| 00:07 | <TabAtkins> | I feel nervous when writing a spec this small. I feel like I'm leaving out something. |
| 00:07 | <boogyman> | lol |
| 00:07 | <TabAtkins> | It's only like 4 screens of text (or probably 8 on a normal persons' monitor) |
| 00:07 | <TabAtkins> | (That includes boilerplate, which is at least half of the height.) |
| 00:07 | <boogyman> | only.. lol |
| 00:19 | <Hixie> | hsivonen: your feedback on http://www.w3.org/Bugs/Public/show_bug.cgi?id=13502 would be quite helpful |
| 00:19 | <Hixie> | TabAtkins: you only have 2 screens' of boilerplate? that can't be right. did you run it through pubrules? ;-) |
| 00:26 | <TabAtkins> | Hixie: 2 of my screens. Actually, 2.5. That's 5 normal screens worth. |
| 00:26 | <TabAtkins> | I've got vertical 30"ers. |
| 00:27 | <Hixie> | still seems short |
| 00:27 | <Hixie> | are you including the conformance section, conformance class definitions, and terminology and typegraphical convention sections in what you call boilerplate? |
| 00:28 | <TabAtkins> | Yes. |
| 00:28 | <Hixie> | definitely sounds short :-) |
| 00:28 | <TabAtkins> | That's the new Module Template, so shrug. |
| 00:58 | <danja> | anyone here working for ISO? |
| 00:59 | <danja> | comments on their html? |
| 01:06 | <TabAtkins> | What, you mean the ISO HTML spec? |
| 01:07 | <aho> | my favorite: http://en.wikipedia.org/wiki/ISO_3103 |
| 01:07 | <aho> | monocle time! |
| 01:07 | <aho> | hm hm. quite. :v |
| 01:08 | <TabAtkins> | danja: If that is what you meant, then that's a fork of HTML4. It's useless. |
| 01:20 | <danja> | hello? |
| 01:20 | <danja> | TabAtkins, I'm feeling a bit cynical |
| 01:24 | <danja> | TabAtkins, thing is, seems like some of the best brains on the planet |
| 01:24 | <danja> | being useless |
| 01:24 | <aho> | fork of html4 = no one gives a f- |
| 01:24 | <danja> | hope hixie heard that |
| 01:24 | <aho> | seriously |
| 01:24 | <TabAtkins> | danja: I have no idea what you mean. |
| 01:24 | <aho> | <: |
| 01:24 | <Hixie> | o_O |
| 01:24 | <danja> | let me explain |
| 01:24 | <aho> | TabAtkins, you did the css @var stuff, didn't you? |
| 01:24 | <TabAtkins> | aho: Yes. I'm writing the actual spec *right now*. |
| 01:24 | <danja> | computers became exciting to me |
| 01:24 | <aho> | i really like how media queries work in scss |
| 01:25 | <aho> | you can put them inside some rule blocks |
| 01:25 | <aho> | http://sass-lang.com/docs/yardoc/file.SASS_REFERENCE.html#media |
| 01:25 | <danja> | hixie! - talk for a mo |
| 01:25 | <aho> | this is the behavior i usually want |
| 01:25 | <aho> | (: |
| 01:25 | <aho> | i.e. having the overrides right next to the default values |
| 01:26 | <aho> | the way media queries currently work is exactly the wrong way around :> |
| 01:26 | <TabAtkins> | aho: Yeah, I agree. It would be useful. One more reason to allow at-rules inside of declaration blocks. |
| 01:26 | <aho> | :)~ |
| 01:27 | <TabAtkins> | Now, I wouldn't say that. The current way works fine, it's just that the other way is *also* useful, when you just need to make a small media-dependent adjustment. |
| 01:28 | <aho> | well, i build the site from the ground up... i create reusable "bricks" which are then used all over the place. everything which belongs together (i.e. everything of this tiny sub tree) is together in one specific place |
| 01:28 | <aho> | media queries totally break that |
| 01:29 | <aho> | s/the site/sites :> |
| 01:30 | <aho> | it's pretty much like override stylesheets for IE via conditinal comments vs IE filters |
