| 00:03 | <TabAtkins_> | ...what the hell? |
| 00:04 | <smokeyrat> | What the hell what? |
| 00:05 | <TabAtkins_> | Your bizarro train-of-thought spew. |
| 00:08 | <smokeyrat> | Bizarro? |
| 00:08 | <smokeyrat> | It is a very good idea. |
| 00:08 | <smokeyrat> | And something I've wanted for ages. |
| 00:09 | <tantek> | Is there a REQUEST-PONY header in HTTP 2.0? |
| 00:10 | <smokeyrat> | Why are you making fun of this? |
| 00:10 | <smokeyrat> | Do you have any idea how this would revolutionize the Internet? |
| 00:10 | <tantek> | is "Windows 95 style commercial jippo" not a humorous reference? |
| 00:10 | <smokeyrat> | I meant that it would be big-scale. |
| 00:10 | <smokeyrat> | Something that everyone hears about. |
| 00:10 | <tantek> | or "all major browsers coming out within a week with full support for" |
| 00:10 | <tantek> | (which has never happened on any spec/feature AFAIK) |
| 00:10 | <smokeyrat> | Yes, "very soon". |
| 00:11 | <tantek> | or "super-safe payments" (what does that mean?) |
| 00:11 | <smokeyrat> | Maybe they have had some kind of secret board meeting and agreed on this and worked behind the scenes. |
| 00:11 | <smokeyrat> | The point is that it would be truly global, instant, no hazzles and one-click. |
| 00:11 | <tantek> | or "totally free from transaction fees" is that a wish? or a requirement in a standard? or an expected market outcome? |
| 00:12 | <smokeyrat> | Or the transaction fee would have to be truly in %, so if the payment is $0.1, the fee truly is just a fraction of that. No minimum charges. |
| 00:12 | <smokeyrat> | This would open so many doors. |
| 00:12 | <tantek> | sounds like you're asking for some kind of micropayment standard |
| 00:12 | <tantek> | searching for that... |
| 00:12 | <smokeyrat> | I would sure as hell pay $0.1 to support a nice game or article, but I would not bother today to send that amount. |
| 00:13 | <smokeyrat> | No... I mean that it should be right into the HTTP protocol, thus in the native browser UI. |
| 00:13 | <tantek> | the HTTP protocol is a standard |
| 00:13 | <tantek> | and $0.1 would be a micropayment |
| 00:13 | <tantek> | so I repeat, sounds like you're asking for some kind of micropayment standard |
| 00:14 | <tantek> | looking up micropayment... |
| 00:14 | <tantek> | http://en.wikipedia.org/wiki/Micropayment#History |
| 00:14 | <tantek> | looks like there is bad news |
| 00:14 | <tantek> | "During the late 1990s, there was a movement to create microtransaction standards,[5] and the World Wide Web Consortium (W3C) worked on incorporating micropayments into HTML even going as far as to suggest the embedding of payment-request information in HTTP error codes.[4] The W3C has since stopped its efforts in this area,[4]" |
| 00:15 | <tantek> | If you'd like to propose a micropayment standard e.g. as an extension of HTTP, perhaps consider attempting to start a W3C Community Group on "HTTP micropayments" and see if you can get other people involved: http://www.w3.org/community/ |
| 00:16 | <smokeyrat> | I can't propose anything. It will have to come from the big boys who people listen to. |
| 00:16 | <smokeyrat> | That they have not sat down in a conference room years ago is a mystery. |
| 00:16 | <smokeyrat> | There must be some kind of evil conspiracy requiring them to not make it easy for people to pay and charge. |
| 00:16 | <tantek> | most successful standards start from individuals with good ideas, not "big boys" |
| 00:17 | <tantek> | from the Wikipedia article, sounds like they did discuss it (whether in a conference room or not is unknown) and eventually stopped efforts. |
| 00:17 | <tantek> | no evil conspiracy is required for inaction. that's simply inertia. |
| 00:19 | <tantek> | anyway, if you want to make it happen, try starting a community group to see if there is interest from others. if no one else is interested, then it means it's not worth creating a standard for it. if there are others interested, then you have to do the hard work of figuring out how it would work. |
| 00:23 | <smokeyrat> | :| |
| 00:24 | <smokeyrat> | It's obvious that it's needed. |
| 00:27 | <TabAtkins_> | Not necessarily. Micropayments have been attempted in lots of forms, both online and off. They don't work very well. Human psychology makes a huge deal of the difference between "free" and "not free", even if the "not free" is "so cheap it's practically free". |
