| 02:06 | <MikeSmith> | reading message about iterators in WebIDL |
| 02:07 | <MikeSmith> | http://lists.w3.org/Archives/Public/public-script-coord/2012AprJun/0239.html |
| 02:07 | MikeSmith | looks around for heycam |
| 02:07 | heycam | has replying to that on his todo list for today |
| 02:25 | <karlcow> | hmmm I was reading the TAG minutes |
| 02:25 | <karlcow> | http://www.w3.org/2001/tag/2012/06/13-minutes#item02 |
| 02:25 | <karlcow> | it seems there is a good will but completely missing the target on frustrations. |
| 02:50 | <tantek> | that's some dry reading karlcow |
| 03:07 | <nessy> | is whatwg.org down? |
| 03:10 | <miketaylr> | seems to be for me, nessy |
| 03:30 | <karlcow> | tantek: I would say completely humid. It is super hot and humid tonight in Montreal |
| 03:33 | <MikeSmith> | nessy, miketaylr, whatwg.org working for me |
| 03:34 | <miketaylr> | MikeSmith: yeah wfm now too |
| 03:34 | <MikeSmith> | k |
| 03:35 | <MikeSmith> | karlcow: good that at least there seems to be some clear recognition that "disconnect from community" |
| 03:35 | <MikeSmith> | but as far as "People don't care about architecture" and "Persuading community that arch. matters", I think that's the wrong way of looking at the problem |
| 03:36 | <MikeSmith> | whatwg and everybody working on HTML5 and APis care about architecture |
| 03:36 | <MikeSmith> | difference is that they become overly religious about architectural purity to the degree that it trumps everything else |
| 03:36 | <karlcow> | yes I have the same feeling, just not the same idea of what the architecture should be. |
| 03:37 | <MikeSmith> | I meant, difference is that whatwg et all *don't* become overly religious about architectural purity |
| 03:37 | <karlcow> | with on top of that, bruise and arguments between people which doesn't help. |
| 03:38 | <karlcow> | Not sure about purity. Each community is purist with regards to its own thought framework. :) |
| 03:38 | <MikeSmith> | karlcow: in the minds of some the idea of what good architecture is keeping everything cleanly separated, "modularized", extensible, all those same principles they want to elevate the importance of |
| 03:39 | <MikeSmith> | of course the problem is, the Web platform is not clean |
| 03:39 | <karlcow> | yup different ideas of what the architecture should be. :) |
| 03:39 | <MikeSmith> | it's messy and hard to separate out things neeetly |
| 03:39 | <MikeSmith> | and their are costs to modularization and extensibility, along with whatever benefits they have |
| 03:40 | <MikeSmith> | modularization can actually make it much harder to get good interoperability |
| 03:41 | <MikeSmith> | and adding extensibility hooks can make things way more complicated than the actual real-world use cases on the table require |
| 03:41 | <karlcow> | I think what the TAG could try would be in the sake of discussion is look at the vernacular architecture. Aka the shelters, farms, dirt walls, etc. and how it could help define something which encompasses more than one way of doing things. :) |
| 03:41 | <MikeSmith> | and extensibility itself is not a use case |
| 03:41 | <MikeSmith> | eh? |
| 03:41 | <MikeSmith> | dirt walls? |
| 03:41 | <karlcow> | let me find the right name |
| 03:42 | <karlcow> | https://en.wikipedia.org/wiki/Pise |
| 03:42 | <MikeSmith> | Web Sockets protocol is a classic case of architecture/extensibility focus have huge complexity costs |
| 03:43 | MikeSmith | reads karlcow link |
| 03:43 | <MikeSmith> | karlcow: I see |
| 03:46 | <karlcow> | [redacted] I had written more non sense about architecture ☺ |
| 03:46 | <karlcow> | 23:48 here |
| 03:46 | <MikeSmith> | heh |
| 03:46 | <MikeSmith> | night night |
| 03:46 | <MikeSmith> | you mean you're heading off for the evening? |
| 03:47 | <karlcow> | in a few minutes. I'm human, I need my share of lullabies and dreams :) |
| 04:53 | <MikeSmith> | hmm |
| 04:53 | <MikeSmith> | «Wouldn't it be great if the power of CSS3 / JS could enable a designer / developer to customize how a browser window looks? Imagine the freedom of expression allowed for a designer / coder who is not limited to a simple viewport encased within a cumbersome looking "Chrome" based window?» |
| 04:54 | <MikeSmith> | no actually, not sure that would be so great at all |
| 04:54 | <MikeSmith> | http://iwantaneff.in/chromeless/ |
| 05:00 | <zcorpan> | wow the people who get those tweets must hate chromeless by now |
| 05:02 | <MikeSmith> | heh |
| 05:03 | <MikeSmith> | I admire the approach for getting attention |
| 05:07 | <zcorpan> | http://software.hixie.ch/utilities/js/live-dom-viewer/saved/1608 chrome doesn't run the w(1) script |
| 06:32 | <zcorpan> | "because only you jeff aids and you just sixteen are guaranteed city use any" http://www.youtube.com/watch?v=nhVfIyFPWAA (15:28 with captions enabled) |
| 06:51 | <Von_Davidicus> | I've been wondering if it would be worth it to leave the old Coach Random website alongside the new one... (the old one being the XML-based website) |
| 07:06 | <Von_Davidicus> | Hello, annevk. |
| 07:07 | <annevk> | mornings |
| 07:08 | <annevk> | "Each community is purist with regards to its own thought framework." is nice :) |
| 07:08 | <Von_Davidicus> | Heh... |
| 07:09 | <Von_Davidicus> | So. Question. Even though XML+XSLT+(XSD+DTD) is a ridiculous setup, is it a bad thing that I'm still kinda proud of pulling it off? |
| 07:10 | <annevk> | not at all |
| 07:10 | <annevk> | it's nice to know the technologies that are out there so you can learn from them |
| 07:10 | <annevk> | at some point in time I thought "generic XML" was the future too |
| 07:11 | Von_Davidicus | is working on changing the Coach Random website to a PHP/SQL setup. |
| 07:11 | <Von_Davidicus> | I must admit--it made for a very small website in terms of file sizes. |
| 07:12 | <Von_Davidicus> | And, frankly, it was fun to do. |
| 07:19 | <annevk> | "it's good having putting documents through the Rec process, because you have to identify WGs from which you need feedback" this TAG stuff is kind of sad |
| 07:20 | <annevk> | "we should liaise for example with WHATWG" well that was unexpected :) |
| 07:20 | <annevk> | although I'm not sure how such a liaison would work... |
| 07:21 | <zcorpan> | maybe they'll join the whatwg list? |
| 07:22 | <annevk> | it would be kind of interesting to have a technical discussion on architecture now and then |
| 07:33 | <kennyluck> | I can't quite imagine how such a liaison would be useful. |