| 01:30 | <aho> | it's ugly either way, but the "inline" option is nicer to use (because it's right there) |
| 01:31 | <danja> | I like the brick metapgor |
| 01:32 | <aho> | .Foo{} .Foo>.bar{} .Foo>.baz{} .Foo>.baz>.bat{} |
| 01:32 | <aho> | everything looks like that :> |
| 01:33 | <aho> | classes, subtree root node is capitalized, child combinators |
| 01:33 | <TabAtkins> | It'll be nice when we can do .Foo{ ... &>.bar{...} $>.baz{ ... &>.bat{...}}} |
| 01:34 | <TabAtkins> | (That looks prettier when actually indented.) |
| 01:34 | <aho> | ye, currently using scss/compass for that |
| 01:34 | <aho> | works pretty well so far |
| 01:34 | <aho> | just the "& > " bit is kinda awkward ;) |
| 01:35 | <aho> | i really wish the child combinator would have been the default instead of the descendant one |
| 01:35 | <aho> | child:desc exceeds 9:1 with current projects |
| 01:36 | <aho> | well, mine anyways .) |
| 01:53 | <TabAtkins> | aho: Oh yeah, nesting @media, at least, is allowed by the Conditionals spec. |
| 01:53 | <TabAtkins> | All three of the conditionals can be nested within each other. |
| 01:53 | <aho> | shiver me timbers :O |
| 04:06 | <danja> | sorry hixie, later |
| 12:00 | <GlitchMr> | http://testcases.glitchmr.pl/html/url-002.html |
| 12:00 | <GlitchMr> | I was bored... |
| 12:00 | <GlitchMr> | I don't know how valid this test is... |
| 12:01 | <GlitchMr> | But I find it interesting how various browsers do it differently... |
| 12:10 | <GlitchMr> | http://www.whatwg.org/specs/web-apps/current-work/multipage/urls.html#parsing-urls |
| 12:11 | <GlitchMr> | I inserted this testcase to this... |
| 12:15 | <GlitchMr> | http://ms2ger.freehostia.com/tests/html5/dynamic-markup-insertion/document.write-01.xhtml |
| 12:15 | <GlitchMr> | Fail, this document is actually text/html :P. |
| 12:16 | <GlitchMr> | OK, seriously... maybe I will make version which actually uses XHTML... |
| 13:28 | <jarek> | Hi |
| 13:28 | <jarek> | is there any point is specifying charset on <script> elements if it has already been defined with <meta charset="utf-8">? |
| 13:29 | <zcorpan> | no |
| 13:29 | <jarek> | If I set <meta charset="utf-8"> in head section, does is it mean that all resources loaded from that page will use utf-8 as well? |
| 13:30 | <zcorpan> | no |
| 13:30 | <zcorpan> | but scripts and style sheets will inherit the encoding from the page unless they declare encoding themselves |
| 13:30 | <jarek> | will it work for CSS files? Should I start my CSS files with @charset 'utf-8'; ? |
| 13:30 | <jarek> | oh, great |
| 13:31 | <zcorpan> | (html in an iframe will need to declare its own encoding) |
| 13:31 | <GlitchMr> | Warning! This method has very idiosyncratic behavior. In some cases, this method can affect the state of the HTML parser while the parser is running, resulting in a DOM that does not correspond to the source of the document. |
| 13:31 | <GlitchMr> | ... |
| 13:31 | <GlitchMr> | this also happens when using DOM... |
| 13:32 | <jarek> | and what if I'm using XHTML5 and declare charset at the top of the document like this: |
| 13:32 | <jarek> | <?xml version="1.0" encoding="UTF-8"?> |
| 13:32 | <zcorpan> | is that about document.write? |
| 13:32 | <GlitchMr> | And older versions of IE also did it without doing anything... |
| 13:32 | <GlitchMr> | zcorpan, yeah |
| 13:32 | <jarek> | do I still have to use <meta charset="utf-8"> in head? |
| 13:32 | <GlitchMr> | jarek, charset declaration in <meta> is ignored in XML mode |
| 13:33 | <GlitchMr> | XML parser must know encoding before processing anything... |
