| 08:03 | <hsivonen> | I'd be OK with a non-validation link checker feature whining about that case, though |
| 08:05 | <Hixie> | k, thanks hsivonen |
| 08:05 | <Hixie> | is anyone working on webidl these days? |
| 08:15 | <hsivonen> | Does the Android browser / Mobile Safari have a way to use the interaction feature of next app view sliding in from the right? |
| 08:15 | <hsivonen> | in Web apps, that is |
| 08:15 | <hsivonen> | seems like a new challenge to CSS |
| 08:16 | <hsivonen> | sort of paginated view |
| 08:16 | <Hixie> | does that need anything more than mouse events and css transitions? |
| 08:16 | <hsivonen> | with swipes between pages |
| 08:16 | <Hixie> | oh you mean between separate pages |
| 08:16 | <Hixie> | not an app simulating this |
| 08:16 | <hsivonen> | Hixie: I mean within an app |
| 08:16 | <hsivonen> | Hixie: however implemented |
| 08:16 | <Hixie> | well just within an app seems easy enough, just detect the swipe and move the content |
| 08:17 | <Hixie> | why would it be hard |
| 08:17 | <hsivonen> | Hixie: though it would be nice for it to be Webby in terms of navigation and styling |
| 08:17 | <Hixie> | (haha famous last words) |
| 08:17 | <hsivonen> | Hixie: is there a way to turn off zooming and to occupy the view port exactly? |
| 08:17 | <hsivonen> | doesn't seem quite media and device-independent... |
| 08:19 | <hsivonen> | (it's quite annoying that in Firefox Mobile, the default viewing model works for typical big site table layouts but sucks big time on single-column stretchy pages like #whatwg logs) |
| 08:20 | <hsivonen> | CSS was supposed to address all this, but when the future came, the future was zooming desktop screen media instead of using a different layout |
| 08:20 | <hsivonen> | or then iPhone-specific pages |
| 08:20 | <hsivonen> | :-( |
| 08:21 | hsivonen | guesses the iPhone screen size and orientation is going to become incredibly sticky for years to come |
| 08:21 | <Hixie> | you can include iphone-safari flags on the iphone |
| 08:21 | <Hixie> | dunno about android |
| 08:22 | <hsivonen> | so either WGs working ahead of implementations lacked the right foresight or the Mobile Safari team suffered from NIH |
| 08:22 | <hsivonen> | (my bet is on the former explanation) |
| 08:24 | <Hixie> | or both |
| 08:24 | <Hixie> | the mobile safari team was operating in stealth mode, which made getting wg help hard |
| 08:25 | <hsivonen> | that, too |
| 08:25 | <Hixie> | and as opera has shown, media queries aren't as great in practice as one might hope |
| 08:26 | <hsivonen> | what Opera has shown, I think, is mainly that it's better to default to zooming a desktopish layout than to reformat the page |
| 08:26 | <hsivonen> | which is a nail in the coffin of semantic markup and device-independence |
| 08:26 | <hsivonen> | because Opera actually tried it the theoretically right way first |
| 08:27 | <Hixie> | yup |
| 08:28 | <hsivonen> | anyway, I should go file a bug about putting the max-width hack used by the Nokia S60 browser, Mobile Safari and Opera Mobile/Mini into Firefox Mobile, too, to make IRC logs suck less |
| 08:40 | <Hixie> | MikeSmith: yt? |
| 08:49 | Hixie | keeps seeing people use hidden="" to hide things from display and not to make them irrelevant |
| 08:50 | <Hixie> | e.g. people use hidden="" to hide things while keeping them in microdata, or to refer to them from ARIA attributes |
| 08:50 | <Hixie> | maybe we should rename it back |
| 09:02 | <Hixie> | hsivonen: dude, your bug reports are vague. :-P http://www.w3.org/Bugs/Public/show_bug.cgi?id=8374 |
| 09:05 | <hsivonen> | Hixie: what's vague about it? there's an "act as if" in there, and that's guaranteed to emit an error |
| 09:05 | <Hixie> | "there"? |
| 09:05 | <Hixie> | where! |
| 09:05 | <Hixie> | the spec has like 15 different places that imply a "select" end tag |
| 09:05 | <hsivonen> | oops :-) |
| 09:05 | <Hixie> | :-P |
| 09:06 | <Hixie> | ok my head just exploded trying to read a parser bug |
| 09:06 | <Hixie> | clearly it's time for me to take a break |
| 09:06 | <Hixie> | (http://www.w3.org/Bugs/Public/show_bug.cgi?id=8375) |
| 09:06 | <Hixie> | only 18 to go tonight |
| 09:07 | <hsivonen> | Hixie: did your head explode at the bug report or at trying to figure out how/if to fix? |
| 09:07 | <Hixie> | it failed to parse the description of the problem :-) |
| 09:07 | <Hixie> | so actually fixing it is clearly out of the question right now :-) |
| 09:08 | <hsivonen> | I guess I suck at writing spec bugs |
| 09:08 | Hixie | loads up his Yes, Minister episodes that he bought yesterday |
| 09:08 | <Hixie> | hsivonen: oh i don't think it was your fault! |
| 09:11 | <MikeSmith> | Hixie: here now |
| 09:11 | <Hixie> | MikeSmith: i figured it out |
| 09:11 | <Hixie> | afk |
| 09:13 | <hsivonen> | Hixie: so, my bug report sucks |
| 09:13 | hsivonen | wonders how I ended up filing it like that |
| 09:15 | <hsivonen> | looks like I mentally or concretely inlined an "as if" piece and then complained about a bad "(fragment case)" assertion that resulted from the inlining |
| 09:24 | <hsivonen> | Hixie: I commented on the bug |