| 07:40 | <annevk> | kennyluck: try saying "yes" more often for a while and see how it turns out ;) |
| 07:42 | <kennyluck> | annevk, yes man. Though I am just trying to be get inspiration in how such a thing would work. |
| 07:42 | <kennyluck> | Appraently Hixie and some other folks don't like telecoms, so there's only either IRC or mailing list. |
| 07:43 | <annevk> | I think Hixie has actually had lunch or some such with the TAG once, but it didn't really go anywhere |
| 07:45 | <annevk> | http://xkcd.com/1071/large/ :) |
| 07:46 | <annevk> | we also had lunch with the W3C AB at some point, but I'm not sure that went much anywhere either, but they're a rather slow group so maybe next year or so :) |
| 07:51 | <annevk> | ugh, vacation auto-responders |
| 07:54 | <annevk> | Hixie: should we change class=domintro to class=intro so it can be used for other things too? e.g. CSS pseudo-elements... |
| 07:57 | <volkmar> | annevk: you said that we need to reference HTML to define the event handler attributes |
| 07:57 | <volkmar> | but can't we move the related part of the spec to DOM ? |
| 07:58 | <volkmar> | a lot of API might use those event handlers attributes |
| 07:58 | <annevk> | no |
| 07:58 | <annevk> | volkmar: though if you figure out how to work around the issues listed in https://www.w3.org/Bugs/Public/show_bug.cgi?id=16477 maybe we can |
| 07:59 | <kennyluck> | Regular lunch/meetings are even more synchronous then telecons, which I consider unfair. |
| 08:00 | <annevk> | kennyluck: well they're equally synchronous; I was just trying to point out that there is some precedent of that happening |
| 08:03 | <kennyluck> | annevk, kk. |
| 08:15 | <jgraham> | Hixie is known to not like telecoms - http://ln.hixie.ch/?start=1263974578&count=1 right? |
| 08:15 | <jgraham> | (sorry ;) |
| 08:17 | <volkmar> | annevk: the question might sounds stupid but can't we just move the minimum so APIs can use onfoo handlers, without carring about all those HTML specfic stuff? |
| 08:17 | <annevk> | volkmar: I haven't found a way |
| 08:18 | <volkmar> | ok :( |
| 08:19 | <annevk> | maybe there can be some kind of split, but I'm not sure if that helps implementors since it should be a logical unit implementation-wise |
| 08:19 | <annevk> | and it does not help developers either trying to understand how they work |
| 08:21 | <volkmar> | it would help specifications to not depend on HTML5 just for event handlers attributes |
| 08:22 | <annevk> | they're the least important when it comes to constituencies |
| 08:23 | <annevk> | besides, most APIs use the event loop, depend on origin, the Window object, Navigator, etc. anyway |
| 08:24 | <annevk> | I'd be interesting to see one that only requires this |
| 08:27 | <MikeSmith> | hey all you Tidy fans and Google Docs fans, I just just pushed some code from Dave Raggett himself that enables you to clean up the unclean HTML that Google Docs exports |
| 08:27 | <MikeSmith> | so now you can do all your HTML authoring in Google Docs! |
| 08:29 | <MikeSmith> | speaking of the TAG, the TAG would be seriously enriched by the presence of Dave Raggett |
| 08:29 | <zcorpan> | wait, Google Docs needs Word HTML cleanup as well? isn't that just FAIL? |
| 08:30 | <MikeSmith> | zcorpan: dunno man I just push the code |
| 08:31 | <MikeSmith> | I don't think Dave has ever been on the TAG |
| 08:31 | <MikeSmith> | there's probably a reason for that... |
| 08:32 | <MikeSmith> | Dave gets into architecture astronaut mode sometimes but at the other times he just writes a lot of good code |
| 08:32 | <MikeSmith> | in C |
| 08:32 | <MikeSmith> | and even in Javascript |
| 08:32 | <MikeSmith> | and probably in all kinds of other stuff |
| 08:32 | <MikeSmith> | and he actually likes Web browsers! |
| 08:33 | <zcorpan> | unlike TAG people? |
| 08:33 | <MikeSmith> | I didn't say that |
| 08:34 | <abarth> | :) |
| 08:35 | <MikeSmith> | spending time with Dave is humbling and even inspiring |
| 08:35 | <MikeSmith> | as it was with DanC |
| 08:35 | <MikeSmith> | and Tim |
| 08:35 | <MikeSmith> | that's one of the really nice things about this job |
| 08:36 | <MikeSmith> | there's old school and then there's old schoo |
| 08:36 | <annevk> | to create this internet thing where you get connected with http://www.reddit.com/r/IAmA/comments/vb53k/my_17_year_old_exgirlfriend_is_now_dating_my_39/ what is this I don't even know |
| 08:37 | <MikeSmith> | abarth: hey, was reading http://abarth.github.com/sysapps/drafts/runtime.html today |
| 08:37 | <MikeSmith> | interesting stuff |
| 08:37 | <abarth> | yeah? |
| 08:37 | <abarth> | i've been working on the security model tonight |
| 08:37 | <MikeSmith> | cool |
| 08:37 | <abarth> | http://abarth.github.com/sysapps/drafts/security.html |
| 08:37 | <abarth> | its still very, very rough |
| 08:38 | <abarth> | but taking shape |
| 08:38 | <MikeSmith> | I don't have any feedback but it's great to have those to get a more concrete idea of where you're going with this stuff |
| 08:38 | MikeSmith | looks at updated security draft |
| 08:39 | <abarth> | i'm hopeful the charter discussion will be over soon so we can start with the technical work |
| 08:40 | <annevk> | hsivonen: you're advising people to go from Nokia to RIM? :) |
| 08:40 | <MikeSmith> | abarth: charter discussion always painful |
| 08:41 | <MikeSmith> | abarth: but seems like after the survey, you all are converging on some agreement about what should be in the phase 1 stuff |
| 08:41 | <hsivonen> | annevk: not advising. I just find it noteworthy that RIM, HP and Canonical are all hiring Qt people right now |
| 08:41 | <annevk> | fair enough |
| 08:41 | <abarth> | MikeSmith: yeah, fingers crossed |
| 08:42 | <annevk> | I guess they want out of Android and Qt is most viable platform? |
| 08:42 | <annevk> | the* |
| 08:42 | <smaug____> | hsivonen: interesting |
| 08:42 | <MikeSmith> | annevk: i'm pretty sure that reddit post is from one of my cousins in Iowa or Nebraska |
| 08:44 | <MikeSmith> | annevk: if you're waiting on me for anything, please remind me. I just today plowed through 1800 or so messages in my inbox, and have 300 or something left |
| 08:44 | <annevk> | MikeSmith: I don't think there's anything there from me |
| 08:45 | <MikeSmith> | OK |
| 08:45 | <annevk> | MikeSmith: I had some questions on URLs, but they're more or less resolved now |
| 08:45 | <MikeSmith> | OK |
| 08:45 | <MikeSmith> | I still need to figure out how I muffed up the backrefs in the URL draft |
| 08:45 | <annevk> | just need to find some time to do the work without getting distracted by useless Fullscreen / CORS discussions |
