| 00:03 | <Hixie> | for the record, i meant "of the implementations" where i said "of the spec" in the e-mail i just sent. oops. |
| 00:04 | <AryehGregor> | This is supposed to alert "3000000000", right? Because that's in range for an unsigned long? Because it doesn't in any browser. data:text/html,<!doctype html><input size=3000000000><script>alert(document.getElementsByTagName("input")[0].size);</script> |
| 00:04 | <AryehGregor> | (at least not that I tested) |
| 00:04 | <Hixie> | what does it alert? |
| 00:05 | <AryehGregor> | 0 in Gecko and WebKit, 2^32 - 1 in Opera. |
| 00:05 | <Hixie> | UAs are allowed to apply arbitrary limits for ridiculous input, maybe they're applying that rule |
| 00:05 | <Hixie> | or maybe they're just not implementing the spec :-) |
| 00:05 | <gsnedders> | Uh, Opera makes it bigger |
| 00:05 | <AryehGregor> | Er, I meant 2^31 - 1. |
| 00:05 | <AryehGregor> | For Opera. |
| 00:06 | <gsnedders> | Ah |
| 00:06 | <AryehGregor> | So it's like it's treating it as signed and capping it. |
| 00:06 | <TabAtkins> | Opera's signing it, yeah. |
| 00:06 | <AryehGregor> | This doesn't seem ridiculous at all, it's within the stated bounds for the type. |
| 00:06 | <AryehGregor> | I'm filing a Firefox bug, but I feel happier filing a bug if other browsers are correct. |
| 00:06 | <Hixie> | a text field big enough for 3000000000 characters is ridiculous. :-) |
| 00:06 | <AryehGregor> | Otherwise we'd want to consider a spec change instead. |
| 00:06 | gsnedders | wonders why it's signed if nothing else |
| 00:07 | <TabAtkins> | Hixie: Not if I'm writing out a large video in binary by hand! |
| 00:07 | <AryehGregor> | I thought the clause is that implementations are allowed to do arbitrary stuff when there are resource limits, not just randomly because they feel like it's ridiculous. |
| 00:07 | <Hixie> | TabAtkins: true, what was i thinking! :-P |
| 00:08 | <AryehGregor> | Gecko and WebKit seem to agree on returning 0 in this case. WebKit is much crazier in other cases, though. |
| 00:08 | AryehGregor | doesn't understand what WebKit is doing at all |
| 00:09 | <gsnedders> | AryehGregor: If they try and allocate a string of that length for the text input, that's 44GiB |
| 00:09 | <AryehGregor> | I think WebKit is just parsing unsigned integers wrong. |
| 00:09 | <gsnedders> | AryehGregor: I guess that's a hardware limitation :) |
| 00:09 | <AryehGregor> | gsnedders, I didn't even add the element to a document. |
| 00:09 | <AryehGregor> | Pure JS test there. |
| 00:09 | <AryehGregor> | Well, my original test didn't. |
| 00:09 | <gsnedders> | (Though presumably this can't be happening, if Opera still goes for 2^31-1 |
| 00:09 | <AryehGregor> | The one I just posted did. |
| 00:45 | <erlehmann> | >0 in Gecko and WebKit, 2^32 - 1 in Opera. |
| 00:45 | <erlehmann> | That's a pretty big implementation difference ;) |
| 00:51 | <erlehmann> | even weirder, for data:text/html,<!doctype html><input size=2147483647><script>alert(document.getElementsByTagName("input")[0].size);</script> webkit creates a tiny input field, hehe. |
| 00:51 | <erlehmann> | with data:text/html,<!doctype html><input size=247483647><script>alert(document.getElementsByTagName("input")[0].size);</script> it is huge in the vertical dimension … |
| 00:52 | <erlehmann> | but apparently at my computer (using epiphany) the webkit canvas ends at about 60% width, hahaha |
| 04:15 | <llrcombs> | http://thedailywtf.com/Articles/Rarely_Just_TRUE_or_FALSE.aspx <-- what, did Alex call HTML5 out on .canPlayType before it even existed? |
| 04:31 | <a-ja> | Hixie: ping |
| 04:31 | <Hixie> | hey |
| 04:33 | <a-ja> | 1 sec....c/p issue...brb |
| 04:34 | <llrcombs> | am I the only one that's just a little annoyed by the .canPlayType situation? |
