| 00:02 | <Hixie> | http://en.wikipedia.org/wiki/Non-English_usage_of_quotation_marks is rather humbling. |
| 00:37 | <hober> | Hixie: indeed, hence dbaron's bug |
| 00:38 | <Hixie> | hm? |
| 00:38 | <dbaron> | http://www.w3.org/Bugs/Public/show_bug.cgi?id=13718 |
| 00:40 | <Hixie> | ah, interesting. hadn't seen that one yet. |
| 00:40 | <Hixie> | can't really argue with the first point. |
| 00:40 | <Hixie> | for the second point, we tried that. people didn't go for it. |
| 00:41 | <Hixie> | or rather, we tried making <q> that. |
| 00:41 | <Hixie> | there isn't really a use case for the described element. |
| 00:41 | <dbaron> | for people who like semantics? :-) |
| 00:42 | <dbaron> | alternatively, deprecate <q> in favor of quotation marks |
| 00:43 | <Hixie> | semantics isn't a use case on its own |
| 00:43 | <Hixie> | the HTML spec already says that there's no reason to use <q> rather than quotation marks |
| 00:43 | <Hixie> | "The use of q elements to mark up quotations is entirely optional; using explicit quotation punctuation without q elements is just as correct." |
| 00:44 | <Hixie> | there's even an example |
| 01:03 | <Hixie> | heycam: are you going to drop [AllowAny]? |
| 01:03 | <Hixie> | heycam: in http://www.w3.org/Bugs/Public/show_bug.cgi?id=12870 you suggest i use it but the spec says it might go away |
| 01:06 | <heycam> | Hixie, probably not going to drop it |
| 01:06 | <Hixie> | k |
| 01:06 | <heycam> | Hixie, it's kind of ugly, but... *shrug* |
| 01:06 | <Hixie> | is there anything in JS that has a [[Call]] thingy but is not a Function? |
| 01:07 | <heycam> | Hixie, in the spec, I don't think so. but isn't there something with RegExp objects? |
| 01:08 | <heycam> | (might be completely wrong here) |
| 01:08 | Hixie | tries setTimeout with a RegEcp |
| 01:08 | <Hixie> | RegExp |
| 01:09 | <Hixie> | ...i have no idea what i would test |
| 01:09 | <heycam> | maybe they're not going to be callable forever... https://bugzilla.mozilla.org/show_bug.cgi?id=582717 |
| 01:11 | <Hixie> | Having the domintro blocks and IDL blocks match isn't a goal, btw. |
| 01:11 | <heycam> | Hixie, ok, I understand that |
| 01:11 | <heycam> | you want to omit the finicky details |
| 01:11 | <Hixie> | I mean, they both have to match the same reality, but it's no problem if they use different arguments and stuff |
| 01:11 | <Hixie> | yeah |
| 01:11 | <Hixie> | (different argument names, that is) |
| 01:12 | <Hixie> | gotta love using more fancy webidl syntax though |
| 01:12 | <heycam> | except when it looks crazy and distracting |
| 02:16 | <gsnedders> | heycam, Hixie: yeah, there's agreement between all major browser vendors to drop [[Call]] from RegExp. Some already have, some haven't, AFAIK. |
| 02:16 | <gsnedders> | (IE never had [[Call]] on RegExp) |
| 02:17 | <gsnedders> | Ah, gone in Fx5. Probably in Saf5.1, too, though I'm not digging through SVN at this time to see if they have. |
| 02:18 | <gsnedders> | Opera 11.50 dropped it too |
| 02:19 | gsnedders | wishes all this stuff happened in public and not emails between JS developers — but it's considered off-topic for es-discuss as it's non-standard |
| 02:20 | <AryehGregor> | Use whatwg. :) |
| 02:21 | <Hixie> | yeah you're welcome to use whatwg |
| 02:21 | <Hixie> | don't forget to update the relevant whatwg wiki page |
| 03:10 | <MikeSmith> | Hixie: I'll split bug 12986 |
| 03:10 | <MikeSmith> | (the big "Last Call comments to HTML5" one) |
| 03:11 | <zewt> | If you're having problems seeing the site, it may be because Grooveshark doesn't support your current zoom level. To remedy this problem, simply press CTRL+0 (CMD+0 for Mac users) to return to your browser's default zoom level. |
| 03:11 | <zewt> | gross |
| 03:22 | <MikeSmith> | Hixie: 13590 too |
| 04:47 | <MikeSmith> | that was fun |
| 05:26 | <Hixie> | MikeSmith: thanks dude |
| 05:27 | <MikeSmith> | no problem man |
| 05:27 | <MikeSmith> | I should have done it from the beginning |
| 05:27 | <MikeSmith> | was just being lazy |
| 05:31 | <Hixie> | heh |
| 05:34 | <MikeSmith> | Hixie: if I missed any others that need splitting, let me know |
| 05:35 | <MikeSmith> | btw, hsivonen can read your mind |
| 05:35 | <MikeSmith> | you made this change recently: |
| 05:35 | <MikeSmith> | http://html5.org/tools/web-apps-tracker?from=6387&to=6388 |
| 05:35 | <MikeSmith> | about the <link> element for microdata |
| 05:36 | <MikeSmith> | but hsivonen already implemented checking for it in validator.nu exactly it that way last year |
| 05:38 | <Hixie> | cool |
| 07:05 | <Hixie> | does css have a way to remove whitespace nodes these days? |
| 07:05 | <Hixie> | e.g. if i have <div><span/> <span/></div> and want to have the two spans be inline-blocks that touch, no space from the space character? |
| 07:21 | <benjoffe> | Hixie: 'whitespace: discard;' should do that I think |
| 07:21 | <Hixie> | oooh |
| 07:21 | <Hixie> | is that implemented anywhere? |
| 07:21 | <benjoffe> | Not sure |
| 07:23 | <benjoffe> | white-space-collapse: discard; |
| 07:26 | <MikeSmith> | Dadhacker: Revolutions That Weren’t - http://www.dadhacker.com/blog/?p=1444 scroll down to the "Ex Em Hell" part |
