| 00:31 | <zewt> | heh, that's a bit user-hostile: the default firefox download page is the mobile page, even though that's obviously not what the overwhelming majority of people want |
| 00:32 | <zewt> | <- not very forgiving of sites intentionally giving people something other than what they want due to ulterior motives |
| 00:34 | <zewt> | strange: <script>history = 1; console.log(history);</script> logs the History object in FF, ignoring the assignment entirely; doesn't happen for "URL", though |
| 00:34 | <zewt> | doesn't happen in webkit; i'd think that would break any site that has a "history" global (which doesn't seem improbable) |
| 00:37 | <zewt> | gripe gripe they totally broke the address bar in ff15 (or maybe 14) |
| 00:37 | <zewt> | this is why i avoid updating firefox--each update means spending time hunting down ways to disable their latest bad ideas in about:config |
| 01:18 | <zewt> | (address bar fix: toggle browser.urlbar.autoFill off) |
| 06:06 | <hsivonen> | Firefox has 4 Korean entries in the Character Encoding menu. Chrome has 1. |
| 06:06 | <hsivonen> | Is Chrome’s Korean an autodetect or is one Korean encoding enough for the legacy Web? |
| 06:08 | <hsivonen> | IE also has just one Korean encoding |
| 06:09 | <hsivonen> | What does the “User Defined” encoding mean in Firefox and IE? |
| 06:34 | <MikeSmith> | hsivonen: I would think that all browsers should at least have two: for EUC-KR and ISO-2022-KR |
| 06:35 | <MikeSmith> | I can't remember ever seeing those other two anywhere else |
| 06:35 | <MikeSmith> | UHC and JOHAB |
| 06:36 | MikeSmith | asks somebody |
| 06:37 | <annevk> | the latter two are not part of the Encoding Standard and definitely not in all browsers |
| 06:38 | <a-ja> | http://support.microsoft.com/kb/170557 |
| 06:41 | <MikeSmith> | annevk: Sangwhan says "Johab is a encoding of its own, which in reality probably doesnt have much of a usecase" |
| 06:44 | <MikeSmith> | a-ja: thanks for the link |
| 06:44 | <MikeSmith> | but weird that page mentions UHC and JOHAB but not EUC-KR and ISO-2022-KR |
| 06:44 | a-ja | asked a search engine |
| 06:45 | <MikeSmith> | and whoah whatta they have going on with the font for the page title |
| 06:47 | <MikeSmith> | hsivonen: our friend Sangwhan from Opera suspects those are there maybe just because a Korean contributor contributed a patch |
| 06:48 | <a-ja> | keep in mind that's a win95-era article |
| 06:48 | <MikeSmith> | ah yeah |
| 06:49 | <MikeSmith> | hmm FF mobile has a much simpler character-encoding menu |
| 06:49 | <MikeSmith> | at least nightly does (mobile) |
| 06:49 | <MikeSmith> | I guess that's to be expected |
| 07:17 | <kmbarnhart> | Hi |
| 07:17 | <kmbarnhart> | I was reading the WebVTT spec at http://dev.w3.org/html5/webvtt/#webvtt-file-using-only-nested-cues |
| 07:18 | <kmbarnhart> | and the example used to show a only nested cue seems to be wrong |
| 07:19 | <kmbarnhart> | specifically the rule "whose cues all have an end time offset x greater than or equal to the end time offsets of all the cues whose start time offsets are less than x is said to be a WebVTT file using only nested cues" |
| 07:19 | <Ms2ger> | Oh dear |
| 07:20 | <kmbarnhart> | the fifth cue in the following example breaks the rule as far as I can tall, although it is quite late |
| 07:22 | <MikeSmith> | matjas: I wrote up a patch the htmlparser code to bring error reporting for ambiguous ampersands into line with the current spec (for the validator bug you filed) |
| 07:22 | <MikeSmith> | no in Henri's hands for review |
| 07:22 | <MikeSmith> | *now |
| 07:48 | <matjas> | MikeSmith: awesome, thank you :) |
| 08:01 | <jgraham> | I notice that "actualy write tests" is slowly dropping off the Core Mobile radar as they realise it's quite a lot of real work |
| 08:15 | <odinho> | jgraham: maaan |
| 08:20 | <jgraham> | odinho: It's not that surprising. For most players writing tests is against their short-term interests (consumes resources, doesn't have short term gains), but in their long term interests (makes the platform better) |
| 08:20 | <jgraham> | So there is way more talking about doing it than there is actually doing it |
| 08:20 | <jgraham> | This is pretty much textbook tragedy of the commons |
| 08:27 | <odinho> | jgraham: Well, they should employ some remote people to just do tests then. We do a lot of them. It's fun-fun-fun. Sometimes. |
