| 01:14 | <Hixie> | if anyone gets the reason behind the reference in the title of my latest blog, i'll be impressed |
| 01:18 | <Hixie> | hsivonen_: does validator.nu have a way to make it validate the page from the Referer field? |
| 01:19 | <Hixie> | hsivonen_: btw #presets in the "the pitch" section of your about page is a broken link |
| 01:45 | <csarven> | Hixie Does it have anything to do with http://www.whatwg.org/issues/data.html ? :) |
| 01:48 | <Hixie> | no |
| 01:59 | <csarven> | BTW, that page is pretty slow in Firefox. Pretty smooth in Opera. |
| 01:59 | <csarven> | The mouse movement on canvas that is. |
| 01:59 | <roc> | Has everyone here read this already? "IE8 standards mode" will be the default "standards mode" http://blogs.msdn.com/ie/archive/2008/03/03/microsoft-s-interoperability-principles-and-ie8.aspx |
| 02:08 | <Hixie> | yeah, i even blogged about it |
| 02:08 | <Hixie> | great news |
| 02:11 | <roc> | sorry, I'm too slow |
| 02:11 | <Hixie> | :-) |
| 02:11 | <Hixie> | i didn't mean it that way :-) |
| 02:25 | mpt | is disconcerted by a picture of Osama bin Laden next to the post about Acid 3 |
| 02:25 | <Hixie> | uri? |
| 02:26 | <mpt> | http://www.flickr.com/photos/41089232@N00/114103905 |
| 02:29 | <Hixie> | either flickr is down or my proxy is acting up |
| 02:30 | <Hixie> | wtf the web just died for me |
| 02:32 | <Philip`> | Didn't you hear about the scheduled maintenance tonight? The web has to be turned off for two hours while it's all upgraded |
| 02:32 | <Hixie> | wouldn't that be nice |
| 02:33 | <Hixie> | mpt: i don't see anything about acid3 on there? *confused* |
| 02:34 | <mpt> | Hixie, that was the photo "randomly selected from Flickr using Yahoo! Pipes based on the words in the latest Web log entry on this site" |
| 02:35 | <Hixie> | oh haha i didn't realise that was the image that had been selected |
| 02:35 | <Hixie> | wow that's funny |
| 02:38 | <Hixie> | so... |
| 02:38 | <Hixie> | if we make <input type=hidden> stay in the table... |
| 02:38 | <Hixie> | what the heck should <table>x<input type=hidden></table> do |
| 02:38 | Hixie | studies |
| 02:39 | <jruderman> | "This decision comes on the heels of a recent leak of a classified NSA National Security Assessment which ranked the Bush Administration itself as the #2 threat to U.S. national security interests." |
| 02:39 | <jruderman> | i hadn't heard about that |
| 02:40 | <jruderman> | Hixie: does html 5 parsing require that where tags go in the DOM depend on attribute values? |
| 02:40 | <Hixie> | it's about to start going down that slippery slope, yes |
| 02:40 | <Hixie> | bz and othermaciej have asked that <input type=hidden> be given special status in <table> |
| 02:41 | <jruderman> | hmm |
| 02:41 | <Philip`> | jruderman: [citation needed] |
| 02:41 | <jruderman> | https://bugzilla.mozilla.org/show_bug.cgi?id=327796#c5 makes me think that will be unpopular |
| 02:41 | <jruderman> | Philip`: hehe |
| 02:42 | <othermaciej> | I don't exactly love the idea of that rule but it seems to matter for compatibility |
| 02:42 | <Hixie> | jruderman: https://bugzilla.mozilla.org/show_bug.cgi?id=390565 |
| 02:53 | <dbaron> | Hixie, where do bug reports in acid3 go? |
| 02:53 | <Hixie> | ian⊙hc or the wasp alias, if you can find it |
| 02:53 | <Hixie> | i think that's acid3⊙wo |
| 02:53 | <Hixie> | not sure |
| 02:53 | <Hixie> | but ian⊙hc is fine |
| 02:54 | <Hixie> | and i try to give quick turnaround, so if you don't get a reply within 12 hours, ping me :-) |
| 02:54 | <othermaciej> | dbaron: what bug did you find? |
| 02:55 | <dbaron> | othermaciej, I think test 42 has an incorrect selector |
| 02:55 | <dbaron> | Fx trunk should pass that test |
| 02:55 | <dbaron> | and it looks wrong by inspection |
| 02:55 | <Hixie> | what failure number do you get? |
| 02:55 | <dbaron> | although I didn't test the modificication |
| 02:55 | <dbaron> | Hixie, the very last failure |
| 02:56 | <Hixie> | 39? |
| 02:56 | Hixie | looks |
| 02:56 | <dbaron> | no, "rule did not start matching after change" |
| 02:56 | <Hixie> | oh the one between 12 and 14? ok |
| 02:56 | Hixie | looks at that instead :-) |
| 02:57 | <dbaron> | oh, I see, the test continues after that function |
| 02:57 | <dbaron> | I think "#div1 ~ div div + div > div" should be "#div1 ~ div + div div > div" |
| 02:59 | <Hixie> | after 310 is in the tree, you get this DOM: |
| 02:59 | <Hixie> | <body><div1/><div2/><div3><div31><div310/><div311><div3111/></div311></div31></div3></body> |
| 02:59 | <dbaron> | oh |
| 02:59 | <dbaron> | hrm |
| 03:00 | <dbaron> | ignore me |
| 03:00 | <dbaron> | for now, anyway |
| 03:00 | <Hixie> | and "#div1 ~ div div + div > div" should match 3111 the same way that "div1 ~ div3 div310 + div311 > div3111" does |
| 03:00 | <Hixie> | seems valid to me :-) |
| 03:01 | <dbaron> | yeah |
| 03:01 | <dbaron> | for some reason I forgot to process the insertBefore mentally even though I saw it |
| 03:01 | <dbaron> | no idea why we fail, though |
| 03:01 | <dbaron> | I guess I'll have to ask someone to write a testcase |
| 03:02 | <Hixie> | it's basically a dynamic variant of your greedy matching test from css1, extended to use the css3 combinators |
| 03:02 | <dbaron> | though maybe it's not too hard to extract just the relevant bits |
| 03:04 | <Hixie> | http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Cstyle%3E%20div1%20~%20div3%20div310%20%2B%20div311%20%3E%20div3111%20%7B%20border%3A%20solid%3B%20%7D%20%3C%2Fstyle%3E%0A%3Cbody%3E%3Cdiv1%3E%3C%2Fdiv1%3E%3Cdiv2%3E%3C%2Fdiv2%3E%3Cdiv3%3E%3Cdiv31%3E%3Cdiv311%3E%3Cdiv3111%3E%3C%2Fdiv3111%3E%3C%2Fdiv311%3E%3C%2Fdiv31%3E%3C%2Fdiv3%3E%3C%2Fbody%3E%0A%3Cscript%3E%0Avar%20div31%20%3D%20document.getElementsByTagName('div31')%5B0%5D%3B%0Adiv |
| 03:05 | <Hixie> | if that got truncated, go to http://software.hixie.ch/utilities/js/live-dom-viewer/ and hit "download" then "permalink" to get the uri |
| 03:05 | <Hixie> | i think that's the same test |
| 03:05 | <Hixie> | and it works fine if you do the mutation manually |
| 03:06 | <Hixie> | so it must be a dynamic thing |
| 03:06 | <dbaron> | yeah, I see the bug |
| 03:06 | <dbaron> | it's a one-liner |
| 03:06 | <Hixie> | cool |
| 03:06 | <dbaron> | now I need to find an excuse to check it in |
| 03:06 | <dbaron> | If only this test had been easier to parse the results of I'd have used it before I checked in my fix for those bugs, and been able to get it in easily... |
| 03:07 | dbaron | hates overly complex tests |
| 03:07 | <Hixie> | i tried to make the tests as self-contained as possible |
| 03:07 | <Hixie> | but really acid tests aren't meant to be processed directly by engineers |
| 03:07 | <Hixie> | when acid2 came out i made minimised tests for each bug that opera had in acid2 |
