| 01:35 | <jwalden> | does webidl say that calling a method on an interface with fewer arguments than it's specified to take (and none are optional) throws a TypeError, as I'm currently reading it? I could have sworn it was supposed to say otherwise |
| 01:38 | <jwalden> | gsnedders: there was a big change to the proxies API which theoretically addressed the [[HasProperty]]/[[GetProperty]] disconnect; I don't know how well it actually does so |
| 02:04 | <heycam> | jwalden, yeah, fewer arguments get TypeError; more arguments get ignored |
| 02:05 | <jwalden> | heycam: did that change at some point? I could have sworn defaults got imputed at one point |
| 02:05 | <heycam> | jwalden, yes but I think it changed a long time ago |
| 02:05 | <jwalden> | heycam: how long, ballpark? |
| 02:06 | jwalden | 's attention is totally fragmented these days, hardly has any time to keep up on any specs |
| 02:06 | <heycam> | jwalden, somewhere between october 2010 and july 2011 -- so I guess not *that* long ago |
| 02:06 | <heycam> | (according to my changes list at the bottom of the spec) |
| 02:06 | <jwalden> | *I* don't even know what "long ago" would mean to me, srsly |
| 02:06 | <heycam> | heh |
| 02:13 | <kennyluck> | Are people here sometimes annoyed by corner cases that are not specced but also unlikely to happen in the real world? |
| 02:17 | <heycam> | kennyluck, sometimes, but I'd be more annoyed knowing at the back of my mind that there are corner cases in my specs that I haven't dealt with |
| 03:22 | <MikeSmith> | so I'm noticing that the command attribute was in the original WF2 spec |
| 03:22 | <MikeSmith> | ah now |
| 03:22 | <MikeSmith> | *ah no |
| 03:23 | <MikeSmith> | Web Apps 1.0 |
| 03:23 | <MikeSmith> | http://web.archive.org/web/20041009144718/http://whatwg.org/specs/web-apps/current-work/#the-command |
| 03:36 | <MikeSmith> | "Any element that can define a command can also, instead, have a command attribute that specifies the ID of a command that the element should defer to." |
| 03:38 | <MikeSmith> | http://www.whatwg.org/specs/web-apps/current-work/multipage/interactive-elements.html#attr-command-command |
| 03:38 | <MikeSmith> | "If a command element slave has a command attribute, and slave is in a Document, and there is an element in that Document whose ID has a value equal to the value of slave's command attribute, and the first such element in tree order, hereafter master, itself defines a command and either is not a command element or does not itself have a command attribute, then the master command of slave is master." |
| 03:38 | <MikeSmith> | wow |
| 03:55 | <Hixie> | yeah that sentence was hard |
| 03:55 | <Hixie> | open to improvements :-) |
| 03:55 | <Hixie> | the old command="" attribute applied everywhere |
| 03:56 | <Hixie> | i figured for the new one i'd try something a bit less radical |
| 03:56 | <MikeSmith> | Hixie: that text is clear enough once I read it |
| 03:56 | <Hixie> | took me forever to write it |
| 03:57 | <MikeSmith> | just .. an interesting collection of words |
| 03:57 | <MikeSmith> | yeah I bet |
| 03:57 | <MikeSmith> | hmm, so where does it appy |
| 03:57 | <Hixie> | i couldn't work out how to refer to one of the command elements rather than the other one |
| 03:57 | <MikeSmith> | *apply |
| 03:57 | <Hixie> | only <command command=""> right now |
| 03:57 | <MikeSmith> | ah |
| 03:57 | <MikeSmith> | OK |
| 03:57 | <Hixie> | might be better as <command ref=""> or something |
| 03:57 | <MikeSmith> | maybe |
| 03:57 | <Hixie> | to avoid the xzibit jokes |
| 03:57 | <MikeSmith> | heh |
| 03:58 | <MikeSmith> | you don't want to avoid those |
| 03:59 | <MikeSmith> | repetition is a key element of poetry |
| 07:07 | <silentimp> | bga: I need you :) |
| 09:06 | <annevk> | HTML WG made a formal decision on a bug that asked to consider something and was considered and rejected by the editor |
| 09:07 | <gsnedders> | annevk: What where? |
| 09:07 | <jgraham> | They formally decided to consider it? |
| 09:07 | <annevk> | The one on code point verbosity |
| 09:08 | <zcorpan> | everyone, let's consider! |
| 09:08 | <annevk> | just struck me how the title of that bug is completely out of touch with its outcome |
| 09:29 | <MikeSmith> | idude complaining about curly apostrophe on the comment part of IDL |
| 09:32 | <jgraham> | iDude? is that some new Apple AI? |
| 09:33 | <MikeSmith> | heh |
| 09:33 | <MikeSmith> | mistyped |
| 09:33 | <annevk> | yeah I thought it would be in the normative part somehow |
| 09:34 | <annevk> | and also those comments are going to be deleted in due course |
| 09:34 | <annevk> | not sure how much I care |
