| 00:00 | <TabAtkins> | annevk: Yo, one more detail. JS Maps now have the set() method return the map, for chaining. They don't currently return the map for delete(), but we're planning to fix that, so URLQuery should match and return the object for both. |
| 00:00 | <Hixie> | i thought chrome implemented <style scoped>. did support get removed or something? |
| 00:00 | <TabAtkins> | I forget, and Dimitri just left. |
| 00:02 | <TabAtkins> | annevk: And append() too, obviously. |
| 00:02 | <TabAtkins> | annevk: That one's easy, so you can just do u.append(k,v).append(k,v2).append(k.v3); |
| 00:02 | <TabAtkins> | But also having the ability to do u.delete(k).append(k,v) is useful. |
| 00:18 | GPHemsley | can't believe W3C doesn't have a mirror system set up to protect against MIT power outages. |
| 00:20 | gavinc | points a very large number of sites that didn't have anything other then their bellow ground datacenter in NYC |
| 00:21 | <TabAtkins> | GPHemsley: On that note, they're back up. |
| 00:22 | <gavinc> | 11/29/12 04:12 PM by IRC at least ;) |
| 00:22 | <GPHemsley> | gavinc: Sites that are crucial to the workings of the Web? (I didn't notice any sites go down.) |
| 00:23 | <gavinc> | GPHemsley: not really "the web" but CBS |
| 00:23 | <GPHemsley> | oh, I don't expect news companies to have that level of redundancy |
| 00:23 | <GPHemsley> | I mean, they should |
| 00:23 | <GPHemsley> | but I don't expect it |
| 00:23 | <GPHemsley> | IT companies, though... |
| 00:24 | <gavinc> | yeah I don't think so |
| 00:25 | <GPHemsley> | for maximum point effectiveness, I'll even limit that to just the W3C |
| 00:26 | <GPHemsley> | I won't presume anything about any other entity |
| 00:26 | <gavinc> | Eh, I remember (was somewhat involved with) turning off a rack at PAIX that ended badly for the West Coast for a few hours, and the overpass fire a few years later that cut most of the fiber in the bay area |
| 00:27 | <gavinc> | disasters are surprising ;) |
| 00:43 | <TabAtkins> | Argh, goddammit, I *really* need to write my own CSS spec preprocessor. >_< Bert's has weird ideas of how to fix up markup it considers invalid. |
| 00:57 | <dbaron> | TabAtkins, can you use one of the other available ones instead? |
| 01:06 | <TabAtkins> | dbaron: I dunno. I use some of Bert's features heavily (namely, the magic around 'foo' and ''foo'', and the text-based linkage of <dfn> and <i>). |
| 01:06 | <TabAtkins> | If something else can reproduce those, or is easy enough to hack them into, I could do so. |
| 01:06 | <TabAtkins> | Oh, and the biblio syntax [[foo]] and [[!foo]]. |
| 01:10 | <tantek> | TabAtkins, that's the one where [[!foo]] means normative (important) foo, right? |
| 01:11 | tantek | wonders if it would have been too hard to just do [[foo!]] instead (or perhaps patch the preprocessor to accept that too) |
| 01:29 | <TabAtkins> | tantek: Yes, it is. |
| 01:30 | <TabAtkins> | tantek: What's the point of moving the !? |
| 01:31 | <Hixie> | ojan_away: yeah the only difference with cross-origin workers vs normal workers is that they're in a different origin |
| 01:44 | <tantek> | TabAtkins, because a trailing ! looks like a shout (of importance), whereas a leading ! looks like an inverter (not). |
| 01:51 | <dbaron> | not in CSS :-P |
| 04:38 | <Hixie> | is there anyone with a bunch of patience who would happens to have some version of IE around who could describe to me the sequence of alerts that IE gets while loading http://damowmow.com/playground/demos/onload/001.html ? |
| 04:57 | <Hixie> | wtf |
| 04:57 | <Hixie> | my mac no longer responds to "f3" |
| 04:58 | <Hixie> | like, the keyboard driven isn't letting that through |
| 04:58 | <Hixie> | the key is physically fine, i can rebind it to "=" and then hitting f3 prints an "=" |
| 04:58 | <Hixie> | and i can rebind "=" to f3, and then "=" does nothing. |
| 04:58 | <Hixie> | (yay programmable keyboards0 |
| 04:58 | <Hixie> | ) |
| 04:58 | <Hixie> | i've tried rebooting my keyboard |
| 04:58 | <Hixie> | guess rebooting the mac is next |
| 08:35 | <annevk> | TabAtkins: ah, so the chaining won? interesting |
| 08:35 | <annevk> | TabAtkins: I guess I'll add that once I receive word from TC39 |
| 09:07 | <annevk> | So URLQuery... When a URLUtils implementing object is created it creates a URLQuery and seeds it with it's url's query. Then whenever url's query is updated (either .href or .search) URLQuery is updated. URLQuery in turn updates the URLUtils implementing object's url's query. |
| 09:08 | <annevk> | DOMTokenList is going to work pretty much the same way, for those interested, and all other objects we're going to introduce (and have introduced, thinking of .style) that need to keep sync with some other object. |
| 09:10 | <annevk> | Can someone one-up Google Translate for me? https://twitter.com/kazuhito/status/274437377290817536 |
| 09:12 | <darobin> | annevk: maybe MikeSmith? |
