| 04:47 | <MikeSmith> | http://item.rakuten.co.jp/icefield/c/0000000218/ |
| 04:49 | <MikeSmith> | http://item.rakuten.co.jp/icefield/10011112/ |
| 04:51 | <MikeSmith> | http://item.rakuten.co.jp/icefield/10010977/ |
| 04:52 | <wirepair> | ha |
| 04:52 | <wirepair> | there's now a brand called html huh |
| 04:52 | <wirepair> | http://item.rakuten.co.jp/icefield/10008676/ is good |
| 05:14 | <MikeSmith> | wirepair: I got to buy some of this stuff |
| 05:16 | <wirepair> | hehe |
| 05:16 | <MikeSmith> | dude must have a shop somewhere in town |
| 05:16 | <wirepair> | i swear though rakuten gives me a seizure |
| 05:16 | <wirepair> | worst website ever |
| 05:17 | <wirepair> | i'm guessing harajuku or shibuya if they have a shop hehe |
| 05:19 | <MikeSmith> | 渋谷PARCO パート1 ６F html渋谷店 |
| 05:24 | <wirepair> | my terms ascii only ;< |
| 05:25 | <wirepair> | btw spoke with Abarth |
| 05:25 | <MikeSmith> | great |
| 05:25 | <wirepair> | looks like the layoutTestController they use is somewhat webkit specific |
| 05:25 | <MikeSmith> | yeah |
| 05:25 | <wirepair> | so i'm just going to incorporate those testcases into my own system |
| 05:25 | <MikeSmith> | knowed that already |
| 05:26 | <wirepair> | which i have working on ie mobile, ie8, opera 9 something, chrome, and firefox |
| 05:26 | <MikeSmith> | that's one reason I suggested talking with Slyvain on #testing |
| 05:26 | <wirepair> | yeah |
| 05:26 | <MikeSmith> | syp |
| 05:26 | <MikeSmith> | he's thought this stuff through quite a bit already |
| 05:26 | <wirepair> | yeah that irc is kinda quiet ;> |
| 05:27 | <wirepair> | i'll ping him there and see when i get a response |
| 05:27 | <MikeSmith> | yeah, if you ping syp there during CET hours, he will likely responde |
| 05:27 | <MikeSmith> | oh man, check out this hipster: http://image.space.rakuten.co.jp/lg01/48/0000990248/97/img760e01c1zikazj.jpeg |
| 05:27 | <MikeSmith> | that is so going to be me after I stop by Parco tonight |
| 05:27 | <wirepair> | haha |
| 05:28 | <wirepair> | Awesome heh |
| 05:28 | <wirepair> | so html isn't copyright'd then |
| 05:28 | <wirepair> | ? |
| 05:29 | <ako> | txt also isn't copyrighted |
| 05:29 | <ako> | *shrug* |
| 05:30 | <wirepair> | well java is isn't it? |
| 05:30 | <wirepair> | even though it could refer to coffee |
| 05:30 | <ako> | "java" the name is trademarked |
| 05:30 | <wirepair> | yeah sorry trademarked, not copyright |
| 05:30 | <ako> | (different thing) |
| 05:32 | <ako> | using "html" as brand name is about as stupid as using "the" as brand name |
| 05:32 | <ako> | it's ungoogleable |
| 05:32 | <ako> | and therefore completely useless |
| 05:32 | <ako> | :> |
| 05:32 | <wirepair> | it won MikeSmith over did it not? ;> |
| 05:35 | <ako> | well, whatever... buh bai :> |
| 05:43 | <MikeSmith> | man, I'm going to grow broke buying HTML goods |
| 07:57 | <MikeSmith> | the single-page version of the HTML5 spec is inching up on 5MB |
| 07:57 | <MikeSmith> | currently 4.8MB |
| 07:59 | <nessy> | ouch |
| 08:00 | <zcorpan_> | MikeSmith: what about complete.html? |
| 09:03 | <MikeSmith> | zcorpan_: dunno, haven't tried it yet |
| 09:04 | <MikeSmith> | I just noticed how big the HTML5 spec had gotten because I needed to sync up the v.nu copy to the latest upstream |
| 09:04 | <MikeSmith> | btw |
| 09:04 | <MikeSmith> | http://www.w3.org/Bugs/Public/show_bug.cgi?id=9595 |
| 09:04 | <MikeSmith> | [[ |