| 07:26 | <MikeSmith> | "What we actually got: Any number of crappy serialization schemes and over-designed and under-implemented replacements for INI files. Undebuggable configuration files, poorly written attempts at replacements to already perfectly awful tools (yes, ANT and MSBuild, I’m thinking of you), and a lot of other smelly garbage littered with angle brackets." |
| 07:27 | <MikeSmith> | "We are still living this nightmare, with no end in sight." |
| 09:11 | <zcorpan> | selectors has [foo^=bar], [foo$=bar] and [foo*=bar] but no way to match case-insensitively |
| 09:12 | <zcorpan> | [foo=/bar/i] |
| 09:13 | <jgraham> | Of course what selectors needs is to look even more like perl |
| 09:14 | <zcorpan> | is there a punctuation character that is usually used to indicate case-insensitivity? |
| 09:15 | <jgraham> | Well, I don't think using random punctuation would be *better*. I mean selectors is already a giant mess |
| 09:15 | <woef> | It would be nicer if html would be case-insensitive in the first place. |
| 09:16 | <jgraham> | But giving people /foo/i would make them think that foo was a regexp |
| 09:16 | <zcorpan> | woef: html *is* case-insensitive for various attribute values |
| 09:16 | <jgraham> | or /bar/i I guess, given your examples |
| 09:17 | <zcorpan> | yeah i was thinking maybe we should just allow full regexps. but maybe that's not acceptable |
| 09:17 | <woef> | zcorpan: so just css selector matching :) |
| 09:18 | <zcorpan> | woef: the problem is that selectors isn't case-insensitive |
| 09:18 | <zcorpan> | except there's an ugly hack to make it so for some attributes in text/html |
| 09:19 | <jgraham> | zcorpan: Well people could always write very slow regexp |
| 09:19 | <zcorpan> | yeah |
| 09:19 | <jgraham> | Or, I guess you could limit the syntax to real regexp and expect people to use something like that google library that is supposed to have good time behaviour in all cases |
| 09:20 | <jgraham> | But that seems like quite a lot of implementation complexity |
| 09:21 | <jgraham> | (re2) |
| 09:21 | <zcorpan> | yeah i'd like just a flag for case-insensitive, but would be nice if it could be used together with ^= $= and *= |
| 09:21 | <jgraham> | (but the site seems to be down) |
| 09:21 | <zcorpan> | and doesn't look like a smiley |
| 09:21 | <zcorpan> | ^_= |
| 09:21 | <zcorpan> | or, what the hell |
| 09:22 | <zcorpan> | ^:= |
| 09:23 | <zcorpan> | that looks ok |
| 09:24 | <woef> | http://images2.layoutsparks.com/1/112359/punisher-black-white-face.jpg -> looks a lit like the Punisher logo :p |
| 09:25 | <zcorpan> | awesome |
| 09:25 | jgraham | doesn't like the fact that CSS loosk increasingly like a cat walking across the number keys whilst holding down shift |
| 09:30 | <woef> | ^_= looks like you just punched someone in the face. Which I guess is exactly what you want to do if you have to worry about case sensitivity in html attributes :x |
| 10:16 | <Philip`> | zcorpan: Maybe [lc(foo)^=bar] so it's case-sensitive matching a forced-to-lowercase value |
| 10:22 | <zcorpan> | wfm |
| 10:23 | <zcorpan> | but wouldn't you do [foo^=lc(bar)] ? |
| 10:25 | <Philip`> | bar is the string the user typed into the selector and they'll have already written that in lowercase, so no |
| 10:26 | <Philip`> | foo is the attribute which they don't know the case of and want to force to lowercase before matching |
| 10:26 | <zcorpan> | fair enough |
| 11:37 | <gsnedders> | jgraham_: re2 is O(n) when matching, though can be O(n^2) compiling the regexp. It's the tradeoff you make for converting the NFA to a DFA. |
| 11:45 | <Philip`> | O(n^2) memory usage too, I think |
| 11:54 | <jgraham_> | gsnedders: But you could compile once per document rather than once per match, so it seems much more acceptable |
| 11:54 | <jgraham_> | Memory usage might be more of an issue |
| 11:54 | <jgraham_> | Not that I am advocating this at all |
| 11:59 | <Ms2ger> | MikeSmith, might be good to redirect http://people.w3.org/mike/web-platform/ |
| 11:59 | <Ms2ger> | Or at least remove D3C :) |
| 12:01 | <Ms2ger> | MikeSmith, and add a link to xml-stylesheet? |
| 12:08 | <MikeSmith> | Ms2ger: done and done |
| 12:08 | <MikeSmith> | thanks |
| 12:08 | <Ms2ger> | Thanks |
| 12:09 | <Ms2ger> | I think that was the reason I've got "<?xml-stylesheet" written on my desk |
| 12:09 | Ms2ger | gets an eraser |
| 12:10 | <MikeSmith> | heh |
| 12:17 | <gsnedders> | Philip`: Indeed. |
| 12:19 | <zcorpan> | Ms2ger: surely you had it on your desk because you wanted to praise me for having worked on it |
| 12:20 | <Ms2ger> | zcorpan, it's a mess of a spec |
| 12:20 | <Ms2ger> | zcorpan, I mean, er.. |
| 12:20 | <Ms2ger> | Huh |
| 12:20 | <Ms2ger> | http://html5.org/tools/web-apps-tracker?from=6403&to=6404 |
| 12:22 | <zcorpan> | Ms2ger: i blame the WG for working against me :P |
| 12:22 | <Ms2ger> | I guess that's actually true |
| 12:24 | Ms2ger | shakes his fist at XMLCore |
| 12:28 | <MikeSmith> | nice, I just committed revision 666 of the validator.nu schema |
| 12:29 | MikeSmith | breaks out some champagne and upside-down crosses |
| 12:34 | <MikeSmith_> | test automation UI that Francois set up is nice: |
| 12:34 | <MikeSmith_> | http://w3c-test.org/framework/suite/nav-timing-default/ |
| 12:35 | <zcorpan> | MikeSmith: what wsa the commit? |
| 12:35 | <Ms2ger> | Section "4.3: The" |
