Options:
- # Session Start: Sun Apr 22 00:00:00 2007
- # Session Ident: #whatwg
- # [00:01] <zcorpan_> there
- # [00:25] * Joins: hober (n=ted@unaffiliated/hober)
- # [00:32] * Joins: SpookyET (i=user@75.138.70.34)
- # [00:34] * Quits: SpookyET (i=user@75.138.70.34) (Client Quit)
- # [00:37] * Quits: aroben (n=adamrobe@adsl-70-231-234-224.dsl.snfc21.sbcglobal.net)
- # [00:47] * Quits: met_ (n=Hassman@r5bx220.net.upc.cz) ("Leaving")
- # [00:49] <zcorpan_> "Simple animation support: Flash makes creating animations extremely easy. Doing the same thing in anything Free (HTML, SVG, even Gtk) requires being a good code monkey. The example above required NO coding." -- http://www.advogato.org/person/company/diary.html?start=39
- # [00:50] <zcorpan_> thus, we need authoring tools for canvas or svg animations
- # [00:50] <zcorpan_> probably svg is more suitable for animations
- # [00:53] * Quits: hober (n=ted@unaffiliated/hober) ("ERC Version 5.2 (IRC client for Emacs)")
- # [00:57] * Parts: hasather (n=hasather@81-235-209-174-no62.tbcn.telia.com)
- # [00:57] * Quits: hasather_ (n=david@81-235-209-174-no62.tbcn.telia.com) ("Lost terminal")
- # [00:58] * Joins: zcorpan (n=zcorpan@84-216-42-9.sprayadsl.telenor.se)
- # [01:01] * Joins: hasather (n=hasather@81-235-209-174-no62.tbcn.telia.com)
- # [01:04] <zcorpan> http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2007-April/010975.html where did that leading space in the table come from?
- # [01:04] * zcorpan ponders
- # [01:07] * om_sleep is now known as othermaciej
- # [01:11] * Parts: hasather (n=hasather@81-235-209-174-no62.tbcn.telia.com)
- # [01:14] * Joins: tantek (n=tantek@adsl-63-203-221-50.dsl.snfc21.pacbell.net)
- # [01:17] * Quits: zcorpan_ (n=zcorpan@84-216-40-107.sprayadsl.telenor.se) (Read error: 110 (Connection timed out))
- # [01:19] * Joins: hasather (n=hasather@81-235-209-174-no62.tbcn.telia.com)
- # [01:20] * Quits: tantek (n=tantek@adsl-63-203-221-50.dsl.snfc21.pacbell.net)
- # [01:20] <Philip`> zcorpan: "<map name> works in all except in Gecko in XHTML and WebKit in XHTML." - but the table's data for 007 seems to indicate it does work in WebKit-XHTML
- # [01:21] <zcorpan> Philip`: ah, yes, thanks
- # [01:24] * Parts: hasather (n=hasather@81-235-209-174-no62.tbcn.telia.com)
- # [01:27] * Joins: hasather (n=hasather@81-235-209-174-no62.tbcn.telia.com)
- # [01:29] * Parts: hasather (n=hasather@81-235-209-174-no62.tbcn.telia.com)
- # [01:31] * Joins: hasather (n=hasather@81-235-209-174-no62.tbcn.telia.com)
- # [01:31] * zcorpan likes the "no alt = image is meaningful" idea
- # [01:38] * Parts: hasather (n=hasather@81-235-209-174-no62.tbcn.telia.com)
- # [01:39] * Joins: hasather (n=hasather@81-235-209-174-no62.tbcn.telia.com)
- # [01:40] * Parts: hasather (n=hasather@81-235-209-174-no62.tbcn.telia.com)
- # [01:42] * Joins: hasather (n=hasather@81-235-209-174-no62.tbcn.telia.com)
- # [01:43] * othermaciej is now known as om_wet
- # [01:46] <jgraham> http://wordsandpictures.dyndns.org/html5/parsetree.html (a HTML5 DOM viewer based on html5lib -- still very rough)
- # [01:47] * jgraham is going to sleep now
- # [01:48] <zcorpan> jgraham: nice!
- # [01:48] <zcorpan> why are the elements uppercased?
- # [01:50] <jgraham> Ah, in innerHTML view they shouldn't be... I'll remove that CSS :)
- # [01:50] <zcorpan> it shouldn't be in the tree either, should it?
- # [01:52] <jgraham> I guess it makes more sense to have everything lowercase (it's also easier)
- # [01:52] <zcorpan> yeah
- # [01:57] <zcorpan> innerHTML view says "<!DOCTYPE: HTML>"
- # [02:00] <zcorpan> perhaps the textarea should be POST?
- # [02:01] * zcorpan stops bugging jgraham now
- # [02:12] * Joins: tantek (n=tantek@dsl001-150-252.sfo1.dsl.speakeasy.net)
- # [02:13] <Lachy> jgraham, #text nodes shouldn't appear outside of the root element.
- # [02:16] <zcorpan> Lachy: i think they should
- # [02:16] <zcorpan> (per spec that is)
- # [02:17] <zcorpan> (if the spec is wrong is another matter)
- # [02:17] <Lachy> I mean in the DOM view
- # [02:18] <zcorpan> yes?
- # [02:18] <Lachy> so the DOM spec applies
- # [02:18] <Lachy> and I'm fairly sure they shouldn't
- # [02:18] <zcorpan> the html5 parsing spec says to append text nodes to the Document object in some cases
- # [02:19] <Lachy> that's weird
- # [02:20] * Parts: hasather (n=hasather@81-235-209-174-no62.tbcn.telia.com)
- # [02:21] <zcorpan> what would you do with whitespace between the doctype and the first tag?
- # [02:21] <Lachy> drop it
- # [02:21] <zcorpan> why?
- # [02:21] <Lachy> outside of the document, it has no meaning and irrelevant
- # [02:22] <Lachy> s/document/root/
- # [02:22] <Lachy> that's what browsers do
- # [02:22] <zcorpan> i think at least firefox appends the whitespace to the HEAD instead
- # [02:23] <zcorpan> not sure about the others
- # [02:24] * Parts: webben_ (n=benh@82.153.134.109)
- # [02:24] <zcorpan> and whitespace after </body> and </html> is appended to the body or whatever unclosed element was open in body... in firefox, and iirc
- # [02:27] <zcorpan> i don't feel strongly about it either way, but the way it's defined now preserves the "source code" after parsing and serializing (fwiw)
- # [02:28] <Lachy> firefox doesn't retain whitespace before the root
- # [02:30] * om_wet is now known as othermaciej
- # [02:31] <zcorpan> Lachy: compare data:text/html,<!doctype html>%0A<html> and data:text/html,<!doctype html><html> in DOM viewer (check the contents of HEAD)
- # [02:32] * Joins: annevk (n=annevk@203.17.70.52)
- # [02:32] <zcorpan> the first has a text node in HEAD, the second doesn't
- # [02:33] <Lachy> which version of FF are you using?
- # [02:33] <annevk> morning
- # [02:33] <Lachy> hi annevk
- # [02:33] <zcorpan> Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a4pre) Gecko/20070417 Minefield/3.0a4pre
- # [02:33] <Lachy> I've only got FF2
- # [02:33] <zcorpan> got the same results last time i checked, which was probably a year ago or something
- # [02:33] <Lachy> and it doesn't show the whitespace
- # [02:33] <zcorpan> really?
- # [02:33] <annevk> testing parsing?
