/irc-logs / freenode / #whatwg / 2009-02-12 / end

Options:

# Session Start: Thu Feb 12 00:00:00 2009
# Session Ident: #whatwg
# [00:00] * Quits: billmason (n=bmason@ip102.unival.com) ("Leaving.")
# [00:07] <Hixie> sicking: http://software.hixie.ch/utilities/js/live-dom-viewer/saved/5
# [00:08] <Hixie> i guess we could remove the check
# [00:08] <Hixie> and just always clone
# [00:08] * sicking is still looking at example
# [00:09] <sicking> Hixie, i need to look at that in henris HTML5 dom-live-viewer right?
# [00:09] <Hixie> wait, hsivonen was asking for a flag for whether it had parser-inserted children
# [00:09] <Hixie> sicking: or webkit
# [00:10] <Hixie> sicking: the AAA in HTML5 is the same as webkit's residual style handling (so long as you only have one block)
# [00:10] <Hixie> hsivonen: when would that algorithm ever _not_ have parser-inserted children?
# [00:10] <sicking> Hixie, "the AAA"?
# [00:10] <Hixie> Adoption Agency Algorithm. residual style in html5.
# [00:10] <sicking> ah
# [00:12] * Quits: mstange (n=markus@aixd3.rhrk.uni-kl.de) ("ChatZilla 0.9.84-2009010213 [Firefox 3.2a1pre/20090210022740]")
# [00:13] <Hixie> (from the spec: "Because of the way this algorithm causes elements to change parents, it has been dubbed the "adoption agency algorithm" (in contrast with other possibly algorithms for dealing with misnested content, which included the "incest algorithm", the "secret affair algorithm", and the "Heisenberg algorithm").")
# [00:13] <Hixie> heisenberg's = mozilla, secret affair = opera, incest = ie
# [00:13] <sicking> yeah, i know what the algorithm does
# [00:13] <sicking> what still confuses me is when it makes a difference which of the two forks it uses
# [00:14] <sicking> i.e. the has-children and don't-have-children forks
# [00:14] <Hixie> i'm actually unsure as to when you can ever not have children without scripting
# [00:14] <Hixie> and why we care about that case
# [00:14] <Hixie> i'm thinking maybe we can remove the if statement altogether
# [00:14] <sicking> that would be nice :)
# [00:15] <Hixie> i hate this algorithm, i only barely understood it when i was writing it and implementing it
# [00:15] <sicking> since i won't have to challange you to a fight of doom over the feedback
# [00:15] <Hixie> hah
# [00:16] <Hixie> i'm gonna remove the if statement and see if anyone complains.
# [00:16] <sicking> haha
# [00:16] <sicking> sold!
# [00:16] <sicking> i'm still not sure how to detect a difference between the two forks
# [00:17] <sicking> but i don't care as long as this thing goes :)
# [00:17] <Hixie> oh, wait...
# [00:17] * Hixie studies further
# [00:18] <sicking> Hixie, in your example, if i remove the script, the DOM still looks the same
# [00:19] <Hixie> yeah i was wrong
# [00:19] <Hixie> here:
# [00:19] <Hixie> it's the difference between:
# [00:19] <Hixie> <!DOCTYPE html><i><em><div></i>xxxx
# [00:19] <Hixie> <!DOCTYPE html><i><em>xxxx<div></i>xxxx
# [00:19] * ojan|eseidel is now known as ojan
# [00:19] <Hixie> in the former case, html5 just moves the original <em> down
# [00:19] <Hixie> in the latter case, it clones it
# [00:20] <Hixie> (the <div> doesn't count as being a child of the <em> because it's removed earlier in the algorithm)
# [00:20] <sicking> so we're trying to avoid creating the extra <em>?
# [00:20] <Hixie> yeah
# [00:21] * sicking start to warm up the fight-of-doom arena
# [00:21] <Hixie> it becomes a big deal when you have a lot of missing end tags (e.g. a lot of <font> with no </font>) iirc
# [00:21] <sicking> right
# [00:21] <Hixie> the problem with keeping track of whether the parser inserted any nodes into each node in the stack is that that's actually not enough
# [00:21] <Hixie> you have to keep track of how many children
# [00:21] <sicking> why is there always an 'i' outside the <div> though?
# [00:22] <Hixie> (0 or 1 or 2)
# [00:22] <Hixie> 2+ rather
# [00:22] <Hixie> the element being closed is always cloned because not doing so would make the algorithm orders of magnitude more complex still
# [00:23] <Hixie> iirc
# [00:23] <sicking> ok
# [00:23] <sicking> that's probably why i failed to figure this out last time
# [00:24] <Hixie> yeah this is definitely one of the most obscure parts of the spec
# [00:24] <Hixie> easily on par with xbl2 in terms of headaches
# [00:24] <sicking> actually, that doesn't seem right either
# [00:24] <sicking> <!DOCTYPE html><i><em><em><em><div></i>xxxx
# [00:24] <sicking> that clones all but the inner most <em>
# [00:24] <sicking> so if you have lots of opened <font>s, you'll still clone all but the last it looks like
# [00:24] * Quits: heycam (n=cam@clm-laptop.infotech.monash.edu.au) (Read error: 104 (Connection reset by peer))
# [00:24] <sicking> unless henri has a bug
# [00:25] * Joins: heycam (n=cam@clm-laptop.infotech.monash.edu.au)
# [00:27] * Quits: eric_carlson (n=ericc@adsl-67-112-12-110.dsl.anhm01.pacbell.net)
# [00:27] <sicking> Hixie, and safari seems to clone them all
# [00:27] <sicking> though maybe we're trying to do better than safari?
# [00:30] <Hixie> hm, maybe it's not an optimisation
# [00:30] <Hixie> yeah maybe we should simplify this by not removing the furthest block node from its parent and then just not checking for children
# [00:30] <Hixie> and always cloning
# [00:31] * Joins: weinig_ (n=weinig@17.203.15.158)
# [00:31] <sicking> Hixie, my first question is what we're trying to accomplish with the check :)
# [00:32] <sicking> Hixie, optimizting away empty element i suspect is actually worth it
# [00:32] <Hixie> yeah, but as you point out, we're not doing it enough
# [00:32] <Hixie> if we want to do it we should do a proper job
# [00:32] <Hixie> which is a whole other problem
# [00:32] <sicking> yup
# [00:32] <sicking> Hixie, i do know that we (=gecko) suffer, or at least used to suffer, on pages with lots of opened-but-not-closed <font>s
# [00:33] <Hixie> yeah
# [00:33] <sicking> but we might never have considered it common enough to fix
# [00:33] <Hixie> ok i've changed it to not do the check
# [00:34] <Hixie> but i think we might want to make a more general fix at one point
# [00:34] <Hixie> to get rid of the empty nodes
# [00:34] <sicking> agreed
# [00:35] <sicking> at which point i'll book the DOOM conference room so we can discuss how :)
# [00:36] <sicking> Hixie, try "<!DOCTYPE html><i><font><font><font><font><div>w</div><div>f</div>" in gecko now
# [00:36] <sicking> Hixie, we create loads of elements
# [00:36] <sicking> Hixie, way more than HTML5 does
# [00:37] * Quits: jruderman (n=jruderma@corp-246.mountainview.mozilla.com)
# [00:37] * Joins: jruderman (n=jruderma@corp-246.mountainview.mozilla.com)
# [00:38] <Hixie> yeah, html5 doesn't move all the nodes to inside the blocks
# [00:38] <Hixie> that isn't affected by the change just now afaict
# [00:39] <sicking> funny thing is, if you just close one <font> after the last div, we dramatically change what we do
# [00:39] <sicking> which is not what i would have guessed
# [00:39] <sicking> right, html5 deals with this stuff dramatically different
# [00:40] <Hixie> the biggest problem with gecko is that what it does differs based on packet boundaries
# [00:41] <sicking> right
# [00:41] <sicking> the lookahead is bad
# [00:41] <Dashiva> "With namespaces, users pay while spec authors get off scott free for their inability to do the hard work of writing specs and forging agreement."
# [00:41] <sicking> i'm not at all suggesting we do what gecko does
# [00:42] <sicking> just saying that dealing with lots of unclosed <font>s might not be critically important to do
# [00:42] <sicking> the packed boundry thing could be fixed in gecko, and still keep the lookahead
# [00:43] <Philip`> Dashiva: But with distributed extensibility, *everyone* is a spec author
# [00:43] <sicking> but we'd still produce much more bloated DOMs than the current spec
# [00:44] <sicking> The only real advantage in what we do is that we can implement a real SAX API even for malformed documents
# [00:45] <Hixie> sicking: yeah, the spec might be good enough at this point
# [00:46] <Hixie> sicking: well, you still can't do a real SAX API for <table>s, right?
# [00:46] <Hixie> or for <html> and <body> nodes
# [00:47] <Hixie> ok bbiab
# [00:48] * Quits: weinig (n=weinig@17.244.17.181) (Read error: 110 (Connection timed out))
# [00:49] <sicking> Hixie, ah, yeah, you're right
# [00:53] <Lachy> Hixie, re the :enabled and :disabled issue, are you now going to make them match hidden inputs based on the changes to Selectors?
# [00:58] * Joins: tantek (n=tantek@c-98-207-116-210.hsd1.ca.comcast.net)
# [01:05] * Joins: starjive (i=beos@213-66-216-93-no30.tbcn.telia.com)
# [01:40] * Quits: millam (n=millam@72.14.227.1)
# [01:44] <Hixie> Lachy: i'll make them do whatever the new selectors spec says
# [01:45] <Hixie> selectors is a decade old now
# [01:45] <Hixie> a decade!
# [01:45] <Hixie> and that's a 32 page spec
# [01:45] <Hixie> it's about as big as the rendering section in html5
# [01:46] <Hixie> and people think the timeline for html5 is unrealistic?
# [01:46] <Hixie> sigh
# [01:46] * Quits: dglazkov (n=dglazkov@nat/google/x-47fea026b48cad31)
# [01:49] <Hixie> annevk5: i actually haven't defined <marquee> yet. :-)
# [01:57] <Lachy> Hixie, I think some people have finally come to realise that the HTML5 timeline is realistic, given it's current size. I think that's one of the many reasons people want to split the spec into smaller specs.