| 12:36 | <MikeSmith> | zcorpan: https://bitbucket.org/validator/syntax/changeset/51b2578b6d58 |
| 12:38 | <zcorpan> | evil |
| 12:39 | <jgraham_> | MikeSmith: Uh, opinions about that might vary :) |
| 12:39 | <MikeSmith> | well, it's a start |
| 12:39 | <MikeSmith> | it was totally manual before |
| 12:39 | <jgraham_> | I was quite surprised that it auto0submitted my results to the W3C |
| 12:40 | <jgraham_> | Maybe it said somewhere that it would, but that makes it pretty toxic for people who don't work entirely in public |
| 12:41 | <jgraham> | I should point this out on the mailing list |
| 12:43 | <MikeSmith> | oh |
| 12:43 | <MikeSmith> | yeah |
| 12:43 | <MikeSmith> | true |
| 12:44 | <MikeSmith> | needs a warning and/or confirmation dialog of some kind |
| 12:47 | <jgraham> | In general I am a test-framework skeptic |
| 12:47 | <Ms2ger> | [x] Report test results automatically (when possible) |
| 12:47 | Ms2ger | missed that as well the first time |
| 12:48 | <jgraham> | I don't think the problem it is trying to solve ("we need test reports to get to CR") is actually the right problem to focus on |
| 12:49 | <jgraham> | The right problem to focus on is "how do we get vendors running all the tests every day" |
| 12:49 | <jgraham> | Once you have that, the "we need the results" thing becomes trivial |
| 12:49 | <Ms2ger> | Working on that... |
| 12:50 | <Ms2ger> | (And IE every three days, I presume) |
| 12:50 | <jgraham> | Heh |
| 12:50 | jgraham | is working on it too |
| 12:51 | <Ms2ger> | Next up: useful tests |
| 12:51 | <jgraham> | Yep. We are getting there, slowly |
| 12:52 | <Ms2ger> | And mostly on Opera's back, or so it seems |
| 12:53 | <Ms2ger> | <noscript><p>Enable JavaScript and reload</p></noscript> <-- Is that common in your tests? :) |
| 12:53 | <jgraham> | That is in our internal guidelines |
| 12:53 | <jgraham> | I have suggested it is silly |
| 12:54 | <jgraham> | But apparently it isn't so silly that the people who have been doing it habitually want to stop |
| 12:55 | <Ms2ger> | The same guidelines that want window.undefined? :) |
| 12:56 | <jgraham> | No, that's just Tarquin :) |
| 12:56 | <zcorpan> | maybe i should start writing window.frames.self.undefined |
| 12:56 | <zcorpan> | you can almost form a sentence with attributes that return window |
| 12:59 | <zcorpan> | hmm there weren't more than those, how disappointing |
| 12:59 | <Ms2ger> | Gecko has _content |
| 12:59 | <Ms2ger> | And content |
| 12:59 | <Ms2ger> | Not sure what they do exactly |
| 13:00 | <zcorpan> | huh, didn't know about those |
| 13:02 | <Ms2ger> | Interestingly enough, if you set window.content yourself, _content will mirror that |
| 13:04 | <zcorpan> | window === content // false |
| 13:04 | <zcorpan> | window === _content // false |
| 13:04 | <zcorpan> | content is enumerable but _content is not |
| 13:05 | <zcorpan> | _content = 1 does nothing |
| 13:06 | <Ms2ger> | Nope |
| 13:07 | <Ms2ger> | We only handle it when getting |
| 13:07 | Ms2ger | wants all this code to die |
| 13:08 | <zcorpan> | is there a bug? |
| 13:09 | <Ms2ger> | For getting bindings that don't suck as much? Yes, probably |
| 13:11 | <zcorpan> | i meant for dropping content and _content |
| 13:12 | <Ms2ger> | Might be something for extensions |
| 13:23 | <zcorpan> | MikeSmith: is it possible to tweak the settings for which emails go to public-html-bugzilla so that if only the cc list is changed, don't send an email? |
| 13:24 | <Ms2ger> | Should be |
| 13:25 | <MikeSmith> | zcorpan: yeah |
| 13:25 | <MikeSmith> | will do that in a bit |
| 13:25 | <MikeSmith> | but now, breakfast |
| 15:00 | <benjoffe_> | I'm surprised that no browser has implemented a special readable stylesheet for html pages that have no styles or scripts |
| 15:01 | <Ms2ger> | I'm not |
| 15:01 | <Ms2ger> | There is lots of other stuff to do |
| 15:02 | <miketaylr> | safari has that reader/readability feature |
| 15:02 | <benjoffe_> | This wouldn't be hard to implement, and there's tonnes of pages out there like this, especially academic content |
| 15:03 | <benjoffe_> | miketaylr: I'm thinking something like that but with all the links and forms styled well inline too |
| 15:05 | <AryehGregor> | You can use a bookmarklet. |
| 15:05 | <AryehGregor> | I have one called Readability. |
| 15:05 | <AryehGregor> | I use it for pages with horrible contrast or such. |
| 15:07 | <benjoffe_> | I know there's ways for me to consume this content, but I'm just surprised none of them have anything decent on by default |
| 15:15 | <AryehGregor> | Someone seriously needs to tell Google Maps about the 184th Street entrance to the 181st Street A station. And the Bennett Avenue entrance to the 190th A. It seems to think there's only one entrance to each station, and it totally messes up the directions it gives. |
| 15:16 | <AryehGregor> | Like thinking it makes as much sense to tell you to walk to 181st and St. Nicholas to get to the 1 as walking to 181st and Fort Washington to get to the A, when you're starting at 184th and Overlook and there's an entrance to the A like fifty feet away. |
| 15:16 | <gsnedders> | AryehGregor: Not as bad as Subway station closest to me, it's ignorant of there is a staircase down from bridge to where it is, so makes you go a huge long way around |
