| 00:06 | <Hixie> | http://www.w3.org/2002/09/wbs/44061/20100225_canvas/ |
| 00:08 | <Philip`> | You are not allowed to see this questionnaire. |
| 00:08 | <Philip`> | I guess that's because it's from the future |
| 00:08 | <Hixie> | actually it's because for some reason it's limited to a11y tf members |
| 00:08 | <Hixie> | no idea why |
| 00:10 | <Philip`> | If it's a reopening of the is-canvas-in-scope decision, I wonder if we could reopen the shall-we-adopt-HTML5 decision too since surely some people still object to that |
| 00:11 | <Philip`> | Oh, it's not |
| 00:12 | <Philip`> | It's http://lists.w3.org/Archives/Public/public-html-a11y/2010Feb/0451.html |
| 00:12 | <Hixie> | right |
| 00:12 | <Hixie> | i replied asking if we could make it public |
| 00:29 | <Hixie> | i cannot work out a way to make Web Socket safe from cross-protocol attacks from HTML pages without sacrificing the HTTP compatibility |
| 00:32 | <Hixie> | unless we can use a new method... maybe that would be acceptable |
| 00:37 | <rauchg> | let's just move to SPDY |
| 00:37 | <Philip`> | http://x264dev.multimedia.cx/?p=292 (about video) seems interesting |
| 00:37 | <Philip`> | (particularly since it's by someone who actually knows about video codecs) |
| 00:40 | <miketaylr> | needs moar contrast |
| 00:46 | <tantek> | Hixie FWIW I agree with making all surveys regarding HTML5, even in TFs, public. |
| 00:52 | <hober> | tantek: indeed. the WG is chartered to do its work in public; the same should apply to its TFs. |
| 01:04 | <bfrantz> | a few days ago i had mentioned that there is disparity in how FF3.5 (+others) and IE8(+others) treat the entity body when encountering a 404. after some testing it turns out they all (opera, chrome, safari, ie, firefox) respect the entity body provided it is over 512bytes. again, i'm unsure if html5 is the place to document or standardize on the "correct" behavior but thought I would provide the update for those interested. |
| 01:27 | <roc> | Philip`: the x264 developers are pretty rabidly in favour of H.264 |
| 01:27 | <roc> | unsurprisingly |
| 01:27 | <roc> | and there are some errors in that article |
| 01:27 | <roc> | > even though H.264 is licensed under RAND terms (Reasonable and Non-Discriminatory, they legally cannot be anti-competitive) |
| 01:27 | <roc> | I don't think that's true at all |
| 03:58 | <rektide> | i'm trying to use SharedWorkers, and have a base case written up that I cant seem to get to do anything. its at http://rektide.voodoowarez.com/2010/2/ping.html |
| 04:00 | <rektide> | it should just post a message to the sahred worker, which should reply back. i cant see any evidence of communication actually happening. not sure whether its my code thats wrong, or whether the browsers silently dont implement any SharedWorker functionality. |
| 04:00 | <MikeSmith> | bfrantz: one thing you could do is to raise a bug report in the HTML5/HTMLWG bugzilla along the lines of, "HTML5 spec should mandate consistent UA behavior for handling of entity bodies for 404 responses" |
| 04:00 | <MikeSmith> | bfrantz: others can/will respond there |
| 04:01 | <MikeSmith> | that way there's some record of the conversation a resolution |
| 04:02 | <MikeSmith> | I would guess that one likely outcome is that the editor will say it's out of scope fo the HTML5 spec and should be covered by the HTTP spec |
| 04:03 | <MikeSmith> | bfrantz: http://www.w3.org/Bugs/Public/enter_bug.cgi?product=HTML%20WG |
| 04:03 | <bfrantz> | sweet, I was just googling that |
| 04:08 | <MikeSmith> | Julian Reschke from the HTTP WG follows HTML5 bug mail pretty closely, so if you raise a bug report there, he will probably see it |
| 04:08 | <MikeSmith> | anyway, I can always give him a heads-up about it too |
| 04:10 | <MikeSmith> | rektide: I don't know what browsers you are trying it in but I know that Shared Workers are supported in Chrome at least |
| 04:11 | <MikeSmith> | there's a page that shows up-to-date status on Chrome standards support |
| 04:11 | MikeSmith | goes to look for it |
| 04:11 | MikeSmith | wonders if cying may have done any app development yet using shared Workers |
| 04:12 | <MikeSmith> | rektide: page is here: http://www.chromium.org/developers/web-platform-status |
| 04:12 | <MikeSmith> | but it looks like it's less up-to-date that I thought |
| 04:12 | <rektide> | MikeSmith: i've been using Chrome, Opera 10.50 and desktop team releases, and firefox trunk. thusfar, my absolutely primitive example code doesnt seem to do anything whatsoever on any browser. |
| 04:12 | <MikeSmith> | seems to have not been updated since December |
| 04:12 | <cying> | MikeSmith: hm, not sure |
| 04:12 | <cying> | MikeSmith: i know the general idea |
| 04:13 | <MikeSmith> | rektide: I don't remember hearing anything about any support at all for workers in Opera yet |
| 04:13 | <MikeSmith> | but maybe I missed something |
| 04:13 | <rektide> | trying other browsers was an act of desperation when i couldnt get Chrome to do anything |
| 04:14 | <MikeSmith> | I know mozilla has workers support, but I don't know if they support shared Workers |
| 04:14 | <MikeSmith> | cying: rektide has a script that uses shared Workers but it seems not to behave as expected in Chrome |
| 04:15 | <rektide> | its http://rektide.voodoowarez.com/2010/2/ping.html |
| 04:15 | <rektide> | it really should be an absolute minimum use case... send a ping to a SharedWorker, get a pong back. |
| 04:15 | <MikeSmith> | cying: so had just been wondering if you might have written anything using shared workers and might have any troubleshooting insight |
| 04:15 | <rektide> | but it doesnt appear to actually be doing anything whatsoever. :/ i cant get any evidence whatsoever the SharedWorker exists / was created |
| 04:16 | <cying> | ah nope |
| 04:16 | <MikeSmith> | rektide: you checking with the error console and/or other script debugging features in Web Inspector? |
| 04:16 | <MikeSmith> | cying: OK, just thought it was worth asking you at least :) |
| 04:16 | <cying> | heh |
| 04:18 | <rektide> | MikeSmith: i have. the debugger seems to be isolated from the SharedWorker scope |
| 04:18 | <MikeSmith> | ah |
| 04:19 | <MikeSmith> | sounds like another area where web inspector could use some improvements |
| 04:20 | <JoePeck> | MikeSmith: mention it in #webkit-inspector I'm sure they would be interested |
| 04:20 | <JoePeck> | although, I think this may have been raised before |