# [01:58] <Lachy> (although CSS3 has been split into separate specs and how many of those have reached rec in the past 10 years?)
# [01:58] <Dashiva> Wasn't CSS kind of a trainwreck already?
# [01:58] <Lachy> I think Selectors and CSS Namespaces are the closest to reaching REC
# [01:58] <Hixie> Lachy: those people don't seem to understand that it doesn't matter how big the spec is, it doesn't make it go any faster if it's smaller
# [01:59] <Hixie> Lachy: (given the interactions between sections)
# [01:59] <Lachy> indeed
# [02:21] * Quits: tantek (n=tantek@c-98-207-116-210.hsd1.ca.comcast.net)
# [02:26] * Joins: eric_carlson (n=ericc@adsl-67-112-12-110.dsl.anhm01.pacbell.net)
# [02:28] * Quits: ojan (n=ojan@72.14.229.81)
# [02:53] * Quits: Amorphous (i=jan@unaffiliated/amorphous) (Read error: 113 (No route to host))
# [02:55] * Joins: Amorphous (i=jan@unaffiliated/amorphous)
# [02:59] <Hixie> jgraham: yt?
# [03:10] <Dashiva> Say, did anything in particular happen that made you stop twittering, Hixie?
# [03:10] * Quits: eric_carlson (n=ericc@adsl-67-112-12-110.dsl.anhm01.pacbell.net)
# [03:11] <Hixie> i stopped twittering?
# [03:11] <Dashiva> There was almost no traffic most of the fall and winter
# [03:11] * Joins: eric_carlson (n=ericc@adsl-67-112-12-110.dsl.anhm01.pacbell.net)
# [03:11] <Hixie> i twittered 15 times in the last 24 hours
# [03:12] <Hixie> what more do you want from me!
# [03:12] <Hixie> :-P
# [03:12] * Quits: slightlyoff (n=slightly@nat/google/x-5f7434e0d11057f2) (Read error: 110 (Connection timed out))
# [03:12] <Dashiva> Maybe I'm looking at the wrong account then!
# [03:12] <Hixie> 32 times if you count the retweets that mike puts out
# [03:12] <Hixie> http://twitter.com/WHATWG
# [03:13] <Dashiva> Right
# [03:13] <Hixie> and http://twitter.com/HTML5
# [03:13] <Hixie> the latter also includes editorial checkins
# [03:13] <Dashiva> Never mind me then
# [03:13] <Hixie> on my Hixie account I don't do much because, well, twitter is dumb :-P
# [03:13] * Joins: doodlewarrior (n=anon-irc@adsl-76-254-59-237.dsl.pltn13.sbcglobal.net)
# [03:14] <doodlewarrior> i'm just starting to peek at the canvas element
# [03:14] <Hixie> it makes sense as a short message broadcast service
# [03:14] <Hixie> but microblogging is just stupid imho
# [03:14] <Dashiva> I guess the clickjacking comment made me think you had all the techie stuff on Hixie
# [03:14] <doodlewarrior> i've heard a lot of people talk about it for charting, but i'm not sure i understand why you'd use it
# [03:14] <Hixie> though i'd use it a lot if someone did the equivalent of bitlbee for twitter
# [03:14] <Hixie> doodlewarrior: http://www.whatwg.org/issues/data.html is an example of <canvas> use for a chart
# [03:14] <doodlewarrior> wouldn't you use positioned divs
# [03:14] <doodlewarrior> thanks ill look
# [03:15] <Philip`> doodlewarrior: Mostly people seem to use it for charts because they're too lazy to write a server-side script that generates an image
# [03:15] <Hixie> right
# [03:15] <Hixie> it grabs the data using XHR and then uses JS to generate the image
# [03:15] <doodlewarrior> thanks guys
# [03:15] <Hixie> Dashiva: i use Hixie for replying to other people
# [03:16] <Hixie> Dashiva: (and for sending SMSes to my partner, since i don't have a phone and they do)
# [03:16] <doodlewarrior> it's amusing to me that theres an HTML element that is literally defined as 'width, height, id' and relies on javascript for everything else
# [03:16] <Hixie> doodlewarrior: there's several of those in html5
# [03:16] <Hixie> doodlewarrior: <eventsource>, <video>, <audio> also
# [03:17] <doodlewarrior> the media ones have a src though
# [03:17] <Hixie> <script> too :-)
# [03:17] <doodlewarrior> thyre just like image
# [03:17] <Hixie> true
# [03:17] <Hixie> true, you can make them show controls
# [03:17] <Hixie> <eventsource> is js-only though
# [03:17] <Hixie> basically
# [03:17] <Dashiva> I seem to recall a request for using an image to "seed" a canvas
# [03:17] <doodlewarrior> i havent heard of eventsource
# [03:17] <doodlewarrior> ill look into it
# [03:18] <doodlewarrior> btw, that issues graph is a better canvas sample than the ones ive seen so far
# [03:18] <Dashiva> But if you're going to have to use js anyway, there's little point
# [03:18] <doodlewarrior> most of those have been bar charts
# [03:18] <Philip`> doodlewarrior: Have you seen http://www.liquidx.net/plotkit/ ?
# [03:18] <Hixie> doodlewarrior: it's not a canvas sample, i actually use it to track feedback :-)
# [03:18] <Hixie> doodlewarrior: it just happens to use canvas :-)
# [03:19] <Philip`> doodlewarrior: (Examples at http://media.liquidx.net/js/plotkit-tests/sweet.html etc)
# [03:20] <doodlewarrior> i had been looking at flotr, which is based on plotkit
# [03:21] <doodlewarrior> maybe for pie or line graphs
# [03:21] <doodlewarrior> but for bar charts, which is what a lot of analytics turns into, CSS seems to be a better way to go
# [03:22] * Joins: slightlyoff (n=slightly@67.218.110.60)
# [03:22] <Hixie> well don't forget that CSS shouldn't be used for content
# [03:22] <Hixie> your page shouldn't change meaning if you turn off the CSS
# [03:22] <Philip`> That seems a pretty pointless argument in practice
# [03:22] <Hixie> (similarly, if you use <canvas>, you should also expose the content inside the element for users that don't see images)
# [03:22] <Philip`> If you're in an environment where you can't see CSS, you almost certainly won't be able to see canvas either
# [03:23] * Joins: tantek (n=tantek@c-98-207-116-210.hsd1.ca.comcast.net)
# [03:23] <doodlewarrior> Philip`s right
# [03:23] <Hixie> the actual correct way of doing bar charts in HTML5 is <meter>
# [03:23] <Hixie> but that's not supported anywhere yet
# [03:23] <doodlewarrior> in that case, it should all be does as images
# [03:23] <doodlewarrior> because both CSS and JS are not *guaranteed*
# [03:23] <doodlewarrior> although in practice theyre pretty ubiq
# [03:23] <Philip`> Images aren't guaranteed either
# [03:24] <Hixie> JS is more guaranteed than CSS or images
# [03:24] <Hixie> (JS is not optional, unlike CSS or images)
# [03:24] <Philip`> I think you should use a Java applet
# [03:24] <doodlewarrior> im a flash guy. id do that, but mobiles dont support it
# [03:24] <doodlewarrior> and id choose flash over java any day of the week
# [03:25] <Philip`> Write an applet that displays a bar, and then include it lots of times with <applet width="..."> for each bar
# [03:25] <doodlewarrior> the only browser ive seen where JS is not optional is chrome
# [03:25] <doodlewarrior> hahahaha
# [03:25] <doodlewarrior> nice
# [03:25] * Quits: tantek (n=tantek@c-98-207-116-210.hsd1.ca.comcast.net) (Read error: 113 (No route to host))
# [03:25] * Joins: tantek (n=tantek@c-98-207-116-210.hsd1.ca.comcast.net)
# [03:26] <Hixie> doodlewarrior: by "optional" i mean "the specs say that you can turn that off without changing the meaning of the page"
# [03:26] <Hixie> turning script off will change the meaning of the page
# [03:27] <doodlewarrior> you can turn off images without changing the meaning?
# [03:27] <doodlewarrior> deviantart is boned
# [03:27] <doodlewarrior> flickr, facebook, et al
# [03:27] <doodlewarrior> youtube
# [03:27] <doodlewarrior> :-p
# [03:27] <Philip`> That depends on the meaning of "meaning"
# [03:28] <Hixie> if those sites don't include the equivalent of the image in the alt="" attribute, they're non-conforming :-)
# [03:28] <doodlewarrior> there are definitely things that break without javascript
# [03:28] <doodlewarrior> the app that has me thinking of all this is one of them
# [03:29] <doodlewarrior> but i would still argue that images are more crucial than javascript
# [03:29] <Philip`> Hixie: Non-conforming HTML pages? That would be unthinkable :-(
# [03:29] <Hixie> hah
# [03:29] <doodlewarrior> (and probably more likely to be supported)
# [03:29] <Hixie> doodlewarrior: a blind user will have JS support but the images won't do him any good
# [03:30] <doodlewarrior> i didnt know text-based clients supported javascript
# [03:31] <Philip`> doodlewarrior: They tend to use a proper graphical browser plus a screenreader, rather than a text browser, apparently
# [03:32] <doodlewarrior> my mac would do that when i was a kid
# [03:32] <doodlewarrior> mouse over something and it would tell you what it is
# [03:33] <Philip`> Aiming the mouse at objects is kind of hard when you can't see the cursor or the objects
# [03:35] <doodlewarrior> i would think so
# [03:36] <doodlewarrior> i imagine it would be easier to surf with firefox + reader than lynx
# [03:36] <doodlewarrior> if for no other reason than all the sites that sniff by browser
# [03:37] <doodlewarrior> Hixie: this is what i'll end up modeling
# [03:37] <doodlewarrior> http://peltiertech.com/WordPress/wp-content/img200805/stack_bar_graded.png
# [03:37] <doodlewarrior> i'm not sure if meter would be right for that
# [03:37] <Hixie> no, <meter> wouldn't really work for a stacked bar
# [03:38] <Hixie> <meter> is just for a single-value gauge
# [03:38] <doodlewarrior> YEAH
# [03:38] <doodlewarrior> bad caps
# [03:39] <doodlewarrior> ill probably end up using HTML + CSS
# [03:40] <doodlewarrior> it's easier to template server-side
# [03:40] <doodlewarrior> and you don't need to worry about rendering the JS
# [03:40] <Philip`> Text in <canvas> is not well supported in current browsers, so it'd probably be easier to just use HTML/CSS