- # [02:34] <zcorpan> annevk: yeah
- # [02:34] <Lachy> testing how whitespace outside of the root should be handled
- # [02:34] <zcorpan> whitespace between doctype and root
- # [02:34] <annevk> it would make sense to contribute to the testsuite format already developed for parsing I think
- # [02:34] <Lachy> annevk, see http://wordsandpictures.dyndns.org/html5/parsetree.html
- # [02:35] <annevk> fancy
- # [02:35] <annevk> what I meant was that there's already a testsuite input and output format
- # [02:36] <annevk> through some scripting that can be used for browser testsuites etc.
- # [02:36] * annevk slowly fetches 208 e-mails
- # [02:37] <zcorpan> Lachy: tested in fx2 now and indeed it gets dropped
- # [02:38] <Lachy> annevk, where have you gone touring around so far? Been to the rainforest yet?
- # [02:39] <Lachy> or sand surfing?
- # [02:39] <annevk> heh, all public-uwa messages get marked as spam
- # [02:39] <annevk> Lachy, I've been doing the rainforest walk
- # [02:39] <annevk> yesterday
- # [02:39] <annevk> it was quite exhausting
- # [02:39] <annevk> we walked over 20km
- # [02:40] <annevk> after about 12km rainforest we found some water
- # [02:40] <annevk> after 20km we found a beer and some chips :)
- # [02:40] <Lachy> :-)
- # [02:40] <Lachy> who did you go with?
- # [02:40] <annevk> the plan was to take a short walk of about an hour
- # [02:41] <annevk> but we missed something underway...
- # [02:41] <annevk> with marcos
- # [02:42] <zcorpan> data:text/html,%3Cp%3Ex%3C%2Fbody%3E%20%3C%21--foo--%3E%20%3C%2Fhtml%3E%20%3C%21--bar--%3E%20x
- # [02:42] <Lachy> jgraham, test FF3 with XML to see if it keeps whitespace outside of the root
- # [02:43] <annevk> that's conforming btw
- # [02:43] <annevk> and should be done per XML
- # [02:43] * Joins: aroben (n=adamrobe@adsl-70-231-234-224.dsl.snfc21.sbcglobal.net)
- # [02:43] <annevk> Opera should have some fix for that for the next major Opera release
- # [02:44] * Quits: aroben (n=adamrobe@adsl-70-231-234-224.dsl.snfc21.sbcglobal.net) (Client Quit)
- # [02:45] <zcorpan> Lachy: fx3 drops whitespace outside root in xml
- # [02:45] * Joins: aroben (n=adamrobe@adsl-70-231-234-224.dsl.snfc21.sbcglobal.net)
- # [02:45] <Lachy> right
- # [02:45] <annevk> dropping?!
- # [02:46] <annevk> it's simply not added per the XML grammar
- # [02:46] <zcorpan> what's the difference?
- # [02:47] * Quits: ajnewbold (n=fax_mach@unaffiliated/chuangtzu)
- # [02:48] <zcorpan> annevk: did you see http://xhtml.se/2007/04/20/im-not-arguing-im-just-making-fun-of-them/ ?
- # [02:51] <Dashiva> I'm famous!
- # [02:51] <zcorpan> lol
- # [02:52] <annevk> no
- # [02:53] <Dashiva> How come I'm the only one with "" around my name?
- # [02:53] <annevk> lol
- # [02:54] <Lachy> probably because he thinks that's not your real name
- # [02:54] <Dashiva> Unlike zcorpan? :)
- # [02:54] <zcorpan> :D
- # [02:54] <zcorpan> it is! promise!
- # [02:54] <Lachy> :-)
- # [02:55] <Dashiva> What's the status of that 'conformant html5' image?
- # [02:55] <zcorpan> ?
- # [02:56] <Lachy> do you mean the Valid HTML5 logo with Hixie's cat on it?
- # [02:56] <Dashiva> Yeah, that one
- # [02:56] <zcorpan> ah
- # [02:56] <zcorpan> http://simon.html5.org/temp/valid-html5.png
- # [02:56] <annevk> you should move that out of temp
- # [02:57] <Dashiva> Is valid the right word? Is using the cat legal? How many submarine patents are expected?
- # [02:57] <Lachy> ha!
- # [02:57] <zcorpan> ftp doesn't work for me today
- # [02:57] * Quits: mpt (n=mpt@canonical/launchpad/mpt) ("This computer has gone to sleep")
- # [02:57] <zcorpan> Dashiva: cat used without permission :)
- # [02:58] * zcorpan expects to get sued by Hixie
- # [02:58] <annevk> whoa
- # [02:58] <Lachy> permission from Hixie is implied, he hasn't complained about it
- # [02:58] <annevk> dhyatt has enough time to be co-editor
- # [02:58] <Lachy> probably just enough time to keep those who think we need 2 editors happy
- # [02:59] <Dashiva> Hixie can send him some diffs so Hyatt can commit them
- # [02:59] <zcorpan> heh
- # [03:00] <Lachy> why is http://simon.html5.org/temp/ forbidden? I want to see what other cool stuff you have hidden in there
- # [03:00] <annevk> it's temp
- # [03:00] <Lachy> so?
- # [03:00] <Dashiva> temp stuff shouldn't be accessible, it leads to people linking it and expecting permanence
- # [03:00] <zcorpan> don't have cool stuff there
- # [03:02] <zcorpan> http://simon.html5.org/temp/program my gym program from last year, just started with it again now
- # [03:02] <zcorpan> in swedish
- # [03:04] <zcorpan> http://simon.html5.org/temp/browsers/ icons used in the web-apps-tracker
- # [03:04] <annevk> heh, collegue of dhyatt proposes dhyatt and then people wonder whether dhyatt actually agreed to it...
- # [03:04] <annevk> i suppose it's possible he didn't...
- # [03:05] * Quits: othermaciej (n=mjs@dsl081-048-145.sfo1.dsl.speakeasy.net)
- # [03:05] <zcorpan> not when the mail proposing it says "he has said he's willing to do it"
- # [03:06] <annevk> oh right
- # [03:06] <annevk> guess that tells how closely I read some of my e-mail initially
- # [03:07] * Quits: tantek (n=tantek@dsl001-150-252.sfo1.dsl.speakeasy.net)
- # [03:08] * Joins: othermaciej (n=mjs@dsl081-048-145.sfo1.dsl.speakeasy.net)
- # [03:23] * moeffju[Away] is now known as moeffju
- # [03:24] * Quits: othermaciej (n=mjs@dsl081-048-145.sfo1.dsl.speakeasy.net) (Success)
- # [03:34] <zcorpan> Lachy: http://www.google.com/search?q=site%3Asimon.html5.org%2Ftemp should bring up anything "cool" i've hidden there :)
- # [03:34] * Dashiva links to hsivonen's checker using valid-html5 image. Go precedent!
- # [03:34] * zcorpan doesn't think that's a good idea
- # [03:34] <Lachy> why not?
- # [03:35] <Dashiva> "Some markup validation/checking services give out a small badge for advertising that a document passed. This service does not."