| 09:34 | <MikeSmith> | I'm sure how much I care :-) |
| 09:34 | <MikeSmith> | I care about it nearly as much as I care about the code-point verbosity problem |
| 09:37 | <MikeSmith> | damn I'm getting an error for the splitter |
| 09:37 | <MikeSmith> | [[ |
| 09:37 | <MikeSmith> | File "html5-tools/spec-splitter/spec-splitter.py", line 413, in <module> |
| 09:37 | <MikeSmith> | term.remove(term.find("./dt")) |
| 09:37 | <MikeSmith> | File "lxml.etree.pyx", line 808, in lxml.etree._Element.remove (src/lxml/lxml.etree.c:36312) |
| 09:37 | <MikeSmith> | TypeError: Argument 'element' has incorrect type (expected lxml.etree._Element, got NoneType) |
| 09:37 | <MikeSmith> | ]] |
| 09:37 | <MikeSmith> | look familiar to anybody? |
| 09:38 | <MikeSmith> | oh |
| 09:38 | <MikeSmith> | that's in code I added |
| 09:41 | <jgraham> | Heh, familiar to someone at least :) |
| 09:41 | <MikeSmith> | heh |
| 09:42 | <MikeSmith> | https://www.w3.org/Bugs/Public/show_bug.cgi?id=15992 is gold |
| 09:46 | <annevk> | :) |
| 09:47 | <annevk> | on twitter the best renaming of quirks mode and all thus far |
| 09:47 | <annevk> | "Sane mode", "Sigh mode" and "Stupid mode" |
| 09:47 | <annevk> | also the only |
| 09:47 | <annevk> | but I like it |
| 09:53 | <charlvn> | annevk: no " |
| 09:53 | <charlvn> | no shit mode? i'm disappointed :( |
| 09:55 | <annevk> | maybe we should rename stupid to silly |
| 09:55 | <annevk> | because quirks mode is mostly just silly |
| 09:56 | <jgraham> | annevk: Well some of the quirks should probably have just been standardised because they make more sense than the standards (don't ask for examples, but I seem to recall there are some where Hixie and dbaron treated the spec as gospel rather than the implemtations) |
| 09:56 | <jgraham> | So "stupid mode" is a bit unfair to the poor quirks |
| 09:56 | <roc> | box-sizing |
| 09:58 | <annevk> | box-sizing, margin collapsing (not sure if that's a quirk though), unitless lengths maybe... |
| 09:59 | <roc> | yeah, unitless lengths probably |
| 09:59 | <roc> | I don't know what quirks there are around margin-collapsing |
| 10:00 | <annevk> | maybe none, but that's the one hixie and dbaron specced based on sentences in 2.0 rather than thought about how it can be done better |
| 10:03 | <jgraham> | So if you looked only at those examples and ignored all the weid stuff in quirks, it would be called "we took the standards too seriously" mode, "no really, we took the standards so seriously that we insist that you use almost all of them even when you really don't want to" mode and "stupid authors don't care about standards" mode |
| 10:07 | <zcorpan> | http://memegenerator.net/instance/14626343 |
| 10:13 | <MikeSmith> | zcorpan: beautiful |
| 10:21 | <charlvn> | zcorpan: sounds like a wrapper, makes no sense whatsoever, very good wrapping therefore |
| 10:28 | <zcorpan> | YO DAWG I HERD YOU LIKE RAPPING SO WE WRAPPED YO RAP CD SO YOU CAN WRAP WHILE YOU RAP |
| 10:29 | <zcorpan> | hmm, a wrapped wrap would be even better |
| 10:30 | <zcorpan> | YO DAWG I HERD YOU LIKE WRAPS SO WE WRAPPED A WRAP AROUND YO WRAP SO YOU CAN EAT WRAPS WHILE U EAT WRAPS |
| 10:32 | <charlvn> | zcorpan: as long as it's a mexican wrap then we're good (i live exclusively off those every time i travel to the usa) |
| 10:33 | <annevk> | I wonder what smaug thinks is wrong with MutationCallback |
| 10:59 | <Ms2ger> | annevk, the usual, not allowing { handleEvent: function() {} } |
| 10:59 | <annevk> | long discussion was had, handleEvent lost |
| 10:59 | <annevk> | so if that's it, he's out of luck |
| 11:00 | <jgraham> | As I believe the kids these days are saying, "web's got 99 problems, that ain't one" |
| 11:00 | <annevk> | oh help, even jgraham starts communicating in memes |
| 11:01 | <Ms2ger> | Anyway, I don't expect us to implement callbacks for anything else than Function |
| 11:02 | <jgraham> | annevk: If it helps I have never actually heard the song |
| 11:02 | <jgraham> | and had to check google to ensure I got the right form |
| 11:03 | <jgraham> | But the message that I actually wanted to convey is "oh no, another trivial issue that people are getting hung up about when they could be using the time to do something that would fix some of the major issues on the platform instead" |
| 11:04 | <Ms2ger> | Indeed, why do people insist on disallowing it instead of doing something useful? :) |
| 11:07 | <annevk> | Ms2ger: outside your role man, Philip` makes those remarks :p |
| 11:07 | <Ms2ger> | Oops, wrong IRC window ;) |
| 11:08 | <annevk> | but euh, because http://brokeninternets.com/wp-content/uploads/2011/10/alens.jpg |