| 08:45 | <MikeSmith> | annevk: btw, any estimate on when we might do the LCWD for Web Notifications? |
| 08:45 | <annevk> | ah, I'll work on that instead |
| 08:46 | <annevk> | MikeSmith: lets publish it next week |
| 08:46 | MikeSmith | reminds himself that PLH needs to get us a charter extension |
| 08:46 | <annevk> | my vacation starts next week, but I can do that |
| 08:46 | <MikeSmith> | annevk: OK thanks |
| 08:46 | <annevk> | two month vacation time and I'm not really sure what I'm gonna do with it all so I guess I'll end up working a bit at least |
| 08:47 | <MikeSmith> | oh |
| 08:47 | <MikeSmith> | you could come to Tokyo |
| 08:47 | <annevk> | yeah I want to, but unfortunately the plans I do have kind of make that impossible :( |
| 08:47 | <MikeSmith> | it's not unbearably humid yet, but it should be by the time you'd get here |
| 08:49 | <annevk> | I don't really have a stretch of two-three weeks with nothing to do where I could come to Tokyo |
| 08:49 | <annevk> | I would have otherwise |
| 08:52 | <annevk> | is |
| 08:52 | <annevk> | "The Web Notifications WG is planning to move Web Notifications to W3C Last Call meaning we don't intend to change it. But we might have missed something and would therefore appreciate your review of http://dvcs.w3.org/hg/notifications/raw-file/tip/Overview.html and any comments you might have at public-web-notification⊙wo" |
| 08:52 | <annevk> | relatively okay English? |
| 08:55 | <MikeSmith> | annevk: looks great to me |
| 09:00 | <Ms2ger> | annevk, sounds like good Dutch to me ;) |
| 09:01 | <annevk> | Ms2ger: je ne compris pas |
| 09:01 | <kennyluck> | doesn't sound like normal annevk-ish :p |
| 09:03 | <Ms2ger> | Oh hey |
| 09:03 | <Ms2ger> | First bug report about Attr.firstChild |
| 09:05 | <annevk> | kennyluck: hopefully it does not sound too much like an actual Chair either |
| 09:06 | <annevk> | MikeSmith: http://lists.w3.org/Archives/Public/public-web-notification/2012Jun/0028.html means we get past the pubteam right? |
| 09:07 | <MikeSmith> | yeah |
| 09:09 | <annevk> | thanks for the reminder |
| 09:09 | <annevk> | I keep forgetting about notifications, I planned to do this Monday :) |
| 09:10 | <MikeSmith> | heh |
| 09:10 | <MikeSmith> | lot of good stuff going on |
| 09:10 | <MikeSmith> | hey we are likely to get the Encoding spec into i18n WG |
| 09:10 | <MikeSmith> | talkd with Richard and PLH about it yesterday |
| 09:11 | <MikeSmith> | Richard needs to write a new draft charter |
| 09:11 | <MikeSmith> | then we go from there |
| 09:11 | <MikeSmith> | I think he'll have it drafted next week |
| 09:14 | <Ms2ger> | Yay, charters |
| 09:15 | <annevk> | MikeSmith: oh interesting |
| 09:16 | <MikeSmith> | annevk: ok, I also just now updated the old cvs link on http://www.w3.org/2010/web-notifications/ |
| 09:16 | <MikeSmith> | per dude's mail |
| 09:17 | <annevk> | cool |
| 09:17 | <annevk> | should maybe create a wiki page for the use cases and such that were originally in the spec |
| 09:17 | <MikeSmith> | I'm happy for somebody to do that :) |
| 09:18 | <MikeSmith> | right now I only got 14 minutes battery left and trying to get github mirroring thing set up for matjas JS tests |
| 09:19 | <annevk> | I like how instead of the title wrapping it just gives you dots at the end on our home page |
| 09:19 | <annevk> | text-overflow:ellipsis ftw |
| 09:20 | <annevk> | MikeSmith: no worries, I can handle that |
| 09:20 | <MikeSmith> | cool |
| 09:22 | <Ms2ger> | annevk, man, more vacationing? ;) |
| 09:22 | <annevk> | Ms2ger: well I skipped part of university so I have some catching up to do |
| 09:29 | <MikeSmith> | matjas or Ms2ger : please push something to the web-ecmscript dvcs.w3.org repo and lemme know what happens |
| 09:29 | <Ms2ger> | Do you have something to test? :) |
| 09:29 | <MikeSmith> | within the next 3 minuts, preferable |
| 09:29 | <MikeSmith> | Ms2ger: nope |
| 09:29 | <MikeSmith> | add a README maybe? |
| 09:29 | Ms2ger | looks |
| 09:32 | <MikeSmith> | 1 minute... I will be back on in 90 minutes or so and will check back then |
| 09:39 | <matjas> | MikeSmith (if you read the logs): this happens http://pastie.org/pastes/4119544/text |
| 09:48 | <matjas> | any feedback so far on http://mathias.html5.org/specs/web-ecmascript/? I copied the Web ECMAScript wiki page, made some corrections, and added some more things |
| 09:51 | <[tm]> | ? |
| 09:52 | <jgraham> | matjas: You incorrectly use RFC2119 language in examples |
| 09:52 | <jgraham> | e.g. "For example, var var; must throw a SyntaxError, but e.g. var v\u0061r; should work — even though strictly speaking, the ECMAScript spec disallows it." |
| 09:52 | <matjas> | s/must/should/, then? |
| 09:52 | <jgraham> | Just makes them statements of fact |
| 09:53 | <matjas> | k |
| 09:53 | <jgraham> | i.e. don't use must.should.etc. at all |
| 09:54 | <matjas> | jgraham: as you’re the one who wrote those algorithms, am I doing `EscapeAttributeValue` (and others) right? |
| 09:54 | <matjas> | (I noticed WebKit escapes single quotes too, although that seems unnecessary) |
| 09:56 | <jgraham> | I think personally I would make those algorithms have immutable variables |
| 09:57 | <jgraham> | And phrase it like: |
| 09:58 | <jgraham> | Let /escaped/ be /value/ with each U+0022 QUOTATION MARK characters replaced by the string """ |
| 09:58 | <jgraham> | Return /escaped/ |
| 09:58 | <jgraham> | Or something |
| 09:58 | <jgraham> | But I don't think that is perfect either |
| 09:59 | <matjas> | what else is needed to make it perfect? |
| 09:59 | <jgraham> | I'm not sure |
| 10:00 | <jgraham> | If you were really pedantic, you could make "replaced by" explicit |
| 10:00 | <jgraham> | i.e. loop through the string and examine each character |
| 10:00 | <matjas> | wow |
| 10:00 | <zcorpan> | i thought they shouldn't do any escaping |
| 10:00 | <jgraham> | see what HTML does for similar things |
| 10:01 | <matjas> | jgraham: I did exactly that before writing that up |
| 10:01 | <jgraham> | Maybe I am giving you bad advice then :) |
| 10:03 | <zcorpan> | matjas: opera and firefox don't escape the value. have you tested ie? |
| 10:03 | <matjas> | zcorpan: yeah but shouldn’t they? (haven’t tested IE) |
| 10:04 | <zcorpan> | matjas: see /topic |
| 10:04 | <matjas> | damn |
| 10:04 | <zcorpan> | unless there's a security problem or something, i think we should go with the majority |
| 10:07 | <zcorpan> | matjas: the comment syntax section is pretty vague, and should probably hook in to ES in some way |
