| 04:56 | <MikeSmith> | TabAtkins: I'm wondering since when did emscripten break the order-of-magnitude barrier |
| 04:57 | <MikeSmith> | last I saw from Joel Webber's figures it still was pretty much an order of magnitude at least |
| 05:12 | <othermaciej> | I didn't think code hand-translated into JS had even broken the order of magnitude barrier |
| 06:25 | <kennyluck> | Why isn't @scoped allowed on <link>. I kind of expect that there was previous discussions about this but I can't find any. |
| 06:37 | kennyluck | found it http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2011-March/031049.html |
| 06:42 | <kennyluck> | Hmm… that didn't say why it is bad to allow @scoped on <link>, besides potentially a lack-of-use-case argument (aka. you can you <style>@import</style>). |
| 06:45 | <Hixie> | the lack-of-use-case argument is sufficient to keep anything out of the spec :-) |
| 06:53 | <kennyluck> | Hixie, well, if implementation cost of adding @scoped to <link> isn't much, not allowing it on <link> confuses authors. (just got a feedback like this) |
| 06:54 | <Hixie> | i don't really understand the use case for using <link> for scoped stylesheets |
| 06:54 | <Hixie> | the whole point is that the style sheet will be embedded in syndication |
| 06:57 | <kennyluck> | Hixie, how would an author know he/she is supposed to use <style>@import</style> instead of <link style> in this use case when the former is not so apparent? |
| 06:57 | <kennyluck> | Or, do you think <style>@import</style> shouldn't be used here too? |
| 06:57 | <Hixie> | i don't think @import makes much sense either |
| 06:57 | <kennyluck> | hmm… ok |
| 06:58 | <kennyluck> | I don't know the details of the use case here but it makes more sense now. |
| 07:02 | <hsivonen> | when people propose something potentially very dangerous, they go "don't be judgemental in *technical* WG" |
| 07:08 | <hsivonen> | DRM systems' Web sites--with smiling stock photo people |
| 07:09 | <hsivonen> | Absent from supported platforms of the Google DRM: desktop Linux |
| 07:10 | <othermaciej> | does Google have a DRM platform of their own? |
| 07:10 | <hsivonen> | othermaciej: http://www.widevine.com/ |
| 07:12 | <hsivonen> | It's interesting that it's considered OK to ask browsers to support unspecified and open-ended DRM but it's not OK to ask Netflix to arrange their CDN to have the capability to set particular HTTP headers |
| 07:14 | <Hixie> | hsivonen: yeah, that's such BS. I called Mark on it, I notice he ignored my e-mail completely. (http://lists.w3.org/Archives/Public/public-html/2012Feb/0497.html) |
| 07:15 | <Hixie> | hsivonen: (the "don't judge, be technical" thing) |
| 07:15 | <Hixie> | hsivonen: i notice nobody ever says we shouldn't be judgemental when the accessibility problem being discussed is for people with disabilities |
| 07:16 | <hsivonen> | Hixie: the most curious thing is people who have branded themselves a11y folks rooting for DRM |
| 07:16 | <Hixie> | yeah that is so completely baffling |
| 07:17 | <hsivonen> | Hixie: one would expect them to resist vigorously and with extreme judgementalness |
| 07:25 | <hsivonen> | Google's demo of their multiplatform DRM tells me: "Your operating system is not supported." |
| 07:31 | <othermaciej> | I wonder what the Director would think of an effortt to standardize a DRM API |
| 07:31 | <hsivonen> | interesting that Google uses Sintel for a DRM demo even though Sintel is CC-licensed and all the CC licenses have an anti-TPM clause |
| 07:32 | <hsivonen> | also interesting that Google's DRM player works in a virtual machine |
| 07:32 | <hsivonen> | lol FAIL? |
| 07:33 | <hsivonen> | they make a big deal about detecting screen recording software and then the thing runs under a VM |
| 07:33 | <othermaciej> | Widevine looks to be a recent-ish acquisition, I would guess Google would add desktop Linux support before putting this in Chrome |
| 07:34 | <othermaciej> | DRM in web video is a thorny issue |
| 07:35 | <othermaciej> | it seems that if it is not offered, providers of Hollywood video content use plugins or platform-specific native apps for delivery, instead of DRM-free Web video |
| 07:35 | <hsivonen> | othermaciej: I think platform-specific native apps are better for the health of the Web in this case |
| 07:36 | <hsivonen> | othermaciej: since they don't poison browser competition even though they retain the status quo as far as OS lock-in goes |
