| 00:28 | <nessy> | Hixie: wow - just wow - I'm totally in awe of the long email on timed tracks - that will take a bit to digest! |
| 00:29 | <gsnedders> | Has Hixie writen another multi-megabyte email? |
| 00:29 | <TabAtkins> | Yes. |
| 00:29 | <TabAtkins> | But it was easy to read. |
| 00:43 | <Hixie> | i've replied to every e-mail reporting problems up to august 3rd or so |
| 00:43 | <Hixie> | i have about 300 e-mails of feature requests that i've not replied to yet, some going back years |
| 00:43 | <Hixie> | but they're not getting replies any time soon |
| 00:43 | <Hixie> | now, bugs |
| 01:46 | <Dashiva> | Hixie: I get the same in MPC |
| 01:48 | <Hixie> | Dashiva: for 018.srt? |
| 01:48 | <Hixie> | Dashiva: cool, thanks |
| 01:48 | <Dashiva> | Although I guess it's wrong to name the player, it's DirectVobSub doing the work |
| 01:48 | <Hixie> | k |
| 01:52 | <Hixie> | one day we should add a TimedTrackCe.play() method that seeks to the cue start time and starts playing |
| 02:00 | <nessy> | oh, that would be nice indeed |
| 02:00 | <nessy> | handn't thought of that yet |
| 02:00 | <nessy> | but then I haven't implemented anything yet - that's next |
| 02:07 | Hixie | checks in the first of hopefully many examples for the timed track apis |
| 02:08 | <Hixie> | btw my twitter posting things are all broken because they turned off basic auth |
| 02:08 | <Hixie> | if anyone actually cares, let me know and i'll send you the code for you to update :-) |
| 06:34 | <TabAtkins_> | boblet: You wanna talk Flexbox/other layout modules? I saw your request a few days ago, but haven't been able to catch you online. |
| 06:47 | <boblet> | hey TabAtkins_ |
| 06:47 | <TabAtkins_> | yo |
| 06:48 | <boblet> | thanks. yeah, just reading up on them and wondering what the dealz is with three layout modules, progress, etc |
| 06:48 | <boblet> | also I’ve heard rumors that flexbox is coming to IE9 but can’t seem to find any proof |
| 06:51 | <TabAtkins_> | Alex Mogilevsky is implementing experimental support for Flexbox in the platform previews, but I have no clue if that'll remain in the final build or not. |
| 06:51 | <TabAtkins_> | I assume the third layout module you're talking about is Grid Positioning? |
| 06:52 | <boblet> | TabAtkins_: also, atm I’m trying to implement the OOCSS AG test http://wiki.github.com/stubbornella/oocss/grids and am surprised that nested boxes seem to affect things when everything has box-flex:1 |
| 06:52 | <boblet> | aah, ok good to know. Here’s hoping IE9 will have it. |
| 06:53 | <boblet> | yeah Grid Positioning and Template Layout |
| 06:53 | <boblet> | I’m guessing they can all do the same stuff, but am wondering if some have strengths for certain types of content eg web apps vs standard document content |
| 06:54 | <TabAtkins_> | They can't quite do all the same stuff. ^_^ |
| 06:54 | <boblet> | hehe |
| 06:55 | <boblet> | if you could highlight diffs and strengths/weaknesses that’d be awesome |
| 06:55 | <TabAtkins_> | Flexbox is in a state of flux. AlexMog and I are reworking it to better suit both of our desires. I hope to have a revised draft out by the end of this week, but it might not show up until next week. |
| 06:56 | <TabAtkins_> | Template is pretty stable, though no one's taken a first step on it yet. I've been pushing it locally in Google, and it appears to have decent support among the other implementors in the CSSWG. |
| 06:56 | <TabAtkins_> | Grid Positioning is eh right now. I think it's a good idea, but AlexMog is unimpressed with it and wants to somehow change how grids are referred to. I don't know what his plans are for it, but I don't think it'll be implemented in the immediate future. |
| 06:57 | <boblet> | I saw the flexbox spec on xanthir.com… |
| 06:58 | <TabAtkins_> | That was a fairly radical redesign, and I've scaled it back in collab with AlexMog. I really liked it, but there were several edge-case usability issues that made it not so great an idea. So now it's back to something much closer to the original draft. |
| 06:58 | <boblet> | huh, interesting. feel free to ping me with the next draft :) |
| 06:59 | <boblet> | ok that’s about what I guessed at from spec completion levels, so what would you say are the strengths/weaknesses of them? |
| 06:59 | <TabAtkins_> | Re: nested boxes in Flexbox - Yeah, each flexbox does its flex calculations separately, so children flexboxes don't interact with the space distribution of their parents. |
| 07:00 | <TabAtkins_> | Flexbox is great for a number of linear-layout related things. You see flexbox-like things used a *lot* in applications' toolbars and such, for example. Firefox uses the XUL translation of flexbox concepts for all of its UI, for example. |
| 07:01 | <TabAtkins_> | However, it's pretty clumsy for doing a whole website in, because you have to deal with some pretty funky nesting issues. Elements become siblings or not based on precise details of how your layout works, rather than on their semantic relationships. |
| 07:01 | <TabAtkins_> | I have an article about why Flexboxes suck for website layout, actually... |
| 07:02 | <boblet> | ohrly? link pls! |
| 07:02 | <TabAtkins_> | http://www.xanthir.com/:kf8 |
| 07:05 | <TabAtkins_> | Overall site layout is what Template is designed for. It's a super-easy WYSIWYG-style grid layout system. |
| 07:06 | <TabAtkins_> | It's basically equivalent to making a "shadow table" and shoving elements into various cells, in terms of what abilities and constraints you have to work with. |
| 07:06 | <TabAtkins_> | Grid Positioning is something that lies on top of everything else. All it does is define a new length unit that can be used for positioning and sizing, the 'gr' unit, which refers to the space between "grid lines". |
| 07:07 | <TabAtkins_> | Grid lines can either be created implicitly by various other layout modules (like Table, Multicol, or Template), or explicitly via grid-rows and grid-columns. |
