| 02:28 | <zcorpan_> | good morning |
| 02:30 | <MikeSmith> | zcorpan_: morning indeed |
| 02:30 | <MikeSmith> | why you up so early? |
| 02:31 | <zcorpan_> | jetlagged |
| 02:32 | <zcorpan_> | came back from las vegas yesterday |
| 02:33 | <zcorpan_> | http://mathildaochsimon.se/brollopsresa |
| 03:08 | <MikeSmith> | zcorpan_: dude |
| 03:08 | <MikeSmith> | congratulations |
| 03:08 | <MikeSmith> | didn't know |
| 03:08 | <zcorpan_> | MikeSmith: thanks |
| 03:09 | <MikeSmith> | sharp suit |
| 03:09 | <zcorpan_> | the picture on that page was from before the wedding |
| 03:12 | <zcorpan_> | see my facebook avatar for the wedding suit |
| 03:17 | <zcorpan_> | so what's happened the past five weeks? |
| 03:32 | <MikeSmith> | zcorpan_: Webkit has an HTML5 parser now |
| 03:32 | <MikeSmith> | I think that happened within the last 5 weeks |
| 03:36 | <zcorpan_> | cool |
| 03:36 | <zcorpan_> | with foreign land support? |
| 03:37 | <MikeSmith> | yeah |
| 03:37 | <MikeSmith> | and entities |
| 03:37 | <MikeSmith> | abarth|pie: and Eric Seidel did that work |
| 03:38 | <MikeSmith> | zcorpan_: so you guys need to get one implemented too |
| 03:38 | <MikeSmith> | I think the Webkit guys produced a number of new test cases as part of their work |
| 03:38 | <MikeSmith> | including for document.write |
| 03:40 | <zcorpan_> | what else happened? i noticed websocket framing changed |
| 03:40 | <MikeSmith> | yeah |
| 03:40 | <MikeSmith> | um |
| 03:40 | <MikeSmith> | there are going to be some further changes there |
| 03:40 | <MikeSmith> | Ian Fette is now editing the protocol spec |
| 03:42 | <zcorpan_> | is the latest protocol spec still part of complete.html? |
| 03:42 | <MikeSmith> | and it seems it could end up becoming a bit more complicated |
| 03:42 | <MikeSmith> | yeah |
| 03:42 | <MikeSmith> | the latest that Hixie produced |
| 03:43 | <MikeSmith> | but the hybi wg is going to be publishing it separately somewhere |
| 03:44 | <zcorpan_> | so when fette makes changes, complete.html will be obsolete? |
| 03:46 | <zcorpan_> | foreign land parsing changed too, right? |
| 03:48 | <zcorpan_> | oh it just reverted my bad suggestion |
| 03:49 | <zcorpan_> | or it seems it's different than before |
| 03:55 | <Hixie> | when ian published his draft i'll take the protocol stuff out of complete.html |
| 03:55 | <Hixie> | publishes |
| 03:56 | <zcorpan_> | ok |
| 08:08 | <zcorpan_> | Hixie: why did you remove support for '/' in targetOrigin? |
| 08:12 | <jgraham> | Oh zcorpan_ is back |
| 08:12 | <jgraham> | yay! |
| 08:16 | <jgraham> | zcorpan_: The hybi list has become more insane with the appearance of Shelby "Axiomatic Proof" Moore |
| 08:17 | <zcorpan_> | jgraham: awesome |
| 08:17 | <zcorpan_> | haven't read the hybi email yet |
| 08:25 | <hsivonen> | zcorpan_: probably not worth the time to read |
| 08:26 | <hsivonen> | zcorpan_: and congratulations |
| 08:28 | <zcorpan_> | hsivonen: thanks |
| 08:35 | <zcorpan_> | hmm, what if a vendor called "data" uses vendor--feature="" |
| 08:37 | <hsivonen> | zcorpan_: I guess we should blacklist vendors called data. |
| 08:37 | <hsivonen> | (or better yet, revert to _vendor-feature) |
| 08:37 | <hsivonen> | maybe Hixie did that already |
| 08:38 | <hsivonen> | awesome. My hotmail account that I've only used for debugging Hotmail is already getting spam |
| 08:39 | <asmodai> | mmm, that's odd |
| 08:39 | <asmodai> | why would Minefield have problems with Cleartype and bold font styling |
| 08:39 | <asmodai> | the showing of the red/green parts on the sides of fonts is annoying :S |
