| 03:14 | <Ernston> | hello |
| 03:15 | <MikeSmith> | Ernston: hi |
| 03:16 | <Ernston> | i have ended up on this page whilst validating a page |
| 03:16 | <Ernston> | i added meta tags which i thought where ok to use |
| 03:16 | <Ernston> | <meta name= "language" content= "english" /> |
| 03:18 | <MikeSmith> | is name=language registered? |
| 03:18 | <MikeSmith> | or wait |
| 03:18 | <Ernston> | am lost |
| 03:18 | <MikeSmith> | the more important question is, Why are you using that to begin with? |
| 03:19 | <Ernston> | primary language of the site is english |
| 03:19 | <Ernston> | so it can be translated |
| 03:19 | <Ernston> | with chrome |
| 03:19 | <MikeSmith> | you know about the lang attribute? |
| 03:20 | <Ernston> | not exactly |
| 03:20 | <MikeSmith> | that's the normal way to indicate the language of the document |
| 03:20 | <Ernston> | oh |
| 03:20 | <MikeSmith> | http://www.whatwg.org/specs/web-apps/current-work/multipage/elements.html#the-lang-and-xml:lang-attributes |
| 03:20 | <MikeSmith> | <html lang="en-US"> |
| 03:21 | <Ernston> | ive found a couple of tags |
| 03:21 | <Ernston> | one being the 1 i showed you |
| 03:21 | <Ernston> | other is the <meta name="author" |
| 03:21 | <Ernston> | gives me error in validator too |
| 03:21 | <Ernston> | thee been depreciated? |
| 03:21 | <Ernston> | these |
| 03:22 | <Ernston> | i am sure ive used these in the past with no probs |
| 03:22 | <Ernston> | using html5 document btw. |
| 03:23 | <MikeSmith> | we recently updated the validator to check name values |
| 03:23 | <Ernston> | arr |
| 03:23 | <MikeSmith> | the spec requires is |
| 03:23 | <MikeSmith> | *it |
| 03:23 | <MikeSmith> | we just were not checking it before |
| 03:23 | <MikeSmith> | Ernston: the thing about most of the meta@name values is that they are just noise |
| 03:23 | <MikeSmith> | junk |
| 03:23 | <Ernston> | this registering of metadata confuses me |
| 03:23 | <MikeSmith> | no applications actually do anything with most of the values |
| 03:24 | <MikeSmith> | so it's usually not worth your time to try to register them |
| 03:24 | <Ernston> | i thought html5 was making way for more semantic elements |
| 03:24 | <MikeSmith> | the better solution is just to not use name values that no applications actually do anything with |
| 03:24 | <Ernston> | what should i do about this lot |
| 03:24 | <Ernston> | <meta name="audience" content="all" /> |
| 03:24 | <MikeSmith> | oh man |
| 03:24 | <MikeSmith> | that is worthless |
| 03:25 | <MikeSmith> | it's pointless |
| 03:25 | <Ernston> | its stuff i pulled from my old code |
| 03:25 | <MikeSmith> | yeah |
| 03:25 | <Ernston> | normally touch meta data last |
| 03:25 | <MikeSmith> | that's where most people pull it from |
| 03:25 | <Ernston> | things have changed since i completed a site lol |
| 03:25 | <MikeSmith> | and that's the reason we check and flag it as an error |
| 03:25 | <Ernston> | 3 more to go |
| 03:25 | <Ernston> | <meta name= "distribution" content= "global" /> |
| 03:25 | <Ernston> | does that do anything? lol |
| 03:25 | <MikeSmith> | worst of all |
| 03:25 | <MikeSmith> | get rid of that, man |
| 03:26 | <Ernston> | they are all getting deleted |
| 03:26 | <Ernston> | i dont like errors in my page on validation |
| 03:26 | <Ernston> | am ocd |
| 03:26 | <MikeSmith> | part of the point of the validator is to let you know about cruft you have in your docs that is not necessary or useful |
| 03:26 | <MikeSmith> | good good |
| 03:26 | <Ernston> | heres a good 1 |
| 03:26 | <MikeSmith> | rip them out |
| 03:26 | <Ernston> | <meta name="address" content="http://www.h |
| 03:26 | <Ernston> | lol |
| 03:27 | <Ernston> | am glad to see junk has been removed |
| 03:27 | <Ernston> | and its not be going more insane |
| 03:27 | <Ernston> | me* |
| 03:28 | <Ernston> | even copyright attribute dead? |
| 03:28 | <Ernston> | took it out anyway due to error |
| 03:29 | <Ernston> | ive failed if i dont pass w3 |
| 03:29 | <Ernston> | yslow and css validtor |
| 03:29 | <Ernston> | validator to the best that i can |
| 03:29 | <Ernston> | :D |
| 03:29 | <Ernston> | damn my head looks naked |
| 03:50 | <Hixie> | is there any sane way to have objects in JS have two constructors? |
| 03:51 | <Hixie> | i need one constructor for when i'm creating the object normally, and one for when i'm creating it from a serialised representation |
| 03:51 | <Hixie> | but there doesn't seem to be a way to do tht... |
| 03:53 | <zewt> | well, that's when you switch to factories... |
| 03:54 | <Hixie> | ugh |
| 03:54 | <Hixie> | but that's so ugly |
| 04:01 | <zewt> | i prefer it, but i guess what you're looking for is overloaded constructors ... don't know if webidl allows that |
| 04:02 | <heycam> | it does |
| 04:02 | <heycam> | but constructors are required to return instances of that interface |
| 04:16 | <Hixie> | webidl isn't accessible from js :_) |
| 04:16 | <heycam> | ah |
| 04:16 | <heycam> | just switch on arguments.length inside your constructor |
| 04:16 | <heycam> | return something different depending on how it was called |
| 04:17 | <Hixie> | what if they both have zero arguments or some such? |
| 04:17 | <heycam> | then… how do you distinguish which one is being called? |
| 04:17 | <Hixie> | by name, in other languages |
| 04:17 | <heycam> | or, are they meant to have different names (like [NamedConstructor]s)? |
| 04:18 | <heycam> | in other languages? |