| 07:08 | <boblet> | hrm, ok |
| 07:11 | <boblet> | so it sounds like flexbox is good for per-line layout, especially for toolbars and stuff |
| 07:12 | <boblet> | and that template layout and grid positioning address full page stuff using table and DTP grid line metaphors |
| 07:13 | <TabAtkins_> | Yeah. |
| 07:14 | <boblet> | I’m a little sad that flexbox can’t be used for page layout as it actually has some implementation, even if it doesn’t separate content and presentation enough |
| 07:15 | <boblet> | but yeah, dimensions not being calculated based on parent dimensions seems like a pretty big problem in addition |
| 07:16 | <TabAtkins_> | Oh, it *can* be used for page layout. It's just kinda sucky for it. You see this in the OOCSS examples for Grid - there's tons of nesting required for any complex nonlinear layout. |
| 07:18 | <boblet> | but my attempt at the OOCSS grid example doesn’t seem to work only using nesting and box-flex:1, it seems like the most nested boxes take lots of space, so less nested units aren’t able to be equally sized due to content requirements |
| 07:18 | <boblet> | (if that makes sense :) ) |
| 07:19 | <TabAtkins_> | Doing that AG test in Template Layout would be: |
| 07:19 | <TabAtkins_> | display: template("abbccd" |
| 07:19 | <TabAtkins_> | "aeffgd" |
| 07:19 | <TabAtkins_> | "aehhhd" |
| 07:19 | <TabAtkins_> | 20% 20% 10% 10% 20% 20%); |
| 07:19 | <TabAtkins_> | (Hopefully you have a fixed-width font.) |
| 07:20 | TabAtkins_ | heads out to shower for a bit. He'll be back in about half an hour. |
| 07:22 | <boblet> | TabAtkins_: thanks for the help :) I’ll see if I can work out why flexbox isn’t doing what I want, and pseudo-code the same example in template and grid positioning CSS |
| 07:23 | <nimbupani> | boblet: http://peter.sh/examples/?/css/flexbox.html |
| 07:30 | <MikeSmith> | anybody here a V8 contributor or committer? |
| 07:32 | <boblet> | nimbupani: :p |
| 07:32 | <nimbupani> | :) |
| 07:32 | <nimbupani> | it has working examples! |
| 07:32 | <nimbupani> | that work in IE9! |
| 07:33 | <boblet> | nimbupani: “TabAtkins_: Alex Mogilevsky is implementing experimental support for Flexbox in the platform previews, but I have no clue if that'll remain in the final build or not.” |
| 07:33 | <nimbupani> | :| |
| 07:35 | <jarib> | MikeSmith: try #v8 |
| 07:36 | <MikeSmith> | jarib: hai |
| 07:36 | MikeSmith | tries now |
| 07:36 | <hsivonen> | I wonder how the HTML WG schedule will work |
| 07:36 | <boblet> | bbiab |
| 07:37 | <hsivonen> | and whether there will be filibuster to trigger Consequences |
| 07:37 | <othermaciej> | what sort of abusive filibuster do you envision? |
| 07:38 | <hsivonen> | othermaciej: the kind that keeps bugs open on Dec 8 to autoescalate to ISSUEs |
| 07:39 | <othermaciej> | e.g. by abusively chain-reopening a bug? |
| 07:40 | <othermaciej> | I think the chairs would put a stop to that, but it also seems like a waste of time compared to marking the bug TrackerRequest |
| 07:55 | <TabAtkins_> | boblet: My flex algorithm computes flexes top-down, rather than bottom-up, which I think should make the OOCSS example work correctly. |
| 07:56 | TabAtkins_ | 's flex algo from his xanthir.com draft should be ported to the spec draft when he does the rewrite. |
| 08:10 | <zcorpan_> | http://www.w3.org/Bugs/Public/show_bug.cgi?id=10556 lol |
| 08:10 | <TabAtkins_> | boblet: Urf. I just looked at the actual source for the AG example with flexbox/OOCSS. Five levels of nesting for two of the boxes. That's as horrid and unmaintainable as table layouts. |
| 08:13 | <annevk> | zcorpan_, also yawn |
| 08:13 | <zcorpan_> | why do we have a registry for canvas contexts anyway? |
| 08:14 | <zcorpan_> | it's not clear to me that it's needed |
| 08:14 | <annevk> | so you can find the contexts defined elsewhere |
| 08:14 | <TabAtkins_> | zcorpan_: Mainly to document what contexts can be used together. |
| 08:14 | <annevk> | and how they work |
| 08:15 | <zcorpan_> | i would use a search engine to find contexts defined elsewhere |
| 08:16 | <zcorpan_> | the contexts we have today can't be used together |
| 08:16 | <TabAtkins_> | Correct. But we kinda expect that to change in the future. |
| 08:16 | <annevk> | whether or not they can be used together does not really matter |
| 08:17 | <annevk> | it's a string that independent bodies can put a claim on and as such a registry is useful |
| 08:17 | <TabAtkins_> | annevk: ?_? It certainly does matter! Or do you mean that it doesn't matter for the registry itself? |
| 08:17 | <annevk> | the latter |
| 08:18 | <TabAtkins_> | I can assure you that the idea of the registry was decided on as a consequence of us talking about the best method to indicate which contexts can be used together. ^_^ |
| 08:18 | <othermaciej> | certainly it would be odd if the registry had a pairwise "can be used together" flag for every pair of contexts |
| 08:18 | <othermaciej> | at least if we ever get to a large number |
| 08:18 | <othermaciej> | which I am not sure will ever be the case |
| 08:19 | <annevk> | I hope not |
| 08:19 | <othermaciej> | but still, O(N^2) and all |
| 08:19 | <TabAtkins_> | Use a table! |
| 08:19 | <Hixie> | it's just a list of disjoint sets |
| 08:20 | <othermaciej> | it would be fortunate if it ends up so well-structured a relation |
| 08:21 | <Hixie> | it has to be a list of disjoint sets by definition |
| 08:21 | <Hixie> | i really don't understand the problem here :-) |
| 08:21 | <TabAtkins_> | It has to be a reflexive, transitive relationship to make any sense at all, so yeah, what Hixie said. |
| 08:23 | <TabAtkins_> | (Non-reflexive uses like a 3d context using a 2d context as a texture are not using the two contexts at the same time. They're two separate canvases, one of which accepts the other somehow in a method call or something.) |