| 04:20 | <MikeSmith> | hey JoePeck |
| 04:20 | <MikeSmith> | I didn't know about #web-inspector |
| 04:21 | <rektide> | just threw them a quick message. figure its worth mentioning. |
| 04:21 | <JoePeck> | #webkit-inspector |
| 04:21 | <MikeSmith> | rektide: maybe also worth trying on #chromium-support |
| 04:22 | <rektide> | i'll probably post to chromium-dev group |
| 04:23 | <MikeSmith> | as far as this channel, I guess it's kind of off-hours for Europe and North America right now |
| 04:23 | <MikeSmith> | it's usually very quiet here this time of day |
| 04:24 | <rektide> | ugg there's a chromium html5 google group. perfect forum! it just looks crazy inactive. |
| 05:01 | <JonathanNeal> | So, in Firefox video there are some keyboard controls, like SPACE, LEFT, and RIGHT, are there any others? |
| 05:12 | <doublec> | JonathanNeal, home and end |
| 05:14 | <doublec> | JonathanNeal, meta-down, meta-up, meta-left, meta-right |
| 05:14 | <doublec> | JonathanNeal, from what I can see perusing the source |
| 05:14 | <JonathanNeal> | oh, what do those do and how do I trigger "meta" down. |
| 05:15 | <doublec> | JonathanNeal, I don't know what the 'meta' key is but meta up/down is mute/unmute. meta left/right is seek 10% increments rather than 15 seconds without the meta. |
| 05:17 | <doublec> | JonathanNeal, apparently the windows key or apple/command |
| 05:17 | <JonathanNeal> | Oh okay. |
| 05:18 | <JonathanNeal> | I'm working on the Flash equiv for a Video For Everybody fallback. |
| 05:18 | <doublec> | nice |
| 05:19 | <doublec> | I'm not sure how those key actions were decided upon - other browsers will be different i'm sure |
| 06:18 | <bfrantz> | anyone know where to find data on the compat hit a user will experience if they config their browser to speak only TLS 1.2 vice TLS 1.1 or 1.0 ? |
| 06:51 | <JonathanNeal_oww> | http://sandbox.thewikies.com/html5-video/ |
| 06:52 | <JonathanNeal> | I know there's no real documentation yet, but I'm looking for folks who can help me write up some documentation. |
| 07:08 | <MikeSmith> | doublec: do I recall correctly that you blogged a few days back about an CVS attribute-naming conflict? |
| 07:10 | MikeSmith | doesn't find mention of it at Bluish Coder.. |
| 07:10 | <MikeSmith> | roc, maybe? |
| 07:10 | <roc> | certainly not me |
| 07:15 | <MikeSmith> | roc, er, I meant SVG attribute-naming conflict |
| 07:15 | <MikeSmith> | freudian slip |
| 07:15 | <roc> | that sounds more like me, but it wasn't :-) |
| 07:16 | <MikeSmith> | ah, found it |
| 07:16 | <MikeSmith> | 'twas Hallvord |
| 07:16 | <MikeSmith> | http://my.opera.com/hallvors/blog/2010/02/16/svg-adds-style-filter-now-where-have-we-seen-that-before |
| 07:17 | <MikeSmith> | shepazu: ↑ |
| 07:17 | <MikeSmith> | " It has been raised as an issue with the SVG working group, who discussed it and decided to sweep the issue under the carpet with an informative warning in the spec." |
| 07:17 | <MikeSmith> | harsh words |
| 07:24 | MikeSmith | comes across "The conformance checker has been configured to assume that the document is an e-mail or document intended for a specific person who is known to be able to view images." in r4795 diff.. |
| 07:26 | MikeSmith | imagines "Valid HTML for e-mail" and "Valid HTML for documents intended for a specific person who is known to be able to view images" badges |
| 07:28 | <othermaciej> | MikeSmith: nothing requires conformance checkers to support such a configuration |
| 07:30 | <MikeSmith> | othermaciej: true that nothing requires it, but the inclusion of it in the spec could be seen as implying it would be a good idea for conformance checkers to provide such options |
| 07:31 | <zcorpan> | MikeSmith: maybe V.nu would be differently configured if used for Mail.app |
| 07:31 | <MikeSmith> | zcorpan: yeah, something like that would make sense, true |
| 07:31 | <zcorpan> | MikeSmith: while the web service is configured to assume that all documents being validated are public web pages intended for anyone |
| 07:31 | <MikeSmith> | yeah |
| 07:55 | <roc> | MikeSmith: we knew about the problem Hallvord talks about |
| 07:56 | <roc> | none of the proposed solutions are very good |
| 07:56 | <roc> | so we decided to tackle the problem head-on by exposing style.filter in Firefox and seeing how bad the compatibility problem is |
| 07:57 | <roc> | if the problem isn't bad, or we can persuade sites to fix their detection logic, then the Web wins. If the problem is really bad and unfixable, then we'll back off and nothing is really lost. Hallvord should be thanking us :-) |
| 07:57 | <MikeSmith> | interesting |
| 07:58 | <MikeSmith> | have you gotten much feedback or data yet on sites breaking? |
| 07:58 | <roc> | not yet |
| 07:58 | <MikeSmith> | ok |
| 07:58 | <MikeSmith> | this is a fairly recent change? |
| 07:58 | <roc> | yes |
| 07:58 | <MikeSmith> | I see |
| 07:58 | <roc> | checked in 12 days ago |
| 07:59 | <MikeSmith> | whoah, very recent indeed |
| 07:59 | <roc> | probably only 10-20K users have that change so far |
| 07:59 | <MikeSmith> | I guess I'm one of those, since I'm running a recent Minefield |
| 07:59 | <othermaciej> | I look forward to hearing the results of the experiment |
| 08:01 | zcorpan | too |
| 08:03 | <roc> | if you guys know of any sites that are likely to be broken, you could let us know in bug 374216 :-) |
| 08:03 | <roc> | The webkit changeset that introduced undetectable 'filter' ( http://trac.webkit.org/changeset/15557 ) only references one site, housingmaps.com, which appears to work now |
| 08:06 | <othermaciej> | roc: I have a strong recollection of there being more but I could not locate them with bug archaeology |
| 08:07 | <othermaciej> | roc: housingmaps we were pretty sure would be fixed soon enough, because even then the copy of Google Maps code they were using was out of date and the latest version had the fix |
| 08:07 | <othermaciej> | but I recall some other sites that had a three-way test for filter, -moz-opacity and opacity, in that order of preference |
| 08:07 | <hsivonen> | JW in the minutes is jwatt, right? |
| 08:10 | <roc> | yes |
| 08:12 | <hsivonen> | do VP6 videos in Flash usually use MP3 for audio? |
| 08:15 | <asmodai> | Jeez, sometimes I think I'm the only techie that abhors conferences. |
| 08:16 | <othermaciej> | there are some people who enjoy going to conferences |
| 08:16 | <Hixie> | asmodai: i hear ya |
| 08:16 | Hixie | hates them too |