| 10:08 | <matjas> | i have some TODOs as HTML comments |
| 10:08 | <jgraham> | Well of course it could be a security problem if someone was doing something dumb |
| 10:10 | <matjas> | IE10 doesn’t escape " and uppercases EVERYTHING |
| 10:10 | <matjas> | same in older IEs |
| 10:11 | <matjas> | would it break backcompat if newer browsers started escaping `"`? are there legitimate scripts out there (i.e. no crazy XSS hacks) that rely on the non-escaped behavior? |
| 10:11 | <Ms2ger> | Probably |
| 10:12 | <matjas> | hard to imagine, and it doesn’t seem to be a problem for WebKit |
| 10:13 | <matjas> | …but that’s not how this thing works, is it. |
| 10:13 | <smaug____> | annevk: so per notification spec tag isn't bound to an origin? |
| 10:13 | <smaug____> | I think it should be |
| 10:16 | <zcorpan> | matjas: i tried grepping in web200904 but it's hard to tell whether ysf=ysf.fontcolor(ac); has ac already-escaped or not |
| 10:17 | <matjas> | any results for ''.link('"') or similar? |
| 10:18 | <zcorpan> | i included link at first but it included so much jQuery and other link()s that it was mostly noise |
| 10:19 | <zcorpan> | but there's some |
| 10:19 | <zcorpan> | document.write(title.big().bold().link(base + link)); |
| 10:19 | <zcorpan> | document.write("View Our Open Positions".link(base)); |
| 10:20 | <zcorpan> | <p>For more information please contact <script>document.write(String.fromCharCode(-186-0x60+0x15b,0144,-0324+0501,0373-0222,0156,-0xe2+0x14b,0xba-0107,-0xce+0502,0x3d+065,-47+0220,-83-155+0x162,0157,-95-0256+0577).link(String.fromCharCode(0155,0xd3-0x72,0151,-166+0x112,0x74,56+0xb4-125,072,0x6c,-0x49+0262,040-52+135,0141,0100,-045+0206,0234-40,-0x90+0340+36,0x6f,0162,0151+0x5,-0xf3-0265+0x20d,-0246+255+32,66+49,0x73+034 |
| 10:20 | <zcorpan> | -0xf2+0377,0156,173-0110,105,0x90-213+0263,230+167-298,0x8d+150-0xf5,0x6c-0x9,245-0x86,109)));</script>.</p> |
| 10:20 | <matjas> | holy… |
| 10:21 | <zcorpan> | seems there was less noise in web200904 compared to the newer data set with "top 10,000 sites" |
| 10:21 | <matjas> | hah, even that one didn’t contain " in the attribute value: <a href="mailto:lisa⊙ac">Administrator</a> |
| 10:22 | <matjas> | why would anyone expect and *rely on* these functions returning invalid HTML in case the value includes `"`? |
| 10:23 | <zcorpan> | the interesting question isn't "why" |
| 10:24 | <zcorpan> | $ grep -aPh "\.(anchor|fontcolor|fontsize|link)\s*\(\s*'[^']*\"" web200904 produced no results |
| 10:24 | <hsivonen> | annevk, smaug____: RIM and Canonical hiring Qt folks is interesting but not *surprising* but HP hiring Qt folks or *anyone* for the division formerly known as Palm is suprising |
| 10:26 | <smaug____> | hsivonen: RIM hiring anyone is surprising |
| 10:27 | <matjas> | the only reason anyone would rely on this (that I can think of) is XSS |
| 10:27 | <hsivonen> | smaug____: well, RIM is known to try to get itself back together by doing Qt stuff |
| 10:27 | <matjas> | ''.link('"><script>alert("h4x")</script>'); // "<a href=""><script>alert("h4x")</script>"></a>" |
| 10:28 | <hsivonen> | smaug____: meanwhile, it seems the old Java folks from RIM are migrating to Android-related jobs |
| 10:29 | <hsivonen> | I wonder how easy it is for HP or RIM to get Norwegian or Finnish ex-Nokia people to relocate to Ontario or California, when the longevity of RIM and the division formerly known as Palm doesn't look so good |
| 10:30 | <hsivonen> | of the companies hiring Qt folks, Canonical has the advantage that they don't require relocation |
| 10:30 | hsivonen | is quite happy not to be an ex-Nokia person forced to look for a new job |
| 10:31 | <matjas> | data:text/html,<script>document.write(''.link('"><script>alert("h4x")<\/script>'));</script> XSS vector for Opera, Firefox and IE |
| 10:31 | <annevk> | smaug____: agreed |
| 10:33 | <annevk> | matjas: jgraham: saying "replaced by" is fine |
| 10:33 | <matjas> | yay |
| 10:34 | <annevk> | matjas: e.g. Return the result of replacing all occurrences of X in Y by Z |
| 10:34 | <zcorpan> | matjas: i can't find any legitimate uses of " in web200904, and since it's a potential xss problem, it seems OK to require escaping the " |
| 10:35 | <zcorpan> | (or any uses at all) |
| 10:35 | <matjas> | zcorpan: ok, I’ll add a note and an HTML comment linking back to these logs |
| 10:40 | <matjas> | thanks for grepping :) |
| 10:40 | <annevk> | matjas: you should require an explicit HTML escape, not allow all of them |
| 10:41 | <annevk> | matjas: I recommend requiring " as HTML uses that too |
| 10:41 | <jgraham> | (that was the main point I was trying to make with my previous comment, but it got lsot a bit under how to phrase it) |
| 10:41 | <annevk> | matjas: see e.g. http://www.whatwg.org/specs/web-apps/current-work/#escapingString |
| 10:42 | <annevk> | matjas: looks cool btw |
| 10:43 | <annevk> | matjas: note that for Date there's a whole lot more to define; basically new Date(stuff) is not defined for a lot of date formats |
| 10:43 | <matjas> | annevk: thanks, will do |
| 10:43 | <zcorpan> | matjas: nice work |
| 10:43 | <matjas> | oh btw i specced String#substr, in case anyone wants to review that |
| 10:44 | <zcorpan> | isn't substr in ES proper? |
| 10:44 | <matjas> | zcorpan: no |
| 10:44 | <matjas> | i wasn’t sure if min() and max() need to be defined as abstract operations as well |
| 10:44 | <zcorpan> | http://es5.github.com/#B.2.3 |
| 10:44 | <matjas> | daaaamn |
| 10:45 | <zcorpan> | seems that whole section is non-normative though |
| 10:45 | <zcorpan> | which is bogus |
| 10:47 | <zcorpan> | but you could reference Annex B and say it must be implemented, or some such |
| 10:48 | <annevk> | "This non-normative annex suggests uniform semantics for such properties without making the properties or their semantics part of this standard." lol |
| 10:50 | <hsivonen> | what's the deal with TC39 refusing to stardardize stuff that needs to be implemented? |
| 10:50 | <hsivonen> | is the necessity of implementation in doubt? |
| 10:51 | <hsivonen> | or do they want non-browser impls without this stuff to be "conforming"? |
| 10:51 | <annevk> | brendan always tell me they have a different approach to standardization and he thinks it is good to have standards bodies compete on strategy |
| 10:51 | <annevk> | I don't necessarily disagree, but I do disagree with the strategy :) |
| 10:52 | <matjas> | they’re working on explicitly forbidding Web ECMAScript identifiers in the spec, too: https://bugs.ecmascript.org/show_bug.cgi?id=277 |
