| 02:09 | <zewt> | gar i hate when i forget to switch to the rich editor before sending mails; mails look ridiculously long when formatted to 80 columns instead of flowed properly |
| 07:41 | <annevk> | given consistent window.onerror implementations it seems you can actually introduce parse-error inducing syntax errors in ecmascript |
| 07:41 | <annevk> | just requires two <script> elements for fallback, which does not seem too bad |
| 07:43 | <annevk> | I wonder how abarth got convinced that CSP is a good idea now |
| 07:43 | <annevk> | still looks waaaaaayyyy complex |
| 07:43 | <annevk> | and more so over time |
| 08:25 | <hsivonen> | Does http://msdn.microsoft.com/en-us/library/ms890014.aspx predate Apple's <meta name=viewport>? |
| 08:31 | <hsivonen> | people suck at registering meta keywords |
| 08:32 | <annevk> | and sending feedback to you know, standards bodies |
| 08:35 | <hsivonen> | in fairness, some of these keywords might lack documentation but still be real |
| 08:36 | hsivonen | is just following the Process like any IANA Designated Expert would |
| 08:36 | <annevk> | for whether or not they're valid? |
| 08:40 | <hsivonen> | annevk: yeah |
| 08:42 | <annevk> | I don't think that has really been the problem with IANA, it's just that the registries lacked everything else that was out there |
| 08:42 | <annevk> | but I do think having IANA is overkill |
| 08:42 | <annevk> | maybe they do some other things, but for maintaining some lists for which you could simply put a piece of software somewhere it seems rather elaborate |
| 08:44 | <hsivonen> | speaking of overkills: http://tools.ietf.org/html/draft-yevstifeyev-disclosure-relation-00 |
| 08:45 | <annevk> | hahaha |
| 08:46 | <annevk> | one paragraph and two list items of normative text |
| 08:46 | <wilhelm> | Beautiful. |
| 08:47 | <annevk> | I'm glad I never got sucked into doing things the IETF way, because it's quite horrible |
| 08:52 | smaug____ | notices that WoW is being advertised heavily, even more than Chrome. |
| 08:52 | <hsivonen> | the person who creates a single-page Google Site for each of his/her meta keyword suggestions knows how to play the game Hixie set up |
| 08:56 | <annevk> | maybe there should be some level before fully registered, where fully registered means it's clearly an important and relatively widely used value |
| 08:56 | <annevk> | I guess if we take the approach where we put registered values into HTML after some time, that might work |
| 08:57 | <annevk> | oh lol, I feel sorry for Marcos |
| 08:57 | <annevk> | re: http://lists.w3.org/Archives/Public/public-webapps/2011OctDec/1613.html |
| 09:08 | <hsivonen> | annevk: someone should remind the "OMG, unversioned normative reference" folks that all specs rely on Unicode and implementations support the latest Unicode that they happen to support instead of artificially hiding the characters that have been defined after whatever spec uses Unicode was finalized |
| 09:11 | <Ms2ger> | Do you think they'll listen? |
| 09:11 | <annevk> | telling someone who's been having this idea since before '98 that he's actually wrong will not lead to much I think |
| 09:12 | <hsivonen> | Ms2ger: no. Hence, "someone"--not "I" |
| 09:12 | <annevk> | in 2007 we had http://annevankesteren.nl/2007/04/html-red-pill |
| 09:12 | <annevk> | now we have the Standards Red Pill |
| 09:13 | <annevk> | which comes down to pretty much the same re-reading that post :) |
| 10:04 | <remysharp> | hi there, EventSource - I'm using onerror to detect when the connection is dropped |
| 10:04 | <remysharp> | but that's not perfect - as it doesn't always fire when the connection is closed |
| 10:04 | <remysharp> | (I closed my laptop, reopened it hours later and the EventSource had readyState=2 but hadn't fired error - correctly I think) |
| 10:05 | <remysharp> | is there any reason why there's no close event - or is there a better way of detecting? |
| 10:07 | <smaug____> | remysharp: file a bug |
| 10:07 | <smaug____> | it is an implementation bug |
| 10:07 | <remysharp> | so you're saying it should fire then? |
| 10:08 | <remysharp> | (as in it's a browser bug?) |
| 10:08 | <smaug____> | ye |
| 10:08 | <smaug____> | yes, error event should fire |