| 08:23 | <othermaciej> | I wouldn't assume transitivity a priori but I grant it's not completely impossible |
| 08:23 | <othermaciej> | s/impossible/implausible/ |
| 08:24 | <Hixie> | it would be a pretty terrible API if the API worked differently based on what order you obtained the contexts in |
| 08:24 | <othermaciej> | that is true to some extent if any two can't be used together |
| 08:27 | <TabAtkins_> | Only insofar as grabbing from one set means you won't fail until you make a call to a context from a different set. |
| 08:27 | <zcorpan_> | Hixie: http://www.whatwg.org/specs/web-apps/current-work/complete/ gives an index listing |
| 08:27 | <TabAtkins_> | Which is a very minor sense of "[works] differently". |
| 08:28 | <Hixie> | zcorpan_: will fix |
| 08:33 | <Hixie> | zcorpan_: it's annevk's fault |
| 08:33 | <zcorpan_> | orly |
| 08:35 | <Hixie> | his server is returning data so slowly i'm getting estimates of 6 hours to get back the complete.zip tarball |
| 08:35 | <Hixie> | er, zip file |
| 08:35 | <Hixie> | whatever |
| 08:36 | <Hixie> | i screwed up the complete/ directory trying to kill the pending wgets :-) |
| 08:37 | <jgraham> | So the new timeline is that we are almost 9 months from LC? |
| 08:38 | <Hixie> | first LC |
| 08:38 | <jgraham> | Sure |
| 08:41 | <zcorpan_> | hmm an example with unbeforeunload in the spec |
| 08:41 | <annevk> | my fault huh |
| 08:42 | <annevk> | html5.org is fast enough for me -- weird |
| 08:42 | <Hixie> | maybe complete.zip is huge? |
| 08:42 | <Hixie> | i didn't investigate closely |
| 08:43 | <boblet> | TabAtkins_: thanks for your comments. really pleased about the proposed change to top-down for Flexbox, as that’d really improve things. re: AG test yeah it’s an extreme example, but atm there’s no way to make it better — rock on Template and/or Grid Positioning! |
| 08:57 | <zcorpan_> | Hixie: http://www.w3.org/Bugs/Public/show_bug.cgi?id=8966 doesn't link to the diff |
| 08:57 | <Hixie> | it happens |
| 08:58 | <Hixie> | sometimes my script fails and i'm so many bugs ahead of the script by that time that i don't notice |
| 08:58 | <Hixie> | other times i forget to paste in the bug number |
| 08:58 | <Hixie> | if you find the revision feel free to paste in a link |
| 08:58 | <zcorpan_> | done |
| 08:58 | <annevk> | whoa, fixing lots of bugs are we now? :) |
| 08:59 | <annevk> | I wish people paid that amount of attention to stuff I wrote |
| 08:59 | <Hixie> | zcorpan_: thanks |
| 08:59 | <Hixie> | annevk: yeah see comments earlier from me about where i stand with respect to e-mails andstuff |
| 08:59 | <Hixie> | plus the chairs just told me i had to be at zarro boogs by oct 1st or something |
| 09:00 | <Hixie> | or nov 8th |
| 09:00 | <Hixie> | i forget the exact date |
| 09:00 | <Hixie> | is there a way to get tracker to show more revs on the index page? |
| 09:00 | <zcorpan_> | ?limit=xyz |
| 09:00 | <zcorpan_> | where xyz is a number... |
| 09:01 | <Hixie> | aha, cool |
| 09:01 | <Hixie> | thanks |
| 09:03 | <annevk> | or -1 for everything |
| 09:03 | <annevk> | which is very slow |
| 09:09 | <annevk> | whoa, someone on the charset list actually suggested updating shift_jis |
| 09:14 | <hsivonen> | annevk: speaking of shift_jis, the answer I got to my question if the Windows-31J tables should be used for ISO-2022-JP and EUC-JP was a "yes" |
| 09:14 | <hsivonen> | (maybe this was news only to me) |
| 09:15 | <hsivonen> | that is, you apply the same decoding delta that you apply to de jure Shift_JIS to make it Windows-31J |
| 09:15 | <hsivonen> | *decoding table delta |
| 09:17 | <annevk> | yeah, except Windows-31J is just a name IANA invented, it is not actually an alias you should recognize (nor does Windows) |
| 09:19 | <hsivonen> | well, hooray to IANA |
| 09:19 | <annevk> | Windows calls it windows-932 I believe |
| 09:19 | <annevk> | or just shift_jis |
| 09:19 | <hsivonen> | well, you apply the delta that makes shift_jis de jure tables into windows-932 tables |
| 09:20 | <annevk> | once Opera implements more Firefox-like encoding matching and I have some more time I was hoping to put some effort into it again |
| 09:20 | <annevk> | I think we should just have one spec you give an alias and a byte stream and it returns a Unicode stream |
| 09:21 | <annevk> | it's all legacy anyway |
| 09:21 | <annevk> | so we just figure out what the various mappings ought to be, write it down, and done.. or some such |
| 09:21 | <Hixie> | could someone sanity check me on http://www.w3.org/Bugs/Public/show_bug.cgi?id=10479 and if I'm right explain it to steven? |
| 09:21 | <annevk> | (also Unicode + alias -> byte stream |
| 09:21 | <annevk> | ) |
| 09:23 | <hsivonen> | the weirdest part is that on OS/2, Gecko uses the IBM delta for shift_jis instead of the Microsoft delta |
| 09:24 | <hsivonen> | I think that's bogus, but it's not a high-priority battle for me |
| 09:24 | <zcorpan_> | does anyone use OS/2? |
| 09:24 | <annevk> | Hixie, I guess the user would still be able to expand the details (ARIA doesn't affect that) but the benefits would only be for AT users? (if there are benefits to using <fieldset> that is...) |
| 09:24 | <hsivonen> | zcorpan_: I believe there are OS/2 users, yes |
| 09:25 | <Hixie> | annevk: how would an AT user expand the details if you overrode the accessibility API mapping of <details>? |
| 09:25 | zcorpan_ | thought OS/2 was about as dead as windows 3.1 |
| 09:26 | <hsivonen> | Hixie: I think the right thing is to let details retain its natural role and have a div inside with role=radiogroup |
| 09:26 | <annevk> | Hixie, if "search" is allowed does that not override it as well? |
| 09:26 | <Hixie> | hsivonen: agreed, but can you explain that to steven? |
| 09:26 | <Hixie> | annevk: ? |
| 09:26 | <hsivonen> | since presumably, you don't want details to lose its default expanding nature |
