| 00:01 | <jwalden> | anyway, Hixie: sounds like |document.domain| should probably be the empty string in contexts where there's no host, and likewise for |MessageEvent.domain| -- I'll send an email to the list about this as a reminder |
| 00:01 | <SadEagle> | fredrikh: if you look at the skewed example, that seems to be right on the interior,I think |
| 00:04 | <Hixie> | it would be funny if browsers looked for the "UA compat" switch and broke rendering if they saw it |
| 00:04 | <Hixie> | (regardless of the value) |
| 00:04 | <Hixie> | jwalden: thanks |
| 00:05 | hober | contemplated telling Emacs to highlight it in `font-lock-error-face' |
| 00:05 | jwalden | also remembers to send the email about open() and origins, while he's at it |
| 00:06 | <SadEagle> | Hixie: making a bug icon crawl over it may be less drastic |
| 00:21 | <Hixie> | i think we need to drop non-native ImageData |
| 00:21 | <Hixie> | and just have a createImageData() method |
| 00:21 | <Hixie> | and require manipulation of the native ImageData objects |
| 00:21 | <Hixie> | for perf |
| 00:21 | <Hixie> | (based on discussions with various people) |
| 00:30 | <SadEagle> | Hixie: the KJS garbage collector (and I am sure the JSC one) like that idea |
| 00:32 | <SadEagle> | Hixie: oh, and am I imagining things, or did html5 have a rule saying to raise an exception in JS if there were not enough arguments... is that gone now? |
| 00:33 | <Hixie> | that kind of contern is gonna be part of dom bindings, i believe |
| 00:33 | <SadEagle> | that spec seem to have the machinery for describing stuff right now, but not the actual info.. |
| 00:34 | <Hixie> | yeah |
| 00:35 | <othermaciej> | SadEagle: I don't think GC is the biggest problem with the current design |
| 00:35 | <SadEagle> | othermaciej: think of GC marking a million-element array.. |
| 00:35 | <othermaciej> | just serialize/deserialize overhead |
| 00:35 | <Philip`> | SadEagle: Think of calculating the contents of a million-element array |
| 00:36 | <Philip`> | (in JavaScript) |
| 00:36 | <SadEagle> | well, my current impl doesn't use a native array object anyway |
| 00:36 | <Philip`> | (and preferably with interesting contents, so you have to do lots of work per pixel anyway) |
| 00:37 | <SadEagle> | Philip`: sure. Just the GC has an additional complication of being hard-to-interrupt native code. |
| 00:38 | <SadEagle> | othermaciej: the way I just did in khtml is to have a special object for ImageData.data, and then if one calls putImageData with a non-ImageData it does the deserialize mess |
| 00:43 | <othermaciej> | SadEagle: yeah, that wouldn't be great (though not totally tragic if they are all immediate values) |
| 00:43 | <Hixie> | http://blogs.zdnet.com/open-source/?p=1925 -- apparently we started in october 2005 |
| 00:44 | <Hixie> | i actually thought we started around november 2003, but you know, details. |
| 01:11 | <othermaciej> | Hixie: that createElement parsing trick is really good news |
| 01:11 | <othermaciej> | Hixie: now if Firefox would just fix parsing of blocks inside unknown elements... |
| 01:12 | <Hixie> | yeah |
| 02:19 | <takkaria> | I wonder how much effort it would take to fix the parsing of blocks inside unknown elements |
| 02:19 | <takkaria> | (in Fx) |
| 02:26 | <Ketsuban> | I love the prevailing opinion in #html. They seem to be pretty heavily in favour of XHTML 2 over HTML5. :P |
| 02:28 | <Hixie> | woe is us |
| 02:28 | <Hixie> | whatever shall we do |
| 02:29 | <Ketsuban> | Maybe we should all post in our LiveJournals about it. :P |
| 03:01 | <AwayEagle> | Hixie: is any browser vendor planning on supporting xhtml2? |
| 03:22 | Philip` | notices that the number of comments on http://blogs.msdn.com/ie/archive/2008/01/21/compatibility-and-ie8.aspx is rapidly approaching the number on http://blogs.msdn.com/ie/archive/2007/12/19/internet-explorer-8-and-acid2-a-milestone.aspx |
| 03:24 | Philip` | also notices that both comment pages are long enough that anybody who posts a comment will have either not read all the previous comments, or will have been driven insane by reading them, so in both cases they shouldn't be adding more |
| 04:51 | <Hixie> | AwayEagle: not that i've heard of, at least not web browser vendors |
| 05:18 | <jruderman> | acid3 is requiring SVG and SMIL? :/ |
| 05:32 | <MacDome> | jruderman: until someone sends Hixie better tests |
| 05:32 | <jruderman> | hah |
| 05:32 | <MacDome> | jruderman: I think he's holding us ransom |
| 05:41 | <Hixie> | pretty much |
| 05:41 | <Hixie> | nobody has sent better tests, though |
| 05:41 | <Hixie> | so... |
| 06:48 | <othermaciej> | Ian Hickson promotes SMIL, podcast at 11 |
| 07:00 | <MacDome> | othermaciej: I expect with all the slashdot coverage, this is just Ian lashing out, like an injured beast, cornered. |
| 08:26 | <hsivonen> | what's the deal with all this access-control PEP location debate? The browser is already the PEP. |
| 08:29 | <Hixie> | hsivonen: dunno, but if you work out the answer, let me know |
| 08:32 | <hsivonen> | I also find it interesting that Web arch people think that /robots.txt is bad and mnot as a HTTP cacheability expert likes it |
| 08:34 | <Hixie> | i found it hilarious that mnot described HTTP headers as "new" |
| 08:34 | <Hixie> | and that using OPTIONS was inappropriate |
| 08:36 | <hsivonen> | it seems to me that OPTIONS is achitecturally the most pure way to deal with this |
| 10:07 | <annevk> | hmm, http://ejohn.org/blog/html5-shiv/ mentions the IE workaround but no workaround for Firefox' bug |
| 10:08 | <jgraham> | annevk: I should perhaps have mentioned the <span> workaround in my comment |
| 10:09 | <annevk> | i don't like that |
| 10:09 | <jgraham> | Do you have a better one? |
| 10:09 | <annevk> | yes, Firefox to fix their damn bug |
| 10:09 | <annevk> | :) |
| 10:10 | <jgraham> | annevk: Fixing the bug is trivial, fiing it without breaking the web less so :) |
| 10:10 | <jgraham> | s/fiing/fixing/ |
| 10:12 | <annevk> | Safari and Opera do it |
| 10:12 | <jgraham> | (although I don't know how many sites rely on residual style working across unknown elements, I guess it's a non-trivial number) |
| 10:13 | <annevk> | maybe if i'll complain on my blog someone will get annoyed and fixes it |
| 10:13 | <jgraham> | annevk: Safari and opera deal with <i><b><section><p></p></section></i></b> differently |
| 10:14 | <annevk> | oh yeah, Opera has some issues with that |
| 10:24 | <jgraham> | annevk: you shoul go ahead and blog if you think it will make any difference |