# [03:41] <doodlewarrior> thanks guys
# [03:41] * Dashiva remembers making bitmap fonts for canvas text
# [03:41] <Dashiva> Bad times
# [03:42] <doodlewarrior> i must admit, im not much for IRC
# [03:42] * Quits: slightlyoff (n=slightly@67.218.110.60)
# [03:42] <doodlewarrior> what's the syntax to do a third person message
# [03:42] <Philip`> /me ...
# [03:43] * doodlewarrior is testing this
# [03:43] * doodlewarrior is wondering how Philip` escaped the /me
# [03:43] <doodlewarrior> /me like this?
# [03:44] <Dashiva> Many ways, depending on the client
# [03:44] <doodlewarrior> anyway, ill stop wasting all your time now
# [03:44] <Dashiva> e.g. mIRC supports ctrl-enter to avoid commands
# [03:44] <doodlewarrior> im in opera
# [03:45] <Philip`> / /me ... escapes it in irssi
# [03:45] <doodlewarrior> i dont really want to install a chat client
# [03:45] <doodlewarrior> although for all practical purposes all opera is to me IS a chat client
# [03:45] <doodlewarrior> although it was my main browser before safari was cross-platform
# [03:45] <doodlewarrior> although i say although a lot
# [03:45] <Dashiva> When all else fails, /msg #channelname /me text should work
# [03:46] <doodlewarrior> gtk
# [03:49] * Quits: dave_levin (n=dave_lev@72.14.227.1)
# [03:56] <Hixie> wow have you guys seen this? http://a.deveria.com/caniuse/
# [03:58] <Dashiva> It's a bit depressing to see features blocking on browser versions that aren't even the next release...
# [03:59] <Hixie> you can uncheck the "Past" row
# [03:59] <Hixie> :-)
# [03:59] <Hixie> bbiab
# [04:01] <deltab> /say usually works too
# [04:03] * Quits: olliej (n=oliver@nat/apple/x-dfce7e84c5fc570d)
# [04:09] * Parts: doodlewarrior (n=anon-irc@adsl-76-254-59-237.dsl.pltn13.sbcglobal.net)
# [04:09] * Lachy is unsure whether he should be happy or sad that <form autocomplete> has now been added to the spec as a result of my data, despite the fact that I think it should never be used by anyone.
# [04:10] <Dashiva> Honor and valor
# [04:11] * Quits: tantek (n=tantek@c-98-207-116-210.hsd1.ca.comcast.net)
# [04:11] <Dashiva> It would probably have been added eventually, in any case.
# [04:30] * Joins: olliej (n=oliver@nat/apple/x-faf3a9e2a43540f3)
# [04:36] <roc> Hixie: that's cool
# [04:36] <roc> someone should tell them that multiple backgrounds is about to land for Firefox
# [04:36] <roc> 3.2 though
# [04:39] <roc> I sort of hate how people assume that Webkit CSS extensions are the future
# [04:39] <roc> what about MY CSS extensions
# [04:40] <roc> hmm, IE8 doesn't support HTML data URLs? FAIL
# [04:41] * Joins: Dashimon (i=Dashiva@46.80-202-223.nextgentel.com)
# [04:41] * Quits: Dashiva (i=Dashiva@wikia/Dashiva) (Read error: 60 (Operation timed out))
# [04:41] * Dashimon is now known as Dashiva
# [04:42] <roc> Ah, I just want to uncheck "Unofficial" and then I'm happy
# [04:43] <Dashiva> I wouldn't mind a "hide stuff that isn't ready" checkbox, just to see what little is actually usable
# [04:46] <olliej> roc: really?
# [04:46] <roc> which part?
# [04:46] <olliej> roc: ie8 w/no data uri support
# [04:46] <roc> that page says it supports data URIs for image and stylesheet loads, but not HTML loads
# [04:46] <roc> or something like that
# [04:47] <olliej> roc: url?
# [04:47] <roc> of course IE also has a ridiculously low URI length limit which cripples data URIs anyway
# [04:47] <olliej> true
# [04:47] <roc> it does give the impression they might have done the minimum needed to pass Acid2
# [04:47] <olliej> roc: it's just a thousand or so character or something iirc
# [04:48] <roc> yeah
# [04:48] <roc> olliej: http://a.deveria.com/caniuse/
# [04:48] <roc> (via Hixie in this channel)
# [04:48] <olliej> roc: at least it's better than the minimum needed to pass some of the acid3 svg tests -- you just needed the element constructors to exist
# [04:48] <roc> yeah well
# [04:48] <olliej> roc: they didn't actually need to work
# [04:48] <olliej> i was like wtf?
# [04:49] <roc> let's not discuss the deficiencies of Acid3 :-)
# [04:49] <olliej> roc: wow, i love the idea that Chrome 0.3 is the old chrome, but safari 2 is the old safari
# [04:49] <roc> yeah
# [04:49] <olliej> ignoring safari 3.0, and 3.1
# [04:50] <roc> I also love the idea that IE9 and Firefox 3.2 will be contemporaneous
# [04:51] <olliej> heh
# [04:51] * Quits: dolske (n=dolske@firefox/developer/dolske)
# [04:52] <olliej> roc: what makes the chrome 0.3 vs. Safari2 comparison especially hilarious is that chrome's initial release was based on Safari 3.1
# [04:52] <olliej> wee!
# [04:52] <roc> yeah I know
# [04:53] <roc> it's iniquitous
# [04:54] <olliej> oh well
# [04:54] * Joins: eric_carlson_ (n=ericc@adsl-67-112-12-110.dsl.anhm01.pacbell.net)
# [04:54] <olliej> shit happens
# [04:54] <olliej> hihi eric_carlson
# [04:54] <olliej> or eric_carlson_ as the case may be
# [04:54] <eric_carlson_> hey olliej
# [05:04] * Quits: eric_carlson (n=ericc@adsl-67-112-12-110.dsl.anhm01.pacbell.net) (Read error: 110 (Connection timed out))
# [05:07] * Joins: dglazkov (n=dglazkov@c-98-207-88-44.hsd1.ca.comcast.net)
# [05:08] * Quits: dimich (n=dimich@72.14.227.1)
# [05:08] * Joins: eric_carlson (n=ericc@adsl-67-112-12-110.dsl.anhm01.pacbell.net)
# [05:09] * Quits: weinig_ (n=weinig@17.203.15.158)
# [05:11] * Quits: dbaron (n=dbaron@corp-246.mountainview.mozilla.com) ("8403864 bytes have been tenured, next gc will be global.")
# [05:12] * Joins: dolske (n=dolske@c-76-103-40-203.hsd1.ca.comcast.net)
# [05:19] * Quits: eric_carlson_ (n=ericc@adsl-67-112-12-110.dsl.anhm01.pacbell.net) (Read error: 110 (Connection timed out))
# [05:29] * Joins: eric_carlson_ (n=ericc@adsl-67-112-12-110.dsl.anhm01.pacbell.net)
# [05:35] * Joins: dbaron (n=dbaron@c-98-234-51-190.hsd1.ca.comcast.net)
# [05:38] * Quits: eric_carlson (n=ericc@adsl-67-112-12-110.dsl.anhm01.pacbell.net) (Read error: 110 (Connection timed out))
# [05:48] * Joins: zcorpan (n=zcorpan@pat.se.opera.com)
# [05:48] <zcorpan> morning
# [05:52] * Quits: starjive (i=beos@213-66-216-93-no30.tbcn.telia.com)
# [05:56] * Joins: dave_levin (n=dave_lev@c-98-203-247-78.hsd1.wa.comcast.net)
# [06:08] * Quits: roc (n=roc@202.0.36.64)
# [06:15] * Quits: dglazkov (n=dglazkov@c-98-207-88-44.hsd1.ca.comcast.net)
# [06:24] * Quits: aroben (n=aroben@unaffiliated/aroben) (Read error: 54 (Connection reset by peer))
# [06:26] * Quits: jwalden (n=waldo@corp-246.mountainview.mozilla.com) ("->apt")
# [06:32] * Quits: doublec (n=chris@202.0.36.64) ("Leaving")
# [06:38] * Joins: dimich (n=dimich@c-98-203-230-54.hsd1.wa.comcast.net)
# [06:48] <zcorpan> http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2009-February/018492.html - what should dom core say?
# [06:50] <zcorpan> right now i have
# [06:50] <zcorpan> A DOMTimeStamp represents a number of milliseconds.
# [06:50] <zcorpan> typedef unsigned long long DOMTimeStamp;
# [06:51] <heycam> zcorpan, i think that is ok
# [06:51] <zcorpan> heycam: ok, thanks
# [06:52] <heycam> the difference in precision between unsigned long longs (which range from 0..2**64-1) and ES Numbers (which can represent integers up to 2**53 - 1) isn't important, iirc.
# [06:52] <heycam> at least for practical uses of time stamps
# [06:52] <zcorpan> hmm maybe the definition should say "...since Unix epoch"
# [06:53] <heycam> yeah. i think dom core says something like not specifying a particular epoch, or that time stamps might always be 0 or something.
# [06:53] <heycam> can't remember exactly
# [06:54] <zcorpan> "The DOMTimeStamp type is used to store an absolute or relative time."
# [06:54] <zcorpan> when is it used for a relative time?
# [06:54] <heycam> don't know of any place in the dom off the top of my head
# [06:55] <zcorpan> dom3 core says "For ECMAScript, DOMTimeStamp is bound to the Date type because the range of the integer type is too small."