| 07:36 | <othermaciej> | I can understand that point of view but I can also see how reasonable people may differ |
| 07:36 | <hsivonen> | othermaciej: also, if 10 years from now Hollywood follows the path of the music industry, the legacy won't be in the browser platform |
| 07:37 | <othermaciej> | Hollywood content coming in native apps is aligned with my corporate financial incentives but notheless I am not sure it is the best thing for the Web |
| 07:38 | <othermaciej> | Hollywood content coming in plugin runtime environments has no real upside for me though |
| 07:38 | <zcorpan> | Hixie: why is download="" not in w3c html5? |
| 07:38 | <Hixie> | same reason ping="" isn't, iirc |
| 07:38 | <Hixie> | but i could be wrong |
| 07:39 | Hixie | barely follows the htmlwg these days |
| 07:46 | <zcorpan> | what was the reason for ping="" again? |
| 07:48 | <annevk> | make link tracking a little easier, not pollute links while doing that, and put users in control |
| 07:48 | <annevk> | it sort of lost to redirects (and thereby polluting links) for now :( |
| 07:48 | <zcorpan> | annevk: no the reason it's not in html5 |
| 07:48 | <annevk> | oh |
| 07:49 | <annevk> | some lobby that hit the right chair keywords? |
| 07:49 | annevk | forgot |
| 07:51 | <annevk> | hsivonen: haha, your twitter reply is golden |
| 07:51 | <annevk> | hsivonen: the one to fantasai |
| 08:39 | <zcorpan> | when did <ol type> become allowed? |
| 08:45 | <zcorpan> | http://www.w3.org/TR/2010/WD-html5-20101019/grouping-content.html#the-ol-element is the earliest WD with ol type |
| 08:50 | zcorpan | doesn't find it in http://html5.org/tools/web-apps-tracker?from=5643&to=5779 |
| 09:11 | <annevk> | https://plus.google.com/109925364564856140495/posts |
| 09:11 | <annevk> | win |
| 09:12 | <kennyluck> | the fact that use cases of @ping can be solved by server redirects reminds me of the fact that the use cases of URI template for <form> can be solved by server redirects. Or is this not a good analogy? |
| 09:12 | <kennyluck> | I like both of them though. |
| 09:22 | <asmodai> | Found a lovely Opera Mobile bug |
| 09:23 | <asmodai> | Just have no clue how to properly describe it in the issue tracker |
| 09:30 | <Velmont> | Hmm. I just saw that microsoft updated their idb tests. |
| 09:30 | <zcorpan> | just go "OMG YOU SUCK ok here's a tc" |
| 09:30 | <Velmont> | So they resemble mine quite much now, -- but they actually redid them. |
| 09:31 | <Velmont> | So we're doubleworking all the way to the bank :P |
| 09:31 | <Ms2ger> | Sounds like something Microsoft would do |
| 09:31 | <zcorpan> | did they make them non-automated in the process? |
| 09:31 | <Velmont> | I'll try to find out any good improvements they've done, and tests added that I don't have written myself. But it's quite the diff. |
| 09:32 | <asmodai> | zcorpan: Not sure if I can fully find a howto reproduce it, I did encounter it a few times, but not sure if logic bug in the application or due to a site's markup in some weird way. |
| 09:32 | <Velmont> | zcorpan: No, they did improve them in all they ways I suggested in my review. |
| 09:32 | <asmodai> | Trying my best in the ticket though XD |
| 09:34 | <annevk> | new home page http://isittacosunday.com/ |
| 09:36 | <gsnedders> | annevk: nice background for the yay :P |
| 09:39 | <Velmont> | Hmm. Maybe I should align with MS on the tests, now that they have fixed the actual problems. -- Although I hate all the boilerplate in the tests. |
| 09:39 | <Velmont> | But it makes it easier to understand an individual test. |
| 09:40 | <asmodai> | There we go ANDMEX-6845⊙boc |
| 09:41 | <Velmont> | I've got db = createdb(t); instead of the multiline big open() onupgradeneeded, setting up failing functions on the other event handler, etc. |
| 09:43 | <annevk> | david_carlisle: no worries man :) |
| 09:43 | <annevk> | david_carlisle: I was unable to spot that two files changed in a ten line diff yesterday |
| 09:44 | <david_carlisle> | still you'd think if I got as far as reading the makefile and seeing i needed a source of that name I might have spotted the file in the same directory. |
| 10:13 | <hsivonen> | does Canonical use DRM in their Ubuntu for TVs that they've demoed with Hollywood movies in the movie rental lens? |
| 10:13 | <hsivonen> | if they use DRM, which DRM? |
| 10:14 | <hsivonen> | and have they already struck deals to carry the movies whose titles/posters appear in their demos? |