| 09:13 | <MikeSmith> | the last part just basically means, "I'm part of the WHATWG." |
| 09:14 | <MikeSmith> | well I assume that's the intended meaning |
| 09:14 | <MikeSmith> | literally it means "I am the WHATWG." |
| 09:14 | <MikeSmith> | the first part I have no idea |
| 09:14 | <MikeSmith> | it's some kind of idiom |
| 09:14 | <MikeSmith> | of which there are a lot in Japanese |
| 09:15 | <MikeSmith> | or at least it's not an expression I'm familiar with |
| 09:15 | <annevk> | yeah, Google says 'Tan ー "I is the WHATWG!!" Hikushi' which is not that great |
| 09:16 | <annevk> | was wondering whether it was something funny that could be retweeted from @WHATWG |
| 09:16 | <MikeSmith> | yeah |
| 09:17 | <MikeSmith> | that would be funny |
| 09:17 | <MikeSmith> | I have no idea what the context is for him having tweeted that |
| 09:17 | <MikeSmith> | but I know him |
| 09:17 | <MikeSmith> | his office is right down that street from my place |
| 09:17 | <MikeSmith> | he's Yakura-san's boss |
| 09:18 | <MikeSmith> | ah |
| 09:18 | <MikeSmith> | Hixie |
| 09:18 | <annevk> | ah yeah, I met him; we went out to ate those beans :-) |
| 09:18 | <MikeSmith> | ひくし is Hixies |
| 09:18 | <MikeSmith> | *Hixie |
| 09:19 | <MikeSmith> | ひくしーたん is Hixie-san |
| 09:19 | <MikeSmith> | maybe he's quoting Hixie? |
| 09:19 | <MikeSmith> | Hixie saying, I am the WHATWG? |
| 09:20 | <darobin> | as in "L'État, c'est moi |
| 09:20 | <annevk> | that would make sense I guess, https://twitter.com/stevefaulkner/status/274427815242391553 |
| 09:20 | <darobin> | ? |
| 09:20 | <annevk> | darobin: yeah |
| 09:21 | <darobin> | it would indeed be fun for you to RT it from @WHATWG :) |
| 09:21 | <MikeSmith> | annevk: ah, OK, yeah |
| 09:21 | <MikeSmith> | well, it's still funny to retweet it |
| 09:21 | <annevk> | darobin: but but but |
| 09:22 | <darobin> | do it do it do it! |
| 09:23 | <annevk> | okay why not, acknowledging dissent is good |
| 09:25 | <darobin> | I don't know if criticising the institution counts as dissent |
| 09:30 | <MikeSmith> | does criticizing Hixie amount to criticizing the institution? |
| 09:30 | <MikeSmith> | does criticizing the institution amount to criticizing Hixie? |
| 09:38 | <MikeSmith> | would be funny to follow it up by re-tweeting https://twitter.com/BrendanEich/status/273715922399805441 |
| 09:38 | <darobin> | MikeSmith: I think that's the question that our Japanese friend is putting to the test |
| 09:39 | <MikeSmith> | yeah |
| 09:43 | <volkmar> | darobin: is there a way to declare an interface with respec without having the attributes + methods definition showing up? |
| 09:43 | <darobin> | volkmar: mmmm, maybe |
| 09:43 | <darobin> | display: none? |
| 09:43 | <darobin> | :) |
| 09:44 | <darobin> | give me a spare afternoon and I'll finish the new IDL support that's a lot nicer... |
| 09:48 | <volkmar> | darobin: I guess I could have a look at the generated HTML |
| 09:49 | <darobin> | if you feel so inclined, be my guest |
| 09:49 | <darobin> | I will be doing some maintenance to integrate patches and stuff on Monday |
| 09:49 | <volkmar> | darobin: oh, I will just have a look at what is generated to hide the part I don't want (if I can) |
| 09:55 | <marcosc> | in anolis... ;) |
| 09:55 | <marcosc> | just sayin' |
| 09:55 | <marcosc> | :) |
| 09:59 | <darobin> | or like marcosc says you could just write the spec by hand, no features -> no problems! |
| 09:59 | marcosc | stops trollin darobin :) |
| 09:59 | <darobin> | just sayin' |
| 10:00 | <darobin> | reminds me of when Lachy was bitching about how it's a pain to generate the W3C boilerplate |
| 10:00 | marcosc | just bitter about <dl> IDL still ... should write a patch for that. |
| 10:05 | <annevk> | might have used the wrong word, indeed |
| 10:18 | <annevk> | if we need to introduce a new XHR constructor for e.g. promises, we could call it Fetch maybe |
| 10:18 | <annevk> | and make Fetch define all the things |
| 10:28 | <annevk> | my GitHub client just fixed a bug where they didn't handle a space in the repository path, I guess it counts as a browser then |
| 11:29 | <annevk> | too late for s/ate/eat/? |
| 12:29 | <annevk> | I think I finally hit that point I've been working towards for the last year and a half or so. That where URLs, Encoding, and <form> submission collide. |
| 13:12 | <odinho> | annevk: That'd be a place for sugar api, doing these fetch things in a nicer way :] |
| 13:12 | <annevk> | yeah that too |
| 13:14 | <annevk> | (new Fetch({url:"lalaland"})).then(teehee, damnit) |
| 13:14 | <annevk> | (default for method would be GET) |
| 13:15 | <Ms2ger> | Oh dear, another API |
| 13:18 | <annevk> | dunno, just sketching |
| 13:40 | <volkmar> | annevk: Fetch could inherits from DOMRequest ^^ |
| 13:40 | <annevk> | what is DOMRequest? |
| 13:40 | <volkmar> | hmmm |
| 13:41 | <volkmar> | annevk: https://mxr.mozilla.org/mozilla-central/source/dom/base/nsIDOMDOMRequest.idl?force=1 |