# [06:55] * zcorpan would like to refer to webidl for ecmascript bindings of dom core :)
# [06:56] <heycam> please feel free :)
# [06:57] <zcorpan> maybe i should put a note in there saying that it's not a Date in ecmascript
# [06:57] <heycam> Dates in ES, iirc, use a Number (i.e. a double) internally
# [06:57] <heycam> so i think the reasoning is bogus
# [06:58] * Joins: zdobersek (n=zan@cpe-92-37-78-186.dynamic.amis.net)
# [06:59] <zcorpan> oh <time> in html5 has relative domtimestamps
# [07:00] * Quits: Dashiva (i=Dashiva@wikia/Dashiva) (Read error: 60 (Operation timed out))
# [07:02] * Joins: Dashiva (i=Dashiva@wikia/Dashiva)
# [07:09] * Joins: zdobersek1 (n=zan@cpe-92-37-76-56.dynamic.amis.net)
# [07:12] * Joins: jwalden_ (n=waldo@c-76-102-160-171.hsd1.ca.comcast.net)
# [07:12] * jwalden_ is now known as jwalden
# [07:12] <zcorpan> Hixie: couldn't html5lib prefix mathml elements with an "M" and svg elements with an "S" in the no-namespace mode, or something?
# [07:18] <Hixie> sure
# [07:19] <Hixie> does it?
# [07:24] <heycam> hmm "layed out" or "laid out"?
# [07:24] <heycam> or "layoutted"? :)
# [07:24] <Hixie> laid
# [07:24] <heycam> surprisingly, google searches for both those terms bring up appropriately the same number of hits
# [07:24] <heycam> ok
# [07:25] <Hixie> except all the hits for "layed" are "is it laid or layed"
# [07:25] <heycam> heh
# [07:26] * Quits: zdobersek (n=zan@cpe-92-37-78-186.dynamic.amis.net) (Read error: 113 (No route to host))
# [07:28] * Quits: zdobersek1 (n=zan@cpe-92-37-76-56.dynamic.amis.net) (Remote closed the connection)
# [07:44] * Quits: dimich (n=dimich@c-98-203-230-54.hsd1.wa.comcast.net)
# [07:47] * Quits: heycam (n=cam@clm-laptop.infotech.monash.edu.au) ("bye")
# [07:49] * Joins: maikmerten (n=merten@ls5dhcp196.cs.uni-dortmund.de)
# [07:53] <zcorpan> Philip`: in XOM, try two attributes with the same namespace and local name but different prefix
# [07:53] <zcorpan> or maybe XOM doesn't let you manage prefixes?
# [08:01] * Joins: ap (n=ap@194.154.88.32)
# [08:04] <hsivonen> sicking: pong
# [08:06] <hsivonen> Philip`: is <?foo ??> well-formed? If it inserts a space unnecessarily, it is probably a bug in the code that tries to avoid the ?> substring
# [08:06] <Hixie> sicking: so do you think mozilla wants the spellcheck="" attribute in html5?
# [08:06] <Hixie> anyone from opera have any opinions on that, also?
# [08:06] <Hixie> ap: how about webkit, do you know if anyone outside google wants spellcheck="" in html5?
# [08:07] <Hixie> i figure i'll add it, more or less as specced here http://damowmow.com/playground/spellcheck.txt
# [08:07] <Hixie> since there seems to be vague positive notions about it
# [08:08] <Hixie> (and it's used by major sites and firefox and chrome both have some level of support)
# [08:08] <Hixie> (though that's all mostly from google)
# [08:09] * Joins: weinig (n=weinig@c-67-180-35-124.hsd1.ca.comcast.net)
# [08:09] <ap> Hixie: https://bugs.webkit.org/show_bug.cgi?id=14552 is not from Google afaict
# [08:09] <zcorpan> hsivonen: <?foo ??> seems to be well-formed
# [08:10] <Hixie> ap: woo, non-google interest
# [08:10] <zcorpan> hsivonen: but <?xml-stylesheet ??> isn't xml+xml-stylesheet-well-formed (if there is such a concept)
# [08:14] <hsivonen> zcorpan: ok.
# [08:16] <zcorpan> speaking of xml-stylesheet, we had a telecon yesterday
# [08:17] <zcorpan> there was push-back in general but they seemed to agree that error handling needed to be defined
# [08:19] * Quits: sverrej (n=sverrej@122.160.12.230) (Read error: 110 (Connection timed out))
# [08:21] <zcorpan> surprising to me was that the xml core wg were apparently unfamiliar with the term "yellow screen of death"
# [08:22] * Quits: virtuelv (n=virtuelv@95.34.27.22.customer.cdi.no) ("Leaving")
# [08:25] <hsivonen> whoa. pretty removed from Web use cases, eh?
# [08:26] <zcorpan> i guess
# [08:29] * Joins: pesla (n=retep@procurios.xs4all.nl)
# [08:32] <hsivonen> Philip`: I think I've now fixed the serializer bugs you found (in svn only so far)
# [08:32] <hsivonen> Philip`: thanks
# [08:41] * Quits: dbaron (n=dbaron@c-98-234-51-190.hsd1.ca.comcast.net) ("8403864 bytes have been tenured, next gc will be global.")
# [08:47] <hsivonen> Philip`: why is <x xmlns="x:&" /> not well-formed?
# [08:47] * Joins: Maurice (n=ano@a80-101-46-164.adsl.xs4all.nl)
# [08:50] * Joins: zalan (n=kvirc@dsl5400EDD3.pool.t-online.hu)
# [08:52] * Quits: nessy (n=nessy@203-158-43-142.dyn.iinet.net.au) ("This computer has gone to sleep")
# [09:00] <Hixie> hsivonen: unescaped ampersand is not valid in AttValue
# [09:00] <hsivonen> ooh
# [09:01] <hsivonen> so obvious now...
# [09:01] <hsivonen> the thread with Adam Barth & others on www-talk is interesting
# [09:24] * Joins: heycam (n=cam@210-84-15-128.dyn.iinet.net.au)
# [09:38] <Hixie> site-meta scares me
# [09:38] <Hixie> i don't really understand the problem it's trying to solve
# [09:43] <hsivonen> Hixie: what kind of parser-level quirks are expected? Is the fragment parsing algorithm meant to be quirk-sensitive?
# [09:44] <hsivonen> <p><table>, I expect
# [09:44] <hsivonen> (and that's Acid2's fault)
# [09:44] <Hixie> hm?
# [09:46] <zcorpan> hsivonen: i think no parser-level quirks are expected (we can maybe get away with doing <p><table> the same in quirks mode)
# [09:46] <hsivonen> Hixie: are more quirks expected than those that have XXX quirks comments? and is the fragment parsing algorithm supposed to honor the quirkiness of the document of the context node?
# [09:46] <Hixie> no idea
# [09:47] <hsivonen> ok. I'll file a bug
# [09:52] <hsivonen> zcorpan: does Opera have any parser-level quirks left?
# [09:53] <hsivonen> fwiw, source code suggests that Gecko inherits quirkiness on innerHTML
# [09:55] <zcorpan> hsivonen: yes (e.g. <p><table>)
# [09:55] <hsivonen> zcorpan: ok, so there's no shipping precedent to ironing out the differences
# [09:56] <zcorpan> hey i thought webkit did the <p><table> the same in quirks mode
# [09:56] <hsivonen> oh
# [09:57] <zcorpan> wonder why they changed it
# [10:04] <hsivonen> what was the magic click that reveals sections linking to a heading in the WHATWG HTML5 spec?
# [10:04] <annevk5> Hixie, Opera wants spellcheck="" btw
# [10:04] <annevk5> Hixie, I thought I mentioned that somewhere on the list
# [10:05] <hsivonen> Hixie: is there a reason why innerHTML & friends are on HTMLElement instead of Element?
# [10:05] <hsivonen> my recollection is that in WebKit and Opera innerHTML works more generally
# [10:05] <Hixie> annevk5: cool
# [10:06] <Hixie> hsivonen: because i don't define Element :-)
# [10:06] * zcorpan finds https://bugs.webkit.org/show_bug.cgi?id=9280
# [10:06] <annevk5> Hixie, but you could define ElementHTML
# [10:06] <annevk5> Hixie, for instance
# [10:06] <Hixie> i could
# [10:08] <hsivonen> it seems silly that the functionality is not available on SVG nodes (in specs & in Gecko)
# [10:12] * Joins: mpt (n=mpt@canonical/launchpad/mpt)
# [10:19] <Hixie> innerHTML is a terrible API
# [10:19] <Hixie> it's the equivalent of eval
# [10:19] <jgraham> Hixie: I'm here now
# [10:19] <annevk5> Hixie, but it does the job
# [10:19] <Hixie> so does eval
# [10:20] <hsivonen> eval is lispish, so it has to be brilliant
# [10:22] * Joins: sverrej (n=sverrej@122.160.12.230)
# [10:23] <annevk5> given that non-HTML elements are marginal in use anyway and there's no other API available I do not really see the harm in making it consistently available
# [10:25] <Hixie> the harm is in making it available at all
# [10:26] * Quits: jruderman (n=jruderma@corp-246.mountainview.mozilla.com)
# [10:27] <Hixie> does anyone know if there's any way to force another computer on the network to flush its dns cache?
# [10:29] <hsivonen> Hixie: when the cat is totally out of the bag if you use innerHTML on a div wrapping svg, what's the harm of allowing it directly on the svg stuff?
# [10:30] <Hixie> innerHTML is bad. when you have something that is bad, you don't make it available _more_
# [10:30] <zcorpan> badness is just an opinion :)
# [10:30] <annevk5> Hixie, you made it available to XML
# [10:31] <annevk5> Hixie, I'm not really convinced
# [10:31] <hsivonen> Hixie: innerHTML is useful for debugging :-)
# [10:31] <Hixie> annevk5: putting it on XML is necessary to allow easier migration to XML
# [10:31] <Hixie> hsivonen: a readonly attribute innerHTML would be fine
# [10:31] <hsivonen> Hixie: wouldn't easy migration require running the HTML parser on setting?
# [10:32] <Hixie> presumably one would fix all the syntax at the same time
# [10:32] <Hixie> i can't imagine what kind of a mess one would have that would make it easier to only upgrade some of the syntax
# [10:32] <annevk5> Hixie, it seems that if we want HTML+MathML+SVG to look lik a single platform innerHTML should be everywhere
# [10:33] <zcorpan> opera uses the html parser on setting and we have bugs about it, iirc
# [10:33] <annevk5> Hixie, also, if innerHTML is bad in the first place, why would you ever want to allow it to migrate to XML
# [10:33] <Hixie> innerHTML is fundamentally bad in the same way eval is bad -- it doesn't get you any compile-time syntax checking, and it encourages poor escaping hygene
# [10:33] <annevk5> Hixie, especially since you increase implementation complexity as it does not do the same thing there!