| 11:41 | <annevk> | http://blog.whatwg.org/weekly-quirks-mode-xml |
| 11:46 | <jgraham> | annevk: The second sentence is very odd. "i.e." should be all lowercase and should never start a sentence |
| 11:46 | <jgraham> | quirks are countable so s/amount/number/ |
| 11:47 | <jgraham> | introduces is in the wrong tense |
| 11:47 | <jgraham> | Possibly s/That was all/That is all/ |
| 11:48 | <annevk> | aah |
| 11:50 | <annevk> | refresh |
| 11:50 | <annevk> | and my apologies |
| 11:50 | <jgraham> | No problem, thanks for writing them ;) |
| 12:02 | <zcorpan> | seems like quirks mode got Warnocked on www-style |
| 12:03 | <MikeSmith> | annevk, Ms2ger - are the dfns in the DOM spec meant to display pop-ups when you click on them? |
| 12:04 | <Ms2ger> | We might have removed that at some point |
| 12:04 | <Ms2ger> | Should work on the ED |
| 12:05 | <Ms2ger> | Oh, except the popupness seems broken |
| 12:05 | <annevk> | http://torrentfreak.com/dutch-parliament-rejects-acta-on-human-rights-violations-120214/ |
| 12:05 | <annevk> | fuck yeah |
| 12:06 | <annevk> | MikeSmith: styles are missing |
| 12:06 | <MikeSmith> | seems so |
| 12:06 | <annevk> | zcorpan: or maybe missed among all the other email? |
| 12:06 | <jgraham> | zcorpan: I have serious difficulties believing that several of the Warnock possibilities exist on www-style. Number 2 in particular seems highly unlikely. |
| 12:07 | <jgraham> | zcorpan: Did you check that the message actually made it all the way to www- style ;) |
| 12:11 | <zcorpan> | it made it to the archives, and i see now responses, that's all i know :-) |
| 12:11 | <zcorpan> | s/w// |
| 12:19 | <zcorpan> | Hixie: whatwg.org/C redirects to complete/ which then redirects to multipage/ |
| 12:41 | <niloy> | the websql was deprecated because everyone used sqlite, what exactly is the problem with single implementation? |
| 12:47 | <annevk> | niloy: search for software monoculture |
| 12:48 | <niloy> | thanks, will do |
| 12:48 | <annevk> | niloy: and SQLite updates were not guaranteed to be backwards compatible |
| 12:49 | <Ms2ger> | jgraham, Warnock's number 2 wouldn't happen on www-style because people *love* replying to utter nonsense? |
| 12:49 | <annevk> | niloy: I still sort of think we should have stuck with SQL, prolly defining some SQL dialect and map that to whatever engine is in use, but apparently inventing a whole new DB engine was thought to be better |
| 12:49 | <niloy> | yes, exactly, pleaseee |
| 12:50 | <annevk> | sorry man, DOM, XHR, and Encoding is what keeps me busy :) |
| 12:51 | <jgraham> | Ms2ger: bingo |
| 12:51 | <niloy> | annevk, nobody else in w3c or whatwg share your feelings? |
| 12:51 | <annevk> | those that do are busy too |
| 12:52 | <annevk> | and those that don't work on Indexed DB :) |
| 12:53 | <niloy> | oh sad |
| 13:02 | <niloy> | I read about software monoculture and realised that the "Internet" guys deprecating websql because of software monoculture is hypocrisy :P |
| 13:04 | <zcorpan> | "yeah i'm so hip because i used the html5 doctype mode", "i used spacer gifs in my tables and noticed this transitional doctype made the stupid margins go away mode", "i really have no idea WTF i'm doing mode" |
| 13:06 | <zcorpan> | would make a great title for my spec at least |
| 13:11 | <jgraham> | You want to write the "I have no idea WTF I'm doing specification" |
| 13:19 | <Ms2ger> | Isn't that all of zcorpan's specs? :) |
| 13:30 | <zcorpan> | yeah |
| 13:31 | <zcorpan> | maybe i should also start the "we have no idea WTF we're doing" working group |
| 13:33 | <Ms2ger> | "We don't Have Any idea wTf we're doing WG" |
| 13:33 | <annevk> | sicking: because those bullet points are the cache check |
| 13:34 | <annevk> | sicking: so in Gecko you can create a cache entry without a matching header or method |
| 13:34 | <annevk> | sicking: the spec doesn't have that |
| 13:34 | <annevk> | sicking: as detailed when the cache entry is created at the end of the preflight request: http://dvcs.w3.org/hg/cors/raw-file/tip/Overview.html#preflight-request |
| 13:35 | <sicking> | annevk: that seems worth fixing in the spec. I don't think i'd want to adjust our implementation for the worse |
| 13:35 | <annevk> | agreed, if I know exactly how |
| 13:36 | <annevk> | I don't really look forward to reverse engineering Gecko's preflight cache implementation |
| 13:36 | <sicking> | annevk: we're open source :-) |
| 13:37 | <sicking> | annevk: can't you just key on the method, even when the method is simple? |
| 13:38 | <annevk> | so you create a cache entry for say the POST method if force preflight is set? |
| 13:39 | <annevk> | and only then? |