| 13:41 | <volkmar> | just look at the first interface |
| 13:41 | <volkmar> | we use that a lot in our new Web APIs |
| 13:41 | <volkmar> | and with sicking, we are thinking of extending the API to make it promise-like |
| 13:41 | <volkmar> | it would have some advantages |
| 13:42 | <annevk> | volkmar: ah yeah, Fetch would work with whatever we end up doing for Promises |
| 13:43 | <annevk> | it seems everyone is designing a promise system |
| 13:43 | <annevk> | but not on any mailing lists |
| 13:44 | <odinho> | cue tantek? |
| 13:45 | <Ms2ger> | Email is a support forum |
| 13:46 | <volkmar> | annevk: we are not really designing anything |
| 13:46 | <volkmar> | annevk: I mean, except DOMRequest |
| 13:46 | <annevk> | looks like a design of sorts to me :) |
| 13:46 | <volkmar> | but it's hard to find a place for this little thing to live in |
| 13:46 | <volkmar> | annevk: I asked you some time ago if it could be added to DOMCore (or Ms2ger, maybe) |
| 13:46 | <volkmar> | and I got a no |
| 13:47 | <volkmar> | a full spec for it would be weird I believe |
| 13:47 | <Ms2ger> | Oh? |
| 13:47 | <volkmar> | Bas? |
| 13:47 | <annevk> | I don't recall that question either |
| 13:47 | <volkmar> | I've been told that because it's using EventHandler it is depending on HTML5, thus can't live in DOMCore |
| 13:47 | <annevk> | that makes sense |
| 13:48 | <annevk> | event queueing makes for a HTML dependency too |
| 13:48 | <annevk> | so HTML might be a logical place |
| 13:48 | <annevk> | since any spec with async stuff will need to depend on HTML anyway |
| 13:48 | <volkmar> | why? |
| 13:48 | <annevk> | event loop |
| 13:48 | <volkmar> | doesn't make much sense |
| 13:48 | <volkmar> | to have event loop defined there |
| 13:52 | <annevk> | sure, but we haven't found anyone to do that work elsewhere |
| 14:33 | <annevk> | so URL now defines application/x-www-form-urlencoded |
| 14:34 | <SimonSapin> | annevk: \o/ |
| 14:37 | <annevk> | maybe I'm able to complete URLQuery today |
| 14:37 | <SimonSapin> | And another mega-email to show how Håkon and Bert are wrong about multicol. |
| 14:37 | <SimonSapin> | Sorry, I mean show how I disagree :) |
| 14:37 | <annevk> | SimonSapin: heh |
| 14:40 | <annevk> | btw, I think we should support constructing URLQuery from a string |
| 14:40 | <annevk> | if you're opposed, let me know |
| 14:40 | <annevk> | (you can seed it from a string already via .search so it only makes sense) |
| 14:47 | <zewt> | yeah, the "make a dummy URL just to make a URLQuery" thing is a pretty silly hack |
| 14:49 | <annevk> | well that was already gone |
| 14:49 | <annevk> | but shit is pretty complicated |
| 14:51 | <annevk> | the amount of encoding overrides going around in URL is also insane :/ |
| 14:59 | <SimonSapin> | annevk: how do I get a account to edit wiki.whatwg.org? |
| 15:00 | <annevk> | SimonSapin: pm me your email and desired username |
| 15:09 | <SimonSapin> | annevk: URLs in non-relative schemes are only made of [scheme, scheme data, query, fragment] right? |
| 15:11 | <annevk> | yes |
| 15:11 | <SimonSapin> | annevk: 5.1 Writing seems to confuse relative/absolute URLs (that need a base URL or not) with URLs in a relative/non-relative scheme |
| 15:12 | <SimonSapin> | "An absolute URL is a scheme, followed by ":", followed by scheme data, optionally followed by "?" and a query. " |
| 15:13 | <annevk> | scheme data is elaborated upon for relative schemes |
| 15:13 | <SimonSapin> | oh I missed that, sorry |
| 15:15 | <SimonSapin> | #url-units has a long list of code point ranges. What is excluded? control chars, %, lone surrogate pairs, private use, …? |
| 15:16 | <annevk> | non-characters |
| 15:16 | <annevk> | private use is in I think |
| 15:16 | <SimonSapin> | wouldn’t a "any code point NOT in the ranges …" definition be better? |
| 15:17 | <annevk> | you get about the same I believe |
| 15:17 | <annevk> | HTML has that |
| 15:17 | <annevk> | note that you don't have to deal with all those code units when parsing |
| 15:17 | <annevk> | because they'll be percent encoded |
| 15:18 | <SimonSapin> | my point is I’d like to know what is excluded and why |
| 15:19 | <SimonSapin> | I have no idea from memory why U+FDD0 is special |
| 15:29 | <annevk> | me neither |
| 15:29 | <annevk> | well actually, I believe Martin told me that's an application internal code point |
| 15:37 | <SimonSapin> | annevk: is an empty query different from no query once a URL is parsed? |
| 15:38 | <annevk> | yes |
| 15:39 | <annevk> | query == null -> "?" |
| 15:39 | <annevk> | euh oops |
| 15:39 | <annevk> | query == null -> ""; query == "" -> "?" |
| 15:39 | <SimonSapin> | ok |
| 15:39 | <annevk> | same for fragment |
| 15:47 | <SimonSapin> | hum, can I do nested lists in the wiki? |
| 15:48 | <SimonSapin> | ah, double stars |
| 15:55 | <hsivonen> | Whoa: http://stackoverflow.com/questions/13521554/xpath-in-internet-explorer-10-gone |