| 09:26 | <annevk> | " |
| 09:26 | <annevk> | Currently the details element has a default ARIA role =group and overrides |
| 09:26 | <annevk> | allowed: Role must be either form, group, navigation, note, or search" |
| 09:26 | <hsivonen> | Hixie: feel free to copy and paste from IRC |
| 09:27 | <annevk> | so do any of those roles not override its native semantics? |
| 09:27 | <Hixie> | hsivonen: i said what you said but he didn't seem to understand it |
| 09:27 | <Hixie> | i was hoping someone else could find another way to explain it :-) |
| 09:27 | <Hixie> | annevk: yes, that sounds like a bug |
| 09:27 | <annevk> | maybe that is why he is confused |
| 09:28 | <hsivonen> | hmm. good question about landmarks overriding native semantic |
| 09:31 | <annevk> | added comments |
| 09:32 | <annevk> | oh god |
| 09:32 | <annevk> | 3023bis is going to require XPointer |
| 09:32 | <hsivonen> | NOOOO! :-) |
| 09:33 | <annevk> | i wonder if IETF is just gonna approve that as RFC without checking implementations again |
| 09:33 | <annevk> | same with this whole IDNA crap |
| 09:33 | <hsivonen> | what I want from 3023bis is text/xml not to default to US-ASCII and not be required to have CRLF breaks |
| 09:34 | <annevk> | paraphrasing: "there was no IETF consensus to consider the needs of web browsers when it comes to IDNA handling" |
| 09:34 | <annevk> | actually |
| 09:34 | <annevk> | "there was IETF consensus to not consider them" was more like it |
| 09:34 | <annevk> | hsivonen, and also not be obsolete |
| 09:35 | <annevk> | look http://www.w3.org/2010/web-notifications/ Anne the Chair! ;p |
| 09:36 | <Hixie> | my condolences |
| 09:37 | <othermaciej> | annevk: welcome to the loser club |
| 09:37 | <annevk> | heh |
| 09:39 | <hsivonen> | annevk: congrats/condolences |
| 09:39 | <hsivonen> | nice background |
| 09:39 | <zcorpan_> | annevk: i'm disappointed that the background image isn't svg |
| 09:39 | <annevk> | MikeSmith can take credit for the design |
| 09:40 | <MikeSmith> | I tried to make it look really obnoxious |
| 09:40 | <MikeSmith> | but I seem to have failed at that |
| 09:41 | <MikeSmith> | anyway, annevk, indeed |
| 09:41 | <MikeSmith> | congrats and welcome to the home of the blues |
| 09:42 | <zcorpan_> | what's up with <div class="box"> ? isn't that a bit 2004? |
| 09:42 | <MikeSmith> | I'm a simple person and I like simple class values |
| 09:43 | <zcorpan_> | i mean as opposed to styling <body> |
| 09:43 | <MikeSmith> | I think i will change that to class=div |
| 09:43 | <MikeSmith> | styling body would be too efficient |
| 09:43 | <MikeSmith> | too obvious |
| 09:43 | <MikeSmith> | I don't like to do the obvious thing |
| 09:43 | <zcorpan_> | maybe you should use a <table> instead |
| 09:44 | <MikeSmith> | now you're talkin' |
| 09:44 | <annevk> | zcorpan_, you have to appreciate the fine line MikeSmith walks between keeping his job and not getting sacked for producing crap markup ;p |
| 09:44 | <MikeSmith> | my markup is an expression of my soul |
| 09:45 | <annevk> | haha |
| 09:45 | <MikeSmith> | when you criticize my markup, you wound me emotionally |
| 09:45 | <Hixie> | annevk: so you're gonna make sure there's no user permissions dialogs or infobars right? :-) |
| 09:46 | <annevk> | Hixie, I already tried that during discussion but I did not quite succeed I think |
| 09:46 | <zcorpan_> | Hixie: you get a notification instead |
| 09:46 | <Hixie> | annevk: you're the chair! |
| 09:46 | <Hixie> | annevk: lay the smack down |
| 09:46 | <Hixie> | annevk: anyone who thinks we should have a user permissions dialog, ban from the group for a week |
| 09:46 | <Hixie> | that'll show them who's boss |
| 09:46 | <annevk> | Hixie, that includes all your colleagues :) |
| 09:47 | <annevk> | Hixie, I think most people want to follow what geolocation has |
| 09:47 | <Hixie> | many of my colleagues are new to web browsers |
| 09:48 | <Hixie> | don't let them bully you into bad UI :-) |
| 09:48 | <jgraham> | "lay the smack down" — are you advocating drugging the WG? |
| 09:48 | <annevk> | Hixie, I tried advocating your model which I think is much nicer, but it did not work out for reasons I cannot recall right now |
| 09:48 | <annevk> | Hixie, the app people wanna know upfront whether it is going to work as notifications often serve as reminders |
| 09:49 | <zcorpan_> | [[ This page wants to know where you are and wants to show you notifications. ( Allow location and notification ) ( Allow location but not notification ) ( Allow location but not location ) ( Allow neither location or notification ) [ ] Remember setting ( Learn more... ) ( Cancel ) ]] |
| 09:49 | <annevk> | Hixie, of course they could show an example one, but that is not quite nice; also, they wanted to be able to combine multiple of these permission dialogs together into something like "grant app priveliges" which I think makes sense |
| 09:49 | <MikeSmith> | annevk: you need to cultivate this kind of any image now: http://www.cinemaretro.com/uploads/walking2.jpg |
| 09:50 | <MikeSmith> | http://www.cinemaretro.com/uploads/walking.jpg |
| 09:50 | <MikeSmith> | Walk Tall |
| 09:51 | <zcorpan_> | s/location but not location/notification but not location/ |
| 09:51 | <jgraham> | zcorpan_: Was that a serious suggestion? |
| 09:51 | <Hixie> | annevk: a "grant app priveliges" [sic] dialog is a terrible idea |
| 09:51 | <Hixie> | annevk: users have no idea what that even means |
| 09:52 | <hsivonen> | are there non-Google "app people" in the group? |
| 09:52 | hsivonen | is too lazy to read the member list |
| 09:52 | <zcorpan_> | jgraham: not really, but it'll end up something like that |
| 09:52 | <annevk> | Hixie, so how do you think the geolocation bar works? |
| 09:52 | jgraham | predicts users will click whatever looks most like the "Fuck off and let me use the website" button |