| 09:04 | <MikeSmith> | I just bumped into a wierd problem in all IE browsers and it turned out to be |
| 09:04 | <MikeSmith> | an HTML comment just before the DOCTYPE. The validator did not recognize it as |
| 09:04 | <MikeSmith> | an error but all IE browsers failed to properly assign css proeprties etc. |
| 09:04 | <MikeSmith> | ]] |
| 09:05 | <MikeSmith> | that's not a validator bug of course |
| 09:05 | <MikeSmith> | but is what that says about IE behavior true? |
| 09:05 | <zcorpan_> | yep, at least up to and including ie8 |
| 09:05 | <zcorpan_> | dunno about ie9 |
| 09:05 | <MikeSmith> | ah, OK |
| 09:05 | MikeSmith | didn't know |
| 09:10 | <MikeSmith> | yee hah |
| 09:10 | <MikeSmith> | complete spec is 5.1MB |
| 09:56 | <annevk> | not a huge difference the splitting out has made then... |
| 10:00 | <MikeSmith> | annevk: I guess not (as far as the file size at least) |
| 10:02 | <zcorpan_> | if removing features from the spec has the consequence of the removed features getting implemented in browsers, then it seems like a good thing |
| 10:12 | <MikeSmith> | zcorpan_: heh |
| 10:51 | <MikeSmith> | http://validator.nu/?doc=http%3A%2F%2Fbikramgreenvalley.com%2FMain_Content%2FFAQ.php |
| 10:51 | <MikeSmith> | error: ILLEGAL_CHARACTER in QUERY |
| 10:51 | <MikeSmith> | src="/Parts/GenerateFiles/MakeFiles.php?name=ToggleDL&type=text/javascript&path=Parts/ToggleDL&js[]=ToggleDL" |
| 10:51 | <MikeSmith> | what's the illegal character? |
| 10:51 | <MikeSmith> | brackets? |
| 10:52 | <MikeSmith> | solidus? |
| 10:56 | <hasather> | MikeSmith: brackets |
| 10:58 | <MikeSmith> | hasather: thanks |
| 11:01 | <craiggles> | I'm playing around with HTML5 canvas element and I'm now thinking of making a game, for a game like super mario bros - http://ian-albert.com/misc/gamemaps/super_mario_bros/full/mario-1-1.gif - would you suggest me using a 2D array for the map? |
| 11:09 | <hsivonen> | MikeSmith: IIRC, if you copy a URL with brackets from the URL bar, the brackets don't get percent escaped |
| 11:09 | <hsivonen> | MikeSmith: I'm not sure if this should be treated as a Firefox bug or as an RFC bug |
| 11:45 | <erlehmann> | hsivonen, i seem to remember that brackets were perfectly legal in URIs, aren't they ? |
| 11:50 | <hasather> | erlehmann: maybe it's only allowed around IP literals even, dunno |
| 11:50 | <erlehmann> | thx. |
| 12:01 | <MikeSmithX> | hsivonen: I see |
| 13:07 | <annevk> | "needed" could maybe be an alternative, but it kind of sucks |
| 13:07 | <annevk> | and might very well have similar issues |
| 13:09 | <annevk> | funny |
| 13:09 | <annevk> | http://blogs.msdn.com/b/ie/archive/2010/06/11/ie-test-center-updates.aspx |
| 13:10 | <annevk> | they updated the tests in a way that IE still passes and can continue to have their silly no whitespace nodes in the DOM bug |
| 13:10 | <annevk> | (foreign content parsing tests) |
| 13:11 | <annevk> | without even acknowledging the problem |
| 13:20 | <MikeSmith> | https://lists.webkit.org/pipermail/webkit-dev/2010-June/013215.html |
| 13:20 | <MikeSmith> | from abarth |
| 13:20 | <MikeSmith> | "HTML5 tokenizer landing soon" |
| 13:24 | <MikeSmith> | and https://spreadsheets.google.com/ccc?key=0AppchfQ5mBrEdDFJUW5DOGNsdmtvZkN0ZmIzMjdaT0E&hl=en |
| 13:40 | <hsivonen> | Interesting. Either WebKit just has broader test suite coverage than Gecko or their test suite depends on more invalid stuff. |