| 13:39 | <sicking> | i suspect so yeah |
| 13:39 | <annevk> | i guess that could work |
| 13:39 | <sicking> | i don't actually remember, but i would assume that's what happens |
| 13:40 | <sicking> | actually we create a cache entry for destination+origin+withCredentials |
| 13:40 | <sicking> | and then in that cache which methods and headers are permitted |
| 13:42 | <annevk> | that's somewhat similar to the spec |
| 13:42 | <annevk> | less redundancy |
| 13:43 | <annevk> | depends on the details if it's the same |
| 14:29 | <annevk> | oh sicking is gone again |
| 14:29 | <annevk> | http://dvcs.w3.org/hg/cors/rev/1376d444aa1b is the fix I hope |
| 14:48 | <Ms2ger> | So, given void foo(ArrayBufferView?); void foo(ArrayBuffer?); void foo(long);, passing an object that isn't ArrayBufferView or ArrayBuffer should throw, correct? |
| 14:52 | <annevk> | or long? |
| 14:53 | <Ms2ger> | Hmm? |
| 14:55 | <jgraham> | or null? |
| 15:13 | <zewt> | it would seem strange if a function with only foo(long) didn't throw when passed an unexpected object type, but one with that plus foo(Object) did |
| 15:14 | <Philip`> | Type-based overloading in a dynamically typed language is inherently strange |
| 15:14 | <jgraham> | Philip`: Tell the jQuery people that |
| 15:15 | <Philip`> | The jQuery people are inherently strange |
| 15:18 | <zewt> | http://dev.w3.org/2006/webapi/WebIDL/#idl-overloading i ... don't feel like reading this at 9am, heh :P |
| 16:22 | <AryehGregor> | "The mark name cannot be the same name as any of the attributes in the PerformanceTiming interface [Navigation Timing]." http://w3c-test.org/webperf/specs/UserTiming/#ut-mark |
| 16:22 | <AryehGregor> | Ouch. |
| 16:22 | <AryehGregor> | Let's create more collision potential than we already have, huh? |
| 16:23 | <jgraham> | That's unbounded collision potential, no? |
| 16:23 | <jgraham> | Assuming that they might update the spec |
| 16:24 | <AryehGregor> | Yes, which is a problem we normally only have with a) Window, b) Node and descendant interfaces (maybe, if we don't institute one of the proposals for fixing that), c) authors who stick properties on objects they shouldn't (thankfully not common enough to worry about in most cases). |
| 16:26 | AryehGregor | does not understand why User Timing doesn't just expose a more accurate Date.now() function, instead of this mark() machinery |
| 16:26 | <jgraham> | AryehGregor: Send feedback? The people in that WG aren't very expert at all the things that people in WebApps take for granted |
| 16:27 | <jgraham> | (I mean about the collision thing) |
| 16:27 | <AryehGregor> | I could, but I'm supposed to be doing transition/animation tests and I've already procrastinated enough on transforms. |
| 16:27 | AryehGregor | was really just looking to see if this stuff would be useful for writing transition/animation tests |
| 16:28 | <AryehGregor> | "If the markName argument is not specified, the method returns an associative Array. Each key in the Array is a mark name. The value associated with the mark name is an Array of timestamps for that mark. In JSON notation, the data structure will look similar to this: { "mark1": [0, 1, 3], "mark2": [2] }" Also known as an object . . . ? |
| 16:28 | <AryehGregor> | I should really send feedback here, or someone should. |
| 16:29 | <bga> | a:visited:after { position: absolute; left: -9999; top: -9999; background: url(attr(href)) } |
| 16:29 | <jgraham> | AryehGregor: I think the point is possibly that Date.now() is not monotonically increasing |
| 16:29 | <bga> | i happy that is not work |
| 16:29 | <AryehGregor> | jgraham, right, so define a new function that returns a monotonically increasing number of milliseconds since page load. |
| 16:30 | <jgraham> | AryehGregor: I agree that would be a simpler solution |
| 16:30 | AryehGregor | is particularly uninterested in subscribing to Yet Another Mailing List |
| 16:30 | <jgraham> | This API seems too high level |
| 16:31 | <jgraham> | But I assume that if I say that I will be told that Microsoft have already implemented it and won't change |
| 16:31 | <AryehGregor> | And I'd really have to read all the historical discussions about the spec before I could suggest in good conscience that it be scrapped and replaced with a single function. |
| 16:31 | <AryehGregor> | Heh. |
| 16:31 | <AryehGregor> | Well, as long as everyone else agrees to scrap it . . . it's only LC after all, right? |
| 16:31 | <jgraham> | I did read some of the discussion about this |
| 16:31 | <AryehGregor> | And still prefixed. |
| 16:31 | <jgraham> | and I think this did come up |
