| 02:39 | <zewt> | https://www.w3.org/Bugs/Public/show_bug.cgi?id=23766 yeah i really don't understand dom3 events at all |
| 02:40 | <zewt> | why is a web spec trying to define every vendor-proprietary key |
| 03:01 | <GPHemsley> | Anyone have any tips for getting a quick understanding of what is valid XSL? |
| 03:04 | <zewt> | *(char*)0 |
| 03:04 | GPHemsley | thinks Hixie_ should give a course on how to write useful specs. |
| 03:05 | <GPHemsley> | This is the closest I've found to the XSL-related specs being useful: http://www.w3.org/TR/xmlschema-0/#index |
| 03:07 | <GPHemsley> | (You'd think XML would be pretty straightforward to document: elements, attributes, possible values—boom, done.) |
| 03:18 | <GPHemsley> | s/XSL/XSD/ |
| 03:18 | <GPHemsley> | I don't know why I keep making that mistake. |
| 08:09 | <Ms2ger> | Hrm |
| 08:09 | <Ms2ger> | So what's [TreatUndefinedAs=Null] good for? |
| 08:21 | <Streusel> | Suggestion to add dc.title for metaextensions. Unless this is the wrong place & I misread it. |
| 08:50 | <MikeSmith> | can't seem to get to http://testthewebforward.org/ from here in Shenzhen |
| 08:51 | <MikeSmith> | great to see the Chinese censors got their priorities straight |
| 08:51 | <Ms2ger> | jgraham, check ^ |
| 08:52 | <MikeSmith> | can get to http://w3cmemes.tumblr.com/ so there's at least one useful thing we'll be able to do while we're here |
| 08:54 | <wilhelm_> | MikeSmith: Confirmed. I've been on VPN all day, so I haven't noticed. |
| 08:59 | <MikeSmith> | wilhelm_: you know if jgraham and annevk-cloud have arrived? |
| 09:02 | <wilhelm_> | MikeSmith: No, don't know which flights they are on. I'm hanging out in the lobby, cursing the wifi. |
| 09:03 | <MikeSmith> | wilhelm_: Wuzhou lobby? |
| 09:03 | <wilhelm_> | Yup. |
| 09:05 | <MikeSmith> | I'll come down |
| 09:05 | <MikeSmith> | building A lobby? |
| 09:06 | <wilhelm_> | Uhh. Main lobby. |
| 09:06 | <wilhelm_> | The one in the middle. |
| 09:07 | <wilhelm_> | jgraham and AutomatedTester just walked in the door. |
| 09:08 | <MikeSmith> | party time |
| 09:09 | <Ms2ger> | Say hi to them, then |
| 09:09 | zcorpan | is now at the guest house |
| 09:16 | <zcorpan> | i guess i'll walk down and see if i can find someone i recognize |
| 09:18 | <annevk-cloud> | MikeSmith: belt seven |
| 09:18 | <annevk-cloud> | In HK... |
| 09:30 | <annevk-cloud> | Ms2ger: I already filed a bug to kill TreatAsUndefined |
| 09:34 | <MikeSmith> | annevk-cloud: you coming to Wuzhou after you arrive? |
| 09:34 | <MikeSmith> | some kind of meetup planned for 8pm |
| 09:35 | <annevk-cloud> | Hopefully I can make that! |
| 09:36 | <hsivonen> | looks like IRC works from China, then |
| 09:38 | <MikeSmith> | hsivonen: yeah |
| 09:40 | <zcorpan> | no dice |
| 09:40 | <wilhelm_> | zcorpan: Some of us are in the sofas in the lobby. |
| 09:40 | <wilhelm_> | Squatting. |
| 09:41 | <zcorpan> | the lobby next to the reception in A? |
| 09:41 | <wilhelm_> | Main lobby. I guess that's A. |
| 09:42 | <zcorpan> | i looked there twice, i guess you must be hiding pretty well |
| 09:42 | <wilhelm_> | Yellow sofas. |
| 09:43 | <zcorpan> | ok i'll try again |
| 09:55 | <sangwhan__> | Don't try to walk out of the hotel, seems to lead straight to some expressway. Almost got killed. |
| 10:02 | hsivonen | makes a mental note of the "other factors" euphemism for prospective Membership fees |
| 10:04 | <MikeSmith> | hsivonen: where's that from? the page about Membership fees I guess? |
| 10:08 | <hsivonen> | MikeSmith: https://twitter.com/w3c/status/398748409018208256 |
| 10:14 | <jgraham> | Hmm, I don't think twitter resolves |
| 10:15 | <jgraham> | Actually, that's not true it's just timing out |
| 10:15 | <hsivonen> | jgraham: the tweet says "@hsivonen We get it. A W3C Host opened there in Jan. Other factors led to holding #TPAC in China; that Twitter is blocked wasn't one of them" |
| 11:20 | <MikeSmith> | annevk-cloud: when you arrive at the Wuzhou, a bunch of us are at the lobby restaurant, straight back from the front entrance |
| 12:05 | <smaug____> | annevk-cloud: ping |