| 09:24 | Lachy | thanks the chairs for holding a straw poll on microdata during the holiday period, while I was away and not monitoring e-mail much, and thus didn't know about it till now. |
| 09:25 | <Lachy> | oh well, I suppose it doesn't matter that much as it's non-binding and my position is already fairly well known |
| 09:25 | <hsivonen> | Lachy: when did your holiday period start? |
| 09:25 | <annevk> | and someone will always be away... |
| 09:26 | <Lachy> | around 6 December |
| 09:28 | <Philip`> | Clearly the group should avoid making any decisions while Lachy is away |
| 09:28 | <annevk> | Lachy, holiday season starts around december 24... |
| 09:29 | <annevk> | (if there is any at all) |
| 09:31 | <hsivonen> | Lachy: dude, you have long holidays even by Nordic standards |
| 09:34 | <Lachy> | annevk, maybe for you northern hemisphere people who are used to taking long holidays during the middle of the year. But for southern hemisphere, December/January is the major holiday period |
| 09:34 | <Lachy> | hsivonen, my holiday was about 3 weeks long |
| 09:35 | <Lachy> | though, admittedly, it did start a little earlier than usual cause I had to get to Aus for my sister's wedding |
| 09:36 | <annevk> | anyway, my point is that you can't expect an entire WG to plan around you |
| 09:36 | <Lachy> | no, that's not what I'm saying. |
| 09:37 | <annevk> | jaja |
| 09:37 | <Lachy> | I'm saying that mid-to-late december, through to about mid-January is a common holiday period for a lot of people, and should generally be avoided for significant things |
| 09:43 | <annevk> | a lot? |
| 09:48 | <Lachy> | annevk, yes. School holidays start around mid-to-late December in Australia, and end around the end of January. I'm sure it's the same in other southern countries too, so that would equal a lot |
| 09:49 | <Lachy> | other people tend to take time of work around the same time too, which is why it's considered a peak holiday season |
| 09:49 | <annevk> | "So Microsoft would not object to us changing the behaviour of onhashchange="" or the semantics of postMessage() to be incompatible with what was implemented in IE8?" |
| 09:49 | <annevk> | lol |
| 09:49 | <annevk> | Lachy, there's not a lot of people there in general |
| 09:49 | <hsivonen> | localStorage FTW! |
| 09:49 | <annevk> | heh |
| 09:51 | <Lachy> | annevk, just because it's most likely fewer people than there are in the northern hemisphere, doesn't mean it's not a lot |
| 09:51 | MikeSmith | wonders when Hixie made the value attribute on meter a required attribute |
| 09:52 | <Lachy> | anyway, arguing with you about this seems pointless |
| 09:58 | <annevk> | even if it is a holiday period for a lot of people (depends on value of "a lot" I suppose), it certainly is not for a lot of people in the HTML WG |
| 09:58 | <annevk> | I'm going on vacation too this Friday for a couple of weeks but I don't really expect that nothing will happen in that time |
| 10:01 | <hsivonen> | MikeSmith: without looking at any logs, my guess is yesterday or the day before |
| 10:01 | <MikeSmith> | hsivonen: ah, OK |
| 10:01 | <MikeSmith> | I had thought it had happened a long time ago and I somehow missed |
| 10:02 | MikeSmith | checks his commit-watchers mail |
| 10:03 | <Philip`> | Giving the difficulty of getting anything decided at all in the HTML WG, I'm probably not a fan of the idea of delaying everything for a month-long period |
| 10:04 | <Hixie> | Lachy: it wasn't a straw poll |
| 10:04 | <Hixie> | Lachy: it was a request for arguments that hadn't been given yet (so why there were any left baffles me, given how long the issue had been open for) |
| 10:05 | <Hixie> | Lachy: and they still haven't announced the result, so if you have a convincing argument, you can still let them know :-) |
| 10:08 | <MikeSmith> | hsivonen: so that same commit also seems to remove all document-conformance requirements on the text content of meter and value |
| 10:08 | <hsivonen> | MikeSmith: which is great |
| 10:10 | <MikeSmith> | hsivonen: is it worth keeping the existing v.nu progress/meter text-content checking code, but emitting warnings instead of errors (with some tweaks to the contents of the existing message)? |
| 10:10 | <annevk> | somewhat annoying that some people just seem to ignore half the questions addressed at them and move on... |
| 10:11 | <hsivonen> | MikeSmith: let's just get rid of the code |
| 10:11 | <MikeSmith> | OK |
| 10:19 | <MikeSmith> | hsivonen: yank the <time> element text-content checking as well? |
| 10:19 | <hsivonen> | MikeSmith: did Hixie remove that, too, from the spec? |
| 10:20 | <MikeSmith> | hsivonen: there don't seem to any longer be any document-conformance requirements on <time> text content in the spec |
| 10:21 | <hsivonen> | MikeSmith: ok. let's get rid of that code as well |
| 10:21 | <MikeSmith> | ok |
| 10:22 | <MikeSmith> | hsivonen: btw, about coding style, I seem to remember a while back you said 2 spaces for indents, rather that 4 spaces |
| 10:22 | <MikeSmith> | or was it the other way around? |
| 10:25 | <Hixie> | MikeSmith: i'm getting a lot of waiting for www-data's lock in /sources/public/html5/spec tonight |