| 10:08 | <remysharp> | and the error is simply to say "the connection has gone", right? |
| 10:08 | <remysharp> | I guess "close" would be, it's closed, and we're not reopening |
| 10:09 | <smaug____> | error even is used also in case there is reconnection happening |
| 10:09 | <smaug____> | but state is then different |
| 10:10 | <smaug____> | s/even/event/ |
| 10:12 | <remysharp> | perfect, then I'm using error correct. Now the hard part: replicating the bug ;-) |
| 10:12 | <remysharp> | cheers for the feedback smaug____ |
| 10:13 | <smaug____> | remysharp: I'm assuming your code doesn't explicitly call close() |
| 10:13 | <remysharp> | smaug____ correct |
| 10:14 | <remysharp> | and yeah, if it called .close() - I *wouldn't* expect the error event |
| 10:14 | <remysharp> | and wouldn't need an event as I triggered it myself |
| 10:55 | <annevk> | seems there's still some discussion going on regarding adaptive images |
| 10:55 | <annevk> | I replied regarding the content negotiation suggestion |
| 10:55 | <annevk> | http://www.brucelawson.co.uk/2011/notes-on-adaptive-images-yet-again/ |
| 11:50 | <bga> | lol setTimeout(function () { alert(123); }, Math.pow(2, 32)) |
| 11:50 | <bga> | <=> setTimeout(_, 0) |
| 12:18 | <annevk> | gsnedders: how does the versioning actually work in ECMAScript though? won't you get an error in older browsers either way? |
| 12:19 | annevk | doesn't get it |
| 12:19 | <gsnedders> | annevk: Either way you get an error in older browsers. Having modes allows us to finally change some of the ugliness that blocks a lot of JS performance optimizations |
| 12:20 | <beverloo> | "use strict" introduces behavioral changes, "use version 6" syntactical changes. There will by syntax errors |
| 12:20 | <beverloo> | But I imagine that "version 6" will also include behavioral changes |
| 12:20 | <beverloo> | s/by/be/ |
| 12:21 | <annevk> | oh, so the syntax is effectively changed |
| 12:21 | <annevk> | sounds pretty bad |
| 12:21 | <annevk> | making things simpler while still allowing the old syntax seems better |
| 12:22 | <annevk> | e.g. allowing if count > 4 { ... } |
| 12:26 | <gsnedders> | annevk: The global scope is no longer in the scope change is the biggest breaking change |
| 12:27 | <gsnedders> | *chain |
| 12:29 | <annevk> | o_O |
| 12:29 | <annevk> | "use version 6" and learn to program again? |
| 12:30 | <annevk> | use <!doctype html> and delimit your tags with [ and ] now; it's better, promise |
| 12:30 | <annevk> | (not quite the same, true true, but this does not really sound like a good idea) |
| 12:31 | <gsnedders> | annevk: Also there are in general problems with adding new keywords |
| 12:31 | <gsnedders> | annevk: As that will easily break things |
| 12:33 | <annevk> | well yes, we have the same problem with adding new members to nodes |
| 12:33 | <gsnedders> | annevk: Not really |
| 12:33 | <annevk> | not really sure that justifies versioning though |
| 12:33 | <gsnedders> | annevk: This causes syntax errors in existing scripts, what that does depends on behaviour of the new member |
| 12:34 | <annevk> | it typically causes existing scripts to break |
| 12:34 | <annevk> | e.g. when we added .min and .max that happened |
| 12:35 | <gsnedders> | I mean, extensions to ES that are where there are currently syntax errors and no vendor implements anything in the same extension space are pretty damned safe. |
| 13:10 | <annevk> | so Mac OS X remembers state of programs so you can sort of reboot and continue (for when they offer a Safari update or something), but it completely breaks history in Terminal |
| 13:10 | <annevk> | which sucks |
| 13:11 | <annevk> | anyone from Apple here who can tell me if that is reported and if not, how I go about that? |
| 13:11 | <annevk> | (this is why I ignore software updates from Apple for weeks on end, I don't like rebooting) |
| 13:17 | <annevk> | zcorpan: do you have an algorithm for decoding UTF-8 encoded text combined with the HTML-defined error handling? |
| 13:32 | <hsivonen> | http://news.cnet.com/8301-30685_3-57342776-264/the-web-in-2012-five-predictions-starting-with-ie10/ wow. an article about browsers that has 5 browser icons and all of them are up to date! |