| 10:26 | <asmodai> | hsivonen: I haven't heard anything about it so far. |
| 10:57 | <annevk> | hmm |
| 10:57 | <annevk> | exposing SVGMatrix on <canvas> might be problematic |
| 10:58 | <annevk> | wasn't there an idea to reconcile that with CSS somehow? |
| 11:10 | <zcorpan> | http://dev.w3.org/cvsweb/html5/html4-differences/Overview.src.html.diff?r1=1.143;r2=1.144 - http://dev.w3.org/html5/html4-differences/Overview.src.html |
| 11:12 | <annevk> | zcorpan: casing of doctype was consistent with HTML until you changed it... |
| 11:14 | <zcorpan> | i like lowercase better :-) |
| 11:16 | <Ms2ger> | Look out, you'll get an ISSUE on that |
| 11:16 | <zcorpan> | heh |
| 13:03 | <annevk> | hmm |
| 13:03 | <annevk> | iso-2022-cn is the worst of -kr and -jp combined it seems |
| 13:03 | <annevk> | and then it doesn't even need to be supported if not for XSS |
| 13:03 | <annevk> | to which IE is vulnerable |
| 13:38 | <annevk> | "Wait — I'm confused. Isn't the primary purpose of media type RFCs to be wilfully violated? What are we doing wrong?" Robin on public-xml-er |
| 13:38 | <annevk> | :) |
| 15:02 | <annevk> | defining UTF-8 is harder than I thought |
| 15:08 | <gsnedders> | annevk: How is it hard? |
| 15:09 | <Ms2ger> | It's an encoding |
| 15:09 | <annevk> | gsnedders: lots of conditions mostly |
| 15:10 | <gsnedders> | annevk: I guess that just makes it awkward in prose more than anything else. |
| 15:10 | <gsnedders> | Just define it in C89 and reference the C spec :P |
| 15:11 | <annevk> | gsnedders: compared with the table-based encodings it's just a little less straightforward |
| 15:12 | <gsnedders> | Well, yeah, all the algorithmic ones are less simple. |
| 15:12 | <gsnedders> | You defining things like GB108030 (or whatever the number is) as a table-based encoding and hence supporting ranges in the tables? |
| 15:13 | <annevk> | yeah that one is hard too |
| 15:13 | <gsnedders> | It is pretty much the most horrible encoding out there. |
| 15:21 | <Ms2ger> | That's a pretty strong statement |
| 15:22 | <Ms2ger> | "CSS Selectors as Fragment Identifiers Community Group" |
| 15:23 | <jgraham> | Wow that's quite soemthing to form a community around |
| 15:23 | <beverloo> | there's a CG to put forward a proposal for <picture> too |
| 15:24 | <jgraham> | Presumably they will hold their meetings in the annex of the parish hall alongside the "insects in local history" society |
| 15:25 | <Ms2ger> | Is that next to the astrophysicists? |
| 15:27 | <jgraham> | Pretty sure astrophysics would draw more of a crowd mathematics |
| 15:27 | <jgraham> | +than |
| 15:29 | <Velmont> | Hmm. Presumably Microsoft will ship with CORS enabled for sync XHR. And then it'll be quite impossible to change that given MS's backwards compat never-break-anything policy. |
| 15:30 | <Velmont> | Has anyone reached out? |
| 15:30 | <divya> | astearns: sup sup |
| 15:30 | <Velmont> | Mozilla wanted to remove it, if that is still happening. |
| 15:38 | <annevk> | Velmont: haven't seen much recently on that front |
| 15:51 | <Velmont> | annevk: Probably needs to happen now, -- before Microsoft releases IE10. |
| 15:52 | <Velmont> | Microsoft was also hatin' on sync XHR in their blog post, -- so maybe they'd be up for removing CORS from sync XHR. |
| 15:55 | <annevk> | hmm yeah |
| 15:55 | <annevk> | maybe email them? |
| 15:55 | <annevk> | hmm |
| 16:26 | <jgraham> | MikeSmith: https://github.com/jgraham <-- any chance we can arrange for these to be moved to the W3C account and a cron job set up to auto sync them with the hg copy (should be OK if we don't ever push directly to master on git) |
| 16:26 | <jgraham> | That is, all pushes much go first to hg master and then to git master |
| 16:26 | <Ms2ger> | AryehGregor, so I've got an issue with Range-intersectsNode.html... |
| 16:26 | <Ms2ger> | It's timing out after 540769ms |
| 16:26 | <jgraham> | But they can ofc live on a branch in git first if the author is sane^W^Wprefers |
| 16:27 | <Ms2ger> | s//in/ |
| 16:40 | <Velmont> | https://github.com/jgraham/testharness.js/pull/1 << ohwell, others might want to review as well: Pull Request #1: Add new DOMExceptions, from DOM4 and from IndexedDB |
| 16:50 | <Ms2ger> | Velmont, that doesn't seem to make sense |
| 16:51 | <Ms2ger> | I doubt we want assert_throws("SYNTAX_ERR", fn) and assert_throws("SyntaxError", fn) to test different things |