| 10:26 | <Hixie> | <time> still has text-content rules |
| 10:26 | <Hixie> | search for "If the element needs a date" |
| 10:29 | <hsivonen> | MikeSmith: 2 spaces in Jing, 4 spaces elsewhere in .java |
| 10:33 | <Hixie> | looks like a bunch of my checkins failed to get the right diff lines |
| 10:33 | <Hixie> | i fixed the problem, but we may have to go through some of the recent changes and add them manually |
| 10:33 | <Hixie> | (the logs list the bug # and the revisions, so it shouldn't be difficult) |
| 10:33 | hsivonen | tries to export Eclipse formatter settings |
| 10:34 | <hsivonen> | MikeSmith: are you using Eclipse? |
| 10:34 | <annevk> | Hixie, I'll look into the XHR hook later |
| 10:35 | <annevk> | Hixie, why does WebSocket not need to use it or EventSource? |
| 10:35 | <annevk> | Hixie, or Web Workers? |
| 10:35 | <Hixie> | i just hardcoded websocket into the html5 spec |
| 10:35 | <Hixie> | i should add eventsource too |
| 10:35 | <Hixie> | i think web workers just works (via garbage collection) |
| 10:38 | <hsivonen> | MikeSmith: http://hsivonen.iki.fi/eclipse-code-style.zip |
| 10:42 | <MikeSmith> | hsivonen: thanks |
| 10:43 | <MikeSmith> | hsivonen: yeah, I am using Eclipse |
| 10:44 | <MikeSmith> | headless Eclipse, with vim (eclim) as the editing frontend rather than the built-in one |
| 10:45 | <hsivonen> | MikeSmith: sounds like an interesting beast |
| 10:51 | <MikeSmith> | hsivonen: I just discovered it recently. It seems quite well done. Makes all the Eclipse completions work in vim, takes all errors and warnings from Eclipse and lets you walk through them in vim and fix them |
| 10:51 | <MikeSmith> | if you use vim, it's definitely worth taking a look at |
| 10:52 | <MikeSmith> | Hixie: thanks, I missed that part about authoring requirements on time with no datetime attribute |
| 10:53 | <MikeSmith> | Hixie: I have been getting those waiting for www-data's lock messages lately too |
| 10:54 | <MikeSmith> | I think something may have changed in my server environment |
| 10:57 | <Hixie> | annevk: how do you want me to resolve http://www.w3.org/Bugs/Public/show_bug.cgi?id=8476 ? |
| 10:58 | <annevk> | i was kind of hoping you'd figure that out... |
| 10:58 | <Hixie> | hah |
| 10:59 | <annevk> | well ideally all the same i guess |
| 10:59 | <annevk> | but if the webgl guys really find that unacceptable we need a way to do something like their spec suggests |
| 10:59 | <annevk> | which is that if you ask for another context the existing one gets detached or some such |
| 10:59 | <Hixie> | can i volunteer you to discuss this with the webgl guys? i don't know who the webgl guys are |
| 10:59 | <annevk> | would be better if the hooks for that were in the base spec |
| 10:59 | <Hixie> | do they have an open list yet? |
| 10:59 | <annevk> | yes, but i haven't subscribed |
| 11:00 | <annevk> | and i'm not at all interested in this |
| 11:00 | <Hixie> | join the club :-) |
| 11:00 | <annevk> | well, enough to find the issue :) |
| 11:00 | <Hixie> | you raised it, your problem :-P |
| 11:00 | <Hixie> | no? :-) |
| 11:00 | <Hixie> | oh i know how to fix it |
| 11:01 | <annevk> | cool |
| 11:01 | annevk | has to go somewhere |
| 11:02 | <Hixie> | wait this is already clear |
| 11:02 | <Hixie> | there's one bitmap according to the spec |
| 11:04 | hsivonen | doubts treating canvas as one bitmap for both the WebGL and 2D contexts is going to fly |
| 11:06 | <Philip`> | There's some currently-active discussion on the WebGL list about the context stuff |
| 11:07 | <annevk> | aah, let Philip` coordinate! |
| 11:08 | <Philip`> | Hixie: Implementors don't want to share a bitmap between 2D and WebGL contexts, because it forces performance penalties in current hardware/APIs |
| 11:08 | <Hixie> | Philip`: ok, please file a bug when you can tell me precisely what the spec should say :-) |
| 11:09 | <Philip`> | so the HTML spec itself is clear but it's incompatible with what the WebGL spec requires |
| 11:09 | <hsivonen> | how does WebGL handle page zooming? |
| 11:09 | Philip` | is kind of hoping the WebGL people will work out what to do |
| 11:11 | <Philip`> | hsivonen: As far as I'm aware (which isn't far since I've never tested it at all), they'd render into a width*height-attribute sized buffer and then stretch it to match CSS and zoom etc |
| 11:11 | <Philip`> | since I don't see what else they could reasonably do |
| 11:12 | <Lachy> | jgraham, is your system clock set to the wrong day? This e-mail of yours was sent yesterday, but dated today. http://lists.w3.org/Archives/Public/public-html/2010Jan/0123.html |
| 11:13 | <hsivonen> | Philip`: can OpenGL ES read back the buffer at the time of zooming, or will everything be kept in a buffer in case the user zooms later? |
| 11:14 | <hsivonen> | anyway, for both 2D and 3D, it seems super-sad to me that Canvas is concuptually a bitmap instead of being a one-way resolution-independent drawing pipeline |
| 11:14 | hsivonen | often runs with non-1:1 zoom |
| 11:14 | hsivonen | doesn't want jaggies |
| 11:15 | <Hixie> | i tried my best to make it the latter |