| 12:05 | <smaug____> | annevk-cloud: are you familiar with https://dvcs.w3.org/hg/webperf/raw-file/tip/specs/Beacon/Overview.html ? |
| 12:06 | smaug____ | kind of likes that idea |
| 12:06 | <smaug____> | but sendBeacon should be mapped somehow to XHR.send |
| 12:51 | <annevk-cloud> | MikeSmith: just got to my hotel, still worth it? |
| 12:51 | <MikeSmith> | yeah |
| 12:52 | <MikeSmith> | annevk-cloud: we just finishing dinner now, and there's some TestTWF thing upstairs we going to from here |
| 12:54 | <MikeSmith> | Huaxia Hall |
| 12:54 | <MikeSmith> | ah now |
| 12:54 | <MikeSmith> | *no |
| 12:55 | <MikeSmith> | that's wrong |
| 12:55 | <MikeSmith> | Longzhou Room |
| 12:55 | <MikeSmith> | is the place |
| 12:58 | <annevk-cloud> | Okay, going to find some money and a cab |
| 12:58 | <annevk-cloud> | Oh, and I guess the name of the hotel in Chinese |
| 13:02 | <MikeSmith> | annevk-cloud: hang on, will get it for you |
| 13:05 | <Ms2ger> | Ohai guys: http://i.imgur.com/57yCyrh.jpg |
| 13:13 | <MikeSmith> | annevk-cloud: 深圳市五洲宾馆 |
| 13:17 | <MikeSmith> | Ms2ger: nice to have something to look forward to |
| 13:21 | <annevk-cloud> | Thanks, going nowish. |
| 13:57 | <sangwhan__> | dbaron: 华强北 (hua chang bay) and 赛格广场 (sei ger guang chang) are the keywords you want (romanization might be incorrect, i don't speak chinese) |
| 14:03 | <hsivonen> | good thing https://bugzilla.mozilla.org/show_bug.cgi?id=214729 was WONTFIXed |
| 15:01 | <SteveF> | marcosc: don't mess with Ms2ger's module |
| 15:02 | <marcosc> | heh |
| 15:02 | <marcosc> | Who is the module owner anyway? |
| 15:12 | <Ms2ger> | jst/peterv |
| 15:12 | <Ms2ger> | SteveF, not mine... I learned not to wontfix bugs the same way :) |
| 15:15 | <SteveF> | Ms2ger: comment was driven by alliteration Mess Ms2ger Module |
| 17:27 | <dglazkov> | good morning, Whatwg! |
| 17:42 | <GPHemsley> | Is there no element selector in Chrome Inspector? |
| 17:43 | Ms2ger | thinks zcorpan got his time zones wrong |
| 18:08 | <karlcow> | are there any browsers reading the "type" information in <link rel="icon" type="blah" />? |
| 18:08 | <karlcow> | My initial gut is none. |
| 18:12 | GPHemsley | thinks that would fall under the purview of mimesniff |
| 18:13 | <GPHemsley> | but I don't know the answer offhand |
| 18:32 | <jorendorff> | there should be a Promise.nowThen(f), equivalent to Promise.fulfill(undefined).then(f) |
| 18:32 | <Hixie_> | nowThen nowThen |
| 18:32 | <Hixie_> | what's all this then |
| 18:33 | Hixie_ | shows himself out |
| 18:33 | <TabAtkins> | jorendorff: What would that do? |
| 18:34 | <jorendorff> | TabAtkins: .then() is a good API, it would be nice to have an extremely similar API to use when I'm creating a new Promise from scratch |
| 18:35 | <TabAtkins> | That's just Promise.fulfill(f()) |
| 18:35 | <jorendorff> | well, no, because f may throw |
| 18:35 | <TabAtkins> | Ah, I see. |
| 18:36 | <jorendorff> | the Promise constructor -- i can see why that is necessary, but those state-mutate-y functions, it's pretty weird using them in combination with the rest of the API which is all combinator-y |
| 18:37 | <TabAtkins> | Yeah, and you can't easily use Promise(f), because f needs to affirmatively resolve the promise with its return value. (But the throwing possibility is taken care of automatically.) |
| 18:38 | <TabAtkins> | Promise.fulfill().then(f) works, of course. |
| 18:38 | <jorendorff> | yeah |
| 18:38 | <TabAtkins> | Promise.call(f) |
| 18:38 | <TabAtkins> | Suggest it as sugar on the repo. |
| 18:38 | <TabAtkins> | Wait no, that doesn't work, because Promise is already a function. ^_^ |
| 18:38 | <Domenic_> | Promise.try is the traditional name. |
| 18:38 | <jorendorff> | heh! good catch :) |
| 18:39 | <TabAtkins> | Domenic_: Sure, that sounds good. |
| 18:39 | <Domenic_> | Pretty sure there's already a closed issue for that |
| 18:39 | <Domenic_> | I have an interesting data point on DRM |
| 18:40 | <Domenic_> | I am on client site for a top-5 US bank. They need training videos in their web app (which they are selling to many other banks). |