| 08:29 | <jgraham> | odinho: I think "they should employ" is the problem :) |
| 08:31 | <jgraham> | Traditionally the way you solve problems liek this is via socialism |
| 08:31 | <hsivonen> | at MS they have a tester for each developer, right? |
| 08:31 | <jgraham> | Really? |
| 08:31 | <jgraham> | Doing *what*? |
| 08:31 | <hsivonen> | so goes the legend |
| 08:31 | <hsivonen> | dunno |
| 08:31 | <jgraham> | Checking their canteen food for poison? |
| 08:32 | <odinho> | They have a brag video about all their hardware and big testsuites at least. |
| 08:32 | <hsivonen> | but I think I’d not having to write tests myself and someone doing it for me |
| 08:32 | <jgraham> | hsivonen: There is am important word missing there |
| 08:32 | <hsivonen> | *I’d like not… |
| 08:33 | <jgraham> | It is kind of the Opera system |
| 08:33 | <jgraham> | It has advantages and disadvantages |
| 08:33 | <jgraham> | Like most things |
| 08:34 | <jgraham> | The main disadvantage is that knowing less about the code can make you oblivious of special cases in the implementation that are unusually likely to contain bugs |
| 08:36 | <jgraham> | But overall I think it is a functional system that is rather good at producing tests that can be contributed to the commons |
| 08:37 | <odinho> | Sometimes joaoe tells me "make a test for X" for some special case. Happily more often than once I already had the test :] |
| 08:38 | <jgraham> | Anyway, the communist solution would be for W3C to employ a dedicated platform test team |
| 08:39 | <jgraham> | And encourage them to work with implementors |
| 08:39 | <jgraham> | Not sure how it would add up financially though |
| 08:40 | <hsivonen> | they already hired darobin, so they are already on the slippery slope |
| 08:40 | <jgraham> | Not really to do that job though |
| 08:41 | <odinho> | Guess that's form "test platform platform team" |
| 08:41 | <jgraham> | There is also an advert for a QA manager |
| 08:41 | <jgraham> | (so if anyone who is first-langauge Japanese and is reading this wants a job, you might want to consider that) |
| 08:42 | <jgraham> | But that also doesn't mention actually developing *tests* anywhere |
| 08:42 | <jgraham> | It says "testing frameworks" |
| 08:43 | <jgraham> | (oh sorry, just fluent Japanese, not necessarily first-language) |
| 08:43 | <jgraham> | +in |
| 08:43 | <MikeSmith> | jgraham: we can hire somebody for that job outside of Japan |
| 08:44 | <MikeSmith> | does it really say fluent japanese is required? |
| 08:44 | <MikeSmith> | if so, it shouldn't |
| 08:44 | <jgraham> | Requirements: [...] Must be fluent in English as the working language of the W3C and Japanese as the first language in Japan where the Keio host is located; |
| 08:44 | <MikeSmith> | stupid |
| 08:45 | <MikeSmith> | somebody added that from when I reviewed it |
| 08:45 | <MikeSmith> | anyway, ignore that part |
| 08:45 | <hsivonen> | the UTF-8 BOM is like namespace. I still need to look the right hex values up every time. |
| 08:46 | <SamB_MacG5> | does it have to do with japanese typesetting/layout? |
| 08:46 | <MikeSmith> | jgraham: lemme put is this way, we can hire somebody to do the same work outside of that Keio position altogether |
| 08:46 | <MikeSmith> | SamB_MacG5: no, not in any way |
| 08:47 | SamB_MacG5 | was guessing at an explanation for the "fluent japanese" requirement having snuck in ... |
| 08:47 | <MikeSmith> | SamB_MacG5: that was almost certainly added just for political reasons |
| 08:48 | <MikeSmith> | to make it difficult for anybody outside of Japan to apply for it |
| 08:49 | <MikeSmith> | anyway, if anybody is interested in leading testing work HTML and related stuff at W3C, feel free to contact me directly if you want |
| 08:50 | <MikeSmith> | if we have the right person I can deal with jumping through the hoops |
| 10:07 | <hsivonen> | annevk: do you happen to recall if there are known XSS vulnerabilities with ISO-2022-JP? |
| 10:41 | <annevk> | hsivonen: only if you don't decode it properly |
| 11:07 | <hsivonen> | I’m very much not a fan of ISO-2002-*. I’d rather not have to learn about it. |
| 11:11 | <MikeSmith> | 21 people in the new SysApps WG, and 13 of them are from Intel |
| 11:11 | <MikeSmith> | https://www.w3.org/2000/09/dbwg/details?group=58119&public=1&order=org |