| 08:16 | <Hixie> | though not as much as teleconferences |
| 08:16 | <asmodai> | My colleagues here were all jumpy about going to a PHPcon |
| 08:16 | <othermaciej> | there might even be some people in that category who produce significant technical achievements |
| 08:16 | <asmodai> | and I am like: meh |
| 08:16 | <roc> | it entirely depends on the conference |
| 08:16 | <asmodai> | would rather go to Pycon at the moment |
| 08:16 | <othermaciej> | the phone is my least-favorite communications medium |
| 08:16 | <asmodai> | And I intensely dislike conferences. |
| 08:17 | <asmodai> | Can't even stand to talk shop outside of my work hours. |
| 08:17 | <othermaciej> | especially in group format |
| 08:17 | <Hixie> | hear hear |
| 08:17 | <othermaciej> | actually I might hate video chat even more, not sure |
| 08:17 | <asmodai> | If there's a group with techies and another with women, I'm over there talking about life with the women |
| 08:17 | <roc> | I'm encouraged that out of the top 100 Google hits for "if style.filter", only one actually is testing style.filter in JS (although several are testing style.filters, whatever that is), and that one tests for 'opacity' first |
| 08:17 | <asmodai> | Hixie: Glad to hear I ain't alone in that :) |
| 08:18 | <hsivonen> | conferences are nice to attend, but traveling there sucks (esp. over multiple time zones) |
| 08:20 | <asmodai> | hsivonen: to each his/her own |
| 08:21 | <othermaciej> | what if there's a group of women techies talking about lifehacking? |
| 08:21 | <asmodai> | othermaciej: Mmm |
| 08:23 | <zcorpan> | othermaciej: that would be a pretty noteworthy event |
| 08:24 | <hsivonen> | othermaciej: I hear telepresence systems are awesome compared to traditional video conferencing |
| 08:25 | <hsivonen> | I haven't tried telepresence systems myself, though |
| 08:25 | <othermaciej> | I don't know what exactly constitutes a "telepresence system" |
| 08:26 | <othermaciej> | my main video chat experience is with iChat |
| 08:26 | <othermaciej> | even when it's pretty good, it's pretty lame |
| 08:26 | <othermaciej> | I could imagine doing better with a lot more bandwidth |
| 08:26 | <othermaciej> | but it kinda creeps me out |
| 08:26 | <hsivonen> | othermaciej: a carefully tuned room where you sit across the table from a quality display that shows the remote people in a similar room elsewhere |
| 08:26 | <othermaciej> | ChatRoulette is basically my worst nightmare |
| 08:27 | <othermaciej> | if anyone ever needs to torture a state secret out of me, make me use ChatRoulette until I break |
| 08:27 | <hsivonen> | othermaciej: where the whole room including the color temperature of the lighting has been built according to Cisco's specs |
| 08:27 | <othermaciej> | sounds fancy |
| 08:27 | <roc> | mmm Cisco |
| 08:27 | <roc> | I've got a story about Cisco |
| 08:27 | <roc> | Worst. Firefox. Bug. Ever. |
| 08:28 | <roc> | "crash in Cairo spline subdivision code while Cisco VPN enabled" |
| 08:31 | <annevk> | the SVG WG should have named it cssFilter |
| 08:31 | <othermaciej> | that sounds pretty awful |
| 08:31 | <Philip`> | roc: Maybe http://philip.html5.org/data/if-style-filter.txt might be relevant? |
| 08:32 | <othermaciej> | annevk: the SVG WG shouldn't have even named the CSS property itself "filter" |
| 08:32 | <othermaciej> | it was lame |
| 08:32 | <othermaciej> | they were warned |
| 08:32 | <annevk> | true |
| 08:32 | <othermaciej> | they were not thinking about tight integration with browsers at the time |
| 08:32 | <roc> | Philip`: ta |
| 08:32 | <Hixie> | roc: wow, what on earth was the story behind _that_?! |
| 08:32 | <annevk> | i guess it is somewhat the fault of the browsers for not participating from the start and then trying to implement it anyway |
| 08:33 | <annevk> | starting with Opera... |
| 08:34 | <othermaciej> | Hixie: you're not jaded enough if you are truly surprised by that |
| 08:34 | <roc> | turns out that Cisco's VPN kernel driver, under rare, random circumstances when processing network traffic, will corrupt the floating point registers of user processes |
| 08:34 | <Hixie> | nice |
| 08:34 | <othermaciej> | I must admit I've never had a browser bug turn out to actually be a kernel bug |
| 08:34 | <othermaciej> | (other than perf issues) |
| 08:35 | <roc> | most of the time, people probably don't notice, since wrong floating point values usually don't cause programs to crash |
| 08:35 | <othermaciej> | however, we *have* had the production compiler from the OS build train miscompile our code |
| 08:35 | <roc> | spline subdivision is one case where they do (the tolerance value got replaced by NaN, and the recursive subdivision never terminated) |
| 08:36 | <roc> | I wonder how many Wall St firms use Cisco VPN on machines running financial simulations, though |
| 08:37 | <annevk> | that could be fun |
| 08:38 | <othermaciej> | yikes |
| 08:38 | <othermaciej> | (though NaN would probably lead to an obvious error) |
| 08:39 | <hsivonen> | has Cisco fixed their code yet? |
| 08:39 | <MikeSmith> | can somebody remind me what does the HTML5 parsing algorithm do for a text/html document that has an svg element as its root element? |
| 08:39 | <roc> | hsivonen: yes, in the latest version |
| 08:39 | <Hixie> | MikeSmith: can't happen |
| 08:39 | <hsivonen> | MikeSmith: it generates <html><head></head><body>before the <svg> |
| 08:39 | <Hixie> | what hsivonen said |
| 08:40 | <hsivonen> | MikeSmith: so as Hixie says, you can't have an svg root |
| 08:40 | <MikeSmith> | OK |
| 08:40 | <MikeSmith> | I meant the serialized document |
| 08:40 | <hsivonen> | It only looks like a root :-) |
| 08:40 | <hsivonen> | there's no spoon there |
| 08:41 | <annevk> | MikeSmith, you mean what happens when you serialize an HTML document that has <svg> as root? |
| 08:41 | <annevk> | nothing special I think |
| 08:41 | <MikeSmith> | I mean what happens if you serve an svg document as text/html |
| 08:46 | <annevk> | what hsivonen said |
| 08:46 | <MikeSmith> | OK |
| 08:48 | <annevk> | so I wonder if IE9 is a) shipping around MIX and b) has SVG in HTML support as seems to be hinted on the IE blog |
| 08:49 | <Philip`> | I don't think they'd be shipping in a month when they haven't even started a beta yet |
| 08:50 | <othermaciej> | If they do have SVG-in-HTML, I wonder if it is per the HTML5 draft, or using IE's namespaces in HTML |