| 10:53 | <matjas> | (or allowing it, depending on the outcome of their research) |
| 11:00 | <hsivonen> | Use case for CSS: (unless you are prone to seizures or are not allowed to make noise at your location), go to http://plone.org/ and type "batman " into the search field |
| 11:02 | <annevk> | matjas: speccing __proto__ sounds cool |
| 11:02 | <annevk> | matjas: and if they actually update the browsers too, then it's cool |
| 11:02 | <matjas> | TIL it’s actually happening in ES6! |
| 11:03 | <annevk> | matjas: but not having the spec match browsers is uncool |
| 11:03 | <annevk> | matjas: yeah, that's what I meant :) |
| 11:03 | <Ms2ger> | hsivonen, heh |
| 11:22 | <gsnedders> | hsivonen: TC39 basically agreed to public something along the lines of Web ECMAScript (whether as an appendix or what is another question). Some things are likely to make it into the spec, some aren't. |
| 11:23 | <gsnedders> | hsivonen: Library stuff (apart from stuff like Sting.prototype.bold, etc., which is obviously very web-specific) is likely to get into the main spec, esp. if it can't legally be implemented as host objects. |
| 11:24 | <gsnedders> | annevk: Annex B is becoming, "optional, but if you implement an extention with this name, required" |
| 11:26 | jgraham | wonders if SVG scripts are getting the force-async flag |
| 11:27 | <annevk> | gsnedders: sounds so very political |
| 11:33 | jgraham | gets a resounding "maybe" |
| 11:35 | <jgraham> | hsivonen: Know anything about the differences bwtween SVG script and HTML script (in gecko?) |
| 11:36 | <hsivonen> | jgraham: we specifically pretend that async and defer attributes don't exist for SVG scripts, since the SVG WG hasn't blessed them |
| 11:36 | <Ms2ger> | If anybody does, he does :) |
| 11:36 | <hsivonen> | which is kinda silly |
| 11:37 | <gsnedders> | annevk: There's not really that much in way of politics mostly. |
| 11:37 | <gsnedders> | At least on the public list. |
| 11:37 | gsnedders | has never been on the private one |
| 11:38 | <jgraham> | hsivonen: Yeah, it is. So you also don't have a force-async flag? |
| 11:38 | <MikeSmith> | annevk: I agree with brendan about it being good to have standards bodies compete on strategy. So let's make an actual separate standard, based on the EcmaScript spec, but in line with what browsers actually implement, with the Web EcmaScript parts and other things that should be normative being normative. And call it the JavaScript Standard. |
| 11:39 | <jgraham> | Standards bodies *usually* only compete for new specs, so the idea of better living through competition is rather weak |
| 11:39 | <annevk> | MikeSmith: matjas: I'm totally in favor of calling it "JavaScript" |
| 11:40 | <jgraham> | Obviously HTML is an exception, but the W3C managed to fuck that up so bad it defies belief |
| 11:40 | <hsivonen> | jgraham: the force-async flag only exists for HTML scripts |
| 11:42 | <hsivonen> | jgraham: I'd be OK with adding the async and defer attributes and the force-async flag to SVG scripts with the exact same semantics as with HTML scripts |
| 11:43 | <MikeSmith> | matjas or Ms2ger : OK please try again a push to the web-ecmascript repo |
| 11:44 | <MikeSmith> | I had forgotten to set something but I think I have it right now |
| 11:44 | <matjas> | MikeSmith: |
| 11:44 | <matjas> | remote: abort: Permission denied: /var/hg/repos/web-ecmascript/.git/objects/40 |
| 11:44 | <matjas> | remote: fatal: 'github' does not appear to be a git repository |
| 11:44 | <matjas> | remote: fatal: The remote end hung up unexpectedly |
| 11:44 | <matjas> | remote: warning: changegroup.github hook exited with status 128 |
| 11:44 | <MikeSmith> | oops |
| 11:44 | <jgraham> | It seems like a good idea to me, but it would be running ahead of the spec a bit. I can't imagine the SVG2 process is going to be very fast |
| 11:45 | <heycam> | hsivonen, jgraham, we want script to have the same features in SVG as it does in HTML |
| 11:45 | <MikeSmith> | matjas: OK, just now fixed that. please try again |
| 11:45 | <MikeSmith> | matjas: oh, hang on |
| 11:45 | <heycam> | maybe not a fast process, but we could at least get that in the spec without too much work |
| 11:45 | <MikeSmith> | I have to set the perms |
| 11:46 | <matjas> | MikeSmith: |
| 11:46 | <matjas> | remote: exporting hg objects to git |
| 11:46 | <matjas> | remote: abort: Permission denied: /var/hg/repos/web-ecmascript/.git/objects/40 |
| 11:46 | <matjas> | remote: ERROR: Permission to w3c/web-ecmascript.git denied to w3c-bots. |
| 11:46 | <matjas> | remote: fatal: The remote end hung up unexpectedly |
| 11:46 | <matjas> | ahh sorry |
| 11:46 | <jgraham> | heycam: I want the spec to be a living standard, so that this kind of update doesn't need to wait on multi-year cycles ;) |
| 11:46 | <heycam> | well it will be in th ED, and that spec will be publicly visible |
| 11:46 | <MikeSmith> | matjas: OK fixed the perms now, ready to be tried again |
| 11:46 | <jgraham> | But good to know that there is enthusiasm for the idea |
| 11:47 | <Ms2ger> | Oh hey, it's heycam |
| 11:47 | <matjas> | MikeSmith: same error |
| 11:47 | <matjas> | remote: exporting hg objects to git |
| 11:47 | <matjas> | remote: ERROR: Permission to w3c/web-ecmascript.git denied to w3c-bots. |
| 11:47 | <matjas> | remote: fatal: The remote end hung up unexpectedly |
| 11:47 | <matjas> | remote: warning: changegroup.github hook exited with status 128 |
| 11:47 | <heycam> | hi Ms2ger |
| 11:47 | <annevk> | smaug____: yt? |
| 11:47 | <heycam> | we're rarely on at the same time these days |
| 11:47 | <MikeSmith> | matjas: OK, will fix that now |
| 11:47 | <annevk> | smaug____: would you implement the origin stuff by basically storing the origin on the Notification object? |
| 11:47 | <annevk> | smaug____: would that be logical? |
| 11:48 | <matjas> | http://gavinsharp.com/irc/whatwg.html (from /topic) seems to 404 |
| 11:48 | <MikeSmith> | matjas: OK, ready for another ry |
| 11:49 | <annevk> | matjas: 403 here |
| 11:49 | <matjas> | it worked! |
| 11:49 | <MikeSmith> | sweet |
| 11:49 | <matjas> | annevk: whoops, yeah 403 |
| 11:49 | <jgraham> | matjas: Yeah, it worked the other day for me |
| 11:49 | <annevk> | gavin: gavin_: your irc/whatwg.html 403's... |
| 11:50 | <jgraham> | matjas: (since you were obviously wondering, no you haven't overtaken annevk on the loser board yet) |
| 11:50 | <matjas> | hah! |