| 11:15 | <smaug____> | :) |
| 11:17 | hsivonen | hadn’t realized the spec now says that encodings that can encode non-markup as US-ASCII markup bytes are ASCII-compatible |
| 11:17 | <hsivonen> | looks like now to be ASCII compatible, the encoding only needs to encode markup as ASCII bytes |
| 11:18 | <hsivonen> | i.e. also encoding something else as the same bytes isn’t banned |
| 11:27 | <MikeSmith> | hmm I see the spec now says the colon character in time offsets is optional |
| 11:28 | <annevk> | hsivonen: in the Encoding Standard universe only UTF-16 is not ASCII-compatible |
| 11:29 | <annevk> | hsivonen: I filed a bug on HTML to be updated to reflect the terminology of that, since the complete picture becomes more coherent |
| 11:31 | <hsivonen> | annevk: do you know what x-userdefined is |
| 11:31 | <hsivonen> | ? |
| 11:31 | <annevk> | last I checked that depends on the context :/ |
| 11:32 | <hsivonen> | what can it be? |
| 11:32 | <annevk> | I haven't really figured that one out completely yet unfortunately |
| 11:32 | <hsivonen> | is it always ASCII-compatible? |
| 11:32 | <annevk> | e.g. in XHR I think it's iso-8859-1 (not windows-1252), but it depends on the implementation |
| 11:32 | <annevk> | hsivonen: I think in general it might be whatever the user prefers, which could be utf-16 I suppose (though maybe implementations should forbid that) |
| 11:33 | <hsivonen> | sigh. who comes up with this stuff? |
| 11:34 | <annevk> | Brendan's old colleagues? |
| 11:35 | <hsivonen> | hmm. what’s x-jis0208 and has it been removed from gecko while I wasn’t looking? |
| 11:36 | <hsivonen> | ooh. it’s a font encoding |
| 11:37 | <annevk> | jis0208 is a Japanese index |
| 11:37 | <annevk> | used by all Japanese encodings |
| 11:37 | <annevk> | dunno about that label though |
| 11:38 | <hsivonen> | I like it how Gecko omits aliases with a colon in them, because the Java property file format does not allow colons in property keys |
| 11:48 | <hsivonen> | annevk: hmm. maybe we should define x-user-defined as zero-extending the bytes to Unicode (i.e. what ISO thinks ISO-8859-1 is) |
| 11:49 | <hsivonen> | looks like we have XHR tests depending on that |
| 11:51 | <hsivonen> | oops no. it seems we map the high bytes to the PUA |
| 11:51 | <hsivonen> | anyway, it's ASCII-compatible |
| 12:04 | <karlcow> | https://developer.mozilla.org/en-US/docs/DOM/XMLHttpRequest/Using_XMLHttpRequest |
| 12:04 | <karlcow> | >// retrieve data unprocessed as a binary string |
| 12:04 | <karlcow> | >xhr.overrideMimeType("text/plain; charset=x-user-defined"); |
| 12:06 | <karlcow> | http://stackoverflow.com/questions/6986789/why-are-some-bytes-prefixed-with-0xf7-when-using-charset-x-user-defined-with-xm |
| 12:07 | <annevk> | hsivonen: but does it work that way for HTML documents? or is it treated as not recognised there? |
| 12:07 | <hsivonen> | annevk: it’s blacklisted for <meta> but it should work with HTTP charset |
| 12:07 | <annevk> | dear god |
| 12:07 | <karlcow> | http://www.alanflavell.org.uk/charset/user-defined.html |
| 12:08 | <karlcow> | a page from 1994 |
| 12:08 | <hsivonen> | I just wrote a patch to remove the <meta> blacklisting, since it’s not really dangerous |
| 12:09 | <hsivonen> | the bogus <meta> blacklisting is entirely my fault |
| 12:10 | <annevk> | so it works in XML too? |
| 12:10 | <hsivonen> | annevk: I don’t see why it would not |
| 12:10 | <annevk> | I can add whatever you want to the Encoding Standard |
| 12:10 | <annevk> | that is, next week |
| 12:10 | <hsivonen> | I’m not sure if I want to add it or try to kill it |
| 12:10 | <hsivonen> | oh, except we can’t kill it for XHR, I suppose |
| 12:10 | <hsivonen> | sigh |
| 12:12 | <hsivonen> | annevk: by code inspection, the ASCII range maps to ASCII and the bytes from 0x80 (inclusive) onwards map to U+F780 (inclusive) and onwards |
| 12:15 | <annevk> | eww |
| 12:16 | <annevk> | I wonder if all browsers do it that way |
| 12:16 | <hsivonen> | annevk: the SO question karlcow linked to suggests WebKit does the same |
| 12:18 | <karlcow> | if we could at least do http://s3.amazonaws.com/files.posterous.com/temp-2012-10-02/hecCsnDHmwksACCpczqbdodzCyJBcyFDsHFoIIgCEsJGlwzFljtHnprDizbH/Submariner53ButtonCrop.jpg.scaled1000.jpg?AWSAccessKeyId=AKIAJFZAE65UYRT34AOQ&Expires=1349266594&Signature=O3Txcgy64Qd5%2FQOdYJyq540PO80%3D |