- # [03:36] <zcorpan> stickies linking to a checker generally gives the impression that the checker result is some sort of status indicator
- # [03:37] <zcorpan> the valid html5 image was mostly a joke on my part and i think hsivonen wants his tool to be a QA tool, not a status indicator
- # [03:37] <annevk> ooh, security
- # [03:38] <Dashiva> I fully agree, but that doesn't stop me from being subversive
- # [03:39] <zcorpan> fair enough :)
- # [03:40] <zcorpan> although i'd suggest that the link point to whatwg.org/html5 if anything
- # [03:40] * annevk wonders when Web Forms 2 will be merged with HTML5
- # [03:42] <zcorpan> Dashiva: also see first paragraph in (4) in http://lists.w3.org/Archives/Public/public-html/2007JanMar/0433.html
- # [03:43] <Dashiva> I think the page is valid pretty much anything in its current state
- # [03:44] <Dashiva> Well, the <img> tag doesn't have the / needed for xhtml
- # [03:44] <annevk> <img> can also be <img></img> in XHTML
- # [03:44] <annevk> fwiw
- # [03:45] * gavins is now known as gavin
- # [03:50] * Quits: aroben (n=adamrobe@adsl-70-231-234-224.dsl.snfc21.sbcglobal.net)
- # [04:03] * Joins: tantek (n=tantek@adsl-63-195-114-133.dsl.snfc21.pacbell.net)
- # [04:06] * moeffju is now known as moeffju[ZzZz]
- # [04:08] <annevk> http://www.w3.org/mid/op.tq008yhnsmjzpq@acer3010.lan
- # [04:08] * Quits: bzed (n=bzed@dslb-084-059-127-023.pools.arcor-ip.net) ("Leaving")
- # [04:10] <annevk> http://therealcrisp.xs4all.nl/blog/2007/04/22/html5-microsoft-and-the-opt-in-catch/
- # [04:11] <annevk> http://labs.google.com/goog411/
- # [04:26] <annevk> + <h4>Origin</h4>
- # [04:26] <annevk> + <!-- Hallowed are the Ori -->
- # [04:26] <annevk> heh
- # [04:33] <Lachy> annevk, which page had that ori comment in it?
- # [04:34] <annevk> HTML5?
- # [04:34] <annevk> Hixie defined origin for scripts
- # [04:41] <Lachy> That solution on crisp's blog won't work as-is, we tried it already
- # [04:42] <Lachy> Though, my latest response from Chris (off list) indicates that he'd listen to a solution that starts out as an opt-in and can transition to an opt-out, once the dependence on opt-ins isn't required
- # [04:43] <zcorpan> i do think making [if IE] return false in ie8 would help them not breaking things, but he didn't reply to that suggestion. perhaps i should bug him about that again
- # [04:43] <annevk> it requires too many fixes
- # [04:43] <annevk> in one go
- # [04:44] <zcorpan> then in ie9
- # [04:44] <annevk> if we got the proper testsuites ready by then...
- # [04:44] * zcorpan will be working on that this summer
- # [04:45] <Lachy> I think the only way we're going to get Chris to compromise is to make HTML4/XHTML1 DOCTYPEs require an opt-in, and use the change in DOCTYPE in HTML to convert to an opt-out system
- # [04:46] <annevk> we shouldn't have to do much
- # [04:46] <annevk> this would only apply to IE
- # [04:46] <annevk> i hope
- # [04:46] <Lachy> that way, IE can use the time it takes to transition from HTML4/XHTML1 to HTML5 to improve its standards support and get to a point where it's close enough to interop with outhers, to not require an explicit opt-in
- # [04:46] <Lachy> yes, it would be an IE only opt-in/out
- # [04:47] <annevk> personally I want to keep fixing quirks mode, documenting quirks, etc.
- # [04:47] <Lachy> yeah, that would be good, but there's no chance IE is going to change it's quirks mode at all
- # [04:48] <zcorpan> in any case, it should be possible for authors to trigger a "latest standards mode", author against opera/firefox/safari, then work around bugs in current IEs without making assumptions about IE.next
- # [04:48] <zcorpan> annevk: yeah, me too
- # [04:53] * annevk wonders if http://dev.w3.org/cvsweb/~checkout~/2006/waf/widgets/Overview.html?content-type=text/html;%20charset=utf-8#autodiscovery is good enough to define a new link type
- # [04:55] * zcorpan thinks it is
- # [04:56] <zcorpan> or how is it handled today in opera?
- # [04:56] <Lachy> annevk, typo: "...to create an hyperlink" s/an/a/
- # [04:58] <zcorpan> ah, Content-Type: application/x-opera-widgets
- # [04:59] <annevk> I'm pretty sure it's always an before h...
- # [04:59] <Lachy> no
- # [04:59] <zcorpan> depends on how it's pronounced
- # [04:59] <zcorpan> an hour
- # [04:59] <Lachy> it doesn't depend on the letter, it depends on the sound
- # [05:01] <annevk> yeah nm me
- # [05:02] <zcorpan> annevk: if there is application/widget, why is there a need for rel="widget"?
- # [05:02] <zcorpan> surely the opera widget page works fine with only the mime type
- # [05:03] <Lachy> same reason there is rel=feed
- # [05:03] <zcorpan> for UI?
- # [05:03] <Lachy> the rel allows the UA to detect the widget before fetching it and offer a better UI
- # [05:04] <zcorpan> ok
- # [05:04] <annevk> right
- # [05:12] <annevk> fixed the typo and some other stuff fwiw
- # [05:14] <annevk> might be back later... food
- # [05:24] * Joins: aroben (n=adamrobe@adsl-70-231-234-224.dsl.snfc21.sbcglobal.net)
- # [05:36] * Quits: annevk (n=annevk@203.17.70.52) (Read error: 110 (Connection timed out))
- # [05:47] <Hixie> anyone got any pet bugs in browsers that they'd want fixed? any bugs. any browsers.
- # [05:47] <Hixie> ideally a DOM2 or HTML4 bug, maybe CSS2.1
- # [05:47] <Hixie> or JavaScript
- # [05:50] <Lachy> I want to see display: table-* properties fixed to work with less bugs in FF
- # [05:51] * zcorpan points to http://simon.html5.org/test/ie7b2-bugs/
- # [05:51] * Hixie takes notes
- # [05:52] <Hixie> Lachy: any specifics in mind?
- # [05:52] <Hixie> zcorpan: wow cool thanks
- # [05:52] <zcorpan> np
- # [05:53] <Lachy> see my site in FF. Sometimes the columns don't work properly, but can usually be fixed by reloading
- # [05:54] <zcorpan> Lachy: oh yeah i get that too on some layouts
- # [05:54] <zcorpan> very unpredictable
- # [05:54] <Hixie> weird
- # [05:54] <Hixie> http://simon.html5.org/test/ie7b2-bugs/003.html <-- wtf is IE doing here
- # [05:55] <Lachy> also, if you create elements with JS and insert them into the DOM with display: table-cell; applied, it doesn't work well either
- # [05:55] <zcorpan> Hixie: no idea
- # [05:55] <Lachy> I'll create a test case
- # [05:56] * Quits: tantek (n=tantek@adsl-63-195-114-133.dsl.snfc21.pacbell.net)
- # [05:57] <zcorpan> Hixie: also http://simon.html5.org/test/opera/
- # [05:57] <Hixie> cool, more tests
- # [06:00] <Lachy> http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C%21DOCTYPE%20html%3E%0A%3Cstyle%3E%0Adiv%20%7B%20display%3A%20table-cell%3B%20border%3A%201px%20solid%20blue%3B%20%7D%0A%3C/style%3E%0A%3Cp%3EThese%20columns%20are%20in%20the%20markup%3A%0A%3Cdiv%3Ecolumn1%3C/div%3E%0A%3Cdiv%3Ecolumn2%3C/div%3E%0A%3Cdiv%3Ecolumn3%3C/div%3E%0A%0A%3Cp%3EThese%20columns%20are%20created%20by%20a%20script%3A%0A%0A%3Cscript%3E%0Avar%20div%3B%0Afor%20%28va
- # [06:00] <Lachy> r%20i%20%3D%201%3B%20i%20%3C%3D%203%3B%20i++%29%20%7B%0A%20%20div%20%3D%20document.createElement%28%22div%22%29%3B%0A%20%20div.innerHTML%20%3D%20%22column%20%22%20+%20i%3B%0A%20%20document.body.appendChild%28div%29%3B%0A%7D%0A%3C/script%3E%0A
- # [06:00] <Hixie> hit the upload link
- # [06:00] <Lachy> I uploaded that to the DOM viewer clipboard, so you can just download it
- # [06:00] <Hixie> cool
- # [06:00] <Hixie> got it
- # [06:03] * Joins: annevk (n=annevk@203.17.70.52)
- # [06:05] * annevk wouldn't mind interop on feature "x"
- # [06:06] <Hixie> any specifics in mind?