| 10:27 | <Philip`> | I've not noticed any Firefox developers objecting to the change - they just haven't bothered working on it |
| 10:30 | <krijnh> | http://www.quirksmode.org/blog/archives/2008/01/toughquiz_viii.html#c10784 |
| 10:31 | <Philip`> | "Just tried this on my valid XHTML4 page." - wow, futuristic |
| 10:31 | <krijnh> | That too |
| 10:32 | <krijnh> | So what happens when everybody starts advicing the HTML5 doctype _already_ ? |
| 10:32 | <krijnh> | Would that be a good thing for HTML5, or a bad thing? |
| 10:33 | <krijnh> | Seems like this issue put lots of attention on HTML5 |
| 10:34 | <annevk> | Philip`, actually, in our world XHTML4 is ancient |
| 10:35 | <annevk> | it would be bad for the IE team |
| 10:39 | <Camaban> | would be kind of weird, and potentially confusing if people start using HTML5 doctype, but are still coding in HTM4/XHTML1.0 |
| 10:39 | <Camaban> | +L |
| 10:43 | <annevk> | some features are becoming increasingly more usable |
| 10:43 | <annevk> | such as <canvas> |
| 10:45 | <Camaban> | partly depends when IE8 comes out I guess |
| 10:45 | <Camaban> | if it's another year away.... |
| 10:49 | <annevk> | excanvas makes <canvas> usable in IE |
| 10:50 | <krijnh> | And things like <svg> and <video> ? |
| 10:53 | <Philip`> | Just implement a Theora decoder in JavaScript, and use <canvas> to output the image date |
| 10:53 | <Camaban> | is there any "differences between HTML4.01 and HTML5" documentation? |
| 10:53 | <hsivonen> | hmm. I've lost track about scripting differences between quirks and standards modes |
| 10:53 | <hsivonen> | I'm pretty sure there are some |
| 10:53 | <Philip`> | Camaban: http://www.w3.org/TR/html5-diff/ |
| 10:54 | <Camaban> | excellent, ty |
| 10:55 | <annevk> | hsivonen, document.compatMode :) |
| 10:55 | <annevk> | some browsers have differences for various layout related DOM attributes, but we might be able to get rid of those |
| 10:56 | <annevk> | getElementsByClassName() is likely different |
| 10:57 | <Philip`> | Firefox does http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Cbody%20id%3Db%20onload%3Dw(b)%3E differently in quirks vs standards - does that count? |
| 10:58 | <Philip`> | (But Opera and IE don't) |
| 10:59 | <hsivonen> | Philip`: thanks. that counts |
| 11:00 | <hsivonen> | annevk: the IE8 thing inspired me to refactor my doctype article roughly like you suggested last year |
| 11:00 | <hsivonen> | (not published yet) |
| 11:01 | <annevk> | k, does it simply talk about Opera instead of Opera 7.5 (and later) now? :) |
| 11:01 | annevk | was already skimming through it to see if it changed |
| 11:01 | <hsivonen> | annevk: no. at least not yet :-) |
| 11:02 | <krijnh> | http://ejohn.org/blog/html5-doctype/#comment-296943 |
| 11:03 | <hsivonen> | krijnh: I'll reply |
| 11:04 | <krijnh> | annevk: why would it be bad for the IE team? |
| 11:06 | <Camaban> | that comment kind of raises the point I'm wondering about, how much effort, beyond just changing the doctype, would people already writing decent code need to make |
| 11:07 | <hsivonen> | krijnh: replied |
| 11:09 | <hsivonen> | Camaban: as the draft currently stands, most people will also hit the internal character encoding declaration syntax change issue |
| 11:12 | <krijnh> | There goes your number one position for 'html5 doctype' btw annevk ;) |
| 11:13 | <hsivonen> | can I get the #1 position for "doctype"? and a pony, too? |
| 11:13 | <zcorpan_> | hsivonen: elm.style.height = 5; |
| 11:13 | <hsivonen> | zcorpan_: thanks |
| 11:15 | <krijnh> | http://snook.ca/archives/browsers/importance_of_being_html5/#c57321 |
| 11:16 | <zcorpan_> | hmm, if the html5 doctype triggers ie8 mode, then that means that my and Hixie's advice effectively is "don't use the html5 doctype" |
| 11:16 | <hsivonen> | all this blog talk about the HTML5 doctype and IE8 has the potential of poisoning the doctype before IE8 ships as far as MS is concerned... |
| 11:17 | <annevk> | krijnh, not many people search for that |
| 11:17 | <annevk> | krijnh, it's too simple |
| 11:17 | <annevk> | "related:youporn.com/"; on the other hand... |
| 11:18 | <krijnh> | Hehe |
| 11:19 | <zcorpan_> | granted, this ie8 thing is going to result in a lot of uptake on html5 doctype usage |
| 11:19 | <krijnh> | hsivonen: that's my point exactly |
| 11:19 | <krijnh> | I hope there's enough information available to guide the newcomers |
| 11:19 | <Camaban> | uptake of HTML5 doctype perhaps, but not HTML5 |
| 11:20 | <hsivonen> | krijnh: I'm working on that |
| 11:20 | <krijnh> | Camaban: that would be bad |
| 11:20 | <Camaban> | krijnh: indeed |
| 11:20 | <Camaban> | but if people can write code mostly like they do now, and use the HTML5 doctype to 'fix' ie8.... |
| 11:20 | <annevk> | what would be bad? |
| 11:20 | <annevk> | actually, what is bad about that? |
| 11:20 | <krijnh> | If the HTML5 doctype becomes a IE7 standards mode trigger |
| 11:20 | <Lachy> | yes, unfortunately, we can't encourage widespread use of the HTML5 DOCTYPE until at least IE9, but more likely not until browsers, especially IE, start shipping support for more HTML5 features. |
| 11:21 | <krijnh> | And NOT the IE8 standards mode trigger |
| 11:21 | <hsivonen> | annevk: if MS decides that the HTML5 doctype is deployed too widely when IE8 ships |
| 11:21 | <krijnh> | Yeah |
| 11:21 | <krijnh> | That's not bad for IE devs, that's bad for authors |
| 11:21 | <annevk> | that depends on what we think of the IE strategy |
| 11:21 | <annevk> | and how much we think it will work, etc. |
| 11:21 | annevk | doesn't see a real problem either way |
| 11:22 | <krijnh> | Well, if the news now is you can use the html5 doctype so you don't have to use the meta.. And that's dismissed when IE8 is shipped.. |
| 11:23 | <Lachy> | maybe I should blog about not actively using the HTML5 DOCTYPE, explaining the problems with doing so |
| 11:23 | <krijnh> | Not that it makes sense to already include the new doctype, but does everybody know that? |
| 11:24 | <annevk> | if IE does lock it down who cares? i don't really see their strategy succeeding |
| 11:24 | <Camaban> | bit of a gamble to take :) |
| 11:25 | <annevk> | acknowledging that we agree that their solution is the way to go and therefore recommending people to not use the HTML5 DOCTYPE seems wrong to me |
| 11:26 | <krijnh> | To not use it _yet_ |
| 11:26 | <krijnh> | It would be great for HTML5 adoption if IE8 comes out and developers don't want to add an IE-only meta tag |