| 00:28 | <TabAtkins_> | So, asking someone to pay 1 cent is psychologically equivalent to asking them to pay a few bucks, in terms of how willing the average person is willing to do it. |
| 00:28 | <TabAtkins_> | This might be changeable - it's not clear if this is something built into our psyche, or if it's a result of microtransactions never really being viable in the real world before the last few decades. |
| 00:28 | <TabAtkins_> | But resistance will be high regardless. |
| 00:30 | <TabAtkins_> | (There were simple experiments with real-world things, like giving people a choice between a 15c truffle and a 1c hershey kiss - 3/4 chose the truffle. But if you drop the kiss to free, 2/3 choose it instead. Most people's brains break when they see "free".) |
| 00:32 | <tantek> | TabAtkins - the positive examples are music downloads (iTunes, Amazon) for < $1, and ringtones as well |
| 00:32 | <tantek> | as well as all those things people buy on Zynga games Farmville etc. |
| 00:32 | <tantek> | so there is some psychological precedent for it |
| 00:33 | <tantek> | especially in the use-case mentioned of (HTML5) games |
| 00:33 | <tantek> | why should Zynga make (nearly?) all the online/web game micropayment $? |
| 00:33 | <smokeyrat> | TabAtkins_: Well, the price example was just an example. |
| 00:34 | <TabAtkins_> | tantek: That's true. We do have a lot more microtransactions in today's world. |
| 00:34 | <smokeyrat> | But even paying $1 is ridiculous today. It's eaten up by fees and is not anywhere near instant or hazzle-free. |
| 00:34 | <smokeyrat> | In fact, it's a nightmare of credit card info, codes, bank hardware dongles and fees. |
| 00:34 | <smokeyrat> | Or paying $10. |
| 00:36 | <tantek> | smokeyrat, Apple, Amazon, Zynga and others have figured out how to make < $1 payments work, perhaps you could research how they've done it and report back your results. |
| 00:37 | <tantek> | oh and please post your research on the web somewhere, preferably at a permalink with a creative commons / public domain license - that helps for work towards a standard |
| 00:39 | <smokeyrat> | For them there isn't anything to figure out. It's a closed system with lots of users. |
| 00:40 | <smokeyrat> | They enter cash and then they have some virtual credits to spend instantly. |
| 00:40 | <smokeyrat> | Couldn't be simpler. |
| 00:40 | <tantek> | smokeyrat - perhaps you could write up your analysis in a blog post |
| 00:40 | <TabAtkins_> | You're thinking of Zynga. Tantek also pointed out Apple and Amazon, which do microtransactions directly off your CC. |
| 00:41 | <smokeyrat> | I was not aware that they do that. But that would require CC entering, etc. |
| 00:41 | <smokeyrat> | So basically they store your CC info. Is that novel? |
| 00:41 | <TabAtkins_> | Storing CC information is trivial - browsers do that already for autocomplete. |
| 00:41 | <smokeyrat> | That went well for Sony... |
| 00:41 | <TabAtkins_> | The difficulty appears to lay elsewhere. |
| 00:46 | <zewt> | TabAtkins_: well, browsers seem to attempt to *not* store CC info |
| 00:46 | <zewt> | (but that's a special case) |
| 00:53 | TabAtkins_ | just realized that </figcaption> isn't an implied end tag. :/ |
| 02:12 | <guzzlefry> | should the xml:lang attribute be used in HTML5? |
| 02:13 | <MikeSmith> | guzzlefry: no not in text/html documents |
| 02:13 | <zewt> | just use @lang |
| 02:14 | <guzzlefry> | k, thanks |
| 05:01 | <guzzlefry> | Does HTML5 have implicit end tags like older versions (pre-XHML)? |
| 05:02 | <guzzlefry> | s/XHML/XHTML |
| 05:23 | <daurnimator> | do you guys know of some way to test a http client? I wrote one yesterday and what to check it against strange corner cases. connection losses, timeouts, shutdowns; broken ssl handshakes and unexpected re-negotiations. etc |
| 05:40 | <MikeSmith> | daurnimator: I thought the HTTPbis WG had a test suite |
| 05:43 | <daurnimator> | MikeSmith: not much of one... and definiently not one that really tests the underlying tcp connection |
| 05:44 | <MikeSmith> | I see |
| 05:44 | <MikeSmith> | maybe could use some tests from browser projects |
| 07:40 | <zcorpan> | TabAtkins_: @import should maybe use the same rules as <link> when it comes to requiring text/css type and quirks mode |