| 03:08 | <Hixie> | i'd expect any qa team to do the same for any acid test |
| 03:08 | <dbaron> | er, no, I don't see the bug |
| 03:10 | <dbaron> | I better leave splitting these bugs up to people who actually think they have enough time to do it, instead of rushing... |
| 03:11 | <Hixie> | i wouldn't recommend trying to do it this late in a release cycle anyway |
| 03:11 | <Hixie> | best to work on standards bugs at the start of a cycle |
| 03:13 | <dbaron> | yeah, except you always release these acid tests a few months before the end of our release cycles |
| 03:14 | <dbaron> | and then we ship a not-much-improved release and look stupid |
| 03:15 | <Hixie> | what would you like me to do? wait til all the browser vendors' release schedules line up? |
| 03:15 | <Hixie> | acid3 has been in production since last april, and mostly came to a head recently because of the IE acid2 announcement |
| 03:15 | <Hixie> | i recommend shorter release cycles, that way you never have a problem like this :-) |
| 03:17 | <Hixie> | in fact, i'd argue that this is the best time for you for an acid test to come out, as it means that as the test finally gets all the bugs shaken out of it, you'll be at the best possible time to start working on standards bugs again |
| 03:17 | <Hixie> | (i.e. it'll be stable just as your release cycle starts) |
| 03:17 | <dbaron> | yeah, I've been pushing for shorter release cycles |
| 03:18 | <dbaron> | We may well do one. |
| 03:18 | <dbaron> | dare I ask where the function expect is defined? |
| 03:19 | <Hixie> | for which test? |
| 03:19 | <dbaron> | er, never mind |
| 03:20 | <dbaron> | The closures confused me about what was calling what |
| 03:20 | <dbaron> | I'm just annoyed about this particular test because I thought I fixed all the bugs that would have caused it to fail. |
| 03:20 | <Hixie> | heh |
| 03:29 | <dbaron> | ah, it's a matching bug, not a dynamic change handling bug |
| 03:35 | <othermaciej> | the next Safari release won't look so good on Acid3 |
| 03:35 | <othermaciej> | not as good as current nightlies |
| 03:39 | <Hixie> | dbaron: really? it only seems to occur with a dynamic change. |
| 03:40 | <dbaron> | Hixie, not sure why you think that -- I tried removing and reinserting the body and the bug was still present |
| 03:40 | <dbaron> | Hixie, simple testcase in https://bugzilla.mozilla.org/show_bug.cgi?id=420814 |
| 03:41 | <Hixie> | odd, i couldn't get the test to fail without dynamic changes |
| 03:41 | <Hixie> | oh well |
| 03:41 | <Hixie> | oh i guess i didn't have elements named the same thing in my version of the test |
| 03:41 | <Hixie> | so maybe there are two bugs here |
| 07:02 | <Hixie> | if this upcoming change doesn't break something, i'll be impressed |
| 07:23 | <Hixie> | so anyone got any hot ideas on how to handle <table><td>x</td> <select><option>a<option>b <td>y</td> in a compatible way? |
| 07:23 | <Hixie> | Philip`? |
| 08:21 | <hsivonen_> | annevk: regarding "UA-specific crap" on the tracker: I'm using the annotations. |
| 08:26 | <hsivonen_> | I think the "ordered" is a red herring when it comes to <ol> and <ul>. It really is about showing the counters. A bulleted list can still have a deliberate order. |
| 08:43 | <othermaciej> | hsivonen_: don't let tantek hear you say that |
| 08:45 | <hsivonen_> | heh |
| 08:51 | <virtuelv> | I haven't read it thoroughly yet, but did Microsoft just turn completely around on the x-ua-compatible stuff? |
| 08:54 | <Philip`> | virtuelv: Not completely - just on the default |
| 08:54 | <virtuelv> | :/ |
| 08:55 | <Philip`> | (which is the part people were most complaining about) |
| 08:57 | hsivonen_ | sees the Osama image at http://ln.hixie.ch/?start=1066145333&count=1 |
| 09:00 | <virtuelv> | hsivonen_: :D :D :D |
| 09:01 | <virtuelv> | the image is based on Hixie's last entry, so you'll see it on all pages |
| 09:12 | <Lachy> | wow!!!! "We’ve decided that IE8 will, by default, interpret web content in the most standards compliant way it can." |
| 09:12 | <Lachy> | http://blogs.msdn.com/ie/archive/2008/03/03/microsoft-s-interoperability-principles-and-ie8.aspx |
| 09:14 | <zcorpan> | Lachy: yes, very good news |
| 09:17 | <zcorpan> | Hixie: dammit, you're working too fast. i can't keep up :P |
| 09:22 | <Hixie> | zcorpan: :-D |
| 09:23 | <Hixie> | i guess i'll have to have a "last insertion mode" variable to do the <table><select><td> thing |
| 09:23 | <Hixie> | anyway bed time |
| 09:25 | <zcorpan> | aha! the precedent is set for looking at attribute values in the tree builder |
| 09:27 | <hsivonen_> | scary |
| 09:27 | <hsivonen_> | ah. it's type=hidden |
| 09:28 | <hsivonen_> | that's good |
| 09:28 | <zcorpan> | why is it good? |
| 09:31 | <hsivonen_> | zcorpan: compat. and in the case of input, the conceptual reality is that the type attribute creates practically separate element types |
| 09:31 | <zcorpan> | hsivonen_: ok. and yep |
| 09:53 | <Philip`> | Hixie: Maybe handle it by copying what WebKit or Gecko does? |
| 09:53 | <Philip`> | I don't really know what they do, though |
| 09:54 | <Philip`> | (except I vaguely remember WebKit running the normal parsing sequence but with an 'is being inserted into table' flag that changes the processing of table elements) |
| 09:59 | <annevk2> | '<p>Ignore the token.</p> <!-- :-( -->' :) |
| 10:49 | <Hixie> | Philip`: it's not that simple; <table><tr><select><tbody> has to do different things than <table><select><tbody> |
| 10:50 | <annevk> | are we now figuring out the parser architecture is not good enough? |
| 10:50 | annevk | hopes not |
| 10:50 | <Hixie> | no, it can be hacked to do this in various ways |
| 10:50 | <Hixie> | i'm just trying to work out which way |
| 10:50 | <Hixie> | i can remember the last mode and treat <select> specially "in table" |
| 10:51 | <Hixie> | and have a bunch of new states "in select" |
| 10:51 | <Hixie> | i can just have a bunch of new states "in select" with those states doing non-trivial examination of the stack |
| 10:51 | <Hixie> | i can have one "in select" state per table state |
| 10:51 | <Hixie> | ("in select in table row", etc) |
| 10:52 | <annevk> | if the last one is not more verbose than the others i might prefer it |
| 10:52 | <Hixie> | i'm leaning towards the seoond of those so far |
| 10:52 | <Hixie> | the last one is massively more verbose |
| 10:52 | <annevk> | k |
| 11:08 | <Hixie> | christ, acid3 just hit something with a lot of readers |
| 11:11 | <Hixie> | http://www.neowin.net/forum/index.php?showtopic=623574 |
| 11:14 | <annevk> | I wonder if everytime MS makes some kind of Web standards announcement there will be another significant announcement related to Web standards on the same day... |