| 13:40 | <hsivonen> | IIRC, the HTML5 *tokenizer* regressed fewer tests in Mozilla's test suites |
| 13:42 | <gsnedders> | Looks like just testing more invalid stuff, not random stuff relying upno it |
| 13:42 | <hsivonen> | Also, good for them that they have the luxury of landing the tokenizer and the tree builder independently |
| 13:43 | <hsivonen> | i.e., their pre-existing abstaction between the tokenizer and the tree builder has to be relatively proper |
| 13:50 | <craiggles> | http://pastie.org/1003752 .. I'm having some problems with my code, perhaps I've done something wrong. well basically what is happening is that one circle isn't being filled green (#008000). |
| 14:19 | <craiggles> | is it possible to check if a certain shape is in a specified area of my canvas? i'm looking through the spec atm |
| 14:23 | <annevk> | craiggles, there's isPointInPath |
| 17:14 | <ap> | hi john! hi everyone! |
| 17:14 | <ap> | sorry, wrong channel |
| 17:17 | <AryehGregor> | Hi anyway! |
| 18:26 | <tabatkins> | Man, switching my monitors to vertical is an improvement, but it'll take some getting used to. I'm not used to seeing this much of a website at once. |
| 18:30 | <miketaylr> | it's always kinda freaked me out |
| 18:43 | <cardona507> | vertical is great for one screen as long as I have another that is horizontal |
| 19:02 | <jgraham> | Yeah that seems like a good setup |
| 19:03 | <jgraham> | I need a second big screen :) |
| 19:03 | jgraham | has one big widescreen and one small 4:3 |
| 19:03 | <TabAtkins> | Two large widescreens here. I think I'd like a small 4:3 to keep horizontal too. |
| 19:04 | <gsnedders> | TabAtkins: That I always found to be the real advantage of a 30" over two 24": vetical space. |
| 19:08 | <jgraham> | gsnedders: With one vertical, 2x24 gives you more vertical space than 1x30 |
| 19:08 | <gsnedders> | Yes, that's true |
| 19:08 | <jgraham> | Really big monitors are useful for some things e.g. image processing |
| 19:08 | <jgraham> | For coding not so much |
| 19:08 | <jgraham> | (the basic question is "do I want one application maximised or several side-by-side) |
| 19:09 | <jgraham> | s/)/")/ |
| 19:09 | <gsnedders> | Really big monitors tend to have less good colour reproduction than some, e.g., 24" screens available, so it's a question of which you care about more |
| 19:09 | <jgraham> | Doesn't that just depend on what type of display technology they use? |
| 19:09 | <gsnedders> | Well, yes. But all the 30" screens available aren't as good as some 24". |
| 19:10 | <jgraham> | (which is _never_ever_ specified) |
| 19:10 | <gsnedders> | Also: why is everything so frickin' expensive here? |
| 19:10 | jgraham | wonders if that is true or if you can get insanely expensive 30" screens |
| 19:10 | <jgraham> | gsnedders: High tax rates, high wages, high expectation of cost? |
| 19:10 | <othermaciej> | doh, I forgot to implement the IDL attribute for @hidden |
| 19:12 | <jgraham> | gsnedders: http://www.necdisplay.com/Products/Product/?product=ec17518a-366e-4f09-8173-ca42d439b357 |
| 19:13 | <TabAtkins> | othermaciej, what the hell. |
| 19:13 | gsnedders | hadn't seen NEC announce any 30" screens |
| 19:13 | <othermaciej> | TabAtkins: sorry man, I know you're disappointed |
| 19:13 | <TabAtkins> | ;_; /wrist |
| 19:14 | <othermaciej> | now I won't be able to say it takes 30 minutes to implement |
