2013-04-01 [00:45:50.0000] hello to everybody [01:57:08.0000] Ah, geocities is restored to the CSSWG [03:29:30.0000] Ms2ger: Wat? They do geocities-css? [03:30:01.0000] http://dev.w3.org/csswg/css-values/ [03:33:10.0000] Ah, nice. I have a site like that somewhere on the intarwebs too, but that's in some cobweb-ridden place where the sun never shines. [03:37:06.0000] lol, NO FEAR [04:15:25.0000] Ms2ger: Looks nice, but needs more background MIDIs [05:26:57.0000] http://lists.w3.org/Archives/Public/public-lod/2013Apr/0014.html [07:57:37.0000] wow, the webkit api on android is in the dark ages (at least whatever version the fire has) [07:58:42.0000] seems like it doesn't even have history.replaceState [08:44:40.0000] good morning, Whatwg! [08:54:22.0000] jgraham: and yet surprisingly few issues, all told :-) [08:57:42.0000] Hixie_, ... yet :) [09:02:04.0000] Can someone give me a pointer regarding a visual (WYSIWYG) editor on linux/BSD for changing a CSS template? [09:02:07.0000] odinho: Glad you caught the NO FEAR easter egg. ^_^ [09:02:40.0000] <3 generated content [09:04:02.0000] TabAtkins: I saw it in the commit :) [09:19:48.0000] TabAtkins: "the" it's quite a few places. Some of it fits, some of it not :-) [09:20:24.0000] odinho: It's just added on p::nth-child(10n+5)::after, so it'll show up everywhere. [09:58:08.0000] Yeah, saw that. :] [09:58:17.0000] But the first time it came up it really made sense :] [09:58:39.0000] Hahaha [09:58:58.0000] So I wondered if it was actually part of the spec :P [09:59:11.0000] Also: oh man, I missed a lot of badass updates to NearlyFreeSpeech over the holiday. ipv6, https, and cron jobs! [10:22:02.0000] anyone around with IE? [10:22:30.0000] i just need to know which OPTION elements in the DOM tree view in http://software.hixie.ch/utilities/js/live-dom-viewer/saved/2190 have #text children [10:22:40.0000] in Chrome it's all of them, in Firefox it's all but the first [10:27:27.0000] Hixie_: In IE9, I see #text on the last two [10:27:34.0000] thanks [10:27:42.0000] three browsers, three behaviours [10:27:43.0000] lovely [10:28:09.0000] IE's and Chrome's can be described in terms of default values [10:28:14.0000] Firefox's... [10:29:33.0000] /me wanders in [10:29:47.0000] What's up? [10:30:22.0000] Firefox has weird behavior for the Option constructor. [10:30:39.0000] It distinguishes between no argument and an empty string. [10:30:46.0000] It just looks at the number of arguments, sure [10:30:57.0000] /me looks at the new code [10:32:03.0000] Same thing [10:32:22.0000] Hixie_, I'm pretty sure we're willing to match Chrome here [10:46:41.0000] Ms2ger: always create a text node? i was thinking of going with the IE behaviour of not creating the empty text node [10:47:00.0000] the second argument is a weirder one, looks like we'll always have to check number of arguments for that one [10:47:22.0000] Always creating would be less code for us [10:47:30.0000] /me asks bz [10:48:25.0000] it's an additional if (val != '') [10:48:49.0000] Right [10:48:51.0000] so let's put the implementation complexity issue aside and look at what authors want :-P [10:48:59.0000] That's the "less code" :) [10:49:22.0000] I suspect (1) most authors don't care at all, but (2) more consistency in produced result is better. [10:51:06.0000] having an extra object around that has no purpose seems pretty weird [10:51:26.0000] you don't get this object if you use createElement() [10:51:30.0000] It's purpose is to contain the text contents! [10:51:36.0000] or if you use innerHTML = '' [10:51:44.0000] Yeah, and you can't fill anything with text with createElement() either. [10:51:56.0000] the text contents can also go in the .label [10:52:18.0000] e4h-style or quasis wouldn't give you an empty text node either [10:52:32.0000] it would be specific to this constructor [10:52:42.0000] I guess textContent wouldn't create one either [10:52:47.0000] yup [10:52:50.0000] in fact it would remove it [10:55:48.0000] Hixie_, alright, you get a "meh" from bz [10:59:14.0000] Ms2ger: heh [11:00:09.0000] Hixie_: optional boolean selected -> could make that default to false and simply set selectedness to it [11:03:19.0000] annevk: already