| 11:27 | <annevk> | if the goal is for browsers to behave the same (and I think it still is, however near impossible it is to get that watertight), x-ua-compatible is not the way to go |
| 11:27 | <krijnh> | But not if the new doctype doesn't change the way the page is rendered anyway |
| 11:28 | <annevk> | i'm not ready to make concessions on that goal |
| 11:30 | <krijnh> | I'm just seeing people unhappy with the meta thing, and totally happy with the HTML5 doctype solution |
| 11:30 | <krijnh> | And getting enthousiastic about HTML5, which is also even a WD this week.. |
| 11:31 | <Camaban> | krijnh: wouldn't say totally happy ;) not everyones keen on HTML5 |
| 11:31 | <Camaban> | more good will towards HTML5 than the meta tag though, by a long way... |
| 11:31 | <krijnh> | Indeed |
| 11:32 | <othermaciej> | I think IE locking in bug mode for the HTML5 doctype would be Microsoft's problem |
| 11:32 | <othermaciej> | I don't think there is any reason to advocate that people shouldn't use it |
| 11:32 | <hsivonen> | annevk: I'm assuming that MS won't change their mind about the number of IE8 modes in the IE8 timeframe, so that needs to be assumed as a constraint on what advice I give. However, I assume that there is a chance that they change their mind in the IE9 timeframe. |
| 11:32 | <krijnh> | I think it would be my problem as well, as an author |
| 11:34 | <Camaban> | well if you're going to advocate HTML5 doctype, that triggers 'edge' mode doesn't it? so they would end up back in the same situation for IE9 anyway |
| 11:34 | <Camaban> | same way if we advocate using the meta tag, but just setting everything to edge |
| 11:34 | <krijnh> | Not if they continue development after IE8, just as other browsers do |
| 11:34 | <krijnh> | Improve their standards support et cetera |
| 11:35 | <Camaban> | still got the issue of things that work in IE8, not working in IE9, causing people to complain that MS 'broke the web' |
| 11:35 | <krijnh> | Wouldn't IE9 still support the same trigger as IE8 ? |
| 11:36 | <krijnh> | Switch that is |
| 11:36 | <krijnh> | Why would they need a different switch again in IE9? The only reason I can think of is if they start development/bugfixing after IE8 again for 5 years |
| 11:37 | <krijnh> | s/start/stop - details :) |
| 11:37 | <Camaban> | the gap between IE7 and 8 won't be 5 years |
| 11:37 | <Camaban> | I don't see your point? |
| 11:38 | <krijnh> | Between IE8 and IE9 |
| 11:38 | <Camaban> | depends entirely on the differences between the 2, not the length of time between them |
| 11:39 | <Camaban> | if there's some equivilent to something like the * html hack breaking between 8 and 9, it'll still be an issue |
| 11:44 | <Lachy> | othermaciej, if MS locks in the HTML5 DOCTYPE to IE8 mode, then while it is certainly their problem, they make it a problem for *every single web developer*. |
| 11:44 | <Lachy> | therefore, reducing the chance of MS encountering problems with the HTML5 DOCTYPE is the only solution until their standards support increases sufficiently |
| 11:45 | <Philip`> | It wouldn't be a problem if HTML5 then changed its doctype to something else |
| 11:45 | <othermaciej> | I think they would be hard pressed to lock it in before they've even implemented any of the spec |
| 11:45 | <krijnh> | Lachy: You mean to IE7 mode? |
| 11:45 | <othermaciej> | or their locked in bug mode for it could never meaningfully support HTTML5 at all |
| 11:45 | <Lachy> | krijnh, no, Chris has already stated that HTML5 DOCTYPE will trigger IE8 mode in IE8. The question is whether they lock it to IE8 mode for IE9, or any other future version |
| 11:46 | <zcorpan_> | <!doctypehtml> |
| 11:46 | <krijnh> | And what if between now and the IE8 release everybody starts using the new doctype? |
| 11:46 | <Philip`> | Lachy: It's only been stated as an example in a comment in a blog post, so it wouldn't be impossible for that to change before IE8 is released |
| 11:47 | <Lachy> | Philip`, from previous conversations I've had with Chris, I'm certain the decision to make <!DOCTYPE html> trigger full standards mode won't change |
| 11:48 | <Philip`> | "Any unknown (i.e. not widely deployed) DOCTYPE. HTML5, for example." will change if the HTML5 doctype becomes known and widely deployed |
| 11:48 | <krijnh> | And it's becoming known right now |
| 11:48 | <krijnh> | Because of the IE-only meta switch we have to include otherwise |
| 11:48 | <othermaciej> | the risk isn't that it will get locked into IE7 mode, more that it will get locked into IE8 mode and therefore not gain any bug fixes or new features in IE9 |
| 11:49 | <othermaciej> | I am curious how Microsoft intends to deal with DOM and JS issues in their versioning scheme |
| 11:49 | <othermaciej> | will different compat modes have different sets of JS and DOM bugs and functionality? |
| 11:49 | <annevk> | and Acid2... |
| 11:50 | <Lachy> | acid2 won't pass with modifying it to include x-ua-compatible |
| 11:50 | <othermaciej> | with their scheme it seems impossible for them to really pass Acid2, short of some sort of blatant cheating which would probably look worse than just not passing |
| 11:50 | <krijnh> | othermaciej: wouldn't that risk be greater if people start using the new doctype now (because of the great news and all), see no difference between browsers and just keep using it because they like it.. |
| 11:51 | <Philip`> | They could make <!doctype html public "-//w3c//dtd html 4.01//en"> (as in Acid2) be treated the same as an unknown doctype (i.e. latest rendering mode), since it's used on less than 0.1% of some set of pages |
| 11:51 | <othermaciej> | krijnh: I don't really know |
| 11:51 | <othermaciej> | I guess I'm glad I'm not a web developer |
| 11:51 | <krijnh> | Heh :) |
| 11:52 | <othermaciej> | I don't know how common it is for the doctype to not mention strict.dtd or loose.dtd or whatever |
| 11:53 | <Philip`> | http://philip.html5.org/data/doctypes-lc.txt |
| 11:53 | <Philip`> | (Not very common) |
| 11:53 | <krijnh> | It's probably not a big deal, there aren't too many developers who like losing their validation mechanism |
| 11:53 | <krijnh> | But if hsivonen starts spreading validator.nu everywhere.. |
| 11:54 | <krijnh> | I don't know what happens, I guess we'll just have to see |
| 11:54 | <hsivonen> | comments? http://hsivonen.iki.fi/doctype/temp.html |
| 11:54 | <othermaciej> | special-casing the Acid2 DTD string would be kind of cheesy, though less so than hardcoding the URL |