| 16:37 | <AryehGregor> | And they said "thanks for the feedback, but we're not going to bother changing it because we've already implemented it", or what? |
| 16:39 | <jgraham> | Well now I look I can't see that |
| 16:39 | <jgraham> | I just see bz pointing out that they should use a monotonic clock |
| 16:40 | <jgraham> | But I could well be missing something, I have only read like 5% of the email to that list |
| 16:40 | <AryehGregor> | The spec says UAs have to use a monotonic clock. |
| 16:40 | <jgraham> | Yeah, they adopted his feedback |
| 17:11 | <remysharp> | isn't there a method for listening for all events against the window? I'm sure I saw it in some spec, but forget what it's called... |
| 17:21 | <AryehGregor> | Is there a way to feature-test for event support? Like, say, what prefix is used for transitionend, if any? |
| 17:21 | <AryehGregor> | Or do I just have to register an event handler for each possible type? |
| 17:21 | AryehGregor | hates prefixes |
| 17:26 | <zcorpan> | so firefox tolerates colors with a length of 1-6 if it starts with a number |
| 17:26 | <zcorpan> | and webkit tolerate colors with a lenght of 1-6 if it doesn't contain a-f |
| 17:26 | <zcorpan> | (apart from the standard lengths 3 and 6) |
| 17:27 | <AryehGregor> | In what context? |
| 17:27 | <AryehGregor> | Oh, quirks. |
| 17:27 | <zcorpan> | while ie and opera only tolerate the lengths 3 and 6 |
| 17:27 | <zcorpan> | yeah |
| 17:27 | <zcorpan> | i'll go with ie/opera |
| 17:27 | <smaug____> | AryehGregor: there isn't any real good way |
| 17:27 | <smaug____> | D3E proposes a way |
| 17:28 | <AryehGregor> | Yay. |
| 17:29 | <zcorpan> | hey i don't need to extend the css grammar |
| 17:30 | <zcorpan> | it's either a NUMBER, IDENTIFIER or IDENT |
| 17:36 | <dglazkov> | good morning, Whatwg! |
| 17:36 | <jgraham> | dglazkov: Remove prefixes from the web and then I will let you call it a good morning |
| 17:44 | matjas | needs textarea@pattern again. grrrr |
| 17:51 | <dglazkov> | jgraham: no worries. It's just a matter of time. |
| 17:51 | <dglazkov> | :) |
| 18:25 | <zcorpan> | ok http://simon.html5.org/specs/quirks-mode the first two quirks should now be a bit closer to reality |
| 18:42 | <zcorpan> | anyone have any brilliant ideas for how to research which properties need those quirks? |
| 18:44 | <AryehGregor> | Argh. How am I supposed to write tests for transition JS APIs when CSSOM interop is so atrocious? |
| 18:45 | <zcorpan> | assume perfect CSSOM compliance and file bugs :-P |
| 18:46 | <AryehGregor> | For that matter, some of these contradict CSS2.1, like "left" always computing to "auto" in some browsers if position is static . . . |
| 18:46 | <AryehGregor> | In some cases, nothing actually defines getComputedStyle()'s values at all. |
| 18:47 | <AryehGregor> | How does 'clip' serialize, with commas or without? |
| 18:47 | <zcorpan> | parsing of 'clip' is also handwavy in css2.1 |
| 18:48 | <AryehGregor> | Tons of computed values in various CSS specs don't give fine details like order. |
| 18:48 | <Ms2ger> | clip(a, b c, d)? |
| 18:48 | <AryehGregor> | Relevant part of definition for computed value of 'clip': "a rectangle with four values". |
| 18:48 | <AryehGregor> | Also, text-shadow serializes with the color first in some browsers, the lengths first in others. |
| 18:48 | <AryehGregor> | SIGH. |
| 18:51 | <AryehGregor> | jgraham, I'm adding an assert_in_array() function, okay? It's needed when there are multiple legitimate results. |
| 18:53 | <zcorpan> | IE is funny with 'font' |
| 18:53 | <zcorpan> | font:100 serif (not applied) |
| 18:53 | <zcorpan> | font:100p serif (applied) |
| 18:53 | <zcorpan> | font:100px serif (applied, gets a different font family!) |
| 18:54 | <nimbu> | is this in quirks mode zcorpan? |
| 18:55 | <zcorpan> | yeah |
| 18:55 | <zcorpan> | or maybe it's the 100p line that gets a different font family |
| 18:55 | <zcorpan> | well, i'll have to leave that for another day |
| 18:56 | <Hixie> | zcorpan: presumably it gets the "p serif" font family |
| 18:57 | <Hixie> | zcorpan: try font:100p serif, monospace; |
| 18:57 | <zcorpan> | Hixie: yeah, turns into monospace |
| 19:01 | <zcorpan> | Hixie: sorry to be such a PITA, but you didn't fix it like i said, so they still aren't equivalent (the webvtt signature thing) |
| 19:02 | <zcorpan> | Hixie: the file "WEBVTT" is now said to be successfully processed but the sniffing requires a space, tab or newline |
| 19:03 | <Hixie> | zcorpan: an empty WebVTT file seems like something pointless to sniff for |
| 19:03 | <Hixie> | zcorpan: but i guess i can mention that an empty WebVTT file would just be the six characters WEBVTT |
| 19:03 | <Hixie> | +/- BOM |