| 08:51 | <Philip`> | (As of a few days ago, they have been "closing down the Internet Explorer 8 Feedback Program in preparation for soliciting feedback for our next version", so maybe the IE9 beta will be around MIX) |
| 08:54 | <asmodai> | I wonder how standards compliant in general it will be anyway. |
| 08:54 | <asmodai> | IE remains a strange beast |
| 08:58 | <zcorpan> | "All you'll have to do is feed |
| 08:58 | <zcorpan> | it into the appropriate compiler (also to be defined in the HTML5 spec) |
| 08:58 | <zcorpan> | and out will come a working browser." |
| 08:58 | <zcorpan> | i guess writing such a compiler has about the same complexity as writing a working browser |
| 08:59 | <roc> | MIX won't have an IE9 beta |
| 08:59 | <roc> | they'll announce a bunch of IE9 features though |
| 09:00 | <roc> | (IE builds won't even be generally available within Microsoft) |
| 09:01 | <annevk> | oh, this is public info? |
| 09:01 | <roc> | not really |
| 09:05 | <asmodai> | Now it is since this channel is logged. |
| 09:05 | <asmodai> | ;) |
| 09:05 | <roc> | that's not exactly sensitive information |
| 09:06 | asmodai | grins |
| 09:09 | <zcorpan> | roc might be making things up |
| 09:09 | <roc> | indeed |
| 09:10 | <Dashiva> | Besides, he works for a competitor :P |
| 09:10 | <roc> | I could be a double agent |
| 09:10 | <roc> | or a triple agent |
| 09:11 | <roc> | or just an agent |
| 09:11 | <zcorpan> | or a user agent |
| 09:11 | <Dashiva> | Is he conforming? |
| 09:11 | <zcorpan> | he's a conforming html4 UA |
| 09:12 | <zcorpan> | assuming he renders quotes around <q>s |
| 09:12 | <Dashiva> | With his fingers, I assume |
| 09:13 | <zcorpan> | yes |
| 09:13 | <Hixie> | does he assume a default encoding? |
| 09:13 | <zcorpan> | no |
| 09:14 | <Hixie> | sounds compliant to me |
| 09:15 | <hsivonen> | the about: URL registration sure takes a long time |
| 09:15 | <hsivonen> | yay for taking stuff off Hixie's plate |
| 09:15 | <Hixie> | was about: registration on my plate? |
| 09:15 | <hsivonen> | Hixie: I thought it might have been |
| 09:16 | <Hixie> | i've registered two schemes in the past year |
| 09:16 | <Hixie> | ws: and wss: |
| 09:16 | <Hixie> | it was relatively easy |
| 09:16 | <MikeSmith> | asmodai: do you know of another channel around here where I can talk about life with the women? |
| 09:17 | <annevk> | there was the whole TAG thing that you should use http://websockets.org/{uri-here} instead or some such... |
| 09:18 | <Dashiva> | >8 pages have longdesc="#". |
| 09:18 | <hsivonen> | annevk: whoa, whoa. what problem does *that* solve? |
| 09:19 | Dashiva | wonders if any AT would be able to create an infinite loop from that |
| 09:19 | <annevk> | hsivonen, beats me |
| 09:19 | <asmodai> | MikeSmith: Mmm -- good one, most of the women I talk to are on my IM. XD |
| 09:19 | <Philip`> | Is that the same argument they used for XRIs? |
| 09:20 | <Dashiva> | hsivonen: It seemed to me that TAG just wants everything to be http |
| 09:20 | <annevk> | Philip`, could be, yes |
| 09:21 | <danbri> | if not the TAG, then a lot of folk on the TAG list (and URI list) |
| 09:21 | <danbri> | there was some discussion lately about having platforms dispatch on substrings of full URI, rather than just the scheme name... wonder if that'll go anywhere |
| 09:22 | <Dashiva> | And then at some later point, they'll have platforms that dispatch on substrings of substrings |
| 09:23 | <hsivonen> | does the TAG have a stated opinion on how things on FTP servers should be addressed from an href? |
| 09:23 | <asmodai> | Is that data URI scheme being mandated in the HTML5 spec, by the way? I remember the first time I wanted to use on-the-fly generated images and it was great using that until you find out it was only supported by one or two browsers. :( (And I still need to find out why the bloody spec sometimes hangs my Firefox.) |
| 09:24 | <Philip`> | Someone should set up http://redire.ct/ftp://ftp.example.com/pub/... |
| 09:25 | <zcorpan> | asmodai: data: support is not required by html5 afaik |
| 09:25 | <asmodai> | zcorpan: A pity |
| 09:25 | <Hixie> | not is http: support :-) |
| 09:25 | <Hixie> | nor, rather |
| 09:25 | <Dashiva> | But all modern browsers support it for images, I believe |
| 09:25 | <zcorpan> | asmodai: except maybe for canvas toDataURL() which requires serializing as data: |
| 09:25 | <Philip`> | Only small ones, in IE8 |
| 09:25 | <asmodai> | Dashiva: Oh really? I hadn't retested that since I last did it. Granted, a few new major ones came out in the mean time. |
| 09:26 | <Dashiva> | Philip`: How small is small? |
| 09:26 | <Philip`> | zcorpan: I don't think the ability to generate data: strings really counts as data: support |
| 09:26 | <asmodai> | Small would be ok if it would be big enough for sparklines. |
| 09:26 | <Philip`> | since you can do that anyway by writing <script>var uri = "data:,";</script> |
| 09:26 | <Philip`> | Reading data URIs is probably the slightly more interesting thing :-) |
| 09:27 | <asmodai> | You wound me with your semantics. :P |
| 09:27 | <Philip`> | Dashiva: Wikipedia says 32KB |
| 09:29 | <Philip`> | Not sure if that's before or after base64 encoding |
| 09:29 | <zcorpan> | i think it's the url's length |
| 09:30 | <Philip`> | http://msdn.microsoft.com/en-us/library/cc848897(VS.85).aspx |
| 09:30 | <Philip`> | "Data URIs cannot be larger than 32,768 characters." |
| 09:32 | <Dashiva> | http://lists.w3.org/Archives/Public/public-html/2010Feb/0765.html |
| 09:32 | <Lachy> | asmodai, if you have AdBlock plus installed, add a filter so that it blocks the script http://www.whatwg.org/specs/web-apps/current-work/status.js . That will resolve the issues with the spec hanging Firefox. |
| 09:32 | <Dashiva> | It's unacceptable to tell people how to do proper accessibility |
| 09:32 | <danbri> | (in some charset...? I guess 32768 japanese chars might be too big?) |
| 09:33 | <asmodai> | Lachy: ahhh! |
| 09:33 | <asmodai> | Lachy: Let me try that |
| 09:33 | <zcorpan> | danbri: i'd assume s/characters/bytes/ |
| 09:33 | <Dashiva> | danbri: It's either base64 or percent-encoded, either way there's no Japanese |
| 09:34 | <zcorpan> | Dashiva: doesn't have to be percent-encoded |
| 09:34 | <zcorpan> | i think |
| 09:34 | <asmodai> | Lachy: I was driving myself nuts with it, because it kept locking up Firefox. |
| 09:34 | <Dashiva> | "If set to base64, this parameter specifies that the data specified by the sResourceData parameter is encoded as base64; otherwise, the data is assumed to be percent-encoded." |