| 15:17 | <zewt> | as much as I like gmaps, I really wish Garmin would make an Android release |
| 15:17 | <AryehGregor> | It does that with the 190th A. |
| 15:18 | <AryehGregor> | There's an entrance on Bennett and one on Fort Washington, and to get from one to the other you have to walk several blocks around because Fort Washington is on top of a cliff overlooking Bennett at that point. |
| 15:18 | <AryehGregor> | Google Maps only knows about the Fort Washington entrance. |
| 15:19 | <gsnedders> | It's really quite amusing around where I used to live: such a random selection of footpaths on it. |
| 15:19 | AryehGregor | tries http://maps.google.com/support/bin/answer.py?answer=162873 |
| 15:21 | <zewt> | is glenn adams trolling? heh |
| 15:21 | <Ms2ger> | zewt, no comment :) |
| 15:25 | <jgraham> | Someone should tell him why W3C publishes "technical recommendations" |
| 15:26 | <Ms2ger> | De Jure Standards, I'm telling you! |
| 15:26 | <gsnedders> | AryehGregor: Not available in the UK, so can't do that for here… |
| 15:26 | <jgraham> | SOmeone whould tell me how the fuck I have managed to break virtualbox so that installing guest additions has no effect |
| 15:27 | <zewt> | through the power of song |
| 16:31 | <dglazkov> | good morning, Whatwg! |
| 17:55 | AryehGregor | does not understand how someone can *consistently* misspell his name as "Areyeh" when communicating in a text medium where his name is presented with the correct spelling on every post |
| 17:56 | <gesa> | Someone doesn't know how to properly utilize copy & paste. Clearly |
| 18:00 | <Philip`> | AryehGregor: Some people consistently can't even spell "Philip" :-( |
| 18:00 | <Ms2ger> | Phillip`: Hmm? |
| 18:01 | Philip` | generally avoids referring to people by name entirely, and sticks with pronouns where possible, else is careful to double-check and/or copy-and-paste their names |
| 18:04 | jgraham | gets Philip wrong at least 50% of the time unless there is tab completion |
| 18:08 | <jgraham> | (otoh, people very often forget that the possesive of James is spelt James') |
| 18:08 | <Ms2ger> | Jame's or James's? |
| 18:09 | <zewt> | james's |
| 18:09 | <zewt> | there is not more than one jame :) |
| 18:09 | <jgraham> | mmm jam |
| 18:09 | <jgraham> | Sorry we were talking about food, right? |
| 18:10 | <jgraham> | I believe that the extra s is also acceptable |
| 18:10 | <zewt> | people regularly call me "glen", even when writing my name in an email where my name, correctly spelled, is right there on screen |
| 18:10 | <zewt> | i'd consider james' flatly incorrect, fwiw |
| 18:11 | Ms2ger | raises a FORMAL COMPLAINT against James' |
| 18:11 | <jgraham> | Well maybe I get it wrong then :p But the mistake I had in mind was not having any suffix at all |
| 18:11 | <gsnedders> | zewt: What is your name? |
| 18:12 | <zewt> | glenn |
| 18:12 | <gsnedders> | I see all kinds of variations of Geoffrey. Most commonly "Geoffery". |
| 18:12 | <Ms2ger> | Are you the Glenn on webapps? |
| 18:12 | <zewt> | there are two Glenns on webapps |
| 18:12 | <zewt> | one of which I am not |
| 18:12 | <jgraham> | zewt: You should tell the other Gleen that he is brining your name into disrepute |
| 18:12 | <Ms2ger> | It would be weird if you were both |
| 18:12 | <zewt> | i'm certainly not the one (seemingly) trolling against whatwg today, heh |
| 18:13 | <Ms2ger> | Good |
| 18:13 | <jgraham> | Not much weirder than the two Philip Taylors on public-html |
| 18:13 | <zewt> | it's sad when it's more important to explain clearly who you are not than who you are |
| 18:13 | <zewt> | it sure is confusing when hixie refers to "Ian" |
| 18:13 | <Ms2ger> | That |
| 18:13 | Philip` | would generally consider "James'" to be probably technically acceptable but old-fashioned and silly, kind of like "an historic" or "the data are" |
| 18:14 | Ms2ger | thinks Philip` is an hero |
| 18:14 | jgraham | thinks that insisting data is plural is... very annoying |
| 18:14 | <zewt> | i'd personally consider it technically incorrect because the "s'" construction is for plural possessives |
| 18:15 | <zewt> | at least "an historic" and "data are" i know where they come from |
| 18:15 | <Ms2ger> | Octopodes, I'm telling you |
| 18:17 | <Philip`> | zewt: It's not childrens' etc - I see it as more related to ending with an s rather than being a plural (and rather than being a plural that ends with an s) |
| 18:17 | <zewt> | and i assume everyone else over 20 is also righteously infuriated over such nonsense as "on accident" |
| 18:18 | <Philip`> | (though I'm just going by what seems right, rather than what is officially right in some sense) |
| 18:19 | <jgraham> | Yeah, I would also say "gsnedders'" but "zewt's" |
| 18:19 | <zewt> | Philip`: but the "'s" is pronounced as a separate "s", where "s'" is not; for example, we do pronounce two s's in "boss's" |
| 18:19 | <jgraham> | (umm, obviously the latter) |
| 18:19 | <zewt> | likewise for "james's", it's not pronounced as "james desk" |
| 18:19 | <Ms2ger> | !summon timeless |
| 18:19 | <Philip`> | zewt: Surely nobody would say something like that by purpose? |
| 18:19 | <zewt> | *smack* |
| 18:20 | <jgraham> | zewt: You seem to be mistakenly assuming a clean mapping between English spelling and pronounciation |