| 07:40 | <zcorpan> | TabAtkins_: opera seems to use the same logic while chrome and firefox allow text/plain also in standards mode |
| 07:46 | <zcorpan> | TabAtkins_: hmm, chrome seems to allow text/html cross-origin @imports which i think is a security problem |
| 07:46 | <zcorpan> | in quirks mode |
| 09:00 | <jgraham> | The internet is funny. CERN live streams a seminar intended for research scientists at which Higgs boson results are announced. Internet complains that a) research level particle physics is hard to follow and b) they used comic sans in the presentation |
| 09:01 | <shwetank> | they wanted an apple keynote of the whole thing |
| 09:02 | <moo-_-> | are we confident that Higgs is smaller than pixels in retina display? |
| 09:11 | <jgraham> | shwetank: I guess. Sadly the real world doesn't work like that |
| 09:11 | <jgraham> | (also I feel great now I got to describe academica as "the real world") |
| 09:18 | <Lachy> | moo-_-, that depends, how do you define the mass of a pixel in the retina display? |
| 09:22 | <Philip`> | Lachy: Mass of screen divided by number of pixels? |
| 09:24 | <Lachy> | do you include the mass of the glass covering it, or just the mass of the phosphorus material? Or you could define it as the energy emitted from the pixel when it's on. |
| 09:34 | <Ms2ger> | var scrollTo = window.scrollTo; |
| 09:34 | <Ms2ger> | w(scrollTo); |
| 09:34 | <Ms2ger> | What should be logged? |
| 09:37 | <jgraham> | Ohn interesting |
| 09:38 | <jgraham> | Presumably it depends on whether var scrollTo creates a new global given that one already exists |
| 09:39 | <jgraham> | on the window object |
| 09:39 | <Ms2ger> | I, too, would presume that :) |
| 09:39 | Ms2ger | looks for gsnedders |
| 09:41 | <ashemedai> | ObligatoryVenting: not being able to discern at on-hover whether a link opens a JavaScript window on top of your existing window, loads content in your current window, or if it opens a new window/tab. |
| 09:41 | <ashemedai> | So frustrating and unclear at times. |
| 09:41 | <Ms2ger> | They used Comic Sans MS?! |
| 09:41 | <ashemedai> | Ms2ger: Yes, ultimate troll :) |
| 09:41 | <Ms2ger> | What, me? |
| 09:41 | <Ms2ger> | Thanks :) |
| 09:41 | <ashemedai> | The Comic Cans |
| 09:41 | <ashemedai> | Sans |
| 09:43 | <ashemedai> | Ultimate Science requires levity in the form of a funny font |
| 09:45 | <Philip`> | Surely real science should be conveyed with a marker pen and transparency and OHP |
| 09:45 | <Ms2ger> | Chalk, dammit |
| 09:45 | <jgraham> | Using comic sans for presentations is a tradition that dates back for as long as people have been making computer-based slideshows |
| 09:46 | <Ms2ger> | Though, the ultimate troll would be using Comic Sans MS in a LaTeX presentation |
| 09:46 | <jgraham> | In an ideal world each slide is a couple of words plus a graph anyway |
| 09:47 | <jgraham> | Nah, the ultimate troll was the guy that used 12pt Times New Roman |
| 10:00 | <Philip`> | Nothing beats a graph with red-squiggly-underlined words that's clearly been screenshotted from Excel |
| 10:03 | <odinho> | annevk: Did you see the svg? |
| 10:03 | <annevk> | odinho: no missed it |
| 10:03 | <Ms2ger> | Nothing beats a maintainer for those canvas tests... |
| 10:05 | <annevk> | odinho: pointer/ |
| 10:05 | <annevk> | ? |
| 10:05 | <odinho> | Computer strugling |
| 10:06 | <odinho> | omfg. some memory leak there! |
| 10:06 | <Ms2ger> | Opera had memory leaks first |
| 10:06 | <odinho> | Ms2ger: You bet'cha |
| 10:06 | <odinho> | Ms2ger: But then Fx came and owned us :-( |
| 10:06 | <Ms2ger> | :D |
| 10:07 | <odinho> | annevk: http://odin.s0.no/tmp/Hodges_odinho-CORS-Diagram-SimpleRequest.svg |
| 10:08 | <annevk> | that looks like the Core-Doxcs t-shirt :) |
| 10:08 | <odinho> | ... |
| 10:08 | <odinho> | Did I save over it :P |
| 10:08 | <Ms2ger> | Zing |
| 10:09 | <annevk> | euh |
| 10:09 | <annevk> | Core-Docxs |
| 10:11 | <annevk> | minor announcement: http://annevankesteren.nl/2012/07/leaving-opera |
| 10:11 | <odinho> | annevk: http://odin.s0.no/tmp/Hodges_odinho-CORS-Diagram-SimpleRequest.svg Tada |
| 10:11 | <annevk> | my IRC client still displays the same picture funnily enough |