| 04:34 | <a-ja> | https://bugzilla.mozilla.org/show_bug.cgi?id=561636#c37 |
| 04:34 | <a-ja> | Hixie: ^^^ |
| 04:34 | <llrcombs> | personally, I preface !! every time I call it |
| 04:35 | <Hixie> | llrcombs: what canPlayType situation? |
| 04:35 | <Hixie> | a-ja: looking |
| 04:35 | <Hixie> | a-ja: not sure what i should be looking at |
| 04:35 | <Hixie> | a-ja: what's the question? |
| 04:36 | <llrcombs> | it returns: "", "maybe", or "probably" |
| 04:36 | <llrcombs> | gimmie true/false |
| 04:36 | <Hixie> | llrcombs: we don't know true/false. |
| 04:36 | <Hixie> | llrcombs: we know "definitely not", "i don't know", and "try me" |
| 04:36 | <a-ja> | example output in spec implies "required", but it's not coded that way |
| 04:37 | <Hixie> | a-ja: how does it imply that? |
| 04:37 | <llrcombs> | so stop not knowing and tell me "try me" or "definitely not" |
| 04:37 | <a-ja> | 2nd line: You cannot complete this form until the field is correct. |
| 04:38 | <llrcombs> | or go for true/false/filenotfound |
| 04:39 | <a-ja> | Hixie: basically....should title be shown when there's a pattern, or validation error msg, or both? |
| 04:40 | <Hixie> | llrcombs: unfortunately that's not how media frameworks work |
| 04:40 | <llrcombs> | well... :( |
| 04:40 | <Hixie> | llrcombs: and changing those is somewhat out of scope for the spec |
| 04:40 | <Hixie> | llrcombs: but if you can get them all to change, i'll be glad to change the spec to match |
| 04:41 | llrcombs | has found that, in practice, every time I've used !!document.createElement("video").canPlayType("video/video_format") and it's returned true, the format's been playable in my browser |
| 04:43 | <Hixie> | a-ja: i've tried to clarify the examine a little, but not sure how to make it much better |
| 04:43 | <Hixie> | a-ja: i'm no UI expert :-) |
| 04:43 | <Hixie> | a-ja: title="" with pattern="" can be shown whenever the user agent thinks the user needs to know about the pattern |
| 04:44 | <Hixie> | llrcombs: the ideal situation is for us to not need canPlayType at all |
| 04:44 | <a-ja> | k....will have a looksee |
| 04:44 | <llrcombs> | Hixie: agreed |
| 04:44 | <llrcombs> | but Apple/Google/Mozilla will never agree to support every format people use |
| 04:44 | <llrcombs> | and M$ won't even agree with itself |
| 04:44 | <Hixie> | Google and Mozilla have agreed on a format |
| 04:44 | <Hixie> | now go convince apple :-) |
| 04:45 | <llrcombs> | what, WebM? |
| 04:45 | llrcombs | puts another bug in RADAR |
| 04:46 | <jamesr> | what i want is document.createElement("img").canDisplayType("image/apng") |
| 04:46 | <llrcombs> | lol |
| 04:47 | <llrcombs> | if a browser can't display a PNG, it's made of fail |
| 04:47 | <llrcombs> | same if it returns maybe |
| 04:47 | <roc> | APNG |
| 04:47 | <jamesr> | APNG is less common |
| 04:47 | <llrcombs> | ohwait |
| 04:47 | <llrcombs> | you mean APNG |
| 04:47 | <jamesr> | the idea is still retarded :P |
| 04:47 | <llrcombs> | my bad |
| 04:47 | <llrcombs> | not strictly |
| 04:47 | <llrcombs> | that's actually valid |
| 04:48 | <jamesr> | it's strictly valid the same way as canPlayType, but it's still pretty dumb |
| 04:48 | <llrcombs> | animated PNG, right? |
| 04:48 | llrcombs | is getting a non-bouncing beachball @ http://en.wikipedia.org/wiki/File:Animated_PNG_example_bouncing_beach_ball.png |
| 04:48 | llrcombs | files a WebKit bug |
| 04:48 | <llrcombs> | PNG > GIF, so I wants APNG support |
| 04:51 | <jamesr> | the point is nobody would do that. they just figure out what image format works everywhere and use that |
| 04:51 | <jamesr> | APNG doesn't work everywhere so nobody uses it |
| 04:51 | <jamesr> | video should be the same way |
| 04:52 | <llrcombs> | but no one format works everywhere |