| 19:14 | <jgraham> | othermaciej: How long does adding an IDL attribute take? I thought you just used some sort of IDL in the source tree? |
| 19:15 | <jgraham> | So it would be a copy+paste? |
| 19:15 | <othermaciej> | jgraham: it will take me probably another 30 minutes, including test case |
| 19:15 | <jgraham> | Well an hour isn't so bad, really :) |
| 19:15 | <othermaciej> | it's actually slightly easier than hooking up a markup attribute to CSS |
| 19:16 | <lifechamp> | Canvas (html5) + SVG libraries VS using SVG directly: theoretically, should there be a big performance difference or no? |
| 19:16 | <TabAtkins> | Depends entirely on how efficient canvas and svg are in the chosen browser. |
| 19:16 | <jgraham> | I guess that depends on the library and what you are doing |
| 19:19 | <lifechamp> | Ok, so answer is no, in some ways. I suppose the one underlying question was whether SVG is hardware supported (whereas a library building SVG atop Canvas) would not be. It seems there is some HW support for SVG, but not widespread: http://en.wikipedia.org/wiki/Scalable_Vector_Graphics#Hardware_rendering |
| 19:19 | <othermaciej> | depends on your library and your use case |
| 19:20 | <lifechamp> | err, sorry, i meant the answer seems to be yes, in some ways, where HW support is used and helpful. |
| 19:30 | <lifechamp> | Although, since WebGL atop Canvas (?) is HW supported, I suppose there can be a HW supported SVG library atop Canvas as well. I'll look for it, but if anyone knows about this, please LMK. |
| 20:06 | <gsnedders> | jgraham: You had some idea of using numpy for computing the average run length, how so? |
| 21:07 | <jgraham> | gsnedders: I hope you were doing something fun earlier to be working now |
| 21:07 | <jgraham> | gsnedders: But I haven't thought about the numpy thing. What does the input look like? |
| 21:09 | <MattCampbell> | html5lib 0.90 isn't handling the </form> end-tag correctly, at least when using simpletree |
| 21:09 | <Hixie> | you sure? |
| 21:09 | <MattCampbell> | given this trivial test document: <form action="http://example.com/blah"; method=post></form><p>test</p> |
| 21:09 | <MattCampbell> | parsing and serializing that moves </form> after </p> |
| 21:11 | <Hixie> | http://james.html5.org/cgi-bin/parsetree/parsetree.py?source=%3Cform+action%3D%22http://example.com/blah%22+method%3Dpost%3E%3C/form%3E%3Cp%3Etest%3C/p%3E |
| 21:11 | <Hixie> | says the <p> is after the <form> |
| 21:11 | <Hixie> | (that's an html5lib dumper) |
| 21:11 | <MattCampbell> | using which tree builder? |
| 21:11 | <jgraham> | Hixie: The code there needs to be updated |
| 21:12 | <jgraham> | MattCampbell: testing... |
| 21:12 | <Hixie> | k |
| 21:12 | <gsnedders> | jgraham: Gah, I'll just ask you tomorrow. I have other things to do now :P |
| 21:15 | <jgraham> | MattCampbell: WFM |
| 21:16 | <jgraham> | At least parsing does |
| 21:16 | <jgraham> | I suppose the serializer could be broken in interesting ways |
| 21:16 | <MattCampbell> | jgraham: Which tree builder are you using? |
| 21:17 | <jgraham> | MattCampbell: SimpleTree (not that it should matter or simpletree should be used for anything serious) |
| 21:17 | <jgraham> | (of course it *could* matter) |
| 21:17 | <MattCampbell> | jgraham: I'm using simpletree too |
| 21:18 | <jgraham> | MattCampbell: You are using the release version, not HEAD |