| 10:11 | <odinho> | Hm. Although I used Ubuntu font, so maybe I should use something neutral like arial if it messes up other places. |
| 10:11 | <odinho> | annevk: IRC client with pictures? Wat? |
| 10:12 | <odinho> | My lingogi shows the correct one now. |
| 10:12 | <annevk> | LimeChat shows images inline |
| 10:12 | <annevk> | including SVG |
| 10:13 | <annevk> | odinho: instead of .com it should prolly say .example or something |
| 10:13 | <annevk> | odinho: also GET lyrics.com/hello.txt is wrong |
| 10:13 | <annevk> | it's GET /hello.txt |
| 10:14 | <annevk> | Host: lyrics.com |
| 10:14 | Philip` | thinks using a non-conventional TLD like ".example" would make the diagram harder to understand |
| 10:15 | <odinho> | annevk: It's not supposed to be real. It also misses the HTTP/1.1 stuff |
| 10:15 | <odinho> | I agree with Philip` |
| 10:15 | <odinho> | I'm just trying to make it easy for them newbies :P |
| 10:15 | <odinho> | But, lunch out in the sun! |
| 10:16 | <Philip`> | (since "webapps.com" is clearly a domain name, but "webapps.example" isn't, so it'd have to be written like "http://webapps.example"; to make it unambiguous but then it'd still look kind of weird) |
| 10:16 | <odinho> | won't anybody think of the children!!11 |
| 10:16 | <annevk> | yeah but there are some unwritten rules against using actual domain names |
| 10:17 | <smaug____> | annevk: Leaving Opera? What will you do? |
| 10:17 | <annevk> | smaug____: nothing much for now :) |
| 10:17 | <smaug____> | oh |
| 10:17 | <smaug____> | that is possibly a good choice |
| 10:18 | <smaug____> | and after that, you'd be welcome to Mozilla ;) |
| 10:20 | <smaug____> | annevk: does this change affect to your role as a spec editor? |
| 10:26 | <annevk> | smaug____: dunno, hope not too much, I like writing specs |
| 10:27 | <MikeSmith> | annevk: http://www.w3.org/Consortium/Recruitment/#techeditor |
| 10:28 | <annevk> | MikeSmith: the link makes it sound interesting, but then there's "for HTML Working Group" on the actual page :) |
| 10:28 | <smaug____> | what... " lead editor for the HTML Working Group" |
| 10:30 | <annevk> | the HTML WG is the new old HTML WG |
| 10:31 | <MikeSmith> | annevk: trail of broken hearts behind you. First you leave the CSS WG, then the HTML WG, and now this |
| 10:32 | <annevk> | MikeSmith: quit while you're ahead? |
| 10:33 | <MikeSmith> | heh |
| 10:33 | <MikeSmith> | are you ahead? |
| 10:33 | <MikeSmith> | annevk: leave them wanting more |
| 10:33 | <annevk> | dunno really, but soon I'm an unemployed bum and I like that idea |
| 10:35 | <annevk> | CSS WG was too much abuse, HTML WG was too much process abuse and irrelevant anyway given the WHATWG |
| 10:36 | <annevk> | I had no conflicts with Opera though, just felt like trying something else |
| 10:36 | <Ms2ger> | WCSSTWG? |
| 10:39 | <annevk> | CSS WG was actually also mostly process nonsense, like how to divide WDs rather than how to solve problems |
| 10:39 | <annevk> | combined with a madman in charge that just gives problems |
| 10:47 | <MikeSmith> | maybe the CSS WG would work better if it were more like the WebApps WG |
| 10:48 | <MikeSmith> | WebApps WG seem to somehow be able to get stuff done without anybody trying to micro-manage it all and layer further process on it |
| 10:49 | <MikeSmith> | anyway, question: Is there really no HTTP test suite anywhere? |
| 10:49 | <annevk> | IETF and testing? mwaha |
| 10:50 | <zcorpan> | is there a HTTP spec that doesn't suck? |
| 10:50 | <annevk> | jreschke has a couple of tests for some features though |
| 10:50 | <annevk> | zcorpan: rhetorical? |
| 10:51 | <MikeSmith> | it seems whacky that IETF culture does not see any need for requiring test suites for RFCs |
| 10:52 | <Ms2ger> | Why would you need tests? They're only requesting comments |
| 10:52 | <MikeSmith> | heh |
| 10:52 | <MikeSmith> | but don't browser projects have HTTP tests? |
| 10:52 | <annevk> | for RFCs it makes sense I guess, but not STDs |
| 10:52 | <jgraham> | We should rename W3C specs RFTs |
| 10:52 | <MikeSmith> | I would think Mozilla would at least have some HTTP tests |
| 10:52 | <jgraham> | I expect everybody has some HTTP tests |
| 10:53 | <jgraham> | Possibly not packaged as a suite |
| 10:53 | <Ms2ger> | Mm |
| 10:53 | <jgraham> | Or easy to reuse |
| 10:53 | <Ms2ger> | And ours run JS on the server |