| 09:34 | <asmodai> | Lachy: Is that a bug within Firefox? |
| 09:35 | <Dashiva> | "The resource data must be properly encoded; otherwise, an error occurs and the resource is not loaded. The "#" and "%" characters must be encoded, as well as control characters, non-US ASCII characters, and multibyte characters." |
| 09:35 | <Philip`> | Windows never uses bytes - I expect it's more likely to be UTF-16 code units than anything else |
| 09:35 | <zcorpan> | Dashiva: sure, doesn't say how much if anything actually is percent-encoded |
| 09:36 | <danbri> | does javascript inside data: URIs ever get interpreted? |
| 09:36 | <zcorpan> | danbri: sure |
| 09:36 | <asmodai> | Sounds like data: could use some clarification then ;) |
| 09:36 | <Dashiva> | zcorpan: non-ascii and multibyte seems pretty clear |
| 09:40 | <zcorpan> | data:,%C3%A5 and data:,å both work but have different length |
| 09:40 | <zcorpan> | um |
| 09:41 | <zcorpan> | the latter was supposed to say "å" |
| 09:41 | <asmodai> | did for me? |
| 09:42 | <zcorpan> | hmm, then it's opera doing something bogus with showing urls in irc |
| 09:42 | asmodai | pats irssi |
| 09:42 | danbri | see a with a bobble-hat in both lines |
| 09:42 | hsivonen | wonders what the "platform" changes to Opera on Mac and Linux are |
| 09:42 | <hsivonen> | Cocoa? newer Qt? Gtk? |
| 09:43 | <annevk> | zcorpan, yeah, in email as well |
| 09:43 | <zcorpan> | annevk: bug reported? |
| 09:43 | <annevk> | no, it's just now that you mention it |
| 09:45 | <hsivonen> | yeah, it seems Opera 10.50 on Mac is Cocoa |
| 09:45 | hsivonen | is a bit surprised Opera made the migration this late |
| 09:46 | <zcorpan> | annevk: could you report it? i'm in the middle of video stuff |
| 09:46 | hsivonen | can't find a blog post explaining what the "platform" changes on Linux are |
| 09:48 | <Rik|work> | hsivonen: http://my.opera.com/desktopteam/blog/happy-new-year ? |
| 09:49 | <hsivonen> | Rik|work: I see. |
| 09:49 | <hsivonen> | Rik|work: interesting choice of making an app that works without either Gtk+ or Qt these days |
| 09:49 | <hsivonen> | and the can optionally use either |
| 09:50 | <hsivonen> | the menus in the latest .deb release of Opera were seriously 1995 |
| 09:50 | <annevk> | zcorpan, yup, done |
| 09:53 | <jgraham> | hsivonen: Which menus? It looks OK to me |
| 10:43 | <annevk> | Hixie, I'm not sure whether XHR origin is always equal to Document origin |
| 10:44 | <Hixie> | when would it not be? |
| 10:44 | <Hixie> | it's not XHR origin that matters, it's script origin, surely |
| 10:45 | <annevk> | if that's true there's no need for an XHR origin |
| 10:45 | <annevk> | annoyingly I forgot the original scenario |
| 10:46 | <annevk> | maybe the origin always matched and it was just the base URL that was a different concept |
| 10:47 | <annevk> | the base URL coming from the Document of which the Window object had the XHR constructor object |
| 10:47 | <annevk> | hmm, so maybe I should kill XHR origin in favor of HTML5 script origin |
| 10:47 | <Hixie> | origin is new, right? |
| 10:48 | <annevk> | new? |
| 10:48 | <Hixie> | like, not implemented before you started adding stuff |
| 10:49 | <annevk> | the Origin header you mean? |
| 10:49 | annevk | is a bit confused |
| 10:49 | <Hixie> | me too |
| 10:49 | <Hixie> | i'm talking about whatever you pass the 'fetch' algorithm |
| 10:49 | <Hixie> | i think that's only for Origin header purposes |
| 10:49 | <Hixie> | so yeah, the header |
| 10:50 | <annevk> | right, that's new, but it would make sense if it matched what I defined to be "XMLHttpRequest origin" |
| 10:50 | <annevk> | I'm realizing the concept of "XMLHttpRequest origin" might not be needed; i.e. just XMLHttpRequest base URL is needed |
| 10:51 | <Hixie> | ah |
| 10:51 | <Hixie> | i'm not familiar enough with XHR to say one way or the other |
| 10:52 | <annevk> | my point was that if "XMLHttpRequest origin" is justified, the Origin header from HTML5 fetch needs to match what it says |
| 10:53 | <Hixie> | not necessarily |
| 10:53 | <Hixie> | a lot of the checks for script origin use the "effective origin" which is made using document.domain |
| 10:53 | <Hixie> | whereas the Origin: header is always made from the original origin |
| 10:53 | <Hixie> | which one does XHR use? |
| 10:53 | <Hixie> | can you XHR to the document.domain domain? |
| 10:54 | <Hixie> | i guess not, since you need a port and once you set document.domain you have no port |
| 10:55 | <annevk> | XHR does not care about document.domain |
| 10:55 | <annevk> | ah, but thanks for bringing that up |
| 10:56 | <Hixie> | even for the cross-document accesses? |
| 10:56 | <annevk> | so you have an iframe foo.example.org and a normal page bar.example.org |
| 10:56 | <annevk> | both set document.domain |
| 10:56 | <annevk> | to example.org |
| 10:56 | <annevk> | bar.example.org has the iframe |
| 10:56 | <Hixie> | so now foo can sent Origin: bar using bar's XHR? |
| 10:56 | <annevk> | bar.example.org then gets the XHR constructor out of foo.example.org |
| 10:56 | <annevk> | the request you make with that XHR is based on origin bar.example.org |
| 10:57 | <Hixie> | ok, good |
| 10:57 | <Hixie> | then you're using the entry script |
| 10:57 | <Hixie> | ('s document's origin) |
| 10:57 | <annevk> | the script runs on foo.example.org, no? |
| 10:58 | <Hixie> | you said bar.example.org then gets the XHR constructor, so i presume the entry script is from bar |
| 10:58 | <annevk> | damn it, I missed up |
| 10:58 | <Hixie> | but what happens if you pass the object to foo? |
| 10:58 | <Hixie> | does the same object then change origin? |
| 10:58 | <annevk> | foo.example.org gets a constructor out of bar.example.org and uses it, the origin is bar.example.org |
| 10:58 | <Hixie> | as in, is it the caller that decides the Origin: header, or the object? |
| 10:58 | <Hixie> | oh |
| 10:58 | <Hixie> | that's not the entry script then |
| 10:58 | <annevk> | right |
| 10:58 | <Hixie> | quite different from how Origin: works in HTML5 |
| 10:59 | <annevk> | the origin and base URL of XHR are determined by where its constructor comes from |
| 10:59 | <Dashiva> | What if you use the XMLHttpRequest object from a different window? |
| 10:59 | <annevk> | yes, same for base URL |
| 10:59 | <Dashiva> | Oh, that's what you meant |