| 12:18 | <karlcow> | but no |
| 12:18 | <karlcow> | :) |
| 12:18 | <hsivonen> | karlcow: I get AccessDenied in unstyled XML |
| 12:18 | <karlcow> | ah! |
| 12:19 | <karlcow> | the finger on the button here http://4cp.posterous.com/bill-everett-sub-mariner-53-page-17-1972# |
| 12:20 | <hsivonen> | still, working on encodings is nicer than working on browsing contexts |
| 12:27 | <Ms2ger> | hsivonen, fwiw, the no-colon-because-.properties has been a lie for a while, they actually work |
| 12:27 | <annevk> | hsivonen: okay, I have tried to avoid mapping to PUA, but if everyone does it maybe I should stop caring about that and just write it down |
| 12:28 | <annevk> | Peter Krefting, Opera's encoding guy, does not like PUA at all; Opera almost always uses FFFD instead |
| 12:29 | <annevk> | I sided with that moral high ground for now, but maybe that's not worth it |
| 12:31 | <hsivonen> | In principle, it would be more wrong *not* to map “user-defined” gunk to PUA |
| 12:31 | <hsivonen> | if you are mapping it to *something* |
| 12:31 | <hsivonen> | i.e. using U+0080… would be more wrong |
| 12:37 | <annevk> | I guess for this particular label you may be correct |
| 12:37 | <annevk> | label/encoding* |
| 12:44 | <blade3> | Hi Everyone, Ive question about Meta name="title" this is a "good" or "bad" meta name ? |
| 12:46 | <Ms2ger> | Bad, just use <title> |
| 12:47 | <blade3> | ho ok thanks alot! |
| 12:47 | <Ms2ger> | Np |
| 12:48 | <jgraham> | annevk: There goes my plan to hack Opera to replace all instances of the letter U+004F to the PUA and ship a font with a single PUA glyph "OPERA LOGO O" |
| 12:49 | <blade3> | why the *** joomla set automatically meta name="title" ><" |
| 12:49 | <jgraham> | Uh, I think s/replace/map/ is one way to make that sentence make sense |
| 12:52 | <zcorpan> | annevk: i meant to cc you on http://lists.w3.org/Archives/Public/public-whatwg-archive/2012Oct/0032.html |
| 12:52 | <hsivonen> | blade3: content management systems seem to manage more when they emit more metadata |
| 12:55 | <blade3> | ^^ |
| 13:28 | <blade3> | I've another question.. about "rel" attribut.. |
| 13:29 | <blade3> | it's not a free attribut ? i've something like rel="albums_201" but w3c doesn't like -_-" |
| 13:30 | <blade3> | on <a> tag : Bad value albums_201 for attribute rel on element a: Keyword albums_201 is not registered |
| 13:31 | <blade3> | i need the rel for my jquery modalbox :-/ how I can fix it? |
| 13:31 | <Ms2ger> | The only allowed values are here: http://www.whatwg.org/specs/web-apps/current-work/multipage/links.html#linkTypes |
| 13:32 | <Ms2ger> | You can use the hidden input, I guess |
| 13:34 | <fkm> | Or the data-attribute. |
| 13:36 | <Ms2ger> | fkm, oh, does it support that? I didn't immediately see that in the documentation |
| 13:36 | <blade3> | fkm: like : <a data-attribute="bleh"></a> ? |
| 13:39 | <fkm> | Ms2ger: Well. Javascript does. So you could do $('whatever').get(0).dataset. Or am I confusing something here? |
| 13:40 | <Ms2ger> | fkm, he's got a library that looks at the rel attribute, or a hidden input, or something to do with forms |
| 13:41 | <fkm> | Well. He could modify it. That's what I did to the Slimbox2 so it would pass the validator. |
| 13:41 | <blade3> | it's not a form. it's just a list of <a><img ...></a> |
| 13:42 | <blade3> | and i've many a rel="albums" to generate a albums with a[rel="albums"] tag |
| 13:43 | <fkm> | blade3: Have a look at the Slimbox part http://dev.snemelk.ch/gallery/README |
| 13:44 | <blade3> | ok i will look that |
| 13:47 | <blade3> | hmm nice, I'll modify my jquery and I'll try.. I come back soon to tell you if it's working |
| 13:59 | <annevk> | zcorpan: hmm yeah, well, browsers are shipping with URL already... |
| 13:59 | <annevk> | zcorpan: unless someone tells me to rename it, I'm not going to care much I think |
| 14:00 | <Ms2ger> | We broke Yahoo maps because of URL at some point, IIRC |
| 14:00 | <Ms2ger> | But bz told me the interface was fine |
| 14:03 | <Ms2ger> | https://www.w3.org/Bugs/Public/show_bug.cgi?id=15765 |
| 14:04 | <zcorpan> | hey, past me suggested renaming :-P |
| 14:06 | <zewt> | yeah, i'm curious as to which discussion led people to decide that the window.URL was OK for web-compat, that's indeed an odd one |
| 14:07 | <zcorpan> | results for "CSS": web200904: 30, stevef-all: 1 |