| 10:54 | <Ms2ger> | So unless everyone wants to run a Gecko server... ;) |
| 10:55 | <MikeSmith> | wonder if it's time to write an HTTP spec |
| 10:56 | <ashemedai> | annevk: Eh? Unemployed? How so? |
| 10:56 | <ashemedai> | MikeSmith: As an update to HTTP or merely a more concise up-to-date spec? |
| 10:57 | <jgraham> | ashemedai: Oslo ran out of bacon cheeseburgers |
| 10:57 | <MikeSmith> | ashemedai: as a real spec to document how it actually works |
| 10:58 | <ashemedai> | jgraham: Noooooes! |
| 10:58 | <annevk> | ashemedai: well you know, when you quit your job and there's no direct follow-up, that's what happens |
| 10:58 | <annevk> | :) |
| 10:58 | <ashemedai> | MikeSmith: I can tell you that I know a bunch of folks who would be Real Happy(tm) if you do that. |
| 10:58 | <annevk> | MikeSmith: gsnedders worked on a part of that a long time ago |
| 10:58 | <ashemedai> | annevk: Ahh ok, it was a choice on your end. :) |
| 10:59 | <annevk> | MikeSmith: but never finished it unfortunately |
| 10:59 | <MikeSmith> | interesting |
| 10:59 | <MikeSmith> | I kind of vaguely remember that now |
| 10:59 | <annevk> | ashemedai: ah yeah, it was, and I'm welcome back ;) |
| 10:59 | <ashemedai> | annevk: Hehe, what kind of work do you want to do though? |
| 11:01 | <jgraham> | Here's a thing |
| 11:01 | <jgraham> | If I have a document with two framesets, is there a way to set an onload listener and tell *which* frameset loaded |
| 11:02 | <annevk> | jgraham: you mean two <frame>s? |
| 11:02 | <jgraham> | Like <frameset rows="*" onload="something"><frameset></frameset> |
| 11:02 | <jgraham> | I mean two framesets |
| 11:02 | <jgraham> | You can nest framesets |
| 11:02 | <annevk> | sure, but they do not load anything |
| 11:03 | <zcorpan> | <frameset onload> reflects window.onload |
| 11:03 | <zcorpan> | right? |
| 11:03 | <jgraham> | Oh wait |
| 11:03 | <jgraham> | I am misunderstanding this test |
| 11:04 | <annevk> | zcorpan: right |
| 11:04 | <jgraham> | I thought it was giving a load event on the frameset when all the subframes had loaded |
| 11:04 | <jgraham> | But it isn't |
| 11:04 | <jgraham> | Ignore me |
| 11:06 | <jgraham> | Right, I now have a less confusing, non-racy test |
| 11:27 | <Ms2ger> | annevk, so did you leave because Facebook is going to buy Opera? :) |
| 11:28 | <jgraham> | Maybe facebook bought annevk |
| 11:31 | <zcorpan> | he left. surely that means something. |
| 11:31 | <icaaq> | zcorpan: hehe |
| 11:32 | <Ms2ger> | Is he not yet allowed to disclose that he's going to work for facebook? |
| 11:38 | <Ms2ger> | matjas, thanks |
| 11:39 | <matjas> | Ms2ger: I’ll leave a comment with some tests results in other browsers |
| 12:08 | <annevk> | Ms2ger: I think those rumors started after I quit |
| 12:12 | <Ms2ger> | Aha! |
| 12:17 | <annevk> | the plot thickens? lol |
| 12:26 | <matjas> | jgraham: “eval.apply({}, code) should throw EvalError” → no browser seems to do this; am I testing this incorrectly? |
| 12:29 | <zcorpan> | isn't EvalError unused? |
| 12:51 | <jgraham> | matjas: I don't remember. Possibly everyone changed that behaviour? |
| 12:51 | <matjas> | jgraham: that’s good news, then |
| 12:53 | <MikeSmith> | do any engines other than gecko have support for the opus audio codec yet? |
| 13:10 | <zcorpan> | hsivonen: MikeSmith: http://validator.keegan.st/ |
| 13:11 | <MikeSmith> | zcorpan: wow |
| 13:11 | <MikeSmith> | that's pretty nice |
| 13:11 | <matjas> | woah i know this guy |
| 13:11 | <MikeSmith> | we should add this upstream |
| 13:14 | <Ms2ger> | assertThrows(function(){ worker.port.onmessage = 1 }, [], TypeError, {}); |
| 13:14 | <Ms2ger> | zcorpan, sounds bogus to me ^ |
| 13:14 | <zcorpan> | Ms2ger: it probably matched the spec when i wrote it |
| 13:15 | <zcorpan> | now it should set to null, right? |
| 13:15 | <Ms2ger> | Right |
| 13:26 | <matjas> | wait, does testharness.js not have an equals() method that compares objects? |
| 13:26 | <Ms2ger> | How would you compare them? |
| 13:26 | <matjas> | iterate over all keys/values in both objects and see if there’s a difference |
| 13:27 | <matjas> | would cover most cases |
| 13:28 | <matjas> | QUnit does it like this: https://github.com/jquery/qunit/blob/b0c3da8d09b1dcbe5dd0877e27386f4b9e8050a1/qunit/qunit.js#L1343-1525 |