| 11:18 | <Philip`> | hsivonen: It's defined in terms of writing to a fixed-size drawing buffer, and glReadPixels reads back from that buffer, and the buffer is passed to "the HTML page compositor immediately before a compositing operation" (where it might be transformed and drawn on top of, etc) |
| 11:19 | Philip` | likes the ability to read pixels back from canvases, because otherwise it'd be impossible to do pretty much everything he's ever done with it |
| 11:20 | <hsivonen> | Philip`: does glReadPixels kill perf? |
| 11:21 | <Philip`> | hsivonen: Its existence doesn't hurt the performance of any other drawing operations |
| 11:21 | <Philip`> | but it's a bit slow if you call it |
| 11:21 | hsivonen | has no idea how GL context virtualization in Quartz Compositor and the like works |
| 11:22 | Philip` | neither |
| 11:22 | <hsivonen> | Philip`: doesn't a readback always stall the GPU pipeline for the duration of the readback? |
| 11:22 | <Philip`> | (though I don't even know what it is) |
| 11:22 | <hsivonen> | or is a readback async? |
| 11:23 | <Philip`> | It's synchronous at the OpenGL API level |
| 11:23 | <Philip`> | so the drivers/hardware would have to flush all current drawing calls and then read all the pixels before carrying on, I presume |
| 11:23 | <hsivonen> | Philip`: as I understand it, Quartz window compositing is done on top of GL, but you can pretend the GL context inside your window is the whole GL world |
| 11:23 | <hsivonen> | or something like that |
| 11:24 | <hsivonen> | without really knowing much about this, it seems that a one-way operation for transferring the pixels onto a texture within the GPU would be a useful operation |
| 11:25 | <hsivonen> | assuming you could queue subsequent ops with the texture into the pipeline right away |
| 11:26 | <Philip`> | You can select a texture as the framebuffer object, so that things are rendered directly onto that texture |
| 11:27 | <Philip`> | http://www.khronos.org/opengles/sdk/docs/man/glFramebufferTexture2D.xml |
| 11:27 | <hsivonen> | Philip`: does using that result in the same black-box-testable effects as reading back the pixels? |
| 11:28 | <hsivonen> | (assuming your CSS compositor allows a gl texture on a layer) |
| 11:30 | <Philip`> | You mean rendering onto the default framebuffer, then reading the pixels from the GPU back to the CPU, then doing the compositing on the CPU; vs rendering onto a new framebuffer that's bound to a texture, keeping it on the GPU, and doing the compositing on the GPU (assuming your GPU compositor is equivalent to the CPU-based one)? |
| 11:30 | <hsivonen> | Philip`: right |
| 11:31 | <Philip`> | Those should be equivalent as far as I'm aware, as long as your framebuffers are the same size and type in each case |
| 11:31 | hsivonen | hasn't used an GL features that weren't available way back when on an SGI O2 |
| 11:31 | <hsivonen> | Philip`: ok |
| 11:31 | <hsivonen> | s/an/any/ |
| 11:31 | hsivonen | also hasn't used all features that were available back then |
| 11:33 | <Philip`> | It seems the problem with 2D/WebGL contexts is that browsers don't have a GPU-based compositor today, and they do 2D canvas on the CPU, but they want to support WebGL today, and they don't want to require a GPU/CPU synchronisation performance penalty |
| 11:33 | <Philip`> | It shouldn't be a problem in the future but they need a solution now |
| 11:33 | hsivonen | thought Safari on Snow Leopard had a GPU-based compositor |
| 11:33 | <Philip`> | (and preferably one that doesn't prevent a better solution in the future) |
| 11:50 | <hsivonen> | I wonder if the non-sensical behavior of the Gecko HTML5 parser on http://fooo.fr/~vjeux/firefox/firefox_bug_fixing_a_tr.html is an implementation bug or a spec bug |
| 11:50 | <hsivonen> | WebKit seems to handle it in a reasonable way |
| 12:01 | <Hixie> | ms2ger is damn incisive |
| 12:01 | <Hixie> | who is this guy |
| 12:01 | <Hixie> | he's been really helpful |
| 12:01 | <Hixie> | like, _really_ helpful |
| 12:01 | <Hixie> | wrote like the whole index and references sections |
| 12:01 | <Hixie> | filed a bunch of cunning bugs |
| 12:02 | <Hixie> | closed a bunch of invalid ones |
| 12:11 | <Lachy> | Hixie, do you have a script that maintains those tables in the indexes automatically, or were they done manually? I could use those tables in the HTML5 Reference |
| 12:12 | <Lachy> | The elements one is especially useful. It's basically like the one I never got around to finishing my script for. |
| 12:12 | <Lachy> | actually, it's a little different from what I planned, but close enough |
| 12:16 | <Hixie> | manually |
| 12:16 | <Hixie> | ms2ger did most of the work |
| 12:17 | <Lachy> | ok. I must finish my scripts so they can be generated automatically for me the way I want them |
| 12:20 | <Hixie> | if you can get a script to generate what's in the spec today, that would be fantastic |
| 12:27 | <Lachy> | I need to check what the state of my scripts are, but I have some scripts that gather data from the spec and generate various tables of elements. |
| 12:29 | Philip` | 's usual conclusion from writing scripts to process the HTML5 spec is that the spec contains so many special cases that it's just as hard to write the script as to extract all the information by hand |