| 14:08 | <zewt> | (aside, personally I don't care about breaking big things like Yahoo, because big, live users like them should be able to fix things quickly and probably before it even hits production; it's the hoard of tiny, dead, unmaintained pages that are the worry) |
| 14:10 | <jgraham> | zewt: You're not a browser QA, are you :p |
| 14:10 | <karlcow> | heh |
| 14:11 | <Ms2ger> | Hey, at least it wasn't hotmail |
| 14:11 | <jgraham> | A motto to live by there |
| 14:11 | <zewt> | i think hotmail should be broken as often and as badly as possible |
| 14:12 | <zewt> | ideally with freak bugs that somehow cause the browser to redirect the user to other mail services |
| 14:12 | <Ms2ger> | How about the people with corporate email hosted by hotmail? |
| 14:12 | <zewt> | they should be updating resumes |
| 14:13 | <Ms2ger> | Students whose college email is hosted by hotmail? |
| 14:14 | <zewt> | transfer time |
| 14:14 | <zewt> | heh, i had a database class in college that turned out to be more like "how to autogenerate SQL using MS Access wizards" :| |
| 14:15 | <Ms2ger> | "Fun" |
| 14:15 | <zewt> | aka "how to learn nothing about databases" |
| 14:15 | <MikeSmith> | help I'm trying to come up with a regexp to express "valid duration string" |
| 14:15 | <zewt> | "duration"? |
| 14:15 | <jgraham> | Oh no! MikeSmith is stuck in a RegExp factory |
| 14:16 | <MikeSmith> | heh |
| 14:16 | <zewt> | sending help |
| 14:16 | <jgraham> | Or, if this is java a RegExpFacrotyFactory |
| 14:16 | <MikeSmith> | help me escape! |
| 14:16 | <zewt> | now you have two problems, etc |
| 14:16 | Ms2ger | managed to avoid the XHTML + VBA class |
| 14:16 | <MikeSmith> | or send somebody in to kill me now |
| 14:16 | <MikeSmith> | so basic pattern is P[dD][T[hH][mM][s[.s]S]] |
| 14:16 | <MikeSmith> | where all the stuff in brackets it optional |
| 14:17 | <MikeSmith> | BUT it must have at least one of the things in brackets |
| 14:17 | <MikeSmith> | but it doesn't matter which one |
| 14:17 | <Ms2ger> | P[dD()] |
| 14:18 | <zewt> | d_D |
| 14:18 | <zewt> | ___ |
| 14:18 | <zewt> | MikeSmith: intuition is "regex is the wrong tool" |
| 14:18 | <MikeSmith> | yeah well |
| 14:18 | <zewt> | maybe there's a non-horrifying way to do it but it isn't occurring to me |
| 14:18 | <MikeSmith> | we're way past that already |
| 14:18 | <jgraham> | Yeah, it sounds like you are way into "now you have two problems" territory |
| 14:19 | <MikeSmith> | I think I just have to OR all the possible combinations |
| 14:19 | <zewt> | (non-horrifying as in not repeating the regex four times, each time with a different item non-optional) |
| 14:19 | <zewt> | afk |
| 14:19 | <jgraham> | OK, two problems was an underestimate |
| 14:20 | <Ms2ger> | N problems |
| 14:20 | <MikeSmith> | private static final Pattern THE_PATTERN = |
| 14:20 | <MikeSmith> | Pattern.compile("^\\p{Zs}*(?:(?:([0-9]{4,})-([0-9]{2})-([0-9]{2})(?:\\p{Zs}*(?:T|\\p{Zs})\\p{Zs}*([0-9]{2}):([0-9]{2})(?::([0-9]{2})(?:\\.[0-9]+)?)?\\p{Zs}*(?:Z|(?:([+-][0-9]{2}):?([0-9]{2})))?)?)|(?:([0-9]{2}):([0-9]{2})(?::([0-9]{2})(?:\\.[0-9]+)?)?\\p{Zs}*(?:Z|(?:([+-][0-9]{2}):?([0-9]{2})))?)|(?:([0-9]{4,})-(W?)([0-9]{2}))|(?:([0-9]{2})-([0-9]{2}))|(?:Z|(?:([+-][0-9]{2}):?([0-9]{2})))|(?:([0-9]+)) |
| 14:20 | <MikeSmith> | (?:(P([0-9]+D)?(T(([0-9]+)H)?(([0-9]+)M)?(([0-9]+)(\\.[0-9]{1,3}S)?))?)))\\p{Zs}*$"); |
| 14:21 | <MikeSmith> | is what the full regexp in the code looks like now |
| 14:21 | <Ms2ger> | r- |
| 14:21 | <MikeSmith> | heh |
| 14:21 | <MikeSmith> | for this bad boy: http://www.whatwg.org/specs/web-apps/current-work/multipage/text-level-semantics.html#datetime-value |
| 14:21 | <MikeSmith> | this stuff is heinous |
| 14:21 | <MikeSmith> | nobody even wants to use the <time> stuff |
| 14:21 | <MikeSmith> | except for weirdos |
| 14:22 | <jgraham> | Really the people that insist on having this stuff in the spec should code up the validation for it themselves |
| 14:22 | <MikeSmith> | their should be parts of the spec marked with some kind of annotation to say, "This feature for OCD people only." |
| 14:22 | <MikeSmith> | jgraham: they're too busy writing ontologies |
| 14:23 | <MikeSmith> | and troubleshooting XML namespace problems |
| 14:23 | <Ms2ger> | AKA "XML namespaces" |
| 14:23 | <MikeSmith> | and pulling their pants up to their chest |
