| 00:02 | <Philip`> | Hixie: y |
| 00:02 | <Hixie> | i commented on the drawImage scaling bug |
| 00:02 | <Hixie> | the issue is going way over my head |
| 00:06 | <Philip`> | "they all look identical to me" - do you see e.g. a faint white line around the blue background, in the first large images in the attachment in Firefox? |
| 00:06 | <Philip`> | (and not in other browsers) |
| 00:10 | <Hixie> | nothing that i wouldn't chalk up to quality of implementation issues |
| 00:10 | <Hixie> | no, no faint anything |
| 00:11 | Philip` | doesn't know how platform-dependent it is |
| 00:11 | <Hixie> | chrome has very slight artefacting going diagonally down the second one |
| 00:12 | <Hixie> | resulting in six little minor blips |
| 00:12 | <Hixie> | but again, nothing more than i'd consider a quality-of-implementation issue |
| 00:14 | <Philip`> | "quality" implies that there's some agreement on the ideal behaviour, and a higher-quality implementation is one that's closer to that |
| 00:15 | <Philip`> | but (except for Chrome seeming buggy) I don't think any of the browsers seem better or worse quality, they're just different |
| 00:15 | <Philip`> | and it's not clear what behaviour they ought to be trying to converge on |
| 00:16 | <Philip`> | and the differences are significant enough to affect authors who want their content to look consistent |
| 00:18 | <Hixie> | so i gathered :-) |
| 00:18 | <Hixie> | i'm not saying we shouldn't fix this, just that the examples all looked the same to me - which isn't really important insofar as fixing the bug goes |
| 00:22 | <Philip`> | http://philip.html5.org/misc/filter.png is what I get from the 4th big panels in http://www.w3.org/Bugs/Public/attachment.cgi?id=915 |
| 00:22 | <Philip`> | using oldish browser versions |
| 00:23 | <Hixie> | my gecko trunk looks identical to your mozilla |
| 00:23 | <Hixie> | s/mozilla/firefox/ |
| 00:23 | <Hixie> | er |
| 00:23 | <Hixie> | ignore everythin i just said |
| 00:23 | <Hixie> | my gecko trunk looks identical to your _opera_ |
| 00:24 | <Philip`> | What platform? |
| 00:25 | <Philip`> | Hmm, how do I find out the version of a new Minefield? |
| 00:25 | <Philip`> | It seems to be missing from the About page :-( |
| 00:26 | <Hixie> | mac |
| 00:26 | <Hixie> | i don't have access to windows currently |
| 00:29 | <Philip`> | An auto-updated-just-now-but-who-knows-what-version-ID Minefield on Linux gives the same as my "Firefox" image (from 3.6.8 on Linux), and matches vlad's description of Firefox |
| 00:35 | <heycam> | Philip`, about: shows it |
| 00:37 | <Philip`> | heycam: Ah, thanks |
| 00:37 | <Philip`> | "about:" seems like a slightly bizarre URL |
| 00:37 | <heycam> | ooh about:about |
| 00:38 | <Hixie> | i love how bot in firefox and chrome some of the links on about:about are 404 |
| 00:39 | <Hixie> | heycam: btw i reassigned a bug about 'octet' to the WebIDL component earlier today |
| 00:39 | <TabAtkins> | he's already responded, hixie. |
| 00:39 | <Hixie> | awesome |
| 00:39 | <Hixie> | heycam: did i hear correctly that you were getting back into the game? |
| 00:39 | <heycam> | i clicked the about:memory link it crashed :/ |
| 00:39 | <heycam> | Hixie, yep, as of yesterday! |
| 00:39 | <Hixie> | woohoo! |
| 00:39 | <Hixie> | fantastic |
| 00:41 | <heycam> | i'll be spending most of this work doing webidl stuff |
| 00:41 | <Hixie> | heycam: where are you gonna be based? and what are you mainly working on? |
| 00:41 | <heycam> | i'm in the auckland moz office |
| 00:41 | <heycam> | not sure what i'll be mainly working on yet, tbd |
| 00:42 | <Hixie> | k |
| 00:42 | <heycam> | brb lunch |
| 01:04 | <cying> | is it possible to dynamically add a stylesheet to a DOM and add an event listener to listen for when the stylesheet has loaded? |
| 01:10 | <paul_irish> | cying: there isnt a clean way of doing that |
| 01:10 | <cying> | paul_irish: dohhh |
| 01:11 | <paul_irish> | most people add a #someuniquethinghere { color: red; } and poll getComputedStyle or whathaveyou to see when it changes |
| 01:13 | <paul_irish> | TabAtkins: in flex the reason box-flex defaults to 0 is so each element essentially has to opt-in to the flexibility of its box, right? |
| 01:14 | <cying> | paul_irish: :( |
| 01:14 | <paul_irish> | i know man. i know. |
| 01:15 | <TabAtkins> | paul_irish: Yeah. It seems like that's a better default. |
| 01:16 | <paul_irish> | TabAtkins: it does. also all demos ive seen have the flexible elements be immediate children of their container box.. is that a requirement? |
| 01:16 | <TabAtkins> | Yes. |
| 01:17 | <paul_irish> | thx |
| 01:17 | <Denicked> | Hi. |
| 01:17 | <TabAtkins> | The flexbox spec only concerns the rendering of the flexbox itself and its immediate children. Any further descendants are covered by the layout model of the child hosting it. |
| 01:20 | <Hixie> | cying: iirc the spec says <link rel=stylesheet onload='' ...> should work |
| 01:20 | <Hixie> | cying: dunno how widely that's supported though |
| 01:26 | <cying> | Hixie: sadly, it is not implemented in webkit |
| 01:36 | <ap> | Hixie: should a document be associated with an appcache if it was redirected? Say, I open foobar.php (which is does not have an entry in appcache) in a new browser window, and it redirects to foobar.html (which does) |
| 01:37 | <ap> | Hixie: I'd say yes, but both Safari and Firefox say no - is there a reason in HTML5 for that? |
| 01:39 | <Denicked> | Is there an offtopic whatwg chan like how there is #notmath and ##c++-social ? |
| 01:40 | <TabAtkins> | #whatwg is the off-topic #whatwg channel. |
| 01:40 | <Denicked> | haha riight |
| 01:40 | <Denicked> | I need advice on motivation |
| 01:41 | <Denicked> | This is my final semester to complete my bachelor of science in computer science |
| 01:41 | <Denicked> | I have an internship, and I hate my job. I think I don't enjoy computers anymore. |
| 01:41 | <Denicked> | I feel like it's too established too |
| 01:41 | <Denicked> | the pioneering aspect of it isn't what it used to be |