| 15:55 | <annevk> | the status field on http://wiki.whatwg.org/wiki/MetaExtensions should maybe be nuked, Hixie? it's kinda useless the way the tables are used; first is proposal, second is rejected proposals, third is crap |
| 15:58 | <hsivonen> | interesting solution: xhr.responseType = "msxml-document"; |
| 16:02 | <zcorpan> | doesn't ie10 support xpath in one of the quirkier modes? |
| 16:05 | <SimonSapin> | annevk: does URL or something else have a "percent decode to bytes" algorithm? I think it would be the same as "percent decode", but without the first and last step |
| 16:05 | <SimonSapin> | it would be used for data: |
| 16:07 | <annevk> | SimonSapin: no, but I can add that to URL |
| 16:07 | <annevk> | SimonSapin: but please file a bug, I'm in the middle of the URLQuery mess |
| 16:07 | <SimonSapin> | ok |
| 16:42 | <SimonSapin> | annevk: wrote up http://wiki.whatwg.org/wiki/URL_schemes#data:_URLs with inline issues |
| 16:44 | <annevk> | cool |
| 16:53 | <dglazkov> | good morning, Whatwg! |
| 16:53 | <dglazkov> | odinho: time to go home |
| 16:54 | <dglazkov> | and feed wife/kids/dog/turtle/evil robot |
| 16:55 | <annevk> | http://url.spec.whatwg.org/#interface-urlquery booyah |
| 16:56 | <annevk> | writing the API bit was surprisingly simple once all the string munging became external to it |
| 16:56 | <odinho> | dglazkov: Okay! :D Actually, friday beer, but I did say I'd be down there before all the chocolate was eaten, so I'll go do that now then. Cheers! |
| 16:56 | <annevk> | hmm chocolate |
| 17:07 | <marcosc> | annevk: it's that crappy Freia (Kraft Foods) chocolate, so don't worry ... you are not missing out :) |
| 17:08 | <annevk> | marcosc: dunno man, I like that stuff |
| 17:08 | <marcosc> | annevk: ok ... I like Heineken |
| 17:08 | <marcosc> | :) |
| 17:09 | <annevk> | your loss really |
| 17:09 | <annevk> | mb |
| 17:09 | <annevk> | bjt |
| 17:09 | <annevk> | wut, but I see your point |
| 18:00 | <GPHemsley> | annevk: I've got a present for you: http://wiki.akosma.com/IPhone_URL_Schemes |
| 18:10 | <GPHemsley> | gavinc_: 2 of the 3 Ws stand for "World Wide". If the world doesn't blow up, there should be a server somewhere that is capable of picking up the slack if something goes wrong somewhere. |
| 18:16 | <gavinc_> | GPHemsley: Fair point. |
| 18:18 | <GPHemsley> | gavinc_: And, frankly, the 3rd W comes into play here, too. What is the Web but a series of interconnected points? |
| 18:19 | <GPHemsley> | Put them all together, and what do you get? A worldwide series of interconnected points... that completely disappears if MIT loses power. |
| 18:21 | <annevk> | GPHemsley: yeah, those don't matter much fortunately |
| 18:21 | <GPHemsley> | are you not dealing with tel:? |
| 18:22 | <GPHemsley> | sms:? |
| 18:22 | <annevk> | don't need to |
| 18:22 | <GPHemsley> | why not? |
| 18:22 | <annevk> | because they're just launching some kind of action |
| 18:23 | <GPHemsley> | So... any scheme that doesn't deal directly with the Web can do what it wants? |
| 18:25 | <annevk> | within the constraints of the URL syntax that is kinda how it works, yes |
| 18:26 | <GPHemsley> | I see |
| 18:26 | <GPHemsley> | so you were only dealing with the issues with URLs only insofar as they affect the Web? |
| 18:26 | <annevk> | hmm, not really |
| 18:26 | <annevk> | URL parsing affects everyone |
| 18:27 | <annevk> | URL schemes I was only interested in those that would affect the upcoming Fetch rewrite |
| 18:27 | <GPHemsley> | what about a URL like this? twinkle://message=text&url=http://foo&image=http://path.to/image.jpg |
| 18:27 | <annevk> | there's a big difference between URL parsing and URL scheme-specific processing |
| 18:28 | <annevk> | gotta go |
| 18:28 | <GPHemsley> | how convenient :P |
| 18:37 | <TabAtkins> | annevk: Promises are being worked on in tc39. Ideally, DOM will wait on them finishing that work, or at least finishing enough of the API decisions that we can engrave them into the web. |
| 18:44 | <TabAtkins> | Based on what I can see at http://www.whatwg.org/specs/web-apps/current-work/multipage/elements.html#embedding-custom-non-visible-data-with-the-data-*-attributes , it looks like el.dataset doesnt' have any setting behavior. |
| 18:45 | <TabAtkins> | Specifically, if you pull el.dataset off the element, then try to assign it to another element, behavior isn't defined? |
| 18:45 | <TabAtkins> | Am I right, or missing something? |
| 18:45 | <Hixie> | it's readonly, if that's what you mean |
| 18:45 | <TabAtkins> | That is what I mean, yes. |
| 18:46 | <TabAtkins> | Any reason for that? |
| 18:46 | <Hixie> | behaviour is defined, it does whatever you do when you assign a readonly attribute a value |
| 18:46 | <Hixie> | generally we don't support assigning objects like that from one to another element, but no reason beyond that that i know of |
| 18:46 | <Hixie> | generally if you wanted to do this you'd just clone the node |