| 11:54 | <hsivonen> | (I noticed that I need to update the table key to cover Safari 3) |
| 11:55 | <othermaciej> | too bad Hixie didn't use a super popular standards mod doctype |
| 11:55 | <hsivonen> | Hixie seems to have a tendency towards minimizing the length of the doctype |
| 11:55 | <Philip`> | hsivonen: s/discused/discussed/ |
| 11:56 | <madmoose> | html6 will probably have <!dctp h>, then ;) |
| 11:56 | <krijnh> | Mr. Turin will handle that for us ;) |
| 11:56 | <hsivonen> | Philip`: thanks |
| 11:58 | <othermaciej> | hsivonen: the Dashboard compat mode in WebKit isn't really a browser mode |
| 11:59 | <othermaciej> | hsivonen: it's not ever exposed in Safari or anything else that would be considered a browser, only in Dashboard itself |
| 11:59 | <othermaciej> | (it would be fair to call it an engine mode but it's even less browser-relevant than WML mode in Opera) |
| 12:00 | <hsivonen> | othermaciej: changed to "engine" |
| 12:01 | <krijnh> | I guess I'm glad I'm not a browser developer |
| 12:03 | <othermaciej> | hsivonen: I didn't know Firefox had scripting quirks - now I feel educated |
| 12:04 | <othermaciej> | (we've tried to avoid them in WebKit but I think Dashboard mode may have some; not the regular quirks mode though) |
| 12:05 | <annevk> | i'd think that style.width = 2 just invokes the CSS parser |
| 12:06 | <annevk> | and that therefore it's not really a scripting quirks |
| 12:08 | <hsivonen> | annevk: ok |
| 12:09 | <hsivonen> | annevk: removed. the goal is to give existence proof--not to document all the quirks anyway |
| 12:10 | <hsivonen> | published |
| 12:12 | <othermaciej> | hsivonen: that's good organization for the document |
| 12:13 | <hsivonen> | othermaciej: thanks. this is the first major reorg since 2000 |
| 12:16 | <hsivonen> | I should test iCab some day to find out if it has modes, too. |
| 12:17 | <zcorpan_> | annevk: true |
| 12:22 | <annevk> | hsivonen, http://hsivonen.iki.fi/engines/ Opera 9.5 and up use "futhark" |
| 12:23 | <Lachy> | http://www.katemonkey.co.uk/article/48/x-ua-lemur-compatible |
| 12:23 | <hsivonen> | annevk: thanks. |
| 12:24 | <annevk> | hsivonen, also, I believe the main reason HTML5 calls it "no quirks mode" and "limited quirks mode" is that the moment you start defining these modes calling one "standards" does not make sense as they are all part of a standard |
| 12:24 | <hsivonen> | annevk: oh. I though the reason I gave was the main reason. |
| 12:24 | <hsivonen> | Hixie: what's the main reason? |
| 12:26 | <hsivonen> | annevk: updated |
| 12:26 | <krijnh> | Lachy: :D |
| 12:27 | <annevk> | hsivonen, you refer to XML mode (which makes sense) but you only define application/xhtml+xml mode |
| 12:29 | <hsivonen> | annevk: do I need to define some document object and createElement craziness, too? |
| 12:30 | <hsivonen> | s/define/document/ |
| 12:30 | <annevk> | seems fine as is |
| 12:32 | <zcorpan_> | Lachy: that was really funny :) |
| 12:32 | <hsivonen> | annevk: I try to balance hair-split correctness against misunderstandings induced by XHTML propaganda |
| 12:32 | <hsivonen> | annevk: I think if I don't put MIME types in the readers' face, some will think they can get an XML mode with doctype |
| 12:33 | <annevk> | just say (XML mode) after it |
| 12:33 | <hsivonen> | annevk: ok |
| 12:34 | <hsivonen> | annevk: done. thanks |
| 12:40 | <MikeSmith> | virtuelv, annevk - are Opera weeklies apt-gettable from anywhere |
| 12:41 | annevk | doesn't know |
| 12:42 | <zcorpan_> | hsivonen: looks good |
| 12:42 | <hsivonen> | zcorpan_: thanks |
| 12:42 | <MikeSmith> | annevk - ok |
| 12:42 | <virtuelv> | MikeSmith: as in "are they in a repository?" -- not as far as I know |
| 12:43 | <MikeSmith> | virtuelv - OK. at least being able to automate download of the latest available weekly would be nice |
| 12:44 | <MikeSmith> | script it I mean |
| 12:44 | <MikeSmith> | I used to just pipe the buildbot messages from my mail client to a script |
| 13:01 | <MikeSmith> | hsivonen, annevk, jgraham - I know I've asked this before, but I'm trying to remember if html5lib and the validator.nu parser are capable of round-tripping a file from html to xml back to html |
| 13:04 | <othermaciej> | MikeSmith: were you looking for me earlier? |
| 13:05 | <MikeSmith> | othermaciej - yeah, but I'll ask over on #webkit because it's an iPhone-related question |
| 13:05 | <othermaciej> | ok |
| 13:24 | <hsivonen> | I wonder if I should shorten the title of my doctype page to something like Activating Browser Modes with Doctype |
| 13:25 | <othermaciej> | hsivonen: we usually refer to Safari's engine as "WebKit" nowadays (regarding your engine page again) |
| 13:25 | <othermaciej> | WebCore is an implementation detail and a somewhat obsolete term for referring to the engine |
| 13:26 | <hsivonen> | othermaciej: fixed. thanks |
| 13:27 | <othermaciej> | hsivonen: this comment below still mentions WebCore: "WebCore and JavaScriptCore are not independent implementations but Apple’s forks of KHTML and KJS" |
| 13:27 | <othermaciej> | hsivonen: (and I'm not sure it's a useful way to think about things any more, even though it's still true that they are forks) |
| 13:28 | <hsivonen> | othermaciej: fixed again. thanks |
| 13:31 | <hsivonen> | Hixie: from your manual pingback client: Failed to ping back 'http://ln.hixie.ch/';: RPC::XML::Client:send_request: parse-level error: |
| 13:31 | <hsivonen> | syntax error at line 1, column 0, byte 0 at /usr/lib/perl5/XML/Parser.pm line 187 (Acknowledged.) |
| 13:31 | <hsivonen> | Hixie: same thing with pinging webkit.org |
| 14:15 | <krijnh> | <blockquote><p>[Quoted]</p><p><cite>[Cited]</cite></p></blockquote> isn't the way HTML5 want it, right? |
| 14:16 | <krijnh> | What's the reason for that? |
| 14:34 | <hsivonen> | MikeSmith: I'm not sure I understand your round-tripping question |
| 14:34 | <hsivonen> | MikeSmith: did you mean parse as HTML5, serialize as XHTML5 without namaspace prefix or CDATA sections and reparse as HTML5? |
| 14:35 | <hsivonen> | I think there'd be some edge cases related to <pre> or <textarea> content starting with a line feed |
| 14:36 | <hsivonen> | MikeSmith: as for whether the adoption agency stuff is stable under reparse, I'm not sure |
| 14:41 | <MikeSmith> | hsivonen - OK. I was thinking of the use case where somebody might want to take an html source file, parse and serialize to xml because they want to run some xml tool on it (to do some minor xslt transformation on), then serialize that (slightly different) result back to html |