| 08:43 | <hsivonen> | does any browser have a debugger that allows breaking on the next click, mousedown or mouseup event handler? |
| 08:44 | <hsivonen> | i.e. breaking without no clue about where the event handler code lives |
| 08:45 | <zcorpan_> | dragonfly has 'stop at new script', dunno if that does what you want |
| 08:45 | <hsivonen> | let's see |
| 08:50 | <hsivonen> | zcorpan_: it seems to break where I want |
| 08:50 | <hsivonen> | zcorpan_: thanks! |
| 08:50 | <zcorpan_> | np |
| 08:50 | <asmodai> | mmm http://www.in-nomine.org/~asmodai/fonts-1.png |
| 08:50 | <hsivonen> | now if Dragonfly had a way to pretty-print minified JS for debugging... |
| 08:51 | <asmodai> | The difference is bizarre |
| 08:51 | <hsivonen> | asmodai: which platform? |
| 08:51 | <asmodai> | Windows 7 |
| 08:51 | <hsivonen> | asmodai: have you enabled D2D in Minefield? |
| 08:52 | <asmodai> | hsivonen: if it is not enabled by default, no |
| 08:52 | <hsivonen> | asmodai: Minefield has undergone all sorts of changes in terms of which rasterizer it uses for what fonts and what the defaults are |
| 08:52 | <hsivonen> | asmodai: I haven't paid close enough attention to say what the current set of rasterizers and defaults is |
| 08:53 | <asmodai> | I don't hope this will be the default though ;) |
| 08:53 | <asmodai> | err |
| 08:53 | <asmodai> | I hope this won't |
| 08:53 | <asmodai> | *sigh* |
| 08:53 | <hsivonen> | whatever the default becomes, someone won't like it :-) |
| 08:53 | <hsivonen> | :-( rather |
| 08:53 | <hsivonen> | I mean some people actually prefer ClearType over Quartz! |
| 08:54 | <asmodai> | Well, this just hurts your eyes after a little while -- and this is the default Windows 7 display way |
| 08:54 | <asmodai> | Heh, both are hacks to accomplish their little goal |
| 08:55 | <asmodai> | also wonder where the little icons (image arrows) went to btw XD |
| 08:56 | <hsivonen> | asmodai: good question. URL? |
| 08:57 | <hsivonen> | asmodai: do you see this font rendering difference on all pages on just this one_ |
| 08:57 | <hsivonen> | ? |
| 08:57 | <hsivonen> | could it be that Minefield has just made everything bold? |
| 09:00 | <hsivonen> | debugging Hotmail's JS seems hopeless due to minification and debuggers stepping by line |
| 09:06 | <hsivonen> | kudos to the Dragonfly team for revealing that a node has a given event handler |
| 09:06 | <hsivonen> | unfortunately, it doesn't show a disassembly of the JS function object |
| 09:07 | <asmodai> | hsivonen: all pages |
| 09:08 | <asmodai> | hsivonen: can give you another example in a sec |
| 09:09 | <hsivonen> | ooh. I get the disassembly in DOM Inspector in Minefield |
| 09:09 | <hsivonen> | too bad the disassembly is rather cryptic |
| 09:10 | <Hixie> | zcorpan_: send mail on the thread about vendor-- saying that data-* clashes. About '/', I'll be putting that back, I was over-eager in my reverts. |
| 09:12 | <asmodai> | hsivonen: http://www.in-nomine.org/~asmodai/fonts-2.png |
| 09:13 | <asmodai> | also curious that the fonts are suddenly 1.2-1.3 times larger |
| 09:14 | <zcorpan_> | Hixie: ok, sent |
| 09:15 | <Hixie> | zcorpan_: ta |
| 09:19 | <hsivonen> | asmodai: I suggest checking your gfx.font_rendering.* prefs in about:config and resetting them to defaults in Minefield if they aren't at the defaults |
| 09:27 | hsivonen | wonders how Opera deals with racy document.writes |
| 09:32 | <zcorpan_> | <iframe class="youtube-player" type="text/html" width="640" height="385" src="http://www.youtube.com/embed/VIDEO_ID"; frameborder="0"></iframe> - yay two invalid attributes |