| 12:29 | <Lachy> | one of my scripts does basically read and store all of the information about elements in some data structures, so it shouldn't be too hard to modify it a bit to output the format needed. |
| 12:30 | <Philip`> | and you have the check the script output carefully by hand anyway, to make sure you didn't miss any special cases |
| 12:30 | <Lachy> | for the stuff I extract from the spec, it does seem fairly consistent. There are a few edge cases, but they seem to be detectable by the script itself and can be easily handled |
| 12:31 | <Lachy> | I also combine it with some data in a separate data file that is maintained by hand, like information about required/optional tags |
| 12:34 | <Hixie> | woo, i hit my target of 100 bugs open. |
| 12:34 | <Hixie> | bed time |
| 12:34 | <Hixie> | nn |
| 12:34 | <gsnedders> | Nighty |
| 12:56 | <peritus> | I just cant get my cache manifest working in firefox. It works excellent in safari, but Firefox never requests new versions of the cached files, even if the cache manifest is updated. The files specified in my cache manifest are all valid and exists. Are there any known issues with firefox? Something i have missed? How can this be debugged? |
| 15:50 | <MikeSmith> | hsivonen: OK to use Java5isms in v.nu code? |
| 15:51 | <MikeSmith> | e.g., for (String s : warnings) {...} |
| 15:51 | <MikeSmith> | thanks for review, btw |
| 15:52 | <MikeSmith> | (going through your review comments now) |
| 16:18 | <hsivonen> | MikeSmith: yes, Java5isms are OK and encouraged |
| 16:18 | <MikeSmith> | hsivonen: OK, thanks |
| 16:22 | <MikeSmith> | hsivonen: btw, I imported your formatters and, going forward, will run everything through that before committing |
| 16:23 | <MikeSmith> | also imported the cleanup stuff and templates |
| 16:36 | <MikeSmith> | hsivonen: about the year part of copyright line, should I change that to 2010, or should I add underneath a copy of the same line with the year as 2010, or change it to, e.g., Copyright (c) 2007-2010, or...? |
| 16:37 | MikeSmith | goes to look how hsivonen has been doing it in htmlparser source |
| 16:39 | <MikeSmith> | hsivonen: nm |
| 16:39 | <MikeSmith> | hsivonen: I see you use ranges in htmlparser, so will change it to that |
| 17:40 | <gsnedders> | UTF-16 and null-terminated strings don't go well together. |
| 17:45 | <Philip`> | Write everything in Chinese and you'll be fine |
| 17:45 | <gsnedders> | That's hard with XML. |
| 17:46 | Philip` | discovered recently that he can set his locale on Linux to use EBCDIC |
| 17:46 | <Philip`> | which makes all my filenames look much more interesting |
| 17:46 | <Philip`> | but I was disappointed that I couldn't set it to UTF-16 |
| 18:21 | <gsnedders> | Hixie: http://html5lib.googlecode.com/hg/ doesn't have </title> |
| 18:36 | <smq-monstros1ty> | can one use <section> instead of <div> now? |
| 18:36 | <Philip`> | Not in all cases, since they have different semantics |
| 18:37 | <smq-monstros1ty> | so section is more of a broad element? |
| 18:37 | <smq-monstros1ty> | div more specific? |
| 18:38 | <Philip`> | I think (but could be wrong) that <section> should be used for a block that has a heading and that should appear in an outline of the document |
| 18:39 | Philip` | looks at the spec |
| 18:39 | <Philip`> | "Note: The section element is not a generic container element. When an element is needed for styling purposes or as a convenience for scripting, authors are encouraged to use the div element instead. A general rule is that the section element is appropriate only if the element's contents would be listed explicitly in the document's outline." |
| 18:39 | <smq-monstros1ty> | oh ok so a page could have <header></header><section></section><footer></footer> |
| 18:40 | <smq-monstros1ty> | oh ok so put the divs inside so those would serve the purpose of a layout |
| 18:40 | <smq-monstros1ty> | and <aside> too i imagine |
| 18:40 | <Philip`> | It could, although in that case I think you could just do <header>header</header>content<footer>footer</footer> and there's no need for <section> (though there would be a need if the content consisted of multiple sections) |
| 18:41 | <smq-monstros1ty> | oh ok so i have one time or a few times had the need to have multiple pages |
| 18:41 | <smq-monstros1ty> | like but have it vertical |
| 18:41 | <smq-monstros1ty> | so a section could serve as a page |
| 18:41 | <Philip`> | If you're using <section>/etc then you should style those directly, rather than writing <section><div>... and styling the div |
| 18:41 | <Philip`> | but if you don't need the semantics of <section> then you should just use <div> by itself (or any more appropriate element) |
| 18:42 | <Philip`> | You might want <article> if the 'pages' are self-contained components |
| 18:43 | <smq-monstros1ty> | yeah i was reading that too yeah and then use pdf and <data> element will leave teh text unchanged too? |
| 18:43 | <smq-monstros1ty> | i like the <video> tag and the <audio> |
| 18:43 | <Philip`> | Uh, what <data> element? |