# [10:33] * Quits: zalan (n=kvirc@dsl5400EDD3.pool.t-online.hu) (Read error: 110 (Connection timed out))
# [10:33] <hsivonen> I agree on escaping hygiene, but compile-time checking is overrated
# [10:34] <zcorpan> annevk5: didn't opera and gecko predate html5 on this matter, though?
# [10:34] <hsivonen> Python has collection literals, Java doesn't. Many Java devs use XML as complex literals...
# [10:34] <Hixie> hsivonen: for this kind of thing, i think it's underrated
# [10:34] * Joins: virtuelv (n=virtuelv@pat-tdc.opera.com)
# [10:34] <hsivonen> people like having literals
# [10:35] <Hixie> anyway
# [10:35] <Hixie> if you want innerHTML on SVG, get zcorpan to do it in Web DOM Core
# [10:35] <Hixie> I have no interest in sticking my neck out on that front
# [10:35] <Hixie> it's one thing to piss off the svg wg, but pissing them off when i disagree that it's a good idea is quite another :-P
# [10:35] <zcorpan> Hixie: what's the difference from saying that all documents must implement HTMLDocument?
# [10:36] * Quits: maikmerten (n=merten@ls5dhcp196.cs.uni-dortmund.de) (Remote closed the connection)
# [10:36] <Hixie> zcorpan: HTMLDocument doesn't encourage poor escaping hygene and suffer from lack of compile-time syntax checking of complex strings?
# [10:36] <annevk5> zcorpan, Opera's implementation actually didn't increase complexity, but yeah, then again, we could remove it
# [10:37] <annevk5> Hixie, HTMLDocument has innerHTML
# [10:37] <zcorpan> Hixie: i meant having an ElementHTML interface that includes useful stuff that's only available on html elements
# [10:38] * Quits: virtuelv (n=virtuelv@pat-tdc.opera.com) (Client Quit)
# [10:38] <zcorpan> i'd be fine with putting stuff in dom core if that's appropriate
# [10:38] <Hixie> annevk5: innerHTML on a Document is a whole different issue -- it's a replacement for DOM3 Load and Save, to convert a self-contained, typically server-provided, raw document string, into a Document
# [10:38] <Hixie> zcorpan: i don't follow. I thought the problem was that people wanted things on nodes _other_ than HTML elements.
# [10:39] * Joins: virtuelv (n=virtuelv@pat-tdc.opera.com)
# [10:39] * Quits: virtuelv (n=virtuelv@pat-tdc.opera.com) (Client Quit)
# [10:39] <zcorpan> Hixie: yes -- then you say that all Elements must also implement ElementHTML
# [10:39] <Hixie> zcorpan: ...but I don't _want_ non-HTML elements to support these things
# [10:40] <zcorpan> Hixie: what about getElementsByClassName?
# [10:40] <Hixie> what about it?
# [10:41] <annevk5> Hixie, poor hugene and no compile-time syntax checking of complex strings though
# [10:41] <zcorpan> Hixie: it's on HTMLElement but would be similarly useful on svg elements
# [10:41] <annevk5> classList too
# [10:41] <Hixie> annevk5: you wouldn't do (in the use case for Document.innerHTML) any string merging, so there's no escape hygene issue. And compile-time checking makes no sense for something you don't have access to until runtime.
# [10:42] <Hixie> zcorpan: if there are things we should hoist to Web DOM Core Element, I'm fine with that too.
# [10:42] * Joins: ROBOd (n=robod@89.122.216.38)
# [10:42] * Joins: maikmerten (n=merten@ls5dhcp196.cs.uni-dortmund.de)
# [10:44] * annevk5 finds another interesting i18n example
# [10:45] <annevk5> there's an armenian list style in CSS but apparently such a thing is never used; they use arabic or roman numbers for lists in practice
# [10:45] <Hixie> there's like 93 different list styles in css3 lists
# [10:46] * Quits: Lachy (n=Lachlan@85.196.122.246) ("This computer has gone to sleep")
# [10:46] <Hixie> many of them are only of historic interest
# [10:46] <Hixie> they're mostly for use with counters rather than lists
# [10:46] <annevk5> why add bunch of codes to browsers for historic interest?
# [10:46] <annevk5> s/codes/code/
# [10:47] <Hixie> it's very little code in most cases, just a table typically
# [10:47] <Philip`> zcorpan: When adding those two attributes, the second one overrides the first one, so you only get one attribute in the output
# [10:47] <Philip`> zcorpan: (XOM does let you manage prefixes, but it does comparisons based on URI+localname and the prefixes are more like hints for serialisation)
# [10:48] <Philip`> hsivonen: <?foo ??> is well-formed as far as I'm aware; I guess the code that adds the space was copied-and-pasted from the code that avoids
# [10:49] <annevk5> Hixie, same question s/code/tables/
# [10:49] <Hixie> tables are cheap
# [10:50] <hsivonen> Philip`: indeed, it looked copied and pasted, so I zapped it
# [10:50] <annevk5> that's hardly a justification :)
# [10:51] <annevk5> especically since it's a bit more than tables for some and testing all that stuff is quite a bit of effort
# [10:51] <Hixie> list styles are pretty easy to test
# [10:51] <Hixie> you just implement the algorithm and output bazillions of automated tests automatically
# [10:51] <jgraham> Unless you implement the algorithm wrong
# [10:52] <jgraham> In which case you output bazillions of misleading tests automatically :)
# [10:52] <Hixie> well then you catch the bug when the "real" implementor tries it
# [10:52] <Hixie> writing wrong tests is always a risk
# [10:52] <Hixie> for lists it's far easier to get them right than, say, margin collapsing
# [10:53] <Hixie> anyway, list styles, even those with only small markets, can be useful
# [10:53] <Hixie> why would we not want to be comprehensive?
# [10:53] <Hixie> since they're cheap and easy to test, it seems like a rare case where we can afford to be
# [10:53] * Joins: nessy (n=nessy@203-158-43-142.dyn.iinet.net.au)
# [10:54] * Joins: jruderman (n=jruderma@c-76-102-160-171.hsd1.ca.comcast.net)
# [10:55] <Philip`> hsivonen: I don't think there's a magic click, you just use the single-page spec and then click the term's definition
# [10:55] <hsivonen> aah. single-page!
# [11:01] <jgraham> Hixie: Speaking of broken tests, you do realise that for every change you make to the AAA, a tiny kitten dies, right?
# [11:01] <Hixie> blame hsivonen
# [11:02] <Hixie> (and sicking)
# [11:02] <jgraham> That doesn't help the poor little kitten now does it? :)
# [11:02] <jgraham> Anyway, I can't really blame hsivonen because I'm counting on him to fix the html5lib test suite before I get around to implementing the change
# [11:03] * Joins: virtuelv (n=virtuelv@pat-tdc.opera.com)
# [11:03] * Joins: jwalden_ (n=waldo@c-76-102-160-171.hsd1.ca.comcast.net)
# [11:04] * Joins: myakura (n=myakura@p3020-ipbf505marunouchi.tokyo.ocn.ne.jp)
# [11:10] * Joins: virtuelv_ (n=virtuelv@213.236.208.247)
# [11:11] * Quits: virtuelv_ (n=virtuelv@213.236.208.247) (Client Quit)
# [11:11] * Quits: jwalden (n=waldo@c-76-102-160-171.hsd1.ca.comcast.net) (Read error: 110 (Connection timed out))
# [11:12] <jgraham> gsnedders: shout when you're around.
# [11:13] * Joins: virtuelv_ (n=virtuelv@213.236.208.247)
# [11:18] * Quits: jruderman (n=jruderma@c-76-102-160-171.hsd1.ca.comcast.net)
# [11:21] * Quits: virtuelv (n=virtuelv@pat-tdc.opera.com) (Read error: 110 (Connection timed out))
# [11:25] * Joins: zalan (n=kvirc@dsl5402683B.pool.t-online.hu)
# [11:34] * Parts: annevk5 (n=annevk@85.196.122.246)
# [11:37] * Quits: Maurice (n=ano@a80-101-46-164.adsl.xs4all.nl) (Read error: 60 (Operation timed out))
# [11:38] * Joins: Maurice (n=ano@a80-101-46-164.adsl.xs4all.nl)
# [11:42] * Joins: yecril71 (n=giecrilj@piekna-gts.2a.pl)
# [11:49] * Joins: arve__ (n=virtuelv@pat-tdc.opera.com)
# [11:51] * Quits: virtuelv_ (n=virtuelv@213.236.208.247) (Read error: 110 (Connection timed out))
# [11:51] * Joins: Lachy (n=Lachlan@213.236.208.247)
# [11:52] * Joins: annevk5 (n=annevk@pat-tdc.opera.com)
# [11:52] <Hixie> i love that the TR/html5 doc is already out of date before even getting published
# [11:54] <annevk5> you should love the point where that stops being the case :)
# [11:57] * Joins: billyjackass (n=MikeSmit@71-218-59-55.hlrn.qwest.net)
# [11:58] * Quits: Lachy (n=Lachlan@213.236.208.247) (Read error: 60 (Operation timed out))
# [11:58] * Joins: Lachy (n=Lachlan@pat-tdc.opera.com)
# [11:59] <Hixie> annevk5: i don't expect it'll be the case before october :-)
# [12:01] * Philip` wonders how Rob Burns can possibly claim that the XML spec says "&#xd800;&#xdc00;" is well-formed
# [12:01] <Philip`> (on www-tag)
# [12:02] <annevk5> Rob Burns also claims that by virtue of referencing Unicode XML requires canonical character comparison rather than codepoint comparison
# [12:02] <annevk5> It's really quite interesting :)
# [12:04] <zcorpan> Philip`: maybe... it says character reference*s* must match Char, so after expanding the two references and serializing the result as utf-16 and parsing it again it would match Char?