| 21:18 | <MattCampbell> | jgraham: And the problem is in the parser or the tree builder; I just traversed the parse tree myself |
| 21:18 | <MattCampbell> | 0.90 |
| 21:18 | <jgraham> | MattCampbell: can you run ./parse.py {your input} --tree from a command line and paste the resulst somewhere? |
| 21:19 | <jgraham> | gsnedders: So were you doing something fun? :p |
| 21:19 | <gsnedders> | jgraham: Yeah |
| 21:19 | <MattCampbell> | jgraham: In the examples directory, right? |
| 21:19 | <jgraham> | MattCampbell: Possibly, yeah |
| 21:20 | <gsnedders> | (and I only did around 30 mins of work) |
| 21:20 | <jgraham> | gsnedders: Sometyhing secret? |
| 21:20 | <gsnedders> | jgraham: Of course :) |
| 21:21 | <jgraham> | Wow, intrigue |
| 21:21 | <jgraham> | I will be sure to make up improbable rumors |
| 21:21 | <MattCampbell> | Any preferred pastebin here? |
| 21:21 | <gsnedders> | jgraham: What was it you made up before? |
| 21:21 | gsnedders | tries to remember |
| 21:21 | <gsnedders> | MattCampbell: not really |
| 21:22 | <MattCampbell> | jgraham: http://pastebin.com/dZi1k9qk |
| 21:23 | <gsnedders> | Odd. |
| 21:23 | <jgraham> | Yeah, odd |
| 21:23 | <MattCampbell> | Remember, in the input, the p element is outside of the form |
| 21:23 | <jgraham> | MattCampbell: Can you try and reproduce with the latest hg trunk? |
| 21:23 | <MattCampbell> | Python 2.6, btw |
| 21:24 | <MattCampbell> | OK |
| 21:24 | <MattCampbell> | What's the repo URL? |
| 21:24 | <MattCampbell> | not on google code, is it? |
| 21:24 | <MattCampbell> | oh ok, it is |
| 21:25 | <MattCampbell> | http://html5lib.googlecode.com/hg/python/ |
| 21:25 | <MattCampbell> | right? |
| 21:25 | <gsnedders> | eah |
| 21:25 | <gsnedders> | *yeah |
| 21:26 | <jgraham> | gsnedders: This time at least one of them will involve sailor hats |
| 21:27 | <jgraham> | ;) |
| 21:28 | <MattCampbell> | works in hg trunk |
| 21:28 | <MattCampbell> | I'll use that then |
| 21:29 | <MattCampbell> | Any particular reason for the switch to setuptools? |
| 21:31 | <jgraham> | There was a switch? |
| 21:31 | <gsnedders> | jgraham: hah |
| 21:31 | <jgraham> | I thought we had always been using setuptools? |
| 21:32 | <MattCampbell> | I think 0.90 used plain distutils |
| 21:32 | <jgraham> | (I think the answer is "python packaging makes me feel queasy every time I have to read about it, so we do not necessarily do the best thing") |
| 21:32 | <gsnedders> | I don't think 0.90 did |
| 21:32 | <jgraham> | If it did it was an accident |
| 21:32 | <MattCampbell> | ah, 0.90 used a try-except |
| 21:32 | <gsnedders> | But maybe it did. All I had to do with that release was nagging jgraham until he made it. |
| 21:32 | <MattCampbell> | so it fell back on distutils in the absence of setuptools |
| 21:33 | <jgraham> | I'm not sure we gain much from setuptools tbh so maybe it is enough to just use distutils |
| 21:33 | <jgraham> | But I don't really know what the tradeoffs are |
| 21:33 | <MattCampbell> | And why do I care? setuptools + py2exe is problematic |
| 21:33 | <MattCampbell> | yes, using html5lib in a Windows app |
| 21:33 | <jgraham> | Have fun :) |
| 21:34 | <jgraham> | Does distutils let you specify dependencies |
| 21:34 | <MattCampbell> | No... I guess that's the main advantage of setuptools |
| 21:34 | <MattCampbell> | 0.90 is 5 months old, right? |
| 21:35 | <MattCampbell> | I just noticed the date on setup.py in that release |