| 04:52 | <llrcombs> | so we have workarounds |
| 04:52 | <llrcombs> | remember when IE borked some image formats? |
| 04:52 | <jamesr> | and the solution is to get apple/microsoft to support webm |
| 04:52 | llrcombs | just filed a QuickTime bug report |
| 04:52 | <llrcombs> | for WebM support |
| 04:53 | <llrcombs> | as for MS, well... |
| 04:53 | <llrcombs> | they're so far from standard as-is |
| 04:53 | <llrcombs> | what formats does <video> support for IE9? |
| 04:57 | <llrcombs> | ahh |
| 04:57 | <llrcombs> | H.264-only |
| 04:57 | <llrcombs> | a good format in itself, but closed! |
| 04:58 | <llrcombs> | maybe apple will pull out some magic dust and open-sourcify H264 to try to compete with Google |
| 05:01 | <a-ja> | iirc, ms has said they'll support webm in ie9 *IF* you already have the codec on your system |
| 05:02 | <llrcombs> | same way Safari works |
| 05:02 | <llrcombs> | it supports QT's builtin codecs |
| 05:02 | <llrcombs> | and then whatever you add (I support FLV, AVI, and OGG) |
| 05:15 | <a-ja> | Hixie: guess that change clarifies a wee bit more. I see that the pattern/title being displayed is a may rather than a should or must in any event |
| 07:40 | <annevk5> | so is per 2.2 in http://en.wikipedia.org/wiki/Strawman#Reasoning Sam Ruby's email also a straw man? |
| 07:41 | <annevk5> | communication would work better :/ |
| 08:41 | <annevk5> | aah |
| 08:42 | <annevk5> | TextWrangler does not recognize svn move |
| 08:46 | <Hixie> | man, i keep having problems with this canvas e-mail i'm writing |
| 08:46 | <Hixie> | i've already lost the entire e-mail once |
| 09:22 | <B1OS> | is it possible to 'send' an audio track to a html5 player without placing it inside the <audio> tags? |
| 09:22 | <B1OS> | so i would click an mp3 file thats outside of the <audio> tag and it will play, somehow, in the player... |
| 09:23 | <annevk5> | if I understand what you mean that'd be possible using some scripting |
| 09:27 | <B1OS> | i'll give it a try and post back |
| 09:29 | <B1OS> | because the mp3 files will be a directory listing using php |
| 09:31 | <B1OS> | like this: http://tjn.no-ip.org/music1 and there will just be a single html5 player in the header |
| 10:04 | <annevk5> | so wtf |
| 10:04 | <annevk5> | is greg proposing changing the framing or not? |
| 10:05 | <annevk5> | his message to me made it seem like he was not, but the last paragraph in his latest makes it seem like he does |
| 10:05 | <annevk5> | blargh |
| 10:07 | <jgraham> | "Can you post a summary?" "I will refrain from reiterating the detail here" |
| 10:07 | <jgraham> | Huh? |
| 10:07 | <jgraham> | Summary implies "no detail" |
| 12:44 | <annevk5> | euh, so now issues can be opened because a bug has not been replied to? |
| 12:47 | annevk5 | asks |
| 12:52 | <Lachy> | I recall people attempting to do that with some issues before. It sucks that some people lack patience, but they still shouldn't be able to do that |
| 15:36 | Lachy | wonders why an editorial issue like issue-109 is considered significant enough for even having a poll |
| 15:37 | <Workshiva> | Because it's less work for the chairs to put up a poll than handle the fallout of refusing? :P |
| 15:41 | <jgraham> | Maybe we should have a competition to try and get the most trivial issue possible taken to a poll |
| 15:41 | <jgraham> | Although I think Julian already won |
| 15:41 | <Workshiva> | Tabs vs spaces for indentation in markup examples, go |
| 15:41 | <Lachy> | I don't know, the wording of a heading seems significantly more trivial than the ascii reference. |
| 15:42 | <jgraham> | Lachy: Well I thought that, but > 0 people will read the heading title |
| 15:42 | jgraham | wonders if he can make it an Issue that Hixie edits in emacs not Vi |