| 09:52 | <annevk> | Hixie, do users understand that? |
| 09:52 | annevk | somewhat doubts they get all the implications |
| 09:52 | <Hixie> | the geolocation bar is a failure imho |
| 09:52 | <jgraham> | In fact maybe there should just be one button and it should be labeled thus |
| 09:53 | <hsivonen> | Hixie: how would you design the geolocation authorization? |
| 09:53 | <Hixie> | i have no idea |
| 09:53 | <Hixie> | hence my not whining about it :-) |
| 09:53 | <Hixie> | but the notification one is far easier |
| 09:53 | <Hixie> | have ui on the notifications to opt-in to having them global, and when they're not opted in, only show them within the viewport, not the desktop |
| 09:54 | <Hixie> | (and have similar ui to opt-out) |
| 09:54 | <Hixie> | and maybe have a special method call you can invoke that pops up one of these widgets with an arrow saying "click here to opt in" (translated into userese) |
| 09:54 | <Hixie> | or some such idea |
| 09:55 | <zcorpan_> | or allow the user to drag the notification |
| 09:57 | <annevk> | Hixie, yeah, though people had concerns about it; would be good if once the joining process is done at the W3C side you just joined the group |
| 09:58 | <Hixie> | remind me when i can |
| 09:58 | <Hixie> | if it's low traffic i can watch it |
| 09:58 | <annevk> | I hope it's low traffic |
| 09:59 | <annevk> | Hixie, you can just subscribe to public-web-notification⊙wo and not bother joining the group; Google will have to join regardless for the editor |
| 10:03 | <Hixie> | k |
| 10:05 | <hsivonen> | how does Opera enforce Same Origin for file: URLs? |
| 10:06 | <annevk> | pm'd you someone's email |
| 10:07 | <MikeSmith> | http://twitter.com/brianleroux/status/23898902961 @brianleroux "way to go growl! you've spawned a w3c working group" |
| 10:08 | <Hixie> | ok bed time |
| 10:08 | <Hixie> | nn |
| 10:10 | <zcorpan_> | http://www.w3.org/mid/581CCCEE-56A0-4C17-BECC-ABC15E4065B0⊙ac lol |
| 10:11 | <annevk> | the HTML variant of that email was more funny |
| 10:12 | <zcorpan_> | i don't see any HTML variant of that email |
| 10:12 | <zcorpan_> | although i got the attachment in the right place in M2 |
| 10:13 | <annevk> | http://blogs.msdn.com/b/ie/archive/2010/09/07/transitioning-existing-code-to-the-es5-getter-setter-apis.aspx great how the "legacy" API is shorter and easier |
| 10:15 | <zcorpan_> | i thought you could do Object.defineProperty(myObject,"p", |
| 10:15 | <zcorpan_> | {get: function() {/* getter function body */}, set: function(v) {/* setter function body */}} |
| 10:15 | <zcorpan_> | ); |
| 10:15 | <annevk> | you can |
| 10:15 | <annevk> | but there's an equivalent of that as well |
| 10:16 | <annevk> | not sure whether it was legacy |
| 10:30 | <annevk> | via bruce: http://isyourmom.partofhtml5.com/ |
| 10:33 | <hsivonen> | http://istheblinktag.partofhtml5.com/? shows a bug in Gecko :-( |
| 10:34 | <zcorpan_> | http://dhtml5.com/ |
| 11:27 | <othermaciej> | insomnia qq |
| 11:57 | <MikeSmith> | othermaciej: :( |
| 12:13 | <Workshiva> | I like this one: http://ishtml5.partofhtml5.com/ |
| 12:28 | <hsivonen> | FAIL. http://isvideo.partofhtml5.com/ is H.264 only |
| 12:30 | <annevk> | defining the nodes model was easier than expected |
| 12:51 | <annevk> | zcorpan_, your insertBefore fu is sort of broken |
| 12:51 | <annevk> | e.g. it doesn't catch the case where the node being inserted is the parentNode |
| 12:51 | <annevk> | (or some other ancestor) |
| 12:52 | <annevk> | and it doesn't remove the node from the other location, I guess adoptNode always has to be implicitly invoked |
| 12:54 | <zcorpan_> | annevk: ok |
| 12:56 | <annevk> | no, not ok |
| 12:56 | <annevk> | :) |
| 12:56 | <annevk> | I guess I'll write some tests for all the modifying the DOM tree stuff |
| 12:57 | <annevk> | and study the original DOM spec and impls some more |
| 12:57 | <annevk> | with the notable exception of mutation events, as I'm not going there for now |
| 12:59 | <zcorpan_> | i like <@philip> |
| 13:02 | <annevk> | yay: http://www.nu.nl/internet/2329069/europarlement-dreigt-anti-piraterijverdrag-blazen.html |
| 13:03 | <Peter`> | wrong channel? |
| 13:05 | <annevk> | see topic |
| 13:05 | <annevk> | :) |
| 13:05 | <Peter`> | pff :p |
| 13:05 | <annevk> | #whatwg is a convenient way to blog |
| 13:06 | <zcorpan_> | if the registries are intended to match reality, maybe they should be automated based on searches on actual content instead of maintained manually |
| 13:06 | <annevk> | and store bits of info |
| 13:06 | <Peter`> | fair enough |
| 13:37 | <zcorpan_> | 0 unread public-html emails |
| 13:43 | <jgraham> | zcorpan_: How many did you actually read though? |
| 13:43 | jgraham | has 7770 unread public-html emails but doesn't mark emails read that he has not actually read |
| 13:46 | <jcranmer> | I ignore most of the emails I get anyways |
| 13:46 | hsivonen | notes that there hasn't been a use case presented you for using the script global object of the evaluation-time document |
| 13:46 | <zcorpan_> | jgraham: i partially read the subject line from all threads at least |
| 13:47 | <hsivonen> | s/you// |
| 13:48 | <jgraham> | zcorpan_: Partially reading the subject line once per thread doesn't count :) |
| 13:48 | <zcorpan_> | why not? |
| 13:50 | <jgraham> | It doesn't impart a sufficient fraction of the information content of the mails for me to consider it "read" |
| 13:50 | <jcranmer> | I generally wait to read it until I'm right about to delete it |
| 13:52 | <zcorpan_> | ok then i probably read somewhere around between 5% and 10% of the emails |
| 14:42 | <hsivonen> | does JS have a way for an anonymous function object to reference itself (as opposed to referencing the object it is a property of) |
| 14:44 | <smaug____> | hsivonen: https://developer.mozilla.org/en/JavaScript/Reference/Functions_and_function_scope/arguments/callee |