| 11:00 | <annevk> | therefore I was saying that fetch should be modified somehow so that Origin matches the XHR origin |
| 11:00 | <Hixie> | Origin? or Referer? |
| 11:00 | <Hixie> | you pass in the value for Origin, so that's not a problem |
| 11:00 | <Hixie> | Referer however comes from the script currently |
| 11:02 | <annevk> | oh ok |
| 11:03 | <Hixie> | to pass an origin you just say "fetch foo from bar" where foo is the URL or resource and bar is the origin |
| 11:04 | <annevk> | http://www.w3.org/Bugs/Public/show_bug.cgi?id=9017 confused me |
| 11:04 | <annevk> | there you said something that indeed is not in the spec |
| 11:05 | <Hixie> | hm? |
| 11:06 | <zcorpan> | http://rishida.net/scripts/uniview/conversion.php - wonder if rishida knows about redirects |
| 11:07 | <zcorpan> | i assumed that his tool was broken since, well, it didn't work anymore |
| 11:08 | <annevk> | Hixie, there you say "The 'origin' parameter is only used for deciding if there should be an Origin header" |
| 11:08 | <annevk> | Hixie, it is also used for its value as it turns out |
| 11:08 | <Hixie> | oh, right |
| 11:09 | <Hixie> | sorry, i was speaking more loosely than i should have |
| 11:10 | zcorpan | sends an email to rishida |
| 11:30 | <zcorpan> | Hixie: shouldn't the categories table include formatBlock candidates? |
| 11:30 | <Hixie> | yeah, forgot about that one for some reason. Can you file a new bug for that? |
| 11:30 | <zcorpan> | sure |
| 11:31 | <zcorpan> | Hixie: could you move focus to the close button when filing a bug? |
| 11:32 | <zcorpan> | Hixie: i've filed dups twice now because i've pressed enter as an attempt to close the dialog |
| 11:33 | <Hixie> | you can just leave the box, it'll disappear after 10 seconds |
| 11:33 | <zcorpan> | i don't want to wait for 10 seconds |
| 11:33 | <Hixie> | if i move focus, it ruins the select() on the text field for keyboard users |
| 11:33 | <Hixie> | maybe i can make ESC hide it or something |
| 11:33 | <Hixie> | file a bug on that too |
| 11:33 | <Hixie> | (i'm about to go to bed) |
| 11:34 | <Hixie> | (thanks) |
| 11:34 | <zcorpan> | maybe you should make it not cover the whole screen so that it's ignorable for 10 seconds |
| 11:35 | <Hixie> | yeah... put that in the bug too, i'll see what i can do |
| 11:35 | <Hixie> | mark it P1 critical |
| 11:35 | <Hixie> | so i remember to look at it |
| 11:40 | <Hixie> | ok, bed time |
| 11:40 | <Hixie> | nn |
| 11:40 | <Hixie> | thanks for filing the bugs |
| 11:40 | <zcorpan> | nn |
| 11:42 | <annevk> | nn |
| 11:43 | <annevk> | (markp agreed with me btw) |
| 11:43 | <annevk> | (on the feed autodisco issue) |
| 11:43 | <gsnedders> | What autodiscovery issue? |
| 11:44 | <annevk> | http://www.w3.org/Bugs/Public/show_bug.cgi?id=9051 |
| 11:44 | <annevk> | Google unfortunately has not properly indexed all of irc-logs |
| 11:44 | annevk | wonders if Bing supports inurl: |
| 11:45 | <annevk> | Bing is even worse |
| 11:48 | <gsnedders> | Ah, if the spec says what it does now then it's incompatible with deployed usage |
| 11:48 | <gsnedders> | In a really big way |
| 11:48 | <gsnedders> | Like, every WordPress blog, every MT blog… |
| 11:56 | <annevk> | doh |
| 12:05 | <asmodai> | Ye gods, now I remember why I disliked zenburn as colour scheme. It just tends to blend into each other. Guess I better implement my own scheme. |
| 12:07 | <asmodai> | Guess colour schemes are as religious as it comes to shells, browsers, MUAs... :) |
| 12:38 | <asmodai> | Now that's pretty interesting: |
| 12:38 | <asmodai> | http://www.4p8.com/eric.brasseur/gamma.html? |
| 12:59 | jgraham | didn't actually read that article but doing one step down from skimming it suggested that his problem was that linear pixel averages should be replaced by gamma-weighted pixel averages |
| 12:59 | <jgraham> | Which I guess web browsers won't do because it would hurt performance |
| 13:00 | <jgraham> | (not that I have any special insight or anything) |
| 13:53 | <Dashiva> | "As others have pointed out: why have so many willfull violations to support legacy content if there is *no legacy HTML5 content to support?*" |
| 13:54 | <TabAtkins> | Where is *that* from, Dashiva? |
| 13:54 | <annevk> | bugs |
| 13:54 | <TabAtkins> | Augh god. |
| 13:54 | <annevk> | just somebody missing the point |
| 13:54 | <annevk> | no big deal |
| 13:55 | <Dashiva> | Well, I'm curious who these others are |
| 13:55 | <TabAtkins> | Sigh, I know. It hurts my head, though, that someone can misunderstand a term so profoundly. |
| 14:03 | <Philip`> | I'm more curious as to why they think an anonymous bug reporting mechanism is a good way to ask such questions |
| 14:03 | <Philip`> | since it makes it pretty inconvenient to see any response |
| 14:04 | <Dashiva> | That's not so important if you assume the response will necessarily be to agree fully |
| 14:09 | <Philip`> | If you explicitly note that others have already pointed out your point, and if the spec hasn't changed in response to them, and you don't offer any new arguments, then I don't see how anyone could assume the response would be to agree and to change the spec |
| 14:10 | <Dashiva> | All the reports going more or less just "add theora" suggest otherwise |
| 14:11 | TabAtkins | is heading to jack-in-the-box for his free sammich for breakfast. |
| 14:16 | <TabAtkins> | I'll mail you some, myakura. |
| 14:16 | <Philip`> | You could get straight fries, and then bend them |
| 14:18 | <myakura> | :S |
| 14:20 | <myakura> | ξ |
| 14:20 | <TabAtkins> | Hehe. |
| 14:30 | <asmodai> | Think our BK still does twister fries at least :) |
| 14:46 | <MikeSmith> | watching a video demo of how easy is to use Hover Mode is in the Nokia 900, it seems instead to do a pretty good job of illustrating Morgan Adams' argument that any possible mechanism for implementing mouse overs on touchscreen device are going to have serious usability problems |
| 14:48 | <MikeSmith> | dude even says at the beginning of the video "many people seems to have problems with it, but it's actually quite simple" and then proceeds to show how completely non-intuitive it is |
| 15:09 | <asmodai> | MikeSmith: Too close to the project :) |
| 15:09 | <asmodai> | MikeSmith: Or something like that. Some people need to take a step back and look at it freshly |
| 15:09 | <annevk> | anyone here know how a Ubuntu OS upgrade suddenly lets me use Wireless-N on a lenovo T60? I'm not even sure I have the hardware for that and yet it works... |