| 18:46 | <TabAtkins> | My friend estelle apparently has a use-case for doing what I described. |
| 18:47 | <Hixie> | tell them to mail the list describing it :-) |
| 18:50 | <Hixie> | sweet, http://software.hixie.ch/utilities/js/live-dom-viewer/saved/1944 reliably crashes chrome when you click the link |
| 18:51 | <Hixie> | safari, too |
| 19:17 | <Hixie> | hm, no darobin and no nessy |
| 19:17 | <Hixie> | anyone here follow the htmlwg commits closely enough to help me understand what patches have been committed and which haven't? |
| 19:19 | <Hixie> | r7534 seems to be in, but r7533 doesn't. I can't work out if that's intentional, if it's documented, or what. |
| 19:19 | <Hixie> | but there's like 15 branches and i've no idea what they are for |
| 19:19 | <Hixie> | so confused |
| 19:21 | <annevk> | TabAtkins: I don't really see how TC39 can work on them without having an event loop |
| 19:22 | <TabAtkins> | annevk: ??? |
| 19:22 | <TabAtkins> | annevk: (I know the context is promises, I just want more elaboration.) |
| 19:22 | <annevk> | Promises are for asynchronous handling. Nothing in ECMAScript has that. That's all defined by HTML. |
| 19:39 | <Hixie> | ah, excellent, darobin |
| 19:40 | <Hixie> | darobin: can you help me understand what patches have been committed to the htmlwg branch(es) and which haven't? specifically, i'm trying to work out if r7533 is in. r7534 seems to be. |
| 19:40 | <Hixie> | darobin: and if it's not, i'm trying to work out if this is documented anywhere. |
| 19:53 | <annevk> | so what's a good name for an IDL attribute that turns a DOMString into a code point sequence and does not make it attractive for API designers to plaster all over? |
| 19:53 | <Hixie> | what interface is it on? |
| 19:53 | <annevk> | I initially thought [Unicode] DOMString but that might be copied quite enthusiastically |
| 19:53 | <Hixie> | oh you mean a webidl annotation thingy? |
| 19:53 | <annevk> | Hixie: all over URL, Encoding, and some of XHR |
| 19:54 | <Hixie> | not an actual idl attribute |
| 19:54 | <Hixie> | ok |
| 19:54 | <Hixie> | um |
| 19:54 | <Hixie> | hmm |
| 19:54 | <annevk> | the other one I had was [UTF] DOMString |
| 19:54 | <Hixie> | wait, what exactly does this do? |
| 19:54 | <Hixie> | aren't all DOMStrings code point sequences? |
| 19:55 | <annevk> | Hixie: it does http://dev.w3.org/2006/webapi/WebIDL/#dfn-obtain-unicode but without invoking that algorithm all over |
| 19:56 | <Hixie> | this is on getting? on setting? |
| 19:56 | <annevk> | setting |
| 19:56 | <Hixie> | so on setting, it cleans up the DOMString so that it's valid Unicode, not just potentially-invalid UTF-16? |
| 19:56 | <Hixie> | hmm |
| 19:57 | <annevk> | yeah, because once you start doing utf-8 operations on it, you kinda want that |
| 19:57 | <Hixie> | [SanitizeUnicodeForTransmissionOnSetting] |
| 19:57 | <annevk> | and then I'll typedef it as UTFString |
| 19:57 | <annevk> | because that makes the IDL unreadable |
| 19:58 | <Hixie> | making the idl unreadable was the point, i thought |
| 19:58 | <Hixie> | you wanted people not to use it |
| 19:58 | <annevk> | yeah, but I also want developers to be able to look at the IDL normally, hmm |
| 19:58 | <annevk> | I guess one of those is not going to work |
| 19:59 | <Hixie> | the cost of the three developers whe make it to the spec having to squint a bit is probably lower than the cost of the 5 spec writers who see it and use it |
| 20:00 | <Hixie> | assuming you mean web devs, not UA devs |
| 20:00 | <MikeSmith> | Hixie: about the commits, I don't follow them closely, but as far as I can tell neither r7534 nor r7533 are in. https://github.com/w3c/html/commits/master The latest change pulled into that branch seems to be r7529 |
| 20:01 | <Hixie> | oh i got confused between ..34 and ..29 |
| 20:01 | <Hixie> | that explains it |
| 20:01 | <MikeSmith> | yeah |
| 20:01 | <MikeSmith> | "Clarify" vs "Clarity" |
| 20:02 | <Hixie> | and nearly the same patch |
| 20:02 | <Hixie> | ok so where's r7528? |
| 20:02 | <MikeSmith> | dunno but maybe Silvia did not pull it |
| 20:03 | <Hixie> | is that documented anywhere? |
| 20:03 | <Hixie> | it's not a controversial one, i mean, it's one she wanted... |
| 20:03 | <Hixie> | i'm so confused |
| 20:04 | <MikeSmith> | need to ask Silvia I guess :) |
| 20:04 | <MikeSmith> | there's basically no discussion ahead of time about what changes to pull |
| 20:04 | <MikeSmith> | Silvia sends a summary message every week listing what she's pulled |
| 20:05 | <MikeSmith> | afterwards |
| 20:05 | <Hixie> | yeah it's kinda ridiculous that the supposedly most public group has two lists so secret not even members can see them, and that's where all their decisions happen |
| 20:57 | <annevk> | Hixie: you use both "file name" and "filename" |
| 20:58 | <annevk> | (my problem is searching for instances) |
| 20:58 | <Hixie> | that's plausible |
| 20:58 | <Hixie> | file a bug if you can't find a pattern :-) |