| 14:44 | <smaug____> | not sure if that is deprecated |
| 14:45 | <zcorpan_> | iirc it isn't (but arguments.caller is) |
| 14:49 | <jgraham> | hsivonen: You are supposed to do function foo() {foo()} |
| 14:49 | <jgraham> | which works even if foo is a functionexpression rather than a FunctionDeclaration |
| 14:50 | <zcorpan_> | what's a functionexpression? |
| 14:51 | <hsivonen> | smaug____: thanks. I decided to do something less fancy to avoid footguns, though. |
| 14:52 | <hsivonen> | jgraham: I don't understand. My use case was an anonymous event handler removing itself after running once |
| 14:57 | <jgraham> | A function expression is something like var foo = function X() {} |
| 14:58 | <jgraham> | Then inside the function you can use X to refer to the function |
| 14:58 | <jgraham> | outside the function you use foo() |
| 14:59 | <zcorpan_> | wow i didn't know that was even allowed |
| 14:59 | <gsnedders> | zcorpan_: n00b |
| 15:01 | <hsivonen> | jgraham: ooh. good to know. thanks |
| 15:19 | <FastJack> | heh, it's funny. I just thought about bringing back <img lowsrc> for mobile devices :) |
| 15:20 | <hsivonen> | over here, mobile devices have faster connections that desktops in some other places |
| 15:20 | <hsivonen> | I wouldn't want my mobile device downloading uglier images than my desktop |
| 15:22 | <FastJack> | you don't have to all the time. but sometimes only edge or gprs are available |
| 15:28 | <zcorpan_> | opera mini/turbo uglifies images for you for slow connections |
| 15:28 | <hendry> | where is the svn interface to HTML5? http://svn.whatwg.org/ doesn't seem to have it |
| 15:29 | <hendry> | i'm interested in running my own "Specification annotation system" |
| 15:29 | <hsivonen> | hendry: this one? http://html5.org/tools/web-apps-tracker?from=5299&to=5300 |
| 15:30 | <hendry> | hsivonen: i want to checkout the HTML5 spec, not a diff interfac |
| 15:30 | <zcorpan_> | hendry: svn.whatwg.org is the svn interface |
| 15:30 | <hsivonen> | hooray. Gecko fails some html5lib tests, because the MathML implementation maintains private data in script-visible _moz- attributes |
| 15:31 | <hendry> | zcorpan_: http://svn.whatwg.org/webapps/complete.html doesn't look right ... ? |
| 15:32 | <zcorpan_> | how doesn't it look right? |
| 15:33 | <annevk> | hendry, it's composed via script |
| 15:39 | <hsivonen> | does DOM TreeWalker try to be fancy about munging white space on the text nodes it exposes? |
| 15:40 | <hsivonen> | aaargh! |
| 15:40 | <hsivonen> | data: URLs!1!! |
| 15:41 | <jgraham> | hsivonen: What is the problem? |
| 15:41 | <hsivonen> | output from the test harness doesn't match expected output on white space details |
| 15:41 | <hendry> | annevk: sorry, getting interrupted, and where is the script again? |
| 15:42 | <hsivonen> | I think the data: URL creation step throws away some unescaped white space |
| 15:42 | <jgraham> | hsivonen: example? |
| 15:42 | <annevk> | hendry, I don't think it's public |
| 15:42 | <jgraham> | hsivonen: Of the input that causes problems |
| 15:42 | <jgraham> | I want to check that my harness doesn't have the same problems |
| 15:42 | <hendry> | annevk: iirc hsivonen showed me it sometime ago. can't recall |
| 15:42 | <hsivonen> | jgraham: http://pastebin.mozilla.org/785408 |
| 15:43 | <annevk> | hendry, well, there's http://pimpmyspec.net/ |
| 15:43 | <annevk> | hendry, but I believe there is some preprocessing going on as well |
| 15:44 | <hendry> | annevk: be good to see how it works. esp. annotation. I'll write an email to Hixie :) |
| 15:45 | <zcorpan_> | hendry: by annotation do you mean the status boxes in the margin? |
| 15:46 | <hsivonen> | here's my problem: data:text/html,<pre>FOO%A0BAR</pre> |
| 15:49 | <hsivonen> | oops. |
| 15:49 | <hsivonen> | data:text/html,<pre>FOO%0ABAR</pre> |
| 15:49 | <hsivonen> | that's better |
| 16:09 | <hendry> | zcorpan_: yes, that too |
| 16:12 | <zcorpan_> | hendry: they're mostly done with javascript |
| 16:39 | <zcorpan_> | annevk: Node.removeNode() seems nice |
| 16:40 | <zcorpan_> | annevk: except it doesn't seem to do what i expect |
| 16:41 | <annevk> | removeNode? |
| 16:42 | <Workshiva> | Node.prototype.removeNode = function() { if (this.parentNode) this.parentNode.removeChild(this); }; |
| 16:42 | <annevk> | oh, no arguments |
| 16:42 | <zcorpan_> | Workshiva: but that's not what it does in opera |
| 16:43 | <zcorpan_> | http://software.hixie.ch/utilities/js/live-dom-viewer/saved/621 |
| 16:43 | <zcorpan_> | also, it's on Element, not Node |
| 16:43 | <Workshiva> | Why? |
| 16:43 | <zcorpan_> | see topic |
| 16:45 | <zcorpan_> | ah removeNode takes a bool argument |
| 16:45 | <annevk> | by default it does not remove the children |
| 16:45 | <zcorpan_> | false inserts the element's children in its place |
| 16:46 | <annevk> | i rather have library authors do this kind of stuff |
| 16:47 | <zcorpan_> | i would have liked it if it was on Node instead and no argument meant true instead of false |
| 16:47 | <zcorpan_> | http://google.com/codesearch?as_q=%5C.removeNode%5C%28 |
| 16:47 | <annevk> | or no argument support at all |
| 16:48 | <zcorpan_> | http://google.com/codesearch?q=%5C.removeNode%5C(++lang:javascript |
| 16:49 | <zcorpan_> | maybe those aren't using the native removeNode |
| 16:49 | <annevk> | seem like custom or IE-only impl |
| 16:49 | <zcorpan_> | does ie support this? |
| 16:49 | <annevk> | think so |
| 16:50 | <annevk> | gotta run |
| 17:38 | <zcorpan_> | what's the impl status of https://cvs.khronos.org/svn/repos/registry/trunk/public/webgl/doc/spec/TypedArray-spec.html ? |
| 17:44 | <zcorpan_> | <http://www.w3.org/mid/9B2DE9094C827E44988F5ADAA6A2C5DACD3E5C⊙Hppc>; - "The AssocSS spec has been approved to go out for review as a PER |
| 17:44 | <zcorpan_> | with the expected publication date being tomorrow." |
| 17:48 | <zcorpan_> | annevk: don't you need a function wrapper here? assert_throws(new TypeError(), document.body.appendChild({'a':'b'})) |