- # [06:07] <annevk> innerHTML
- # [06:07] <annevk> HTML table model
- # [06:07] <Hixie> zcorpan: these are really good tests
- # [06:07] <zcorpan> :)
- # [06:08] <annevk> heh, why do you think we hired him?
- # [06:08] * annevk doesn't have much tests for either of his suggestions
- # [06:11] <annevk> what are the feature requests for?
- # [06:12] * annevk wonders why Opera doesn't do the dynamically generated <div> stuff correctly
- # [06:13] <annevk> ooh, Firefox has problems too
- # [06:13] <Lachy> annevk, are you talking about the test I made? What does opera get wrong?
- # [06:13] <annevk> some build i have doesn't render the generated divs at all
- # [06:13] <zcorpan> same here
- # [06:14] <annevk> actually, all builds so far...
- # [06:14] <Hixie> i'm preparing for the writing of acid3
- # [06:14] <Lachy> that's weird. It renders them as soon as I click download, but they disappear after the source is edited
- # [06:14] <Hixie> trying to get bugs lined up
- # [06:14] <annevk> (this would be covered by my "HTML table model" suggestion above which really should be named "HTML table layout model")
- # [06:14] <Lachy> oh, awesome!
- # [06:15] <Hixie> especially interested in DOM tests
- # [06:15] <Hixie> and JS
- # [06:15] <Lachy> when do you expect acid3 to be released?
- # [06:16] <Hixie> when it's ready
- # [06:16] <Lachy> fair enough
- # [06:16] <Hixie> http://hixie.ch/www/tests/evil/acid/003/ is where things will happen at first
- # [06:16] <annevk> http://tc.labs.opera.com/html/canvas/getContext/2d/ still shows quite some differences between browsers
- # [06:16] <annevk> although I suppose <canvas> may be too new
- # [06:17] <Hixie> yeah this is going to cover only things that were specced around 2003 or earlier
- # [06:17] <annevk> k
- # [06:17] <annevk> "PNG that contains some HTML"?!
- # [06:18] * Quits: psa (n=yomode@posom.com) (Remote closed the connection)
- # [06:18] <annevk> IE still has a bunch of CSS parsing issues
- # [06:19] <annevk> http://annevankesteren.nl/2007/02/ie7-css-hacks contains a list of copy & paste testcases
- # [06:20] <annevk> oh, \n\r in document.write() and scripts in general might be good
- # [06:20] <annevk> or \r\n whatever it is
- # [06:21] <Hixie> has to be only things that can be seriously argued out of existing specs, unfortunately
- # [06:23] <zcorpan> having ie support the events model correctly would be nice, i think... and addEventListener, event.target, etc.
- # [06:23] <Hixie> yeah, gonna test that for sure
- # [06:23] <annevk> handling of <i><b></i> is to new too?
- # [06:23] <annevk> or is it too new too?
- # [06:24] <Lachy> I didn't realise FF3 passed acid2 already. That's awesome! :-)
- # [06:24] <Hixie> parsing of html is too new if it's error handling
- # [06:24] <Hixie> Lachy: yeah, hence it being time for acid3
- # [06:24] <zcorpan> Lachy: better late than never :)
- # [06:26] <zcorpan> getAttribute("for") in IE
- # [06:26] <Lachy> Hixie, add this: <!--[if IE]>FAIL<![endif]--> :-)
- # [06:26] <zcorpan> (IE needs getAttribute("htmlFor"))
- # [06:27] <Lachy> same for "className"
- # [06:27] <zcorpan> yeah
- # [06:27] <annevk> <style type="">
- # [06:27] <Lachy> and there's something else too
- # [06:27] * Lachy checking the DOM ref...
- # [06:27] <annevk> versus <style type="unknown"> versus <style> versus <style type="text/css">
- # [06:28] <annevk> maybe in the CSSOM? for float? which becomes .cssFloat as attribute
- # [06:28] <Lachy> I think getAttribute("http-equiv"); also has the problem
- # [06:28] <zcorpan> likely
- # [06:29] <zcorpan> using getAttribute on event handler attributes doesn't work right either iirc
- # [06:29] <annevk> events on body / document / window
- # [06:31] <Lachy> event object isn't passed to the handler in IE, needs window.event
- # [06:31] <annevk> already mentioned
- # [06:31] <Lachy> oh, sorry
- # [06:33] <Lachy> http://developer.mozilla.org/en/docs/Core_JavaScript_1.5_Reference:Global_Objects:Array#Methods some of those aren't supported in IE
- # [06:33] <annevk> i suppose investigating in what Dean Edwards has to work around in his library might give some good suggestions on to what to include
- # [06:34] <Lachy> I assume you'll test the NS methods in the DOM too?
- # [06:35] <Hixie> already looked at dean's stuff
- # [06:35] <Hixie> nah, no point testing namespaces really
- # [06:35] <Lachy> ok
- # [06:35] <Hixie> since we don't particularly want anyone to use them
- # [06:35] <Lachy> alright
- # [06:36] <annevk> nasty SVG bits?
- # [06:39] <Lachy> Hixie, <link rel="icon"> - IE uses rel="shortcut icon"
- # [06:40] <Lachy> although, technically it wasn't specified till recently, icons have been widely supported for a while
- # [06:45] <Hixie> yeah but there's no spec that says it has to be "icon"
- # [06:46] <Lachy> I suppose you're not counting HTML5 then?
- # [06:46] <annevk> before 2003
- # [06:46] <annevk> HTML5 _started_ in 2004
- # [06:46] <Lachy> I know that, but it's been widely supported for much longer than that
- # [06:46] <annevk> the other requirement is for it to be in a spec
- # [06:47] <Hixie> unless i can point at a REC with clear conformance criteria, it doesn't count
- # [06:47] <Lachy> ok
- # [06:47] <Hixie> since you can argue that it isn't required for compliance
- # [06:47] <annevk> that leaves out testing HTML features :p
- # [06:50] * Quits: aroben (n=adamrobe@adsl-70-231-234-224.dsl.snfc21.sbcglobal.net)
- # [06:52] <Hixie> mostly, yeah
- # [06:53] <virtuelv_> One thing that definetly would need to go into acid3 is how 'apply' works
- # [06:54] <Hixie> which spec?
- # [06:54] <Lachy> virtuelv_, in JavaScript?
- # [06:54] <virtuelv_> Lachy: yup
- # [06:54] <virtuelv_> I've run into cases where the behavior is rather exotic across browser
- # [06:55] <virtuelv_> +s
- # [06:55] <Lachy> is it defined in ECMA 262?
- # [06:57] <virtuelv_> iirc, yes
- # [06:57] <virtuelv_> lemme check, though
- # [06:57] <Lachy> yep, it's on page 87 (section 15.3.4.3
- # [06:58] <Hixie> if you can give me a testcase and explain how it tests behaviour described in the spec i can add it, sure
- # [07:00] <Hixie> anne, wouldn't rel=widget be better as a type="" value?
- # [07:00] <annevk> similar to rel=feed
- # [07:01] <Lachy> Hixie, why?