| 01:42 | <TabAtkins> | Do webapps instead. |
| 01:42 | <Denicked> | what kind of webapps? |
| 01:42 | <Denicked> | at my internship I do CRM stuff |
| 01:42 | <TabAtkins> | Just in general. There's a lot going on to inspire a pioneering spirit. |
| 01:43 | <Denicked> | At my internship I'm learning this web management system called SugarCRM. I h8 it |
| 01:44 | <TabAtkins> | Yeah, CRM is usually going to be disgusting enterprise-y stuff. |
| 01:44 | <Denicked> | What makes you say "disgusting" ? |
| 01:45 | <Denicked> | It's just boring. I don't get any satisfaction out of it. I don't think I'm making the world a better place or contributing positively to society. The boss just tries to sucker people. He gives me full projects and has me going through anxiety attacks because I have to learn and do the project by set deadlines. It's not fair to me nor the client. |
| 01:49 | <TabAtkins> | BigThen do something else? |
| 01:50 | <Denicked> | I was thinking of going to lawschool |
| 01:55 | <Denicked> | TabAtkins what do you do for a living |
| 01:56 | <Hixie> | bummer, missed ap |
| 01:56 | <Denicked> | What's that? I don't quite follow you Hixie |
| 01:56 | <Hixie> | ap asked me a question above, but he quit before i could answer |
| 01:56 | <Denicked> | Oh interesting. |
| 01:57 | <Denicked> | well not quite, but I'm just making conversation hehe |
| 01:57 | <Hixie> | i wonder if the chairs are planning on making any decisions with all these polls we've taken |
| 01:57 | <Denicked> | Hixie what do you do for a living |
| 01:58 | <Hixie> | i write the html5 spec |
| 01:58 | <Hixie> | well, html spec |
| 01:58 | <Denicked> | for a living!? |
| 01:58 | <Hixie> | yeah, sweet huh? |
| 01:58 | <Denicked> | Yes sir |
| 01:59 | <Denicked> | To be part of the team that sets the standards for html, you have to be the top of your field |
| 02:01 | <Hixie> | actually it turns out that all i had to do was do it |
| 02:01 | <Hixie> | nobody checked that i was top of anything |
| 02:01 | <Hixie> | it's just that nobody else was doing it |
| 02:01 | <Hixie> | so... |
| 02:02 | <Philip`> | You had to get dozens of other people to believe in what you were doing, too |
| 02:03 | <Philip`> | (which doesn't require being the top, but it does generally require being sufficiently adequate) |
| 02:03 | <Hixie> | i guess |
| 02:04 | <heycam> | HTML5: It's sufficiently adequate |
| 02:05 | <TabAtkins> | Denicked: Sorry, was away for a bit. |
| 02:06 | <TabAtkins> | I write CSS specs, and work on the Chrome team. |
| 02:06 | <TabAtkins> | And give developer talks and stuff, sometimes. |
| 02:06 | <TabAtkins> | Before that, though, I was a web developer for a small software company, which was very fulfilling. |
| 02:06 | <TabAtkins> | (Google just offered to pay me a bunch more money.) |
| 02:14 | <Hixie> | TabAtkins: working on css specs isn't fulfilling? :-P |
| 02:15 | <TabAtkins> | That wasn't a statement about the relative merit of my jobs. ^_^ Both my current and previous job are and were awesome. |
| 02:17 | <Denicked> | TabAtkins what can I do to reach your level |
| 02:17 | <Denicked> | and Hixie's level |
| 02:17 | <TabAtkins> | If you want to work on standards? Start working on standards. ^_^ Get on mailing lists and become a useful contributor. Standards are run like an open-source project (at least, the good ones are). |
| 02:18 | <TabAtkins> | If you want to do more fulfilling comp sci work? Get a job with a smaller company. That doesnt' guarantee non-asshole bosses, but if you get a good boss, I think being in a smaller environment with relatively more responsibilities is a lot of fun. |
| 02:19 | <Denicked> | I'm 24 making 12 dollars an hour while zuckerberg was a billionaire at like 22 |
| 02:19 | <TabAtkins> | And? |
| 02:19 | <Denicked> | TabAtkins the company I'm with is 6 people. |
| 02:20 | <TabAtkins> | Then get a better company. |
| 02:20 | <Denicked> | The boss walks around with a shirt that says on the front "I see stupid people" |
| 02:20 | <TabAtkins> | "I work with assholes" isn't a good reason to quit your field. It's a good reason to quit your company. |
| 02:21 | <Denicked> | good point |
| 02:22 | <Denicked> | TabAtkins you say And? I say and I feel poor while other people have the money and girls :-P |
| 02:23 | <TabAtkins> | I said "And?" because you were comparing yourself to Zuckerberg. He's an extreme outlier. |
| 02:23 | <TabAtkins> | *I'm* an outlier - I got lucky with several of my life choices, and had a good wife who made some good decisions for me. |
| 02:24 | <TabAtkins> | But you can definitely do better for yourself. Just jump out and try. |
| 02:24 | <TabAtkins> | If you're young and single, it's hard to starve. You can live remarkably cheaply. |
| 02:24 | <Denicked> | I think it's psychological. When you have a lot of bad luck or bad decisions it starts to have a psychological impact |
| 02:25 | <TabAtkins> | Agreed. So you have to recognize the source consciously, so you can work around it. |
| 02:25 | <TabAtkins> | s/had a good wife/have a good wife/ |
| 02:30 | <Denicked> | TabAtkins do you live in cali? |
| 02:31 | <TabAtkins> | Yeah, now I do. (I was in Texas before Google hired me.) |
| 02:31 | <Denicked> | My cousin works for google too. |
| 02:31 | <TabAtkins> | I'm out for the night, though. See you later. |
| 02:31 | <Denicked> | Bye |
| 02:33 | <TabAtkins> | My conclusion, dude, is that you don't need advice. You already know what the problem is. Make sure you consciously admit that. And thus, you know what your next steps can be. Do so. That's all you need. |
| 02:33 | <Denicked> | That's actually the best advice I've ever gotten. |
| 03:11 | <heycam> | shepazu, ping |
| 03:11 | <dgouldin> | I'm not seeing html5 audio/video events bubbling up the dom ... anybody have an idea why that might be? |
| 03:47 | <karlcow> | http://tools.ietf.org/html/rfc6068 |
| 03:47 | <karlcow> | mailto URI scheme |
| 06:25 | <itissid> | hi was full text search carried over to webstorage api for sqlite? |