| 17:50 | <jgraham> | yes |
| 17:52 | <zcorpan_> | annevk: doctypes are special. deal with it :) |
| 19:53 | <Hixie> | hober: thanks dude |
| 20:42 | <hober> | np |
| 21:38 | <annevk> | meh |
| 21:38 | <annevk> | insertBefore et al need to use adoptNode |
| 21:38 | <annevk> | but adoptNode throws for DocumentType |
| 21:39 | <gsnedders> | Who said DOM Core was nice? |
| 21:39 | <annevk> | Opera does not support DocumentType in this way it seems though |
| 21:39 | <annevk> | maybe we should drop support for it apart from document.doctype |
| 21:40 | <annevk> | oh I see simon said "<zcorpan_> annevk: doctypes are special. deal with it :) " |
| 21:42 | <annevk> | i guess i could define how it works now and then ask people to make it simpler |
| 22:05 | <annevk> | I wonder if Gecko/WebKit are willing to nuke inputEncoding |
| 22:05 | <annevk> | clearly charset, characterSet, and defaultCharset is enough for everyone :) |
| 22:06 | <zcorpan_> | annevk: have you covered this case? foo.insertBefore(foo, other) |
| 22:08 | <annevk> | it does not cover several cases atm, but thanks for mentioning that one |
| 22:12 | <annevk> | also applicable to appendChild/replaceChild |
| 22:13 | <zcorpan_> | yes |
| 22:14 | <annevk> | I wonder if the null/null passing should instead through some kind of ECMAScript error |
| 22:14 | <annevk> | types are not nullable if they do not have a trailing question mark so I guess Web IDL covers that and I do not have to |
| 22:14 | <annevk> | but then the question is how stable Web IDL is |
| 22:14 | <annevk> | when is cameron starting at Moz again? |
| 22:15 | <paul_irish> | cameron who? |
| 22:16 | <annevk> | Mr. Web IDL: http://dev.w3.org/2006/webapi/WebIDL/ |
| 22:16 | <TabAtkins> | annevk: I do my best to make people aware of WebIDL when they're designing APIs around here, so they just consistently defer to the spec in matters like this. |
| 22:17 | <paul_irish> | ah |
| 22:18 | <annevk> | so I guess that ought to be a type error too then |
| 22:19 | <annevk> | jgraham, new TypeError() does not work |
| 22:20 | <annevk> | I'm getting |
| 22:20 | <annevk> | Error: invalid array length |
| 22:20 | <annevk> | Source File: file:///Users/annevk/Work/w3c-dvcs/web-dom-core/tests/testharness.js |
| 22:20 | <annevk> | Line: 802 |
| 22:20 | <TabAtkins> | Has Chrome removed validation from the unimplemented input types yet? I thought we had somewhere above the public channel, but I'm suddenly struck by doubt. |
| 22:20 | smaug____ | thinks he wanted to change EventTarget handling in WebIDL |
| 22:20 | <annevk> | that reminds me, is EventTarget already defined in such a way that it no longer assumes nodes? |
| 22:21 | <smaug____> | I think so |
| 22:21 | <smaug____> | when did it assume nodes? |
| 22:21 | <annevk> | just in describing most things |
| 22:27 | <AryehGregor> | TabAtkins, pretty sure it has, but it's a one-line test, no? |
| 22:27 | <TabAtkins> | If I know which version has it turned off, sure. |
| 22:32 | <annevk> | btw, I killed all <table>s from Web DOM Core |
| 22:39 | <zcorpan_> | annevk: http://html5.org/tools/web-apps-tracker?from=5432&to=5433 doesn't show the commit message if i have 'show editorial changes' unchecked |
| 22:39 | <jgraham> | annevk: What actual code is causing the error |
| 22:39 | <jgraham> | It isn't new TypeError(0 :) |
| 22:39 | <jgraham> | s/0/) |
| 22:41 | <annevk> | jgraham, I thought you said that was it |
| 22:41 | <annevk> | zcorpan_, is the latest code not in subversion? |
| 22:42 | <zcorpan_> | annevk: what? |
| 22:42 | <annevk> | zcorpan_, usually you just commit fixes and I propagate |
| 22:43 | <zcorpan_> | ah. i haven't checked |
| 22:43 | <jgraham> | annevk: I am very confused/ |
| 22:43 | <jgraham> | What did I say? |
| 22:43 | <annevk> | to use new TypeError() |
| 22:44 | <jgraham> | Right |
| 22:44 | <annevk> | i.e. |
| 22:44 | <annevk> | assert_throws(new TypeError(), function() { document.body.insertBefore({'a':'b'}, document.body.firstChild) }) |
| 22:44 | <annevk> | but if I use that things fall apart |
| 22:44 | <jgraham> | OK, that is now enough for me to debug :) |
| 22:44 | <jgraham> | thanks |
| 22:44 | <jgraham> | (is this in the repo somewhere?) |
| 22:45 | <annevk> | if you take http://bitbucket.org/ms2ger/web-dom-core/src/tip/tests/Node-insertBefore.html |
| 22:45 | <annevk> | and change the first NOT_SUPPORTED_ERR to new TypeError() |
| 22:45 | <annevk> | you'll get it |
| 22:51 | <Hixie> | is it bad that i have no idea what i was talking about in http://www.w3.org/Bugs/Public/show_bug.cgi?id=10240 |
| 22:51 | Hixie | starts rereading the algorithm in question |
| 22:51 | <Hixie> | i'm often irritated that people file the most terse bugs and i often can't work out wtf they're talking about, but if even I'm doing it... |
| 22:52 | TabAtkins | initially read that as "irradiated". |
| 22:52 | <Hixie> | oh, i know what i meant |
| 22:53 | <zcorpan_> | we should get rid of xml:id... |
| 22:53 | <zcorpan_> | (in opera) |
| 22:54 | <annevk> | yes |
| 22:54 | <MikeSmith> | zcorpan_: yeah |
| 22:54 | <annevk> | i had this "radical" idea of introducing a global id and class attribute in DOM Core |
| 22:54 | <MikeSmith> | heh |
| 22:55 | <MikeSmith> | breaking new ground |
| 22:55 | <annevk> | it's a pretty neat idea I think |
| 22:55 | <MikeSmith> | scare quotes implied |
| 22:55 | <MikeSmith> | what's so neat about it? |
| 22:55 | <annevk> | so my old phone is kind of broken |
| 22:55 | <jgraham> | annevk: Another instance of the same bug. I really need to sit down and work out how to stop this happening |
| 22:56 | <MikeSmith> | annevk: btw, I think I remembered cam saying he was starting in October |
| 22:56 | <annevk> | the zero key does mostly not work and getting it on at all is problematic |
| 22:56 | <annevk> | and my passcode is four zeros |