| 09:37 | <hsivonen> | zcorpan_: we should make frameborder=0 conforming, IMO |
| 09:41 | <zcorpan_> | hsivonen: file a bug :P |
| 09:42 | <hsivonen> | hmm. Chromium nightlies have the HTML5 parser but not the newly-announced MathML support |
| 09:43 | <hsivonen> | abarth|fileapi: Do Chromium nightlies have some kind of hack that allows random scripts document.write() content magically before EOF in a spec-incompliant way? |
| 09:44 | hsivonen | wonders why http://uglyducklinghouse.blogspot.com/ doesn't break in Chromium but breaks in Minefield |
| 09:45 | <asmodai> | hsivonen: gfx.* is all default -- just checked |
| 09:46 | <asmodai> | hsivonen: So didn't have to reset anything. |
| 09:47 | <hsivonen> | asmodai: OK. I don't have an explanation. |
| 09:47 | <asmodai> | hsivonen: Guess I should log a bug for it. |
| 09:47 | <hsivonen> | yeah |
| 10:11 | <asmodai> | hsivonen: https://bugzilla.mozilla.org/show_bug.cgi?id=591887 |
| 10:11 | <asmodai> | done |
| 10:22 | <hsivonen> | asmodai: thanks |
| 10:57 | <asmodai> | hsivonen: If I need to add anything, please let me know what's missing |
| 11:24 | <asmodai> | hsivonen: interesting, I should try that with Aero off as well |
| 13:34 | <zcorpan_> | annevk: thanks |
| 14:29 | <annevk> | hmm so scrollX/scrollY are the same as pageXOffset/pageYOffset |
| 14:29 | <annevk> | how did we end up with two?! |
| 14:32 | <Rik`> | annevk: isn't there a difference on iframe ? |
| 14:38 | <annevk> | they're on Window |
| 15:36 | <adactio> | gsnedders: do you have time to answer a quick question about your outliner tool? |
| 15:36 | <gsnedders> | adactio: I guess. |
| 15:37 | <adactio> | Thanks. I'm running this page through the outliner: http://huffduffer.com/popular and getting this outline: http://gsnedders.html5.org/outliner/process.py?url=http%3A%2F%2Fhuffduffer.com%2Fpopular |
| 15:37 | <adactio> | Now, I'm expecting to see an Untitled Section for the nav element at the top of the page, but I'm surprised by the placement of the Untitled Section in the outline. |
| 15:38 | <adactio> | Is the Untitled Section in the outline that nav element even though it comes *before* the first heading in the source order? |
| 15:38 | <gsnedders> | Hmm, interesting |
| 15:39 | <annevk> | your h1 is for the whole document |
| 15:39 | <adactio> | annevk: yes |
| 15:39 | <annevk> | well, actually, let me stay out of this, I'm not fully up to speed |
| 15:39 | <jgraham> | adactio: Yes |
| 15:39 | <jgraham> | It makes sense |
| 15:41 | <adactio> | jgraham: yes, it does make sense. I was just surprised by the cleverness of it ...that having an Untitled Section *before* the first heading of the page still appears *under* that heading. |
| 15:41 | <gsnedders> | I'm not fully up to speed either |
| 15:41 | <gsnedders> | I just wrote the code |
| 15:41 | <gsnedders> | (And a long time ago) |
| 15:41 | <adactio> | gsnedders: and very useful it is, too. |
| 15:41 | <jgraham> | adactio: It doesn't care about source order, only logical order (which depends on source order somewhat) |
| 15:42 | <adactio> | jgraham: Cool. The outline algorithm is smarter than I thought. |
| 15:42 | <gsnedders> | It's smart enough that it confuses me when I try and implement it |
| 15:42 | <adactio> | gsnedders: :-) |
| 15:43 | <gsnedders> | So really, Hixie is the person to ask about how it works. I only care if my code is b0rked. :P |
| 15:44 | <adactio> | So, I *think* if I had a document like: <body><aside><h1>Bar</h1></aside><h1>Foo</h1></body>, the outline would be Foo -> Bar. |
| 15:44 | <adactio> | gsnedders: I fear your code may be b0rked for the file upload part (works fine by URL though). |