| 06:25 | <itissid> | From google gears that is.. |
| 06:26 | <itissid> | aka fts3 |
| 08:17 | <zcorpan> | Hixie: here's a patch for the issues graph, the month labels are becoming unreadable |
| 08:17 | <zcorpan> | var months = ['Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun', 'Jul', 'Aug', 'Sep', 'Oct', 'Nov', 'Dec']; |
| 08:27 | <hsivonen> | what's the address of the graph? |
| 08:28 | <zcorpan> | http://www.whatwg.org/issues/data.html |
| 08:28 | <hsivonen> | thanks |
| 08:29 | hsivonen | wishes it were SVG instead of <canvas> |
| 08:29 | <hsivonen> | it looks fuzzy with my zoom level |
| 09:12 | <hsivonen> | Did Opera change pragma no-cache as a meta handling in the standards mode in 10.62 relative to 10.60? |
| 09:13 | <annevk> | assigning to readonly values does not throw a TypeError, right? |
| 09:13 | <annevk> | hsivonen, probably not |
| 09:14 | <jgraham> | No, unless you are in strict mode, maybe |
| 09:14 | <jgraham> | Actually that's an interesting point |
| 09:14 | <hsivonen> | Am I going crazy? my notes say Opera 10.61 didn't honor pragma no-cache as a meta in the standards mode--only in the quirks mode |
| 09:14 | <hsivonen> | now I see 10.62 honoring it in both modes |
| 09:24 | <hsivonen> | OK. my previous testing of 10.61 was wrong |
| 09:24 | <hsivonen> | (or Opera had differences between operating systems, which is unlikely) |
| 09:39 | <annevk> | http://www.w3.org/Bugs/Public/show_bug.cgi?id=10977 -- some properties are directly on the object right? such as onclick et al? |
| 09:46 | <zcorpan> | why don't we have a policy to only accept automated tests for the html5 testsuite? |
| 09:49 | <jgraham> | zcorpan: We do, sort of |
| 09:49 | <jgraham> | Or at least everyone agreed apart, possibly, from Microsoft |
| 09:50 | <zcorpan> | and microsoft is submitting non-automated tests |
| 09:50 | <jgraham> | Well as long as everyone else agrees not to accept them, there is no problem |
| 09:51 | <annevk> | they're pretty shitty tests too |
| 09:51 | <zcorpan> | there's a problem if people are wasting time reviewing non-automated tests |
| 09:51 | <annevk> | don't they see this themselves? |
| 09:52 | <annevk> | it's kind of worrisome |
| 09:52 | <annevk> | and similar to the crappy CSS 2.1 situation |
| 09:52 | <annevk> | zcorpan, right |
| 10:43 | <hsivonen> | why is MS writing non-automated tests in the first place? do they have an army of people running tests manually on every code change? |
| 10:44 | <jgraham> | I believe they have a screenshot-based system |
| 10:44 | <micheil> | hsivonen: well, naturally, right, how else would you do it? |
| 10:44 | <micheil> | :P |
| 10:44 | <jgraham> | I guess they have an army of people to deal with the scaling issues of such a system |
| 10:45 | <jgraham> | (like marking tens of thousands of screenshots as pass/fail because the font rendering changed slightly) |
| 10:50 | <micheil> | hmm.. anyone know why people wanted compression on websockets? |
| 10:51 | <micheil> | or frame level meta-data? |
| 10:51 | <jgraham> | annevk: I always get very confused about where DOM properties live on the prototype chain. It seems to be rather browser dependent. |
| 10:51 | <jgraham> | micheil: I assume people wanted compression in order to transfer less data |
| 10:52 | <jgraham> | micheil: The frame-level metadata stuff I didn't get at all |
| 10:52 | <micheil> | jgraham: was that without any prior look into how people were and are using the protocol? |
| 10:52 | <jgraham> | Was what? |
| 10:53 | <micheil> | well, I'm guessing the average message size for most applications would be < 1024bytes |
| 10:53 | <micheil> | or 1 Kb. |
| 10:53 | <micheil> | you wouldn't surely save anything by compressing 1 Kb. |
| 10:53 | <jgraham> | That seems like a big assumption |
| 10:53 | <micheil> | jgraham: my examples from which I draw that assumption are things like socket.io |
| 10:54 | <micheil> | where the most of applications are things like games, chats, etc. |
| 10:54 | <micheil> | I so far don't know of one example in development or production where people are sending large data over websockets |
| 10:54 | <jgraham> | It seems dangerous to draw conclusions about how the protocol will be used in the future based on how it is used before it is in any production browsers |
| 10:55 | <jgraham> | Well Chrome I guess |
| 10:55 | jgraham | is discounting Safari 5 for being too old |
| 10:55 | <micheil> | Chrome, Chromium, Safari |
| 10:55 | <micheil> | Safari 5.02 is draft 76. |
| 10:55 | <jgraham> | Oh, nice |
| 10:55 | <micheil> | Firefox 4 is draft 76. |
| 10:55 | <jgraham> | So -75 is dead |
| 10:55 | <micheil> | http://github.com/miksago/node-websocket-server/wiki/Browser-Support |
| 10:55 | <jgraham> | Anyway, it is hardly mainstream |
| 10:56 | <micheil> | pretty much |
| 10:56 | <jgraham> | Get back to me when youtube are using websockets for video-related stuff |
| 10:56 | <jgraham> | for example |
| 10:57 | <micheil> | well, I know a few startups are using it at core level of their applications. |
| 10:57 | <micheil> | youtube is old in all. |
| 10:57 | <jgraham> | anyway, I agree that compression is not really a v1 feature |
| 10:57 | <micheil> | I think compression should be end-developer level. |
| 10:57 | <jgraham> | Youtube was an example of something mainstream that involves transporting lots of data |
| 10:57 | <micheil> | if you're exceeding average tcp-frame sizes, then I think you'll have other problems |
| 10:58 | <jgraham> | Why? |
| 10:58 | <micheil> | also, as for video, shouldn't you be using already existing transports; eg, RTSP |
| 10:58 | <micheil> | well, you're going to probably break the parsers in browsers. |
| 10:58 | <micheil> | or in the servers. |
| 10:58 | <jgraham> | Well that would be a bug |
| 10:58 | <micheil> | true |
| 10:59 | <jgraham> | I mean you don't hear people say "web documents have to be small because browsers might have buggy network stacks" |
| 10:59 | <micheil> | true |
| 10:59 | <jgraham> | No readon websockets is anay different |