| 19:04 | <zcorpan> | in the text/vtt registration? WFM |
| 19:05 | <zcorpan> | although i'd rather have the parser match the current sniff rules, i think |
| 19:06 | <Hixie> | why? |
| 19:07 | <Hixie> | seems weird to make a trailing newline be the difference between onload and onerror |
| 19:07 | <Hixie> | AryehGregor: looking at window.find()... you suggest in one of these e-mails that we should make execCommand work on Range objects, is that still something you're looking at? |
| 19:07 | <AryehGregor> | Hixie, not actively, no. |
| 19:08 | <zcorpan> | i dunno. either way works |
| 19:09 | <Hixie> | AryehGregor: i'm trying to work out if i should spec window.find(), spec a subset of window.find() that just does the "destroy selection to let people manipulate the dom" thing people use it for, or punt on window.find() on the assumption that we'll add an API specifically for getting Ranges for search strings and then make execCommand() work on those |
| 19:09 | <Hixie> | AryehGregor: is there more to the Range/execCommand() thing than just putting execCommand() on the Range interface? |
| 19:10 | <Hixie> | (laughing at my naivete here is acceptable ;-) ) |
| 19:10 | <AryehGregor> | Hixie, potentially, yes. Not a *lot* more. The issue that comes to mind is that the range mutation algorithms don't work as we want in all cases, so there are plenty of places where execCommand() resets the selection to something specific -- these might not all be trivial to make work if the selection can be something arbitrary. |
| 19:10 | <AryehGregor> | But we could just let execCommand() called on a Range do random weird things to the selection if we want, I guess. :) |
| 19:11 | <Hixie> | AryehGregor: or we could just make certain commands only work on selections |
| 19:11 | <AryehGregor> | Wouldn't be very useful, IMO. |
| 19:11 | <AryehGregor> | E.g., if you do insertText, the selection is first deleted, then some text is inserted, then the selection is put before the start of the text (which may or may not have any relationship to the start of the original selection). |
| 19:11 | <AryehGregor> | I guess range mutation rules might work acceptably. |
| 19:12 | <AryehGregor> | I don't know offhand. |
| 19:12 | <Hixie> | k |
| 19:16 | <MikeSmith> | somebody needs to write a mail client in canvas |
| 19:16 | <StoneCypher> | i wrote a polygon image evolver in canvas the other day because i was bored |
| 19:16 | <MikeSmith> | cool |
| 19:16 | <AryehGregor> | Okay, so IE10 Developer Preview sometimes serializes text-shadow like "1px 2px 0px rgb(0,0,255)" (note lack of space after commas) and sometimes as "1px 2px 0px 0px rgba(0, 0, 255, 1)" (note *four* lengths). |
| 19:16 | <AryehGregor> | . . . |
| 19:17 | <Hixie> | rgb() vs rgba(), too |
| 19:17 | <Hixie> | does it depend on the input colour syntax? |
| 19:17 | <Hixie> | or only the number of lengths? |
| 19:18 | <zcorpan> | AryehGregor: does it support hashless hex colors or unitless lengths in text-shadow? :-) |
| 19:18 | <AryehGregor> | Hixie, it changes the output format right after I set it to a different value, with a transition applied (so the computed value shouldn't have changed yet). |
| 19:19 | <Hixie> | AryehGregor: funky |
| 19:19 | <AryehGregor> | jgraham, can I go ahead and commit my assert_in_array thing, or do you want to review it first? |
| 19:19 | <jgraham> | AryehGregor: Well I hate how bloated the API is becoming, but I don't know what to do about it, so sure |
| 19:19 | <Hixie> | AryehGregor: i wonder if it's working out that it'll become rgba() in due course, and so preemptively adding alpha to its concerns, and then just rgb() has a differnet serialiser than rgba() |
| 19:19 | <AryehGregor> | Hixie, no, it just generally serializes colors as rgba(..., 1). |
| 19:20 | <AryehGregor> | :) |
| 19:21 | <Hixie> | so what's with the rgb() case before then? |
| 19:21 | <Hixie> | weird |
| 19:21 | <jgraham> | AryehGregor: I mean sure you can write one |
| 19:21 | <Hixie> | AryehGregor: you speccing the c12n algorithm for cssom? |
| 19:21 | <jgraham> | AryehGregor: If you make a pull request against github I will review it |
| 19:21 | <AryehGregor> | IE is generally the weirdest browser, followed by Opera, then WebKit, then Gecko. |
| 19:21 | <AryehGregor> | Hixie, I'd be happy to, but no. Just writing some transition tests. |
| 19:21 | <AryehGregor> | jgraham, okay. |
| 19:21 | <Hixie> | AryehGregor: ah, cool |
| 19:22 | <AryehGregor> | I already wrote a bunch of transforms tests. Transitions are much more of a pain. |
| 19:22 | <zcorpan> | AryehGregor: with hashless hex color, the weirdness was reversed, i think, possibly having webkit at the end of the scale |