| 17:01 | <arv> | annevk: Why are WebIDL platform arrays not quite JS arrays? |
| 17:03 | <Ms2ger> | Because they can be live? |
| 17:05 | <dglazkov> | good morning, Whatwg! |
| 17:05 | <Ms2ger> | Good day to you too |
| 17:28 | <kennyluck> | shit… is http://krijnhoetmer.nl/irc-logs/whatwg/ dead already? |
| 17:29 | <Philip`> | kennyluck: Works for me |
| 17:30 | <kennyluck> | Philip`, OK, good to know. Let me see if I can make it work for me. |
| 17:49 | <TabAtkins> | What's the incantation to summon david_carlisle? Do I just have to mutter "MathML" under my breath? |
| 17:50 | <Ms2ger> | You could try "LaTeX", and hope he comes in to express his distress |
| 17:52 | <TabAtkins> | I can't parse the way that different services often use different communication means even |
| 17:52 | <TabAtkins> | Let's try that again. |
| 17:53 | <TabAtkins> | I can't parse "the way that different services often use different communication means even |
| 17:53 | <TabAtkins> | when sharing the same underlying protection technology. |
| 17:53 | <TabAtkins> | " |
| 17:53 | <TabAtkins> | Dammit, replaying linebreaks. |
| 17:54 | <Ms2ger> | DRM? |
| 17:54 | <TabAtkins> | Yes, latest email from Adrian Bateman. |
| 17:54 | <TabAtkins> | I just don't understand his sentence structure. |
| 17:54 | <gsnedders> | Yeah, I can't parse it at all. |
| 17:54 | <Ms2ger> | That's because you're not on a supported platform |
| 17:56 | <gsnedders> | [[(the way (that different sevices)) (often) (use) (different communication)] (means) (even when) (sharing) (the same (underlying) (protection) technology)]] maybe? |
| 17:56 | <gsnedders> | s/]]/]/ |
| 17:56 | <TabAtkins> | Your diagramming is not enlightening. |
| 17:56 | <Ms2ger> | (different communication means)? |
| 17:57 | <annevk> | arv: ask heycam|away |
| 17:57 | <gsnedders> | Ms2ger: No, noun phrase followed by another noun phrase |
| 17:57 | <Ms2ger> | Why? |
| 17:57 | <gsnedders> | Ms2ger: Maybe a noun phrase pre-modifying another noun phrase, actually… |
| 17:57 | <gsnedders> | i.e., ((different communication) means) |
| 17:58 | Ms2ger | should not argue with linguists |
| 17:58 | <gsnedders> | Ms2ger: I can't tell you exactly for the very reason NLP is hard |
| 17:58 | <Ms2ger> | Is it NP-hard? |
| 17:59 | <Ms2ger> | TabAtkins, fwiw, hg rebase is known to delete random files occasionally |
| 17:59 | <gsnedders> | Depends on the language, but typically far harder, if it can indeed be solved by a Turing machine |
| 17:59 | <TabAtkins> | Ms2ger: Interesting. |
| 17:59 | <gsnedders> | (which is a questionable subject) |
| 17:59 | <TabAtkins> | I only know git rebase, so I assumed it was similar. |
| 18:00 | <Ms2ger> | I dunno, I presume git is at least as buggy :) |
| 18:00 | <TabAtkins> | Also, OMG only four more days: http://www.youtube.com/watch?v=bG2mdZ23eP8 |
| 18:00 | <TabAtkins> | A lot of us in webkit use rebase workflows, and I havent' heard complaints, so shrug. |
| 18:00 | <gsnedders> | (English for example *likely* only parsable unambiguously by a machine more powerful than a Turing machine — probably an Oracle machine, or at least that's my opinion) |
| 18:01 | <TabAtkins> | So you're implying that humans can't parse english unambiguously? |
| 18:01 | <gsnedders> | TabAtkins: Yes. |
| 18:01 | <gsnedders> | TabAtkins: The fact humans often misunderstand each other suggests that. |
| 18:01 | <TabAtkins> | I would put out the alternative hypothesis that English is fundamentally ambiguous is some circumstances. |
| 18:02 | <Ms2ger> | http://24.media.tumblr.com/tumblr_m06xeh4nf21rqvy12o1_500.jpg < how did webkit memes get a picture of bz's brain? |
| 18:02 | <gsnedders> | TabAtkins: English is fundementally ambiguous in some circumstances, so you need an Oracle machine to unambiguously parse it. |
| 18:02 | <gsnedders> | (Which obviously is impossible) |
| 18:02 | <TabAtkins> | "Fundamentally ambiguous" does not jive with "unambiguous to a sufficiently powerful machine". |
| 18:03 | <gsnedders> | TabAtkins: An Oracle machine can solve the halting problem. |
| 18:03 | <TabAtkins> | Yes? |
| 18:03 | <TabAtkins> | The halting problem is perfectly solveable if you have infinite time. |
| 18:03 | <TabAtkins> | It's not ambiguous. |
| 18:03 | <gsnedders> | TabAtkins: An Oracle could determine intent, no? |
| 18:04 | <TabAtkins> | A TM could determine intent, no? |
| 18:04 | <gsnedders> | Only in a subset of cases. |