| 14:42 | <MikeSmith> | xslt part being just an example |
| 14:42 | <MikeSmith> | general case being just to be able to run it through xml tools |
| 14:42 | <hsivonen> | MikeSmith: that's possible but the HTML serialization may not parse back as the same DOM |
| 14:43 | <MikeSmith> | right, Ok |
| 14:43 | <hsivonen> | depends on what kind of tree comes out of the XML tool |
| 14:43 | <hsivonen> | conceptually, that's what the XLST4HTML5 sample program that comes with the V.nu parser does |
| 14:44 | <MikeSmith> | yeah, this is coming back to me now from the previous time(s) I asked the same question |
| 14:44 | <hsivonen> | it just skips the parts of serializing to XML and parsing the XML back |
| 14:46 | <zcorpan_> | Hixie: test 4 is wrong |
| 14:47 | <zcorpan_> | expect(i.nextNode(), document.getElementById('instructions').firstChild); |
| 14:47 | <zcorpan_> | expect(i.nextNode(), document.links[0]); |
| 14:47 | <zcorpan_> | there's a <span></span> in there which you have forgotten in the script, it seems |
| 14:52 | <MikeSmith> | hsivonen - anyway, the reason I was thinking of it was also because of Norm's blog entry about HTML5 and his mention that 'One of the two “authoring formats” described by the HTML 5 specification is a custom one.' and 'The goal of the custom parser, as I understand it, is that it imposes an unambiguous HTML 5 interpretation on any random stream of characters." |
| 14:56 | <zcorpan_> | Hixie: also, other tests seem to create and insert links, so document.links[0] doesn't point to the link to reference.html. i'm not sure if this is a bug in the test or not, yet |
| 14:56 | <zcorpan_> | (i.e. when checking against document.links[0] in test 4) |
| 15:07 | Lachy | successfully upgraded blog.whatwg.org to wordpress 2.3.2 |
| 15:13 | <krijnh> | Code is poetry :) |
| 15:18 | <inimino> | *some* code is poetry |
| 15:29 | <krijnh> | Yeah, like http://www.joomla.org/content/view/4488/1/ (also released on January 22 btw) |
| 15:33 | <annevk> | that was not a successful upgrade Lachy |
| 15:33 | <annevk> | http://blog.whatwg.org/html-5-published-as-w3c-first-public-working-draft |
| 15:35 | <krijnh> | Details :) |
| 15:40 | <Lachy> | what? |
| 15:40 | <Lachy> | oh crap. I forgot the damn .htaccess |
| 15:41 | <Lachy> | I knew I should have kept the orignal folder around longer :-( |
| 15:42 | <Lachy> | fixed |
| 15:44 | <Lachy> | I must write some sort of upgrade script to make that easier |
| 15:46 | <Lachy> | should be possible: just download latest.zip from wordpress.org, unzip, cp required files across to new directory, replace old dir with new dir (keeping backup of old dir) and then wget upgrade.php |
| 15:49 | <zcorpan_> | Hixie: oh, and document.links also includes <area href>s |
| 15:52 | <annevk> | you'd think someone has written such a script already |
| 15:56 | <Lachy> | google reveals several different plugins and shell scripts. Just need to pick one that will work |
| 16:07 | <gsnedders> | "Note that whatever solution we have today is not necessarily final. We've explained (I hope) why we need a solution. We can debate if a solution is in fact needed (I am convinced it is) and what the best solution might be..." — Alex Mogilevsky on www-style (about IE8 Standards Mode opt-in) |
| 16:13 | <annevk> | http://lists.w3.org/Archives/Public/public-xml-core-wg/2008Jan/0053.html |
| 16:13 | <annevk> | hmm |
| 16:15 | <krijnh> | What's wrong with it? |
| 16:23 | gsnedders | sighs |
| 16:24 | <gsnedders> | Time to sit down and write http-parsing |
| 17:30 | <annevk> | http://lists.w3.org/Archives/Public/public-html-comments/2008Jan/0023.html CSS5 heh |
| 17:30 | <annevk> | didn't notice that :) |
| 18:00 | <gsnedders> | anyone have a clue how many sites use the wrong Content-Encoding? |
| 18:01 | <Philip`> | Since http://http-parsing.gsnedders.com/ indicates 1 content-encoding header in total, I have no clue |
| 18:02 | <gsnedders> | 1 out of ~15k I find amazing to believe, FWIW |
| 18:02 | <gsnedders> | I normally hit issues trying to follow the spec after a few hundred |
| 18:03 | <annevk> | text/html;ISO-8859-1; charset=UTF-8 nice |
| 18:03 | <SadEagle> | heh. |
| 18:04 | <SadEagle> | annevk: nothing beats charset=UCS-2 or such in an ascii file :-) |
| 18:05 | <SadEagle> | gsnedders: what % of the pages you tested were missing content-type? |
| 18:05 | <gsnedders> | SadEagle: no idea :P |
| 18:06 | <SadEagle> | man, I need to write myself a todo list of things to write whatwg list or hixie about.. |
| 18:06 | <gsnedders> | I just create a draft email with a subject line and no more |
| 18:06 | <Philip`> | gsnedders: Maybe I need to send an accept-encoding request header? |
| 18:06 | <gsnedders> | Philip`: maybe :) |
| 18:08 | <gsnedders> | SadEagle: the data set comes from Philip` |
| 18:08 | <annevk> | SadEagle, please do cc some list :) |
| 18:09 | <gsnedders> | annevk: didn't you say something about chunked transfer encoding and trailing headers? |
| 18:10 | <annevk> | might be |
| 18:11 | <SadEagle> | looking through kio_http sources, the only 2 quirks it has for content-encoding: some servers like[d] to make tar.gz files into content-encoding:gzip mimetype:tar, and some site out there used content-encoding:8bit |
| 18:12 | <gsnedders> | annevk: "it's not entirely clear to me how that works though" is what you said, so nothing useful :) |
| 18:13 | <gsnedders> | SadEagle: yeah, Mozilla seems to do nothing more too |
| 18:15 | <zcorpan_> | hsivonen: hmm, "[Image of a nutshell]" seems like a bit weird alt text to me |
| 18:15 | <zcorpan_> | "In a nutshell: " or just "" seems to read better |
| 18:16 | <hsivonen> | zcorpan_: thanks. that has probably been there since 2000 when I didn't know better |
| 18:16 | <hsivonen> | zcorpan_: fixed |
| 18:17 | <hsivonen> | I find it amusing that SVG is sometimes marketed as being big on mobile. Today I read planet intertwingly on my phone and saw the text "embedded svg failed to render" or something to that effect |
| 18:18 | <hsivonen> | all of my desktop browsers support SVG but at least three out of my four mobile browsers don't |
| 18:22 | <hsivonen> | too bad that I seem to have lost the high-res alpha channel version of the nutshell photo in a hard disk crash or something |
| 18:35 | <gsnedders> | knowing when to trust Content-Length seems slightly more complex |
| 18:36 | hsivonen | naïvely thought that Content-Length worked |