| 15:23 | <Hixie> | i would hope UTF-32 and EBCDIC would be non-ASCII compatible... |
| 15:38 | <Hixie> | MikeSmith: tantek had some reasonably good use cases for <time> |
| 15:38 | <Hixie> | MikeSmith: have him do it :_) |
| 15:47 | <MikeSmith> | Hixie: I got most of it done now. The first part ended up being not quite as nauseating as I thought'd be |
| 15:47 | <MikeSmith> | now working on the "duration time component" part |
| 15:48 | <MikeSmith> | Hixie: hey btw you saw my question about IANA and application/json+microdata ? |
| 16:27 | <dglazkov> | good morning, Whatwg! |
| 16:30 | <Hixie> | MikeSmith: rings a bell, but i don't recall what the question was |
| 16:32 | <MikeSmith> | Hixie: IANA replied to me directly with some questions and comments |
| 16:32 | <MikeSmith> | so I wanted to know if you could respond |
| 16:32 | <MikeSmith> | if I forward the reply to you |
| 16:32 | <MikeSmith> | and Cc some list I guess |
| 16:32 | <Hixie> | sure, though depending on the question i may not give answers they like :-) |
| 16:32 | <MikeSmith> | OK |
| 16:33 | <Hixie> | (e.g. "oh well, it's ok, i don't think we'll be using the iana registry for long" :-P) |
| 16:34 | tantek | scrolls up |
| 16:34 | <tantek> | hello MikeSmith and Hixie |
| 16:34 | <MikeSmith> | hey tantek |
| 16:36 | <tantek> | all the features of <time> were based on actual real world publishing examples where things either broke (accessibility problems), or people got them wrong because the alternatives were even more complicated. |
| 16:37 | <tantek> | all the experience came from the microformats community - and we try to avoid ontologies. |
| 16:40 | <MikeSmith> | OK |
| 16:48 | <Hixie> | hober: btw if you just right "bug 12345" bugzilla will automatically generate the link |
| 16:48 | <Hixie> | hober: you can also do "bug 12435 comment 6" |
| 16:50 | <Ms2ger> | *write |
| 16:50 | <Hixie> | uh |
| 16:51 | <Hixie> | yeah |
| 16:51 | <Hixie> | -_- |
| 16:51 | <Ms2ger> | And https://www.w3.org/Bugs/Public/page.cgi?id=linkify.html for previews |
| 16:52 | <Hixie> | huh |
| 16:52 | <Hixie> | when would you want to use that? weird |
| 16:54 | <Hixie> | MikeSmith: i disagree that it would be unwise to inherit parameters |
| 16:55 | <Ms2ger> | Hixie, to test syntax without spamming bugs? :) |
| 16:55 | <MikeSmith> | Hixie: OK |
| 16:55 | <Hixie> | Ms2ger: what is there to test? |
| 16:56 | <Hixie> | MikeSmith: and their comment about the security implications are bogus. If application/json doesn't define them properly, then that should be fixed. There's nothing new in microdata+json re securit. |
| 16:56 | <Hixie> | y |
| 16:56 | <Hixie> | MikeSmith: (in particular, i've no idea what tehy're talking about regarding "active content") |
| 16:56 | <Ms2ger> | To see if "bp-1166c631-aa95-4077-a103-90eed2120413" turns into a link, for example |
| 16:56 | <MikeSmith> | Hixie: me nt |
| 16:57 | <MikeSmith> | *me neither :) |
| 16:57 | <Hixie> | MikeSmith: the specification link of course is wrong, it should point to the whatwg spec. :-) |
| 16:57 | <tantek> | Hixie, btw - I've normalized the microformats2 JSON data model to also use the same "items:" array structure as well, type, properties etc. Pretty sure they're now aligned. |
| 16:57 | <MikeSmith> | I had been hoping you'd have some idea |
| 16:57 | <Hixie> | MikeSmith: the "applications" coment makes sense. let me fix that... |
| 16:57 | <tantek> | btw - I have to ask, what's the use-case for a new mimetype for this? |
| 16:58 | <tantek> | that you get above and beyond just sending it as application/json? |
| 16:58 | <hober> | Hixie: thanks |
| 16:58 | <Hixie> | tantek: distinguishing it in drag-and-drop |
| 16:59 | <tantek> | really? enough people are using drag-and-drop to have run into this? |
| 16:59 | tantek | thought web developers had given up on using drag-and-drop |
| 16:59 | <tantek> | due to the unintuitive model etc. |
| 16:59 | <Hixie> | i'm not aware of anyone having even implemented this, so no, but it's hard to see how you could use it if it didn't have its own type. |
| 16:59 | <Hixie> | and having a new type is essentially free |
| 16:59 | <Hixie> | so... |
| 16:59 | <Hixie> | brb |
| 17:00 | <tantek> | I guess I try to avoid creating new things until someone comes forward with a real world need for it. Even if it does seem theoretically necessary. |