| 18:40 | <Domenic_> | When presented with the choice between flash + encryption or HTML5 + no encryption, they said HTML5 was a no-brainer |
| 18:40 | <Domenic_> | They would just protect the videos with login etc. |
| 18:41 | <Domenic_> | (story over) |
| 18:41 | <TabAtkins> | Yeah, because that's the *right answer*. |
| 18:41 | <TabAtkins> | But cool to hear. ^_^ |
| 19:36 | <GPHemsley> | Why can't I get a ::before pseudo-element to not inherit text-decoration? ;_; |
| 19:45 | <Hixie_> | GPHemsley: it isn't inheriting it |
| 19:45 | <Hixie_> | GPHemsley: it's just being drawn by the parent |
| 19:45 | <Hixie_> | across all its children |
| 19:45 | <GPHemsley> | but I can't undo it, even with !important |
| 19:46 | <Hixie_> | it's not inherited at all, so, yeah |
| 19:47 | <Hixie_> | there's nothing to undo |
| 19:48 | <GPHemsley> | so how do I do what I want? |
| 19:53 | <Hixie_> | GPHemsley: you want to underline the contents of the element and not underline its ::before contents? |
| 19:53 | <GPHemsley> | line-through, but yeah |
| 19:53 | <Hixie_> | GPHemsley: i don't think that's possible without either my elaborate ::outside stuff, or extra elements in the DOM |
| 19:54 | <GPHemsley> | why?? ;_; |
| 20:37 | <TabAtkins> | GPHemsley: text-decoration doesn't inherit. Text decorations do, in a separate inheritance mechanism. |
| 20:37 | <TabAtkins> | There are values in Text Decor that suppress decorations coming from your ancestors. |
| 21:15 | <jorendorff> | Domenic_: would you mind joining #jslang on irc.mozilla.org? |
| 22:30 | <Hixie_> | TabAtkins: one option for the img thing that might be better than src-* or scripts or srcset is to have a way to declare how you pick images in general, which you could stash in a block in the <head>, and then you wouldn't need to repeat all the details in each <img> but could still prime the preloaer appropriately |
| 22:33 | <TabAtkins> | Hixie_: Yeah, I've been over this in my head, and I can't come up with a way to program the preloader that's sufficiently general to accomplish what src-N does, let alone goes further. |
| 22:34 | <Hixie_> | well, you can move most of the logic from src-n to the preloader, right? i mean, in practice, there's gonna be a lot of replication |
| 22:34 | <Hixie_> | even if it was just to declare a "template", basically, and then srcfoo="" could just invoke a template and then list the urls to put into that template |
| 22:41 | <TabAtkins> | Problem is that you can't *just* provide extraction logic. You need to have sufficient primitives to program it with so it can *decide* which to load. |
| 22:45 | <Hixie_> | TabAtkins: sure. i'm saying move from <img srcset="foo 2x, bar 1x"> to some block that declares a simple template taking two URLs and mapping them to 2x and 1x respectively, and then <img srcfoo="foo bar"> |
| 22:45 | <Hixie_> | TabAtkins: obviously with one image it's not great, but with a huge grid of images on some complex multi-"breakpoint" layout, it'd be a huge win |
| 22:46 | <Hixie_> | no need for multiple attributes per element, for example |
| 23:02 | <TabAtkins> | Hixie_: Right, but what *I'm* saying is that that's not enough. You lose the ability to MQ. You lose the compactness and readability of the Variable-Sized Images grammar. |
| 23:03 | <TabAtkins> | You keep trying to solve the problems you were presented with back when you wrote srcset, and discounting the use-cases that have been collected since. :/ |
| 23:03 | <TabAtkins> | Those earlier problems are easy. |
| 23:03 | <Hixie_> | i'm not trying to solve any problems right now except the extreme ugliness of src-* :-) |
| 23:04 | <Hixie_> | i don't see why you lose any of those abilities |
| 23:04 | <Hixie_> | i'm saying keep the same abilities |
| 23:04 | <Hixie_> | just factor them out so you don't have to keep repeating them |
| 23:17 | <Hixie_> | anyone got IE around able to test http://www.hixie.ch/tests/adhoc/html/script/callbacks/002.html for me? |
| 23:27 | <annevk> | Hixie_: you should really get VMWare :-) |
| 23:27 | <annevk> | 0: outer |
| 23:27 | <annevk> | 1: inner |