- # [07:01] <virtuelv_> Hixie: will do, bug me at the end of the coming week
- # [07:01] <Hixie> virtuelv_: k
- # [07:01] <Lachy> rel=widget is easier for authors to type and remember, compared with MIME types
- # [07:01] <Hixie> i guess
- # [07:02] <virtuelv_> inability to add stuff to the Element prototype is also, IMHO one thing browsers should get right in 2007
- # [07:03] <virtuelv_> s/in//
- # [07:03] <Hixie> unfortunately prototypes for elements aren't defined anywhere (yet)
- # [07:04] <Hixie> so that's not a standards compliance issues
- # [07:04] <Hixie> issue
- # [07:04] <annevk> It's more consistent with rel=stylesheet rel=feed etc. too. It means that authors have to learn one less trick (the trick that type= might lead to special handling too).
- # [07:04] <Lachy> but Elements are just Objects in ECMAScript, and Object is defined
- # [07:04] <virtuelv_> Hixie: that's rather unfortunate, really
- # [07:05] <Hixie> Lachy: if you can find a spec that says that, let me know :-)
- # [07:05] <Lachy> http://www.w3.org/TR/DOM-Level-2-Core/ecma-script-binding.html binds all the DOM objects to Object in ECMAScript
- # [07:06] <Lachy> ECMA 262 defines .prototype
- # [07:06] <Lachy> and it's how all other major browsers implement it, AFAIK
- # [07:07] <Hixie> but where does it say that the object has to have a prototype?
- # [07:08] <virtuelv_> Hixie: all objects that have constructors must have a prototype
- # [07:08] <Hixie> but Element doesn't have a constructor
- # [07:10] <virtuelv_> sigh. Neither does Node
- # [07:10] <Lachy> all Objects have constructors though. It inherits from Object, and so it has a constructor. It's just not available to scripts
- # [07:14] <annevk> but that Element is available somewhere in the global scope as attribute is also not defined
- # [07:20] <Lachy> border-collapse should be tested too (or was it tested in acid2 already?) http://lachy.id.au/dev/css/tests/bordercollapse/bordercollapse.html
- # [07:21] <Lachy> I think IE is the only one to fail those tests
- # [07:22] <Hixie> not really interested in css
- # [07:22] <Lachy> ok
- # [07:23] <annevk> location.hash when the page contains a single # is sometimes a compat issue iirc
- # [07:23] <annevk> oh, nm
- # [07:24] <Hixie> location was specced post 2003
- # [07:24] <Hixie> namely, last week
- # [07:24] * annevk was under the illusion location was somehow part of old dom specs
- # [07:24] <annevk> did you and Maciej agree that the Window spec will be merged into HTML5?
- # [07:32] * Quits: zcorpan (n=zcorpan@84-216-42-9.sprayadsl.telenor.se) (Read error: 110 (Connection timed out))
- # [07:39] <Lachy> OMG WTF? "Long release cycles are a very good thing" on public-html
- # [07:40] <annevk> citing IE6 as an example
- # [07:40] <Lachy> Frequent releases of increasingly buggy browsers may be a bad thing, but the quicker we can get more standards compliant browsers into the market the better
- # [08:04] <Hixie> annevk: haven't asked
- # [08:05] <annevk> k
- # [08:05] <Hixie> annevk: but since Window was no longer being developped, and I needed to spec things that depended on it...
- # [08:05] <annevk> yeah, just wondering
- # [08:06] <annevk> now the chairs sort of "accepted" HTML5 will you merge in WF2? or let dhyatt do it? :p
- # [08:14] <Hixie> they have?
- # [08:16] <Lachy> Hixie, Chris responded fairly positively about the HTML5 spec
- # [08:17] <Hixie> uri?
- # [08:17] <Lachy> http://www.w3.org/mid/5C276AFCCD083E4F94BD5C2DA883F05A27D9CDDC38@tk5-exmlt-w600.wingroup.windeploy.ntdev.microsoft.com
- # [08:20] <annevk> whoa, schepers is asking for a third editor...
- # [08:20] <Lachy> annevk, I was just typing the same thing :-)
- # [08:21] <Lachy> 2 editors is more than enough
- # [08:23] <Hixie> Lachy: it wasn't clear to me that he was accepting the proposal maciej put forward exactly. but yeah, that was positive regarding the spec itself.
- # [08:23] <Hixie> (my reply was http://lists.w3.org/Archives/Public/public-html/2007Apr/1241.html)
- # [08:24] <Hixie> i'm curious as to what people think "editor" means
- # [08:25] <Hixie> why does it matter what the opinions of the editor are?
- # [08:25] <Lachy> yeah, I don't understand why he wants the whatwg to stop either
- # [08:26] <Hixie> chris?
- # [08:26] <Lachy> yes
- # [08:30] <Lachy> I have no idea how I should respond to this about selectors API http://lists.w3.org/Archives/Public/public-webapi/2007Apr/0025.html
- # [08:30] <Lachy> I already said it's not wise to explicitly introduce undefined behaviour
- # [08:32] <annevk> Does anyone know what the last dot http://www.w3.org./ can possibly do?
- # [08:33] <Hixie> Lachy: well, what you have now doesn't guarentee interop, so what they're asking for doesn't really change that
- # [08:33] <Hixie> annevk: guarentees you're not going to do local dns lookup (with a suffix, e.g.)
- # [08:33] <annevk> yeah, just say that authors should not assume a specific prefix invocation algorithm or something
- # [08:33] <Hixie> bit like the leading / in /foo/bar
- # [08:34] <Hixie> saying authors "should not" is just a waste of time :-)
- # [08:34] <Hixie> won't do anything
- # [08:34] <Hixie> especially for APIs
- # [08:34] <annevk> should create a side-effect resolver...
- # [08:34] <Lachy> but I don't see what problems it solves by me taking it out
- # [08:35] <annevk> it makes things simpler
- # [08:35] <Lachy> so you think I should?
- # [08:35] <annevk> well, i was convinced
- # [08:35] <annevk> so yes
- # [08:36] <annevk> it's unfortunate jonas is so late with catching up with e-mail otherwise I would have fixed this ages ago (or never specced it)
- # [08:37] <Lachy> I suppose it doesn't really help authors much knowing their function is only going to be invoked once per prefix either
- # [08:38] <annevk> i suppose it's nice to point it out in interop concerns so people who are designing something similar know what to do
- # [08:39] <annevk> (same for people who wonder why it differs between browsers)
- # [08:48] <Hixie> annevk: "the ToASCII algorithm must apply successfully (without errors)"
- # [08:48] <Hixie> how can it fail?
- # [08:49] * Quits: csarven (n=nevrasc@modemcable081.152-201-24.mc.videotron.ca)
- # [08:54] * Joins: mpt (n=mpt@canonical/launchpad/mpt)
- # [09:01] * Joins: aroben (n=adamrobe@adsl-70-231-234-224.dsl.snfc21.sbcglobal.net)
- # [09:01] * Quits: aroben (n=adamrobe@adsl-70-231-234-224.dsl.snfc21.sbcglobal.net) (Remote closed the connection)
- # [09:02] * Joins: aroben (n=adamrobe@adsl-70-231-234-224.dsl.snfc21.sbcglobal.net)
- # [09:07] * Joins: peepo (n=Jay@host86-129-174-162.range86-129.btcentralplus.com)
- # [09:07] * Quits: annevk (n=annevk@203.17.70.52) (Read error: 110 (Connection timed out))
- # [09:21] <Lachy> Hixie, this has a good list of IE bugs you could add to acid3 http://easy-designs.stikipad.com/ie-next-wishlist
- # [09:23] <Hixie> IE bugs are easy to find :-)
- # [09:23] <Hixie> it's other browsers that are the problem
- # [09:24] <Lachy> right
- # [09:24] <Hixie> in fact, i've added all of _one_ test to the page so far and it breaks in IE drastically.