| 18:43 | <smq-monstros1ty> | i have wanted to have an easy way to put a song on a page |
| 18:43 | <Philip`> | Also I'm not sure how PDF is relevant :-) |
| 18:43 | <smq-monstros1ty> | i might have mistyped it heheh |
| 18:43 | <smq-monstros1ty> | well for a document |
| 18:44 | <smq-monstros1ty> | that has many many pages |
| 18:44 | <smq-monstros1ty> | probably berter to embed a pdf? |
| 18:44 | <smq-monstros1ty> | say a document that has like 50 pages |
| 18:44 | <smq-monstros1ty> | but you want it on the page viewable |
| 18:44 | <smq-monstros1ty> | internet explorer doesnt support it lol but <audio> and <video> i surely will make use of |
| 18:45 | <Philip`> | I'd usually prefer using HTML to PDF |
| 18:45 | <smq-monstros1ty> | yeah |
| 18:45 | <smq-monstros1ty> | me too i read a tag that stated that the html wont modify it |
| 18:45 | <Philip`> | (except in cases where the exact layout is critical, or where it's intended for printing, in which case PDF is probably better) |
| 18:48 | <smq-monstros1ty> | why did the align="center" get removed? |
| 18:48 | <smq-monstros1ty> | <table align="center"> is quite convenient |
| 18:49 | <Philip`> | Because CSS is recommended for presentational settings |
| 18:49 | <smq-monstros1ty> | ah ok |
| 18:49 | <Philip`> | <style>table.whatever { margin: auto }</style> would center-align it in its container, I think |
| 18:49 | <smq-monstros1ty> | nice |
| 18:50 | <smq-monstros1ty> | thanks |
| 18:50 | <smq-monstros1ty> | back a while ago |
| 18:50 | <smq-monstros1ty> | alt="" would make a title over the img as well |
| 18:50 | <smq-monstros1ty> | then it stopped working |
| 18:50 | <smq-monstros1ty> | title="" i found it the other day |
| 18:50 | <smq-monstros1ty> | sometimes its weird somethign small like that |
| 18:50 | <Philip`> | alt only made tooltips in IE, not any other browser |
| 18:50 | <smq-monstros1ty> | oh |
| 18:50 | <smq-monstros1ty> | for a while i was using ie only |
| 18:51 | <Philip`> | then IE changed to be consistent with the others, by using title for tooltips |
| 18:51 | <Philip`> | and leaving alt for its proper purpose (alternative text for non-graphical browsers) |
| 18:51 | <smq-monstros1ty> | i like firefox and i like opera |
| 18:51 | <smq-monstros1ty> | yeah i always put in an alt in imgs now |
| 18:52 | <smq-monstros1ty> | opera's widgets are cool the 2d ruler |
| 18:52 | <smq-monstros1ty> | and the screenruler |
| 18:52 | Philip` | wonders if anyone has a made a real plastic ruler for measuring pixels (calibrated to common monitor DPIs) |
| 18:53 | <smq-monstros1ty> | I need to read more on the em's |
| 18:54 | <smq-monstros1ty> | i understand someehwat on px to use but em is a bit vague for me |
| 18:54 | <smq-monstros1ty> | i understand it from emphasis concept |
| 18:54 | <smq-monstros1ty> | but distancewise i |
| 18:54 | <smq-monstros1ty> | its basically the browser that needs to be compatible to use display:inline-block; right? |
| 18:55 | <smq-monstros1ty> | often i would like to use an inline element as a block too |
| 18:56 | <Philip`> | inline-block should work in all modern browsers |
| 18:56 | <Philip`> | according to http://www.quirksmode.org/css/display.html |
| 18:57 | <smq-monstros1ty> | interesting |
| 18:57 | <smq-monstros1ty> | its probably my error somewhere that made it not work then heheh |
| 18:59 | <smq-monstros1ty> | nice thanks for the link |
| 18:59 | <smq-monstros1ty> | i find using outline: helpful when constructing a page |
| 19:01 | <smq-monstros1ty> | thanks for the information :) got to go do some unfun stuff hahaha |
| 19:44 | <cardona507> | http://wiki.whatwg.org/wiki/FAQ#What_are_the_various_versions_of_the_spec.3F - wow - thats rather convoluted |
| 20:14 | <jgraham> | Lachy: Yes I had been doing a HTTP test that had a step "set the clock orwrd by a day" and forgot to set it back |
| 21:33 | <annevk> | Hixie, did Tyler explain already how UM would work with <canvas> and <img>? |
| 21:34 | annevk | is confused now |
| 21:39 | <othermaciej> | annevk: is this even relevant to discuss? UM could be useful even if it does not cover all the use cases of CORS |
| 21:39 | <othermaciej> | annevk: so whether UM covers all useful use cases is only really a relevant question if someone argues to rescind CORS or fail to advance it |
| 21:42 | <annevk> | that's their position |
| 21:42 | <annevk> | so I think it's relevant to discuss |
| 21:48 | <othermaciej> | annevk: would the correctness or wrongness of their position affect our willingness to publish the spec as FPWD? |
| 21:48 | <othermaciej> | or even to advance it to Last Call? |
| 21:49 | <annevk> | if we just want UM features we can do that easily with CORS already |
| 21:49 | <annevk> | by just passing an anon origin if the right flags are passed to XHR2 |
| 21:49 | <annevk> | and setting the credentials flag to false |
| 21:50 | <annevk> | so if they're wrong we don't need UM as we can just put it on top XHR2/CORS with a paragraph of extra text or so |
| 21:51 | <othermaciej> | sure, CORS could define a UM mode by having suitable mode flags |
| 21:51 | <othermaciej> | (besides the credentials flag I guess it would need an origin flag) |