| 15:44 | <annevk> | why is there no textarea option for the outliner? |
| 15:44 | <gsnedders> | annevk: Because I'm lazy. |
| 15:45 | <annevk> | typical |
| 15:46 | <gsnedders> | annevk: I'm a teenager, remember? :P |
| 15:47 | <adactio> | Yup, that's one clever algorithm: http://gsnedders.html5.org/outliner/process.py?url=http%3A%2F%2Fadactio.com%2Ftest.html from http://adactio.com/test.html |
| 15:48 | <Rik`> | adactio: maybe https://developer.mozilla.org/en/Sections_and_Outlines_of_an_HTML5_document can help |
| 15:52 | <adactio> | This *does* make sense to me: I just have to keep reminding myself that the body element is sectioning content. If I mentally replace body with section, it becomes clearer to me: <section><aside><h1>Bar</h1></aside><h1>Foo</h1></section> = Foo -> Bar |
| 15:53 | <annevk> | that foo is higher level makes sense to me |
| 15:53 | <annevk> | but that it comes before bar... |
| 15:53 | <annevk> | meh |
| 15:54 | <Rik`> | annevk: there should be an untitled section maybe |
| 15:55 | <Rik`> | 1. Untitled section 1.1 Bar 2. Foo |
| 15:55 | <adactio> | annevk: It makes sense (though it's not exactly intuitive): you could have a bunch of sectioning content things (navs, asides, etc.) *before* you introduce the heading of your content. |
| 15:56 | <adactio> | Rik`: that's what I initially expected but the more I thought about it, the more it made sense that it's 1. Foo 1.1 Bar |
| 15:57 | <adactio> | It isn't exactly easy to explain, though. But maybe I'll give it a shot with a blog post. |
| 15:58 | <jgraham> | It feels very intuitive to me... If it didn't work like this it would be broken for common cases like adactio's original example |
| 15:59 | <Rik`> | yeah, I guess so |
| 15:59 | <adactio> | jgraham: for me, it's initially unintuitive but then becomes sensible ...if that makes sense. |
| 16:01 | <jgraham> | adactio: Yeah :) |
| 16:05 | <annevk> | so scrollIntoView() and #target use the same algorithm |
| 16:45 | <annevk> | can someone help me out with a definition for scrollIntoView? |
| 16:46 | <annevk> | basically, you need to go through all ancestor scrolling containers (in order), queue tasks to dispatch events and abort the moment you hit a scroll container which is not same-origin with the current one |
| 16:47 | <annevk> | scrolling containers can be either element or viewport |
| 17:12 | <annevk> | i have something now... |
| 17:12 | <annevk> | For each ancestor element or viewport that establishes a scrolling box starting with the innermost scrolling box, in order: |
| 17:12 | <annevk> | if the Document associated with the element or viewport is not same origin with the Document associated with the element on which the method was invoked terminate these steps. |
| 17:12 | <annevk> | if the align to top flag is set align the top of border box of the element with the top of the scrolling box |
| 17:12 | <annevk> | otherwise, align the bottom of the border box of the element with the bottom of the scrolling box |
| 17:12 | <annevk> | left/right? |
| 17:12 | <annevk> | queue a task ... |
| 17:13 | <annevk> | not very good but better than what is in HTML5 |
| 17:15 | <nimbupani> | "If the document associated with the element on which scrolltoView was called is not same origin as document associated with the viewport or parent element, do not continue" |
| 18:00 | <gsnedders> | Gah, my Xbox 360 after being unused for five months has decided to start crashing on me… in a I'm going to die way. |
| 18:05 | <AryehGregor> | In my day, there was no such thing as a console "crashing". |