| 20:58 | Hixie | figures out why his f3 key is acting weird |
| 20:58 | <Hixie> | i seem to have somehow bound it to the "rewind" key |
| 20:59 | <Hixie> | i wonder how and how to fix it |
| 21:00 | <annevk> | I first thought the pattern was plural, but it appears to be mood of the day instead |
| 21:01 | <Hixie> | nessy! maybe you can answer the question. i've been trying to work out if r7528 is in the htmlwg spec, and if not, if that's documented anywhere (r7529 seems to be in) |
| 21:05 | <annevk> | Hixie: btw http://url.spec.whatwg.org/#application/x-www-form-urlencoded |
| 21:18 | <Hixie> | annevk: does the encoding only take effect for values after the _charset_ is seen? and you can set multiple _charset_s? |
| 21:18 | <Hixie> | annevk: i would have thought the first _charset_ would affect all values |
| 21:19 | <annevk> | Hixie: thanks, will fix |
| 21:21 | <Hixie> | annevk: my algorithm had that right already dude :-P |
| 21:22 | <annevk> | Hixie: your algorithm didn't skip empty strings (e.g. ?test&&test) dude :-P |
| 21:22 | <annevk> | but yes, I should have copy-and-pasted-and-fixed better :) |
| 21:23 | <Hixie> | you also seem to have lost isindex support |
| 21:24 | <annevk> | yeah, I thought the idea was HTML would handle that? |
| 21:24 | <Hixie> | my algorithm seems to handle && fine |
| 21:24 | <Hixie> | in parsing? |
| 21:25 | <annevk> | hmm yeah, maybe for parsing it should be handled here |
| 21:25 | <Hixie> | also you don't define "splitting" |
| 21:26 | <Hixie> | interesting, your definition of "percent decode" does a lot more than just decoding %s |
| 21:27 | <TabAtkins> | Where is the handling of isindex actually explained? |
| 21:27 | TabAtkins | looked in the parser and in the obsolete features section. |
| 21:27 | <Hixie> | TabAtkins: in about five places in HTML, currently, just search for "isindex", you'll find them all |
| 21:27 | <TabAtkins> | Ugh, I've gotta load single-page then. ;_; |
| 21:27 | <Hixie> | yes |
| 21:28 | <Hixie> | google-provided hardware and software handle the single-page fine. :-P |
| 21:28 | <TabAtkins> | Yeah yeah, it still takes about 5s to load the page. |
| 21:28 | <Hixie> | i just leave it open |
| 21:29 | <Hixie> | annevk: for serialisation, we said i'd just keep the html spec's, right? |
| 21:29 | <annevk> | Hixie: so your preference is that && results in a pair whose name and value are both empty? |
| 21:29 | <SimonSapin> | loading single-page is one of the reasons I might start up chromium |
| 21:29 | <Hixie> | annevk: i don't really have a preference, but that's what the html spec does |
| 21:29 | <Hixie> | annevk: depends what problem we're trying to solve, really |
| 21:29 | <annevk> | Hixie: I think the idea was for serialization to reuse the generic part and fill in some specific bits |
| 21:30 | <annevk> | Hixie: so HTML would do some pre-processing on the name-value pairs it supplies |
| 21:33 | <Hixie> | annevk: your algorithm does encoding on a per-character basis |
| 21:33 | <Hixie> | annevk: that doesn't work for stateful encodings. |
| 21:34 | <annevk> | it does work, but it's not what impls do |
| 21:34 | <Hixie> | well it returns strings that are way longer than ideal |
| 21:35 | <Hixie> | also your algorithm sometimes results in strings that contain %20 instead of + |
| 21:35 | <Hixie> | (if the encoding encodes a character to a stateful string that contains 0x20) |
| 21:36 | <Hixie> | i'm starting to have the same concerns abarth mentioned the other day :-P |
| 21:36 | <annevk> | no kidding |
| 21:36 | <annevk> | but thanks for the feedback |
| 21:36 | <Hixie> | :-:) |
| 21:37 | <Hixie> | er |
| 21:37 | <Hixie> | :-) even |
| 21:37 | <Hixie> | right, back to working out poster frames |
| 22:03 | <SimonSapin> | "If you look at Microformats just right, it really is RDF." 2008, http://standardssuck.org/grddl-bridging-the-interwebs |
| 22:04 | <Hixie> | everything is rdf if you "look at it right", that's what the RDF people think makes RDF so powerful, and is what I think makes RDF so pointless :-) |
| 22:04 | <SimonSapin> | "GRDDL is the red pill for the web" |
| 22:06 | <TabAtkins> | MikeSmith: Could you add the Display module to the CSS component in bugzilla? |
| 22:07 | <MikeSmith> | sure |
| 22:07 | <MikeSmith> | TabAtkins: put you at the default assignee? Or somebody else? |
| 22:10 | <stevefaulkner> | MikeSmith: pinged u about some advice |
| 22:10 | <MikeSmith> | yup |
| 22:11 | <stevefaulkner> | did you get my mail? |
| 22:12 | <TabAtkins> | MikeSmith: Me, please. |
| 22:17 | <Hixie> | jgraham: yt? |
| 22:18 | <TabAtkins> | Hixie: Note that you don't need to report the intrinsic width/height as 300/150. The sizing algorithm in CSS fills that in automatically. |
| 22:18 | <Hixie> | ah, cool. you got a link to that text? i'll file a bug on me to remove any mentions of those dimensions where possible |