| 18:20 | <zewt> | people under around 20 seem to be saying "on purpose" a lot, which makes me sad (there are some others, though I'm having trouble recalling them off-hand) |
| 18:21 | <zewt> | there's a reasonably consistent mapping between "s'" and "s's" and one or two s sounds, in all cases I can think of |
| 18:21 | <zewt> | english certainly isn't a consistent language, but that doesn't mean it has no consistencies at all :) |
| 18:21 | <Ms2ger> | Those it has are probably accidents |
| 18:23 | <MikeSmith> | it would be very cool to actually rescind a Rec |
| 18:23 | <MikeSmith> | for a number of reasons |
| 18:23 | <MikeSmith> | including having a precedent |
| 18:23 | <zewt> | show that it can be done? heh |
| 18:25 | <Philip`> | Let's rescind them all and start again from scratch |
| 18:26 | <Ms2ger> | Philip`, want to write two paragraphs to rescind Views? |
| 18:27 | <Philip`> | Since I don't know what it is, no |
| 18:27 | <timeless> | anyone here familiar w/ <input type=number> ? |
| 18:30 | <Ms2ger> | Ah, timeless |
| 18:30 | <Ms2ger> | timeless, want to write two paragraphs to rescind Views? :) |
| 18:33 | <Philip`> | I guess that's a no? |
| 18:34 | <Ms2ger> | :( |
| 19:10 | <Ms2ger> | So delete window.Node should do what? |
| 19:18 | <AryehGregor> | Nothing good, I suspect. |
| 19:18 | <Ms2ger> | >.< |
| 19:19 | <Ms2ger> | It seems like the answer is "Nothing", but I can't figure out why that is |
| 19:21 | AryehGregor | recommends summoning heycam|away |
| 19:22 | <Ms2ger> | I tried |
| 19:22 | <Ms2ger> | He's always away |
| 19:22 | <Ms2ger> | I mean, it's past 6AM in NZ |
| 19:32 | <gsnedders> | Ms2ger: Well, I guess window.Node.[[Configurable]] = false |
| 19:33 | <Ms2ger> | No |
| 19:33 | <gsnedders> | Ms2ger: But I guess you mean in IDL terms? |
| 19:33 | <Ms2ger> | The property has the attributes { [[Writable]]: true, [[Enumerable]]: false, [[Configurable]]: true }. |
| 19:34 | <Ms2ger> | http://dev.w3.org/2006/webapi/WebIDL/#es-interfaces |
| 19:35 | <gsnedders> | Ms2ger: The global object is odd and defined nowhere in real terms :( |
| 19:35 | Ms2ger | sighs |
| 19:36 | Ms2ger | files a bug |
| 19:38 | <gsnedders> | Ms2ger: Look at self and parent for the *real* gfun. |
| 19:38 | <gsnedders> | *fun |
| 19:38 | <Ms2ger> | No thanks |
| 19:38 | <w3ztb0y> | is there any indonesian people?? |
| 19:39 | <Ms2ger> | Presumably there are |
| 19:39 | <Ms2ger> | AFAIK, Indonesia hasn't been abandoned |
| 19:39 | <AryehGregor> | Even if it had, wouldn't there still be Indonesian expatriates? |
| 19:40 | <Ms2ger> | Would you call them Indonesian people in that case? Presumably yes. |
| 19:41 | <AryehGregor> | Sure. |
| 19:41 | <TabAtkins> | Hixie: whitespace-collapse:discard isnt' yet implemented anywhere afaik. But it's definitely awesome, for precisely the case you mention. |
| 19:42 | <AryehGregor> | it does sound awesome. |
| 19:42 | <AryehGregor> | w3ztb0y, no, no one here is Indonesian. |
| 19:42 | <TabAtkins> | zcorpan, others: Selectors 4 introduced a way to match attributes case-insensitively a few weeks ago. |
| 19:42 | <Hixie> | TabAtkins: yeah once i saw it i realised it was what we were talking about back in 2003 or so |
| 19:42 | <Hixie> | TabAtkins: i still think we should get rid of the three properties and still only use white-space, with whatever keywords make sense |
| 19:42 | <Ms2ger> | Hixie, then it should be implemented in 5 years or so |
| 19:43 | <Hixie> | TabAtkins: since the vast majority of the combinations make no sense at all |
| 19:43 | <Ms2ger> | Send email to www-style |
| 19:43 | <AryehGregor> | TabAtkins, won't that force browsers to keep a separate case-normalized copy of every attribute value? Do they actually want to do that? |
| 19:43 | <Hixie> | Ms2ger: dude, you obviously aren't familiar with the csswg's specs' schedules (especially fantasai's...) |
| 19:43 | <Ms2ger> | The CSSWG apparently thinks you don't know about www-style |
| 19:44 | <Hixie> | by "the CSSWG" do you mean daniel? :-) |
| 19:46 | <TabAtkins> | Hixie: Nah, fantasai was making fun of you for saying that you weren't sure how to ask the CSSWG to handle ::paragraphs. |
| 19:46 | <Hixie> | what i wasn't sure about was whether there was some convenient way to just reassign the bug |
| 19:47 | <TabAtkins> | No, we don't use the bugtracker for feature requests. |
| 19:47 | <Hixie> | lame |
| 19:47 | <TabAtkins> | Those are solely handled on the mailing list. |
| 19:47 | <Hixie> | where are they tracked? |
| 19:47 | <Ms2ger> | The editor's heads |
| 19:47 | <Hixie> | i'm going to go back to "lame" |
| 19:47 | <Ms2ger> | Hmm |
| 19:47 | <TabAtkins> | Various places. We're consolidating to *tracking* on a bugtracker. |
| 19:48 | <TabAtkins> | s/a bugtracker/bugzilla/ |
| 19:48 | <Ms2ger> | That's a bit of a strange picture |
| 19:48 | <TabAtkins> | But discussion shouldn't ever happen on bugzilla. |
| 19:48 | <Ms2ger> | *The editors' heads |
| 19:48 | <TabAtkins> | /lunch |
| 19:49 | <Hixie> | well where discussion happens is not really my concern |
| 19:49 | <Hixie> | i just wanted to get the bug off my list :-P |
| 19:49 | <zewt> | when you run out of space, do you have to buy a razor |
| 21:01 | <TabAtkins> | Hixie: I think that wrapping and whitespace collapsing and newline collapsing are all distinct, and useful to control separately, so the separate properties are good. |