| 11:20 | <Hixie> | heh |
| 11:23 | <zcorpan> | what was announced at the same time as the ie8 versioning switch thing (first time around)? the acid3 competition? |
| 11:25 | <annevk> | zcorpan, HTML5 |
| 11:25 | <annevk> | (it was published as FPWD) |
| 11:26 | <takkaria> | how strange, Ff3b3 got 60/100 the first time I ran it and 59 the next |
| 11:27 | <takkaria> | though it's Fx now, I live in the past |
| 11:27 | <zcorpan> | annevk: ah, right |
| 11:27 | <Hixie> | sure getting a lot of hits from http://pro.tweakers.net/nieuws/52227/web-standards-project-geeft-acid3-test-vrij.html |
| 11:28 | <annevk> | tweakers.net is pretty huge yeah |
| 11:31 | <annevk> | http://browsershots.org/http://acid3.acidtests.org/ is nice |
| 11:33 | <takkaria> | heh, poor old MSIE4 |
| 11:34 | <Hixie> | i hear NS4 crashes |
| 11:34 | <annevk> | how unexpected :) |
| 11:34 | <annevk> | it's funny how many people wonder how the reference rendering was made |
| 11:35 | <zcorpan> | obviously Hixie made a superbrowser that perfectly implements html5 and css2.1 to create the reference rendering |
| 11:35 | <annevk> | it's not like anything else would make sense |
| 12:23 | <Philip`> | Hixie: Could you have a generic "in select in <var>T</var>" and then say "process the token as if the insertion mode was 'in select in <var>table row</var>' etc? Then then text wouldn't be verbose, and it would be clear that you can implement it as lots of states instead of needing another variable to remember where you were |
| 12:23 | <Philip`> | s/' /'" / |
| 12:24 | <Philip`> | (I don't like variables, because it makes it hard to work out what's going on) |
| 12:45 | <Philip`> | It's a little odd how everyone seems to be congratulating/thanking Microsoft for deciding to make IE8 act like how everyone used to take for granted it was going to act |
| 12:50 | <zcorpan> | it's not uncommon to be thankful when someone decides against doing something bad even if you didn't expect the bad thing being planned from the beginning |
| 12:51 | <zcorpan> | but the questions on how the rendering engines interact across frames still remain |
| 12:51 | <zcorpan> | and security or crash fixes to the old engine |
| 12:53 | <Philip`> | They would have to do security fixes to IE6 and IE7 and IE8, regardless of the modes within each browser release, so I'd guess they will share code between IE7 and IE8's IE7-compatibility mode, so it's no more work than they'd need anyway |
| 13:06 | <Philip`> | Hixie: Is http://www.whatwg.org/specs/web-apps/current-work/multipage/section-parsing.html#determining broken, like where it says "2. Otherwise, return to step 2 in these inner steps." that seems to be an infinite loop? |
| 13:06 | <Philip`> | and the "two step" algorithm goes up to step 5 |
| 13:21 | <Philip`> | hsivonen_: Does your MetaSniffer code accurately / partially / not at all match the current spec? |
| 13:45 | <hsivonen_> | Philip`: it matches the spec as of June 2007 |
| 14:08 | <Philip`> | http://www.modellbausieghard.de/ - <meta http-equiv="content-style-type" content="text/css; charset=iso-8859-1" /> - the meta charset detection thing will get misled by that |
| 14:10 | <Philip`> | About 60% of sniffable charsets can be sniffed from the first 256 bytes, 85% from 512 bytes, 93% from 1024 bytes |
| 14:13 | <Philip`> | http://groups.msn.com/GraftonRifleClub/ sends "Content-Type: text/html; charset=utf-8;" which HTML5 thinks is a charset named "utf-8;" |
| 14:15 | <annevk> | what does HTML5 do with Content-Type:text/html;foo=bar;charset=utf-8;foo=bar; ? |
| 14:15 | <Philip`> | http://www.whatwg.org/specs/web-apps/current-work/multipage/section-content-type-sniffing.html#algorithm3 - it returns nothing if the first parameter is not charset |
| 14:17 | <Philip`> | I don't see any pages sending anything quite like that, but there's one group with "text/html;ISO-8859-1; charset=UTF-8" |
| 14:18 | <Philip`> | (and one page with "text/html;image/gif;image/jpeg; Charset=iso-8859-1;text/javascript") |
| 14:19 | <Philip`> | (That one page is only half the number that use apostrophes) |
| 14:19 | <Philip`> | (and only 21 use double-quotes) |
| 14:19 | <Philip`> | (around the charset string) |
| 14:20 | <zcorpan> | in emails, the charset parameter is often not the first one |
| 14:21 | <zcorpan> | and it would be nice if the same code could be used for parsing content-type for emails and for content-type of resources in a browsing context |
| 14:23 | <annevk> | heh, Hixie used a dotted line in http://www.whatwg.org/issues/data.html |
| 14:23 | <Philip`> | It would be nice if it parsed valid HTTP headers correctly when there's no reason not to |
| 14:24 | <Philip`> | Hixie: Is it a bug in your code that http://www.whatwg.org/issues/data.html gets the mouse position wrong if you scroll down the page, or is that Firefox's fault? |
| 14:25 | <annevk> | fault of the script |
| 14:25 | <annevk> | it should use pageX/Y instead of clientX/Y |
| 14:27 | <annevk> | or alternatively it could use getBoundingClientRect().top and getBoundingClientRect().left instead of offsetTop and offsetLeft |
| 14:28 | Philip` | finishes processing his pages, and find's it more like 57% getting sniffed in the first 256 bytes, 82% in 512, 92% in 1024 |
| 14:31 | <Philip`> | Does HTML5 say that if the server sends HTTP charset=utf-16, it should be ignored? |
| 14:32 | <Philip`> | Oops, I don't mean that |
| 14:32 | <Philip`> | I was looking at the meta charset thing, which does say that... |
| 14:34 | <Philip`> | http://www.vorem.com - "Content-Type: text/html; charset=UTF-32" |
| 14:34 | <annevk> | shit |
| 14:35 | <annevk> | i hoped utf-32 could be dropped |
| 14:35 | <Philip`> | The page is actually iso-8859-15 |
| 14:36 | <Philip`> | (or something close to that) |
| 14:36 | <Philip`> | The top unknown charsets are "", "none", "x-windows-874" and "iso-8559-1" |
| 14:36 | <annevk> | that makes it nice |
| 14:37 | Philip` | wonders who was trying to make a page with no character set |
| 14:39 | <annevk> | oh I see |
| 14:39 | <Philip`> | Oh, looks like it's probably Apache users with "AddDefaultCharset none" instead of "AddDefaultCharset Off" |
| 14:39 | <annevk> | vorem.com has two Content-Type headers |
| 14:39 | <Philip`> | I only see one... |
| 14:40 | <annevk> | never mind |
| 14:40 | <annevk> | GET -ed is not trustworthy |
| 14:40 | <annevk> | i should remember that |
| 14:41 | <Philip`> | Ah, that's probably the LWP module doing fancy automatic meta-http-equiv extraction |
| 14:41 | Philip` | uses 'HEAD http://www.vorem.com'; instead |
| 14:43 | <annevk> | why does Firefox 3 not show it as UTF-32? |
| 14:44 | <Philip`> | Does FF3 support utf-32? |
| 14:45 | <annevk> | the encoding meny claims it does |