| 18:05 | <gsnedders> | (Adrian's sentence is likely a fragment, as far as I can tell, with only what should be an embedded clause present) |
| 18:05 | <TabAtkins> | You're asserting that intent determination is sometimes equivalent to the halting problem, rather than simply being ambiguous at times. |
| 18:05 | <TabAtkins> | I don't find that obvious. |
| 18:06 | <gsnedders> | TabAtkins: On the whole, intent can likely be solved using a TM, but in practically infeasible time, but such knowledge of psychology needed for it is likely beyond us. |
| 18:06 | <gsnedders> | But this is all massively theoretical. |
| 18:07 | <TabAtkins> | If we're talking about arbitrary brains, I'll agree with you. But human brains probably require at most exponential simulations for a good reading of intent. |
| 18:07 | <TabAtkins> | And since human brains are limited in size, we're okay. |
| 18:07 | <TabAtkins> | exponential-in-the-size-of-the-mind |
| 18:07 | <gsnedders> | Ms2ger: Also, trying to parse stuff like that off the top of my head without paper to scribble on is hard :) |
| 18:07 | <gsnedders> | TabAtkins: Also likely related to time. |
| 18:08 | <gsnedders> | The older someone is the more knowledge they have to complicate it. |
| 18:08 | <gsnedders> | So it's not just the size of the mind, but what it can recall. |
| 18:08 | <TabAtkins> | Not if you're bruteforcing. A hard-drive with more data isn't any heavier. |
| 18:09 | <gsnedders> | Right. |
| 18:10 | <TabAtkins> | I expect the whole-mind-simulation problem to break down in the next century at most. |
| 18:10 | <TabAtkins> | And from there it's an engineering problem to, for practical purposes, solve human psychology. |
| 18:10 | <gsnedders> | (A trivial and typical example of NLP being hard is a discourse of two people A and B, as follows: A: "Do you want to go for lunch?"; B: "Sorry, haven't got paid yet this week.") |
| 18:11 | <gsnedders> | How the two sentences relate is hard to computational determine |
| 18:11 | <gsnedders> | s/sentences/utterences/ |
| 18:11 | <gsnedders> | TabAtkins: Leads to interesting questions of free-will, though. |
| 18:11 | <TabAtkins> | Free-will is an incoherent concept, so we're in the clear. |
| 18:12 | <gsnedders> | TabAtkins: How so? If the mind can be simulated, is it not inherently predicable? Yes, the system is that of the whole universe pretty much given inter-personal communication and subjects such as astronomy. |
| 18:13 | <gsnedders> | If it is entirely predicable, that implies there is no free-will, as every event within the mind in the future is predicable. |
| 18:13 | <TabAtkins> | The universe is a combination of deterministic and random processes. Nowhere in there is there space for free-will. So, it doesn't exist. |
| 18:14 | <gsnedders> | Another hard problem with NLP is the effect of culture on a language. |
| 18:15 | <gsnedders> | An Indian English speaker will rarely say, "No", preferring "Yes, but [some reason for saying no |
| 18:15 | <gsnedders> | ]" |
| 18:15 | <gsnedders> | If an English person said the same thing, it would imply they would do the action but without agreeing with it. |
| 18:16 | <gsnedders> | If an Indian person said it, it would imply they wouldn't do it. |
| 18:16 | <gsnedders> | That's a fairly radical difference. |
| 18:20 | <astearns> | the English version of "yes, but…" meaning "no" is "that's a good idea" |
| 18:21 | <TabAtkins> | "imma let you finish, but" |
| 18:21 | <gsnedders> | astearns: Nah, that depends on tone-of-voice. |
| 18:23 | GPHemsley | grumbles something about the W3C validator considering "HTML5 Validator Error" a single type... |
| 18:25 | <gsnedders> | (FWIW, I'm probably actually less interested in NLP than one might expect given my degree) |
| 18:26 | <TabAtkins> | Man, if it weren't for all the crazy stuff, SVG fonts would be pretty awesome. |
| 18:26 | <shepazu> | which specific crazy stuff? |
| 18:27 | <TabAtkins> | The fact that you can use all of SVG in a <glyph>. |
| 18:27 | <StoneCypher> | gsnedders: what's your degree? |
| 18:27 | <shepazu> | everyone has their own list of "crazy stuff" |
| 18:27 | <TabAtkins> | If it was just the basic geometry stuff and transforms, it'd be simple and great. |
| 18:27 | <Ms2ger> | StoneCypher, linguistics + CS |
| 18:28 | <StoneCypher> | ah. neat. |
| 18:28 | StoneCypher | often wonders why that cluster of people doesn't generate more programming languages |