| 21:35 | <jgraham> | Already? |
| 21:35 | <MattCampbell> | January 17 |
| 21:35 | <gsnedders> | jgraham: yes, already |
| 21:35 | <gsnedders> | jgraham: Obviously with your busy social life time flies |
| 21:36 | <jgraham> | gsnedders: My busy social life? I'm not th one having secret rendevousz ;) |
| 21:36 | <jgraham> | sp? |
| 21:36 | <jgraham> | Yeah I guess it is that old |
| 21:36 | <jgraham> | We should make a new release soon then |
| 21:36 | <gsnedders> | rendezvous |
| 21:38 | <MattCampbell> | Are there any current release-critical bugs? |
| 21:38 | <jgraham> | Well we need to fix some issues that prevent all the tests passing |
| 21:39 | <jgraham> | Nothing too serious, just little spec changes |
| 21:39 | <gsnedders> | Did you get all the table fostering tests passing? |
| 21:39 | <jgraham> | Yeah I think they pass |
| 21:39 | <gsnedders> | I had a look at that a while ago, and it wasn't easy with the current code which is quite different to the spec |
| 21:39 | <gsnedders> | (current when I looked) |
| 21:39 | <jgraham> | It is just a couple of things with <button> iirc |
| 21:39 | <gsnedders> | (this was last year) |
| 21:48 | <hober> | html5lib peeps, have you seen http://github.com/aredridel/html5 ? |
| 21:49 | <hober> | ( impl. of html5 parsing algorithm in js for node.js ) |
| 21:52 | <jgraham> | Seems to be using a way out of date testsuite |
| 21:52 | <jgraham> | Dunno quite how it managed that |
| 21:53 | <jgraham> | hober: Or, in long, no |
| 21:55 | <jgraham> | """Excessive blink rates, such as those over 3 times a second can cause epileptic seizures in some users. To address this issue the user will configure the operating system blink rate for carets or selections to one which does not induce a seizure. User agents must support this method so that authors can determine the system blink rate and adjust their canvas blink rate to match that set in the operating system. """ |
| 21:56 | <jgraham> | Doesn't it make more sense just to have a lower blink rate always |
| 21:56 | <jgraham> | ? |
| 21:56 | jgraham | finds no evidence that the blink rate can be changed on OSX |
| 21:57 | <jgraham> | Oh it can be via the command line |
| 21:57 | <zcorpan_> | does a blinking cursor really cause epileptic seizures in some users? |
| 21:57 | <jgraham> | So I guess some a11y thing could expose that |
| 21:59 | <zcorpan_> | and yeah, why would you want the cursor blink over 3 times a second anyway |
| 21:59 | <TabAtkins> | So you can cause seizures. |
| 21:59 | <zcorpan_> | need to make sure the cursor is at least 200x200 pixels also |
| 22:00 | <TabAtkins> | Really, though, I'll just include a <video> of Porygon if I wanted to make people seizure. |
| 22:00 | <jgraham> | zcorpan_: There are at least reports on the internet of blinking cursors causing seziures |
| 22:00 | <jgraham> | although few seem to be first hand |
| 22:01 | <AryehGregor> | Maybe they're fatal. |
| 22:02 | <jgraham> | (but I agree it seems odd to specifically optimise for that case when there are so many other things that could cause seziures if you are photosensitive) |
| 22:07 | <jgraham> | http://www.amazon.com/Photosensitive-Epilepsy-Clinics-Developmental-Medicine/product-reviews/1898683026 appears to be an actual human claiming to have seziures triggered by blinking cursors. Although the reporter is a little zealous in his call to ban blinking in all forms |