| 13:34 | <zcorpan> | GlitchMr: file a bug saying that the note should be clearer about what it means |
| 13:34 | <zcorpan> | jarek: xml defaults to utf-8 so doesn't need a declaration at all |
| 13:34 | <jarek> | GlitchMr: but will <?xml version="1.0" encoding="UTF-8"?> be also applied to scripts and stylesheets that are loaded from that page? |
| 13:34 | <zcorpan> | yeah |
| 13:34 | <GlitchMr> | jarek, I don't know, really :P. |
| 13:35 | <jarek> | ok, let's check :P |
| 13:35 | <GlitchMr> | zcorpan, I through that XML detects encoding basing on BOM and uses UTF-8 if BOM wasn't found... |
| 13:35 | <zcorpan> | no declaration in xml will still let css and scripts inherit the used encoding |
| 13:35 | <zcorpan> | GlitchMr: it's not that simple. :) |
| 13:35 | <GlitchMr> | But I've never had need need to specify charset in CSS. The only reason I could think is having Unicode characters in content and even that is possible without specifying charset. |
| 13:37 | <jarek> | it looks like UTF-8 is assumed by default for CSS, JS and XHTML |
| 13:37 | <jarek> | at least on Chrome |
| 13:38 | <GlitchMr> | Even with stuff like ISO-8859-1? |
| 13:38 | <jarek> | yeah |
| 13:38 | <GlitchMr> | Makes sense... |
| 13:38 | <zcorpan> | jarek: css and js inherits the encoding from the embedding page if they don't declare their own encoding |
| 13:39 | <jarek> | I tried setting ISO-8859-1 in head section and as xml tag attribute, but I can still print UTF8 characters |
| 13:39 | <zcorpan> | what does your content-type http header say? |
| 13:41 | <jarek> | I can't find this header in Chrome Dev Tools |
| 13:41 | <jarek> | I can only see "Accept-Charset:ISO-8859-1,utf-8;q=0.7,*;q=0.3" |
| 13:42 | <zcorpan> | that's a request header |
| 13:42 | <zcorpan> | content-type is a response header |
| 13:42 | <zcorpan> | or at least i'm interested in the response header :) |
| 13:43 | <GlitchMr> | http://glitchmr.pl/private/cssencoding.html |
| 13:43 | <jarek> | ahh, it shows up in Firebug |
| 13:43 | <jarek> | "Content-Type application/xhtml+xml" |
| 13:45 | <zcorpan> | jarek: do you have a URL to your test case? |
| 13:45 | <jarek> | zcorpan: not yet, I'm running it on my local server |
| 13:48 | <zcorpan> | jarek: i can think of two reasons why your xhtml uses utf-8 instead of what you declared in the xml decl |
| 13:48 | <zcorpan> | jarek: either you have a UTF-8 BOM |
| 13:49 | <zcorpan> | jarek: or your server sends content-type: application/xhtml+xml; charset=utf-8 |
| 13:56 | <jarek> | do I get any performance gains by using XHTML5 instead of HTML5? |
| 13:57 | <jarek> | I heard that XML parsers are much faster than SGML parsers |
| 13:57 | <GlitchMr> | jarek, I don't think so |
| 13:57 | <GlitchMr> | Browsers skip more complex SGML tricks anyway... |
| 13:57 | <zcorpan> | jarek: no |
| 13:57 | <GlitchMr> | <b/This is considered wrongly by browsers.../ |
| 13:57 | <GlitchMr> | And browsers have to check for errors in XML mode... |
| 13:57 | <zcorpan> | jarek: also, browsers don't use sgml parsers |
| 13:58 | <GlitchMr> | <!-- -- hello this -- is -- error -- which -- needs -- to be -- detected -- in -- xml --mode --> |
| 13:58 | <jarek> | zcorpan: yeah, but parsing HTML still requires a lot more trickery than regular XML |
| 13:58 | <GlitchMr> | jarek, actually, it's pretty easy to parse HTML |
| 13:59 | <zcorpan> | jarek: xml also requires a lot of trickery |
| 13:59 | <zcorpan> | jarek: like namespaces and checking for invalid characters in tags |
| 13:59 | <GlitchMr> | Yeah, namespaces are problematic |