| 15:42 | <Lachy> | Hixie does edit in Emacs already. |
| 15:43 | <jgraham> | Yes, that is the Issue |
| 15:43 | <Workshiva> | He should alternate between editors |
| 15:43 | <Workshiva> | That's the only way to reach consensus |
| 15:46 | <Philip`> | History shows that it's okay to rely on a single editor, as long as there's a token appearance from another editor from another organisation to assuage people's fears that the first editor will monopolise the process |
| 15:46 | <Philip`> | then you can drop the second editor after a couple of years when nobody cares any more |
| 15:47 | <Philip`> | So I suggest the issue could be resolved by just using Vi for a little bit and then switching back to Emacs for all the serious work |
| 15:48 | <jgraham> | It's not possible to use vi for a little bit; it takes a lot of concerted effort just wo work out how to insert a character |
| 15:50 | Philip` | wonders if it's bad to use Opera's bug wizard when he's too lazy to figure out how to use the proper BTS entry form |
| 15:50 | <jgraham> | Philip`: Not really, unless you want to do something you can't do with the bug wizard like CC someone |
| 15:51 | <gsnedders> | Philip`: I'm not really sure it's less of a mess |
| 18:44 | <annevk5> | we appreciate all your bug entries :) |
| 18:47 | <annevk5> | time to land some xhr tests and make everyone cry |
| 19:04 | <annevk5> | ah damn |
| 19:04 | <annevk5> | Philip`, how did you configure hg again to push? |
| 19:04 | <annevk5> | it was something simple but I forgot :/ |
| 19:05 | <Philip`> | annevk5: Do you mean configuring the username/password? |
| 19:05 | <annevk5> | yeah |
| 19:05 | <annevk5> | not the ui bit, but the actual bit that matters |
| 19:05 | <Philip`> | [auth] |
| 19:05 | <Philip`> | w3c.prefix = dvcs.w3.org/hg/ |
| 19:05 | <Philip`> | w3c.username = ... |
| 19:05 | <Philip`> | w3c.password = ... |
| 19:05 | <Philip`> | That bit? |
| 19:05 | <annevk5> | yay |
| 19:08 | <annevk5> | Philip`, that should get read from repository/.hg/hgrc as well no? |
| 19:09 | <annevk5> | aah, https |
| 19:09 | <annevk5> | doh |
| 19:11 | <Philip`> | I expect it should |
| 19:12 | <annevk5> | how do you set a default repository to pull and push from? |
| 19:13 | <Philip`> | [paths] |
| 19:13 | <Philip`> | default = https://dvcs.w3.org/hg/html/ |
| 19:13 | <Philip`> | in the .hg/hgrc |
| 19:13 | <annevk5> | i should write a blog post on this too |
| 19:13 | <annevk5> | this is the second time I needed this info and couldn't find it in the IRC logs |
| 19:14 | <jgraham> | You heard of Google, right :) |
| 19:14 | <annevk5> | ah yeah, and then I can put the [auth] / [ui] stuff in a higher level .hgrc |
| 19:14 | <annevk5> | jgraham, didn't help |
| 19:24 | <henrikbjorn> | Hey i have a question about header (the page header) and multiple nav elements. I have two main navigation parts on my site. One being a sections like "News, Gallery" etc. and the other being a based on the choice of section "submenu". would a <header><nav /><nav /></header> be the solution or a <nav> with to <ul /> in it ? |
| 19:32 | <TabAtkins> | Man, people shouldn't use the world "halting problem" for things that aren't the halting problem. It confuses my brain. |
| 19:33 | <TabAtkins> | henrikbjorn: Don't overthink it. Just wrap your navigations blocks in <nav>, and place them where it feels appropriate. |
| 19:33 | <henrikbjorn> | so one nav with two lists? |
| 19:34 | <henrikbjorn> | cant seem to understand if its valid or not by the spec |
| 19:34 | <TabAtkins> | One <nav> with two <ul>s is fine, or two <nav>s with their own <ul>. Whichever. |
| 19:35 | <henrikbjorn> | okay thanks :) |