| 13:30 | <jgraham> | matjas: Because there are a huge number of weird edge cases |
| 13:30 | <jgraham> | e.g. loops in the object graph |
| 13:30 | <jgraham> | and wanting to compare particular keys with something other than assert_equaks |
| 13:30 | <jgraham> | *assert_equals |
| 13:31 | <matjas> | jgraham: why not add a simple method and document that it won’t work for every possible use case? |
| 13:31 | <jgraham> | Because people will be surprised when it doesn't work for their use case |
| 13:32 | <matjas> | fair enough |
| 13:32 | <jgraham> | And because it's hard to get right even in simple cases |
| 13:32 | <jgraham> | and doesn't promote good quality tests (a single assert can only give a single error message) |
| 13:33 | <jgraham> | (what was your use case?) |
| 13:34 | <matjas> | comparing fn.arguments with the arguments object in a function (it should be ===) |
| 13:34 | <matjas> | i just use assert_equals(argumentsProp === argumentsObject, true) now |
| 13:35 | <matjas> | but in other cases that wouldn’t be possible |
| 13:35 | <jgraham> | But if they are === then assert_equals should work |
| 13:35 | <jgraham> | assert_equals(argumentsProp, argumentsObject) |
| 13:35 | <matjas> | oh, right |
| 14:19 | <matjas> | do any of you know if this is a compatibility measure, or just a bug that all engines appear to have? https://mail.mozilla.org/pipermail/es-discuss/2012-July/023888.html |
| 14:22 | <Ms2ger> | Heh, people still care about acid2 |
| 14:23 | <gsnedders> | Ms2ger: var scrollTo = window.scrollTo; doesn't create a new variable binding |
| 14:24 | <Ms2ger> | In the meantime, I found https://bugs.ecmascript.org/show_bug.cgi?id=78#c0 |
| 14:26 | <gsnedders> | Ms2ger: I believe var self = 1; would mutate Window.prototype.self to equal 1. |
| 14:26 | <gsnedders> | (I could be wrong, I've never been that happy with this stuff) |
| 14:27 | <Ms2ger> | That sounds rather unlikely |
| 14:27 | <gsnedders> | That's what I think. |
| 14:27 | <jgraham> | Huh, how would that work? |
| 14:27 | <gsnedders> | But you get a reference for "scrollTo", which is a reference to Window.prototype.scrollTo, and then put a value on that reference. |
| 14:28 | <gsnedders> | Or am I being stupid? |
| 14:28 | <zcorpan> | that sounds rather likely :-P |
| 14:28 | <gsnedders> | Yes, it does. :P |
| 14:29 | <Ms2ger> | No comment :) |
| 14:37 | <matjas> | Ms2ger: heh, that’s the link from http://mathias.html5.org/specs/javascript/#top-level-var-statements |
| 14:38 | <Ms2ger> | \o/ |
| 14:45 | <Ms2ger> | matjas, <jorendorff> Ms2ger: if you look at ES5.1 �10.5, arguments object is created in steps 6-7; variables aren't bound until step 8 |
| 14:48 | <matjas> | Ms2ger: cool, filed it here: Thanks for confirming. I filed it here: https://bugs.ecmascript.org/show_bug.cgi?id=440. |
| 14:48 | <matjas> | whoops copypasta |
| 14:49 | <Ms2ger> | So the spec is correct, right? |
| 14:50 | <Ms2ger> | Also... |
| 14:50 | <Ms2ger> | A test! |
| 14:50 | <matjas> | Ms2ger: the spec text is wrong, but the steps are right |
| 14:51 | <Ms2ger> | Oh, �10.6 disagrees with �10.5? |
| 14:52 | <matjas> | unless i misread the text |
| 14:53 | <matjas> | “an `arguments` object is created unless […] the identifier `arguments` occurs as […] the Identifier of a VariableDeclaration […] contained in the function code.” seems to conflict with 10.5 |
| 14:54 | <Ms2ger> | Looks like we all agree on that point :) |
| 14:59 | <matjas> | so in the case of function fn() { var arguments = undefined; return typeof arguments; }; fn(); // "undefined" |
| 15:00 | <matjas> | the `arguments` object is created in step 7 |
| 15:01 | <matjas> | but then overwritten by the variable in step 8, iiuc |
| 15:02 | <matjas> | heh |
| 15:04 | <SimonSapin> | I’m now passing Acid2 :) http://weasyprint.org/#acid2 |
| 15:05 | <matjas> | SimonSapin: nice work |
| 15:07 | <SimonSapin> | matjas: thanks |
| 15:07 | <SimonSapin> | We’re a bit late to the party but we only started 14 months ago |
| 15:15 | <Ms2ger> | matjas, no, it's not overwritten due to 8.c |
| 15:18 | <matjas> | Ms2ger: oic. but how is `var arguments;` different from `var arguments = undefined;`, then? |
| 15:18 | <matjas> | aren’t both VariableDeclarations? |
| 15:18 | <Ms2ger> | `var arguments = undefined;` is `var arguments; argument = undefined;` |