| 17:00 | tantek | wouldn't be surprised if drag-and-drop was something sufficiently ignored as to eventually be dropped and/or rewritten with a new model. |
| 17:01 | <tantek> | every new thing created has a cost (dialog, review, maintenance, support) - having a new XYZ is never essentially free |
| 17:13 | <Hixie> | tantek: a new mime type for drag and drop is exceedingly cheap even by web standards |
| 17:13 | <Hixie> | tantek: i'd be much less surprised to see microdata (and rdfa and so forth) going away than drag and drop :-) |
| 17:13 | <tantek> | FWIW in my experience, nothing to do with IETF/IANA is "cheap". |
| 17:14 | <Hixie> | "nothing to do with IETF/IANA" is exactly right :-) |
| 17:17 | <Hixie> | MikeSmith: ok, i updated the applications line. I disagree with everything else. |
| 17:17 | <MikeSmith> | OK |
| 17:17 | <MikeSmith> | thanks |
| 17:17 | <Hixie> | MikeSmith: i don't intend to worry about it since i think the ietf/iana mime type registration thing is dead |
| 17:17 | <Hixie> | MikeSmith: i expect to do something about it next year sometime |
| 17:17 | <MikeSmith> | hai |
| 17:42 | <MikeSmith> | OK I finished writing up a validator patch to support datetime value in all its baroqueness |
| 17:42 | <MikeSmith> | and updated the wiki too |
| 17:42 | <MikeSmith> | http://wiki.whatwg.org/wiki/MicrosyntaxDescriptions#time-datetime |
| 17:42 | <MikeSmith> | will get it into Henri's review queue |
| 17:47 | <JonathanNeal> | At what point would https://github.com/scottjehl/picturefill be more appropriate to polyfill with a <picture> element? Never? Once the spec is at X stage? |
| 17:47 | <TabAtkins> | When at least one browser implements <picture>. |
| 17:48 | <SamB_MacG5> | what the heck is picture? |
| 17:49 | <MikeSmith> | SamB_MacG5: alternative proposal for same use case as srcset |
| 17:49 | <MikeSmith> | responsive images |
| 17:49 | <JonathanNeal> | SamB_MacG5: http://dvcs.w3.org/hg/html-proposals/raw-file/tip/responsive-images/responsive-images.html |
| 17:49 | <MikeSmith> | aka adaptive images |
| 17:52 | <JonathanNeal> | aka flossy may swanson aka lulu pearl mccloud aka tita chicita juanita |
| 17:55 | <SamB_MacG5> | oh, for retina and such! |
| 17:57 | <SamB_MacG5> | huh, very much under construction I see |
| 17:59 | <tantek> | MikeSmith - yes, dates and times are difficult, welcome to the club. :) |
| 17:59 | <SamB_MacG5> | also, monolithic regular expressions are unreadable |
| 18:00 | <MikeSmith> | tantek: I'm only a temporary member of the club |
| 18:00 | <MikeSmith> | Killroy was here |
| 18:01 | <tantek> | MikeSmith, glad to have you, even ephemerally :) |
| 18:01 | <MikeSmith> | :-) |
| 18:01 | <tantek> | we are all temporary members of the club, some more or less so. |
| 18:01 | <SamB_MacG5> | Emacs has at *least* three different libraries for building regular expressions in pieces |
| 18:03 | <jgraham> | SamB_MacG5: Since that was obviously a setup line, I will oblige: emaces has at least 3 different libraries for everything (but no good text editor) |
| 18:03 | <jgraham> | (actually I use emacs and like it as text editor ;) |
| 18:04 | <jgraham> | (I'm not really an emacs user though; I don't know any elisp at all) |
| 18:04 | <jgraham> | (note the subtle semantic difference) |
| 18:05 | <SamB_MacG5> | You're supposed to build your own text editor out of it ;-P |
| 18:06 | <SamB_MacG5> | what it really needs is a decent syntax highlighting framework |
| 18:06 | <say2joe> | i have to agree on emacs |
| 18:06 | <SamB_MacG5> | you know, one that actually uses grammars, rather than just highlighting stuff that matches certain regular expressions |
| 18:07 | <jgraham> | With that in mind hober promised a html-mode based on an actual parser *years* ago |
| 18:07 | <jgraham> | :p |
| 18:07 | <SamB_MacG5> | hey, at least now there's a spec ... |
| 18:08 | <jgraham> | (nxml is famously based on an actual elisp xml parser) |
| 18:08 | <SamB_MacG5> | yes, I know, but that's the *only* such mode I'm positive exists |
| 18:08 | <SamB_MacG5> | er. wait ... yes, js2-mode does parse |
| 18:08 | <jgraham> | Yeah |
| 18:08 | SamB_MacG5 | was having trouble remembering whether it did or not |
| 18:08 | <jgraham> | (it is also kind of broken, at least versions I used) |
| 18:09 | <jgraham> | (and sometimes decides that you are in the middle of a string) |