| 10:59 | <jgraham> | *reason |
| 10:59 | <micheil> | but either way, you're going to probably increase latency |
| 11:00 | <jgraham> | Right, there might be good reasons to favour packets of a certian size |
| 11:00 | <jgraham> | None of this means that compression isn't a good idea |
| 11:00 | <micheil> | anyway. I'm going to try some tests with large and small packets to see if browsers will actually join messages into one tcp packet or split |
| 11:00 | <jgraham> | also, people have been talking about stream-level compression |
| 11:00 | <jgraham> | rather than per-message compression |
| 11:00 | <micheil> | I still think it's something that the end developer should do. |
| 11:01 | <jgraham> | Which would mean the packet size wouldn't matter |
| 11:01 | <micheil> | stream level compression? |
| 11:01 | <jgraham> | It seems insane to have to implement gzip in each protocol |
| 11:01 | <jgraham> | And also in javascript |
| 11:01 | <jgraham> | When browsers already have it built in |
| 11:02 | <micheil> | I'm just trying to work out the fastest way to parse messages. |
| 11:02 | <micheil> | because I'm finding that's a bottleneck in my server. |
| 11:03 | <micheil> | for instance, if my server receives say, 2 Mb's of data (multiple frames/messages) |
| 11:03 | <micheil> | then it actually has to read every byte in that 2 Mb's just to figure out where to emit out a new client to the end-developer |
| 11:03 | <micheil> | * emit out a new message to the end developer |
| 11:05 | <jgraham> | Hope that the length-framed version gets traction :) |
| 11:06 | <jgraham> | (right about now you might want to start getting annoyed at IETF for being slow) |
| 11:10 | <micheil> | jgraham: I gave up trying to keep up with the mailing list banter about 4 months ago, |
| 11:16 | <jgraham> | At the moment, the tl;dr is 1) People: we should make the handshake like x. 2) Adam / Maciej: That seems weak against cross-protocol attacks since all the security properties depend on assumption y. 3) GOTO 1 |
| 11:17 | <jgraham> | (well there are a few more steps inside the loop where people discuss whether it is sufficient to have security that depends only on y |
| 11:17 | <jgraham> | ) |
| 11:50 | <hsivonen> | sigh. JF reopened the DRM bug. :-( |
| 11:55 | <annevk> | I don't see it on http://lists.w3.org/Archives/Public/public-html-bugzilla/2010Oct/ ? |
| 11:55 | <annevk> | I guess Mike forgot to turn email back on like he said he thought he would? |
| 11:58 | <Lachy> | hsivonen, which is the DRM bug? |
| 11:59 | <zcorpan> | http://www.w3.org/Bugs/Public/show_bug.cgi?id=10902 |
| 12:10 | <annevk> | whoa |
| 12:10 | <annevk> | 1600 messages |
| 12:10 | <annevk> | just noticed that |
| 12:13 | <annevk> | I always forget |
| 12:13 | <annevk> | is it "to dispatch a HTTP ... event" or "to dispatch an HTTP ... event"? |
| 12:13 | <Philip`> | "an" |
| 12:13 | <Philip`> | Other people may disagree but they are wrong |
| 12:13 | <annevk> | great |
| 12:17 | <zcorpan> | Philip`: those that pronounce "HTTP" as "huttup"? |
| 12:18 | <Philip`> | I was thinking of people who say "haitch" |
| 12:26 | <hsivonen> | bah. silly me. I used a local hg rev number in a commit message like a n00b who thinks he is using svn. :-( |
| 13:56 | <hsivonen> | uh oh. another batch of "pre-LC1" email coming in |
| 13:57 | Philip` | notes the lack of optimism in naming it "LC1" instead of "LC" |
| 13:59 | <annevk> | MikeSmith, maybe next time give a heads up about these administrative changes that flood the inboxes? |
| 13:59 | <MikeSmith> | which mailing list? |
| 14:00 | <MikeSmith> | I thought I had mail turned off |
| 14:00 | <annevk> | everyone who is cc'd still gets email |
| 14:01 | <MikeSmith> | ah |
| 14:01 | <MikeSmith> | sorry |
| 14:02 | <MikeSmith> | fwiw, I think I have them all done now |
| 14:03 | <annevk> | k |
| 14:04 | <MikeSmith> | annevk: we don't need to set up a separate component for the HTML5-HTML4 differences doc, right? |
| 14:07 | <annevk> | rather not, no |
| 14:07 | <annevk> | also, I don't think it should go to LC |
| 14:08 | <MikeSmith> | ok |
| 14:08 | <hsivonen> | have the Chairs given any indication of a schedule for Decisions on ISSUEs that have had polls? |
| 14:08 | <hsivonen> | one would think it's a no-brainer to evaluate which proposal drew the weaker objections at http://www.w3.org/2002/09/wbs/40318/issues-4-84-objection-poll/results |
| 14:09 | <hsivonen> | yet, it's been over 2 months without a Decision |
| 14:16 | <Workshiva> | Maybe they're saving the easy ones for last |
| 15:47 | <mathrick> | Hixie: not that you can really do anything about it at this point, but amusingly Chromium/Webkit sort of deadlocks if you call executeSql from inside a statement callback, which is the only way to execute a query based on the results of a previous one |
| 15:48 | <paul_irish> | mathrick: that might be worth filing a webkit ticket for. |
| 15:48 | <mathrick> | I think the proliferation of callbacks in Web APIs has got out of control, and certain things are simply impossible if all you have are callback-based APIs |
| 15:49 | <mathrick> | paul_irish: yes, I'm reducing it to the minimal testcase now |
| 15:49 | <paul_irish> | :) |
| 15:50 | <mathrick> | still, any API that could possibly ever be called from an event listener should have a sync version, or else things go wrong |
| 15:50 | <jgraham> | mathrick: Any suggestions for good API patterns that handle the requirement to be async without callbacks and with current js language features? |
| 15:51 | <jgraham> | "it should be sync" doesn't really work if sync calls can block the browser for an arbitary amount of time |
| 15:52 | <jgraham> | (Opera handles this unusually well, but still) |
| 15:52 | <mathrick> | sure, but there's openDatabaseSync already, so that argument doesn't really fly in the specific case of webSQL |
| 15:52 | <jgraham> | That's only for workers |
| 15:52 | <jgraham> | Isn't it? |
| 15:53 | <mathrick> | oh, it doesn't really say it in the spec that I've noticed |