| 04:18 | <heycam> | btw you can return whatever you want from a constructor function, and it doesn't have to be related to the constructor itself |
| 04:18 | <heycam> | function A() { } function B() { return new A() } |
| 04:19 | <heycam> | then new B() will be an A |
| 04:19 | <Hixie> | like ObjectPascal |
| 04:19 | heycam | was a big Delphi fan a long time ago |
| 04:20 | <Hixie> | you call Foo := TFoo.Create(); or Foo := TFoo.MakeFromStream(stream); or whatever |
| 04:20 | <Hixie> | and they can be both constructors |
| 04:20 | <heycam> | ah right, that seems familiar |
| 04:21 | <Hixie> | (you declare them as constructor Create(); or constructor MakeFromStream(); instead of function Foo(): Integer; or whatever) |
| 04:21 | <Hixie> | perl is similar |
| 04:21 | <heycam> | you could do: function A() { } A.FromStream = function() { var a = new A(); a.something = 123; return a } |
| 04:21 | <heycam> | then new A.FromStream() |
| 04:21 | <zewt> | what does it matter if it's actually a constructor? as far as users are concerned it just means you don't say "new" |
| 04:21 | <heycam> | looks nice |
| 04:21 | <Hixie> | yeah, that might work |
| 04:21 | <zewt> | x = A.FromStream(); |
| 04:21 | <heycam> | either way |
| 04:21 | <zewt> | that's all i meant by factory |
| 04:22 | <zewt> | (not that "factory" is actually anything special in dynamically-typed languages) |
| 05:01 | <Hixie> | this may seem like a silly question but... what purpose does the 'constructor' property in JS have? |
| 05:02 | <Hixie> | i can't find anything in the JS spec that defines anything that it _does_ |
| 05:02 | <heycam> | Hixie, it's so you can do myInstance.constructor and find out which constructor (probably) constructed the object |
| 05:02 | <heycam> | not accurate of course because you can return whatever froma constructor function |
| 05:03 | <Hixie> | but does anything actually ensure that a constructor function sets 'constructor' on the instance? |
| 05:03 | <heycam> | you should get it from the prototype |
| 05:03 | <heycam> | i thought |
| 05:03 | <heycam> | or maybe I misremember, I guess you are meant to set it yourself? |
| 05:04 | <heycam> | I remember now writing this kind of code: function MyClass() { } MyClass.prototype.constructor = MyClass; |
| 05:07 | <Hixie> | yeah it's a line like that that made me wonder |
| 05:13 | <benvie> | the constructor property is automatically set when a function is created, along with the default protoype. It's rare for people to overwrite the prototype but if they do then they have to manually set the constructor property or else it'll just point to Object.prototype. The constructor property serves no purpose aside from that fact that people rarely change it or realize they can, so it |
| 05:13 | <benvie> | usually is accurate. It is used by most dev tools for introspection in order to give a useful name to objects based on the constructor's name |
| 05:13 | <benvie> | er point to Object, Object.prototype.constructor |
| 05:17 | <heycam> | oh yes, benvie is right -- `function f(){}` results in f.prototype.constructor == f |
| 05:17 | <heycam> | so I'm not sure why I would've been doing that manually |
| 05:18 | <heycam> | was a long time ago, /me probably just cargo culted it |
| 05:18 | <benvie> | something like `Ctor.prototype = { constructor: Ctor, ..methods }` |
| 05:18 | <heycam> | ah, if you overwrite prototype, good point |
| 05:19 | <benvie> | basicaly if you want to set the new prototype methods en masse you just overwrite it, in which case you need to manually set it again or else you show up in dev tools labelled as 'Object' |
| 05:19 | <heycam> | as it's often more convencient to write that style rather than a sequences of assignments to Ctor.prototype.blah |
| 05:20 | <benvie> | or in cases where you're inheriting. `F.prototype = Object.create(Super.prototype); F.prototype.constructor = F;` |
| 05:28 | <delight> | there is an interesting discussion going on @ http://code.google.com/p/chromium/issues/detail?id=44116 about i18n, floats and the html5 type="number" input field |
| 05:30 | <delight> | were would be the appropriate place to address such concerns esp. the additional attribute localized="true" needed for a smoother transition ? |
| 05:52 | <Hixie> | is there any difference between "new foo" and "new foo()" ? |
| 06:31 | <jgraham> | Hixie: No, unless I am horribly misremembering |
| 06:31 | <jgraham> | Oh look. Opera 12 got released |
| 06:32 | <hsivonen> | jgraham: about time |
| 06:57 | <AryehGregor> | Hixie, in what way is contenteditable a lower-level primitive than input/textarea? |
| 07:13 | <hsivonen> | why are the HW accel and WebGL prefs in Opera 12 integers and not booleans? |
| 07:46 | <sm0lm> | where can i find a reference of the default styles in CSS each markup element implies without any explicit CSS declarations? ie, what CSS would i need to put in `div { **HERE** }` to make my <div>s behave like <pre>s? |
| 07:47 | <zcorpan> | sm0lm: http://www.whatwg.org/specs/web-apps/current-work/multipage/rendering.html#rendering |
| 07:47 | <jgraham> | HTML spec, rendering section |
| 07:49 | <zcorpan> | how is target=_unrelated different from rel=noreferrer ? |
| 07:50 | <jgraham> | It would be totally disconnected AIUI |
| 07:50 | <jgraham> | So the opened window would have window.opener == window |
| 07:51 | <jgraham> | (or null? I forget how that one works) |
| 07:51 | <zcorpan> | i thought noreferrer does that as well |