| 11:52 | <hsivonen> | (aside: since SVG scripts can nest, supporting async and defer for them will complicate the HTML parser a bit) |
| 11:53 | <hsivonen> | oh, also: at one point, src="" (empty string value) on HTML and xlink:href="" on SVG had different behavior |
| 11:53 | <hsivonen> | let's see if that's still the case... |
| 11:54 | <jgraham> | Sigh, I didn't mean to end up testing this stuff today |
| 11:55 | <jgraham> | I only meant to do document.write in SVG, and then realised there was a lot more scope |
| 11:56 | <jgraham> | (what is the problem with async and defer in SVG? Does it also apply to XHTML?) |
| 11:57 | <hsivonen> | Hixie's live DOM viewer seems broken |
| 11:57 | <hsivonen> | (or Gecko is very broken) |
| 11:59 | <hsivonen> | jgraham: looks like both Firefox and Opera have a bug here, but different bugs. WebKit works correctly: http://software.hixie.ch/utilities/js/live-dom-viewer/?saved=1613 |
| 11:59 | <hsivonen> | jgraham: but it looks like Firefox now gives the same treatment to empty src in HTML and empty xlink:href in SVG |
| 12:01 | <jgraham> | Nice bugs you have there |
| 12:01 | <jgraham> | s/have/found/ |
| 12:12 | <matjas> | how can I tell Anolis to use data/references.json? I’m getting “SyntaxError: Specification not found: ECMASCRIPT” although there’s an entry for it |
| 12:13 | <matjas> | /spam gsnedders Ms2ger jgraham |
| 12:13 | <Ms2ger> | Are you sure? :) |
| 12:13 | <matjas> | Ms2ger: what am I missing? |
| 12:14 | Ms2ger | looks |
| 12:14 | <matjas> | is there a command-line flag I need to pass to `anolis` explicitly or smth? |
| 12:15 | <Ms2ger> | So |
| 12:15 | <Ms2ger> | references.json is for the References section |
| 12:16 | <Ms2ger> | That error comes from the cross-references, I think |
| 12:16 | <matjas> | I expect <span data-anolis-spec=ECMASCRIPT>…</span> to create a new entry in the References section, is that correct? |
| 12:16 | <Ms2ger> | No |
| 12:17 | Ms2ger | looks for the incantation |
| 12:17 | <Ms2ger> | <span data-anolis-ref class=informative>DOM2TR</span> |
| 12:17 | <matjas> | Ms2ger: Thanks! |
| 12:17 | <Ms2ger> | Np |
| 12:17 | <Ms2ger> | I really really should write more documentation :) |
| 12:18 | <matjas> | yes please, for n00bs like me :( |
| 12:18 | <matjas> | gsnedders’s example files are very minimal |
| 12:22 | <zcorpan> | matjas: data-anolis-spec= is for cross-spec cross-references |
| 12:23 | <zcorpan> | matjas: e.g. if you want a link directly to ES's ToString |
| 12:23 | <annevk> | Ms2ger: add it to http://wiki.whatwg.org/wiki/Howto_spec so we get all specs aligned |
| 12:23 | <annevk> | Ms2ger: also, maybe we should start supporting <span data-spec= and <span data-ref> |
| 12:23 | <annevk> | Ms2ger: because why type -anolis- when we really don't have to |
| 12:23 | <Ms2ger> | Or hey, <span spec> :) |
| 12:23 | <zcorpan> | <spec>! |
| 12:24 | <Ms2ger> | <s> |
| 12:24 | <annevk> | yeah if we strip it out anyway |
| 12:24 | <Ms2ger> | Oh |
| 12:24 | <matjas> | zcorpan: what should that look like then? <span data-anolis-spec=ecmascript-tostring>…</span> and then edit some file in the data/ folder? |
| 12:24 | <annevk> | matjas: yes |
| 12:24 | <annevk> | not sure we have one for ecmascript yet |
| 12:24 | <Ms2ger> | We don't |
| 12:25 | <annevk> | matjas: should look like the data/xref files |
| 12:25 | <matjas> | cool, thanks all |
| 12:25 | <Ms2ger> | And add a line to data/specs.json |
| 12:25 | <zcorpan> | except it'd be <span data-anolis-spec=ecmascript title=tostring>…</span> i think |
| 12:25 | <Ms2ger> | Right |
| 12:25 | <zcorpan> | or <span data-anolis-spec=ecmascript>ToString</span> |
| 12:25 | <Ms2ger> | ^ |
| 12:26 | <annevk> | matjas: ^ |
| 12:27 | <matjas> | roger that |
| 12:29 | <Ms2ger> | matjas, oh, and Mozilla is always hiring... ;) |
| 12:53 | <annevk> | matjas: http://opera.jobs/ |
| 12:53 | <annevk> | Ms2ger: ^ |
| 12:53 | <Ms2ger> | annevk, boo :) |
| 12:53 | <annevk> | Ms2ger: also, how can you offer first |
| 12:54 | <annevk> | Ms2ger: Mozilla is sloppy seconds, Opera offers first |
| 12:54 | <matjas> | I like where this is going |
| 12:54 | <Ms2ger> | "Microsoft Wins Congressional Backing For Do-Not-Track Default In IE10" |
| 12:59 | <hsivonen> | Ms2ger: URL? |
| 13:00 | <hsivonen> | Why doesn't the Congress just legislate the default without having to send any bytes then? |
| 13:00 | <Ms2ger> | politics.slashdot.org/story/12/06/20/1231225/microsoft-wins-congressional-backing-for-do-not-track-default-in-ie10 |
| 13:01 | <hsivonen> | http:// adding not working when copying, eh? |
| 13:01 | <hsivonen> | Ms2ger: but thanks |
| 13:01 | <Ms2ger> | Heh |
| 13:01 | <Ms2ger> | I blame Firefox |
| 13:01 | <Ms2ger> | Or myself, this is a pretty old build |
| 13:22 | <annevk> | matjas: if you <dfn title=ToHTMLTag>...</dfn> you can do <code title=toHTMLTag>... later on and it will hyperlink them together |
| 13:22 | <annevk> | matjas: in addition the fragment identifier will be #tohtmltag rather than the longish string it is now |
| 13:22 | <matjas> | annevk: awesome! |
| 13:23 | <matjas> | i was wondering about that |
| 13:23 | <annevk> | matjas: basically it either stringifies the contents of a <dfn> but you can override that by specifying a title="" |
| 13:25 | <zcorpan> | <dfn title=ToHTMLTag>...</dfn> <code>ToHTMLTag</code> also works |
| 13:31 | <annevk> | right |
| 13:36 | <jacobrask> | What is supposed to happen when a browser encounters pattern=""? |
| 13:40 | <zcorpan> | http://www.whatwg.org/specs/web-apps/current-work/multipage/common-input-element-attributes.html#compiled-pattern-regular-expression |
| 13:41 | <annevk> | jacobrask: anything in particular you are wondering about? |
| 13:41 | <jacobrask> | yes, I'm reading that, so does an empty string mean the same thing as "has no such attribute"? |
| 13:41 | <zcorpan> | ah. no, the attribute is still present even if the value is the empty string |
| 13:42 | <jacobrask> | annevk: Well we could take it in #staff ;) but Opera seems to fail any value for pattern="" while Chrome accepts any value |
| 13:43 | <zcorpan> | afaict from the spec it should compile, so only the empty string would be valid for the field |
| 13:43 | <jacobrask> | right, Opera accepts the empty string |
| 13:44 | <zcorpan> | bug in chrome then |
| 13:44 | <annevk> | sitecompat? |
| 13:44 | <jacobrask> | yep. Firefox is same as Opera |
| 13:44 | <jacobrask> | nope, a site I'm working on, wondered where/if I should report a bug.. |
| 13:45 | <annevk> | jacobrask: https://bugs.webkit.org/ it is |