- # [09:25] <Lachy> yeah, I saw that.
- # [09:25] <Lachy> I don't understand what causes it to break though
- # [09:26] <Lachy> is it because there's no font-size specified for body?
- # [09:26] * Quits: peepo (n=Jay@host86-129-174-162.range86-129.btcentralplus.com) ("later")
- # [09:27] <Hixie> i'm guessing it doesn't support inline-block
- # [09:27] <Hixie> though i thought it did
- # [09:28] <Lachy> it does for inline level elements only
- # [09:30] <Lachy> do you have some sort of plan for the final rendering, or are you just making it up as you go?
- # [09:35] <Hixie> it'll look mostly like it does now in firefox
- # [09:35] <Hixie> but i'm making it up as i go along
- # [09:35] <Hixie> and i want to add animation
- # [09:45] <Hixie> why is opera not using arial
- # [09:45] <Hixie> wtf
- # [09:48] <Lachy> heh, I keep forgetting to switch to FF3 when looking at test cases :-)
- # [09:49] <Lachy> maybe opera doesn't support :root
- # [09:50] <Hixie> seems unlikely
- # [09:50] <Hixie> but maybe
- # [10:14] <Hixie> last chance to see what the test is intended to look like before i break it
- # [10:14] <Hixie> http://www.hixie.ch/tests/evil/acid/003/index.html
- # [10:14] <Lachy> yeah, I just looked at it. It looks ok
- # [10:15] <Hixie> it already shows bugs in every browser except firefox
- # [10:15] <Hixie> and the only reason it works in firefox is i changed 'document' to 'window'
- # [10:15] <Hixie> for hte load event
- # [10:15] <Lachy> are you going to change it back to document?
- # [10:15] <Hixie> yes
- # [10:15] <Hixie> there's no window in 2003 standards
- # [10:16] <Lachy> yeah
- # [10:49] * Joins: hasather (n=hasather@81-235-209-174-no62.tbcn.telia.com)
- # [10:52] * Quits: hasather (n=hasather@81-235-209-174-no62.tbcn.telia.com) (Remote closed the connection)
- # [10:53] * Joins: hasather (n=hasather@81-235-209-174-no62.tbcn.telia.com)
- # [11:23] * Joins: bzed (n=bzed@dslb-084-059-125-165.pools.arcor-ip.net)
- # [11:36] * Joins: tantek (n=tantek@adsl-63-195-114-133.dsl.snfc21.pacbell.net)
- # [11:54] <Hixie> http://www.hixie.ch/tests/evil/acid/003/ and http://www.hixie.ch/tests/evil/acid/003/reference.html
- # [11:54] <Hixie> should look the same everywhere
- # [11:54] <Hixie> i haven't even tried adding tests yet and they don't
- # [11:54] * Lachy checks every browser he has...
- # [11:56] <Lachy> nice, FF3 gets it right :-)
- # [11:56] <Hixie> it does?
- # [11:56] <Lachy> it did a second ago, not now
- # [11:56] <Hixie> i just get 0%
- # [11:56] <Hixie> oh you probably tested it before i switched it back to document
- # [11:56] <Lachy> yeah
- # [11:57] <Hixie> the reference rendering should be pixel-perfect, including font rasterisation
- # [11:58] <Hixie> but really the reference rendering is a test on to its own
- # [11:58] <Lachy> Opera 9 is doing surprisingly poorly
- # [11:59] <Hixie> opera 9 is hitting two bugs that i know of
- # [11:59] <Hixie> one, it's getting screwed by not having a <head> in the DOM, because one of the selectors goes through the <head>
- # [11:59] <Hixie> and two, it doesn't seem to do :root
- # [12:01] * Hixie throws in some :first-child/:last-child tests
- # [12:01] <Lachy> which selector goes through head?
- # [12:01] <Hixie> *:first-child + * > * > p
- # [12:01] <Hixie> that's equivalent to head + body > div > p
- # [12:01] <Lachy> ah, I see
- # [12:02] <Hixie> as in, the only ps that it matches will match because the first simple selector there matches the head
- # [12:02] <Hixie> and since opera has no head...
- # [12:05] <Hixie> anyway i should go to bed
- # [12:05] <Hixie> nn
- # [12:05] <Lachy> cya
- # [12:06] * Joins: Toolskyn (n=toolskyn@adsl-dc-266ef.adsl.wanadoo.nl)
- # [12:21] * Joins: ROBOd (n=robod@86.34.246.154)
- # [12:27] * Joins: zcorpan (n=zcorpan@84-216-43-71.sprayadsl.telenor.se)
- # [12:59] * moeffju[ZzZz] is now known as moeffju
- # [13:13] * Quits: aroben (n=adamrobe@adsl-70-231-234-224.dsl.snfc21.sbcglobal.net)
- # [13:15] * Joins: met_ (n=Hassman@r5bx220.net.upc.cz)
- # [13:26] <gsnedders> anyone know what percentage of XHTML docs are served as XML?
- # [13:26] <gsnedders> well, per-mille may be more useful, actually :P
- # [14:38] * Quits: zcorpan (n=zcorpan@84-216-43-71.sprayadsl.telenor.se) (Read error: 110 (Connection timed out))
- # [14:48] * Parts: hasather (n=hasather@81-235-209-174-no62.tbcn.telia.com)
- # [14:49] * Joins: hasather (n=hasather@81-235-209-174-no62.tbcn.telia.com)
- # [14:57] * Joins: ajnewbold (n=fax_mach@unaffiliated/chuangtzu)
- # [15:08] * Joins: zcorpan (n=zcorpan@84-216-42-223.sprayadsl.telenor.se)
- # [15:24] * Joins: Toolskyn88 (n=toolskyn@adsl-dc-266ef.adsl.wanadoo.nl)
- # [15:29] * Quits: bzed (n=bzed@dslb-084-059-125-165.pools.arcor-ip.net) ("Leaving")
- # [15:33] <zcorpan> Hixie: acid2 was easier to see if it was correct or not (without looking at the reference rendering)
- # [15:39] <hasather> zcorpan: agreed
- # [15:39] <Lachy> zcorpan, any suggestions for a better picture?
- # [15:39] * Quits: Toolskyn (n=toolskyn@adsl-dc-266ef.adsl.wanadoo.nl) (Read error: 110 (Connection timed out))
- # [15:40] <hsivonen> what test are you talking about?
- # [15:40] <Lachy> acid 3
- # [15:40] <Lachy> http://www.hixie.ch/tests/evil/acid/003/
- # [15:40] <Lachy> it's still a work in progress though
- # [15:42] <zcorpan> Lachy: hm, dunno... a questionmark inside a circle?
- # [15:42] * zcorpan hides
- # [15:42] <Lachy> I thought maybe a cat
- # [15:43] * Joins: SpookyET (i=user@75.138.70.34)
- # [16:01] <zcorpan> "Awareness of future trends." ?
- # [16:10] <hsivonen> zcorpan: indeed
- # [16:15] <hsivonen> It'll be interesting to see the diff for Web Apps 1.0 from the day the HTML WG was announced to the day the WG accepts HTML5 as the starting point (assuming that happens)
- # [16:24] <SpookyET> Opera allready supports some of it according to Wikipedia.