| 15:53 | mathrick | rereads |
| 15:53 | <mathrick> | jgraham: but transactions are already run async |
| 15:54 | <grantg> | Does anyone know if chrome 6 supports the file api? |
| 15:54 | <annevk> | I wrote http://dev.w3.org/2006/webapi/progress/ today |
| 15:54 | <mathrick> | so if you add another level of executeSql working async, it becomes quite unmanageable |
| 15:54 | <grantg> | I know 7 does. |
| 15:54 | <annevk> | feedback welcome before I post to public-webapps |
| 15:54 | annevk | needs to do some shopping first |
| 15:55 | <mathrick> | jgraham: ah, right, WorkerUtilsDatabase |
| 15:56 | <grantg> | Can anyone figure out how to make webworkers work with graphics bound to the CPU core. |
| 15:56 | <grantg> | I tried that with the gbc emu, but it made performance slower, since the framebuffer was being passed back to the UI. |
| 15:57 | <mathrick> | jgraham: one way to alleviate it a bit would be to introduce checkpoints / barriers that could be inserted optionally and would essentially put the caller to sleep until the results are available |
| 15:58 | <grantg> | WebWorks seem to also block any onmessage event inside the worker when it's running (Feature?). |
| 15:58 | <jgraham> | mathrick: That would be a language feature,no? |
| 15:58 | <mathrick> | so transaction.barrier() would make sure all transaction steps queued until now have been performed |
| 15:58 | <grantg> | *WebWorkers |
| 15:58 | <jgraham> | co-routines? |
| 15:58 | <mathrick> | jgraham: it doesn't really need to be exposed in the language |
| 15:58 | <jgraham> | Well it doesn't work if you want to get data back, does it? |
| 15:59 | <jgraham> | It sounds like |
| 15:59 | <jgraham> | foo = (yield something) |
| 15:59 | <mathrick> | it would if your callback stuffed the result in a known location, which is not hard to do with closures, which JS thankfully does have |
| 15:59 | <jgraham> | in python |
| 15:59 | <mathrick> | ah, you're looking at it from the other side |
| 16:00 | <mathrick> | I mean more like this: |
| 16:00 | <karlcow> | Very early alpha state pure JavaScript implementation of the RDFa API draft spec released -- http://lists.w3.org/Archives/Public/public-rdfa-wg/2010Oct/0029 |
| 16:01 | <mathrick> | function myTransaction(tx) { var result; tx.executeSql("...",[], function(tx, r) { result = r }); tx.checkpoint(); alert(result) } |
| 16:07 | <mathrick> | jgraham: related is http://chumsley.org/jwacs/ |
| 16:07 | <mathrick> | which is a dialect with continuations that transforms to regular JS |
| 16:08 | <mathrick> | though I don't believe you can actually use it to fix things like "can't use any async APIs from inside an event listener" |
| 16:12 | <jgraham> | Anyone know what ES5-strict-mode-enabled browsers do when setting a readonly IDL property? |
| 16:12 | <jgraham> | per the current letter of WebIDL they should not throw since it is ignorant of strict mode, but per the spirit of Strict Mode they should throw |
| 16:42 | <annevk> | no feedback? lame |
| 16:45 | <Philip`> | annevk: You forgot all the semicolons in the example code |
| 16:47 | <annevk> | wontfix :) |
| 17:18 | <matjas> | https://bugzilla.mozilla.org/show_bug.cgi?id=591981 made me wonder, where in the HTML5 spec does it state that the execute-in-insertion-order for DOM inserted <script>s SHOULD NOT be enforced |
| 17:21 | <Hixie> | http://www.whatwg.org/specs/web-apps/current-work/complete.html#running-a-script |
| 17:22 | <annevk> | hairy |
| 17:22 | <annevk> | Hixie, took over Progress Events |
| 17:22 | <annevk> | Hixie, http://dev.w3.org/2006/webapi/progress/ |
| 17:22 | <Hixie> | cool |
| 17:24 | <annevk> | I was reminded once again that because of SVG 1.2 Tiny it has the most suckiest interface members |
| 17:24 | <annevk> | made me sad |
| 17:25 | <Hixie> | oh? |
| 17:25 | <annevk> | lengthComputable, loaded, total... |
| 17:26 | <annevk> | hasMax, value, max, or some such would've been nicer |
| 17:26 | <Hixie> | *shrug* |
| 17:27 | <annevk> | I guess |
| 17:43 | <annevk> | I should probably add the reference to SVG back in, I actually forgot it came from there while writing |
| 17:43 | <annevk> | so not "back in" but "in" |
| 17:57 | <annevk> | damn it, http://lists.w3.org/Archives/Public/public-html-bugzilla/2010Oct/ is still not getting any email |
| 17:58 | <annevk> | and it's way too late in Tokyo to fix that now |
| 17:58 | <micheil> | hmm? |
| 17:58 | <micheil> | it's only 4am AEST |
| 17:58 | <micheil> | (not sure whether that's too close to tokyo) |
| 17:59 | <annevk> | I got 2AM |
| 17:59 | <annevk> | for "time in tokyo" on Google |
| 17:59 | <micheil> | yeah |
| 17:59 | <micheil> | I'm obviously up late. |
| 18:01 | <annevk> | well, I suppose "late" is a personal thing |
| 18:02 | <annevk> | but 2AM is late for Mike afaik :) |
| 18:03 | <micheil> | true |
| 18:06 | Ms2ger | was summoned by annevk |
| 18:07 | <annevk> | oh hey |
| 18:07 | <annevk> | I should improve my summoning skills |
| 18:07 | <jgraham> | Seriously ;) |
| 18:09 | <annevk> | Ms2ger, so I actually need to fix a few more things and then hopefully you can generate a copy and then we can hopefully get Art to convince the rest of the W3C we should publish Web DOM Core with a shortname of domcore |
| 18:09 | <annevk> | I forgot to put in the whole patent policy disclaimer part |
| 18:10 | <Ms2ger> | Sure |
| 18:11 | <Ms2ger> | Did I miss anything besides someone dumping 300 bugmails in my inbox? |
| 18:13 | <annevk> | Ms2ger, I wrote http://dev.w3.org/2006/webapi/progress/ and nobody cares ;p |
| 18:13 | <gsnedders> | annevk: You wrote it, why would they? |
| 18:13 | <annevk> | oh, and I believe the DRM bug got reopened, Fronteers conference is this week, and some other stuff I do not recall right now |
| 18:14 | <annevk> | gsnedders, excellent point |
| 18:16 | <annevk> | Ms2ger, try it now? |
| 18:16 | <annevk> | after hg pull / hg update |
| 18:18 | <annevk> | hah |
| 18:19 | <annevk> | I guess I better start summoning now, for it takes a long time :) |