| 07:51 | <jgraham> | It does? Pointer? |
| 07:52 | <zcorpan> | "This keyword also causes the opener attribute to remain null " http://www.whatwg.org/specs/web-apps/current-work/multipage/links.html#link-type-noreferrer |
| 08:01 | <jgraham> | Only works in WebKit, it seems |
| 08:05 | <zcorpan> | wider deployed than target=_unrelated then :-) |
| 08:06 | <jgraham> | Yes :) |
| 08:19 | <Ms2ger> | Yay, HTMLElement.itemType isn't web-compatible |
| 08:24 | <hsivonen> | Ms2ger: huh? bug#? |
| 08:24 | <Ms2ger> | https://bugzilla.mozilla.org/show_bug.cgi?id=763229 |
| 08:24 | <hsivonen> | thanks |
| 08:33 | <annevk2> | whoa |
| 08:36 | <abarth> | annevk2: thanks for writing http://wiki.whatwg.org/wiki/Anolis and http://wiki.whatwg.org/wiki/Howto_spec |
| 08:36 | <abarth> | annevk2: they're super useful |
| 08:38 | <annevk> | thanks |
| 08:39 | <zcorpan> | Ms2ger: i hope that can be evang'ed |
| 08:39 | <annevk> | abarth: if you have any feedback or get feedback from others, let me know; I'd love to improve them, especially the latter |
| 08:39 | zcorpan | files an opera OTW bug about it |
| 08:39 | <abarth> | annevk: it's missing a description of class=switch :) |
| 08:39 | <abarth> | i had to crib that from domcore |
| 08:40 | <Ms2ger> | zcorpan, say so in the bug if you prefer evang? |
| 08:40 | <abarth> | it's also not clear to me how to contribute my xref JSON files back to the specification-data repo |
| 08:40 | <Ms2ger> | abarth, create an account on bitbucket |
| 08:41 | <annevk> | abarth: ah yeah, we should mention that |
| 08:41 | <abarth> | darn! abarth already exists on bitbucket |
| 08:41 | <annevk> | abarth: what are you working on? |
| 08:41 | <abarth> | https://github.com/abarth/sysapps |
| 08:42 | <Ms2ger> | Oh, evil stuff ;) |
| 08:42 | <abarth> | mostly i'm trying to get some skeleton specs ready so the read editors can take over |
| 08:43 | <abarth> | Ms2ger: ok, i've created the account adambarth on https://bitbucket.org/ |
| 08:43 | <zcorpan> | Ms2ger: done |
| 08:43 | <Ms2ger> | You've got access |
| 08:43 | <abarth> | cool, thanks |
| 08:43 | <Ms2ger> | Np |
| 08:43 | <abarth> | when should I push the JSON files? |
| 08:44 | <abarth> | I probably should wait until the charter is approved etc |
| 08:44 | <Ms2ger> | Whenever someone wants to use references to them, I'd say |
| 08:45 | <abarth> | ok. right now these specs are all in one repo, so it's easy for them to cross-reference each other |
| 09:18 | <sm0lm> | can datalist elements be used for fetching server-generated search suggestions in a searchbox like google suggest or youtube? |
| 09:21 | <zcorpan> | there used to be a declarative way to do that, but it was dropped |
| 09:21 | <zcorpan> | you have to fetch it yourself and build the datalist <option>s yourself |
| 09:23 | <sm0lm> | zcorpan, wtf, why would it be dropped? |
| 09:23 | <sm0lm> | any historical docs that refer to it you might be able to link me to, zcorpan ? |
| 09:24 | <annevk> | sm0lm: http://dev.opera.com/articles/view/an-html5-style-google-suggest/ |
| 09:25 | <sm0lm> | annevk, it just occurred to me that you probably wrote that :P |
| 09:25 | <annevk> | I did write that, some time ago :) |
| 09:25 | <annevk> | and zcorpan wrote the first demo I think |
| 09:26 | <sm0lm> | are you guys norwegian then? |
| 09:26 | <annevk> | Dutch and Swedish |
| 09:28 | <sm0lm> | cool. |
| 09:28 | <sm0lm> | i'm a boring american. |
| 09:29 | <sm0lm> | annevk, do u know what happened to the in-markup declaration? |
| 09:29 | <sm0lm> | what was the rationale for killing it? |
| 09:30 | <annevk> | it was both a declaration and an external XML format |
| 09:30 | <sm0lm> | annevk, what do u mean? |
| 09:31 | <annevk> | we (WHATWG) dropped it because it was only implemented by us (Opera) and it was easy enough to do using a little bit of scripting |
| 09:31 | <sm0lm> | annevk, but why not make it easy forpeople to see and notice in the whatwg docs? |
| 09:32 | <annevk> | sm0lm: well we weren't quite sure it was the correct solution anymore |
| 09:33 | <sm0lm> | annevk, well that's what i'm wondering: what about it made it not the correct solution? seems pretty great to me.. |
| 09:33 | <annevk> | sm0lm: you might not want new network requests whenever the user types something new, but rather have an ongoing connection via WebSocket and populate the <datalist> through that |
| 09:34 | <annevk> | sm0lm: we also learned that XML is something we rather not require from web developers because it always goes wrong |
| 09:35 | <sm0lm> | annevk, json? |
| 09:36 | <sm0lm> | <datalist src="http://my/dyn_dlist?q=%s"; /> |
| 09:36 | <annevk> | yeah that would probably be a better format, but does not address the continuous connection issue |
| 09:38 | <sm0lm> | annevk, right. i don't understand the connection side of things that well, but couldn't you side step an issue of "[some people] might no want...." by giving two parallel ways of doing it and letting webdevs take it or leave it? |
| 09:41 | <annevk> | sm0lm: sure, but we should only do that if the new feature addresses the problem for the majority of developers |
| 09:41 | <sm0lm> | sm0lm, wouldn't it? |
| 09:41 | <sm0lm> | haha |
| 09:41 | sm0lm | talks to himself. |
| 09:41 | <sm0lm> | annevk, |
| 09:41 | <sm0lm> | mulligan^ |
| 09:41 | <annevk> | sm0lm: that's not entirely clear :) <datalist> without this feature is not widely implemented or used yet |
| 09:41 | <sm0lm> | annevk, wouldn't it, though? |