| 21:53 | <annevk> | that's already there |
| 21:53 | <annevk> | both are already there |
| 21:53 | <annevk> | all that's needed is a new constructor for xhr |
| 21:54 | <annevk> | which would override those flags |
| 21:54 | <othermaciej> | I see |
| 21:55 | <othermaciej> | so UM is basically the no-credentials no-origin profile of CORS |
| 21:55 | <othermaciej> | (except that the current version does not do preflight either |
| 21:55 | <othermaciej> | ) |
| 21:56 | <annevk> | yeah, preflight would automatically work with "CORS-UM" |
| 21:56 | <annevk> | I still don't really see how the UI and all will work for UM though |
| 21:56 | <annevk> | those threads are never really clarified |
| 21:57 | <annevk> | in other news, "BiDirectional or Server-Initiated HTTP" HTTP?! |
| 21:57 | <annevk> | at least the text is somewhat less misleading |
| 21:57 | <annevk> | although it does talk about HTTP clients... |
| 22:00 | <annevk> | also, I'm letting http://www.w3.org/Bugs/Public/show_bug.cgi?id=8476 be somebody else's problem |
| 22:00 | <annevk> | too much work |
| 22:00 | <annevk> | already |
| 22:01 | <othermaciej> | I expect the spec requirement doesn't work for WebGL |
| 22:03 | <annevk> | the WebGL spec conflicts with the spec at the moment |
| 22:21 | <othermaciej> | I have heard from our 3D folks that it's not really practical to do 3D and 2D drawing freely mixed into a shared bitmap |
| 22:21 | <othermaciej> | although I am not sure what happens when you try to switch |
| 22:24 | <Hixie> | i'll be happy to change teh spec to whatever the webgl guys think it should say |
| 22:25 | <Hixie> | whether that is that getContext('3d') should blow away the bitmap, or getContext() for anything should do so, or getContext() should switch to a different bitmap, or whatever |
| 22:26 | <annevk> | i guess you should just have a concept of a context that reserves the bitmap |
| 22:26 | <annevk> | other specs define whether they are such a context |
| 22:26 | <annevk> | lets call the context x |
| 22:27 | <annevk> | you have to worry about getContext(x) on something with a non-x context and about invoking non-x or different x on an already active x |
| 22:28 | <annevk> | in the latter case, should be able to attach the removed context again later somehow? |
| 22:28 | <annevk> | the object is still around... |
| 22:28 | <annevk> | actually, in both cases you have that |
| 22:28 | <annevk> | doh |
| 22:28 | <othermaciej> | should it be only on getContext(x), or if you save a context and use its methods does that also affect the bitmap? |
| 22:29 | <othermaciej> | like if you save copies of 2d and webgl contexts, and alternately call methods from one or the other |
| 22:29 | <annevk> | it only affects if the context is "active" i believe |
| 22:29 | <annevk> | no idea what happens with the inactive one though |
| 22:30 | <annevk> | maybe they should throw INVALID_STATE_ERR |
| 22:30 | <annevk> | on anything |
| 22:31 | <Hixie> | so i'm all in favour of people agreeing with me, but in this instance i'm a little concerned that i might be part of some spammer's script: http://thestatbot.com/2009/01/17/the-entire-louis-gray-greader-leaderboard/#comment-28682654 |
| 22:31 | <Hixie> | the comment that mentions my name comes after a bunch of clearly spammy comments, and is a complete non-sequitur |
| 22:32 | <Hixie> | looks like in fact it is copied straight from http://playtheweb.org/?p=124 |
| 22:32 | <Hixie> | oh and the comment has a spammy name |
| 22:32 | <Hixie> | that's pretty funny |
| 22:33 | <annevk> | wasted 30s of my time! |
| 22:33 | <Philip`> | https://www.khronos.org/webgl/public-mailing-list/archives/1001/msg00031.html |
| 22:33 | <Philip`> | (Current thread about getContext) |
| 22:33 | <annevk> | I remember that we can actually do 2d and 3d on the same context at some point |
| 22:34 | <annevk> | same bitmap |
| 22:34 | <annevk> | doh |
| 22:34 | <annevk> | I should ask Tim again |
| 22:34 | <annevk> | would definitely be vastly superior |
| 22:35 | <othermaciej> | I don't know how to evaluate the claim that it's hard or can't be done |
| 22:35 | <othermaciej> | it seems like you could always render down to a bitmap when you switch |
| 22:35 | <Hixie> | seems to me like it could just be implemented as two bitmaps over each other |
| 22:35 | <Hixie> | if we wanted to just allow overlays |
| 22:36 | <Hixie> | or yeah, forcing a flattening when switching |
| 22:36 | <Philip`> | It seems hard to come up with useful semantics since the contexts have different buffer formats, e.g. 2D has RGBA but 3D has depth too |
| 22:36 | <Hixie> | i'm not especially worried about people wanting to use two contexts on one canvas though |
| 22:37 | <annevk> | 3d for some graphics, 2d for some text oh well |
| 22:37 | <othermaciej> | I'm not sure overlay is a logically expected semantic, given the non-retained-mode behavior of canvas |
| 22:38 | <Hixie> | annevk: [ms] is ms2ger |
| 22:38 | <annevk> | aah |
| 22:54 | <GPHemsley> | Does anyone know off-hand if IE6 supports @id anchors? |
| 22:55 | <annevk> | it does |
| 22:55 | <GPHemsley> | I thought so |
| 22:55 | <annevk> | IE5 does too |