| 18:53 | <Ms2ger> | you need the image to load in order to get a HTMLImageElement object |
| 18:53 | <Ms2ger> | Really? |
| 18:54 | <Hixie> | new Image(); and document.createElement('img') both give you an HTMLImageElement object with no image loaded |
| 18:55 | Ms2ger | expects jgraham to correct himself on the list |
| 18:57 | <Ms2ger> | Also, annevk, why'd you mark bug 10832 as VERIFIED instead of CLOSED? |
| 18:58 | <annevk> | dunno, I can mark it CLOSED |
| 18:58 | <annevk> | good to see you're back so soon |
| 18:58 | <annevk> | can you generate a DOM Core copy? |
| 19:04 | <annevk> | cool |
| 19:04 | <annevk> | lets see what pubrules says |
| 19:05 | <Ms2ger> | http://dvcs.w3.org/hg/domcore/rev/4e81c4313c3e |
| 19:05 | <annevk> | oh |
| 19:06 | <annevk> | Anolis does not understand [STATUS] or [LONGSTATUS]? |
| 19:06 | <annevk> | bah |
| 19:06 | <Ms2ger> | Looks like it should |
| 19:07 | <annevk> | hmm |
| 19:07 | <annevk> | I'll fix those manually for now in the generated copy so I have something to give to Art |
| 19:09 | <Ms2ger> | I'm fixing it |
| 19:09 | <annevk> | ok |
| 19:09 | <annevk> | i'll wait |
| 19:10 | <gsnedders> | annevk: --w3c-compat-options or something like that? |
| 19:11 | <gsnedders> | (the design of it in Anolis 1 means you basically have to compute the replacement string before doing any replacements, and they're way more expensive to compute than other replacements) |
| 19:12 | <annevk> | I don't know Anolis unfortunately (or why I cannot get it to work) but these can be replaced whenever |
| 19:14 | <Ms2ger> | annevk, could you push to w3.org? |
| 19:15 | <annevk> | sure |
| 19:16 | <annevk> | done |
| 19:16 | <Ms2ger> | Apparently I lost the right incantation once more |
| 19:17 | <annevk> | looks very official now: http://dvcs.w3.org/hg/domcore/raw-file/tip/Overview.html |
| 19:17 | <annevk> | ta |
| 19:17 | <Ms2ger> | Except for the IDL blocks |
| 19:19 | <annevk> | hmm, lets copy the style block from XHR |
| 19:20 | <annevk> | pushed |
| 19:25 | <Ms2ger> | And pushed |
| 19:26 | <annevk> | hmm, reference section is now split? I guess that's fine |
| 19:27 | <annevk> | and all <var> are suddenly linked |
| 19:28 | <TabAtkins> | OMIGOD I HATE AIRLINES >_< |
| 19:28 | <annevk> | and <a> is inside <code>, but I guess that is ok |
| 19:28 | <jgraham> | TabAtkins: Yes |
| 19:28 | <annevk> | Ms2ger, so should we do s/<var>/<var title>/ ? |
| 19:29 | <jgraham> | Ms2ger: How do you get a reference to a HTMLImageElement for a SVG image if you don't support SVG elements in image? |
| 19:30 | <jgraham> | Or am I not understanding the correction that you are looking for? |
| 19:30 | <Ms2ger> | I might be misunderstanding your email |
| 19:31 | <annevk> | Ms2ger, likely publication date is Oct 7 |
| 19:32 | <jgraham> | Ms2ger: The point I was trying to make is that it is OK to test that e.g SVG in drawImage works if you set up the test so that it doesn't fail if SVG in <img> elements doesn' work at all |
| 19:32 | <jgraham> | Since it seems to be a requirement of the spec that drawImage supports any format for which you can get a HTMLImageElement |
| 19:33 | <jgraham> | Which I think is equivalent to "any format supported by <img>" |
| 19:34 | <romeo_> | cool. The idiotic mega spec killed my browser and all my tabs once again... |
| 19:35 | <jgraham> | romeo_: Which browser? |
| 19:36 | <romeo_> | jgraham, will tell you when I manage to launch it again |
| 19:37 | <jgraham> | You're not sure which browser you use? |
| 19:37 | <jgraham> | Actually I guess I sometimes realise that things are a bit odd because I am not using the browser I hink I am |
| 19:37 | <jgraham> | *think |
| 19:38 | <romeo_> | some Firefox 3.6 |
| 19:39 | <romeo_> | sessionstore.js is a single line of condensed js. *sigh* |
| 19:40 | <Ms2ger> | annevk, I hope it's almost ready now ;) |
| 19:40 | <jgraham> | romeo_: I hear that Firefox 4 improves spec loading time a great deal |
| 19:41 | <TabAtkins> | romeo_: Just stripped whitespace, or variables shortened as well? There are good prettifiers that'll fix the former. |
| 19:42 | <annevk> | Ms2ger, thanks |
| 19:42 | <annevk> | Ms2ger, not sure what to do with the <var> stuff |
| 19:42 | <annevk> | seems only <var>node</var> is affected atm |
| 19:42 | <Ms2ger> | Oh, duh |
| 19:44 | <smaug____> | romeo_: you should use Firefox 4 |
| 19:45 | <romeo_> | Yeah yeah yeah, but I am not |
| 19:45 | <annevk> | romeo_, you can use whatwg.org/C |
| 19:46 | Ms2ger | crosses fingers |
| 19:46 | <annevk> | Ms2ger, heh, I actually think we should fix that in Anolis |
| 19:46 | <annevk> | I don't think that should be in W3C compat |
| 19:46 | <annevk> | but this works |
| 19:47 | <Ms2ger> | Hmm? |
| 19:47 | <annevk> | that <var> can be an xref |
| 19:47 | <annevk> | I don't think you ever want that |
| 19:47 | <annevk> | and if you do, just wrap a span around it |
| 19:47 | <Ms2ger> | I suppose that's true |
| 19:48 | <jgraham> | Doesn't HTML5 use that? |
| 19:48 | <jgraham> | Or maybe I am confused |
| 19:49 | <Ms2ger> | Bah, it does |
| 19:50 | <annevk> | it does? |
| 19:50 | <annevk> | weird |
| 19:50 | <Ms2ger> | <var>the script block's type</var> |
| 19:50 | <gsnedders> | annevk: It does. It was on list of elements Hixie wanted to xref, and thought he could justify semantically. |
| 19:50 | <annevk> | grr |
| 19:50 | <smaug____> | jgraham: do you happen to know if it is a known problem that there are some painting/rendering problems when scrolling html5 draft Opera/linux |
| 19:51 | <jgraham> | I am not sure |
| 19:51 | <Ms2ger> | jgraham, surely you can get an HTMLImageElement regardless of the format? |
| 19:51 | <gsnedders> | smaug____: With the comment box, or anything else? |
| 19:51 | <jgraham> | Ms2ger: How? |
| 19:52 | <Ms2ger> | Uh |
| 19:52 | <smaug____> | gsnedders: with the comment box |
| 19:52 | <Ms2ger> | createElement("img") |
| 19:52 | <Ms2ger> | img.src = "foo.mng" |
| 19:52 | <smaug____> | gsnedders: sounds like a known bug then |