# [12:04] <Philip`> zcorpan: You can't serialise U+D800 to UTF-16
# [12:05] <zcorpan> bummer :(
# [12:05] <Philip`> So I guess you have to expand the two references, serialise as UCS-2, then reinterpret it as UTF-16 when parsing, and then it would match Char
# [12:05] <Philip`> though that would break whenever you use >= U+10000 in your document
# [12:07] <zcorpan> you'd only do it for character references that expand to the surrogate range?
# [12:07] * arve__ is now known as virtuelv
# [12:07] <Philip`> "If the character reference begins with " &#x ", the digits and letters up to the terminating ; provide a hexadecimal representation of the character's code point in ISO/IEC 10646."
# [12:08] <Philip`> so he can't claim that &#xd800;&#xdc00; should be parsed into a single codepoint
# [12:08] <hsivonen> Philip`: sure he *can* and just did :-(
# [12:08] <heycam> right, and there's no such character as U+D800
# [12:09] <Philip`> hsivonen: Hmm, I didn't see him claim that (yet)
# [12:09] <yecril71> krijnhoetmer is down again
# [12:09] * Quits: billyjackass (n=MikeSmit@71-218-59-55.hlrn.qwest.net) ("Tomorrow to fresh woods, and pastures new.")
# [12:09] <Philip`> yecril71: Works for me
# [12:09] * heycam always pronounces "krijnhoetmer" as "chronometer" in his mind
# [12:10] <hsivonen> Philip`: ok. perhaps he didn't claim exactly that, but he claims a lot of things
# [12:10] <annevk5> yecril71, if krijnh is actually in the channel it's unlikely to be down
# [12:10] <yecril71> Take it back, I expected Goto to refresh in IE.
# [12:11] * yecril71 takes it back
# [12:11] <Philip`> hsivonen: If he did claim that then at least he'd be more internally consistent, even if it made him more wrong :-)
# [12:11] * hsivonen thinks the issue timeless pointed out is a bug in cell phone carriers in Canada
# [12:12] <yecril71> If the user�s language does not match the site�s, how on earth can the user actually use the site?
# [12:13] <annevk5> yecril71, because the user is e.g. Dutch but can read English?
# [12:13] <yecril71> For the life of me, I cannot figure out anything on Chinese sites.
# [12:13] <jgraham> yecril71: Google translate or e.g. having some but not good comprehension of the language
# [12:13] * jgraham has to use Swedish sites occasionally
# [12:13] <yecril71> Does Google translate allow you to submit forms?
# [12:13] * annevk5 can quite easily use oslokino.no
# [12:14] <annevk5> without speaking Norwegian
# [12:14] <hsivonen> yecril71: I tend to use Google Translate in another tab when I need to submit forms e.g. in German
# [12:14] <annevk5> s/spreaking/knowing/
# [12:14] <zcorpan> yecril71: gmail.com might be in english but you want to write a mail in a different language
# [12:15] <yecril71> Annevk5! How do you use oslokino.no?
# [12:17] <yecril71> It seems like a film review site, the reviews are in Norwegian.
# [12:17] <Lachy> JohnResig, yt?
# [12:18] <annevk5> among other things it lists which movies run in theaters in Oslo, which is useful when I'm there
# [12:18] <hsivonen> yecril71: it seems the main use case for the site is buying tickets
# [12:18] <yecril71> Even if I want to write a mail in a different language, my language and the GMail�s language are still the same.
# [12:19] <yecril71> It is the recipient�s language that is different.
# [12:19] <yecril71> Do you also translate your feedback into German with Google first?
# [12:20] <yecril71> In my ideal world, everyone would use the language she knows best,
# [12:21] <yecril71> and the recipient�s software would have the task to translate it.
# [12:21] <hsivonen> yecril71: does "use" mean write, read, or both?
# [12:21] <yecril71> In this case, write, mostly.
# [12:22] <yecril71> Passages for reading do not need spell checking.
# [12:22] <yecril71> But you have to be able to read in order to know what to write (and where).
# [12:23] <hsivonen> I wouldn't hold my breath with that vision. Even with human translators, after some threshold of reading proficiency in a foreign language, people are better off reading original text instead of translations
# [12:23] <yecril71> The translation would only be a prothesis, the original text being available as well.
# [12:25] <yecril71> I am unhappy about breaking the link between @lang and spell checking.
# [12:25] * hsivonen wonders how the IETF mechanism of getting security review works compared to getting Adam Barth to read a draft
# [12:26] <yecril71> Hopefully, the browsers can still do it, regardless that Hixie decided to ignore it.
# [12:27] <Hixie> did you read what the spec says on the subject?
# [12:29] <yecril71> The spec says that spellcheck is used to turn spell checking on for text input controls.
# [12:29] <Hixie> i urge you to read the entire section :-)
# [12:29] <krijnh> I'm down?
# [12:30] <krijnh> yecril71: No, I'm not :)
# [12:30] <yecril71> No, I am dumb :-(
# [12:30] <yecril71> Or rather MSIE is.
# [12:31] <krijnh> (heycam: it's more like "cryin' hoodmer", but who cares :)
# [12:31] <yecril71> Why does MSIE want to load Microsoft HTML Viewer to display the spec?
# [12:32] <annevk5> why do people use MSIE?
# [12:33] <yecril71> Because their employers want them to use Microsoft Windows, I presume.
# [12:33] <yecril71> Because it is the best choice for incompetent losers.
# [12:34] * Philip` notes that other web browsers work on Microsoft Windows too
# [12:34] * yecril71 notes that that doubles the maintenance cost
# [12:35] <krijnh> Because of shitty IE-only intranet apps and sysadmins not doing their jobs
# [12:36] <yecril71> OK, I disabled CSS so perhaps I shall be able to read it now
# [12:40] <yecril71> Except that disabling CSS is not permanent, and I have to wait for ages until I will be able to disable it again :-(
# [12:47] <yecril71> I am sorry, I am unable to find spellcheck in the contents.
# [12:47] <yecril71> http://www.whatwg.org/specs/web-apps/current-work/multipage/index.html#contents
# [12:47] <yecril71> Find "spellcheck".
# [12:49] * Joins: robburns (n=robburns@adsl-99-180-247-184.dsl.chcgil.sbcglobal.net)
# [12:50] <robburns> Philip` with your quote, I see how you're reading XML 1 now.
# [12:50] <robburns> Philip`: "If the character reference begins with " &#x ", the digits and letters up to the terminating ; provide a hexadecimal representation of the character's code point in ISO/IEC 10646."
# [12:51] <yecril71> The attribute values should be
# [12:51] <yecril71> spellcheck="on/off", not "true/false"
# [12:52] <robburns> My reading overlooked the subtle different there where it expands to the CHARACTER'S code point
# [12:52] <robburns> Philip`: so I was thinking that the two character references expanded to two code points (surrogates) that when adjacent to one another signified an astral code point
# [12:54] <hsivonen> robburns: did you read the BNF comment at http://www.w3.org/TR/REC-xml/#NT-Char ?
# [12:54] <robburns> Philip`: At one point all the major browsers supported that reading
# [12:54] <robburns> see https://bugs.webkit.org/show_bug.cgi?id=6446
# [12:54] <robburns> and later https://bugs.webkit.org/show_bug.cgi?id=22210
# [12:54] <hsivonen> robburns: browser behavior is not normative over XML :-)
# [12:54] <Philip`> robburns: As far as I'm aware surrogates only signify astral code points in UTF-16, and UTF-16 doesn't seem relevant to character references at all
# [12:55] <robburns> hsivonen: Yes, I did. However as I said in that email that makes it clear that you cannot have  but not so clear that you can't have two valid surrogates adjacent to one another.
# [12:55] <robburns> Both IE and Safari still support this.
# [12:56] <yecril71> Why is spell checking disabled by default?
# [12:56] <robburns> Philip`: yeah, but having those surrogates as assigned code points (rather than a UTF issue) certainly contributes some ambiguity. After all we don't have UTF-8 code points assigned too.
# [12:57] <yecril71> Spell checking should be enabled IMHO, except for text input controls.
# [12:58] <robburns> yecril71: in my opinion, spell checking shouldn't be up to authors at all
# [12:58] <robburns> spellchecking is just not an authors concern
# [12:58] <yecril71> (INPUT[type=TEXT], that is)
# [12:58] <robburns> meaning the author of the page, not the author now using that page to input html or other content
# [12:58] <yecril71> But the latest specification says it should be off by default.
# [12:59] <Philip`> http://philip.html5.org/misc/surrogate-charrefs.xml
# [13:00] <Philip`> Gives error in Firefox 2, Opera 9.6, Safari 3.1; displays with no error in IE6
# [13:00] <robburns> Philip`: you're right. I guess for XML Safari doesn't handle that. I had just tested in HTML, but I thought I had done it before in XHTML.
# [13:00] <Philip`> IE8 does show an error message, but also seems to carry on parsing anyway
# [13:01] <Philip`> http://philip.html5.org/misc/single-surrogate-charref.xml
# [13:01] <Philip`> IE6 is happy with that too
# [13:01] <hsivonen> robburns: may I suggest more careful testing next time before claiming unusual things about XML conformance on the mailing lists
# [13:01] * Joins: zalan_ (n=kvirc@dsl54007086.pool.t-online.hu)
# [13:01] <robburns> Philip`: well I think a year or so ago, IIRC Safari, FireFox and IE were all handling those surrogates.
# [13:01] <Philip`> (and IE8 does the same as the other example)
# [13:02] <jgraham> gsnedders: Specifically I would like anolis to support anolislib.generator.fromFile(file, **kwargs) and anolislib.generator.toString(tree, **kwargs) which I implemented but left the patch elsewhere
# [13:02] <robburns> hsivonen: well with the flip-flopping on this (as it is described a the WebKit bugzilla) I don't think I should be held responsible for the most up-to-date state of this.
# [13:02] * jgraham thinks about implementing it now
# [13:03] <robburns> hsivonen: who knows tomorrow it might all work again in those browsers. There's some people out there who like to ensure there are as many fatal errors in XML as can be even if they aren't really in the spirit of fatal error handling of XML.