| 13:33 | <annevk> | Stephen Shankland is the only reporter I know that is follows the browser and standards sphere somewhat actively and closely |
| 13:35 | <annevk> | though I don't pay much attention to journalism related to browsers/standards |
| 13:36 | <hsivonen> | annevk: I'm not surprised that if one reporter gets the icons right, it's stshank |
| 13:36 | <hsivonen> | but getting all the icons right is still rare in journalism |
| 13:43 | <zcorpan> | annevk: nope |
| 13:47 | <zcorpan> | hsivonen: i think the opera logo is still the wrong one. there's one logo for the company name and one for the browser product |
| 13:51 | <hsivonen> | zcorpan: whoa. you are right. You guys are making it hard to do the right thing. |
| 13:51 | <hasather> | zcorpan: correct, there is the logo, and the icon: http://www.opera.com/press/resources/ |
| 13:51 | hsivonen | hangs head in shame; was just looking for non-toilet seat |
| 13:54 | <Ms2ger> | That would be a nice slogan |
| 13:54 | <Ms2ger> | "Opera - No longer a toilet seat" |
| 13:55 | <hsivonen> | the right set is http://paulirish.com/lovesyou/new-browser-logos/all-browser_logos/browser_logos-256.png |
| 13:56 | hsivonen | wonders if the Safari icon has stayed constant or had subtle refreshes |
| 14:14 | <hsivonen> | gsnedders: if ES6 takes the global object out of the scope chain, what about the document object? what about Element? |
| 14:14 | <hsivonen> | for inline event handlers that is |
| 14:17 | <bga> | pythonic js |
| 14:33 | <annevk> | is there some kind of preference for specifying strings in standards? |
| 14:34 | <annevk> | I sometimes find myself using "..." and sometimes ... |
| 14:34 | <annevk> | e.g. the name of an encoding is a string too, but I'd like to just refer to it without code |
| 14:34 | <annevk> | whereas for a label it might make sense to use <code>? |
| 14:34 | <annevk> | or maybe not |
| 14:48 | <annevk> | http://dvcs.w3.org/hg/encoding/raw-file/tip/Overview.html#encodings |
| 14:49 | <Ms2ger> | "The encoding" |
| 14:49 | Ms2ger | likes |
| 14:54 | <gsnedders> | hsivonen: I believe the window object will be accessable from the global object, just not the global object |
| 15:30 | <hsivonen> | annevk: I hope you will put UTF-16 under a heading "legacy multi-octet encodings" |
| 16:09 | <erlehmann> | this is 404? http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2010-December/029459.htm |
| 16:09 | <erlehmann> | no, it is 200 ok |
| 16:09 | <erlehmann> | it may be lying? |
| 16:18 | <miketaylr> | ah, erlehmann add an l to the end :) |
| 16:18 | <miketaylr> | that confused me a little bit ago too |
| 16:18 | <erlehmann> | this is 404? http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2010-December/029459.html |
| 16:18 | <miketaylr> | wfm |
| 16:22 | <Ms2ger> | Ah, isn't the web fun |
| 16:22 | <Ms2ger> | Removing navigator.taintEnabled(): at least 3 broken sites |
| 16:41 | <smaug____> | Ms2ger: did anyone else but Gecko have taintEnabled() |
| 16:41 | <Ms2ger> | No idea |
| 17:15 | <dglazkov> | good morning, Whatwg! |
| 17:16 | <jwalden> | good afternoon, dglazkov |
| 17:16 | <jwalden> | ;-) |
| 17:17 | <dglazkov> | timezones are irrelevant. They are just clever marketing ploys invented by clock companies. |
| 17:18 | <jwalden> | gotta watch out for those companies... |
| 17:32 | <annevk> | hsivonen: very much the plan |
| 17:32 | <annevk> | hsivonen: I think I will work on that now, to at least make it clear what all the allowed labels are, so people can start complaining about those |
| 17:32 | <annevk> | and then work out the various difficult encodings, one by one |
| 17:50 | <annevk> | man these people in the document.characterSet bug don't make much sense |
| 17:51 | <annevk> | euh, document.charset bug |
| 17:52 | <Ms2ger> | They don't |
| 18:01 | <annevk> | http://parislemon.com/post/14330528072/over-the-next-few-weeks-well-figure-it-out sounds like the disabling H264 in Chrome story |
| 18:12 | <annevk> | oooh |
| 18:12 | <annevk> | that Jean-Claude |
| 18:12 | <annevk> | from my very early standards stint in the CDF WG |
| 18:26 | <annevk> | pretty cool these links from mattur |
| 18:27 | <annevk> | how standards were developed in '94 http://www.w3.org/People/Connolly/events/www94/trip-report.html |