| 18:09 | <SamB_MacG5> | hmm, I think maybe it doesn't actually turn off font-lock mode? |
| 18:11 | <SamB_MacG5> | oh, huh, hober seems to have gotten as far as starting on the tokenizer :-( |
| 18:11 | <SamB_MacG5> | two years ago |
| 18:39 | <hober> | SamB_MacG5: yeah, sorry. it's tough to find time to work on it. |
| 18:39 | <SamB_MacG5> | hober: that was not intended as a complaint |
| 18:40 | <SamB_MacG5> | I only looked at the repository because of the vague description in el-get's listing |
| 18:41 | SamB_MacG5 | had been wondering if there was some kind of "look up <foo> in HTML5" tool |
| 18:43 | <hober> | SamB_MacG5: something like (defun html5-lookup (element) (interactive "sElement: ") (browse-url (format "http://www.whatwg.org/specs/web-apps/current-work/#the-%s-element"; element))) should work |
| 18:43 | <hober> | untested of course |
| 18:44 | <SamB_MacG5> | hober: yeah, I had noticed that there was such a pattern... |
| 18:45 | <SamB_MacG5> | ... too bad there isn't one for the multi-page versions |
| 18:45 | <Hixie> | there's an index :-) |
| 18:45 | <Hixie> | SamB_MacG5: actually for teh multipage version just go to http://whatwg.org/htm/#the-%s-element |
| 18:45 | <Hixie> | er |
| 18:45 | <Hixie> | /html/ |
| 18:45 | <SamB_MacG5> | oh really? |
| 18:46 | <Ms2ger> | Yep |
| 18:46 | <SamB_MacG5> | I was thinking that there should be one of those, but I hadn't considered the possibility of it being at the root ;-) |
| 18:47 | <Hixie> | also http://whatwg.org/c |
| 18:47 | <Hixie> | ("c" for "complete" because for a while /html/ went to a subse) |
| 18:47 | <Hixie> | subset |
| 18:49 | <SamB_MacG5> | I expect the W3C version doesn't have the fancy javascriptery? |
| 18:53 | <Hixie> | w3c version is missing a bunch of stuff, see the intro to the whatwg version for a summary |
| 18:59 | <pablof> | are there any examples of overloaded functions in the dom that anyone can think of, besides setRangeText? |
| 19:00 | <pablof> | overloaded as in different definitions of the same function, which vary in the number of parameters |
| 19:01 | <Hixie> | yeah, there's lots of them |
| 19:01 | <Hixie> | e.g. window.open(), document.open() |
| 19:01 | <Hixie> | document.open() is especially eggregious |
| 19:05 | <pablof> | Hixie: indeed, thanks, any others, off the top of your head? |
| 19:06 | <jgraham> | A whole bunch of canvas functions |
| 19:06 | <jgraham> | I'm pretty sure |
| 19:07 | <jgraham> | (document.open would be funny if it wasn't so sad) |
| 19:08 | <pablof> | canvas indeed, thanks |
| 19:08 | <jgraham> | "what's that? Three parameters? Hmm... I guess you just typoed 'window' to 'document'! Let's do that instead!" |
| 19:11 | <Hixie> | pablof: yeah, e.g. XMLHttpRequest.send() |
| 19:12 | <Hixie> | pablof: setTimeout() and setInterval() |
| 19:13 | <Hixie> | HTMLSelectElement.add() |
| 19:13 | <pablof> | found what i needed, thanks everyone! |
| 19:13 | <Hixie> | the Option constructor |
| 19:13 | <Hixie> | there's lots :-) |
| 19:42 | SamB_MacG5 | boggles at the mention of Netscape 2 and IE 3 in http://www.quirksmode.org/js/strings.html |
| 19:48 | <say2joe> | @samb_macg5 not sure that quirksmode.org is updated much these days despite it being a valuable resource… Netscape 2 (I can only remember back to 3)? |
| 19:49 | <SamB_MacG5> | I started out with IE 3.2 and then Netscape 4, I think... |
| 20:02 | <TabAtkins> | People who started webdev in the last 8 years or so have only had to worry about IE6, I think. |
| 20:06 | <Hixie> | gotta love how the canonical way of installing "make" on a mac involves downloading a 1.6 gig file |
| 20:17 | <ap> | Hixie: Just noticed that HTML spec doesn't ever mention respecting Content-Language HTTP header field. Is that intentional? Are browsers allowed to respect it? |
| 20:18 | <Hixie> | it's mentioned somewhere, iirc |
| 20:18 | <ap> | Hixie: I can only find http-equiv |
| 20:18 | <Hixie> | search for "then language information from a higher-level protocol (such as HTTP)" |
| 20:19 | <ap> | Hixie: thanks |
| 20:20 | <ap> | Hixie: surprising that pragma wins over HTTP here |
| 20:20 | <Hixie> | *shrug* |
| 20:20 | <Hixie> | i've long stopped being surprised :-P |
| 20:22 | <ap> | Hixie: last occurrence of "surprise" on May 8th , 2012, it appears :-P |
| 20:23 | <Hixie> | :-) |