| 15:09 | <MikeSmith> | he's probably the guy who wrote the implementation |
| 15:10 | <MikeSmith> | asmodai: ↑ |
| 15:13 | <Philip`> | annevk: How can you tell it's using Wireless-N? |
| 15:13 | <Philip`> | (e.g. does it actually connect?) |
| 15:13 | <Philip`> | (and is it really connecting with N, or is the UI just broken?) |
| 15:14 | Philip` | wonders if Ubuntu has /sbin/iwconfig to tell you wireless informations |
| 15:15 | <annevk> | Philip`, I have set up my Airport Extreme in such a way that it broadcasts two different networks; one being just for Wireless-N |
| 15:16 | <annevk> | Philip`, before the OS upgrade the Wireless-N variant was not detected and now it is and it works fine |
| 15:16 | <asmodai> | MikeSmith: Heh, most likely. |
| 15:20 | <asmodai> | GAH, some people should NOT be allowed near a text editor. |
| 15:20 | <asmodai> | *kcccht* Step away from that editor, do not hurt that HTML any more |
| 15:21 | <Philip`> | HTML doesn't feel pain, it's only markup |
| 15:21 | <Philip`> | If those people want to communicate, and use HTML (badly) to do so, why should you stop them? |
| 15:21 | <asmodai> | Because I have to maintain it now :( |
| 15:22 | <Philip`> | Oh, okay |
| 15:23 | <asmodai> | Some application. With a lovely mix of PHP and inline HTML. And the HTML cannot make up its mind whether it wants to be HTML 3, HTML 4, or XHTML 1 |
| 15:25 | <Philip`> | Sounds like it wants to be HTML5 |
| 15:26 | <Philip`> | since that allows (a subset of) the union of HTML and XHTML syntax :-) |
| 15:29 | <asmodai> | Philip`: Yea |
| 15:30 | <asmodai> | Philip`: Right now I am rewriting it to HTML 4.01 Strict and keeping in mind HTML 5/XHTML constructs for easy moving over whenever we want. |
| 15:36 | <TabAtkins> | Yay, I just reduced 23 images to only 2 in conforming browsers! |
| 15:37 | <TabAtkins> | (read: in latest Firefox, because I'm using -moz-linear-gradient()) |
| 15:38 | <Philip`> | What do you mean by "conforming"? |
| 15:38 | <TabAtkins> | "Conforming to my personal set of specs that I think should be supported right now". |
| 15:38 | <Philip`> | By "specs" you mean "non-standard vendor-prefixed extensions"? |
| 15:38 | <TabAtkins> | No, they're standardized. |
| 15:39 | <Philip`> | Not with the -moz-, I hope :-p |
| 15:39 | <TabAtkins> | Just not in CR yet, so the vendor prefix is appropriate. |
| 15:39 | <jgraham> | Well something can have a spec without being a standard |
| 15:39 | <jgraham> | Indeed traditionally the W3C didn't publish "tandards" |
| 15:39 | <TabAtkins> | Bah. Semantics. |
| 15:39 | <jgraham> | huh, what happened to the "S" |
| 15:39 | <jgraham> | it was there, I saw it |
| 15:40 | <TabAtkins> | The w3c doesn't publish with an S. |
| 15:40 | <TabAtkins> | That is, "The w3c doesn't publih with an S." |
| 15:41 | <jgraham> | afaict the w3c doesn't publih |
| 15:42 | <TabAtkins> | Exactly. |
| 15:42 | <jgraham> | *doen't |
| 15:42 | <Philip`> | TabAtkins: So what you mean is "in browsers which use Mozilla's vendor-prefixed implementation of a spec that hasn't even reached CR yet"? |
| 15:42 | <annevk> | jgraham, your "s" is broken |
| 15:42 | <TabAtkins> | Philip`: Yes, that's what I mean. |
| 15:42 | <annevk> | jgraham, at least somewhat |
| 15:42 | <Philip`> | TabAtkins: I guess "conforming" is an easy typo for that |
| 15:45 | <jgraham> | annevk: you think there i omething trange? Everything eem perfectly normal a far a I can ee"" |
| 15:47 | <annevk> | maybe you are getting blind? |
| 15:48 | <Philip`> | Seems more like a case of hallucinating "s"s |
| 15:50 | <jgraham> | emm to me that I'm not the illy one having trange ight iue |
| 15:51 | <Philip`> | s/emm/eem/ ? |
| 15:53 | <jgraham> | ye |
| 16:05 | <Philip`> | It' urpriingly uneay aying entence o everal "" character exit inide everal word |
| 16:08 | <TabAtkins> | I ee. |
| 17:09 | <annevk> | mobile browser compat test already tests HTML5: http://www.w3.org/2010/01/wctmb2/ |
| 17:11 | <wycats> | Hixie: well... my last group of bugs were a bust :P |
| 17:12 | <Philip`> | annevk: All of HTML5? |
| 17:13 | <Philip`> | Opera 10.10 score 42% :-( |
| 17:22 | <Philip`> | (Oh, it's just a mixture of feature-existence tests and very basic functionality tests) |
| 17:22 | <Philip`> | (Useful for seeing what's implemented, but not whether it's implemented correctly) |
| 17:27 | <rektide> | i'm having a hard time getting a base case of Shared Worker code working |
| 17:28 | <rektide> | i wrote a post describing my attempt at http://groups.google.com/a/chromium.org/group/chromium-html5/browse_thread/thread/8a1a0042c8c32de3# |
| 17:28 | <rektide> | and the sample code is running at http://rektide.voodoowarez.com/2010/2/ping.html |
| 17:43 | <TabAtkins> | What's the most kosher way to indicate on <body> whether js is enabled? Right now I'm thinking <body class=nojs><script>document.body.className="js";</script> |
| 17:44 | <TabAtkins> | I think that will swap the class out immediately, before the rest of the page is processed, so any CSS that qualifies on .js will apply immediately without a flash of wrong styling? |
| 17:44 | <Philip`> | <html><noscript><body class=nojs></noscript><p>... |
| 17:44 | <Philip`> | (Warning: bad idea) |
| 17:44 | <jgraham> | No, really? |
| 17:44 | <TabAtkins> | I was asking for things that *weren't* bad ideas. ^_^ |
| 17:45 | <annevk> | seems like a cool idea to me |
| 17:45 | <annevk> | shorter too |
| 17:45 | <miketaylr> | TabAtkins: paul_irish has an article on that here, http://paulirish.com/2009/avoiding-the-fouc-v3/ |
| 17:46 | <paul_irish> | aye. i think using the html elem for the class is best, removes the neccessity of a script tag at the top of your <body> |
| 17:46 | <TabAtkins> | Ah, so he pops it onto <html> instead. Sure, works for me. |
| 17:46 | <paul_irish> | and css is all peachy with it. |
| 17:46 | <annevk> | "I prefer to write unique css for the no-javascript user" -- aren't those users obsolete by now? meh |
| 17:46 | <Philip`> | It's not valid HTML5, because noscript in head can only contain link/style/meta |
| 17:47 | <Philip`> | (and it's a parse error) |
| 17:47 | <TabAtkins> | annevk: No? Plenty of people run around with javascript off. |
| 17:47 | <TabAtkins> | They suffer, but I'd like to minimize the amount of that suffering that I cause. |
| 17:47 | <annevk> | weirdos |
| 17:47 | <miketaylr> | :D |
| 17:47 | <TabAtkins> | I agree, but still. ^_^ |
| 17:48 | <annevk> | time to get some ingredients for food |