| 18:27 | <annevk> | with workshops given by the editor |
| 18:29 | <Ms2ger> | <hr>Dinner On the Boat<hr> |
| 18:29 | <smaug____> | CDF... |
| 18:30 | <smaug____> | I never quite understood CDF |
| 18:30 | <smaug____> | although Lasse was my colleague |
| 18:31 | <annevk> | if you view CDF as a group that made sure cross-language barriers were better defined in CSS/SVG/HTML, they were somewhat successful |
| 18:31 | <annevk> | but they did much more, which was rather questionable (and people did not like me for saying so) |
| 19:04 | <annevk> | how do you tell whether a UTF-16 label is supported and there's no heuristics going on? |
| 19:05 | <Yuhong> | annevk: BTW, here is a good test for multi-octet encodings: http://ha.ckers.org/weird/variable-width-encoding.cgi |
| 19:06 | <annevk> | It seems to just show you need to have certain error handling otherwise you run into issues |
| 19:07 | <annevk> | presumably that with a two-level octet you should not replace the first octet with U+FFFD and then match the second one again... |
| 19:08 | <Yuhong> | While most browsers fixed their decoding long time ago, IE only fixed it in August: http://technet.microsoft.com/en-us/security/bulletin/MS11-057 |
| 19:09 | <annevk> | this is why we need this standard |
| 19:23 | <annevk> | does anyone know if big5-hkscs is a superset of big5? |
| 19:23 | <annevk> | looks like at least some people want them to map to the same thing: http://www.google.com/support/forum/p/Chrome/thread?tid=466c210af3fb6d08 |
| 19:28 | <annevk> | http://en.wikipedia.org/wiki/HKSCS mess |
| 19:29 | <Ms2ger> | http://en.wikipedia.org/wiki/World_Wide_Web mess |
| 19:29 | <annevk> | new meme \o/ |
| 19:33 | <annevk> | Ms2ger: you guys make some sweet beer :) |
| 19:33 | Ms2ger | isn't into beer |
| 19:33 | <annevk> | (or the guys running your proxy) |
| 19:36 | <annevk> | the Web Encodings wiki page claims Microsoft simply makes big5-hkscs a label for big5, but Mozilla and Opera have it as a distinct encoding |
| 19:36 | <annevk> | and then you have those users complaining... |
| 19:37 | <Ms2ger> | http://mxr.mozilla.org/mozilla-central/source/intl/uconv/tools/gen-big5hkscs-2001-mozilla.pl |
| 19:39 | <annevk> | euh |
| 19:43 | <annevk> | so http://coq.no/character-tables/mime/locale-specific/en suggests it is indeed a superset |
| 19:43 | <annevk> | combined with the support thread above it seems that might be somewhat of an issue for Gecko / Chrome / Opera adoption in that area... |
| 19:52 | <annevk> | so it seems there are some differences |
| 19:53 | <annevk> | but IE does not seem to care and win |
| 19:53 | <annevk> | quote I can prolly share: "For instance, =C6BF..=C6C2 get converted to U+2F02..U+2F07 (⼂⼃⼅⼇) in the -2003 table, and to U+4E36..U+4EA0 (丶丿亅亠) in the -HKSCS table (look-alike mappings, apparently). The differences seem to range up to =C6D9 (with =C6D8 being mapped to U+00A8 (¨) in both tables) (＾ vs ˆ for =C6D9)." |
| 20:12 | <annevk> | I thought the single-octet stuff was sad |
| 20:31 | <annevk> | Most Firefox' utf-16 labels are for utf-16be, but HTML states "When a user agent is to use the self-describing UTF-16 encoding but no BOM has been found, user agents must default to little-endian UTF-16." |
| 20:32 | <annevk> | Either Firefox has not been updated to take HTML into account, or Web Encodings is out of date with what Firefox does... (Probably both.) |
| 20:32 | <gsnedders> | Well, Firefox is right per normal Unicode behaviour. |
| 20:33 | <annevk> | Per normal behavior utf-16 has no other labels than utf-16 |
| 20:34 | <Ms2ger> | There is no normal behaviour on the web |
| 20:34 | <annevk> | You don't trust IANA? Come on! |
| 20:34 | <gsnedders> | Normal behaviour is there are three labels: UTF-16, UTF-16LE, and UTF-16BE. The latter two, if the first codepoint is U+FFEF, it represents a ZWNBSP and is not a BOM. |
| 20:34 | <annevk> | those are considered separate encodings |
| 20:34 | <annevk> | three separate encodings that is |
| 20:34 | <gsnedders> | Given UTF-16, if it starts with a BOM, then it is whatever endian that is. Otherwise, it is big endian. |
| 20:35 | <gsnedders> | That's what the Unicode spec says, at lesat. |