| 14:45 | <annevk> | but maybe it's not enabled for Web content |
| 14:49 | <Philip`> | http://www.liniehoeve.nl/ - <meta name="Title" charset="Beagle Kennel van der Liniehoeve"> - ? |
| 14:49 | <Philip`> | I'm not sure how they managed to mix up 'charset' and 'content' |
| 14:50 | <annevk> | heh |
| 14:55 | <zcorpan> | so vorem.com is an example of where it would be better to not support utf-32 |
| 14:55 | <zcorpan> | that's interesting |
| 14:57 | <Philip`> | http://www.rotondadepacifico.com/ has <META http-equiv="Content-Type" content="text/html; charset=UTF-32"> |
| 14:57 | <Philip`> | Those are the only two I've found that claim to use UTF-32 |
| 14:57 | <Philip`> | (There's also two claiming to use UTF-9) |
| 14:58 | <Philip`> | (and also one ISO-10646-UTF-1 which actually seems to be a real encoding) |
| 15:05 | <zcorpan> | that uses a window-1252 copyright sign, so utf-32 doesn't need to be an alias for utf-8 like utf-16 is when found in <meta charset> |
| 15:06 | <annevk> | just needs to be ignored |
| 15:06 | <Philip`> | It's hard to make good decisions about how to handle UTF-32 pages based on a sample size of one :-) |
| 15:07 | <Philip`> | s/one/about two/ |
| 15:07 | <zcorpan> | indeed :) |
| 15:08 | <virtuelv> | is the feed for http://reddit.com/r/browsers/ something that should/could go on planet html5? |
| 15:09 | <virtuelv> | (that's also a general request to post HTML5-related contents that doesn't go on the planet there) |
| 15:31 | <annevk> | http://www.unicode.org/Public/MAPPINGS/ is interesting |
| 15:31 | <annevk> | virtuelv, talk to MikeSmith |
| 15:32 | <MikeSmith> | annevk, virtuelv - I can go ahead and add it |
| 15:34 | <annevk> | virtuelv, btw, planet HTML5 does do a content substring so not all entries would show up |
| 15:36 | <virtuelv> | annevk: I'd guess tagging entries with "HTML5" should suffice? |
| 15:36 | <virtuelv> | fwiw, everyone is free to post there |
| 15:36 | <virtuelv> | I just think it's a good place to catch what doesn't get caught by the planet |
| 15:39 | <virtuelv> | btw, I think the filtering on planet html5 may be a bit strict |
| 15:39 | <virtuelv> | for instance, it totally missed MS' announcment |
| 15:40 | <annevk> | how is that HTML5 news? |
| 15:40 | <virtuelv> | hm, you might be right |
| 15:41 | <virtuelv> | I guess I just think the planet is a bit unpopulated |
| 15:42 | <annevk> | true, but html5 is not exactly daily news |
| 15:42 | <annevk> | well, except in this channel :) |
| 15:46 | <MikeSmith> | virtuelv - I added it |
| 15:46 | <MikeSmith> | btw, filtering on Planet HTML5 is very simple |
| 15:47 | <MikeSmith> | it looks for the strings "HTML5" or "HTML 5" anywhere in the content of feed item |
| 15:47 | <virtuelv> | MikeSmith: notede |
| 15:47 | <virtuelv> | -e |
| 15:47 | <MikeSmith> | so I guess the MS announcement didn't mention HTML5 at all |
| 15:47 | <virtuelv> | it didn't |
| 15:49 | <zcorpan> | MikeSmith: what about "HTML 5"? |
| 15:50 | <zcorpan> | <abbr>HTML</abbr> 5 |
| 15:50 | <Philip`> | HTML<ins> </ins>5 |
| 15:50 | <MikeSmith> | any blog posting that does <abbr>HTML</abbr> will be summarily rejected |
| 15:51 | <zcorpan> | so much for semantics ;) |
| 15:52 | <Philip`> | abbr/acronym are fatally flawed since they can't represent recursive acronyms |
| 15:56 | <Philip`> | We need <abbr name="gnu">GNU</abbr> (<expansion for="gnu"><abbr name="gnu">GNU</abbr>'s not <abbr name="unix">UNIX</abbr></expansion>) ... |
| 15:58 | <zcorpan> | "GNU (GNU's not UNIX)" ... |
| 15:59 | <Philip`> | Plain English won't work for mutually recursive acronyms Hurd - you'd need to add ugly explanations around it to say what's going on, rather than elegantly encoding the expansion graph using metadata markup |
| 15:59 | <Philip`> | s/ / like / |
| 16:01 | <zcorpan> | who's going to benefit from the metadata markup? |
| 16:01 | <zcorpan> | and is it easier to understand than plain english? |
| 16:02 | <Philip`> | Google could use the metadata so you can ask it for "define:hurd" and it will give you the correct answer |
| 16:02 | <virtuelv> | outside of GNU and (historically) PHP, how many recursive initialisms are there? |
| 16:02 | <zcorpan> | Philip`: what is the answer? |
| 16:02 | <Philip`> | http://en.wikipedia.org/wiki/Recursive_acronym - lots :-) |
| 16:03 | <virtuelv> | "The GNU Hurd project is named with a mutually recursive acronym: "Hurd" stands for "Hird of Unix-Replacing Daemons," and "Hird" stands for "Hurd of Interfaces Representing Depth." |
| 16:03 | <Philip`> | zcorpan: It's "<a href=search?q=define:Hird>Hird</a> of Unix-Replacing Daemons" |
| 16:03 | <virtuelv> | I think I can live with not being able to express those :P |
| 16:05 | <Philip`> | virtuelv: But there may be other people in the world who cannot live with such a dire situation |
| 16:06 | <zcorpan> | <dl><dt><dfn><abbr>Hurd</abbr></dfn><dd><abbr>Hird</abbr> of Unix-Replacing Daemons<dt><dfn><abbr>Hird</abbr></dfn><dd><abbr>Hird</abbr> of Interfaces Representing Depth</dl> |
| 16:07 | <zcorpan> | elsewhere: <p>... <abbr>Hurd</abbr> ... |
| 16:09 | <Philip`> | That breaks down if Hird can stand for multiple things on the same page |
| 16:10 | <Philip`> | and e.g. for feed readers which don't know if one term is going to be used multiple times, so they have to expand everything out into title attributes just in case |
| 16:11 | <Philip`> | (Er, feed aggregators in particular) |
| 16:11 | <zcorpan> | use title='' to disambiguate |
| 16:11 | zcorpan | doesn't read carefully |
| 16:12 | <Philip`> | <abbr title='Hurd of ...'>Hird</abbr> would lose the markup around Hurd |
| 16:12 | <Philip`> | so I don't see how you'd disambiguate it |
| 16:13 | <zcorpan> | <span title><abbr> |
| 16:13 | <zcorpan> | :P |
| 16:13 | <Philip`> | Oh, it seems <dfn title="hurd-1"><abbr>Hurd</abbr></dfn> would do useful things |
| 16:13 | <Philip`> | (in terms of disambiguating) |
| 16:14 | <Philip`> | Automatic cross-referencing still seems a bad idea since it introduces lots of invisible dependencies between separate parts of the page, though |
| 16:15 | <zcorpan> | why would they be invisible? |
| 16:15 | <Philip`> | They're invisible in the source |
| 16:15 | <zcorpan> | i guess |
| 16:15 | <Philip`> | and so they're likely to break if you copy-and-paste chunks of one document into another |
| 16:16 | <Philip`> | or if you concatenate two documents |
| 16:16 | <zcorpan> | seems like we're searching for problems to make our solution more complicated, when what we really need is gloves |
| 16:17 | <Philip`> | The indirection between elements and <style>s causes enough problems when you're trying to mix documents, so I'd probably be happier if no more indirections were added |