- # [16:26] <Lachy> SpookyET, yes, it has support for Web Forms 2.0 and a few things from Web Apps 1.0 like <canvas> and <event-source>
- # [16:29] <hsivonen> SpookyET: I meant it'll be interesting to see how much real work Hixie and the folks on the WHATWG list got done while public-html generated heat
- # [16:32] * Quits: zcorpan (n=zcorpan@84-216-42-223.sprayadsl.telenor.se) (Read error: 110 (Connection timed out))
- # [16:33] <SpookyET> hsivonen: http://en.wikipedia.org/wiki/Comparison_of_layout_engines_(WHATWG)
- # [16:33] <SpookyET> see all that green on the right?
- # [16:33] * Joins: zcorpan (n=zcorpan@84-216-40-169.sprayadsl.telenor.se)
- # [16:33] <Lachy> SpookyET, what's your point? we know that already
- # [16:36] <SpookyET> Lachy: I was talking to hsivonen
- # [16:36] <Lachy> I realise that, but he knows too
- # [16:37] <SpookyET> Let him speak for himself.
- # [16:39] <Lachy> I'm trying to determine what your point is by telling people things we already know?
- # [16:40] <zcorpan> hsivonen: yeah, that's something to blog about :)
- # [16:48] <Dashiva> At the moment it looks like a lot of colorblind scandinavian flags in Opera
- # [16:48] <hsivonen> SpookyET: I don't understand what point you are trying to make to me
- # [16:48] <SpookyET> No point. I was stating a fact.
- # [16:49] <hsivonen> ok
- # [16:49] <Lachy> you can get it to work in Firefox 3 if you just change document.addEventListener to window.addEventListener. (that is, until Hixie adds more tests to it later)
- # [16:51] <hsivonen> gotta respect the people who send patches for Lynx. last time I looked the source wasn't pretty
- # [16:54] <Dashiva> Acid3 uses HTML 4.0 doctype... am I missing something, or was it a random choice?
- # [16:55] <zcorpan> Dashiva: Hixie probably copied it from damowmow.com or something
- # [16:55] <hsivonen> that's what happened with Acid2, IIRC
- # [16:56] <zcorpan> or wait, damowmow uses 4.01
- # [16:56] <zcorpan> then hixie.ch or something :)
- # [16:58] <hsivonen> the comment in the damowmow source is funny
- # [16:58] <zcorpan> heh
- # [16:59] <SpookyET> Adobe's Apollo is pretty interesting. I just stumbled upon it. It seems to be a competitor to Microsofts XAML.
- # [17:00] <hsivonen> I don't understand the Adobe-internal politics that produced Apollo considering that they have Flash
- # [17:00] <SpookyET> It uses Flash, Acrobat, HTML, CSS, JavaScript.
- # [17:00] <zcorpan> the new Ajax?
- # [17:01] <SpookyET> The cool thing about Apollo is that it uses WebKit, while Microsoft's solution uses IE's rendering engine.
- # [17:01] <SpookyET> http://www.niallkennedy.com/blog/archives/2007/03/adobe-apollo.html
- # [17:01] <SpookyET> zcorpan: It's not AJAX. It's designed to merge desktop and web apps.
- # [17:02] <Lachy> it's like their own version of Widgets
- # [17:02] <SpookyET> It's not widgets, but full apps.
- # [17:02] <zcorpan> SpookyET: ok
- # [17:03] <SpookyET> https://ssl.blogs.zdnet.com/blogs/Stewart/images/apollo_mac_library_large.png
- # [17:04] <SpookyET> It looks like you create a hybrid desktop + web app instead of a desktop client that connects to a web service.
- # [17:04] <Lachy> Interesting. I just found out that for DOCTYPE sniffing, IE ignores the language code at the end of <!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">, whereas in FF it needs to be EN or it triggers standards mode
- # [17:05] <hsivonen> Lachy: interesting indeed. I wonder why the IE team chose to look inside the identifier
- # [17:05] <Dashiva> They're big on localization? :)
- # [17:05] <hsivonen> Lachy: Gecko lowercases the string and compares it against a list of strings
- # [17:05] <Lachy> probably because there are silly people around who change the "EN" to their own native langauge
- # [17:06] <Lachy> it makes sense to do what IE does
- # [17:06] <hsivonen> Lachy: yeah, but shouldn't those silly people get the quirks mode
- # [17:06] <Lachy> yes, they do in IE. They don't in FF, that's my point
- # [17:06] <hsivonen> Lachy: considering that sniffing is a heuristic for deciding if the author had clue
- # [17:08] <Philip`> I'd like a "-//W3C//DTD HTML 4.01 Transitional//EN-GB" with <centre> and <font colour="grey">
- # [17:08] <Lachy> and in IE, the number needs to be 2, 3 or 4 to trigger quirks mode (decimal points ignored)
- # [17:09] <zcorpan> ie's doctype sniffing is insane. the way the system id also matters iirc
- # [17:09] <zcorpan> er, the say the system id is written...
- # [17:09] * zcorpan -> food
- # [17:10] * Quits: Toolskyn88 (n=toolskyn@adsl-dc-266ef.adsl.wanadoo.nl) (Remote closed the connection)
- # [17:10] <Lachy> Philip`, the langauge code is basically irrelevant even in SGML. It's supposed to refer to the language that the DTD is written in, but I have no idea why that's relevant at all
- # [17:14] <Philip`> (Incidentally, I was actually surprised when seeing that CSS3 Color supports "grey" and that it actually works in browsers - I'm certain it used to be interpreted as "green" in the past (like, uh, Netscape 3 or something))
- # [17:15] <Lachy> wow
- # [17:15] <Dashiva> I remember some early Opera canvas bug which interpreted random strings as colors. E.g. "bogus" became pink, if I recall correctly
- # [17:16] <Philip`> I just want to know how anybody is ever going to implement the "flavor" colour
- # [17:16] <hasather> "crap" became brown in IE before, I think. Dunno if that still works
- # [17:17] <Lachy> hasather, yep, still works!
- # [17:17] <Lachy> bogus becomes ping in IE too!
- # [17:18] <Dashiva> I suspect it has to do with trying to interpret it as hex, but the details elude me
- # [17:19] <Lachy> it seems that it throws away any letters other than [A-Fa-f]
- # [17:21] <Lachy> then replaces them with 0, so "bogus" becomes "b00000"
- # [17:21] <Lachy> and "crap" becomes "c0a000"
- # [17:22] <Lachy> I wonder if that's specced in HTML5...
- # [17:22] <Philip`> It work for "bogo" but not "bog" - it seems to only do that for >= 4 characters
- # [17:22] <Lachy> for the various color attributes
- # [17:23] <Dashiva> If there's a site relying on that, I will personally hunt down the author
- # [17:23] <Philip`> At least in FF3, it's quirks-only
- # [17:23] <SpookyET> Maybe, it should support non-us spelling:-)
- # [17:23] <Dashiva> What does 'beos' give you, Lachy?
- # [17:23] <SpookyET> background-colour
- # [17:24] <Lachy> Dashiva, "beos" gives "be0000"
- # [17:24] <Philip`> Maybe the browser developers were trying to be nice to authors who misread books with bad fonts, and tried writing <body bgcolor="ffooff">?
- # [17:24] <Lachy> SpookyET, doesn't work for CSS, only for the HTML attributes
- # [17:24] <Dashiva> Lachy: Hmm, in the bug I mentioned it became green
- # [17:25] <Philip`> (Quirks-only in Opera 9 too)
- # [17:25] <Lachy> Dashiva, which bug?
- # [17:25] <Lachy> and which browser?
- # [17:27] <Dashiva> Opera in the first stages of canvas support
- # [17:27] <Dashiva> Just after the "blue is red on windows" bug :)
- # [17:27] <Philip`> (Oh, it's doing bo=#b00000, bog=#0b0000, bogo=#b00000, and cra=#0c0a00 - I can see why it's called "quirks"...)