| 22:55 | <Hixie> | is there an ETA on the microdata decision yet? |
| 22:56 | <annevk> | NN4 didn't |
| 22:56 | <annevk> | but I never cared about that |
| 22:56 | <GPHemsley> | :) |
| 22:56 | <annevk> | Hixie, it seems they reached a decision but are working on wording |
| 22:57 | <annevk> | "The Chairs met last week and today to discuss the results of the Microdata poll. We are formulating our response and will provide it to the WG ASAP." |
| 22:57 | <Hixie> | maciej: any ETA on the microdata decision's wording being published? |
| 22:59 | <Hixie> | er, othermaciej_ ^ |
| 23:00 | <othermaciej_> | Hixie: no firm ETA but I am hoping for sometime this week |
| 23:01 | <Hixie> | i'm getting close to having no bugs left for which i'm comfortable giving an editor's decision without hte microdata thing resolved |
| 23:01 | <Hixie> | so i might defer the remaining bugs to next week |
| 23:36 | <othermaciej> | Hixie: it looks like there's still 100 open and most do not seem removal-related |
| 23:38 | <daedb> | If I do a CSS transition on background-color from #dfdfdf to #eee, why does the background become darker before it goes lighter? |
| 23:41 | <annevk> | browser bug? |
| 23:41 | <annevk> | did you try another browser? |
| 23:42 | <daedb> | It's the same in both Chrome and Opera 10.5... |
| 23:42 | <annevk> | not sure about Firefox, but Opera 10.5 should support it |
| 23:42 | <annevk> | oh |
| 23:42 | <annevk> | you're way ahead of me then :) |
| 23:43 | <AryehGregor> | daedb, have a minimal test case? |
| 23:44 | <daedb> | AryehGregor: I can make one. |
| 23:47 | <daedb> | hmmm, it doesn't do it when I make a minimal test... odd. I must've screwed up somewhere :-/ |
| 23:49 | <daedb> | oh, it's because the background-color was inherited from the parent... |
| 23:49 | <gsnedders> | payman: Wow, it's tiny! |
| 23:51 | <annevk> | async IRC reactions FTW |
| 23:51 | <annevk> | or was it just wc? |
| 23:52 | <gsnedders> | annevk: Related to stuff he'd asked in private. |
| 23:52 | <AryehGregor> | annevk, what does "What did you do? You tried to pass the comment system with an invalid comment! (Or you used an entity that cannot be used. Sorry about that, PHP is a mess.)" mean? I have a comment I want to post but it won't let me. |
| 23:52 | <annevk> | ah, it might mean you used & or so |
| 23:52 | <Philip`> | AryehGregor: It means your comment was invalid |
| 23:53 | <AryehGregor> | annevk, should I use numeric references instead, or what? |
| 23:53 | <annevk> | there's a bug somewhere in the system that I haven't caught, so I just worsened usability |
| 23:53 | <annevk> | just say AMP or something and I'll fix it |
| 23:53 | <gsnedders> | annevk: We should get PHP html5lib to a state in which you can use it |
| 23:53 | <AryehGregor> | What? |
| 23:53 | <AryehGregor> | gsnedders, <3 |
| 23:53 | <annevk> | replace & with AMP |
| 23:53 | <daedb> | AryehGregor: http://software.hixie.ch/utilities/js/live-dom-viewer/saved/344 |
| 23:54 | <gsnedders> | annevk: Then you can just check for parse errors |
| 23:54 | <annevk> | yeah |
| 23:54 | <annevk> | or ignore them :) |
| 23:54 | <annevk> | though that might require some spam control |
| 23:54 | <AryehGregor> | annevk, I'm not using &, I'm using < and >. Same principle? |
| 23:54 | <Philip`> | Why not use the Python html5lib, and call it with IPC from PHP? |
| 23:55 | <Philip`> | Seems easier than making PHP html5lib work |
| 23:55 | <annevk> | AryehGregor, yes |
| 23:55 | <annevk> | AryehGregor, though > is never needed |
| 23:55 | <annevk> | AryehGregor, just use the raw character there |
| 23:55 | <AryehGregor> | It is in attributes in XML, right? |
| 23:55 | <annevk> | it's only the predefined entities |
| 23:55 | <annevk> | AryehGregor, oh yeah, though I've no idea why |
| 23:55 | <gsnedders> | AryehGregor: no |
| 23:55 | <annevk> | oh lol |
| 23:55 | <annevk> | XML5 allows it all :) |
| 23:56 | <AryehGregor> | gsnedders, no? |
| 23:56 | <gsnedders> | '"' ([^<&"] | Reference)* '"' |
| 23:56 | <gsnedders> | | "'" ([^<&'] | Reference)* "'" |
| 23:56 | <gsnedders> | That's AttValue |
| 23:56 | <AryehGregor> | annevk, doesn't seem to have worked too well. http://annevankesteren.nl/2010/01/optimizing-html#comment-6872 |
| 23:56 | <annevk> | I said I'd fix it |
| 23:56 | <annevk> | that means manually |
| 23:56 | <AryehGregor> | annevk, ah, I see. |
| 23:57 | <AryehGregor> | gsnedders, all I know is when I used > in attribute values, XML parsers started dying on me. |
| 23:58 | <annevk> | you're not using attributes... |
| 23:58 | <AryehGregor> | The spec does seem to agree with you, though. |
| 23:58 | <AryehGregor> | annevk, no, it was a hypothetical point in the case of my blog post. |
| 23:58 | <AryehGregor> | I stuck in > to see if that would make your confusing error go away. |
| 23:59 | <annevk> | yeah it sucks |
| 23:59 | <annevk> | I should really invest some time in da blog |
| 23:59 | AryehGregor | would probably just nl2br(htmlspecialchars()) comments if he wrote his own hacky blog software |
| 23:59 | <annevk> | but it'll have to wait until February at the earliest |
| 23:59 | gsnedders | wouldn't use PHP ;P |