| 09:45 | <sm0lm> | so declarative remote server-gen'd datalists aren't necessarily gone forever? |
| 09:46 | <sm0lm> | annevk, or what about just have it implicitly open and keep open a socket for fetching dlists? |
| 09:48 | <annevk> | sm0lm: back later, lunch |
| 09:49 | <sm0lm> | annevk, enjoy. |
| 10:28 | <annevk> | sm0lm: no, not gone forever |
| 10:28 | <sm0lm> | annevk, ok-- cool. |
| 10:29 | <annevk> | sm0lm: I think once <datalist> is more widely adopted we can study how developers use it and standardize the most common pattern |
| 10:29 | <sm0lm> | what do u think of having an implicit instruction to start a socket that stays open? |
| 10:29 | <annevk> | not really sure how you see that working |
| 10:29 | <sm0lm> | annevk, makes sense. how long do you expect that to take, roughly? |
| 10:30 | <annevk> | dunno, I think <datalist> is now in WebKit, so it might see some more adoption |
| 10:31 | <sm0lm> | annevk, yeah, sorry. i really don't know mucch about using web societs or how they work. |
| 10:34 | <annevk> | Ms2ger: is there a way in bitbucket to follow you, but not get everything you do on my homepage? |
| 10:34 | <annevk> | Ms2ger: or should I just stop following you and only follow some projects... |
| 10:34 | <annevk> | Ms2ger: the Gecko commits are interesting, but I'm way more interested in changes to standards |
| 10:34 | <Ms2ger> | I've got the same issue :/ |
| 10:37 | <Ms2ger> | Not following me probably works bets |
| 10:37 | <Ms2ger> | best* |
| 10:41 | <annevk> | :( |
| 10:41 | <annevk> | hmm, limechat has no smileys |
| 10:46 | <jgraham> | That seems like a win |
| 10:47 | <odinho> | Smileys are essensials, I like how they are in irssi. All ascii goodness :] |
| 10:49 | <zcorpan> | sm0lm: http://dev.opera.com/articles/view/introducing-web-sockets/ |
| 10:50 | <sm0lm> | zcorpan, thanks, i'll have a read. |
| 10:50 | <sm0lm> | to whatever degree that this question isn't loaded (i know you both work for opera): how good is the dev.opera.com documentagtion for someone who really doesn't have any interest in opera itself? |
| 10:50 | <zcorpan> | sm0lm: unfortunately opera 12 still has the old websocket protocol which is disabled by default |
| 10:51 | <sm0lm> | is stuff only relevant to opera clearly marked? |
| 10:51 | <zcorpan> | sm0lm: usually dev.opera.com articles aren't necessarily opera-specific |
| 10:51 | <jgraham> | odinho: Well yes, the ascii representation is of course fine :) |
| 10:51 | jgraham | starts the countdown to someone using a unicode smilie |
| 10:52 | <zcorpan> | sm0lm: if other browsers implement the same feature, it should apply equally to them |
| 10:52 | <odinho> | ☺ |
| 10:52 | <zcorpan> | sm0lm: we also try to make demos cross-browser |
| 10:53 | <sm0lm> | zcorpan, great. thanks : |
| 10:53 | <sm0lm> | ) |
| 10:53 | <odinho> | We follow standards ;] |
| 10:53 | <sm0lm> | odinho, imagine that. |
| 10:53 | <zcorpan> | odinho: except when we don't :-P |
| 10:54 | <odinho> | zcorpan: Well, yes, but then it's often a bug, and we def. don't do it to be mean :P |
| 10:55 | <sm0lm> | hah, imagine that: not being mean. |
| 10:55 | <sm0lm> | that's revolutionary... /me thinks it should get codified in the spec. |
| 10:56 | <zcorpan> | odinho: not mean, just evil |
| 10:56 | sm0lm | proposes a strict ban on meanness and evilness in the spec. |
| 10:57 | zcorpan | goes back to writing evil tests |
| 10:57 | <odinho> | zcorpan: Someone else has copyright on that, and they kinda watered it down. :P |
| 10:57 | <sm0lm> | "in order to be compliant with the HTML specification, you must not be mean, or evil. |
| 10:57 | <sm0lm> | mehh.... |
| 10:57 | <sm0lm> | i mean, on the one hand google is probably like the most terrible internet company there is right now, or at least up there. |
| 10:58 | <sm0lm> | on the other hand, in the context of this whole gig, their product *is* FOSS at least... |
| 10:58 | <sm0lm> | i'd be way more interested in opera if it weren't proprietary. |
| 10:58 | <sm0lm> | it seems like high-quality software... |
| 10:58 | <nonge> | yay, go the FLOSS way, Opera! |
| 10:58 | <sm0lm> | but who knows what kinda shit u guys put in there? |
| 10:59 | Ms2ger | yawns |
| 10:59 | <zcorpan> | sm0lm: if you're curious, apply for a job :-) |
| 10:59 | <sm0lm> | nonge, i'm not necessarily saying that, but if you're going to crack "evil" jokes or make jabs at other companies' social responsibility practices, you know... |
| 10:59 | <sm0lm> | "let the one without sin throw the first stone" |
| 11:00 | <sm0lm> | is Ms2ger a googler? |
| 11:00 | <zcorpan> | mozilla |
| 11:01 | <sm0lm> | zcorpan, its true. |
| 11:01 | <jgraham> | Maybe he uses google though |
| 11:01 | <sm0lm> | in the social responsibility dept. and values integrity dept, they are. |
| 11:01 | <sm0lm> | but IMO distributing shitty sw is another form of evil. |
| 11:01 | <jgraham> | One who uses google should be a googler. One who is google shoudl be a googlee. |
| 11:01 | Ms2ger | is getting terribly confused here |
| 11:02 | <Ms2ger> | jgraham, Larry Page? |
| 11:02 | <sm0lm> | jgraham, i used to think that too... but i gave up and accepted the common usage long ago. |
| 11:02 | <jgraham> | Ms2ger: For all we know you are Larry Page |
| 11:03 | <Ms2ger> | Well, that's true |
| 11:03 | <sm0lm> | for all you know, sm0lm is Larry Page. |
| 11:04 | <Ms2ger> | That, too is true |
| 11:04 | <Ms2ger> | For all they know, we're both Larry Page |