| 23:28 | <annevk> | 2: inner |
| 23:28 | <annevk> | 3: nothing happens |
| 23:29 | <annevk> | 4: nothing happens |
| 23:29 | <annevk> | For both 3 and 4 I get this in the console: |
| 23:29 | <annevk> | > SCRIPT65535: Invalid calling object |
| 23:29 | <annevk> | > 002a.html, line 1 character 1 |
| 23:30 | <annevk> | Hixie_: this is on IE10, Windows 7 |
| 23:30 | <Hixie_> | thanks |
| 23:31 | <annevk> | Hixie_: also, your cute ASCII art made this take longer :p |
| 23:31 | <Hixie_> | why? |
| 23:31 | <annevk> | Hixie_: distraction :-) |
| 23:31 | <Hixie_> | it's supposed to make it way quicker, since you can just recognise which it is without reading |
| 23:32 | <Hixie_> | (try 001.html instead... it has 33 subtests, so you get tired of reading pretty quick) |
| 23:32 | <annevk> | hah, I guess if you looked at things for a while, but I only got 3 actual result pages |
| 23:33 | <Hixie_> | yeah, for 002 it's not a big deal. i just copied the 'a's from 001 |
| 23:38 | <TabAtkins> | Hixie_: I can't tell if you don't understand the problem, or are trying to say something subtle. Please explain how a template + preloader instructions can handle literally anything more complex than "list of sources and densities". |
| 23:40 | <TabAtkins> | More properly, as far as I can tell, you don't understand what src-N does, so your suggestions are intentionally too weak because you think that's all that's necessary. |
| 23:40 | <coldwind> | annevk-cloud: are you around? |
| 23:40 | <coldwind> | annevk-cloud: could you look at this path for the URL spec? https://github.com/smola/url/commit/00f0abd934f619e86281083466b65d23cc3f23cc |
| 23:40 | <coldwind> | it seems to me the parsing rules can be simplified a little bit |
| 23:40 | <Hixie_> | TabAtkins: say proposal A results in markup of the form "AAA 111 BBB 222 CCC 333". You could form a different proposal that results in markup of the form "AAA BBB CCC" "111 222 333" with no loss of generality. |
| 23:40 | <coldwind> | patch* |
| 23:41 | <TabAtkins> | Hixie_: So yes, you're assuming that all we need to solve is "list of sources with densities", which is not correct. |
| 23:41 | <coldwind> | annevk-cloud: forget it :p |
| 23:41 | <Hixie_> | TabAtkins: the statement i just made applies to pretty much any feature, it's not specific to images or anything. |
| 23:41 | <TabAtkins> | The problem with src-N is *not* "I have to repeat the densities multiple times". |
| 23:42 | <TabAtkins> | But that's what you're solving. |
| 23:42 | <Hixie_> | i don't know why you keep mentioning densities, they're not relevant to my argument? |
| 23:42 | <TabAtkins> | I'm trying to figure out how your argument is relevant to what we're discussing by casting it to actual syntax and what authors would do. |
| 23:43 | <Hixie_> | TabAtkins: you'll typically have many images with very similar descriptions but different image files, right? |
| 23:43 | <Hixie_> | many <img>s |
| 23:43 | <TabAtkins> | "descriptions"? |
| 23:44 | <Hixie_> | by descriptions i mean method of picking one from many |
| 23:44 | <Hixie_> | media queries, srcset syntax, whatever |
| 23:44 | <Hixie_> | the stuff that isn't file names, that you put in src-* attributes. |
| 23:44 | <TabAtkins> | Oh, you're trying to solve something completely different - "multiple <img> elements use the same basic source patterns". |
| 23:44 | <TabAtkins> | Which is irrelevant for src-N discussions - it's an orthogonal problem. |
| 23:44 | <Hixie_> | as i said above, the problem i'm trying to solve is "img-* is incredibly ugly" |
| 23:44 | <Hixie_> | src-*, rather |
| 23:45 | <TabAtkins> | This doesn't do a thing for that problem. |
| 23:45 | <TabAtkins> | Literally nothing. |
| 23:45 | <Hixie_> | k |
| 23:46 | <TabAtkins> | (Also, note that most people's suggested "fix" for "src-N is ugly" is "jam them all into a single attribute with another level of delimiters".) |
| 23:48 | <Hixie_> | yeah, i don't think that's particularly gonna help :-) |
| 23:52 | <Hixie_> | annevk-cloud: actually, would be interesting to see if you got different results on 001 than i did with IE9 |