| 19:24 | <AryehGregor> | jgraham, done. |
| 19:24 | <AryehGregor> | <3 github. |
| 19:29 | <Ms2ger> | Bah, github |
| 19:29 | <zcorpan> | boo, logs are down |
| 19:30 | <Ms2ger> | Where does testharness.js live nowadays? |
| 19:33 | <AryehGregor> | Ms2ger, apparently https://github.com/jgraham/testharness.js |
| 19:40 | <jgraham> | Not relly |
| 19:40 | <jgraham> | *really |
| 19:41 | <jgraham> | It really lives in http://dvcs.w3.org/hg/resources/ |
| 19:41 | <jgraham> | But the W3C infrastructure sucks |
| 19:42 | <tomer> | Hi. Is there any spec for how Unicode Emojis should work in a web browser? For example, if and how these icons should be replaced with graphical/colorful icons. Currently it is implemented in iOS and Safari on OSX10.7, and I'm looking for how we can push it to other browsers as well. |
| 19:44 | <tomer> | … For example, the following bug mention how it can be controlled by a special control character, but we might want a better way to control its representation directly from the HTML/CSS sources, like how we are dealing with the spellchecking on input fields in Firefox. |
| 19:44 | <tomer> | https://bugzilla.mozilla.org/show_bug.cgi?id=727276 |
| 20:29 | <Ms2ger> | "This is a very boring kind of report this does not take each & every thing into consideration HTML 4.0.1 was good infact was better way better.." |
| 20:35 | <annevk> | whoa |
| 20:35 | <annevk> | locked myself out of the house |
| 20:35 | <annevk> | inbox 400 again |
| 20:35 | <bga> | ctrl + a, del |
| 20:36 | <bga> | so easy |
| 20:37 | <annevk> | heh |
| 20:37 | <annevk> | I was not alone https://twitter.com/#!/slicknet/status/169840384502009856 |
| 20:50 | <isherman> | tantek: ping |
| 21:09 | <Ms2ger> | https://cvs.khronos.org/svn/ |
| 21:09 | <Ms2ger> | ... |
| 21:10 | <TabAtkins> | git://cvs.khronos.org/svn/ |
| 21:10 | <Ms2ger> | TabAtkins-- |
| 21:11 | <Ms2ger> | Anybody know how to submit tests to WebGL's test suite? |
| 21:11 | <TabAtkins> | I can ask. |
| 21:12 | <Ms2ger> | That would be nice |
| 21:12 | <TabAtkins> | File a bug on the public khronos bugzilla, attach the test as a patch. |
| 21:12 | <Ms2ger> | Good |
| 21:17 | <Ms2ger> | Though I have no idea how to test this... |
| 21:29 | <annevk> | it's kind of amusing how congress is interested in apps being to "steal" your addressbook data on phones, but nobody is asking about traditional operating systems |
| 21:30 | <annevk> | but I guess some interest in protecting privacy is better than none whatsoever |
| 21:31 | <TabAtkins> | Congress is half corrupt and half pig-ignorant about tech. :/ |
| 21:37 | <jgraham> | It's probably not as common to have large a number of personal details on your non-portable device |
| 21:38 | jgraham | is pretty sure the whole thing must be illegal in the EU already |
| 21:39 | <jgraham> | Because AFAIK if I give you my personal information, that doesn't give you the right to do what you like with it |
| 21:39 | <jgraham> | At least it doesn't for businesses |
| 21:40 | <jgraham> | If it does for individuals it seems like a loophole that will be closed |
| 21:40 | <jgraham> | (IANAL, TINLA) |
| 22:06 | <jwalden> | privacy interests in information disclosed to third parties are weird in the US |
| 22:07 | <jwalden> | rather bizarrely, reasonable expectations often are not |
| 22:22 | <TabAtkins> | Who has control over the W3C's github account? |
| 22:22 | <TabAtkins> | And how can I get in on that action? |
| 22:24 | <jgraham> | The W3C has a github account? |
| 22:24 | <TabAtkins> | Yes. |
| 22:25 | <TabAtkins> | https://github.com/w3c |
| 22:25 | <jgraham> | TabAtkins: the email address on the profile page is a clue |
| 22:25 | <TabAtkins> | Hurp durp, yes. |
| 22:25 | annevk | was about to say |
| 22:25 | <annevk> | members too |
| 22:33 | <Hixie> | annevk, TabAtkins: i don't really want to jump into the perf thread, but fwiw, readyState is already a string on Document |
| 22:34 | <TabAtkins> | Cool. |
| 22:34 | <TabAtkins> | So it's *entirely* inconsistent across the platform. |
| 22:35 | <Hixie> | welcome to the web |
| 22:35 | <TabAtkins> | Yup. |
| 22:35 | <Hixie> | readyState in particular was a microsoft invention both on Document (DOMString) and XHR (unsigned short) |
| 22:36 | <TabAtkins> | That's awesome. |
| 22:37 | <annevk> | Hixie: oh right, well, if someone replies I'll point that out |
| 23:19 | <jwalden> | clearly readyState on XHR2 should switch to being a string |
| 23:26 | <annevk> | http://www.reddit.com/r/fifthworldproblems nice |
| 23:27 | <annevk> | jwalden: cannot make incompatible changes; but you can just use the events and forget about readyState |