| 11:05 | Ms2ger | wrestles with idlharness.js instead |
| 11:06 | <sm0lm> | Ms2ger, we're fooling them so well. |
| 11:06 | <Philip`> | I thought Google didn't want people genericising its trademarked name by verbing it, so surely "Googler" should be replaced by "Employee of Google Inc" |
| 11:06 | <sm0lm> | googler isn't verbing it in the way that they're trying to avoid though. |
| 11:06 | <Philip`> | (Not quite as catchy, admittedly) |
| 11:07 | <sm0lm> | but "google it" isn't as catchy as "perform a google search for it" either. |
| 11:07 | <sm0lm> | zcorpan, btw, how does opera make its $$? |
| 11:07 | <Ms2ger> | Bah |
| 11:07 | <Ms2ger> | "Duplicate test name: FormData interface: operation append" |
| 11:07 | <Ms2ger> | "Cached metdata not present. Click for source code." |
| 11:07 | <Ms2ger> | Not helpful, plinss |
| 11:08 | <Philip`> | sm0lm: By selling itself to Facebook? |
| 11:08 | <sm0lm> | Philip`, s/\?$/ until now?/ |
| 11:09 | <Ms2ger> | Selling its products to Nintendo? |
| 11:09 | <zcorpan> | sm0lm: http://my.opera.com/chooseopera/blog/2011/01/03/how-does-opera-make-money-aka-our-most-asked-question-ever |
| 11:11 | <sm0lm> | zcorpan, So..., how does that revenue model conflict with opening the sources? |
| 11:12 | <sm0lm> | or i mean what's the reasoning for keeping it closed? |
| 11:12 | <Philip`> | How would the revenue model be enhanced by opening the source? |
| 11:13 | <Ms2ger> | Philip`, three open source hippies would contribute bug fixes |
| 11:13 | <Ms2ger> | Not me, I hear they use git |
| 11:13 | <sm0lm> | Philip`, how would it be diminished by opening the source? |
| 11:13 | <sm0lm> | Ms2ger, do you work for any of the big vendors? |
| 11:14 | <Ms2ger> | Not currently |
| 11:14 | <Ms2ger> | As zcorpan mentioned, I'm a Mozillian |
| 11:14 | <jgraham> | Ms2ger: Three open source hippies? I thought RMS was the only open source hippie? The idea that there might be two others is terrifying |
| 11:14 | <sm0lm> | jgraham, wtf? |
| 11:14 | <hsivonen> | jgraham: how insensitive to call RMS an open source hippie. If he is a hippie, he is an Free Software hippie. |
| 11:15 | <sm0lm> | don't diss open source hippies. |
| 11:15 | sm0lm | makes one more. |
| 11:15 | <Ms2ger> | jgraham, there's been some guy hanging around irc.m.o working on an extension to block JS scripts without a license header |
| 11:15 | <Ms2ger> | I think that makes two, at least |
| 11:15 | sm0lm | laughs out loud at hsivonen 's comment. |
| 11:15 | <Philip`> | sm0lm: I imagine there would be e.g. conflicts between having an open bug database and source repository, and having Opera developers work on patches for unannounced embedded devices which have to be kept confidential |
| 11:15 | <sm0lm> | Ms2ger, that actually is really dumb. |
| 11:15 | <jgraham> | hsivonen: Egg and my face are in alignment |
| 11:16 | <Philip`> | sm0lm: (Presumably not unsolvable, but probably not entirely trivial) |
| 11:16 | <sm0lm> | Philip`, Presumably? Seems like "clearly" to me. |
| 11:16 | <sm0lm> | Just separate your private clients' sensitive data and unannounced work into another repo/tracker... problem solved, no? |
| 11:17 | <Ms2ger> | "Just" |
| 11:17 | <jgraham> | No |
| 11:17 | <Ms2ger> | Anyone feel like implementing enum support in idlharness? |
| 11:17 | <jgraham> | Having multiple trackers makes it hard to move things between trackers, for example |
| 11:18 | <jgraham> | Ms2ger: you? |
| 11:18 | <Ms2ger> | Not particularly |
| 11:19 | <Ms2ger> | I'm interested in someone doing it, just not me :) |
| 11:19 | <sm0lm> | When netscape opened theirs they had a team of lawyers approving each src file as safe for public ation wit no vulgar or slanderous comments... so its a job, but its not that big of a deal. |
| 11:20 | <Ms2ger> | Of course, all those vulgar comments were released afterwards |
| 11:20 | <jgraham> | sm0lm: I imagine that the biggest impediment to making Opera open source would be convincing people (the management, the board) that it is a good idea :) |
| 11:21 | <sm0lm> | Ms2ger, ofc they were. |
| 11:21 | <sm0lm> | on jwz.org i believe. |
| 11:22 | <jgraham> | (so discussing technical difficulties seems rather silly) |
| 11:24 | <Ms2ger> | Ohai, it's graememcc |
| 11:26 | graememcc | discovers Ms2ger is everywhere |
| 11:27 | <Ms2ger> | So are you! :) |
| 13:51 | <annevk> | so to properly resolve URLs you need a hardcoded list |
| 13:51 | <annevk> | teehee |
| 13:52 | <annevk> | of schemes |
| 13:53 | <annevk> | in other news, [tm] we got http://www.w3.org/TR/notifications/ published |
| 13:53 | <annevk> | [tm]: I guess that means we get a charter extension |
| 13:54 | <zewt> | cool, another TR for people to trip over :P |
| 13:55 | <jgraham> | It's OK in this case progress is so slow that TR will only be a little bit totally misleading at any given time |
| 14:03 | <annevk> | zewt: heh /TR/ash /TR/ip |
| 14:03 | <annevk> | jgraham: it has been pretty bad unfortunately |
| 14:04 | <annevk> | jgraham: but now I have granted myself full editing power (muhahaha) it should go more smoothly |
| 14:04 | <jgraham> | annevk: Well it is partially my fault, so I can't really complain |
| 14:05 | <mhausenblas> | annevk congrats re the Notifications TR, now a quick question: are you aware of http://tools.ietf.org/html/draft-ietf-core-observe-05 …? |
| 14:05 | <annevk> | mhausenblas: I'm not |
| 14:06 | <annevk> | mhausenblas: is it somehow related to end-user notifications? |