# [13:03] * Joins: pauld (n=pauld@217.41.236.36)
# [13:03] <Philip`> robburns: Those bug reports are about HTML, which is not XML
# [13:04] <hsivonen> robburns: when in doubt, you may assume that XML bugs Philip` has found in Validator.nu and I have fixed are actual XML violations
# [13:04] <robburns> Philip`: "HTML/XML character set (independent of actual character encoding of a document)
# [13:04] <robburns> is Unicode/ISO 10646 and NCRs represent Unicode code points. They do not
# [13:04] <robburns> represent '2byte code units' of UTF-16. So, NCRs with surrogate code points
# [13:04] <robburns> should not be allowed whether they are paired or not. " Says XML there.
# [13:04] <robburns> hsivonen: I don't agree.
# [13:05] <robburns> hsivonen: I think this is open to interpretation and I don't think treating those as fatal error is really in keeping with the reasons for fatal error handling in XML (on the other hand I think you're right that it shouldn't be serialized that way)
# [13:05] <Philip`> robburns: It does literally say "XML", but it's only talking about changes to the HTML parser (hence it being the "HTML DOM" component, and referring to a bug which involved a change to the HTML parser)
# [13:05] <hsivonen> robburns: I'm not suggesting taking me or Philip` as authorities. Just that checking specs carefully is a good idea before suggesting that someone else didn't grok the specs.
# [13:06] * Philip` is certainly not an authority, and is often mistaken :-)
# [13:06] <hsivonen> robburns: when *you* think something is open to interpretation, it usually helps to check if multiple XML parser implementors happen to agree on the interpretation (they might have it right)
# [13:07] <robburns> hsivonen: I quoted the spec and gave you my interpretation of it. I can see how Philip read it differently now. However, there is room for interpretation there.
# [13:07] <yecril71> Editorial: http://html5.org/tools/web-apps-tracker?from=2800&to=2801 features "asits"
# [13:08] <robburns> Philip`: well I would count you here as an authority, but even authorities can get too draconian at times.
# [13:08] * Quits: zalan_ (n=kvirc@dsl54007086.pool.t-online.hu) ("KVIrc 3.4.0 Virgo http://www.kvirc.net/")
# [13:09] <robburns> hsivonen: I have checked multiple XML parsers (at least from my recollection). However, as I said the implementations are flip flopping on this because there's a group of interested parties who like to make sure XML is more draconian then it is intended to be.
# [13:09] <Philip`> robburns: I'm not interested in being draconian, just in understanding the spec in the same way that everyone else understands it
# [13:09] <robburns> hsivonen as Philip` demonstrated IE does parse surrogate pair character references.
# [13:10] <Philip`> and I haven't seen any evidence of people being indecisive about how to implement it in XML (as opposed to in HTML where it's not been specified before HTML5)
# [13:10] * jgraham wonders if gsnedders realises that fromFile will sometimes work if the input is actually a string
# [13:10] * Joins: pesla\work (n=retep@procurios.xs4all.nl)
# [13:10] <Philip`> except for IE starting to flag it as an error, following what other parsers seem to do
# [13:10] * Joins: zalan_ (n=kvirc@dsl54007086.pool.t-online.hu)
# [13:12] <robburns> Philip`: here's another bug report https://bugs.webkit.org/show_bug.cgi?id=18039
# [13:12] <robburns> seems to be xml related
# [13:13] <hsivonen> robburns: also doesn't support your interpretation :-)
# [13:13] * Joins: erlehmann (n=erlehman@86.59.25.121)
# [13:13] <annevk5> "because there's a group of interested parties who like to make sure XML is more draconian then it is intended to be" really? :)
# [13:14] <zcorpan> robburns: the test case has "𝌴𝌴𝌴" - doesn't look like surrogate character references to me
# [13:14] * Quits: pesla (n=retep@procurios.xs4all.nl) (Read error: 60 (Operation timed out))
# [13:14] <annevk5> it's an interesting conspiracy theory
# [13:14] <robburns> yeah, I see that now
# [13:15] <robburns> zcorpan: ^
# [13:16] * Philip` notes that his testing in IE6 was actually in Wine, and he's not sure if that would affect the results
# [13:18] * jgraham tries upgrading PimpMySpec.net to a version that allos you to set options for URL-based (rather than upload-based) requests
# [13:19] <robburns> Philip`: I don't think that would change anything.
# [13:19] <jgraham> It is much uglier than it used to be and has approximately no QA
# [13:19] * Quits: zalan (n=kvirc@dsl5402683B.pool.t-online.hu) (Read error: 110 (Connection timed out))
# [13:22] <robburns> annevk5: "because there's a group of interested parties who like to make sure XML is more draconian then it is intended to be" really? yes really!
# [13:23] <Philip`> Ooh, there's even a test
# [13:23] <Philip`> http://dev.w3.org/cvsweb/2001/XML-Test-Suite/xmlconf/xmltest/not-wf/sa/145.xml?rev=1.1.1.1&content-type=text/x-cvsweb-markup
# [13:23] <Philip`> (where the "not-wf" directory is used for not-well-formed documents)
# [13:24] <hsivonen> clearly, the XML test suite folks are trying to make sure XML is more Draconian than it is intended to be!
# [13:24] <annevk5> every year I file these bugs on browsers to fail on non-NFC content, but nobody implements :(
# [13:26] <annevk5> robburns, kidding aside, anything in particular?
# [13:27] <robburns> Philip`: that test is a single surrogate unless I'm missing someting
# [13:27] <Philip`> robburns: Oh, good point
# [13:27] <robburns> hsivonen: I guess we're all too quick to find urls to prove the others wrong
# [13:28] * Quits: Maurice (n=ano@a80-101-46-164.adsl.xs4all.nl) (Read error: 60 (Operation timed out))
# [13:29] <hsivonen> http://diveintomark.org/archives/2004/08/16/specs
# [13:29] <annevk5> http://ajaxian.com/archives/frames-are-back oops
# [13:31] * Joins: Maurice (n=ano@a80-101-46-164.adsl.xs4all.nl)
# [13:31] <Philip`> http://dev.w3.org/cvsweb/2001/XML-Test-Suite/xmlconf/oasis/p66fail6.xml?rev=1.1.1.1&content-type=text/x-cvsweb-markup
# [13:31] <Philip`> ("<TEST TYPE='not-wf' SECTIONS='4.1 [66]' ID='o-p66fail6' URI='p66fail6.xml'> no references to non-characters </TEST>")
# [13:34] <robburns> Philip`: I guess since that test came after IE's implementation they didn't know about that.
# [13:35] <annevk5> it's in the spec...
# [13:35] * Joins: virtuelv_ (n=virtuelv@213.236.208.247)
# [13:36] <robburns> annevk5: no it's not (at least not clearly), we've already discussed that.
# [13:37] <robburns> annevk5: so then we've looking to the implementations and, at least for xml, its IE on once side and everything else on the other (until IE8).
# [13:42] <Philip`> See o-p66fail6 in http://xmlconf.sourceforge.net/xml/reports/report-msxml2-val.html vs http://www.xml.com/2000/08/30/msxml/msxml3-val.html
# [13:42] <robburns> and for HTML firefox flip flopped and safari joined firefox (before they flipped back) and IE
# [13:42] <Philip`> It was broken in a May 2000 preview, and fixed a few months later
# [13:43] <robburns> Philip`: I'm not following what you mean. Broken as in they didn't support surrogate references or they did?
# [13:43] <Philip`> robburns: I don't see the relevance of HTML to the parsing of XML
# [13:43] <hsivonen> annevk5: in the first edition, even
# [13:43] <Philip`> robburns: "Broken" as in they failed the test (and did not reject the document as not-well-formed)
# [13:44] * Quits: virtuelv (n=virtuelv@pat-tdc.opera.com) (Read error: 110 (Connection timed out))
# [13:46] <robburns> Philip`: I"m not saying that HTML is relevant to XML, but rather it is relevant to the understanding that two numeric character references that reference valid surrogate pairs can be understood as referencing an astral character
# [13:46] * Quits: dave_levin (n=dave_lev@c-98-203-247-78.hsd1.wa.comcast.net)
# [13:47] <Philip`> robburns: The only things that seem relevant to the parsing of character references in XML are the XML spec and current XML parser implementations
# [13:47] <robburns> Philip`: Of course, but the way any reader reads a spec is colored by their understanding of everything they've read before (or understood before).
# [13:48] * Quits: sverrej (n=sverrej@122.160.12.230) (Read error: 110 (Connection timed out))
# [13:48] <robburns> Philip`: Unless you think that specs are written with the precision of a programming language
# [13:48] <robburns> and I've yet to see one that where I would say that.
# [13:51] <hsivonen> anyone got test cases for innerHTML?
# [13:51] <Philip`> robburns: Ah, indeed, it's relevant in terms of providing a rough mental model that lets people understand the specs more easily; it just doesn't seem helpful to use that rough mental model when discussing precise details of the spec's requirements
# [13:52] <zcorpan> hsivonen: yep, hold on
# [13:52] <robburns> Philip`: I would also say that where you would treat the implementations as the arbiter of what the spec actually says, I would rather think about what the spirit of the spec
# [13:53] <zcorpan> http://simon.html5.org/test/html/serializing/001.htm http://simon.html5.org/test/html/serializing/002.xht http://simon.html5.org/test/html/parsing/fragment/content-model-flag/
# [13:53] <robburns> Philip`: and in this case I would say that the IE treatment of surrogate pair references is more in keeping with the spirit of the xml spec regardless of what others have implemented or produced test cases to demonstrate
# [13:53] <hsivonen> zcorpan: great! thanks.
# [13:54] <Philip`> robburns: I would treat the apparently unanimous interpretation of the spec by the experts who've implemented the spec and written and reviewed test cases, as being the only relevant interpretation of the spec
# [13:55] <robburns> Philip`: I'd probably be with you if you didn't have to use the word "apparently" before "unanimous"
# [13:55] <zcorpan> hsivonen: some tests might be wrong by now though
# [13:55] <hsivonen> zcorpan: ok
# [13:55] <robburns> Philip`: since one implementation didn't treat it that way.
# [13:56] <Philip`> robburns: It's only "apparently" because I can only judge it based on all the evidence I've seen
# [13:56] <zcorpan> hsivonen: http://tc.labs.opera.com/apis/innerHTML/xml/
# [13:56] <zcorpan> and ../
# [13:56] <Philip`> robburns: I've not seen anything to indicate that was anything other than a bug, which got quickly fixed once it was identified
# [13:56] <hsivonen> robburns: what's ambiguous about: "Characters referred to using character references must match the production for Char." and "Char ::= #x9 | #xA | #xD | [#x20-#xD7FF] | [#xE000-#xFFFD] | [#x10000-#x10FFFF]"? (quotes from XML 1.0 1st ed.)