| 18:28 | <gsnedders> | StoneCypher: Technically English Language (i.e., linguistic study of English) & Computing Science, but for the former I'm mostly doing generic linguistics options. |
| 18:28 | <TabAtkins> | StoneCypher: They get too caught up in ambiguities. |
| 18:28 | <StoneCypher> | gsnedders: neat |
| 18:28 | <StoneCypher> | TabAtkins: makes sense |
| 18:29 | <miketaylr> | somehow i got put on an unremovable NLP mailing list from NYU... |
| 18:29 | <StoneCypher> | gsnedders: does that mean you have a special place in your heart for infocom games? |
| 18:29 | <TabAtkins> | shepazu: Thoughts about hosting SVG2 on XML-ER? |
| 18:29 | <gsnedders> | StoneCypher: Heh, not really. |
| 18:30 | <gsnedders> | StoneCypher: My main interest in them is actually more in the implementation. |
| 18:31 | <StoneCypher> | well |
| 18:31 | <StoneCypher> | there's an old language called AGT, or a more modern one called Inform 7, which may each be of interest |
| 18:31 | <StoneCypher> | inform has a fascinatingly deep debugger |
| 18:40 | <zewt> | miketaylr: the shorthand for "unremovable mailing list" is "spam", btw |
| 18:41 | <miketaylr> | heh |
| 18:41 | <Philip`> | Perl was explicitly influenced by linguistics, but that seems rarer in other languages |
| 18:42 | <zewt> | perl and one of the most absurd structures in programming: "x if y" |
| 18:42 | <zewt> | python, for some reason, decided to one-up it with "y if x else z" |
| 18:42 | <zewt> | it's like someone took yoda and stuck him in a blender |
| 18:43 | <Philip`> | Perl's version sounds perfectly natural if you want to emphasize the action over the condition |
| 18:44 | <zewt> | flow control should be written in order of evaluation |
| 18:44 | <Philip`> | If you want to emphasize the condition over the action, then you can do it that way too |
| 18:46 | <Philip`> | (Python's version seems much more rarely natural to me) |
| 18:46 | <zewt> | c's ternary expressions are much saner; python's are unnatural, hard to read and hard to write |
| 18:46 | <zewt> | the possible results are on the outside, sandwiching the condition in between--wtf? |
| 18:47 | <zewt> | "x()? 5: 2" is much easier to read (especially at a glance) than "5 if x() else 2" |
| 18:47 | <zewt> | every language has at least one wart; python has a couple, and that's one of them :) |
| 18:49 | Philip` | agrees about preferring ?:, though it might not hurt to have nicer syntax |
| 18:52 | <Philip`> | I think I use "x if y" most commonly in situations like 'print "foo: $foo\n" if $verbose;', where the condition would get in the way and disrupt the flow when reading the code if it was the most prominent thing on the line (or even worse if it was forced onto multiple lines) |
| 18:55 | <zewt> | that annoys me, because it means as I'm reading the code I'm thinking "it does A, B, C, then prints $foo, then"--then all stop because suddenly there's flow control retroactively injected into something earlier |
| 19:12 | <kennyluck> | Is test262 public to public submission? |
| 19:12 | <Ms2ger> | No |
| 19:12 | <Ms2ger> | ECMA members only |
| 19:12 | <Ms2ger> | My repo on bitbucket, otoh, is open to everyone ;) |
| 19:19 | <kennyluck> | Ms2ger, huh that sucks. Thanks for your redirection by the way. |
| 19:22 | <kennyluck> | Ms2ger, I couldn't find your js tests though. Do they fall under web-tests? |
| 19:22 | <Ms2ger> | Yeah |
| 19:23 | <Ms2ger> | There's only one of them, though :) |
| 19:36 | <bga> | http://www.w3.org/TR/css3-transforms/ |
| 19:37 | <bga> | examples w/ errors |
| 19:37 | <bga> | .container > :last-child |
| 19:37 | <bga> | should be .container > ::last-child or .container :last-child |
| 19:38 | <kennyluck> | bga, I you can file the bug here → https://www.w3.org/Bugs/Public/enter_bug.cgi?product=CSS |
| 19:38 | <kennyluck> | s/I// |
| 19:38 | <TabAtkins_> | bga: ::last-child isn't a thing. |
| 19:38 | <bga> | ah |
| 19:39 | <TabAtkins_> | That example is completely correct. |
| 19:39 | <bga> | isnt pseudo element |
| 19:39 | <kennyluck> | ah, I missed that. |
| 19:43 | <TabAtkins_> | Sigh at Glenn trying (badly) to misdirect the conversation. |
| 19:43 | <TabAtkins_> | Bringing up a no-op DRM module as it was relevant to the concerns Henri's outlining? Really? |
| 19:45 | <TabAtkins_> | s/as/as if/ |
| 19:46 | <Ms2ger> | DRM? Really? |
| 19:46 | <TabAtkins_> | Is that sarcasm? |