| 13:59 | <zcorpan> | see http://hsivonen.iki.fi/cost-of-html/ |
| 13:59 | <GlitchMr> | <html xmlns="http://www.w3.org/1999/xhtml">;... |
| 14:01 | <GlitchMr> | Note: The charset attribute on the meta element has no effect in XML documents, and is only allowed in order to facilitate migration to and from XHTML. |
| 14:01 | <zcorpan> | jarek: hsivonen's finding showed that namespaces cost 21% perf hit for xml while html's "fixups" only costs 7% perf hit |
| 14:02 | <GlitchMr> | I find it interesting through that implementation of XML in IE9 doesn't fail on invalid documents. I don't even mean more complex XML rules which needs to be find intentionally (like "<" in attribute values)... |
| 14:03 | <zcorpan> | IE doesn't fail? |
| 14:04 | <GlitchMr> | Yes |
| 14:04 | <zcorpan> | o_O |
| 14:04 | <GlitchMr> | Whatever you do, it won't fail |
| 14:04 | <GlitchMr> | But it implements rules like <div/>. |
| 14:04 | <GlitchMr> | And it doesn't generate automatically elements like <tbody> or <body>... |
| 14:05 | <zcorpan> | i thought it aborted parsing but didn't show an error message |
| 14:05 | <zcorpan> | what if you give it data:text/xml,<foo>test </bar> test</foo> ? |
| 14:05 | <zcorpan> | does it render "test test"? |
| 14:06 | <GlitchMr> | ... |
| 14:06 | <GlitchMr> | Weird... it inserts some random error message |
| 14:06 | <GlitchMr> | http://www.webdevout.net/articles/beware-of-xhtml-examples/1.xhtml |
| 14:07 | <GlitchMr> | This could be perfectly seen in IE9 through |
| 14:07 | <zcorpan> | that page doesn't have any more content after the wellformedness error |
| 14:07 | <zcorpan> | so it doesn't tell whether it aborts parsing or continues parsing |
| 14:08 | <GlitchMr> | About your example, DOM of it looks like <foot>test... |
| 14:08 | <GlitchMr> | Weird... |
| 14:09 | <zcorpan> | that means it aborted parsing |
| 14:10 | <GlitchMr> | Makes sense... |
| 14:10 | <GlitchMr> | I through it ignored errors |
| 14:11 | <GlitchMr> | But I was testing it with stuff like <p>Hello<p>World |
| 14:11 | <GlitchMr> | Which actually gave <p>Hello<p>World</p></p> DOM... |
| 14:12 | <GlitchMr> | I think that Chrome does it similarly but it gives error messages... |
| 14:13 | <GlitchMr> | http://glitchmr.pl/pastebin/?f=rlaptin-2011-09-24 |
| 14:14 | <GlitchMr> | I think that "type" is overused in HTML... |
| 14:14 | <zcorpan> | is it a problem? |
| 14:15 | <GlitchMr> | Nope, it's just kinda weird... |
| 14:15 | <GlitchMr> | type having multiple meanings... |
| 14:15 | <GlitchMr> | But it kind of makes sense... |
| 14:15 | <GlitchMr> | It's easy to type word. |
| 14:15 | <zcorpan> | there are more attributes like that |
| 14:17 | <GlitchMr> | I went through my list. I've only found VALUE and (possibly) ALIGN... |
| 14:17 | <zcorpan> | href |
| 14:17 | <GlitchMr> | 'HREF'=>'on <a>, <area>, <base> or <link> |
| 14:17 | <GlitchMr> | Specifies the destination of a link.', |
| 14:17 | <GlitchMr> | I think I should make more detailed instructions :P. |
| 14:17 | <GlitchMr> | lol |
| 14:17 | <zcorpan> | <base> is not a link :) |
| 14:18 | <GlitchMr> | Yeah... |
| 14:18 | <zcorpan> | name="" |
| 14:18 | <GlitchMr> | It's from my IRC bot list of tags... |
| 14:18 | <GlitchMr> | Yeah, I will agree with name. |
| 14:19 | <GlitchMr> | Possibly also MULTIPLE, it has three different meanings... |
| 14:20 | <zcorpan> | even attributes like width, height, src are subtly different |
| 14:20 | <zcorpan> | and alt="" |
| 14:20 | <zcorpan> | span="" |