- # [17:28] <Lachy> Philip`, which browser are you getting those results?
- # [17:28] <Lachy> there shouldn't be a leading 0 on the values for bog and cra
- # [17:28] <Philip`> Oh, that was Opera 9
- # [17:29] <Philip`> but FF3 doesn't do that
- # [17:29] <Philip`> (Yay, non-interoperability)
- # [17:30] <Philip`> (FF3 just gives #000000 for anything <= 3 characters, it seems)
- # [17:30] <Dashiva> Time to run a bruteforce scan of all possible combinations on 1-6 letters, then!
- # [17:31] <Lachy> Dashiva, you only need to do 3 to 6, and letters A-F and one other
- # [17:31] * Quits: SpookyET (i=user@75.138.70.34) ("Client Exiting")
- # [17:32] <Lachy> e.g. bgcolor="azz", "bzz", ...
- # [17:32] <Philip`> Lachy: Not quite that simple - Opera ignores up to one leading #, and any spaces
- # [17:32] * Joins: briansuda (n=briansud@dsl-219-147.hive.is)
- # [17:34] <Philip`> In Opera, "f f f f f" == #fffff0; in FF, it's #f00ff0 (?!)
- # [17:34] <Lachy> oh, 1 and 2 letters give weird results too
- # [17:36] <Lachy> Safari does it in standards mode
- # [17:38] <Philip`> With <body bgcolor=" f f f "> I get grey in FF, almost-black in Opera, and green in Konqueror
- # [17:41] <Philip`> zcorpan: You said you wanted to spec quirks mode? ;-)
- # [17:41] <Lachy> nice, <body bgcolor="rainbow"> give a diff colour in nearly every browser :-)
- # [17:43] <Philip`> http://lxr.mozilla.org/mozilla/source/gfx/src/nsColor.cpp#126
- # [17:43] <Philip`> so it always does black for <= 3 chars (excluding the optional leading #)
- # [17:43] <Lachy> someone want to check webkit source, it seems to be closer to IE
- # [18:32] * Quits: virtuelv_ (n=arve@ti132110a341-1121.bb.online.no) ("Leaving")
- # [18:48] <zcorpan> Philip`: yes, test cases welcome
- # [18:53] * Joins: SpookyET (i=user@75.138.70.34)
- # [18:53] <SpookyET> I keep getting disconnected.
- # [18:55] <SpookyET> Has anyone seen Flashtml? :-) www.pierinc.com Thank you for killing my scrollbar and back button. But, for a second, it fools you into thinking that it's html and dhtml animations.
- # [18:58] * Quits: briansuda (n=briansud@dsl-219-147.hive.is)
- # [18:58] * zcorpan will use http://del.icio.us/zcorpan/quirks to pile up quirks
- # [19:18] * Quits: SpookyET (i=user@75.138.70.34) ("Client Exiting")
- # [19:20] * Joins: dbaron (n=dbaron@c-71-198-189-81.hsd1.ca.comcast.net)
- # [19:40] * Joins: csarven (n=nevrasc@modemcable081.152-201-24.mc.videotron.ca)
- # [19:43] * Quits: csarven (n=nevrasc@modemcable081.152-201-24.mc.videotron.ca) (Client Quit)
- # [19:43] * Joins: csarven (n=nevrasc@modemcable081.152-201-24.mc.videotron.ca)
- # [19:46] * Quits: tantek (n=tantek@adsl-63-195-114-133.dsl.snfc21.pacbell.net)
- # [19:47] * Joins: tantek (n=tantek@adsl-63-195-114-133.dsl.snfc21.pacbell.net)
- # [20:05] * Quits: tantek (n=tantek@adsl-63-195-114-133.dsl.snfc21.pacbell.net)
- # [20:36] <zcorpan> yay! ftp works again. i have *no* idea why it didn't work before
- # [20:44] * zcorpan considers moving valid-html5.png out of temp/
- # [20:51] <zcorpan> done
- # [21:21] * Joins: aroben (n=adamrobe@adsl-70-231-234-224.dsl.snfc21.sbcglobal.net)
- # [21:32] <zcorpan> started to draft on my presentation http://simon.html5.org/temp/html5-geekmeet.html
- # [21:32] <zcorpan> (in swedish)
- # [21:34] <Philip`> "Cross-document messanging" - spurious n?
- # [21:34] <zcorpan> oops
- # [21:35] <zcorpan> fixed
- # [21:41] * hsivonen learns that boilerplate is boilerplate in Swedish :-)
- # [21:42] <zcorpan> probably isn't
- # [21:42] <zcorpan> don't know a good swedish word for it :)
- # [21:42] <hsivonen> zcorpan: looks good
- # [21:43] <zcorpan> thanks. have two canvas demos iframed, commented out for now
- # [21:51] <Philip`> Be careful that the second doesn't suck up all your CPU while you're trying to get on with the rest of the presentation :-)
- # [21:52] <zcorpan> yeah, fortunately it can be paused :)
- # [21:52] <zcorpan> not sure i'll use those at all though
- # [21:53] <zcorpan> but having a "3d" fps game in 2d canvas is cool, so i want to include it
- # [21:54] <Philip`> Oh, can it?
- # [21:54] <Philip`> Aha, the "pause" key - that makes sense
- # [21:54] <zcorpan> there's also a pause button
- # [21:55] <Philip`> Oh, okay, on the 84 version since I didn't remove the debug bit at the bottom
- # [21:56] * Quits: aroben (n=adamrobe@adsl-70-231-234-224.dsl.snfc21.sbcglobal.net)
- # [21:57] <Philip`> (The 2d canvas is quite nasty for doing this 3d stuff in, though :-p )
- # [21:58] <Philip`> (It's nice that you can do it, but it isn't really playing to the strengths of the canvas)
- # [21:59] <zcorpan> indeed. it's like using GIF for video
- # [21:59] <Philip`> ((It is quite fun, though))
- # [22:01] <Philip`> Perhaps it's more like using ASCII art over Telnet for video
- # [22:02] <zcorpan> heh, ok
- # [22:04] <Philip`> (That reminds me, I did a (rubbish) text-based multiplayer deathmatch FPS game over the IRC protocol (though not a real IRC server) once - it would be nice to integrate that with the canvas code and get browser-based multiplayer deathmatch...)
- # [22:15] * Joins: tantek (n=tantek@c-71-198-74-90.hsd1.ca.comcast.net)
- # [22:43] <Philip`> (Hixie: I think the spec has a spare > in Scripting/Origin with "<span>browsing context</span>>")
- # [22:43] * Joins: aroben (n=adamrobe@adsl-70-231-234-224.dsl.snfc21.sbcglobal.net)
- # [22:45] * Joins: othermaciej (n=mjs@dsl081-048-145.sfo1.dsl.speakeasy.net)
- # [22:47] * Quits: aroben (n=adamrobe@adsl-70-231-234-224.dsl.snfc21.sbcglobal.net) (Client Quit)
- # [22:56] * Quits: ajnewbold (n=fax_mach@unaffiliated/chuangtzu)
- # [23:03] * Quits: ROBOd (n=robod@86.34.246.154) ("http://www.robodesign.ro")
- # [23:23] * Quits: met_ (n=Hassman@r5bx220.net.upc.cz) ("Leaving")
- # [23:29] * Quits: dbaron (n=dbaron@c-71-198-189-81.hsd1.ca.comcast.net) ("8403864 bytes have been tenured, next gc will be global.")
- # [23:48] * Quits: tantek (n=tantek@c-71-198-74-90.hsd1.ca.comcast.net)
- # Session Close: Mon Apr 23 00:00:00 2007
The end :)