| 18:05 | <AryehGregor> | Crazy people, putting operating systems on gaming consoles. You get what you deserve. |
| 18:05 | <annevk> | defined |
| 18:06 | <TabAtkins> | AryehGregor: You're in your mid-20s. Today is "back in [your] day". |
| 18:06 | <AryehGregor> | Early 20s, not mid-20s. |
| 18:06 | <AryehGregor> | I'm 22. |
| 18:07 | <TabAtkins> | Ah, you're a year younger than I thought. |
| 18:07 | <AryehGregor> | My day vis-a-vis gaming consoles was circa 1998. |
| 18:07 | <AryehGregor> | I had a Super NES, and some Gameboys. |
| 18:08 | AryehGregor | is about 22.6276 years old, in fact. |
| 18:09 | <annevk> | except for left/right scrolling |
| 18:35 | <ojan> | is there a good page that outlines the life of a w3c spec? |
| 18:35 | <AryehGregor> | http://www.w3.org/2005/10/Process-20051014/ |
| 18:35 | <AryehGregor> | Particularly: http://www.w3.org/2005/10/Process-20051014/tr.html#Reports |
| 18:36 | AryehGregor | wonders why Hixie's bug-response script posts the diff in a separate comment from the response |
| 18:36 | <ojan> | AryehGregor: thx |
| 18:38 | <jgraham> | gsnedders: You want a Wii |
| 18:42 | <Hixie> | AryehGregor: completely unrelated scripts |
| 18:42 | <Hixie> | my response is a bookmarklet |
| 18:42 | <Hixie> | the diff is my commit script |
| 18:42 | <AryehGregor> | Why aren't they the same script? |
| 18:43 | <Hixie> | i don't commit if i reject |
| 18:43 | <AryehGregor> | Hmm, makes sense. |
| 18:44 | <Hixie> | they also run in different computers in different cities |
| 18:44 | <Hixie> | my response is done in my web browser on my laptop |
| 18:44 | <Hixie> | the commit is done on a computer I ssh to |
| 18:44 | <AryehGregor> | That part isn't an a priori reason why they should be separate, though. |
| 18:45 | <jgraham> | http://thewildernessdowntown.com/ is kinda awesome |
| 18:45 | <jgraham> | Need to work out why it isn't working in Opera |
| 18:45 | <ojan> | Is there any chance of getting a change to DOM3 Core? |
| 18:45 | <ojan> | Hixie: ^^^ |
| 18:46 | <ojan> | WebKit is planning on changing to match Gecko https://bugs.webkit.org/show_bug.cgi?id=19524 |
| 18:46 | <ojan> | which violates the DOM3 Core spec |
| 18:46 | <AryehGregor> | Why does it open a billion windows in different places? |
| 18:46 | <jgraham> | ojan: Web DOM Core FTW |
| 18:46 | <Hixie> | AryehGregor: well somewhere i have to have a browser so i can see the bug in the first place, and i'd rather not to that part over ssh |
| 18:46 | <Hixie> | AryehGregor: and the commit can't be done anywhere by the machine where i edit |
| 18:47 | <Hixie> | AryehGregor: so... the simplest solution is what i have :-) |
| 18:47 | <jgraham> | AryehGregor: That is part of the experience |
| 18:47 | <Hixie> | ojan: yes, you just have to find an editor first |
| 18:47 | <hober> | I noticed the other day that Emacs' html-mode has some special highlighting for a <rev> element. |
| 18:47 | <AryehGregor> | They got opened in random hard-to-see places, though. |
| 18:47 | <AryehGregor> | Thanks, metacity. |
| 18:47 | <jgraham> | It seems like they could have not used OS windows though |
| 18:47 | <hober> | Any idea what bizarre side-road of HTML history had such an element? |
| 18:47 | <jgraham> | Just <div>s |
| 18:48 | <ojan> | Hixie: that sounds...complicated |
| 18:48 | <Philip`> | hober: Maybe it's got element names and attribute names in the same list? |
| 18:49 | <jgraham> | emacs' html-mode is quite dreadful |
| 18:49 | AryehGregor | remembers that he's supposed to mark bugs he filed as CLOSED if he likes the resolution |
| 18:49 | TabAtkins | forgets that he should do that. |