| 14:06 | <mhausenblas> | reason is: I'm getting more and more involved in the IoT stuff (see also http://scriptogr.am/mhausenblas/post/coap-and-the-rest) and interested in how aligned these things are |
| 14:07 | <mhausenblas> | that's what I'm trying to figure ;) |
| 14:07 | <mhausenblas> | on the surface, I'd say no, but need to dig deeper re overlaps |
| 14:07 | <annevk> | mhausenblas: if we take Mac as an example all this does is give a web page the ability to ping Growl |
| 14:07 | <mhausenblas> | right |
| 14:07 | <annevk> | mhausenblas: or the native notification system coming in 10.8 |
| 14:08 | <annevk> | mhausenblas: not much to do with HTTP APIs |
| 14:08 | <mhausenblas> | yup, understood |
| 14:14 | <annevk> | Lachy: how about referencing DOM4? |
| 14:14 | <Lachy> | I do in Selectors API 2 |
| 14:15 | <Lachy> | But I left it as DOM3Core in v1 because of W3C bureaucracy, which requires normative references to stable specs |
| 14:15 | <annevk> | Lachy: also, I guess you can list me as former editor or in the acknowledgments somehow |
| 14:15 | <annevk> | Lachy: stick it to the man |
| 14:17 | <Lachy> | annevk, I'm hoping I can just let v1 die and v2 supersede it, at which point it doesn't really matter what v1 references. But it's more unnecessary work to change it to reference DOM4 in v1 that will likely have to be reverted anyway if it progresses to PR and REC |
| 14:28 | <annevk> | oh right |
| 14:28 | <annevk> | mike was not going to be around |
| 14:45 | <annevk> | so HTML has hierarchical URL and authority-based URL |
| 14:45 | <annevk> | is there any example of an hierarchical URL that is not authority-based? |
| 14:57 | <annevk> | So Safari (not Chrome) is the only browser that aside from having a magical list to go down the authority-based URL route also considers // after the scheme |
| 16:03 | <SimonSapin> | python -m weasyprint.browser http://i.imgur.com/KQy8R.png |
| 16:19 | <Lachy> | Hixie, yt? |
| 16:26 | <Lachy> | Hixie, I'm being assigned to more spec work now. Do you have any suggestions for specs that need editors, or does anyone else have/know of any specs that are in need of some work? |
| 16:27 | <Lachy> | I've looked at http://wiki.whatwg.org/wiki/Specs_todo already. Still considering options though |
| 16:31 | <Ms2ger> | 66767 INFO Passed: 61610 |
| 16:31 | <Ms2ger> | 66768 INFO Failed: 0 |
| 16:31 | <Ms2ger> | 66769 INFO Todo: 4999 |
| 16:31 | Ms2ger | approves of test suites |
| 16:32 | <jwalden> | bah, who needs them, all they create is more work |
| 16:35 | <Ms2ger> | Lachy, table layout ;) |
| 16:45 | <jonlee> | annevk: ping |
| 16:52 | <annevk> | jonlee|afk: I'm around |
| 16:53 | <annevk> | jonlee|afk: well somewhat, helping Lachy with dinner a bit |
| 17:02 | <annevk> | ah you emailed |
| 17:07 | <Lachy> | Ms2ger, yeah, that's one I thought of already. It's a possibility. |
| 17:09 | <TabAtkins_> | Man, who wrote that page? When was "CSS Gradients" ever in need of editors? |
| 17:09 | <TabAtkins_> | Animations is out of data now, too - we finally got active editors for it again. |
| 17:09 | <TabAtkins_> | I'll fix. |
| 17:11 | <TabAtkins_> | Okay, I can blame Hixie for that. |
| 17:12 | <annevk> | there's more than table layout to fix though |
| 17:12 | <annevk> | e.g. imperative model of box construction |
| 17:12 | <TabAtkins_> | Sure, put it in. |
| 17:13 | <TabAtkins_> | Or I guess I will. |
| 17:21 | <annevk> | ah yeah, hit testing |
| 17:21 | <annevk> | there's a lot of stuff |
| 17:22 | <Ms2ger> | Oh hey, D3E published |
| 17:22 | <annevk> | but I keep forgetting it |
| 17:22 | <TabAtkins_> | annevk: A few days ago when you last mentioned it, you disappeared before I could ask what you actually meant by "hit testing is underdefined". Can you elaborate? |
| 17:23 | <annevk> | it's not defined which element is returned for a given x,y within a document |
| 17:23 | <TabAtkins_> | Okay. Is it really as trivial as I think, where you just take the geometry of all the elements in the documents (transformed as appropriate), order them by z-order, then test them for containment and take the first? |
| 17:24 | <annevk> | well, you have to go through margin boxes, fully transparent elements, maybe semi-transparent elements? and such |
| 17:24 | <annevk> | Leif Arne had a pretty good write up at one point in 2010 |
| 17:24 | <annevk> | then tantek was going to specify it |
| 17:24 | <annevk> | and it ended there... |
| 17:25 | <TabAtkins_> | Last I saw, it's not clear that transparent elements actually need to have special handling. |
| 17:25 | <jamesr_> | IE does |
| 17:25 | <tantek> | yeah - it is sitting in my CSS tasks queue text document :/ |
| 17:25 | <jamesr_> | in bizarre ways |
| 17:25 | <jamesr_> | IE handles transparent elements differently, that is |
| 17:25 | <annevk> | yeah, the write-up was about IE |
| 17:25 | <TabAtkins_> | Yeah, I know it does something kinda weird. |
| 17:25 | <tantek> | I was going to put it into CSS3-UI (since hit-testing *does* have to do with UI) |
| 17:25 | <jamesr_> | i think they be on crack |
| 17:26 | <annevk> | but once you implement pointer-events you need to do similar things |
| 17:26 | <TabAtkins_> | But based on my reading of the Firefox bug, it's not clear there's a web-compat concern. |
| 17:26 | <annevk> | this hit testing algorithm is influenced by the pointer-events property |