| 19:52 | <annevk> | jgraham, i = new Image() |
| 19:53 | <Philip`> | Just check for .complete = true, I guess |
| 19:54 | <annevk> | not just that anymore |
| 19:54 | <annevk> | checking width/height might be more reliable |
| 19:55 | <gsnedders> | smaug____: Known (at least insofar as QA have seen it before, whether we have a bug report for it is a separate question). :P |
| 19:57 | <jgraham> | Ms2ger: OK, I guess you need to check that width !=0 and height != 0 |
| 19:57 | <annevk> | but assuming SVG works in <img> is fine imo |
| 19:58 | <annevk> | no need to make things complex now for something everyone will have in one/two years |
| 19:58 | <jgraham> | I note my original email doesn't mention how you decide if the image format is supported |
| 19:58 | <jgraham> | annevk: I disagree |
| 19:59 | <jgraham> | The spec is very careful not to require SVG support |
| 19:59 | <annevk> | yeah well, impls support it so we might as well test it to some extent |
| 20:00 | <annevk> | the spec does not require support for ECMAScript either |
| 20:00 | <jgraham> | This is true |
| 20:00 | <annevk> | http://www.w3.org/2010/09/13-ab-minutes -- W3C Member-only (also interesting) |
| 20:01 | <jgraham> | There has to be a line somewhere |
| 20:01 | Philip` | wonders if SVG drawImage should be required to be high-quality vector rendering, or if it's acceptable to flatten it to a bitmap and then render the bitmap onto the canvas |
| 20:01 | <jgraham> | It is not clear to me why SVG is in but PDF is out, for example |
| 20:01 | Philip` | also wonders what should happen with SVG drawImage and security |
| 20:02 | <Ms2ger> | Taint it |
| 20:02 | <Ms2ger> | At least, I understand that's what Mozilla is doing |
| 20:02 | <aho> | well, you would flatten it to the size you're drawing it at |
| 20:02 | <jgraham> | Whereas at least png and gif and ECMAScript have a long history of being supported |
| 20:03 | <Philip`> | Ms2ger: In all cases, even if it's just dumb shapes and colours? |
| 20:04 | <aho> | well, you explictly ask for a bitmap representation :> |
| 20:04 | <Ms2ger> | Your dumb shapes and colours might be my trade secrets |
| 20:04 | <Philip`> | Ms2ger: If that was a problem, it would be equally a problem for drawImage of PNGs and JPEGs |
| 20:05 | <Ms2ger> | I should not that I don't know what I'm talking about |
| 20:05 | <Ms2ger> | note, even |
| 20:05 | <Philip`> | (If that's not a problem, SVG is still a problem since a same-origin SVG image can embed different-origin images) |
| 20:05 | <Philip`> | (and <foreignObjects>) |
| 20:06 | <Philip`> | (but I guess it's kind of risky for browsers to allow some SVGs and not others) |
| 20:06 | <Ms2ger> | All SVGs, with an XXX comment in the source to maybe make it more permissive |
| 20:07 | <annevk> | jgraham, PDF does not have native support in all user agents, SVG does |
| 20:08 | <annevk> | jgraham, it seems unlikely PDF will have that |
| 20:08 | <Philip`> | What is "all"? |
| 20:08 | <Ms2ger> | WebKit, Gecko, Trident and Presto |
| 20:08 | <Philip`> | Canvas seems to be supported by some non-desktop browsers too, which may be less supportive of SVG |
| 20:08 | <annevk> | the usual suspects, indeed |
| 20:09 | <annevk> | so they note n/a for the test |
| 20:10 | <jgraham> | annevk: Right. With the proviso that we implement a N/A state that checks a precondtion, I am happy to have this kind of test |
| 20:10 | <jgraham> | I will even implment support in jsharness.js |
| 20:10 | <annevk> | no |
| 20:10 | <annevk> | no |
| 20:10 | <jgraham> | No? |
| 20:10 | <annevk> | it can just fail in their impl and they can mark it as n/a |
| 20:10 | <jgraham> | No |
| 20:11 | <annevk> | no need for tricks |
| 20:11 | <jgraham> | That is a trick |
| 20:11 | <jgraham> | Manually changing results |
| 20:11 | <annevk> | it's avoiding needless complexity |
| 20:11 | <annevk> | we're not having pre-conditions for all tests that use ECMAScript or XML either |
| 20:11 | <annevk> | it'd be insane |
| 20:12 | <jgraham> | Well for ECMAScript it is clearly insane |
| 20:12 | <jgraham> | We already had explicit agreement that we could assume a mostly-functional ECMSScript engine |
| 20:13 | <jgraham> | *ECMAScript |
| 20:14 | <annevk> | I don't really get the philosophical debate about whether assuming SVG support matters |
| 20:14 | <annevk> | tests make all kinds of assumptions all the time about other specs |
| 20:14 | <annevk> | introducing complexity for that is not at all worth it |
| 20:14 | <annevk> | especially if the top-five browsers plan on supporting it anyway |
| 20:32 | <jgraham> | annevk: Well fel free to disagree with me on the list |
| 20:34 | <jgraham> | I still think the idea of adding tests that have the sole purpose of testing something that is specifically not required by the spec is suspect |
| 20:35 | <jgraham> | Moreso than relying on things that are known to be universially supported for testing other things |
| 20:41 | <annevk> | someone else already disagreed, no? |
| 20:41 | <annevk> | and as Kris said, someone needs to test whether drawing SVG on <canvas> works, as the combination of specs requires that |
| 20:43 | <jgraham> | It seems to me that it is for the SVG WG to test |
| 20:44 | <jgraham> | Especially as they define(d) that SVG had to work in <img> |
| 20:44 | <Philip`> | I think canvas tests need to test it, because there's lots of non-trivial interactions with canvas behaviour |
| 20:44 | <Philip`> | (security and scaling etc) |
| 20:45 | <annevk> | yeah, and we have a <canvas> test suite which has all the infrastructure already |
| 20:45 | <jgraham> | It's not clear that affects who's responsibility it is |
| 20:46 | <jgraham> | They could after all reuse our infrastructure |
| 20:46 | <Philip`> | My original concern was just whether a non-SVG-supporting browser should be considered a pass or fail |
| 20:46 | <annevk> | really, responsibility is something for people who care about Process to bring up |
| 20:46 | <annevk> | rather than progress |