| 19:46 | <Ms2ger> | I hope the DRM proposal is sarcasm, if that's what you mean |
| 19:47 | <Ms2ger> | hsivonen, btw, did you review http://dvcs.w3.org/hg/webcomponents/raw-file/tip/spec/templates/index.html#parsing ? |
| 20:02 | <AryehGregor> | Ms2ger, the actual original Range-intersectsNode.html running in testharness.js times out, or some version adapted to mochitest does? |
| 20:03 | <Ms2ger> | Mochitest |
| 20:03 | <Ms2ger> | This might just be because I'm on a debug build |
| 20:04 | <AryehGregor> | Running the actual file takes like two or three seconds for me. |
| 20:04 | <AryehGregor> | 50-whatever seconds seems amazingly excessive. |
| 20:05 | <AryehGregor> | I'm not sure I'm any better situated to debug the fact that it's taking a long time than you are. |
| 20:07 | <TabAtkins_> | Scumbag linux. >_< |
| 20:08 | <TabAtkins_> | nvidia |
| 20:08 | <TabAtkins_> | /nvidia |
| 20:10 | <Ms2ger> | Blame nvidia |
| 20:10 | <Ms2ger> | Linux is FOSS, it can't be buggy |
| 20:14 | <gsnedders> | No, that's seL4. |
| 20:15 | <gsnedders> | (not that that is open-source) |
| 20:16 | <jamesr__> | Ms2ger, try some open source graphics drivers on linux for a laugh |
| 20:19 | <gsnedders> | How? My browser blacklists them. |
| 20:19 | <gsnedders> | :P |
| 20:37 | <jwalden> | kennyluck: test262 isn't even open to some ECMA members due to the legal-policy idiocy that's complicating submitting our existing tests to them :-\ |
| 20:37 | <kennyluck> | how sad |
| 20:39 | <Ms2ger> | jwalden, btw, SM's score on my ES test suite is horrible ;) |
| 20:44 | <gsnedders> | Ms2ger: Your testsuite? |
| 20:45 | <Ms2ger> | Yeah, one test I wrote this week |
| 20:45 | <gsnedders> | I was more looking for URLs :P |
| 20:45 | <Ms2ger> | bitbucket.org/ms2ger/web-tests off the top of my head |
| 20:47 | <gsnedders> | Anyway to run it without clioning the repo? |
| 20:48 | gsnedders | is lazy |
| 20:48 | <gsnedders> | Or just tell me how we do :P |
| 20:49 | <kennyluck> | Ms2ger, let me guess…. SM scores 0 right? :p |
| 20:49 | <Ms2ger> | gsnedders, you pass everything, actually |
| 20:49 | <gsnedders> | Ms2ger: That's what I suspected, on the whole |
| 20:49 | <Ms2ger> | Heh |
| 20:50 | <gsnedders> | We tend to be good at following the spec in general. |
| 20:50 | <gsnedders> | The fact we were basically written against ES3.1 drafts probably helps compared with decade old engines. :) |
| 20:51 | gsnedders | wonders if Chakra development only started after ES5's publication. |
| 20:51 | <Ms2ger> | ms2ger.freehostia.com/tests/various/js/Array.prototype.join-order.html |
| 20:52 | <Ms2ger> | http://ms2ger.freehostia.com/tests/various/js/Array.prototype.join-order.html when Firefox is helpful |
| 20:52 | <Hixie> | anyone want to edit an http+aes:// url scheme spec? |
| 20:52 | <Ms2ger> | No |
| 20:52 | <Hixie> | i suppose i could just stick it in the HTML spec |
| 20:54 | <Hixie> | it's basically "do exactly what http says except decode the message body (after decoding any transfer-encoding applied to the entity body) using AES-CTR with the key given in the url" |
| 20:55 | <gsnedders> | Hixie: Isn't that practically another transfer-encoding? |
| 20:55 | <Hixie> | more or less |
| 20:55 | <Hixie> | but the key comes from the url |
| 20:55 | <Hixie> | so we can't use another t-e |
| 20:56 | <gsnedders> | Does RFC2616bis require a specific handling for userinfo, or could you use that? |
| 20:56 | <Hixie> | zewt: what's the spec out of sync with the whatwg spec that doesn't mention 'missed cues'? |
| 20:56 | <Hixie> | gsnedders: 2616 doesn't mention userinfo. haven't checked 2616bis. |
| 20:57 | <gsnedders> | Hixie: Well, yeah, when 2616 was written the URI spec didn't have such a concept. |
| 20:57 | <Hixie> | what's hte url for the latest http spec? |
| 20:58 | <gsnedders> | http://tools.ietf.org/wg/httpbis/ |
| 20:59 | <Hixie> | christ |
| 20:59 | <Hixie> | what a mess |
| 20:59 | <gsnedders> | End of 2.7.1, in messaging |
| 21:00 | <Ms2ger> | Hixie, you must be new here |
| 21:00 | <zewt> | http://dev.w3.org/html5/spec/video.html vs http://www.whatwg.org/specs/web-apps/current-work/multipage/the-video-element.html |
| 21:01 | <gsnedders> | Hixie: Using userinfo and t-e seems cleaner than using a new scheme, IMO |
| 21:01 | <Hixie> | gsnedders: oh, interesting approach |
| 21:02 | <gsnedders> | (I could be speaking of something infeasible here, I've been in JS-land for a while.) |