| 20:45 | <TabAtkins> | In Paper's email on the datatime <input> thread, anyone know what is meant by "Christian/Hebrew (Sun–Sat, no splits), US (Sun–Sat, splits)". I don't know what a "split" is or how the US and Christain calendars differ. |
| 20:50 | <Philip`> | http://en.wikipedia.org/wiki/Seven-day_week#Week_numbering |
| 20:50 | <aho> | TabAtkins, in some calendars the weeks starts with so - in others with mo |
| 20:50 | <Philip`> | TabAtkins: Sounds like it might mean whether the first week is split across two years |
| 20:50 | <Philip`> | or whether the first week always starts on 1 Jan |
| 20:50 | <aho> | eh... su :> |
| 20:51 | <TabAtkins> | aho: Nah, that's referring to a different distinction (which he makes between another calendar system running Mon-Sun). |
| 20:51 | <TabAtkins> | Philip`: That makes sense. |
| 20:51 | <aho> | i'd like a more logical calendar... 13 months with 4 weeks each... and one extra day (or 2) :> |
| 20:51 | <TabAtkins> | No, 13 months of 28 days is horrible. |
| 20:51 | <TabAtkins> | You want 12 months of 30 days, with 5/6 intercalary days. |
| 20:51 | <aho> | why? would you miss quarters that much? |
| 20:52 | <TabAtkins> | No, I just hate 13 and 28. They're very bad numbers for mental math. |
| 20:52 | <aho> | well... the first of each months is a monday |
| 20:52 | <TabAtkins> | tantek has a great calendar scheme up on his site. |
| 20:52 | <aho> | and so on |
| 20:52 | <aho> | would make things a lot easier |
| 20:52 | <TabAtkins> | Oh yeah, in the 12/30 calendar, each week is 5 days. |
| 20:52 | <Philip`> | Regular calendar systems are boring, because your birthday would be on the same day every year |
| 20:52 | <TabAtkins> | So every month starts on Monday too. |
| 20:53 | <Philip`> | so some people would be stuck on weekdays and some on weekends |
| 20:53 | <aho> | 3 days work, 2 days weekend... sound great to me |
| 20:53 | <aho> | ;D |
| 20:53 | <Philip`> | which wouldn't be fair |
| 20:54 | <TabAtkins> | aho: It's not *quite* as good as it sounds, since you have more weeks. But that's is a good bit more weekend. To get close to the current weekend amount, you can do alternating 3/4 day workweeks. |
| 20:55 | <AryehGregor> | Philip`, just adopt the convention that birthdays are all celebrated on some fixed day of the week. Like Monday holidays. |
| 20:57 | <aho> | while we're at it... something regular for time would be nice too (no timezones) :> |
| 20:58 | <aho> | sorta like swatch beats |
| 20:58 | <aho> | i really liked that idea |
| 20:58 | <TabAtkins> | Check out www.xanthir.com |
| 20:58 | TabAtkins | loves beats. |
| 20:58 | TabAtkins | just wishes they'd standardized on GMT, rather than GMT+1. |
| 20:59 | <TabAtkins> | <3 millidays. |
| 20:59 | <aho> | heh |
| 20:59 | <aho> | (i see ripped mario sprites...) :> |
| 21:00 | <Philip`> | We need to get a group of people to agree on how we should redesign dates and times and timezones in an ideal world, and then get one member of the group to become a world dictator who can force everyone to adopt the new system |
| 21:00 | <Philip`> | Otherwise it's just going to be a waste of effort :-( |
| 21:00 | <TabAtkins> | aho: You should see the current time in beats, written on the hill. |
| 21:00 | <TabAtkins> | And yes, ripped mario sprites. |
| 21:00 | <aho> | ISO 8601 is nice http://en.wikipedia.org/wiki/ISO_8601 |
| 21:00 | <TabAtkins> | Philip`: Agreed. |
| 21:00 | <aho> | (the date formatting part) |
| 21:00 | <aho> | @879 it is :) |
| 21:02 | <TabAtkins> | Honestly, who needs more than 90s resolution in the base value? Not me. And decimal beats are just fine for when you do. |
| 21:02 | <TabAtkins> | Also: no leap seconds or other ridiculousness. The length of a beat naturally changes with the length of a day. |
| 21:03 | <Philip`> | Variable-length time units sound fun when you're trying to do any kind of precise measurements of anything |