| 16:21 | <annevk> | did IE support UTF-32? |
| 16:21 | <zcorpan> | no |
| 16:21 | <zcorpan> | afaict |
| 16:22 | <zcorpan> | it rendered vorem.com like firefox |
| 16:43 | <gsnedders> | I wonder how much broken XHTML we'll get if IE8 supports it. |
| 16:44 | <webben> | well, MS already have a halfway usable XML parser and a halfway usable MSAA DOM representation and a quarter-way usable DOM |
| 16:45 | <webben> | so I wouldn't have thought it would be all that bad |
| 16:45 | <webben> | oh wait, you mean broken XHTML from authors |
| 16:45 | <gsnedders> | webben: yeah |
| 16:45 | <webben> | I guess that depends on what IE did with borked XHTML |
| 16:46 | <gsnedders> | I mean, every man and his dog will probably try and jump on XHTML |
| 16:46 | <gsnedders> | even with fatal error handling I bet a heckuva lot of broken XHTML will appear |
| 16:46 | <webben> | producing non-broken XHTML isn't an especially difficult task, if you're producing XHTML at all. |
| 16:47 | <webben> | I suspect many of the situations where you're doomed to produce broken XHTML, you're going to produce it sooner rather than later |
| 16:47 | <webben> | e.g. as soon as you have a page with an ad. |
| 16:48 | <webben> | at which point you're probably going to revert to text/html |
| 16:48 | <gsnedders> | I wait with abated breath |
| 16:49 | <webben> | I suspect very few people will switch because so few people understand that serving XHTML tends to involve server configuration as well as just changing a META element. |
| 16:52 | <annevk> | "producing non-broken XHTML isn't an especially difficult task, if you're producing XHTML at all." you should check with the very few people who try |
| 16:55 | <webben> | I have tried, I switched back to HTML because it was pointless. |
| 16:55 | <webben> | I still don't think it's especially difficult, FWIW. |
| 16:56 | <webben> | the problem is one of quality control more generally; XHTML just exposes how appalling poor quality control on the web is |
| 16:57 | <webben> | (that is, it was pointless for me, as I don't need Ruby/SVG/MathML, etc.) |
| 17:05 | <zcorpan> | using xhtml because of needing ruby annotations is a bit backwards |
| 17:12 | <webben> | zcorpan: Other than the fact that Moz only supports them with a plugin and IE only supports them in text/html, I don't see how. |
| 17:15 | <zcorpan> | webben: those were the reasons i had in mind |
| 17:15 | <webben> | I wouldn't call that "backwards" exactly. |
| 17:16 | <zcorpan> | ie is the only browser with native ruby support, ie doesn't support xhtml |
| 17:16 | <zcorpan> | hence, using ruby in xhtml doens't work for anyone |
| 17:16 | <zcorpan> | using ruby in html works for ie |
| 17:16 | <webben> | where anyone excludes people with Fx and the Ruby extension? |
| 17:16 | <zcorpan> | (assuming that "no-one" uses mozilla with the plugin) |
| 17:17 | <webben> | I don't think that's a sensible assumption. |
| 17:17 | <zcorpan> | in any case, i bet firefox with the extension supports ruby in both html and xhtml |
| 17:18 | <webben> | Probably, but then one could just content-negotiate for IE. |
| 17:18 | <zcorpan> | and so using xhtml *because* of ruby (like i said) is, if not backwards, very unnecessary |
| 17:18 | <webben> | XHTML is the only standard way to publish ruby annotations. |
| 17:19 | <webben> | I'd be happy to use a non-standard way as a hack to support the most popular browser on earth. |
| 17:19 | <zcorpan> | that will be fixed in due course when html5 includes ruby |
| 17:20 | <webben> | That day is not today. |
| 17:21 | <webben> | and a couple years is a long time |
| 17:21 | <zcorpan> | meanwhile you can be non-compliant with useless specs :) |
| 17:21 | <webben> | It's no more or less useless than HTML5 will be. |
| 17:21 | <webben> | A spec's a spec. |
| 17:22 | <zcorpan> | specs can surely be useless |
| 17:22 | <zcorpan> | html4 is useless |
| 17:22 | webben | doesn't agree. |
| 17:22 | <zcorpan> | let's leave it at that |
| 17:23 | zcorpan | should go home |
| 17:27 | <Philip`> | webben: It appears to be very difficult to produce non-broken XHTML when you include content from potentially-malicious users |
| 17:27 | <Philip`> | e.g. if you commit to a Git repository with a message that contains U+FFFE, then the web interface will become ill-formed and only IE users will be able to read it |
| 17:28 | <webben> | Philip`: Yes, but once you've included malicious content your site is probably a public health hazard. |
| 17:28 | <webben> | that's what I mean by grouping this stuff under general quality control |
| 17:28 | <webben> | if you're not in control of your markup, then you're probably not effectively stopping XSS etc |
| 17:29 | <Philip`> | This is when people just do s/&/&/ and s/</</ (like what their language / template system's HTML escape function does) |
| 17:29 | <Philip`> | which is perfectly safe in HTML |
| 17:29 | <Philip`> | s/people/developers/ |
| 17:30 | <Philip`> | and maybe replace funny characters with &#nnn; too |
| 17:30 | <webben> | I'm not sure what "This" refers to |
| 17:30 | <webben> | but this all sounds like it comes under the rubric of input validation |
| 17:30 | <webben> | which as we all know is generally poorly done |
| 17:31 | <webben> | XHTML just makes that visible in a really annoying way. |
| 17:31 | <Philip`> | "This" = "The situation where malicious users can break the web site of someone who's tried to be careful at escaping input, but hasn't been quite careful enough for XML" |
| 17:31 | <webben> | hmm yes, well, there's been a lot more energy put into developing solutions for escaping HTML than XML. |
| 17:32 | <webben> | for understandable reasons |
| 17:32 | <Philip`> | XML escaping requires much more energy than HTML, because there are more ways it can go wrong |
| 17:33 | <webben> | I'm not sure that's true. |
| 17:34 | <Dashiva> | Well, you need some way to ensure it's well-formed, that's a big step up just by itself |
| 17:34 | <webben> | if you're parsing the input anyways, I'm not sure it is. |
| 17:35 | <Dashiva> | If |
| 17:36 | <annevk> | that would require the usage of a compliant XML parser |
| 17:36 | <annevk> | that's a big step up |
| 17:36 | <Philip`> | I'm mostly thinking of just plain text (not trying to allow a 'safe' subset of HTML or anything fancy like that), like when you say "You search for <? escape_html(query) ?> returned <? n ?> results", and escape_html in HTML is quite trivial because you care about & and <, but in XHTML it's hard because there's lots of forbidden characters |
| 17:36 | <Philip`> | s/You/Your/ |
| 17:37 | <webben> | annevk: For input validation? Depends on the sort of input; for (X)HTML-esque input I'd probably use a tag soup parser, strip unwanted stuff, and then reserialize into the database. |