| 22:19 | <TabAtkins> | http://dev.w3.org/csswg/css3-images/#default-sizing |
| 22:19 | <Hixie> | ta |
| 22:26 | <Hixie> | gecko spins the event loop during the onbeforeunload prompt? what? |
| 22:26 | <Hixie> | meanwhile opera just ignores my onbeforeunload entirely |
| 22:26 | <Hixie> | ok. zero interop. |
| 22:26 | <Hixie> | guess i can do what i want then! |
| 22:27 | <Hixie> | -_- |
| 22:43 | <Hixie> | ok i need a test where the page takes a long time to load, has an onload handler, and a link. |
| 22:43 | <Hixie> | should be easy enough... |
| 22:56 | <TabAtkins> | annevk: What do you think about a two-arg has(), same as delete()? |
| 22:57 | <Hixie> | ok seriously |
| 22:57 | <Hixie> | how do i dispable compression on apache via .htaccess files |
| 22:57 | <Hixie> | disable, even |
| 22:58 | <Hixie> | oh wait, maybe this isn't compression that's screwing me |
| 22:58 | <TabAtkins> | RewriteRule . - [E=no-gzip:1] |
| 22:58 | <TabAtkins> | First result in google for me? |
| 22:58 | <Hixie> | yeah i have that already |
| 22:58 | <TabAtkins> | http://stackoverflow.com/questions/3414243/how-to-disable-compression-in-apache-on-some-response-header-from-mod-jk-proxied |
| 22:59 | <TabAtkins> | First answer, that line didn't work for him either. |
| 22:59 | <Hixie> | also tried that |
| 22:59 | <TabAtkins> | unsetting Accept-Encoding? |
| 22:59 | <Hixie> | also tried: |
| 22:59 | <Hixie> | <IfModule mod_gzip.c> |
| 22:59 | <Hixie> | mod_gzip_on No |
| 22:59 | <Hixie> | </IfModule> |
| 22:59 | <TabAtkins> | Read the rest of the first answer. |
| 22:59 | <Hixie> | i don't think i can unset that from .htaccess |
| 23:00 | <TabAtkins> | Dude, *read the answer*. There's an htaccess line for doing so right in it. |
| 23:01 | <Hixie> | http://damowmow.com/playground/demos/navigation/dot-htaccess is what i've tried, which as far as i can tell is everything from that page |
| 23:01 | <Hixie> | (and every other page from a number of other google queries i did) |
| 23:01 | <TabAtkins> | Okay, that's everything I've seen. |
| 23:01 | <TabAtkins> | Are you outputting from PHP? |
| 23:01 | <Hixie> | perl |
| 23:01 | <Hixie> | and i have $| = 1 |
| 23:01 | <TabAtkins> | Then no clue. |
| 23:02 | <TabAtkins> | (PHP has some output encoding you can turn off too.) |
| 23:02 | <jsbell> | sicking: ping? |
| 23:02 | <jsbell> | sicking: just curious re: Arthur's mail about moving the IDB spec along |
| 23:03 | <Hixie> | TabAtkins: k |
| 23:03 | <Hixie> | TabAtkins: btw "*read the answer*" isn't helpful when i found and read the answer long before i asked here |
| 23:04 | <TabAtkins> | Hixie: You were telling me you'd already done the things in the answer, but only describing precisely the things that the answer said didn't work, and not mentioning the line the answer said *did* work, which was why I pointed you to it in the first place. ^_^ |
| 23:05 | <Hixie> | i only said "also tried that" |
| 23:05 | <Hixie> | meaning that page |
| 23:06 | <TabAtkins> | Okay, that was unclear. It was muddled further when you said "i don't think i can unset that from .htaccess", when the answer specifically provides an htaccess line for unsetting that. ^_^ |
| 23:06 | <TabAtkins> | It sounded a lot like you had just skimmed the page and missed that last part. |
| 23:06 | <Hixie> | no it doesn't, it provides an httpd.conf line for setting that |
| 23:06 | <Hixie> | VirtualHost only appears in httpd.conf |
| 23:06 | <Hixie> | i'm fine with you not being able to help, just a bit offended that you just assume i don't know how to read :-P |
| 23:07 | <TabAtkins> | Oh, I forgot that VirtualHost is only settable in httpd.conf |
| 23:07 | <Hixie> | seriously though, wtf apach |
| 23:07 | <Hixie> | e |
| 23:07 | <TabAtkins> | Hixie: I was as surprised at your apparent lack of ability as you are offended. ^_^ |
| 23:07 | <Hixie> | :-P |
| 23:08 | <annevk> | TabAtkins: yeah, can do, I was actually wondering if we might be able to get away with not having the two-arg delete() |
| 23:08 | <Hixie> | ok so it looks like the problem isn't compression, apache is just buffering stuff up for reasons i cannot phathom |
| 23:09 | <annevk> | TabAtkins: we can always add, not so much remove |
| 23:10 | <TabAtkins> | annevk: Sure. Feel free to simplify for now, and only add the absolute minimum of MultiMap features. |
| 23:10 | <TabAtkins> | annevk: I'm writing an email to es-discuss right now laying out our conclusions and seeing if they're okay. |
| 23:10 | <annevk> | TabAtkins: is MultiMap ordered btw? |
| 23:10 | <TabAtkins> | Map is ordered, so yes. |
| 23:11 | <annevk> | oh, Map is ordereD? |
| 23:11 | <TabAtkins> | (It was that, or require randomization, and people wouldn't bite that bullet.) |
| 23:11 | <annevk> | interesting |
| 23:11 | <TabAtkins> | (Go requires its map iterators to randomize.) |
| 23:11 | <annevk> | because otherwise developers start depending on implementation order? |