| 18:37 | <virtuelv> | exactly what does IE do here? http://ejohn.org/blog/html5-shiv/ ? |
| 18:37 | <gsnedders> | hsivonen: oh, of course not. |
| 18:37 | <gsnedders> | hsivonen: why assume anything works? :) |
| 18:37 | <virtuelv> | does it prevent IE from autoclosing the unknown element, so that the DOM is autocorrected? |
| 18:38 | <gsnedders> | hsivonen: that said, I've never seen anything that isn't a number greater than or equal to 0 being returned (though when zero you have to discard the header, I think) |
| 18:39 | <annevk> | virtuelv, it sets different parsing behavior |
| 18:40 | <zcorpan_> | virtuelv: the parser treats that tag the same as it treats tags that have colons in them |
| 18:41 | <virtuelv> | zcorpan_: but you get a "correct" DOM, at least for non-empty elements? |
| 18:41 | <virtuelv> | s/DOM/document tree/ |
| 18:42 | <zcorpan_> | the element ends up in the DOM with the original case preserved, /> is treated "like XML", the element can have contents, and so forth |
| 18:43 | <virtuelv> | Neat. IE's inane behavior is my last stumbling block for using certain HTML5 features |
| 18:44 | <zcorpan_> | stray end tags are ignored instead of generating empty "/FOO" elements |
| 18:51 | <annevk> | http://www.infoworld.com/article/08/01/22/W3C-offers-HTML-5-draft_1.html |
| 18:51 | <zcorpan_> | hsivonen: hmm, i found a bug in validator.nu |
| 18:51 | <zcorpan_> | Info: Using the schema for HTML5. |
| 18:51 | <zcorpan_> | The document validates according to the specified schema(s). |
| 18:51 | <zcorpan_> | <!DOCTYPE html>↩ |
| 18:51 | <zcorpan_> | <title></title>↩ |
| 18:51 | <zcorpan_> | <body>↩ |
| 18:51 | <zcorpan_> | </span> |
| 18:56 | <annevk> | parsing bug? |
| 18:57 | <zcorpan_> | seems so |
| 19:00 | <zcorpan_> | html5lib doesn't have this bug |
| 19:02 | <zcorpan_> | same with </foobar> |
| 19:07 | <virtuelv> | huh? http://ejohn.org/blog/html5-doctype/ |
| 19:13 | <gsnedders> | virtuelv: Just follow the quote of Chris Wilson: unknown doctypes trigger IE8 Standards Mode (or maybe edge, dunno, not clear) |
| 19:16 | <Philip`> | zcorpan_: "tags that have colons in them" aren't normally handled specially by IE |
| 19:17 | <Philip`> | (It's only if they're in a document.write or are after an <html xmlns:foo> or after some PI thing) |
| 19:17 | <virtuelv> | gsnedders: I certainly _hope_ they will go for edge mode |
| 19:20 | <webben_> | It's a bit problematic that they (or at least the alistapart author) are recommending people avoid edge |
| 19:21 | <virtuelv> | indeed |
| 19:21 | <virtuelv> | and I have problems just mentioning this |
| 19:21 | <gsnedders> | it's logical, though, if you look at the argument |
| 19:24 | webben_ | fails to see how forcing browsers to have multiple engines like some sort of Matryoshka doll for ever more is a good idea. (It's going to be hard enough dealing with multiple formal standards, like HTML and XHTML (and whatever eventually replaces those).) |
| 19:24 | <annevk> | differences between html and xhtml are trivial |
| 19:24 | <annevk> | as far as browsers are concerned |
| 19:24 | <gsnedders> | both are dealt with from a single DOM |
| 19:25 | <gsnedders> | you parse them into a DOM, and have a couple of if statements. It really isn't very major. |
| 19:25 | <Philip`> | webben_: It's unlikely to be "for ever more" - it's just until the cost of maintenance outweighs the cost of hurting paying customers |
| 19:25 | <webben_> | Philip`: That's not the argument put forward. |
| 19:25 | <webben_> | Philip`: Note: "Sure, most trips through the “Wayback Machine” don’t suffer in modern browsers because the DOCTYPE switch still serves them well, but what about a site built to IE6’s implementation of “standards” mode?" |
| 19:26 | <webben_> | Like Hixie (to some degree), these folks are worried about long-term archival renderings. |
| 19:26 | Philip` | wonders where that is quoted from |
| 19:26 | <webben_> | Philip`: soz: http://www.alistapart.com/articles/beyonddoctype |
| 19:26 | <Philip`> | Ah, thanks |
| 19:27 | <annevk> | long-term archival renderings tied to a single proprietary rendering engine is just plain silly |
| 19:27 | <Philip`> | Who are "these folks"? I don't think Microsoft is going to be interested in preserving the Wayback Machine, because it is of no value to them |
| 19:27 | <annevk> | especially given prior experience with the vendor on a product that used a similar strategy and failed: MS Office |
| 19:28 | <webben_> | If someone could demonstrate that browser developers could reliably create mock-IE5, -IE6, -IE7, -IE8 engines I'd be a bit less worried. But the general consensus as expressed by their participation in WHATWG seems to be that creating multiple rendering engines is /hard/ and faking IE is /really/ hard. |
| 19:28 | <Philip`> | Actually, I suppose "these folk" are the author of the article plus maybe others |
| 19:28 | <Philip`> | so that answers that question |
| 19:29 | <webben_> | I think anyone with an interest in the transformative powers of the web would have /some/ sort of interest in the archival question. ... That doesn't translate into being a big business goal for MS, necessarily. |
| 19:30 | <webben_> | But it would be part of the developer buy-in that keeps them operational as a tech company. |
| 19:30 | <hsivonen> | hmm. interesting. even parsetree.validator.nu doesn't report an error in zcorpan's case |
| 19:30 | <webben_> | (internal and external) |
| 19:30 | <Philip`> | I imagine a future historian will have more severe difficulties if they want to e.g. see what Google and Hotmail were like centuries ago - the CSS layout bugs will be a very minor problem |
| 19:31 | <SadEagle> | webben_: IE does some freakish stuff, but fortunately it's so freaking, I don't think it's used that much |
| 19:31 | webben_ | isn't so sure about that. |
| 19:31 | <SadEagle> | not too many sites using vml or whatever it's called, e.g. :-) |
| 19:31 | <webben_> | You can easily get away with doing freakish stuff if you're catering to 90% or whatever of the market by doing so. |
| 19:32 | <webben_> | VML's potentially implementable (there's an open spec submitted as a note to W3C) |
| 19:32 | <SadEagle> | OTOH, the window object is just a disaster |
| 19:32 | <hsivonen> | webben_: submitted to ISO, too |
| 19:32 | <webben_> | hsivonen: Did it make it to ISO Standard status? |
| 19:33 | <virtuelv> | " It's unlikely to be "for ever more" - it's just until the cost of maintenance outweighs the cost of hurting paying customers" |