| 17:39 | <webben> | annevk: and I'd do that whether I was going to serve HTML or XHTML. |
| 17:40 | <Philip`> | (and escape_xhtml is clearly hard because I haven't yet found a web site that does it correctly - they all forget about some of the characters that can make their output ill-formed) |
| 17:40 | <webben> | Has anyone compiled a simple list of them? |
| 17:40 | <webben> | (of the characters ...) |
| 17:41 | <annevk> | it's not about what you do, it's about what people are doing now and where they have to move |
| 17:41 | <annevk> | unless you changed the rules |
| 17:42 | <Philip`> | webben: There's http://www.w3.org/International/questions/qa-forms-utf-8 but that's wrong because it misses FFFE and FFFF |
| 17:42 | <webben> | annevk: Most people aren't producing XHTML now. |
| 17:42 | <Philip`> | (according to http://www.intertwingly.net/blog/2008/01/02/Keeping-On-Your-Toes ) |
| 17:43 | <webben> | annevk: Even if you were to count the people who produce tag soup XHTML, which I don't really. |
| 17:43 | <annevk> | that's besides the point |
| 17:43 | <annevk> | (both) |
| 17:44 | <webben> | what are the people who are producing real XHTML doing now? |
| 17:44 | <annevk> | they fail and plug holes, mostly |
| 17:47 | <webben> | Philip`: I don't fully grok that ... the W3C page looks like a regex for sniffing UTF-8 rather than a simple list of characters not allowed in XHTML. |
| 17:48 | <webben> | annevk: Sounds similar to what happens when we try to produce HTML. |
| 17:49 | <webben> | just with a different failure result. |
| 17:50 | <annevk> | sure, but it doesn't affect the end user |
| 17:50 | <webben> | yes it does |
| 17:50 | <webben> | when their data is stolen it definitely does! |
| 17:51 | <webben> | when the page renders but doesn't work properly |
| 17:51 | <webben> | when characters are rendered incorrectly |
| 17:51 | <webben> | etc. |
| 17:52 | <annevk> | those are not the result of using HTML incorrectly |
| 17:52 | <gsnedders> | webben: they can equally happen with XHTML |
| 17:52 | <annevk> | that too |
| 17:53 | <webben> | gsnedders: Of course. That's not my point. |
| 17:53 | <annevk> | your point got lost to me 10 minutes ago |
| 17:53 | <webben> | gsnedders: My point is rather that publishers' problems with XHTML are in part a warning signal that they don't have enough control over markup quality to protect user experience and data. |
| 17:55 | <webben> | XHTML doesn't cure the problem or anything. |
| 17:58 | <Philip`> | webben: About regex thing: Hmm, true, I guess it's not entirely relevant - it does forbid the control characters below 0x20 that XML doesn't allow, but it doesn't say it's doing that on purpose, and it wouldn't work in a proper Unicode string environment |
| 17:58 | <Philip`> | http://www.w3.org/TR/REC-xml/#charsets lists the actual characters that are allowed, which is more useful when trying to write code |
| 18:01 | <webben> | Philip`: Right, and if folks just used /that/ then they wouldn't have probs with FFE and FFFF. |
| 18:01 | <webben> | *FFFE |
| 18:02 | <Philip`> | It seems that in practice people don't use that, or they use it on the wrong side of a UTF-8 <-> Unicode conversion, or do something else wrong |
| 18:03 | <Philip`> | Maybe it's just something that the world needs to be taught before using XHTML |
| 18:03 | <Philip`> | and which hasn't been taught yet |
| 18:03 | <webben> | well, it's not just XHTML |
| 18:03 | <webben> | also RSS etc |
| 18:04 | <Philip`> | (the same as how it's taken a while for people to realise that SQL injection and XSS etc are problems, but they're fairly well known now among certain groups of people) |
| 18:04 | <webben> | Indeed. |
| 18:04 | <webben> | That's very much along the lines of what I was saying. |
| 18:06 | <webben> | I wonder how much of the problem does or doesn't stem from poor language support for unicode too. |
| 18:06 | <webben> | And bad defaults. |
| 18:06 | <webben> | e.g. PHP now has filter: http://uk.php.net/filter |
| 18:06 | <webben> | but it's off by default because turning it on would break a lot of the web |
| 18:18 | <gsnedders> | what's browser support of IRIs like (esp. in paths)? |
| 18:21 | Philip` | sees that Opera doesn't like http://.../...#%65 etc, but Firefox does, and wonders who's wrong |
| 18:21 | <annevk> | Opera |
| 18:21 | <annevk> | known issue |
| 18:21 | <Philip`> | Oh, works in 9.5 |
| 18:21 | Philip` | doesn't fix his code, in that case |
| 18:22 | <annevk> | fixed issue, even, nice :) |
| 18:22 | annevk | needs to keep up |
| 19:58 | <Hixie> | Philip`: i can't find a step 2 that says to go to step 2, and the two step algorithm does have two steps. (well 7, if you include the 5 inner steps of step 1) |
| 20:02 | <Philip`> | Hixie: It's "1. If position points to: <stuff with two "a sequence" things, the second with 5 steps> 2. Otherwise, return to step 2 in these inner steps." |
| 20:03 | <Philip`> | and then it has "A sequence of bytes starting with a 0x3C byte (ASCII '<'), optionally a 0x2F byte (ASCII '/'), and finally a byte in the range 0x41-0x5A or 0x61-0x7A (an ASCII letter)" outside the list it should probably be in (introduced by "If position points to:") |
| 20:04 | <Hixie> | uh |
| 20:04 | <Hixie> | yes clearly that step 2 is wrong |
| 20:04 | <Philip`> | perhaps because the "case-insensitive ASCII '<meta' followed by a space or slash" bit has an <ol> with a <p> child |
| 20:04 | <Hixie> | oh |
| 20:04 | <Hixie> | i see |
| 20:04 | <Hixie> | massive markup error |
| 20:04 | <Hixie> | will gix |
| 20:04 | <Hixie> | fix |
| 20:06 | <annevk> | not catched by the validator? |
| 20:06 | <Hixie> | the preprocessor "fixed" it |
| 20:07 | <annevk> | ah yeah, that happens |
| 20:07 | <annevk> | maybe run it over the source? |
| 20:07 | <Philip`> | http://validator.nu?doc=http://www.whatwg.org/specs/web-apps/current-work/source |
| 20:07 | <Philip`> | annevk: Oh, good idea |
| 20:07 | <annevk> | well, you need to prefix the source with something first I guess |
| 20:08 | <Philip`> | I assume it'd recover after complaining it was missing the doctype and title |
| 20:08 | <Hixie> | ok i should have fixed that list issue |
| 20:08 | <Hixie> | looks now |
| 20:08 | <Hixie> | i'll fix the rest later |
| 20:09 | <Philip`> | so just ignore the first four messages and fix the next near-infinite errors |
| 20:09 | <Philip`> | Hixie: Okay, thanks! |
| 20:09 | <annevk> | Philip`, true |
| 20:10 | <annevk> | heh |
| 20:10 | <annevk> | "Fatal Error: Too many messages." |
| 20:10 | <annevk> | hsivonen, could you add "HTML is hard." to that as easter egg? :) |
| 20:11 | <Hixie> | hsivonen: you output three messages when one end tag has been substituted for another |