| 13:45 | <odinho> | jgraham: So Paul Irish imported the docs from testharness.js to readme.md. -- Don't really like to have to update docs two places. https://github.com/paulirish/testharness.js/commit/7bdd2d76a5afeb2d80b5cd48ce4c403c37f47c18 |
| 13:46 | <odinho> | paul_irish: Were you thinking about pushing that to mercurial, or is it mostly to have a github.com-page that's easy to link to etc? |
| 13:46 | <zcorpan> | <label for="i51">Ваше имя:</label><input type="text" maxlength="50" name="answer_5" id="i51" pattern="string" notice="Проверьте, пожалуйста, ответили ли Вы на все вопросы анкеты"> |
| 13:46 | <zcorpan> | <label for="i52">Телефон:</label> <input type="text" maxlength="16" name="answer_6" id="i52" pattern="string" notice="Проверьте, пожалуйста, ответили ли Вы на все вопросы анкеты"> |
| 13:48 | <odinho> | jacobrask: It has been proposed to rename this channel to #opera :P |
| 13:48 | <jacobrask> | :) |
| 13:50 | zcorpan | finds no empty-value pattern attributes in his data |
| 13:52 | <jacobrask> | I create input elements and set some attributes conditionally |
| 13:52 | <jacobrask> | So really I expected/hoped for pattern="" to behave as in WebKit |
| 13:53 | <jgraham> | odinho: Agreed, I think we need to reformat the inline docs in markdown or so and have a script that can automatically extract them on push |
| 13:54 | <jgraham> | I also think we need w3c-test.org to become a pretty website with all the info about how to write tests for W3C |
| 13:54 | <odinho> | jgraham: Yeah. :D' |
| 13:55 | <odinho> | jgraham: BTW, I deleted my testharness.js-repo that was forked from you, to make one forked from w3c instead, -- but now it won't let me do that. -- I guess GitHub has some old reference lying around crashing it. |
| 13:56 | <jgraham> | Yeah, seems to be a problem with the forking system |
| 13:56 | <odinho> | jgraham: Oh, tobie did the same as paul_irish :] Common theme that. |
| 13:57 | <jgraham> | (that it is hard to change the upstream) |
| 13:57 | <odinho> | The "docs" on stackoverflow said delete + do it again would be a strategy. Guess it failed :P |
| 14:07 | <MikeSmith> | hahaha Opera guys trying to do an "Occupy whatwg" |
| 14:08 | <MikeSmith> | speaking of charters I just finished writing a draft of a new charter for the HTML WG |
| 14:09 | <MikeSmith> | which is scheduled to be unleashed on the world for review tomorrow |
| 14:09 | <MikeSmith> | good times ahead |
| 14:17 | <jgraham> | MikeSmith: Yeah, W3C team are the 1%, with their private mailing lists and close connections that they use to oppress the masses |
| 14:18 | <MikeSmith> | jgraham: we prefer the term "illuminati" |
| 14:18 | <MikeSmith> | and we're not oppressing the masses, we're cleverly manipulating and exploiting them |
| 14:32 | <matjas> | jgraham, odinho, paul_irish: that’s why you need to have the docs in one place, being the source code, and use a tool that automatically generates the Markdown-formatted docs for you |
| 14:33 | <odinho> | matjas: We have it one place now, - in the testharness.js-file. |
| 14:33 | <matjas> | odinho: yeah, but a “readable” version would be useful, like the one paul_irish created manually |
| 14:34 | <odinho> | matjas: Yeah, so we should get some commit-hook to make stuff like that. |
| 14:34 | <matjas> | a README.md file that is autogenerated from source, and visible in the GitHub repo, would be perfect |
| 14:34 | <matjas> | *cough* http://jonathan.tang.name/code/pyjsdoc *cough* |
| 14:34 | <odinho> | matjas: Yea. Although I like jgraham's proposal of making w3c-test.org a really nice page better. But they don't exclude eachother. |
| 14:35 | <matjas> | exactly |
| 14:36 | <odinho> | http://jonathan.tang.name/files/pyjsdoc/apidocs/ << kinda empty though ;-) |
| 14:38 | <matjas> | I personally use docdown (PHP; https://github.com/jdalton/docdown) but Ms2ger seemed to prefer Python |
| 14:38 | <odinho> | ...and odinho and jgraham and possibly more ;-) |
| 14:39 | <odinho> | So agree with Ms2ger there, -- but not in dvcs choice :-) |
| 14:39 | <Ms2ger> | YOU'RE WRONG |
| 14:39 | <Ms2ger> | :) |
| 14:39 | <odinho> | ^^ |
| 14:41 | <Ms2ger> | MikeSmith, does the new HTMLWG charter give you unlimited power to override any decisions? :) |
| 14:43 | <MikeSmith> | Ms2ger: no, but it does introduce a new role |
| 14:44 | <MikeSmith> | along with editor, chair, team contact |
| 14:44 | <Ms2ger> | Werewolf? |
| 14:44 | <MikeSmith> | Enforcer |
| 14:45 | <zcorpan> | MikeSmith: now you might as well just leak the whole thing |
| 14:45 | <MikeSmith> | wanted to call it Punisher but got overruled |
| 14:45 | <MikeSmith> | heh |
| 14:46 | <Ms2ger> | Oh, you mean it wasn't you who put it on wikileaks? |
| 14:46 | <MikeSmith> | the butler did it |
| 14:47 | <MikeSmith> | or rather I should say my gentleman's gentleman |
| 14:47 | <Ms2ger> | Said the actress to the pope |
| 14:49 | <MikeSmith> | anyway, the clock will soon strike midnight here and I figure I have done my quota of damage for today, so I shall drop from my PC soon |
| 14:49 | <MikeSmith> | if anybody needs anything, ping me at [tm] |
| 14:50 | <Ms2ger> | Good night, then :) |
| 14:51 | <MikeSmith> | cheers |
| 14:51 | <annevk> | nn MikeSmith |
| 14:51 | <MikeSmith> | g'night |
| 14:57 | <annevk> | saying Last Call is imminent is a good way to get comments |
| 14:58 | <annevk> | maybe we should say "Last Call is imminent" about a "random" chosen spec each week (filtering duplicates) |
| 14:59 | <Ms2ger> | DOM4? :) |
| 14:59 | Ms2ger | checks when the charter claims DOM4 will be a rec |
| 14:59 | <annevk> | sh |
| 14:59 | <AryehGregor> | "immediately before the charter expires" |
| 15:00 | <annevk> | if they find someone to maintain a branch of DOM4 while the rest of us edit DOM I guess that works |
| 15:00 | <Ms2ger> | Q1 2013 |
| 15:00 | <AryehGregor> | When does the charter expire? |
| 15:00 | <Ms2ger> | 31 May 2014 |
| 15:00 | <AryehGregor> | Okay, so I guess I was wrong. |
| 15:21 | <gavin> | annevk: looking into it |
| 15:21 | <gavin> | I think dreamhost broke something :( |
| 15:58 | <odinho> | https://github.com/w3c/testharness.js/pull/2 |
| 16:15 | <odinho> | https://github.com/w3c/testharness.js/pull/3 |
| 16:18 | <Ms2ger> | Can you give a less stupid example too? :) |
| 16:27 | <gavin> | matjas: stats fixed |
| 16:30 | <odinho> | Ms2ger: {toString:function(){return "hello";}} vs "hello" --- I didn't really hit this myself, it was someone rambling on an Opera channel, I just wrote the patch since I was already in the mood :P |
| 16:31 | <odinho> | Ms2ger: Not 100% convinced it's really needed. |