| 20:47 | <jgraham> | I am mildly concered that we end up with a clear idea of what the scope of the testsuite should be |
| 20:47 | <jgraham> | It is highly unclear that SVG-in-canvas is in scope more than PDF-in-canvas |
| 20:48 | <annevk> | i think this is exactly the kind of thing we should be testing |
| 20:48 | <annevk> | one of the things, as Ian now and then mentions, we really care about is closing the gaps between the various specifications |
| 20:48 | <annevk> | this is one of those things |
| 20:48 | <aho> | i think svg in canvas already works in one browser |
| 20:48 | <aho> | i gues it was opera... mmmh |
| 20:48 | <aho> | +s |
| 20:48 | <annevk> | jgraham, it is not at all highly unclear, geez |
| 20:48 | <jgraham> | If the criteria is "have all four major browsers agreed to implement it" then its now |
| 20:48 | <jgraham> | *not |
| 20:49 | <jgraham> | If the criteria is "what does the spec say" then it is |
| 20:53 | <romeo_> | How do I generate something like whatwg.org/C from source? |
| 20:53 | <romeo_> | The svn repo seems to miss a script for that job |
| 20:54 | <jgraham> | For the benefit of the logs: anne, seriously if "four engines plan to implement it" is your criterion for "is this in scope", please say so on the list |
| 20:54 | <jgraham> | If you have a different criterion, please say that |
| 20:58 | <aho> | mine is... if it makes sense (i.e. if there is some use case for this) it should work :> |
| 21:01 | <jgraham> | That's a good policy for a QA testsuite. Not so much for a conformance testsuite for a particular specification |
| 21:02 | gsnedders | would add, "and if it doesn't make sense, it shouldn't crash" |
| 21:02 | <Hixie> | is there harm having more than necessary in the test suite? |
| 21:02 | <gsnedders> | (because someone will do it anyway) |
| 21:02 | <jgraham> | Hixie: It depends what the additional tests are and why they were submitted |
| 21:02 | <gsnedders> | Hixie: If a test is failed by every browser because they don't support it because the spec doesn't require it, then is the test useful? |
| 21:03 | <jgraham> | If they are for some proprietry image format only supported by one browser, for example, it doesn't seem good |
| 21:04 | <jgraham> | Also having tests for things like "does this corrupt PNG file work as expected" seem out of place in a HTML testsuite |
| 21:04 | <jgraham> | Although they might be very useful tests |
| 21:06 | <jgraham> | (also, we have to reach two fully interoperable implementations, so the more stuff we have that won't be implemented by all players who have a realistic chance of being one of those implementations, the less the chance of actually achieving that) |
| 21:07 | <Hixie> | sure, i'm assuming we're only testing stuff that everyone agrees should be implemented in the first place :-) |
| 21:07 | <jgraham> | Hixie: Well we're already not, in fact |
| 21:08 | <jgraham> | Since we need to use some video codecs |
| 21:08 | <Philip`> | I already have tests for APNG |
| 21:08 | <jgraham> | But of course we try to work around that |
| 21:08 | <Philip`> | (in canvas) |
| 21:08 | <Philip`> | but they're designed to pass if APNG isn't supported |
| 21:08 | <Hixie> | oh well some tests may need alternatives to test the same thing with different underlying techs |
| 21:08 | <Hixie> | that's still "one test" for the purposes of this discussion |
| 21:11 | <jgraham> | Anyway, I can agree that if everyone of interest plans to implement SVG-in-<img> and we think that we will do a better job of the tests than the SVG WG then having the tests in the HTML testsuite is OK |
| 21:14 | <Hixie> | i don't really care where the tests are so long as they are somewhere; the worst situation is having two groups both saying that the test should be "over there" |
| 21:14 | <Hixie> | also, having the test in both places is harmless |
| 21:16 | <aho> | ye |
| 21:17 | <aho> | svg+webgl also makes lots of sense. if the resolution is high you can then easily use high res textures :> |
| 21:18 | <aho> | svg in canvas doesnt seem to work anywhere yet... just tried it in the latest oepra, minefield and chrome dev :l |
| 21:19 | <Ms2ger> | There's a patch for Gecko |
| 21:19 | <TabAtkins> | <svg> in canvas, or SVG-in-<img> in canvas? |
| 21:19 | <aho> | latter |
| 21:19 | <aho> | img = new Image(); img.src = '...svg' |
| 21:20 | <aho> | i don't even get onload then :l |
| 21:25 | <romeo_> | The spec-splitter.py script is wonderful. It made a multipage version of the spec within 3 seconds. |
| 21:26 | <romeo_> | Is there a script like that available for the WA1.0 spec? |
| 21:27 | romeo_ | is going to study the spec-splitter.py script |
| 21:28 | <Philip`> | The script should work for any version of the spec |
| 21:28 | <Philip`> | (with varying levels of goodness) |
| 21:28 | <jgraham> | (but not for other specs, right?) |
| 21:28 | <Philip`> | It'd work for anything with the same markup structure |
| 21:28 | <romeo_> | Maybe I have an old or obsolete version? |
| 21:29 | <Philip`> | e.g. same number of elements in head, and same nestedness |
| 21:29 | <romeo_> | Here it is! http://html5.googlecode.com/svn/trunk/spec-splitter/ |
| 21:30 | <romeo_> | I don't think it has a WA1 option |
| 22:19 | <Hixie> | ~. |
| 22:20 | <heycam> | Hixie, enter, not space :) |
| 22:21 | <Hixie> | yeah it turns out the connection wasn't dead after all ;-) |
| 22:28 | <jgraham> | But maybe some people just learnt a good ssh trick |
| 22:29 | <jgraham> | (~ at the start of the line is the escape character "." means close the connection) |
| 22:29 | jgraham | only learnt that quite recently and so guesses some people don't know |
| 22:30 | <Hixie> | it's in the ssh man page, so i'm sure most people who use ssh know about it :-) |
| 22:30 | <jgraham> | Umm. Yeah |
| 22:30 | <jgraham> | Man pages aren't noted for being fun light reading |
| 22:31 | <romeo_> | For every man page there should be a human page. |
| 22:36 | <romeo_> | ... with e.g. common examples. |
| 22:53 | <Philip`> | Woah, there's an escape character? |