| 20:11 | <Hixie> | as in, <span>...</code>...</p> |
| 20:12 | <Hixie> | missing </span>, unexpected </code>, and <span> still open at </p>. |
| 20:32 | <Hixie> | in fact the >1000 errors are only at most 50 or so |
| 20:32 | <Hixie> | all fixed now |
| 20:37 | <Hixie> | is there an output format for the validator that outputs text/plain? |
| 20:38 | <annevk> | there's documentation on the whatwg wiki |
| 20:38 | <annevk> | (there is, just not sure how to get it) |
| 20:38 | <takkaria> | Hixie: http://wiki.whatwg.org/wiki/Validator.nu_Web_Service_Interface |
| 20:39 | <Hixie> | cool thanks |
| 20:39 | <Philip`> | &out=text |
| 20:39 | <Hixie> | out=gnu is probably what i want |
| 20:39 | <Hixie> | well maybe out=text is ok |
| 20:41 | <annevk> | for out=gnu you could simply check if the amount of lines is greater than 4 |
| 20:41 | <Hixie> | out=text uses wacky ass quotation marks |
| 20:42 | <Hixie> | annevk: ? |
| 20:42 | <annevk> | if the return value has more than 4 lines you have hit a regression |
| 20:42 | <annevk> | and need to fix your markup |
| 20:42 | <Hixie> | nope. out=text still uses silly quotation marks |
| 20:42 | <Hixie> | hsivonen: any chance i can get a US-ASCII output mode? :-) |
| 20:43 | <Hixie> | heh, the validator caught my en-GB-hixie error |
| 20:43 | <Hixie> | guess i'd better fix it |
| 20:45 | <annevk> | Hixie, I'm talking about out=gnu, it uses those quotation marks too, but i'm not sure why that's relevant if you simply want to catch regressions |
| 20:46 | <hsivonen> | Hixie: out=gnu is an UTF-8 output mode. (It'll change a bit soon as the GNU spec is amended.) UTF-8 is the new US-ASCII. ;-) |
| 20:47 | <hsivonen> | Hixie: out=text is for human reading. you shouldn't try to write a program to parse it. |
| 20:47 | Hixie | accidentally outputs the w3c version to the whatwg site |
| 20:48 | <Hixie> | annevk: yeah i was just confused about why you wanted to skip 4 lines? shouldn't i want zero errors? |
| 20:49 | <hsivonen> | Hixie: how seriously do you need a US-ASCII mode and what should it do with non-ASCII characters? |
| 20:49 | <annevk> | Hixie, oh, I was assuming you wouldn't prefix the source file with the WHATWG header file |
| 20:49 | <Hixie> | hsivonen: i don't want to parse it. i have my emacs configured to US-ASCII with any 8 bit bytes displayed in octal, so that i can easily edit test cases with malformed byte sequences |
| 20:49 | <Hixie> | hsivonen: so your quotation marks look like \123\123\123 |
| 20:49 | <Hixie> | really it's just a "no fancy quotes" mode that i'd like :-) |
| 20:50 | <Hixie> | annevk: in my script it's already done, so that's not a problem |
| 20:51 | <hsivonen> | Hixie: out=gnu is meant to be emacs-friendly, but only when emacs is configured to do UTF-8 |
| 20:52 | <Hixie> | specifically i get: |
| 20:52 | <Hixie> | http://www.whatwg.org/specs/web-apps/current-work/source-whatwg:1.51-2.25: error: Bad value \342\200\234en-GB-hixie\342\200\\ |
| 20:52 | <Hixie> | 235 for attribute \342\200\234lang\342\200\235 on element \342\200\234html\342\200\235: Bad variant subtag. |
| 20:52 | <Hixie> | i guess it's not a big deal, just looks ugly and makes it hard to work out what the error is |
| 20:53 | <hsivonen> | Hixie: can you give me a spec for some processing magic before the UTF-16 to UTF-8 conversion? |
| 20:53 | <hsivonen> | Hixie: just replacing the quotes? |
| 20:54 | <hsivonen> | Hixie: out of curiosity, what code do you use for integrating emacs and the Web service? |
| 20:58 | <Hixie> | curl |
| 20:58 | <Hixie> | and yeah, just replacing quotes with " would solve my immediate problem |
| 21:03 | <met_> | checking acid3 in last Webkit, strange that test 78 sometimes fails, otherwise not. So I have 87/100, but from time to time it ends 86/100 |
| 21:03 | <met_> | from console it's Test 78: expected: 90, got: 0 - getRotationOfChar(0) failed. |
| 21:03 | <hsivonen> | Hixie: ok. |
| 21:04 | <Hixie> | met_: yeah it's probably not always loading the font |
| 21:05 | <hsivonen> | Hixie: it's http://bugzilla.validator.nu/show_bug.cgi?id=135 now |
| 21:06 | <Hixie> | do the fancy quotes provide any direct benefit to users? |
| 21:06 | <Hixie> | apart from looking arguably prettier in certain contexts? |
| 21:07 | <Hixie> | because frankly they also cause me other problems, like i can't easily copy and paste them across ssh sessions |
| 21:07 | <hsivonen> | Hixie: they disambiguate start and end better and they don't themselves occur in XML/HTML syntax |
| 21:08 | <Hixie> | fair enough |
| 21:14 | <met_> | Hixie, yeh and it differs from time to time, now failed other svg test -79. It's difficult co compare, when other browsers doesn't support remote fonts |
| 21:15 | <annevk> | met_, Opera 9.5 does, no? Although we have some failures there as well I think |
| 21:16 | <met_> | annevk, Opera 9.5 support remote fonts? Holly father, I have missed it. |
| 21:17 | <gsnedders> | interesting. I have hits from a browser claiming to be Firefox 9.0 |
| 21:23 | <annevk> | met_, I believe we do, but I no next to nothing about SVG |
| 21:23 | <annevk> | s/no/know/ |
| 21:24 | <met_> | annevk, just downloading last opera weekly to check |
| 21:33 | <met_> | annevk, no examples from http://www.alistapart.com/articles/cssatten still don't work in public build |
| 21:34 | gsnedders | decides he better not describe himself as Hixie's minion protégé in a silly autobiography, as someone is bound to take it seriously |
| 21:35 | <annevk> | met_, I thought you were talking about SVG specifically |
| 21:35 | <annevk> | met_, there's no "generic" support for downloadable fonts |
| 21:35 | <met_> | from the acid test yes, i have tried it generally now, so for svg should work? |
| 21:35 | <met_> | ok |
| 21:40 | <roc> | hmm, should I fix IFRAME drawing inside SVG filters on Mac, just so I can adapt reftest-analyzer to work with live testcases? |
| 21:40 | <roc> | yeah! |
| 21:54 | <eseidel> | ? |
| 21:54 | <eseidel> | ah, he's talking about FF SVG |
| 21:55 | <jgraham> | eseidel: With that kind of enthusiasm, who cares about the tangentialness of the relevance ;) |
| 21:57 | eseidel | would like someone to fix rightsizing in webkit |
| 21:58 | <roc> | oops wrong channel |
| 21:59 | <roc> | but it *is* kind of cool that with SVG filters, IFRAMEs and foreignobject you can create a pixel-level page comparison tool in a Web app |
| 21:59 | <annevk> | rightsizing?! |
| 21:59 | annevk | was hoping that terminology would not be used outside the CDF WG |
| 22:00 | <roc> | oh, have they got a spec for that? |
| 22:00 | jgraham | wonders what rightsizing is if it's not a euphemism for sacking everyone |
| 22:00 | <eseidel> | annevk: we need a term for those bugs |