| 21:02 | <Hixie> | zewt: dunno how you got that dev.w3.org url, but that's an ancient file |
| 21:02 | <Hixie> | zewt: you want http://dev.w3.org/html5/spec/media-elements.html |
| 21:03 | <zewt> | Hixie: then it should be removed, since that's where I've ended up every time I've viewed that spec in a long time |
| 21:03 | <Ms2ger> | Hixie, http://dev.w3.org/html5/spec/, click "the video element" |
| 21:03 | <Hixie> | zewt: ask mike? |
| 21:03 | <zewt> | i just type "video" in firefox to get there |
| 21:03 | <Ms2ger> | Oh |
| 21:03 | <Ms2ger> | Nvm me |
| 21:03 | <Ms2ger> | The bug is assigned to Mike, IIRC |
| 21:04 | <Hixie> | anyway, use the single-page whatwg copy, you'll avoid these problems :-) |
| 21:04 | <zewt> | Hixie: that hardly addresses the problem |
| 21:04 | <Ms2ger> | Then you've just got other problems :) |
| 21:04 | <Hixie> | gsnedders: finally got there... |
| 21:04 | <Hixie> | but lunch now |
| 21:04 | <zewt> | as much as people say "use the editor's drafts", TR's are still a problem |
| 21:04 | <Hixie> | bbl |
| 21:27 | <Velmont> | Ms2ger: SYNTAX_ERR is legacy, SyntaxError is new. -- And I think the old code didn't test for name, -- and new should test for type=SyntaxError. |
| 21:52 | <TabAtkins_> | Oh jeez, my computer's broken now. |
| 21:52 | <TabAtkins_> | Wedged in some inconsistent state after attempting to update graphics drivers. |
| 21:53 | <zewt> | fyture |
| 21:53 | <zewt> | f |
| 21:53 | <zewt> | future |
| 22:12 | <bga> | http://www.fxitech.com/products/ |
| 22:32 | <Hixie> | gsnedders: i think requiring CDNs to send transfer encoding headers may be too high a bar |
| 22:39 | <TabAtkins_> | Okay. All done. Back to where I was before, I think. And now Chrome has its accelerated compositing disabled, so my browsing won't be interrupted by a transforms-heavy page crashing metacity. |
| 22:54 | <Hixie> | how do you even refernece this httpbis beast |
| 22:54 | <TabAtkins_> | in verse |
| 23:10 | <jamesr__> | TabAtkins_, how do i make a new CSSRule to stick in this list? |
| 23:11 | <TabAtkins_> | That's an excellent question. |
| 23:11 | <jamesr__> | lawl "To parse a CSS rule ..." |
| 23:11 | <jamesr__> | http://dvcs.w3.org/hg/cssom/raw-file/tip/Overview.html/#parse-a-css-rule |
| 23:12 | <jamesr__> | ok i'm pretty sure i inserted a rule but nothing happened |
| 23:13 | <TabAtkins_> | Just insert a <style> and set its textContent. |
| 23:13 | <TabAtkins_> | Fuck all the rest. |
| 23:14 | <TabAtkins_> | Every part of the CSSOM is horrible. |
| 23:28 | <Hixie> | zewt: why https://www.w3.org/Bugs/Public/show_bug.cgi?id=16157 ? |
| 23:34 | <Hixie> | "I'm not talking about any kind of lock-in" coming from someone pushing a DRM scheme is rather hard to swallow |
| 23:39 | <zewt> | Hixie: because the failure cases is too exceptional, and will lead to too much buggy code |
| 23:39 | <zewt> | also, are |
| 23:40 | <zewt> | also, most of the time there's nothing sane you can do in response to that exception anyway, except to manually remove the unpaired surrogates and try again |
| 23:42 | <Hixie> | you should never hit that exception unless you have a pretty fundamental design bug |
| 23:42 | <Hixie> | in the latter case you'll hit it relatively early in the design |
| 23:42 | <Hixie> | (if you're not unlucky) |
| 23:42 | <zewt> | it'll happen if the user pastes in an unpaired surrogate, or if you try to chop off the first 100 letters of user-entered text naively, etc |
| 23:42 | <Hixie> | the user can't paste in unpaired surrogates |
| 23:43 | <Hixie> | your latter point is an interesting one though |
| 23:43 | <zewt> | sure he can, at least in firefox last i tried it (maybe around 6 or 7) |
| 23:43 | <Hixie> | file a bug on firefox, that's a bug |
| 23:43 | <zewt> | but granted that's a rare case |
| 23:49 | <Hixie> | hmmm |
| 23:49 | <Hixie> | Path and 2DRContext |
| 23:50 | <Hixie> | i don't want to duplicate all this material in both interface blocks |
| 23:50 | <Hixie> | [NoInterfaceObject] and foo implements bar; results in the same as just copying the idl in both places right? |
| 23:51 | <Hixie> | i wonder if i should do an idl block for each related block of content, or if i should just have one big one imported twice |
| 23:51 | <Hixie> | i'm gonna reuse the transformation stuff for patterns, so probably at least some smaller-sized blocks would make sense... |
| 23:52 | <Hixie> | this is gonna be a pain to spec |