| 14:20 | <GlitchMr> | 'WIDTH'=>'on <applet> (Transitional), <col> (HTML4), <colgroup> (HTML4), <embed> (HTML5), <hr> (Transitional), <iframe> (Transitional), <img> (All), <object> (All), <pre> (HTML4), <table> (HTML4), <td> (Transitonal), <th> (Transitional) or <video>. (HTML5) |
| 14:20 | <GlitchMr> | Specifies width of element', |
| 14:20 | <GlitchMr> | alt? I don't think so... |
| 14:20 | <GlitchMr> | 'ALT'=>'on <applet>, <area>, <img> or <input type=image> |
| 14:20 | <GlitchMr> | Sets alternate text to display if image couldn\'t be displayed/', |
| 14:20 | <GlitchMr> | s/image/object/; |
| 14:20 | <zcorpan> | max="" |
| 14:20 | <GlitchMr> | I think I should replace it |
| 14:21 | <zcorpan> | title="" |
| 14:21 | <GlitchMr> | About span=""? It has just one meaning. |
| 14:21 | <GlitchMr> | For <col> and <colgroup>... |
| 14:21 | <GlitchMr> | And title="" has this same meaning for every tag... |
| 14:22 | <GlitchMr> | Of course I know about <span> and <title>, but that's different construct. |
| 14:22 | <zcorpan> | <col span> and <colgroup span> have different requirements |
| 14:23 | <zcorpan> | title="" has different meaning on <style> and <link rel="alternate stylesheet"> |
| 14:23 | <GlitchMr> | zcorpan, oh right, I forgot about it |
| 14:23 | <zcorpan> | but i guess it depends on how strictly correct you want your definitions to be |
| 14:24 | <zcorpan> | being exactly correct would just be a copy of the spec :) |
| 14:24 | <GlitchMr> | :P |
| 14:24 | <GlitchMr> | It's IRC bot, I try to avoid very long descriptions. |
| 14:24 | <GlitchMr> | I don't want every description to activate my pastebin :P. |
| 14:28 | <GlitchMr> | KEYTYPE for me is funny (for now...). As for now it looks like - if you specify it you might make <KEYGEN> not work :P. |
| 15:03 | <GlitchMr> | http://testcases.glitchmr.pl/html/marquee-002.html |
| 15:03 | <GlitchMr> | I find it interesting. Opera and Chrome fail it in different ways |
| 15:03 | <GlitchMr> | In Opera, onbounce="" event doesn't activate. |
| 15:04 | <GlitchMr> | In Chrome it doesn't move at all... |
| 15:04 | <GlitchMr> | Is it possibly bug in Chrome? |
| 15:05 | <GlitchMr> | It looks like behavior="alternate" isn't implemented in WebKit or something |
| 16:24 | <zewt> | heh |
| 16:24 | <zewt> | unpaired surrogates completely hose input boxes in chrome |
| 16:27 | <Philip`> | UTF-16 really doesn't seem to have worked out well in practice |
| 16:27 | <zewt> | well, it's fundamentally absurd, so that's little surprise :) |
| 16:29 | <zewt> | (i don't know which is worse, the fact that it combines the undesirable properties of multibyte and wide encodings, or the fact that the cases where it needs special handling are also exceptionally rare) |
| 16:43 | <zewt> | ouch, chrome is broken even for legal surrogate pairs |
| 16:44 | <zewt> | (deja vu--maybe I hit this before) |
| 16:44 | <zewt> | (can't tell if it's that or just general "utf-16 never works") |
| 16:57 | <Philip`> | Computers are surprisingly rubbish at fundamental things like text |
| 16:57 | <Philip`> | and numbers |
| 16:58 | <zewt> | heh |
| 16:59 | <zewt> | they should teach kids NaN and overflow flags in grade school |
| 17:12 | <GlitchMr> | "has some support but crashes" |
| 17:12 | <GlitchMr> | O_o |
| 17:13 | <GlitchMr> | But that <b></b> while parsing invalid HTML <table><b> is weird... |
| 17:18 | <GlitchMr> | Should I remove testcase which reports 404? |
| 17:51 | <GlitchMr> | http://testcases.glitchmr.pl/html/img-001.html |
| 17:51 | <GlitchMr> | First version of that had race condition :(. |
| 17:51 | <GlitchMr> | So I think I fail... |