| 21:04 | jgraham | blinks |
| 21:05 | <TabAtkins> | jgraham: Don't stop blinking! |
| 21:06 | <jgraham> | TabAtkins: It's OK I think that I will keep blinking for as long as I keep getting email about the W3C test framework thing. |
| 21:07 | <jgraham> | It is being built with a set of assumptions that are quite alien to me |
| 21:07 | jgraham | should try to ignore it |
| 21:28 | <Hixie> | TabAtkins: i think giving authors switches where the majority of possible setting combinations are bad when we can enumerate the good ones is bad UI |
| 21:28 | <matjas> | is there a point in using @defer when you only use a single <script> and it’s at the bottom, right before </body>? |
| 21:28 | <Hixie> | not really |
| 21:30 | <matjas> | not really or not at all? |
| 21:30 | <matjas> | what is the point? |
| 21:31 | <Hixie> | there's no point that i can think of |
| 21:31 | <Ms2ger> | Being fancy! :) |
| 21:31 | <Hixie> | there are some subtle minor differences, but nothing useful i don't think |
| 21:35 | <matjas> | Hixie: it seems to make a difference in rendering, at least in Chrome |
| 21:35 | <matjas> | compare http://stevesouders.com/cuzillion/?c0=bi1hfff2_0_f&c1=bj1hfff2_0_f&t=1313008395 and http://stevesouders.com/cuzillion/?c0=bi1hfff2_0_f&c1=bj1hfft2_0_f&t=1313008412 |
| 21:36 | <matjas> | first test is blank until the script is loaded, second test (with @defer) displays the table “immediately”, then shows the last line after the script is loaded |
| 21:36 | <Ms2ger> | QoI |
| 21:36 | <Hixie> | that's a browser detail |
| 21:40 | <Hixie> | jgraham: 504ing |
| 22:20 | <jgraham> | Hixie: I really don't know what to do. The only message in the logs is about the ekkepalive connection being closed which afaict from the docs is an effect rather than the cause |
| 22:21 | <jgraham> | *keepalive |
| 22:21 | <Hixie> | any idea _which_ connection? |
| 22:21 | <Hixie> | when does the problem occur? |
| 22:28 | <jgraham> | Well I assume that what's happening is that the host has changed their setup to use nginx as a frontend for all requests and then send them to apache if needed. And sometimes the script takes a little bit too long and nginx gets bored waiting for a response |
| 22:28 | <Hixie> | aaah |
| 22:28 | <Hixie> | interesting |
| 22:28 | <Hixie> | no way to change the timeout? |
| 22:28 | <Hixie> | maybe make your script send stuff back regularly? |
| 22:29 | <Hixie> | i ignore what you send back, iirc |
| 22:29 | <Hixie> | oh no wait |
| 22:29 | <jgraham> | I'm looking to see if there is a timeout |
| 22:29 | <Hixie> | that's for the multipage thing |
| 22:29 | <Hixie> | i actually use your output |
| 22:29 | <Hixie> | you could send back something in a comment at the top that i then strip out |
| 22:29 | <Hixie> | <!-- progress... line 1000... line 2000... line 3000... |
| 22:29 | <Hixie> | or whatever |
| 22:34 | <jgraham> | I probably can |
| 22:34 | <jgraham> | But it is going to be a crazy hack |
| 22:35 | <Hixie> | up to you :-) |
| 22:35 | <Hixie> | on my end the entire pipeline is one long series of crazy hacks |
| 22:36 | <Philip`> | Only a long series, not a complex interconnected DAG? |
| 22:37 | <Philip`> | (or even a cyclic one) |
| 22:38 | <gsnedders> | Philip`: Then it wouldn't be a DAG |
| 22:38 | <jgraham> | gsnedders: That's presumably why he said "or even" |
| 22:38 | <Philip`> | I didn't mean to imply it would be |
| 22:40 | <jgraham> | Hixie: I think it is too late at night for crazy hacks right now. Hopefully I will wake up with a better idea than I have right now |
| 23:11 | <Hixie> | anyone know if the css rules define how to serialise rgba()'s alpha value? |
| 23:15 | <tantek> | as in how many decimal places? |
| 23:15 | <nimbu> | http://www.w3.org/TR/2003/REC-SVG11-20030114/masking.html#SimpleAlphaBlending ? |
| 23:16 | <nimbu> | (from http://www.w3.org/TR/2011/REC-css3-color-20110607/#alpha ) |
| 23:16 | <TabAtkins> | Hixie: They do not. |
| 23:16 | <Hixie> | hmm |
| 23:16 | <TabAtkins> | In general, serialization is a big pile of undefined crap. |
| 23:17 | <TabAtkins> | ;_; |
| 23:17 | <Hixie> | i thought anne had largely fixed that using my proposal |
| 23:18 | <tantek> | Hixie, he did |
| 23:18 | <TabAtkins> | anne hasn't done anything with it, as far as I know. cssom still just has an issue for both <color> and <number> |
| 23:18 | <tantek> | but then we introduced new things into CSS |
| 23:18 | <tantek> | like rgba |
| 23:18 | tantek | figures Anne will do something sensible. |
| 23:19 | <TabAtkins> | Now that fantasai and I are co-editors on the Values & Units spec, we intend to try and define serialization there. |
| 23:19 | <Hixie> | man i wish values & units would be implemented by browsers already |
| 23:19 | <Hixie> | i want vh and vw so bad |
| 23:20 | <Hixie> | "User agents must express the fractional part of the alpha value, if any, with the level of precision necessary for the value, when reparsed, to be interpreted as representing the same alpha value. |
| 23:20 | <Hixie> | " |
| 23:20 | <Hixie> | still doesn't quite define what the value should be |
| 23:21 | <Hixie> | e.g. if the underlying precision is a 2-bit precision (0%, 33%, 67%, 100%) it doesn't say whether to express 33% as 0.2, 0.3, or 0.4... |