| 19:33 | <hsivonen> | webben_: the OOXML saga ain't over yet |
| 19:33 | <virtuelv> | it's not that long ago I had to deal with people who maintained NS4.x-compatibility |
| 19:33 | <webben_> | If one can write translations of SVG to VML, one presumably could write translations the other way round. |
| 19:33 | <webben_> | well, presumably's too strong, let's say: hopefully |
| 19:34 | <hsivonen> | what was wrong with VML that caused SVG to be developed instead? |
| 19:34 | hsivonen | doesn't know VML |
| 19:34 | <SadEagle> | webben_: let me rephrase... I think stuff like getElementById returning things by name tends to cause a lot more problems than things that are very different |
| 19:35 | <webben_> | Philip`: I think for that sort of thing (experiencing Gmail), a historian's likely to use some sort of VM. It's actual content that you'd most want to extract I suspect. |
| 19:35 | <webben_> | SadEagle: I think that's true. |
| 19:36 | <Philip`> | SadEagle: VML isn't used nowhere - it's more common than e.g. <bdo> |
| 19:36 | <Philip`> | http://canvex.lazyilluminati.com/survey/2007-07-17/analyse.cgi/tag/v%253Astroke is on 6 out of 7739 pages |
| 19:37 | <SadEagle> | OTOH, IE does have all sorts of oddness in how JS and the DOM and the DOM state vs. e.g. form element state mesh together |
| 19:37 | <SadEagle> | Philip`: how much of that is word output? |
| 19:37 | <hsivonen> | part Word Google Maps rest? |
| 19:37 | <hsivonen> | hmm. the </span> occurs IN_BODY phase, so it is not an end-game bug |
| 19:38 | <Philip`> | http://www.magneticsforyou.com/ is broken in Firefox since none of the VML images come through |
| 19:38 | <Philip`> | SadEagle: Probably all of it, but I'm too lazy to check :-) |
| 19:38 | <SadEagle> | <html xmlns:v="urn:schemas-microsoft-com:vml" xmlns:o="urn:schemas-microsoft-com:office:office" xmlns="http://www.w3.org/TR/REC-html40">; |
| 19:39 | <SadEagle> | nice one on an html document |
| 19:39 | <Philip`> | http://canvex.lazyilluminati.com/survey/2007-07-17/analyse.cgi/attr/xmlns:o |
| 19:39 | <Philip`> | It's pretty common |
| 19:40 | <webben_> | it's not much weirder than serving XML as HTML in the first place. |
| 19:40 | <SadEagle> | is that sort of stuff why you folks were reverse-engineering the : behavior in IE? |
| 19:40 | <SadEagle> | webben_: people seem confused about that |
| 19:40 | <webben_> | SadEagle: about what? |
| 19:40 | <webben_> | (or rather: how?) |
| 19:40 | <SadEagle> | about sering XML as HTML |
| 19:41 | <webben_> | Well, yes, people are confused by it. |
| 19:41 | <SadEagle> | Philip`: seems like FrontPage makes it, too |
| 19:41 | <webben_> | Oh I see what you mean. |
| 19:41 | <webben_> | SadEagle: What I meant is it's weird; but then XML-as-HTML is a weird thing. |
| 19:42 | <Philip`> | SadEagle: People were reverse-engineering the : behaviour in IE because it would be nice to use that syntax for e.g. embedding SVG, not because anybody wants to implement VML or <o:p> |
| 19:42 | <SadEagle> | Philip`: right. But I am guessing there may be something to reverse-engineer due in part to this stuff... |
| 19:42 | <Philip`> | but IE is insane so now people want to avoid using : wherever possible :-) |
| 19:43 | <webben_> | although presumably it would path the way to doing so |
| 19:43 | <webben_> | (which wouldn't necessarily be a bad thing ... if lots of sites use VML, maybe it's worth implementing as a backwards compat feature) |
| 19:49 | <blooberry> | philip`: what is the best way to get the full list of alexa 500 sites? |
| 19:50 | <Philip`> | blooberry: I did it by downloading http://www.alexa.com/site/ds/top_sites?ts_mode=global&lang=none and the next four pages, then regexping the site list from each page |
| 19:51 | <blooberry> | *sigh* been there and done that for lycos' top lists in the past. 8-} ok. Thanks. |
| 19:52 | <hsivonen> | zcorpan: fixed. Thank you. |
| 19:52 | <hsivonen> | clearly, the html5lib test suite could have better coverage |
| 20:06 | <blooberry> | whoa...over 80,000 URLs from dmoz (out of a URL space of ~4 million) declare the VML namespace. That's a lot higher than I expected |
| 20:10 | <Philip`> | 2%? |
| 20:10 | <Philip`> | That matches http://canvex.lazyilluminati.com/survey/2007-07-17/analyse.cgi/attr/xmlns:v and the list of pages indicates it's spread over lots of independent sites |
| 20:10 | <hsivonen> | blooberry: what are you using as your analysis tool? |
| 20:11 | <blooberry> | Something I wrote. I'm writing up the results to put up somewhere right now. |
| 20:12 | <Philip`> | Are you using an existing library for parsing HTML? |
| 20:12 | <blooberry> | no. I wrote it, and at times it is not the prettiest thing. 8-} |
| 20:13 | <blooberry> | and it evolved from some rather simplistic checking |
| 20:14 | <blooberry> | I eschewed using an existing parsing library at one point because I didn't know how well it would react to the sorts of brain-dead markup I had been running into. |
| 20:16 | <blooberry> | I figure that when analyzing a document, you have 3 main parsing styles to allow for - very strict for the script, pretty strict for the CSS, and not-so-strict for the markup. |
| 20:18 | <Philip`> | You could consider there to just be one main parsing style, which is to be compatible with what web browsers do |
| 20:18 | <SadEagle> | there are some quirks in ES parsing as well |
| 20:19 | <blooberry> | yes, but to reflect what web browsers actually do, it seems to me that the above "styles" of parsing seem the most realistic. (10,000 ft view, of course) |
| 20:20 | <blooberry> | sadeagle: most certainly. 8-} |
| 20:21 | <gsnedders> | there are quirks everywhere :P |
| 20:23 | <Philip`> | Are there GIF-handling quirks? |
| 20:23 | gsnedders | would be amazed if there weren't |
| 20:24 | <SadEagle> | Philip`: yes. |
| 20:24 | <SadEagle> | Philip`: actually, it's hard to say if there are or aren't, since GIF is specified in a way that's not very web-relevant |
| 20:25 | <zcorpan_> | SadEagle: is handling of GIF in browsers different in quirks mode and standards mode? |
| 20:25 | Philip` | wasn't thinking of that type of quirk |
| 20:26 | <SadEagle> | zcorpan_: don't believe so |
| 20:27 | <SadEagle> | Philip`: e.g. gif images can provide geometries, and sometimes the one that should be the entire image's size isn't trustworthy. There is also an undocumented/illegal animation mode in pretty wide use |
| 20:28 | <Philip`> | Hmm, sounds fun |
| 20:28 | <Philip`> | Isn't there anything old enough that all the quirks have been shaken out by now? |