| 23:12 | <TabAtkins> | If you don't randomize, you're just speccing it as de factor (and non-interoperably) ordered, anyway. |
| 23:12 | <TabAtkins> | s/factor/facto/ |
| 23:13 | <TabAtkins> | (Or rather, Go's map iterators start from a random *position*. I don't think it specs the actual order, but impl-order should be less dangerous if you can't depend on the starting position.) |
| 23:13 | <annevk> | i'm removing two-arg delete right now |
| 23:13 | <TabAtkins> | kk |
| 23:13 | <annevk> | so TC39 can look at the simplified thing |
| 23:14 | <Hixie> | ok wtf apache. |
| 23:14 | <Hixie> | surely it can't be buffering _megabytes_ of data |
| 23:14 | <TabAtkins> | I fail to see why you find that surprising. |
| 23:14 | <TabAtkins> | Buffers and huge and everywhere. |
| 23:15 | <TabAtkins> | s/Buffers and/Buffers are/ |
| 23:15 | <Hixie> | well mainly i find it surprising because the buffering extension is optional and defaults to 128KB and doesn't seem to be installed |
| 23:15 | <TabAtkins> | Hahaha |
| 23:17 | <Hixie> | also because none of the apache processes are as big as the amount of data i've output |
| 23:17 | tantek | scrolls up to see SimonSapin's quoting something from http://standardssuck.org |
| 23:18 | <SimonSapin> | yeah, just found out about it |
| 23:18 | <tantek> | n.m. that post was from 4+ years ago! |
| 23:18 | <tantek> | "August 4, 2008" |
| 23:18 | <SimonSapin> | indeed |
| 23:18 | <Hixie> | oooh, it _is_ caching |
| 23:18 | <tantek> | GRDDL was horribly impractical, needed XSLT to work etc. |
| 23:19 | <Hixie> | just one of my apache processes just reached like 100mb |
| 23:19 | <Hixie> | and rising |
| 23:19 | <Hixie> | that definitely seems suboptimal |
| 23:19 | <tantek> | SimonSapin, now it's even simpler for the RDF folks get an RDF-like-view of microformats: http://microformats.org/wiki/microformats2-parsing-rdf |
| 23:19 | <tantek> | (if that kind of thing floats your boat) |
| 23:20 | <SimonSapin> | tantek: thanks for the link. I’m not really using/implementing any of these, it was more anecdotal. Seemed appropriate given recent discussions |
| 23:21 | <SimonSapin> | I mean, funny that the same things were being said 4 years (aka. forever) ago. |
| 23:22 | <TabAtkins> | annevk: What happens if you do "u.set('foo', 'bar'); u.append('foo', 'bar');"? Set twice? |
| 23:22 | <tantek> | SimonSapin, yes, such conversations will continue until sufficient conversion tools exist to make the conversations pointless |
| 23:22 | <TabAtkins> | I guess you have to allow that, since you can create a query with "?foo=bar&foo=bar", I think. |
| 23:23 | <annevk> | yeah that works |
| 23:23 | <annevk> | append is just append, nothing special |
| 23:23 | <tantek> | And to that extent, we are making progress. |
| 23:23 | <TabAtkins> | Kk, I was just wondering if "u.delete('foo', 'bar');" would then remove both of them, or the first one, or what. |
| 23:23 | <annevk> | the definition I had would kill both |
| 23:24 | <SimonSapin> | annevk: any idea around of doing standardssuck.org again? :) |
| 23:24 | <TabAtkins> | That's fine with me. Just wondering if there was an obvious answer, for my email. |
| 23:24 | <tantek> | microformats2 parses to pretty much the same JSON that microdata does, but is less markup for authors (and fewer attributes to learn which does what etc.). once there's implementations of such conversion tools for the RDF folks, they can just use that, rather than asking authors to bend over backwards with ugly markup. |
| 23:25 | <annevk> | TabAtkins: yeah dunno, depends on what you expect I guess, maybe some of these system guarantee name-value pair uniqueness? not sure what the benefit of that would be though |
| 23:25 | <tantek> | annevk - you should interview yourself for standardssuck.org on the subject of licenses and (not) being an invited expert for W3C. just splice together video of you asking and answering the questions |
| 23:25 | <tantek> | e.g. sitting in opposing chairs |
| 23:25 | <annevk> | heh |
| 23:25 | <TabAtkins> | That would be funny. ^^_ |
| 23:26 | <SimonSapin> | definitely |
| 23:26 | <tantek> | there aren't enough funny/snarky posts (and video at that) about licensing issues at W3C |
| 23:45 | <sicking> | jsbell: ugh, yeah, i don't think that i realistically will be doing any IDB editing this year :( |
| 23:45 | <jsbell> | sicking: this year = 2012 or this year = next 12 months? Wondering how we can help. |
| 23:46 | <sicking> | 2012 :) |
| 23:46 | <sicking> | jsbell: if you want to help editing that would rock |
| 23:46 | <sicking> | jsbell: the big task is to integrate the latest respec |
| 23:46 | <sicking> | jsbell: that way we can fix up a lot of the webidl, which will tighten things up a lot |
| 23:48 | <jsbell> | sicking: not sure if I'll have cycles over the next few weeks either, but if you can send me pointers maybe I can take a look (I haven't touched one of these specs before) |
| 23:59 | <sicking> | jsbell: cool, will do |