| 17:26 | <tantek> | but decided that putting in that much new spec text / functionality in a last call was probably a bad idea and so dumped it along with the 'pointer-events' property |
| 17:26 | <TabAtkins_> | Yeah, in a relatively trivial way. |
| 17:26 | <tantek> | right |
| 17:26 | <tantek> | TabAtkins - there is nothing trivial about either hit-testing, nor pointer-events, nor their relation. |
| 17:26 | <annevk> | tantek: it's a shame really, the WG should have let Leif Arne define it as I suggested |
| 17:26 | <annevk> | instead of letting it die in CSS3 UI |
| 17:27 | <TabAtkins_> | tantek: I'm not really seeing the complication. :/ |
| 17:27 | <annevk> | we could have had something by now |
| 17:27 | <tantek> | annevk - sorry, that's probably my fault. |
| 17:27 | <TabAtkins_> | At least, from a conceptual standpoint. Efficient implementation is a different matter. |
| 17:27 | <tantek> | and more like punted/postponed from CSS3 UI rather than "died" |
| 17:27 | <tantek> | though I suppose it's hard to tell the difference until something new is published |
| 17:28 | <tantek> | annevk - the last time we discussed h-t/p-e was in the Seattle CSS WG f2f that you were in |
| 17:28 | <jamesr_> | you basically run this: http://www.w3.org/TR/CSS21/zindex.html and then figure out bounds for everything, no? |
| 17:28 | <tantek> | which I believe was your last CSS WG f2f |
| 17:28 | <jamesr_> | figuring out exact bounds for everything is interesting |
| 17:28 | <annevk> | that was a nice meeting |
| 17:28 | <tantek> | jamesr_ - odd things happen with background, borders, overflow, scrolling etc. |
| 17:28 | <annevk> | I like Seattle |
| 17:28 | <TabAtkins_> | I think exact bounds is kinda underdefined for inline layout, but well-defined for everything else. |
| 17:29 | <tantek> | annevk - it was nice up until the point where glazman berated you - but I suppose you have blocked that out ;) |
| 17:29 | <TabAtkins_> | (Underdefined because it's basically "whatever the hell your text layout engine spits out".) |
| 17:29 | <annevk> | tantek: guess so - just fond memories of hanging out with Alex :) |
| 17:30 | <tantek> | I remember something about a boat |
| 17:30 | <tantek> | and water |
| 17:30 | <tantek> | though that was secondhand |
| 17:30 | tantek | knows better than getting on a boat during CSSWG meetings. |
| 17:30 | <annevk> | also a wet iPhone :p |
| 17:30 | <annevk> | haha |
| 17:31 | <TabAtkins_> | Drunken boating was a lot of fun. |
| 17:31 | <TabAtkins_> | Mainly because dbaron took over immediately and barely drinks. ^_^ |
| 18:40 | <dbaron> | TabAtkins_, um, I took over because I seemed to be the only one who realized that the appropriate way to steer those boats was to turn the wheel about a quarter turn (and then wait ten seconds to see how the boat responds) rather than turning the wheel six times and then being surprised that you end up going in a tight circle |
| 18:45 | <Hixie> | o_O |
| 18:49 | <Ms2ger> | dbaron, having forgotten what TabAtkins_ said last, I thought that was a reference to css3-transitions :) |
| 19:14 | <matjas> | Hixie: http://www.whatwg.org/specs/web-apps/current-work/multipage/entities.json is incorrect, Unicode escape sequences must have exactly 4 hex digits (not 5) |
| 19:15 | <matjas> | Hixie: what you want is basically a simplified version of http://mathias.html5.org/tests/html/named-character-references/data.json — I could quickly generate it for you if needed |
| 19:17 | <TabAtkins_> | I *also* thought that dbaron was talking about Animations. ^_^ |
| 19:20 | <Ms2ger> | http://www.w3.org/community/blog/2012/06/14/proposed-group-html5-mobile-test-community-group/ |
| 19:20 | Ms2ger | inserts Condescending Wonka |
| 19:23 | <matjas> | Hixie: here you go http://mathias.html5.org/tests/html/named-character-references/hixie.json |
| 19:29 | matjas | files https://www.w3.org/Bugs/Public/show_bug.cgi?id=17490 while Hixie’s afk |
| 19:40 | <Hixie> | matjas: how do i give astral plane characters then? |
| 19:40 | <matjas> | Hixie: surrogate pairs, as I’ve done in the file |
| 19:41 | <Hixie> | seriously? |
| 19:41 | <matjas> | Hixie: more info here http://mathiasbynens.be/notes/javascript-encoding |
| 19:41 | <matjas> | (applies to JSON too) |
| 19:41 | <Hixie> | i know JS is broken but making JSON broken too seems dumb |
| 19:41 | matjas | shrugs |
| 19:41 | <Hixie> | i'll just output U+..... then |
| 19:42 | <matjas> | Hixie: you were also missing commas after each key/value pair |
| 19:42 | <Hixie> | oops, will fix |
| 19:42 | <Hixie> | (i can't just use your version, this file is autogenerated from a python script) |
| 19:43 | <matjas> | is that python script public? i’d love to make it spit out surrogate pairs |
| 19:43 | <Hixie> | it can be :-) |
| 19:44 | <matjas> | Hixie: make it so and i’ll fix it up for you |
| 19:46 | <Hixie> | http://damowmow.com/temp/entity-processor-json.txt is what i have at the moment |
| 19:46 | <Hixie> | it's mid-way through a fix |
| 19:46 | <Hixie> | i'm changing it to output an array of unicode codepoints |
| 19:46 | <Hixie> | but currently it still outputs zero-padded numbers |
| 19:47 | <matjas> | Hixie: in a nutshell… https://github.com/mathiasbynens/unicode-data/blob/447d59809cb7d78a6113389176ff6dc15942557e/scripts/utils.py#L8-13 |
| 19:47 | <Hixie> | how do i convert a string to an integer in python? |
| 19:47 | <Hixie> | i don't think i really want the surrogates in the output |
| 19:48 | <Hixie> | surrogates should only ever exist in UTF-16, using them elsewhere is just bogus |