| 23:22 | <Hixie> | and it doesn't disallow 0.33333333333333333333 |
| 23:22 | <Hixie> | well i guess it does disallow it |
| 23:22 | <Hixie> | since that level of precision isn't necessary |
| 23:22 | <Hixie> | hmm |
| 23:23 | <Hixie> | ok well we'll go with that for now |
| 23:24 | <TabAtkins> | Hixie: IE implements the viewport-relative uits. |
| 23:24 | <Hixie> | wow, really? |
| 23:24 | <Hixie> | go IE |
| 23:24 | <TabAtkins> | Yup. |
| 23:24 | <Hixie> | how about calc()? anyone got that yet? |
| 23:25 | <TabAtkins> | (Half-unfortunately, they did so unprefixed, which makes it somewhat more painful to change vm to vmin (so we can allow vmax)). |
| 23:25 | <Hixie> | i specced that like half a decade ago already |
| 23:25 | <TabAtkins> | Yes, Firefox and IE have it, Webkit is getting it. |
| 23:25 | <Hixie> | what's vm? |
| 23:25 | <zewt> | i wonder if the algorithm used by python for representing floats as strings is clearly specified |
| 23:25 | <TabAtkins> | vm = min(vh,vw) |
| 23:25 | <Hixie> | ah |
| 23:25 | <Hixie> | just add vb |
| 23:25 | <Hixie> | for "big" |
| 23:25 | <Hixie> | (if you really want it) |
| 23:25 | <Hixie> | (what's the use case for vb?) |
| 23:26 | <Hixie> | or just use min()! |
| 23:26 | <TabAtkins> | Dunno yet. I suspect there's one there, but I wouldn't add it yet. |
| 23:26 | <Hixie> | calc(min(1vh,1vw)) |
| 23:26 | <TabAtkins> | Yes, just using min/max is fine too. |
| 23:26 | <Philip`> | Clearly max should be abbreviated to "vx" |
| 23:26 | <TabAtkins> | And remember, min/max dont' ahve to occur within calc() anymore! |
| 23:26 | <Hixie> | ah, cool |
| 23:26 | <Hixie> | well, cool once implemented |
| 23:27 | <Hixie> | i really wish there was just a single css spec i could keep up with instead of this mass of drafts |
| 23:27 | <Hixie> | anyway |
| 23:27 | <Hixie> | i should take a week to learn where browsers are with css |
| 23:28 | <jamesr> | some 2.1, some 3, and some 4 |
| 23:28 | <jamesr> | of course |
| 23:28 | <TabAtkins> | You can just watch the yearly snapshots. |
| 23:29 | <Hixie> | when was the last one of those? |
| 23:29 | <TabAtkins> | That's by definition fairly behind, but still. |
| 23:29 | <TabAtkins> | last year. Fantasai's publihsing one per year now. |
| 23:29 | <Hixie> | link? |
| 23:29 | <TabAtkins> | Or rather, intends to (she just started last year). |
| 23:29 | <TabAtkins> | http://www.w3.org/TR/css-2010/ |
| 23:29 | <TabAtkins> | It's just the specs CR or later. |
| 23:30 | <Hixie> | oh |
| 23:30 | <Hixie> | well |
| 23:30 | <Hixie> | that's no good |
| 23:30 | <TabAtkins> | Sure, not right now. css-2011 will be quite a bit bigger. |
| 23:30 | tantek | entertains the idea of an automatically combined spec of latest less-than-year-old WDs + CRs + PRs + RECs of CSS3 modules |
| 23:30 | <Hixie> | what i want to know is what's implemented, and what's still fiction, and apparently what the prefixes are. |
| 23:30 | <TabAtkins> | Hixie: Yeah, that's something else. caniuse.com is a partial example |
| 23:31 | <tantek> | Hixie, putting things in one spec won't tell you "what's implemented, and what's still fiction", after all, HTML5 doesn't necessarily tell you what's implemented, and what's still fiction |
| 23:31 | <Hixie> | tantek: well we have hte little status boxes that try to tell you that |
| 23:31 | <Hixie> | tantek: but sure |
| 23:31 | <tantek> | Hixie, yeah, I like the status boxes |
| 23:31 | <Hixie> | tantek: it would, however, help me know what was out there |
| 23:32 | <tantek> | they're a definite improvement |
| 23:32 | <tantek> | I agree, to know "what was out there" |
| 23:33 | <tantek> | I think Eric Meyer complained rather vociferously on Twitter a few months ago about how hard it was for him to track down "all" the CSS3 specs and ascertain their relative levels of status (as opposed to claimed status in the document status section) |
| 23:33 | <Hixie> | yeah |
| 23:33 | <jamesr> | yeah i have no idea how to do that, and i'm pretty sure i've implemented some of it |
| 23:33 | <tantek> | I'm personally thinking of just starting to mark features as at risk or not based on implementation status |
| 23:33 | <tantek> | in the specs I edit |
| 23:33 | <tantek> | even before CR |
| 23:34 | <Hixie> | another way we could do this is by expanding the css section of platform.html5.org |
| 23:34 | <tantek> | I'm not sure centralized works |
| 23:34 | <tantek> | I think we have to upgrade what's expected of spec editors/authors |
| 23:34 | <tantek> | that scales better |
| 23:34 | <Hixie> | no argument from me there |
| 23:35 | <tantek> | is there a "how to" written up somewhere for how a spec editor could add the little status boxes? |
| 23:35 | <TabAtkins> | I'm happy to help out here. I don't think it's appropriate to add an "Implementation Status" section to the specs (it'll be outdated shortly after publication), but maintaining a wiki page or something like that is easy. |
| 23:35 | tantek | is into writing-up how-to's for spec editors. e.g. http://wiki.csswg.org/spec/cvs |
| 23:36 | <tantek> | Tab - I'm suggestion per property notes, just like the status boxes |
| 23:36 | TabAtkins | still references that regularly. |
| 23:36 | tantek | does too. |