| 17:48 | annevk | is hungry |
| 17:49 | <jgraham> | If you are hungry now, you ought to have got them earlier |
| 17:49 | <TabAtkins> | Benefit of putting it on <body>, though, is that I actually have a use for the <body> element other than just preventing IE from slurping all the new html5 elements that occur before a recognized html4 element into the head. |
| 17:49 | <TabAtkins> | But meh. It wont' kill me to just put a lonely, useless <body> in there. |
| 17:49 | <Philip`> | Ingredients for food are often themselves food |
| 17:49 | <Philip`> | It's a handy shortcut if you're in a rush |
| 17:51 | <TabAtkins> | Wait, nm, I already have an <a> as the first element in the body, so I don't need <body>. <html> it is, then. |
| 17:51 | <annevk> | jgraham, ideally that's how it works |
| 18:22 | karlushi | wonders if TTML could be of any use in combination of video element http://www.w3.org/TR/2010/CR-ttaf1-dfxp-20100223/ |
| 18:22 | <karlushi> | I guess it will be implemented in Quicktime. |
| 18:22 | <annevk> | jgraham, http://gist.github.com/312344 something for Web ECMAScript? (via Sjoerd Visscher) |
| 18:22 | <karlushi> | longtime I have checked the support of SMIL in webkit |
| 18:22 | <roc> | karlushi: i's huge, underspecified, and duplicates a lot of HTML+CSS features |
| 18:23 | <annevk> | oh yeah, I commented on that spec long ago saying something like that |
| 18:23 | <annevk> | like a couple of years back at least |
| 18:24 | <roc> | unfortunately there's a lot of momentum behind it because there are several shipped implementations |
| 18:24 | <karlushi> | roc, your comment relates to SMIL or TTML? or both ;) |
| 18:24 | <roc> | DFXP specifically |
| 18:24 | <karlushi> | ok |
| 18:24 | <roc> | extra-unfortunately, those implementations all implement different subsets of DFXP |
| 18:24 | <annevk> | oh, mine is about TTML (and also SMIL) |
| 18:24 | <annevk> | roc, guess everyone got their proprietary features in, hurray |
| 18:25 | <roc> | SMIL is huge and underspecified, but at least it doesn't duplicate HTML+CSS features :) |
| 18:25 | <karlushi> | my tests of SMILs for creating slideshow in the past have been frustrating. Low interoperability between real, quicktime, and some other products |
| 18:26 | <roc> | yeah, that's what you'd expect |
| 18:28 | karlushi | should try to test it again. Last time was in 2006 |
| 18:34 | <TabAtkins> | OMG Why is CSS so underspecified? >_< |
| 18:37 | <karlushi> | most commons failed tests on http://www.w3.org/2010/01/results-wctmb2 |
| 18:37 | <karlushi> | * contenteditable |
| 18:37 | <karlushi> | * <input type='date'> |
| 18:37 | <karlushi> | * <video> |
| 18:43 | <karlushi> | ah except pour opera mac 10.5 beta RC http://www.w3.org/2010/01/result-wctmb2/156 |
| 18:45 | <gsnedders> | And karlushi is doing what I did earlier and speaking two languages at once :) |
| 18:51 | <rektide> | i cant get my extremely simple ping pong SharedWorker code working |
| 18:51 | <rektide> | http://rektide.voodoowarez.com/2010/2/ping.html |
| 18:52 | <rektide> | ping.html sends "ping" to a pong.js shared worker, the shared worker should send "pong" back.... i'm not getting any errors from Chrome, but its not doing anything either. |
| 18:52 | <rektide> | if someone could put eyes on it, this should be a super-dead-simple absolutely minimimal SharedWorker implementation |
| 18:55 | <annevk> | does a spec example work? |
| 19:06 | <zcorpan> | rektide: you need to call postMessage on the 'connect' event's port; the 'message' event doesn't have .ports |
| 19:08 | <zcorpan> | rektide: you could store a reference to the port by doing onconnect = function(e) { self.port = e.ports[0]; |
| 19:13 | <zcorpan> | rektide: or maybe 'message' has ports |
| 19:14 | <zcorpan> | but it might be null |
| 19:18 | <zcorpan> | onconnect = function(event) { |
| 19:18 | <zcorpan> | var port = event.ports[0] |
| 19:18 | <zcorpan> | port.onmessage = function(event) { |
| 19:18 | <zcorpan> | port.postMessage('pong') |
| 19:18 | <zcorpan> | } |
| 19:18 | <zcorpan> | port.postMessaage("connect") |
| 19:18 | <zcorpan> | }; |
| 19:20 | <dimich> | Interesting that WebKit nightly actually reports the correct errors for rektide's URL |
| 19:20 | <dimich> | But Chrome does not. A bug. |
| 19:22 | <zcorpan> | opera says |
| 19:22 | <zcorpan> | Type mismatch (usually non-object value supplied where object required) |
| 19:22 | <zcorpan> | stacktrace: In function .onconnect |
| 19:22 | <zcorpan> | [source code not available] |
| 19:22 | <zcorpan> | ... |
| 19:22 | <zcorpan> | URL: http://rektide.voodoowarez.com/2010/2/pong.js |
| 19:22 | <zcorpan> | Context: Web Worker exception |
| 19:22 | <zcorpan> | Time: Tue Feb 23 2010 20:27:55 GMT+0100 |
| 19:43 | <zcorpan> | rektide: btw, you also need to call worker.port.start() if you're using addEventListener (as opposed to worker.port.onmessage, which calls start() implicitly) |
| 19:44 | <zcorpan> | rektide: otherwise the messages are on a queue waiting for the port to be enabled |
| 19:53 | <zcorpan> | is there a bugzilla component for web workers? |
| 19:57 | <annevk> | zcorpan, html5 |
| 19:58 | <zcorpan> | annevk: i thought workers was a webapps deliverable |
| 19:58 | <annevk> | zcorpan, yeah, but the whatwg.org script uses only a single component for everything |
| 19:59 | annevk | fills stuff in from complete.html |
| 19:59 | zcorpan | tried loading complete.html but gave up after a minute when the comment box still hadn't appeared and his macbook started to make noise and get warm |
| 20:31 | <zcorpan> | "I suspect that there are 10x as many HTML documents as SVG documents on the Web" -- http://www.w3.org/mid/4B842FC4.8040103⊙wo |
| 20:37 | <zcorpan> | rektide: i've posted a shared workers example to the whatwg list |
| 20:52 | <karlushi> | http://www.google.com/search?q=*+filetype%3Asvg 76,300,000 |
| 20:52 | <karlushi> | http://www.google.com/search?q=*+filetype%3Ahtml 25,270,000,000 |
| 20:53 | <karlushi> | http://www.google.com/search?q=*+filetype%3Axhtml 787,000,000 |
| 20:53 | <Dashiva> | Maybe he meant xhtml then, the numbers seem to fit |
| 20:54 | <annevk> | it was a joke |
| 20:58 | <objectivegiant> | *rimshot* |
| 20:59 | <karlushi> | annevk, shepazu making jokes? <grin/> |
| 21:03 | <annevk> | just saying... |
| 21:45 | <othermaciej> | Hixie: can you field http://www.w3.org/Bugs/Public/show_bug.cgi?id=9057 -- Julian replied to your suggestion |