done, see next checkin [11:03:27.0000] well, the default is false [11:03:34.0000] the text isn't changed [11:04:18.0000] can you default to a value if the next argument does not default to a value? [11:04:38.0000] i hope so [11:05:47.0000] hmm, oh well, not going to read up on that today [11:09:19.0000] no optional value is presumably just the same as " = " [11:09:24.0000] annevk: Most languages with positional argument disallow that - once one argument gains a default value, everything following has to have one (because otherwise it's not really a default value at all, since you're forced to fill it in). [11:09:29.0000] s/attributes/argument/ [11:09:46.0000] JS is a bit different, of course, since it fills in missing argument with undefined. [11:10:18.0000] (Python is also different, because all positional arguments are *also* keyword arguments.) [11:10:21.0000] TabAtkins: why are you forced to fill it in? i'm confused [11:10:33.0000] oh you mean non-optional arguments [11:10:41.0000] with default values [11:10:52.0000] as opposed to optional arguments with default values like here [11:10:52.0000] ok [11:11:35.0000] Hixie_: In most languages, there's no such thing as "non-optional arg with default values". ^_^ [11:11:40.0000] right [11:11:57.0000] well, there is [11:12:00.0000] it's all of them [11:12:16.0000] it's more that there's no such thing as optional arg with default values [11:12:22.0000] If it's non-optional, that presumably means that you must fill it in. So, when would the default value ever happen? [11:12:26.0000] lolwut? [11:12:34.0000] there's three states [11:12:43.0000] present, absent but has a default value, and absent without a default value [11:12:58.0000] some languages only have two of these [11:13:06.0000] which i assume is what you mean [11:13:07.0000] right? [11:13:11.0000] whereas here we have all three [11:13:11.0000] Yes. [11:13:24.0000] Well, because in most, the third case is an error. [11:13:29.0000] right [11:13:40.0000] or it's equivalent to an overload [11:13:48.0000] or rather, done as an overload [11:14:05.0000] so you can't have arguments after an optional one without a default, since it's not there to be after [11:14:09.0000] anyway [11:14:11.0000] Yeah. [11:29:53.0000] what is with people filing bugs whose values consist only of copy-and-pasted text from the spec? [11:29:56.0000] i don't get it [11:30:02.0000] i wonder how to catch that [11:30:21.0000] They forgot to put "^^^ THIS" afterwards. [11:30:22.0000] i'm guessing document.documentElement.textContent.match() isn't a good plan on a 5MB document... [11:30:43.0000] TabAtkins: i mean e.g. https://www.w3.org/Bugs/Public/show_bug.cgi?id=21488 [11:31:03.0000] Yeah, I know. I was being silly. ^_^ [11:31:10.0000] But seriously, that's just dumb. [11:32:17.0000] i don't understand how it happens [11:48:49.0000] TabAtkins: var() fallbacks can contain commas, right? [11:49:51.0000] either way, it should be explicit with both normative text and an example, not just implied by "same syntax as custom properties" [11:51:38.0000] And if it can, is it really useful to have the parser split function arguments on commas? Isn’t it easier if functions arguments are a single list of tokens, and everything (other than var()) can use the "Parse a comma-separated list" algorithm? [12:21:57.0000] SimonSapin: Maybe? [12:22:07.0000] Oh, and yes to your first question. [12:23:26.0000] arguments are a single list of tokens in tinycss [12:32:35.0000] Yeah, that probably makes more sense. [12:50:49.0000] Hi! Do you know when meta tag I added yesterday will be recognized by html5 validators? [12:55:08.0000] SoWhat: what exacctly did you add? [13:01:06.0000] web_author meta tag to the wiki [13:01:44.0000] http://www.metatags.info/meta_name_webauthor [13:03:37.0000] Hixie_: big thanks for adding the "Tag omission in text/html" stuff [13:04:16.0000] /me looks at SoWhat link [13:05:24.0000] SoWhat: why not just use name=author ? [13:05:35.0000] author is totally different [13:05:51.0000] http://www.metatags.info/meta_name_author [13:07:32.0000] MikeSmith: np [13:08:34.0000] SoWhat: The description of "author" and "webauthor" appear to be copy-pasted with only minor