| 22:09 | <jgraham> | http://www.justdamn.com/?p=1627 |
| 22:09 | <TabAtkins> | No blinking rates between .1Hz and 1000Hz? |
| 22:09 | <TabAtkins> | If anyone's injured by something blinking on off for 5 seconds each, they're pretty much screwed anyway. |
| 22:10 | <jgraham> | Although that one could be hyperbole |
| 22:10 | <jgraham> | TabAtkins: That was part of the zealousness |
| 22:10 | <jgraham> | I have never seen anyone else worry about rates < 2Hz |
| 22:11 | <TabAtkins> | I believe the blog post you just linked to is hyperbole. |
| 22:11 | <zcorpan_> | the range .1Hz .. 1000Hz probably excludes all monitors |
| 22:12 | <TabAtkins> | I'm pretty certain it does. Most monitors run between 50 and 100 Hz. |
| 22:14 | AryehGregor | is fairly sure 50 and 100 lie between .1 and 1000 |
| 22:15 | <zcorpan_> | i read about a new type of monitor being used for read pads that don't blink at all |
| 22:15 | <jgraham> | 3-60 says http://www.epilepsy.org.uk/info/photosensitive/intro |
| 22:15 | <jgraham> | which I tend to trust more than some random guy on amazon |
| 22:16 | <zcorpan_> | instead it moved around some ink-like stuff and didn't require any battery to show the image so long as it doesn't change |
| 22:16 | <AryehGregor> | It's funny how some people in the HTMLWG seem to all think that the WHATWG is some nonentity that will go away if the W3C ignores it. |
| 22:22 | <AryehGregor> | Random note: whoever decided that the Nexus One should make a noise by default every time they received an e-mail, obviously didn't subscribe to many mailing lists. |
| 22:22 | <jgraham> | Hmm, I don't think the drawFocusRing(x,y,w,h) works very well |
| 22:23 | <jgraham> | Although I might be wring |
| 22:23 | <jgraham> | *wrong |
| 22:23 | <jgraham> | Because the algorithm doesn't convert w,h accoring to the current transform matrix |
| 22:24 | <jgraham> | (which makes sense if the result has to be a rectangle) |
| 22:24 | <jgraham> | (but means that the area defined by the rectangle might not cover what you expect) |
| 22:26 | <jgraham> | But presumably the UA can calculate the best-fitting rectangle from the path anyway, so I don't see why the author has to supply it |
| 22:28 | <jgraham> | I assume there is some reason it would be bad (perf?) to make drawCustomRing a callback that is called if the UA needs to draw a custom focus ring? |
| 22:30 | jgraham | wonders what "editable fallback content" means |
| 22:32 | <zcorpan_> | <canvas contenteditable>? |
| 22:32 | <jgraham> | Maybe |
| 22:32 | <zcorpan_> | or <canvas><textarea> |
| 22:32 | <jgraham> | It feels like there is some design assumption hidden in that phrase |
| 22:33 | <jgraham> | that is not explicitly stated |
| 22:33 | <zcorpan_> | i guess it's 'bespin made accessible' |
| 22:33 | <zcorpan_> | though my guess is as good as anyone's |
| 22:35 | <jgraham> | Yeah, this whole set of changes reads like a reaction to bespin |
| 22:35 | <jgraham> | But bespin deliberately avoided also using the DOM for content, at least initially |
| 22:35 | <jgraham> | Did the reasons for that evaporate? |
| 22:37 | <lifechamp> | Flash HTML5 compiler coming? How about this for now? http://csharksgames.blogspot.com/2009/07/jangaroo-html-5-killer-as3-to-js.html |
| 22:37 | <jgraham> | In the nicest posible way, this is not camera ready spec text |
| 22:38 | <jgraham> | I am finding it quite hard to follow |