| 22:56 | <annevk> | so screwed |
| 22:56 | <annevk> | MikeSmith, cool |
| 22:56 | <zcorpan_> | sounds like you should get a new phone |
| 22:56 | <Hixie> | where's cam starting in october? |
| 22:56 | <MikeSmith> | Mozilla |
| 22:57 | <Hixie> | cam as in heycam? |
| 22:57 | <annevk> | getting one tomorrow, but I was hoping to back up to my sim card one more time |
| 22:57 | <MikeSmith> | Hixie: yeah |
| 22:57 | <Hixie> | sweet! |
| 22:57 | <MikeSmith> | yup |
| 22:57 | <MikeSmith> | cam rocks mightily |
| 22:57 | <Hixie> | any idea if he's gonna be doing spec stuff? |
| 22:57 | <annevk> | MikeSmith, so it's cool because a) it doesn't bother the XML people that much and b) simplifies code as id/class handling will be the same for all elements everywhere |
| 22:57 | Hixie | stares longingly at webidl |
| 22:57 | <MikeSmith> | Hixie: he's going to start back up on WebIDL |
| 22:57 | <Hixie> | woot |
| 22:58 | <Hixie> | that's awesome |
| 22:58 | zcorpan_ | would like a bugzilla component for webidl |
| 22:58 | <MikeSmith> | The force of Cam will once again be felt around the standards world. |
| 22:58 | <Hixie> | zcorpan_: there is one |
| 22:58 | <zcorpan_> | oh. good then |
| 22:59 | <annevk> | i would like someone to make a decision on fricking null handling |
| 22:59 | <Hixie> | at least i'm pretty sure there is |
| 22:59 | <MikeSmith> | if there's not I can make one |
| 22:59 | <zcorpan_> | annevk: and too few arguments |
| 22:59 | <annevk> | null -> "" as default and undefined -> "undefined" (always?) is prolly best |
| 22:59 | <annevk> | zcorpan_, I guess that should throw |
| 23:00 | <zcorpan_> | annevk: some people prefer the webkit behavior since it's native JS behavior |
| 23:00 | <annevk> | jgraham, make the display code less magic |
| 23:00 | <TabAtkins> | I prefer webkit behavior. |
| 23:00 | <annevk> | zcorpan_, hmm fair enough |
| 23:00 | TabAtkins | commonly designs functions for himself which have omittable arguments on purpose, and expects the rest of the platform to follow suit. |
| 23:01 | <annevk> | it seems kind of silly though |
| 23:01 | <annevk> | if they are optional mark them as such |
| 23:01 | <annevk> | if they are not throwing seems better |
| 23:02 | <annevk> | e.g. xhr.open() makes no sense |
| 23:02 | <annevk> | or node.addEventListener() |
| 23:02 | <annevk> | (the third argument of that method should be optional though) |
| 23:02 | <TabAtkins> | Problem is that then you have silly people like Hixie who refuse to mark trailing boolean arguments as optional when they clearly should be. ^_^ |
| 23:02 | <TabAtkins> | (In the arcTo functions, specifically.) |
| 23:03 | <annevk> | you can override hixie by convincing all implementors of your solution |
| 23:03 | <annevk> | or just a couple |
| 23:03 | <TabAtkins> | Well, I've got webkit down. |
| 23:04 | <annevk> | that is the way I'm killing HTMLBaseFontElement :) |
| 23:04 | <jgraham> | annevk: Yeah, or less broken |
| 23:04 | <zcorpan_> | i like throwing better, personally, since it is easier to debug when you make a mistake and it can be easier to extend the spec in the future if people don't rely on missing arguments "working" |
| 23:04 | <jgraham> | anyway, more bandages applied |
| 23:04 | <jgraham> | in W3 repository |
| 23:04 | jgraham | hates having two copies of the file in different repositories |
| 23:04 | <zcorpan_> | but i don't care enough to fight over it |
| 23:05 | <jgraham> | since I just accidentially overwrote a copy with an older copy |
| 23:06 | <annevk> | would be neat if you could mark files as being from other repos |
| 23:06 | <annevk> | and hg would sort it all out |
| 23:06 | <annevk> | zcorpan_, I might |
| 23:06 | jgraham | likes throwing for too many arguments |
| 23:07 | <jgraham> | Not sure about too few |
| 23:07 | <annevk> | woot: http://w3future.com/2010/js1k.html |
| 23:07 | <annevk> | including music |
| 23:07 | <jgraham> | Certianly addEventListener should be changed to accept 2 arguments only |
| 23:08 | <zcorpan_> | jgraham: we have interop on too many arguments already |
| 23:08 | <Hixie> | Philip`: i don't understand http://lists.w3.org/Archives/Public/public-html/2008Apr/0770.html -- did you mean the radius to not be zero? Is the spec changed from what it was already? Did I miss that e-mail altogether? |
| 23:09 | <Hixie> | Philip`: oh, i see, you meant there to be an 'r' |
| 23:09 | <jgraham> | zcorpan_: What happens? I don't remember |
| 23:09 | <zcorpan_> | annevk: hmm, that crashed opera |
| 23:09 | <annevk> | oh, not here |
| 23:09 | <zcorpan_> | jgraham: excess arguments are ignored |
| 23:10 | <jgraham> | zcorpan_: Oh. Well that seems to suck for forward-compat |
| 23:10 | <jgraham> | or am I missing someting? |
| 23:10 | <annevk> | anyway, nn |
| 23:10 | <zcorpan_> | jgraham: yeah, it does |
| 23:10 | <jgraham> | Seems to suggest that too-few arguments should act like undefined as you would expect in js |
| 23:11 | <jgraham> | rather than throwing |
| 23:11 | <jgraham> | modulo special cases |
| 23:11 | <jgraham> | Not that I really like it |
| 23:11 | <jgraham> | but yeah, bed time |
| 23:11 | <zcorpan_> | nn |
| 23:41 | <Philip`> | Hixie: Yeah, there was meant to be an r there |
| 23:42 | <Hixie> | i commented on the relevant bug |
| 23:42 | <Hixie> | looks like i'd already fixed the spec though |
| 23:42 | <Philip`> | The spec was changed, though I forget what it was changed to |
| 23:42 | <Hixie> | it just compares the delta now |
| 23:43 | <Philip`> | and I think I need to concentrate a lot in order to understand it without getting myself mixed up |
| 23:43 | <Hixie> | as you increase endAngle (assuming endAngle > startAngle and the right direction is set) the arc can never shrink again |
| 23:44 | <Philip`> | What if it's the wrong direction? |
| 23:44 | <Hixie> | then it's the other way around |
| 23:45 | <Hixie> | bbiab |