edits. I can't see any difference in them. [13:10:08.0000] SoWhat: what TabAtkins said. This http://www.metatags.info/ site doesn't really look authorative. It sorta looks like on guy's personal wish list or something. [13:10:50.0000] MikeSmith, when have you ever seen an authorative .info site? [13:11:15.0000] then again, meta tags themselves aren't very authoritative, so perhaps that makes sense [13:11:46.0000] annevk: any idea if adam is planning on updating WebOrigin to use the new URL spec? [13:12:09.0000] you are probably right.. maybe I should invent developer meta tag instead of using web author? [13:12:26.0000] SoWhat: what problem are you trying to solve? [13:12:30.0000] tantek: true about .info sites [13:12:45.0000] annevk: i'm in particular interested in how the host portion of origins, as used in document.domain, will handle IPv6 urls [13:13:15.0000] Hixie_: I just want to put my credits in source of websites which I have created [13:13:34.0000] Hixie_: "[::1]" ? [13:14:13.0000] Hixie_: I don't know Adam's plans. I've been thinking of putting all origin-related requirements into Fetch [13:14:46.0000] Some people, when confronted with a problem, think "I know, I'll use a meta tag." Now they have a meta problem. [with apologies to jwz] [13:15:54.0000] Hixie_: oh wow I see you added the attribute short descriptions too [13:16:02.0000] thanks thanks thanks [13:16:07.0000] this is great [13:19:15.0000] so how is that done? are they added dynamically or via some template? [13:19:21.0000] they didn't appear in the diff [13:19:53.0000] so what do you think about adding ? [13:20:38.0000] MikeSmith: my pleasure :-) [13:20:45.0000] MikeSmith: you using them for something specific? :-) [13:20:58.0000] SoWhat: what's wrong with a ? [13:21:55.0000] there is no standart for web crawlers to index [13:22:15.0000] you didn't say anything about crawlers indexing anything :-) [13:22:16.0000] I want all my sites to be found in google later [13:22:30.0000] ah, well, there's no way to guarantee that [13:22:49.0000] but maybe look at google authorship markup? [13:23:08.0000] gotta go, hopefully tantek or someone will take over again :-) [13:23:09.0000] bbiab [13:23:57.0000] Hixie_: :) I don't want to see my face next to these sites :D [13:24:45.0000] there is huge difference between person who is writing content for website and person/company who created website [13:25:09.0000] SoWhat - the fastest way to get your sites found in google is by publishing with open standards, in particular make sure each page has its updates posted in the HTML, with permalinks, and an equivalent feed in Atom, and the updates use PubsubHubbub to ping the Google hub. [13:25:33.0000] using rel author is nice but not required [13:25:55.0000] http://microformats.org/wiki/rel-author [13:26:14.0000] TabAtkins: Separate ident/ident-rest states, isn’t that redundant with checking "starts with an identifier" before switching to the ident state? [13:26:15.0000] ehh [13:26:38.0000] tantek: do you know Wordpress? [13:26:45.0000] ie. could remove ident, rename ident-rest to ident, keep equivalent behavior [13:27:06.0000] In wordpress you have lots of authors who are writing their posts [13:27:26.0000] TabAtkins: same with number/number-rest, maybe [13:27:51.0000] and in each post wordpress writes its author in rel=author [13:27:53.0000] SimonSapin: Probably. I was overly cautious with some of my states in the initial draft. [13:28:21.0000] well, the "starts with an identifier" arrived much later than ident/ident-rest [13:28:42.0000] but now that it’s there, I guess the rest can simplify [13:29:11.0000] Yeah, I haven't fully simplified a lot of the stuff related to the "start with..." stuff. [13:32:17.0000] SoWhat - yes that makes sense for a multi-author / group blog [13:32:53.0000] it's best if they use rel=author to link from their posts to their own websites, and then use rel=me to link from their own website to their various social network profiles, Twitter, FB, G+, etc. [13:33:33.0000] tantek: but there is only one person/company who has written code for this blog and I am this person for lots of websites [13:34:25.0000] shouldn't need any "code" other than the