| 16:31 | <Ms2ger> | Mm |
| 16:35 | <Ms2ger> | Commented |
| 16:40 | <odinho> | Ms2ger: I like it. Did it. |
| 16:42 | <Ms2ger> | Hmm |
| 16:52 | <Ms2ger> | odinho, string concatenation and ${} things together? |
| 16:52 | Ms2ger | isn't entirely convinced |
| 16:55 | <odinho> | Ms2ger: Yeah, can't do ${} because it uses format_value which brings in the "'s. |
| 16:55 | <odinho> | ("number") 1 |
| 16:56 | <odinho> | I did that originally, but it looked ugly. :-) |
| 16:56 | <Ms2ger> | OK |
| 16:56 | <Ms2ger> | I was wondering if it did that |
| 17:27 | <Hixie> | there's still no way to change [[Class]] in JS, right? |
| 17:28 | <Hixie> | JS being so dynamic makes it oddly hard to do certain things |
| 17:29 | <Hixie> | (such as in this case, walk a tree of objects and serialise it to a string in such a way that it can be reconstructed later, but in a manner where the serialiser/parser code doesn't need to know about all the possible types of objects involved and what their constructors are) |
| 17:34 | <jsbell> | FYI In the ES6 drafts [[Class]] goes away, replaced by [[NativeBrand]] |
| 17:35 | <Ms2ger> | Are you introducing something else you'll call [[Class]]? |
| 17:36 | <jsbell> | (if you = me, I'm not part of TC39 I just lurk; but not to my knowledge) |
| 17:38 | <Ms2ger> | Maybe only the CSSWG is that silly |
| 17:38 | <Hixie> | jsbell: either way :-) |
| 17:39 | <Hixie> | jsbell: what i really want is a reflection of the object's identity such that i can later resurrect the object |
| 17:40 | <Hixie> | currently i have each object override a getClass() method and it returns the string which is the member of |window| that can be used as a constructor for that object |
| 17:42 | <Hixie> | but that's pretty hacky |
| 17:42 | <jsbell> | What's the motivation? (I care about the "structured clone" algorithm as it relates to IndexedDB) |
| 17:42 | <Hixie> | the motivation is in a game i'm writing, how to bring back the UI widgets as they were when the user last logged out |
| 17:43 | <Ms2ger> | Why are you writing games and not specs, sir! |
| 17:43 | <Hixie> | to test the specs |
| 17:44 | <Ms2ger> | Acid4: FPS? |
| 17:44 | <Hixie> | acid tests test the browsers :-) |
| 17:45 | <jsbell> | Doing this via the JSON API reviver functions, or separately? |
| 17:45 | <jsbell> | reviver/replacer |
| 17:46 | <Hixie> | separately |
| 17:46 | <Hixie> | it's not a single data structure i'm rebuilding, it's also all the backend network connections, etc |
| 17:46 | <Hixie> | the whole thing is very async |
| 17:48 | <jsbell> | Hrm. You could ask each object to serialize itself to a script fragment that reconstructs itself (and dependents), then parsing is just an eval(). ISTM that anything else will require external assumptions about the objects. |
| 17:49 | <Hixie> | that's even more hacky :-P |
| 18:32 | <Ms2ger> | Doesn't Opera have unprefixed border-image? |
| 18:37 | <miketaylr> | Ms2ger: i think it did once upon a time, but then went back to prefixed |
| 18:37 | <miketaylr> | not sure why |
| 18:37 | <Ms2ger> | Huh |
| 18:40 | <gsnedders> | We had all of CSS Background and Borders unprefixed. |
| 18:45 | <Ms2ger> | Do continue |
| 18:49 | <hsivonen> | Opera could benefit from repainting everything after a Linux box wakes up from hibernation |
| 18:49 | <hsivonen> | I'm seeing garbage pixels in an iframe |
| 18:49 | <hsivonen> | I'm not posting a screenshot, since I don't know what private data that memory range contains |
| 18:57 | <hsivonen> | Kinda weird that WAI is starting work on intentional events. http://www.w3.org/WAI/IndieUI/ |
| 18:57 | <hsivonen> | (as opposed to WebApps WG) |
| 19:00 | <Hixie> | there's no good way to test if an element is an HTML <section> element other than checking both nodeName/localName/tagName and the namespace or instanceof HTMLElement, right? |
| 19:01 | <hsivonen> | sounds about right |
| 19:02 | <Ms2ger> | QI to nsIContent and do ->IsHTML(nsGkAtoms::section) |
| 19:02 | <hsivonen> | Gnome Memes seems to be enjoying the new Angry Linus meme |
| 19:02 | <Hixie> | yeah i suppose we could add isHTML(tagName), isSVG(tagName), and isMathML(tagName) to Element |
| 19:03 | <Hixie> | Ms2ger: i'll let you and anne ponder whether it's worth it :-) |
| 19:03 | <hsivonen> | Hixie: Gecko has that internally |
| 19:03 | <Ms2ger> | Wontfix |
| 19:03 | <Ms2ger> | ;) |
| 19:03 | <hsivonen> | yeah, seems like WONTFIX material for the Web |
| 19:03 | <hsivonen> | authors won't dare to use pure sugar like that |
| 19:04 | <Hixie> | i would :-P |
| 19:04 | <hsivonen> | will worry about it not working in IE9 for the next 20 years |
| 19:04 | <hsivonen> | Hixie: you aren't under normal compat constraints or under normal unfamiliarity with the platform details constraints |
| 19:04 | <Hixie> | it is true |
| 19:05 | <Ms2ger> | In other words, you are abnormal :) |
| 19:12 | <dbaron> | Are the contents of the <video> element supposed to be fallback content? |
| 19:12 | <Ms2ger> | No |
| 19:12 | <Ms2ger> | Well, fallback for browsers that don't implement <video> |
| 19:13 | <dbaron> | ok |
| 19:41 | <Ms2ger> | annevk, you're wrong :) |
| 19:43 | <annevk> | Ms2ger: that would be nice, because all my stuff is in Mercurial |
| 19:44 | <annevk> | Ms2ger: but I think I'm right |
| 19:44 | <paul_irish> | odinho jgraham: Yeah mostly just finished the markdown port as a POC; just to see how good it'd look. I agree next step is to write a script to output from script to readme. ++ on pyjsdoc or docdown |
| 19:54 | <Ms2ger> | http://w3c-test.org/framework/suite/web-storage-dev/ says "NOTICE: Undefined offset: 1" |
| 19:54 | <Ms2ger> | [tm], ^ |
| 20:33 | <jgraham_> | pyjsdoc isn't any use since jsdoc isn't any use |
| 21:11 | <bencc> | are pointer events for pen and digitizer standard in html5? |
| 21:11 | <Hixie> | html5 doesn't even define events for mice... |
| 21:12 | <bencc> | Hixie: I'm talking about this: http://blogs.msdn.com/b/ie/archive/2011/09/20/touch-input-for-ie10-and-metro-style-apps.aspx |
| 21:12 | <bencc> | canvas.addEventListener("MSPointerMove", paint, false); |
| 23:53 | <heycam> | hsivonen, what do you mean SVG scripts can nest, like you can have nested <script> elements? |
| 23:54 | <gavinc> | Anyone who cares about RDF and is bored in the next few days http://dvcs.w3.org/hg/rdf/raw-file/default/rdf-turtle/index.html Turtle is going to Last Call next week! Feedback is awesome |
| 23:54 | <heycam> | hsivonen, we should just do whatever HTML does for that too |