| 15:18 | <Ms2ger> | <jorendorff> function fn() { alert(typeof arguments); var arguments = undefined; alert(typeof arguments); } |
| 15:18 | <Ms2ger> | <jorendorff> alerts "object" then "undefined" |
| 15:18 | <Ms2ger> | <jorendorff> All the steps in 10.5 happen before the function body starts running. |
| 15:21 | <Ms2ger> | Is anyone else doing proxies, btw? |
| 15:22 | zcorpan | awaits an xkcd about higgs |
| 15:23 | <jgraham> | Ms2ger: Not yet afaik |
| 15:23 | <annevk> | zcorpan: me too, I wonder if it'll be about the font :) |
| 15:25 | <matjas> | Ms2ger: thanks! http://ecma-international.org/ecma-262/5.1/#sec-12.2 VariableDeclaration : Identifier vs. VariableDeclaration : Identifier Initialiser |
| 15:31 | <matjas> | i guess i was thrown off by this: function fn() { alert(typeof arguments); var arguments; alert(typeof arguments); }; fn(); // alerts "object" twice |
| 15:38 | <Ms2ger> | That's consistent with what I said, I think :) |
| 16:03 | <matjas> | sure, it all makes sense in YOUR head! |
| 16:05 | <Ms2ger> | :D |
| 16:08 | <matjas> | Ms2ger: http://ecma-international.org/ecma-262/5.1/#sec-12.2 says “Variables are initialised to `undefined` when created” — so I expected `var foo;` to set it to undefined upon creation |
| 16:09 | <Ms2ger> | But 8.c means there's no variable |
| 16:11 | <matjas> | …as varAlreadyDeclared is false, because the `arguments` object is already created at that point |
| 16:11 | <matjas> | so only when you do `[var] arguments = undefined` it overwrites that |
| 16:11 | <matjas> | FINALLY |
| 16:12 | <matjas> | I don’t know why I keep trying to read specs… not even all this Higgs boson talk makes me feel this stupid :) |
| 16:14 | <matjas> | /s/false/true/ a couple of messages back |
| 16:15 | <Ms2ger> | ES is horrible, I can only do this thanks to a smart guy in another channel :) |
| 16:16 | <jgraham> | ES is just quite complicated |
| 16:17 | <matjas> | the spec is a bit weirder than other specs, too |
| 16:17 | <jgraham> | I'm not sure that any other programming language spec would be simpler |
| 16:17 | <matjas> | e.g. “If experValue is null or undefined, return (normal, empty, empty).” — what the hell is (normal, empty, empty)? |
| 16:17 | <jgraham> | That's defined elsewhere |
| 16:18 | <matjas> | undoubtedly. but in WHATWG/W3C specs you’d get a clickable link to the explanation/definition |
| 16:18 | <matjas> | stuff like that helps make things a bit more n00b-friendly |
| 16:18 | <jgraham> | http://es5.github.com/#x8.9 |
| 16:18 | <matjas> | although i can see how that is not the goal of any spec |
| 16:19 | <jgraham> | One major problem with ECMAScript is that it doesn't use HTML natively so obvious features like hypoerlinks of definitions are missing |
| 16:20 | <jgraham> | e.g. if that was phrased If experValue is null or undefined, return the [completion type] (normal, empty, empty).", with the bit in square brackets being a link it would be a lot easier to work with |
| 16:20 | <matjas> | jgraham: exactly |
| 16:20 | matjas | ♥ Anolis |
| 16:22 | <Ms2ger> | :) |
| 16:25 | <gsnedders> | matjas: We love you too. |
| 17:03 | <matjas> | Ms2ger: ask jorendorff to accept https://github.com/jorendorff/es-spec-html/pull/8 please |
| 17:03 | matjas | wants to delete fork |
| 17:06 | <Ms2ger> | Breaking news: CSSWG is renaming flexbox properties again |
| 17:10 | <Ms2ger> | ... I guess that wasn't a surprise to anybody |
| 17:21 | <shepazu> | krijnh: hey, when did your IRC logs change? they look like they've lost flagging functionality, or is that just me? |
| 17:22 | <Ms2ger> | Just you |
| 17:23 | <zewt> | been missing for me for a while too |
| 17:24 | <shepazu> | it's weird… and all of the comments by Ms2ger are now rendered in white text on a black background, with little unicode skulls at the end of each line |
| 17:25 | <Ms2ger> | Oh really ? ☠ |
| 17:26 | <shepazu> | took you a while to find that skull, Ms2ger, I'm disappointed :) |
| 17:26 | Ms2ger | patpats shepazu |
| 17:27 | <shepazu> | srsly, do you have to be one of the in-crowd, or signed in, or has the functionality just be removed or broken? |
| 17:28 | <gsnedders> | WFM, and it doesn't have my nick saved or anything |
| 17:30 | <zewt> | definitely doesn't work for me |
| 17:30 | <gsnedders> | Weird. |
| 17:30 | <zewt> | in ff13 or chrome 19 |