| 23:36 | <Hixie> | TabAtkins: in the case of the html spec the "how to" is "write a server-side CGI script that interfaces with a MySQL database, then write a JS front-end that interfaces with the CGI script" |
| 23:36 | tantek | has run into CVS problems, and googled for the error and found that document. Scary. |
| 23:36 | <Hixie> | er |
| 23:36 | <Hixie> | tantek: ^ |
| 23:36 | <Hixie> | tab-completion fail |
| 23:36 | <TabAtkins> | tantek: Yes, I understand. They just can't be part of the spec itself. Having a script that grabs from something external is fine. |
| 23:36 | <Hixie> | oh, a pun! Tab-completion fail! |
| 23:36 | <TabAtkins> | Sigh. |
| 23:37 | <tantek> | lol |
| 23:37 | <zewt> | D: |
| 23:37 | TabAtkins | thinks he should go ahead and write that. |
| 23:37 | <tantek> | TabAtkins - that would be great if you added support for status boxes to CSS specs |
| 23:37 | <Hixie> | how widespread is CORS support? |
| 23:37 | tantek | would use it in his specs. |
| 23:37 | <Hixie> | i don't mind hosting the CGI side of this |
| 23:38 | <Hixie> | and then the scripts could just pull from that |
| 23:38 | <Philip`> | If there were status boxes in the CORS spec you'd be able to answer that yourself |
| 23:38 | <Hixie> | tis true |
| 23:38 | <Hixie> | we have a bootstrapping problem |
| 23:38 | <tantek> | I'll gladly test it. For now I've been putting in various notes into css3-ui. And yeah, it's ad hoc but I'm capturing it there because there is an expectation of imminent return to LCWD(2) and then CR (thus tracking "at risk" is important) |
| 23:39 | Hixie | wonders where the code for the status cgi script is... |
| 23:40 | <TabAtkins> | Hixie: No need for a cgi. Just a json file would work. |
| 23:40 | <Hixie> | ah, here we go |
| 23:40 | <Hixie> | TabAtkins: how would you update it? |
| 23:40 | <TabAtkins> | ...by editting it? |
| 23:40 | <Hixie> | oh |
| 23:40 | <Hixie> | well that's lame |
| 23:41 | <Hixie> | the html spec can be edited by anyone who has sent feedback! |
| 23:41 | <Hixie> | crowd-sourcing |
| 23:41 | <Hixie> | it's fashionable |
| 23:41 | <TabAtkins> | Bah. |
| 23:41 | <Hixie> | hah |
| 23:41 | <TabAtkins> | If we check it into dev.w3.org, then anyone with CVS access can update it. |
| 23:42 | <Hixie> | ok well in that case, no need for me to update the status.cgi script, excellent |
| 23:42 | TabAtkins | is squeamish around SQL these days, anyway. |
| 23:43 | <Hixie> | hah |
| 23:43 | <Hixie> | did it bite you? |
| 23:43 | <Philip`> | How many people have made substantial updates to the HTML spec stat boxes? |
| 23:43 | <TabAtkins> | Hixie: I've been bitten by corruption before. Binary formats are the devil. |
| 23:43 | <Hixie> | ah well yeah |
| 23:43 | <Hixie> | i make daily text-form backups |
| 23:44 | <Hixie> | Philip`: please hold |
| 23:44 | <Philip`> | (If it's only like two people then a static file seems good enough) |
| 23:46 | <Hixie> | Philip`: how many changes is "substantial" |
| 23:47 | <Peter`> | ideally we'd have one central place to track implementation status of web platform features |
| 23:47 | <Peter`> | combining the stability of the specification (or rather, chapter in the specification) with the various rendering engines and associated browsers |
| 23:47 | <Peter`> | some kind of (semi-)official caniuse.com |
| 23:48 | <Philip`> | Hixie: Probably the shape of the distribution is more important than the numbers |
| 23:50 | <Hixie> | the answer seems to be "4". http://junkyard.damowmow.com/493 |
| 23:51 | <TabAtkins> | 4 besides you? I'd rate zcorpan as substantial. |
| 23:51 | <Hixie> | 4 including me |
| 23:51 | <Hixie> | zcorpan == simonp |
| 23:51 | <TabAtkins> | Yes. |
| 23:51 | <TabAtkins> | Oh, wait, didn't realize he was further down the list as well. |
| 23:52 | <Hixie> | (the database stores every change so i can nuke a griefer pretty easily by just deleting their changes and the status just reverts to the previous state) |
| 23:54 | <Philip`> | Looks like the long tail of contributors is actually a pretty small tail, so the crowdsourcing probably isn't hugely useful |
| 23:55 | tantek | likes wikis for crowd sourcing that kind of thing. |
| 23:56 | <hober> | indeed |
| 23:56 | <Hixie> | in this particular instance i think the real trouble is discoverability |
| 23:56 | <Hixie> | first you have to send feedback, then you have to log in, then you have to magically guess that you have to alt-double-click |
| 23:56 | <Hixie> | it's pretty obscure |
| 23:56 | <Hixie> | i've had several people ask me about it |
| 23:57 | <Hixie> | including a number of people trying to update the values by submitting bugs |
| 23:57 | <Hixie> | i think we could massively improve matters just by making it easier to update values |
| 23:57 | <Philip`> | Alt-double-click doesn't even work in my browser (since the window manager interprets it as the start of drag-moving a window), though I think shift-alt-double-click worked okay |
| 23:58 | <Hixie> | yeah i worked pretty hard to find a test that worked everywhere with some combination or other |
| 23:58 | <Hixie> | but anyway, we could e.g. have non-logged-in be able to submit values that displayed their IP instead of them having to log in first |
| 23:58 | <Hixie> | and we could put an "edit" link in the boxes |