| 22:01 | <roc> | I think it means "give IFRAMEs a meaningful intrinsic size" (or height at least) |
| 22:01 | <eseidel> | so I stole the CDF WG's term |
| 22:01 | <eseidel> | oh, no that's another set of bugs roc |
| 22:01 | <eseidel> | roc: that's just implementing CSS 2.1 :) |
| 22:01 | <roc> | what is? |
| 22:02 | <eseidel> | CSS 2.1 requires inline replaced elements to support relative intrinsic sizes |
| 22:02 | <eseidel> | which we in webkit at least don't support |
| 22:02 | <roc> | you mean the aspect-ratio-preserving stuff? |
| 22:02 | <eseidel> | so things like <object> don't corretly pick up the height="100%" off of their svg child doc |
| 22:02 | <othermaciej> | I don't think current browsers ever give <iframe> a useful intrinsic size |
| 22:02 | <eseidel> | roc: the aspect-ratio-preserving stuff is "right sizing" |
| 22:02 | <othermaciej> | even if you put a fixed-size raster image in it |
| 22:03 | <othermaciej> | so it's not really the same issue as <object? |
| 22:03 | <othermaciej> | > |
| 22:03 | <roc> | CSS2.1 specifies that IFRAME intrinsic sizes should depend on the child document? news to me |
| 22:03 | <annevk> | eseidel, aspect ratio is also CSS 2.1 |
| 22:03 | <othermaciej> | (it would be handy if <iframme> could size appropriately to a child HTML document, but you'd need some way to ask for that in CSS) |
| 22:03 | <annevk> | eseidel, in fact, if it's not, rightsizing would be a violation of CSS 2.1, which is not what you want... |
| 22:04 | <roc> | yeah, that's what I want (and the Web needs I think) |
| 22:04 | <annevk> | (having said that, CSS 2.1 should just say whatever is most fancy and works) |
| 22:04 | <roc> | I actually had a patch for it in Gecko, but there's a concern about information leakage |
| 22:04 | <roc> | for cross-domain frames |
| 22:05 | <eseidel> | rightsizing entered into my lingo from the WICD tests http://bugs.webkit.org/show_bug.cgi?id=15836... I might be using it wrong though |
| 22:05 | <eseidel> | annevk: this is the CSS 2.1 bug of which I speak: http://bugs.webkit.org/show_bug.cgi?id=15849 |
| 22:05 | <eseidel> | which causes us to fail hixie's tests: http://bugs.webkit.org/show_bug.cgi?id=15473 |
| 22:18 | <gsnedders> | Hixie: with <http://bugs.webkit.org/show_bug.cgi?id=17023>; you better paste it all again with the changes :P |
| 23:08 | <jgraham> | Highly unpolished and quite probably wrong at least sometimes: http://james.html5.org/outliner.html |
| 23:14 | <Hixie> | does it work on the spec? :-) |
| 23:15 | <jgraham> | Hixie: Define work :) |
| 23:15 | <annevk> | cool stuff |
| 23:16 | annevk | plyed with <h1>1</h1><h5>2</h5><section><h3>3</h3><section><h2>4 |
| 23:16 | <annevk> | played, even |
| 23:19 | <jgraham> | Oh and fwiw, I have serious doubts about making <td> a sectioning root element. My guess is that that will make the algorithm useless for a significant amount of real-world content |
| 23:19 | <annevk> | the algorithm is already useless for that if it only takes into account <h1>-<h6> |
| 23:20 | <annevk> | for real world content you want to base this stuff on computed style information |
| 23:20 | <jgraham> | annevk: A non-trivial number of pages use <h1>-<h6> in a semi-sane way |
| 23:20 | <annevk> | so given that you might as well discourage table for layout madness |
| 23:21 | <annevk> | jgraham, inside <table>? |
| 23:21 | <Hixie> | jgraham: i would agree, if it wasn't for the problem of legitimate uses of headers inside cells being cases where you _would_ want it treated a sectioning root |
| 23:21 | <jgraham> | Hixie: Example? |
| 23:22 | jgraham | can't really think of why you would want headers in tables at all |
| 23:22 | <Hixie> | a character sheet where one of the cells is the characters's background story |
| 23:23 | <jwalden> | a page using tables for layout |
| 23:23 | <jgraham> | And the story has multiple subheadings? |
| 23:23 | <Hixie> | yeah |
| 23:23 | <Hixie> | jwalden: we already know that's non-conforming :-) |
| 23:24 | <jwalden> | just stating the practical response :-) |
| 23:24 | <jgraham> | jwalden: That's the case that is currently unsupported by the algorithm |
| 23:24 | <jwalden> | er, pragmatic |
| 23:24 | <jgraham> | and that I think needs to be supported |
| 23:25 | <Hixie> | if i made it a sectioning element it still wouldn't work |
| 23:25 | <Hixie> | for real content |
| 23:26 | <jgraham> | I think for real content it would have to be ignored. |
| 23:27 | <Hixie> | right, which isn't compatible with conforming content |
| 23:27 | <jgraham> | I'm not sure how to deal with the character sheet example other than to say "don't use a table for that", which seems a stretch as a table isn't obviously wrong |
| 23:27 | <Hixie> | and as much as we want to make legacy pages work, valid semantic pages have to be a higher priority imho |
| 23:29 | <jgraham> | I guess it would be nice to look at how headings as descendants of table content are actually used in the real world |
| 23:31 | <jgraham> | On an (not) entirely different topic, is there a good reason to call the current outline element /current outlinee/ rather than, say, /current outline element/ |
| 23:33 | <jgraham> | which looks less like a typo and prevents me from having to make the same translation every time in my head |
| 23:33 | <Hixie> | heh |
| 23:33 | <Hixie> | well it's not the "current outline element" |
| 23:34 | <Hixie> | it's the "element that is currently having an outline created for it" |
| 23:34 | <Hixie> | i don't mind changing it to something else if you have a suggestion that's correct :-) |
| 23:34 | <Hixie> | so supporting <table><select><table> turned out way simpler than i expected |
| 23:35 | <Hixie> | i must have misunderstood something |
| 23:39 | <jgraham> | Oh I hadn't read outlinee like that at all. I guess that's wwhat happens when you use made up words even if they do follow standard morphological rules :) |
| 23:40 | <Hixie> | :-) |
| 23:42 | <annevk> | <input> closes <select>, interesting |
| 23:45 | <Hixie> | yeah a webkit bug was just filed on this |
| 23:46 | <annevk> | btw, browsers other than IE7 also reprocess <select> when found in "in select" |
| 23:47 | <annevk> | (tested Firefox and Opera) |
| 23:47 | <annevk> | nn |
| 23:47 | annevk | -> bed |
| 23:49 | <Hixie> | doesn't the spec already say that? |
| 23:51 | <Hixie> | heh, i think my script handles "<!DOCTYPE html> <table border> <b><p>a</b> <td> b" as we'd want because it pushes a <#bucket> node onto the stack when doing foster parenting |
| 23:59 | <Hixie> | jgraham: i pasted the html5 spec into the outliner algorithm and it didn't do anything (After hanging safari for a few minutes) |
| 23:59 | <Hixie> | just said "updating..." |