| 20:42 | <SadEagle> | Philip`: to be honest, I didn't quite figure out all the animation modes, either... May be I can find some mozilla guy to explain it to me, though |
| 21:09 | <ed_home> | MacDomeOut: I was just browsing the irclogs, and was wondering about something you said about iframes with dataURI:s in the svg acid-tests, for the record you can see the ACID3 submissions I made on the public www-svg list...no data-uri:s in there I can assure you |
| 21:09 | <ed_home> | ...http://lists.w3.org/Archives/Public/www-svg/2008Jan/0055.html |
| 21:11 | <annevk> | ed_home, hey! |
| 21:11 | <ed_home> | hey anne :) |
| 21:21 | <Hixie> | zcorpan_: i actually would encourage people to use <!DOCTYPE HTML> a lot before IE8 comes out, so that when IE8 comes out either they are forced to make that DOCTYPE trigger IE7 mode, or lots of sites break and they realise that their idea didn't work anyway |
| 21:21 | <Hixie> | zcorpan_: (i would rather everyone was using the same mode, IE7-mode, than have multiple modes in use.) |
| 21:22 | <Hixie> | hsivonen: i called the modes "no quirks" and "limited quirks" and "quirks" because as annevk says, it makes no sense to refer to a "standards" mode when all the modes are in the standard |
| 21:22 | <Hixie> | othermaciej: acid3 uses a common doctype. acid2 uses a somewhat common doctype. |
| 21:22 | <Hixie> | hsivonen: my pingback stuff is probably broken. i don't plan to fix it any time soon. |
| 21:22 | <othermaciej> | Hixie: someone showed me some distribution stats |
| 21:23 | <Hixie> | zcorpan_: please send mail about acid3, i am swamped right now |
| 21:26 | <Hixie> | holy crap, roy fielding agreed with me on something |
| 21:27 | <Hixie> | othermaciej: acid2's is low, but it still affects in the order of 1% of pages, iirc |
| 21:27 | <Hixie> | (50% of pages have no dcptype, and the majority of hte rest have xhtml or transitional doctypes, so...) |
| 21:34 | <hsivonen> | Hixie: ok. I updated my doctype page not to ascribe a motivation to the nomenclature in HTML 5. |
| 21:35 | <hsivonen> | my own Pingback impl. is years late... perhaps I should use telnet to port 80... |
| 21:36 | gsnedders | was using telnet for port 80 a few minutes ago |
| 21:45 | <zcorpan_> | "If you want to be more annoying, use IE=${random(9,12)}" -- http://annevankesteren.nl/2008/01/ie-lock-in#comment-6424 |
| 21:46 | <webben_> | lol |
| 21:56 | <cgriego> | IE=2 |
| 21:56 | <Hixie> | that would just make them change the syntax |
| 21:57 | <zcorpan_> | indeed |
| 21:57 | <Hixie> | the only way to actually make any difference, as far as i can tell, is to not use it |
| 21:57 | <Hixie> | as i discussed on ln.hixie.ch |
| 21:58 | <zcorpan_> | Hixie: if <!doctype html> triggers ie8 mode, then that means not using <!doctype html>, too. no? |
| 21:59 | <zcorpan_> | Hixie: btw, did you see my comments on acid3 earlier? |
| 22:00 | <zcorpan_> | http://krijnhoetmer.nl/irc-logs/whatwg/20080124#l-584 |
| 22:01 | <virtuelv> | IE=1 |
| 22:01 | gsnedders | comes across <http://bugs.php.net/bug.php?id=24083>; |
| 22:05 | <Hixie> | zcorpan_: if enough people use it early enough, they'll be forced to make it not trigger IE8 mode |
| 22:05 | <Hixie> | (or, lots of pages will break, which is fine too, since it will suggest that this technique doesn't work) |
| 22:06 | <Hixie> | zcorpan_: i mailed your comments to myself |
| 22:06 | <zcorpan_> | k |
| 22:08 | <annevk> | "maxage" is used in cache-control |
| 22:33 | <Hixie> | annevk: so? |
| 22:33 | <Hixie> | annevk: it's not used as part of a hypenated word |
| 22:33 | <annevk> | fn |
| 22:34 | <Hixie> | annevk: the point is just that you should either have a phrase that is consistently mashedtogether, or is-hyphenated, or isCamelCase, or TitleCase, or whatever |
| 22:34 | <Hixie> | mixing-them justLooks quitesilly |
| 22:34 | <Hixie> | or rather, "mixing-ThemJustLooks-quitesilly" |
| 23:01 | <annevk> | http://annevankesteren.nl/2008/01/firefox-html5#comment-6433 |
| 23:01 | <annevk> | <input type=date> styling in IE |
| 23:01 | <Dashiva> | Wow, that last mail about ogg really made me rethink everything said so far |
| 23:02 | <Dashiva> | The last PS of anti-microsoft paranoia really closed the deal |
| 23:11 | <Hixie> | "Can we make standards ISO compliant" -- isn't MPEG an ISO standard? |
| 23:12 | <Hixie> | or at least ISO standards track? |
| 23:13 | <dbloom> | Moving Picture Experts Group is a working group of ISO |
| 23:13 | <billmason> | There are various mpeg references in http://standards.iso.org/ittf/PubliclyAvailableStandards/index.html if that's significant. |
| 23:13 | <dbloom> | so i would hope so |
| 23:13 | <Hixie> | k |
| 23:15 | <takkaria> | annevk: that's pretty awesome |
| 23:24 | <hubick> | Any HTML5 Java programmers here that know if it's possible to get a javax.xml.transform.Transformer to output a <!DOCTYPE HTML> header, being that there doesn't seem to be an OutputKeys.DOCTYPE_NAME I can set on it? |
| 23:27 | <annevk> | I don't know about Java, but I do know that with XSLT you need to output literal text using <xsl:text> or something to get the DOCTYPE out correctly |
| 23:28 | <annevk> | As XSLT is currently not HTML5 aware |
| 23:29 | <hubick> | annevk: hmm... I am feeding this thing a DOM tree, and if that contains such literal text, I think it would be encoded upon output |
| 23:29 | <annevk> | you also need disable-output-escaping or something like that |
| 23:30 | <hubick> | in my simple case, the transformer is used with no XSLT, just simply as an XML serialization tool (hack) |
| 23:31 | <annevk> | I see |
| 23:31 | <hubick> | hmm, i may be forced to look into DOM Load/Save code... though I don't know if that actually ships with Java yet in 1.6 or 1.7 |
| 23:31 | <annevk> | http://about.validator.nu/htmlparser/ might have a better serializer |
| 23:32 | <annevk> | that's the only HTML5 Java project I know of |
| 23:32 | <hubick> | annevk: good idea. I'm already using Henri's stuff on the input side |
| 23:33 | <annevk> | I suspect he went to bed already, in about 8-9 hours he might be online |
| 23:45 | <weinig> | Hixie: If you have some time, I would be curious as to what your thoughts on http://bugs.webkit.org/show_bug.cgi?id=16968 are? |
| 23:56 | <Hixie> | yeah it's a known issue i will deal with before acid3 is done |
| 23:59 | <hubick> | if (user agent Accept's HTML5) then output <header> else output <div class="header">; Is there a way to know if the user agent accepts html5? |