| 20:35 | <annevk> | yeah, HTML points out it violates that |
| 20:35 | <annevk> | apparently all of Windows prefers LE |
| 20:35 | <annevk> | hello Microsoft \o/ |
| 20:35 | <gsnedders> | annevk: I'll bet that's because x86 is LE |
| 20:36 | <annevk> | the new encoding standard will attempt to correct the nonsense |
| 20:37 | <gsnedders> | annevk: Any idea how far UTF-16 being LE by default spreads? |
| 20:38 | <annevk> | I have not done any testing yet on UTF-16 |
| 20:38 | <annevk> | just looked at a bunch of information |
| 20:41 | <smaug____> | "jquery 1.6 uses navigator.taintEnabled for browser detection" o_O |
| 20:41 | <smaug____> | I knew jQuery is strange... |
| 20:47 | <annevk> | I sort of expected that since encodings are in a way somewhat trivial there would not be this clusterfuck |
| 20:53 | <Ms2ger> | annevk, dunno if you saw the xmlVersion bug? That one is fun too :) |
| 21:18 | <annevk> | Ms2ger: did :) |
| 21:18 | <annevk> | I checked in some legacy multi-octet stuff, but there's not much there now |
| 21:19 | <annevk> | as evidenced by the log I mostly studied the various encodings a bit and then wrote up some labels based on IE/Gecko/HTML |
| 21:19 | <annevk> | I think tomorrow I'll define label -> encoding (including the trimming of leading and trailing whitespace and case-insensitive stuff) |
| 21:20 | <annevk> | and maybe ping whatwg and iana charsets lists |
| 21:40 | <lmarcetic> | mic check. |
| 21:49 | <lmarcetic> | Question about DOM events: How hard would it be to construct an event logger? Must one create and add handlers for each and every event? How complex would a js library have to be to be able to record and replay events? |
| 21:49 | <annevk> | yes, there's no * |
| 21:50 | <annevk> | and euh, somewhat complex I reckon |
| 21:50 | <lmarcetic> | how about a few basic ones, like point, click and scroll? would that be easy enough? |
| 21:51 | <lmarcetic> | pointing/dragging would probably be complex (alternatively, jittery) |
| 21:51 | <annevk> | I'd just try it out :) |
| 21:52 | <annevk> | prolly with capture event listeners on window or some such so you are before everyone else and then create some synthetic events |
| 21:53 | <annevk> | the problem you'll run into is that a synthetic event does not always have the same privileges (e.g. cannot open a new window); not sure if that's a big problem |
| 21:55 | <lmarcetic> | annevk, oh it's you Anne! Didn't even notice :-) I was the one who bugged you about this via e-mail. Anyway, not being able to open windows is understandable. Is there anything else? |
| 21:57 | <annevk> | oh, http://www.iana.org/assignments/charset-reg/CP50220 mentions http://wiki.whatwg.org/wiki/Web_Encodings |
| 21:58 | <annevk> | lmarcetic: I suspect stuff like :hover will not work |
| 21:59 | <annevk> | lmarcetic: I have not really played with hit testing and such much |
| 22:00 | <annevk> | and I should really do something else before I start dreaming about encodings again |
| 22:01 | <lmarcetic> | annevk, hover is important, that's too bad. |
| 22:05 | <lmarcetic> | annevk, thanks. afk. |
| 22:16 | <Danny_Joris> | hi all - I'm trying to figure out if this is valid html? http://jsfiddle.net/Danny_Joris/sc96c/ I used thead twice in html. If there's an alternative available I'd love to learn: http://jsfiddle.net/Danny_Joris/sc96c/ |
| 22:16 | <Danny_Joris> | hmm pasted that twice |
| 22:18 | <Danny_Joris> | I'm working on a site which I have to theme this page for: http://dl.dropbox.com/u/12905785/DGI/ceacs%20subpage-author.png |
| 22:18 | <Danny_Joris> | in the rest of the site where you see 'authors' this is themed with the caption |
| 22:20 | <Danny_Joris> | I probably gave a few of you a heart attack by showing that snippet. |
| 22:36 | <smaug____> | Alac in JS |
| 22:36 | <smaug____> | http://codecs.ofmlabs.org/ |
| 22:54 | <Danny_Joris> | ok - fear not! This is how I'm going to solve it: http://jsfiddle.net/Danny_Joris/sc96c/2/ |
| 23:16 | <Danny_Joris> | actually this might be better: http://jsfiddle.net/Danny_Joris/sc96c/3/ Just had a facepalm moment. |
| 23:58 | <bencc> | in websockets, why does the client sends origin header while the server responds with Sec-WebSocket-Origin header? |