| 19:48 | <Ms2ger> | int() |
| 19:49 | <Hixie> | aha, int(..., 10) |
| 19:49 | <matjas> | Hixie: basically http://pastie.org/pastes/4088113/text and then call codePointToString() on the numeric code points |
| 19:50 | <matjas> | Hixie: yeah, but if you want to offer JSON this is the only way, other than using the raw glyphs |
| 19:50 | <Hixie> | there are many other ways :-) |
| 19:51 | <matjas> | Hixie: i don’t think having the numerical code points in JSON is nearly as useful |
| 19:51 | <Hixie> | e.g. what i'm doing, which is having the numerical codes |
| 19:52 | <Hixie> | it's unclear to me why it would be less useful |
| 19:52 | <Hixie> | it's the same data |
| 19:52 | <Hixie> | it's more useful, imho, since you don't have to worry about the surrogate nonsense |
| 19:52 | <matjas> | the numerical code points you can get from the HTML version |
| 19:52 | <Hixie> | what HTML version? |
| 19:52 | <Ms2ger> | The spec |
| 19:52 | <matjas> | of the spec |
| 19:53 | <Hixie> | there's no machine-readable version of this unless you mean the DTD data: URL... |
| 19:53 | <matjas> | the raw characters not, because of the undiscovered bug in the python script |
| 19:54 | <matjas> | …or you could please everyone by taking the http://mathias.html5.org/tests/html/named-character-references/data.json route + adding a `codepoint` property with the numerical value there |
| 19:54 | <matjas> | hah |
| 19:55 | <Hixie> | could have both i guess, true |
| 19:59 | matjas | refreshes http://www.whatwg.org/specs/web-apps/current-work/multipage/entities.json |
| 20:02 | <matjas> | heads up: so far it only lists 2098 out of 2231 named entities |
| 20:02 | <Hixie> | yeah, trying to work out why |
| 20:02 | <Hixie> | specifically, the surrogate pairs are missing |
| 20:03 | <Hixie> | check the damowmow link again |
| 20:03 | <Hixie> | see anything wrong? |
| 20:06 | <Hixie> | oh |
| 20:06 | <Hixie> | forgot to import math |
| 20:06 | <Ms2ger> | Math is everywhere! |
| 20:06 | <Hixie> | god i hate late-binding languages |
| 20:06 | <matjas> | heh |
| 20:09 | <matjas> | in `codepoints` you’re listing the surrogate halves separately, rather than the astral code point — i had expected the numerical astral code point value there |
| 20:10 | <matjas> | now it is indeed the same data as the string |
| 20:10 | <Hixie> | hm? |
| 20:10 | <Hixie> | "𝔄": { codepoints: [0x1D504], characters: "\uD835\uDD04" }, |
| 20:10 | <Hixie> | seems to be what i expect? |
| 20:10 | <matjas> | Hixie: ah, you’re right, of course |
| 20:11 | <Hixie> | no "of course" there, given how i've been screwing this up so far today :-) |
| 20:11 | <matjas> | i was looking at an entity that results in two symbols |
| 20:11 | <Hixie> | heh |
| 20:11 | <matjas> | nice work! |
| 20:11 | <Hixie> | thanks for the help! |
| 20:11 | <matjas> | you may want to get rid of that trailing comma at the end |
| 20:12 | <matjas> | after the last key/value pair |
| 20:12 | <Hixie> | please don't tell me json can't handle it |
| 20:12 | <Ms2ger> | Nope |
| 20:12 | <Ms2ger> | Syntax error |
| 20:12 | <Hixie> | are you kidding me |
| 20:12 | <Hixie> | jesus |
| 20:12 | <matjas> | Hixie: append each line to a list `buffer`, then ',\n'.join(buffer) |
| 20:13 | <matjas> | instead of `print |
| 20:13 | <Ms2ger> | Or, well, use the builtin json module |
| 20:13 | <Hixie> | the python generates only part of the file, the whole thing is later piped through 'sort' in a bash script |
| 20:13 | <Hixie> | but it's ok! |
| 20:13 | <Hixie> | i shall use a perl script to remove the comma |
| 20:13 | <Hixie> | for surely such a simply task can only be complete once it uses perl, python, and bash |
| 20:13 | Ms2ger | approves |
| 20:14 | <Ms2ger> | Note that the json module has an option to sort the keys, btw |
| 20:14 | <matjas> | sed -i '' -e '$s/,$//' entitites.json |
| 20:15 | <matjas> | Hixie: throw some VBScript in there and you’re golden |
| 20:17 | <matjas> | Hixie: also, the keys (`codepoints`, `characters`) need to be quoted, using double quotes |
| 20:17 | <Ms2ger> | Indeed so |
| 20:20 | <Hixie> | i am starting to hate json :-P |
| 20:23 | <Hixie> | in other news: |
| 20:23 | <Hixie> | <template> and company. |
| 20:23 | <Hixie> | does anyone want the HTML spec updated for this other than google? |
| 20:24 | <Hixie> | hsivonen: yt? |
| 20:24 | <Hixie> | anyone from opera or webkit? |
| 20:24 | <Hixie> | (or IE?) |
| 20:26 | <TabAtkins_> | What's up? |
| 20:26 | <TabAtkins_> | Oh, webkit other than google. |
| 20:27 | <TabAtkins_> | Moz and IE are getting pretty interested in Components, but I can't speak for them. |
| 20:36 | <Hixie> | TabAtkins_: i'm trying to work out if they're "pretty interested" the same way they were in xbl2, or if they actually want me to update the spec to support it... |
| 20:37 | <TabAtkins_> | Ask them yourself, but I'm pretty sure it's a stronger commitment than xbl2. |
| 20:38 | <Hixie> | TabAtkins_: that's whyat i am asking them :-) |
| 20:38 | <TabAtkins_> | That's why I said I can't speak for them. ^_^ |
| 20:40 | <Hixie> | right, hence why i didn't ask you :-P |
| 20:40 | <TabAtkins_> | FINE |
| 20:42 | <Hixie> | does <template> have a corresponding insertion point element? i can't find it in http://dvcs.w3.org/hg/webcomponents/raw-file/tip/spec/templates/index.html |
| 20:43 | <Hixie> | aha, <content> http://dvcs.w3.org/hg/webcomponents/raw-file/tip/spec/shadow/index.html#content-element |