| 17:30 | <shepazu> | gsnedders: I'm on FF13 |
| 17:31 | <shepazu> | gsnedders: you can flag something, and it persists when you reload? it's not just the highlighted line behavior? |
| 17:32 | <zewt> | it doesn't even give me the UI to do it |
| 17:32 | <gsnedders> | shepazu: I didn\t try reloading. |
| 17:32 | <gsnedders> | Why would I do something so radical? |
| 17:33 | <shepazu> | gsnedders: well, you are an iconoclast, I thought it would be second nature to doubt |
| 17:33 | <gsnedders> | Right, persisted here, at least. |
| 17:33 | <shepazu> | zewt: same here |
| 17:34 | <gsnedders> | You should be using Opera, obviously. :) |
| 17:34 | <shepazu> | I guess sgsnedders is one of the in-crowd :) |
| 17:34 | <shepazu> | zewt: … well, I'm sorry... |
| 17:34 | <shepazu> | you didn't make the cut |
| 17:35 | <gsnedders> | shepazu: I guess not, given krijnh has no way to tell me from anyone else. :) |
| 17:38 | <zewt> | looks like something broken on the php side, since it's just not outputting the span that the script is looking for |
| 17:54 | <odinho> | I thought we had to use usercss and stuff. |
| 19:01 | <jgraham> | shepazu: You aren't trying to flag something from not today, are you? |
| 19:02 | <jgraham> | krijnh changed it so that flagging gets disabled after the day is no longer today |
| 19:02 | <shepazu> | ahhhh |
| 19:02 | <shepazu> | why's that? |
| 19:02 | <Philip`> | For caching reasons, I thought |
| 19:03 | <Philip`> | (making all archived logs be static HTML instead of running slow PHP each time) |
| 19:03 | <Philip`> | though I may just be imagining remembering that |
| 19:03 | <jgraham> | Oh, I thought it was just so that people living in silicon valley would have to try double hard to get things in yellow |
| 19:03 | <jgraham> | Since the cutoff is halfway through the day there |
| 19:03 | <shepazu> | oh, that's too bad, usually flagging doesn't take place the same day, I'd have thought |
| 19:04 | <jgraham> | If it was me I would have let things be flagged for a few days |
| 19:04 | <jgraham> | But it's not, and I don't have to pay the bandwidth so... |
| 19:06 | <shepazu> | I'm looking at improving the W3C IRC bot stuff |
| 19:06 | <shepazu> | including something like krijn's logs |
| 19:19 | <Philip`> | Is there really no decent off-the-shelf IRC logging software that could be used? |
| 19:25 | <jgraham> | It does sound like a solved problem, doesn't it |
| 20:36 | <annevk> | jgraham: it's not bandwidth, it's performance |
| 20:37 | <annevk> | jgraham: but maybe krijnh is willing to tweak it to make it possible to change it for the last three days or so |
| 20:37 | <annevk> | jgraham: dunno how flexible the setup is |
| 21:02 | <gsnedders> | Fairly agressive caching could help without making it immutable. |
| 21:17 | <annevk> | webkit-dev is amusing |
| 21:18 | <annevk> | Intel guy #1: Look, user timing API implementation! |
| 21:18 | <annevk> | Intel guy #2: Nice feature! |
| 21:18 | <annevk> | </thread> |
| 21:31 | <guzzlefry> | Is it preferable to set image width and height in CSS or the actual HTML? |
| 21:31 | <annevk> | guzzlefry: if you set them in CSS you might as well not set them |
| 21:31 | <guzzlefry> | Why is that? |
| 21:32 | <annevk> | guzzlefry: because by the time the style sheet is loaded the images are prolly too |
| 21:32 | <guzzlefry> | ah okay, thanks |
| 21:38 | <jgraham> | annevk: Could make the highlights a seperate JSON file and applied in client script so that you always serve static files and the only dynamic bits are in response to POSTs that update the highlighting JSON |
| 21:38 | <annevk> | I'm sure krijnh has plenty of free time :) |
| 21:39 | <annevk> | Which reminds me, how is updating pimpmyspec.net going? |
| 21:39 | <jgraham> | Sure, I realise it is work |
| 21:40 | <jgraham> | Not at all. The main missing feature is the xrefs which I don't really know how to implement |
| 21:44 | <zewt> | annevk: there's no way there's a performance problem with toggling highlights in an IRC log, heh |
| 21:44 | <zewt> | unless there's some catastrophically bad code involved |
| 21:46 | <annevk> | it's not about the highlighting itself |
| 21:46 | <annevk> | it's about generating the pages |
| 21:47 | <zewt> | it's not like there are thousands of highlights being toggled a second, heh |