| 22:40 | <jgraham> | Is there enough information in the text API to make computing the height/width of characters reasonable? |
| 22:42 | <jgraham> | I think the setCaretSelectionRect stuff doesn't make much sense without measureText providing all the required information |
| 22:42 | <Hixie> | jgraham: it has more fundamental problems, like nobody will use the selection part of the API since it doesn't get them anything |
| 22:42 | <Hixie> | jgraham: (other than accessibility in ATs, which they can't test, so it'll always be wrong) |
| 22:42 | <jgraham> | Hixie: Yes, that is also a problem |
| 22:43 | <Hixie> | it's better than usemap="" though! |
| 22:44 | <jgraham> | I don't see that we need to block on this stuff; it addresses problems that only a small percentage of apps will encounter |
| 22:44 | <jgraham> | I expect violent disagreement on that point though |
| 22:46 | <jgraham> | Why does setCaretSelectionRect return true? |
| 22:48 | <jgraham> | Actually I'm lost |
| 22:48 | <jgraham> | How does one know the selection is a rectangle? |
| 22:48 | <Hixie> | it won't be |
| 22:48 | <Hixie> | i'm guessing the rectangle is the zoom rectangle |
| 22:49 | <jgraham> | Yeah that makes sense |
| 22:49 | <jgraham> | But I don't think the spec actually says that |
| 22:51 | <jgraham> | The caretBlinkRate thing specifically notes that the default is 2Hz |
| 22:52 | <jgraham> | (on typical operating systems) |
| 22:52 | <jgraham> | Given the figures quoted earlier, I wonder why it is needed |
| 22:53 | <jgraham> | (actually I still wonder why we care so much about the specific case of carets when so many other things could cause similar but worse problems) |
| 22:54 | <jgraham> | (since they won't typically be at 2Hz) |
| 22:55 | <jgraham> | I guess I will send this feedback to the list tommorow, so if I am totally wrong about anything it would be nice to let me know |
| 23:10 | <MikeSmith> | http://dev.opera.com/articles/view/css3-show-and-hide/ is nice |
| 23:10 | <MikeSmith> | http://devfiles.myopera.com/articles/1861/css-menu-ex4.html |
| 23:11 | <zcorpan_> | hmm, basing it on :focus seems weird |
| 23:12 | <zcorpan_> | :target seems better if you want css-only |
| 23:13 | <MikeSmith> | dunno about the details, but for one thing, it seems like this could be used to emulate support for the display element |
| 23:13 | <MikeSmith> | until such time as we get native implementations |
| 23:13 | <TabAtkins> | Yeah, you really don't want :focus - it'll rehide or reshow things as focus moves, which is weird. |
| 23:13 | <zcorpan_> | you mean details element? |
| 23:13 | <MikeSmith> | zcorpan_: erp, yeah |
| 23:13 | <MikeSmith> | details |
| 23:13 | TabAtkins | still likes using the hidden :checked hack. |
| 23:14 | <MikeSmith> | would make a good html5doctor article maybe |
| 23:14 | MikeSmith | looks around for some HTML5 doctors in the house |
| 23:14 | <zcorpan_> | as TabAtkins says, :focus doesn't work except for a minimal demo page |
| 23:14 | <zcorpan_> | the article acks this at the end |
| 23:15 | <MikeSmith> | http://dev.opera.com/forums/topic/600712?t=1276554067&page=1#comment5792922 |
| 23:15 | <MikeSmith> | yeah |
| 23:15 | <zcorpan_> | an infobox instead of a menu would be more appropriate for showcasing :focus |
| 23:16 | <zcorpan_> | but maybe even that would be annoying |
| 23:38 | <zcorpan_> | does safari not fire any events for the arrow keys? |