| 18:50 | jgraham | too |
| 18:50 | <hober> | jgraham: indeed. I've looked into updating its lists of elements & attributes to match HTML5, which is when I encountered this weird <rev> thing. |
| 18:50 | <hober> | Philip`: that's what I first thought, but the special highlighting happens for <rev>foo</rev> and not <a rev=foo href=bar>baz</a> |
| 18:51 | <hober> | I couldn't find such an element in HTML 3.0 or HTML+, which seemed like the best bets in terms of elements-that-never-were |
| 18:52 | <jgraham> | You already tried the common source of insane elements :) |
| 18:52 | <jgraham> | Maybe it's just a bug? |
| 18:53 | <hober> | Yeah, it's probably a vestige of some other format the sgml-mode.el's author commonly edited at the time |
| 19:14 | <AryehGregor> | Enums are a mess. |
| 19:31 | <abarth> | hsivonen: i think our document.write blow-away-the-document behavior should match the spec |
| 19:31 | <abarth> | hsivonen: maybe we should look at a reduction? |
| 19:53 | AryehGregor | finds a spec bug in WebKit reflection code that's only reproducible if you set particular attributes in a particular *order*. Sigh. |
| 19:57 | <cheeser> | fun |
| 20:33 | AryehGregor | finds it interesting that his test suite has so far found more WebKit bugs than Gecko bugs, even though he's not done filing WebKit bugs yet. |
| 20:33 | AryehGregor | isn't sure if this is because Gecko follows the specs, or because the specs follow Gecko. |
| 20:34 | <AryehGregor> | (both are low, though, compared to IE or Opera) |
| 20:34 | <TabAtkins> | Could be both, to be honest. |
| 20:34 | <AryehGregor> | (especially compared to IE) |
| 20:49 | <othermaciej> | what are the specs in question/ |
| 20:49 | <othermaciej> | ? |
| 20:50 | <AryehGregor> | This is on my reflection tests, http://aryeh.name/tests/reflection.html |
| 20:51 | <AryehGregor> | http://aryeh.name/tests/bugs_filed.txt |
| 20:52 | <AryehGregor> | I've filed basically every bug I found in Gecko, but only some of the WebKit bugs so far. |
| 20:53 | <AryehGregor> | (ignoring things that have pending spec bugs open) |
| 20:55 | <othermaciej> | most of why I see reported in Safari is dir on various elements |
| 20:55 | <AryehGregor> | Yes, put "dir" (with quotes) in the box to hide that. |
| 20:56 | <AryehGregor> | I haven't reported most of the enum stuff yet, since browsers disagree a lot and some of it's changed very recently or might change soon. |
| 20:56 | <AryehGregor> | I was told that a bunch of the enum stuff changed recently, Firefox nightlies have it but not betas yet. |
| 20:57 | <AryehGregor> | Anyway, I'm filing bugs on all of it, at least for Gecko and WebKit. |
| 22:58 | <annevk> | ojan, Hixie, I plan to update DOM 3 Core |
| 22:58 | <annevk> | rename it Web DOM Core or something |
| 22:58 | <annevk> | but CSSOM CSS Values is first |
| 23:28 | <AryehGregor> | Okay, so, why is Google recommending people use SPF softfail? Peculiar. http://www.google.com/support/a/bin/answer.py?answer=178723 |
| 23:28 | <AryehGregor> | Hmm. |
| 23:29 | <AryehGregor> | gmail.com, yahoo.com, hotmail.com, aol.com, microsoft.com all use either softfail or not even that. |
| 23:29 | <AryehGregor> | Is hard fail unsafe or something? |
| 23:30 | <AryehGregor> | Aha, Facebook uses hard fail. |
| 23:31 | <AryehGregor> | So do Blogger, Twitter, and YouTube. |
| 23:31 | <AryehGregor> | Can't be totally unsafe. |
| 23:31 | <AryehGregor> | Oh, of course, they must set it to softfail because it breaks forwarding. |
| 23:31 | <AryehGregor> | Makes sense. |
| 23:32 | <AryehGregor> | So I'll use hard fail for my domain. |