| 21:03 | <TabAtkins> | That's what seconds are for. |
| 21:03 | <Philip`> | where by "fun" I mean "terrible" |
| 21:04 | <Philip`> | So we've got to have two independent time units, and remember how to translate between them? |
| 21:04 | <TabAtkins> | Yes, though switching between them should rarely be necessary in casual contexts. |
| 21:04 | <Philip`> | where the translation depends on continually-updated observations of the earth's rotation so you can't even write a static algorithm for it? |
| 21:05 | <TabAtkins> | And in any case, we still keep the timestamp, which requires similar translation already. |
| 21:06 | <Philip`> | People are lazy and will write software that uses the same units internally as it uses on its UI, and if the human-readable UI units are something crazy then that'll infect the internals of the software too and cause all sorts of bugs |
| 21:07 | <TabAtkins> | Might as well make the conversion somewhat harder so people will just store dates using timestamps anyway. Yay! |
| 21:08 | <TabAtkins> | Converting from timestamp -> current human-readable time is roughly equivalent to converting from timestamp -> millidays. |
| 21:11 | <TabAtkins> | (In other words, pretending that we currently keep time with seconds is silly in the first place. We keep time with a mixed-radix system based on both days (and composites of such) and seconds (and composites of such). Converting from seconds-based time to days-and-seconds-based time is approximately equally difficult to just converting from seconds-based to days-based.) |
| 21:15 | <TabAtkins> | (At least, for computers. For humans, 1 beat is about 100 seconds (roughly 15% off, but close enough for casual work). That's easier to do mentally than juggling multiples of 60 and 24.) |
| 23:24 | <AryehGregor> | IE9PP4 is a real trainwreck on my reflection tests, compared to any other browser I've tested. Except maybe Opera, but Opera doesn't work on it anymore. (Any progress, gsnedders?) |
| 23:24 | <AryehGregor> | (I mean, Opera doesn't work on the test, it gives some crazy JS error) |
| 23:24 | <TabAtkins> | pp4 is doing plenty of weird js things, I've found. |
| 23:24 | <TabAtkins> | apparently [].indexOf doesn't exist in it at the moment. |
| 23:25 | <AryehGregor> | Is it easier to write cross-browser JS than with previous IE versions, though? |
| 23:25 | <AryehGregor> | My tests don't even run in IE8. |
| 23:25 | <AryehGregor> | At least they complete in IE9PP4. |
| 23:25 | <TabAtkins> | No clue yet. I don't write cross-browser js. |
| 23:25 | <AryehGregor> | You just use jQuery? |
| 23:25 | <TabAtkins> | I either target decent browsers, or write jquery. |
| 23:25 | <AryehGregor> | Heh. |
| 23:26 | <AryehGregor> | Looks like IE9 will still be an odd duck, but way less than ever before. |
| 23:26 | <gsnedders> | AryehGregor: Didn't have time to look |
| 23:26 | <AryehGregor> | I doubt I'll ever have the patience to track down all its errors on my reflection tests, though. |
| 23:27 | <gsnedders> | TabAtkins: Uh, if that was true nothing would work |
| 23:27 | <gsnedders> | TabAtkins: Like, jQuery would be totally b0rked |
| 23:27 | <AryehGregor> | I'm doing it for WebKit and Gecko, but IE is way more failures. |
| 23:27 | <TabAtkins> | gsnedders: Hey, that's the error it gives me when I try to run my manufactoria emulator. |
| 23:27 | TabAtkins | goes to duplicate it, to be sure. |
| 23:27 | <gsnedders> | TabAtkins: Odd |
| 23:30 | <TabAtkins> | ... |
| 23:30 | <TabAtkins> | I must be crazy. It works just fine now. |
| 23:31 | <gsnedders> | TabAtkins: Totally. |
| 23:31 | <TabAtkins> | Still wont' animate SVG, but at least the presence of an animation no longer does weird scaling things. |