| 23:27 | <jwalden> | annevk: I am tempted to string things along further, but I was not being serious |
| 23:30 | <annevk> | jwalden: we've been low on sarcasm lately |
| 23:30 | <annevk> | apparently it's all about memes these days |
| 23:33 | <annevk> | http://eric.van-der-vlist.com/blog/2012/02/15/xml-prague-2012-the-web-would-be-so-cool-without-the-web-developers/ is a pretty neat summary of the XML thing past weekend |
| 23:34 | <gsnedders> | jwalden: seeming you're around, know what the semantics of Object.prototype.__proto__ are? Just a accessor property that claims data property which mutates the [[Prototype]] of this? |
| 23:34 | <gsnedders> | (in SM) |
| 23:35 | <jwalden> | gsnedders: nobody represents it as an accessor, but it just reflects [[Prototype]] |
| 23:35 | <jwalden> | and lets you change it |
| 23:36 | <gsnedders> | And there's nothing magic about it apart from the fact that it is secretly an accessor, thus the fact it can be deleted fine? |
| 23:37 | <gsnedders> | (or have I misunderstood what be said on es-discuss about its deletability) |
| 23:42 | <gsnedders> | The more and more I think about it, the less I'm convinced my gut reaction of not wanting to ever make it an accessor is right, and the more I want to make it an accessor. |
| 23:46 | <TabAtkins> | jwalden: I thought that we (chrome) represent it as an accessor in recent builds? |
| 23:48 | <jwalden> | TabAtkins: might have changed; last I remembered the internal API implemented it as a CALLBACK |
| 23:48 | <jwalden> | er |
| 23:48 | <jwalden> | INTERCEPTOR, sorry, I think |
| 23:48 | <gsnedders> | TabAtkins: AFAIK only JSC has changed recently |
| 23:49 | <TabAtkins> | I only know what I overheard in the __proto__ thread on es-discuss. |
| 23:49 | <jwalden> | gsnedders: well, we mark it as non-configurable, so it can't be deleted; I think we should change this, and probably make it a getter/setter pair, but I haven't looked too closely |
| 23:50 | <gsnedders> | jwalden: The consensus on es-discuss seemed to be against making it an accessor, because the setter function could then be got. But you don't seem to lose much, seeming you can't statically determine the lookup anyway… |
| 23:50 | <gsnedders> | jwalden: And I thought someone claimed it could be deleted. Definite consensus that it much be deletable. |
| 23:50 | <gsnedders> | s/delete/configurable/ |
| 23:51 | <gsnedders> | Bleh, now what I wrote just makes no sense, ignore that change :P |
| 23:51 | <jwalden> | frankly I say we nuke it from orbit, and we all do it at once and suck up the pain |
| 23:51 | <gsnedders> | jwalden: That's what I've been arguing for too. |
| 23:51 | <jwalden> | IE is totally in the right here |
| 23:53 | <gsnedders> | jwalden: Well, I've more been arguing for nuking the setter, I have no issue with Object.defineProperty(Object.prototype, "__proto__", {get:function(){return Object.getPrototypeOf(this);}, configurable:true}); given those exact semantics |
| 23:53 | <gsnedders> | (i.e., it's something which has obvious native semantics and is contained in a single place) |
| 23:54 | <jwalden> | that would be better than nothing, certainly |
| 23:54 | <gsnedders> | I think the ship has sailed with the getter — it looks more and more like mobile is forcing the setter :( |
| 23:54 | <gsnedders> | At the very least it'd be nice to poison the __proto__ in strict. |
| 23:54 | <jwalden> | stupid idiot web developers |
| 23:55 | <jwalden> | __proto__-setting is poisoned, at least |
| 23:55 | <gsnedders> | jwalden: In what? |
| 23:55 | TabAtkins | won't mention his own use of __proto__. |
| 23:56 | <jwalden> | er |
| 23:56 | <jwalden> | no, sorry, I was thinking of __proto__-setting on non-extensible objects |
| 23:56 | <gsnedders> | jwalden: But still, in what? |
| 23:56 | <gsnedders> | Or is that required by spec? |
| 23:56 | gsnedders | guesses test262 doesn't test that seeming __proto__ is non-standard |
| 23:56 | <jwalden> | "use strict"; Object.freeze({}).__proto__ = {} |
| 23:57 | <jwalden> | TypeError: Object.freeze({}).__proto__ is not extensible |
| 23:57 | <jwalden> | if you have a non-extensible object, you can't change its prototype |
| 23:57 | <jwalden> | and a TypeError gets thrown if you try in strict mode |
| 23:57 | <jwalden> | I think most engines implement this |
| 23:57 | <gsnedders> | Carakan seems to at least |
| 23:58 | gsnedders | didn't remember that being the case |
| 23:58 | <gsnedders> | http://trac.webkit.org/changeset/107498 is the JSC change |
| 23:58 | <gsnedders> | __proto__ is an accessor in JSC now. |
| 23:59 | <gsnedders> | And reverted. |