# [13:57] <robburns> Philip`: which if I understand correctly had IE as doing something different than other implementations, right? So by "apparently unanimous" you meant "nearly unainimous"
# [13:58] * Quits: virtuelv_ (n=virtuelv@213.236.208.247) (Read error: 110 (Connection timed out))
# [13:58] <robburns> hsivonen: I already addressed that. If it said "A character referred to using a character reference must math the production for Char" However, it says "Characters referred to using character references must match the production for Char" and it seems reasonable to interpret two surrogate paris (one high and one low) as matching that BNR notation.
# [13:59] <Philip`> robburns: I mean that (apparently) nobody argued that they interpreted the XML spec as claiming &#xd800;&#xdc00; was valid, regardless of whether they had implementation bugs
# [13:59] <hsivonen> robburns: no, it is not reasonable to interpret it in any other way except restricting the expansion of each NCR individually
# [13:59] <robburns> Philip`: well whoever read it while implementing the IE parser must have read it that way.
# [14:00] <robburns> hsivonen: It's sounds insane when you say that
# [14:00] <robburns> hsivonen: Philip`'s quote was more to the point in favor of no surrogates.
# [14:01] <robburns> "If the character reference begins with " &#x ", the digits and letters up to the terminating ; provide a hexadecimal representation of the character's code point in ISO/IEC 10646."
# [14:01] <Philip`> robburns: More likely they didn't read it carefully at all, and just wrote something that seemed close enough
# [14:02] <Philip`> or based it on ancient pre-first-edition version of XML that didn't have the same restrictions
# [14:02] <robburns> the word "character" there leans it away from my interpretation. But that's still a a slight lean
# [14:02] <Philip`> because that's how most spec-compliance bugs seems to come into existence
# [14:02] * Quits: eric_carlson_ (n=ericc@adsl-67-112-12-110.dsl.anhm01.pacbell.net)
# [14:02] <robburns> Philip`: Well one could always discount any and all interpretations with that argument. I could say maybe the IE developers were the only ones to read it carefully.
# [14:03] <robburns> Philip`: the point is that fatal errors in the case of surrogates is unlike any of the fatal errors. It doesn't break the rest of the page the way others do.
# [14:04] <Philip`> robburns: If they read the spec in accordance with its intent, why would they have changed their implementation when finding a test failed, rather than having the test be fixed and the spec clarified to match what was intended?
# [14:04] * zcorpan wonders why he is reading this discussion
# [14:04] * Parts: zcorpan (n=zcorpan@pat.se.opera.com)
# [14:04] <robburns> Two well-formed numeric character references that happen to be an illegal character shouldn't break rendering of the rest of the page (the way other well-formedness errors should) That's what I'm talking about with the spirit of the spec. I agree a serializer shouldn't produce those, but that's a different situation.
# [14:05] * Quits: nessy (n=nessy@203-158-43-142.dyn.iinet.net.au) ("This computer has gone to sleep")
# [14:05] <robburns> Philip`: I'm not sure. I didn't really follow that web page you sent. I thought you were testing recent IE and it wasn't "fixed"
# [14:05] * Quits: pauld (n=pauld@217.41.236.36)
# [14:06] <robburns> zcorpan: maybe it's like a train wreck :-)
# [14:07] <Philip`> robburns: The two pages that gave conformance test results for some old (2000) versions of MSXML, and showed that a preview release failed and a slightly later release passed the test?
# [14:07] <robburns> Philip`: Yeah, that's what I thought it said. But I also thought you said you tested IE and the surrogate pair references worked in 2009.
# [14:08] <robburns> Philip`: and I couldn't reconcile those two things.
# [14:08] <Philip`> That was IE6 (from 2001), running in Wine (so I've got no idea what version of MSXML it's using, or if it emulates it using some other code)
# [14:09] <robburns> Philip`: OK I see. But didn't you say that IE8 was changing it?
# [14:10] <Philip`> robburns: I said I tested in IE8 (on Vista) and it did report the error
# [14:10] <Philip`> (Well, I didn't say the "on Vista" part)
# [14:10] <robburns> but didn't wasn't fatal right?
# [14:12] <robburns> Philip`: I mean "but it wasn't a fatal error, right?
# [14:13] <robburns> "
# [14:14] <Philip`> robburns: It's an error (and causes an error dialog box to appear), but (at least when I'm just navigating to the XML file directly in IE) it continues parsing the document and displays the rest of it in its usual DOM tree thing
# [14:14] <Philip`> ( has exactly the same effect)
# [14:16] <robburns> Philip`: well that too is a different reading of the spec.
# [14:18] <Philip`> If I access it via XMLHttpRequest, then responseXML is empty if I have  or &#xd800;&#xdc00; etc
# [14:19] <Philip`> so the error prevents the XML file from being used
# [14:19] <Philip`> so it seems to be a fatal error in that context
# [14:19] <Philip`> (It might just be an artifact of the default XML document display that makes it continue past some errors)
# [14:21] * Joins: virtuelv (n=virtuelv@pat-tdc.opera.com)
# [14:22] * Quits: virtuelv (n=virtuelv@pat-tdc.opera.com) (Remote closed the connection)
# [14:27] * Joins: mstange (n=markus@aixd3.rhrk.uni-kl.de)
# [14:30] <robburns> Philip`: which is another way of saying, a different reading of the spec :-)
# [14:30] <annevk5> robburns, your way of reading specs is strange
# [14:31] <robburns> annevk5: I'm sure it would to you.
# [14:31] <annevk5> it seems I'm not alone
# [14:31] <robburns> annevk5: no sadly.
# [14:32] * jgraham thinks robburns would do well as a professor of literature or simesuch
# [14:32] <jgraham> *somesuch
# [14:33] * Joins: virtuelv (n=virtuelv@pat-tdc.opera.com)
# [14:34] * Quits: mstange (n=markus@aixd3.rhrk.uni-kl.de) (Remote closed the connection)
# [14:35] * Joins: mstange (n=markus@aixd3.rhrk.uni-kl.de)
# [14:35] * Quits: robburns (n=robburns@adsl-99-180-247-184.dsl.chcgil.sbcglobal.net)
# [14:36] * Joins: pauld (n=pauld@217.41.236.36)
# [14:37] <annevk5> philosophy maybe?
# [14:37] <annevk5> oh, he left
# [14:37] <annevk5> :/
# [14:37] <jgraham> I woner if I often have that effect on people
# [14:37] * Parts: annevk5 (n=annevk@pat-tdc.opera.com)
# [14:38] * Joins: annevk5 (n=annevk@pat-tdc.opera.com)
# [14:38] <takkaria> rob is a professor of literature or somesuch, he's a PhD with interests in Marxist philosophy and history of thought
# [14:39] <annevk5> ah, so I was close
# [14:41] <jgraham> takkaria: I knew that bit don't know if it has roughly the same job requirements as Literature
# [14:41] <takkaria> well, it's all much of a sameness
# [14:42] * Quits: virtuelv (n=virtuelv@pat-tdc.opera.com) (Remote closed the connection)
# [14:43] <jgraham> (specifically Literature professors seem to get known for having a specific type of interpretation that they like e.g. Marxist or Feminist or whatever. Then they write about how literature fits their mode of interpretation)
# [14:43] * Joins: virtuelv (n=virtuelv@pat-tdc.opera.com)
# [14:44] * Quits: shepazu (n=schepers@c-24-8-217-41.hsd1.co.comcast.net) (Read error: 110 (Connection timed out))
# [14:45] <JohnResig> Lachy: hey
# [14:45] * wilhelm would expect a more pragmatic, practice-based approach from a fellow marxist.
# [14:46] * Joins: virtuelv_ (n=virtuelv@pat-tdc.opera.com)
# [14:47] * Quits: virtuelv (n=virtuelv@pat-tdc.opera.com) (Client Quit)
# [14:49] <Lachy> JohnResig, I just wanted to check with you that it's OK I incorporate your selectors api tests directly into the official test suite, just for licencing reasons. (I'm not sure what licence the test suite will use yet, but I'll try and make it MIT or something)
# [14:50] <JohnResig> Lachy: I think that's ok. I'm curious though - was there really a significant number of other tests to warrant the creation of a new suite?
# [14:50] <Lachy> JohnResig, I created the stuff in CVS shortly before you created your tests
# [14:51] <Lachy> then I haven't touched it much since, because I was using yours
# [14:51] <JohnResig> Lachy: sure - so couldn't we just switch and use the one I created as the official one?
# [14:51] <Lachy> possibly, but we would have to make it work in IE8
# [14:52] <JohnResig> k, one sec
# [14:52] <Lachy> that means we can't use constants like DOMException.SYNTAX_ERR, or tree walkers, and other stuff
# [14:52] <JohnResig> Lachy: wait - *passes* in IE8 or *runs* in IE8?
# [14:53] <Lachy> it currently doesn't run at all in IE8
# [14:53] <JohnResig> lemme see
# [14:53] * Quits: virtuelv_ (n=virtuelv@pat-tdc.opera.com) ("Leaving")
# [14:54] <JohnResig> ok, let me fiddle around with it
# [14:54] <Lachy> oh, and where you set the css and ecss variables, you would need to use .innerHTML rathern than .firstChild.nodeValue to get the CSS from the style elements
# [14:54] <JohnResig> yeah, I just got to that line, as well :)
# [14:55] * Joins: virtuelv (n=virtuelv@pat-tdc.opera.com)
# [15:02] <JohnResig> Lachy: eww... removing the tree walker is going to make things dicey
# [15:02] * Joins: aroben (n=aroben@unaffiliated/aroben)
# [15:02] <JohnResig> Lachy: since this was designed to work on document fragments as well
# [15:02] <JohnResig> hmm
# [15:02] <Lachy> yeah, that's the bit I got stuck on too
# [15:02] <JohnResig> I'll fiddle with it
# [15:03] <JohnResig> I've already fixed the CSS loading and the DOMException stuff
# [15:07] * Quits: aroben (n=aroben@unaffiliated/aroben) (Read error: 104 (Connection reset by peer))
# [15:07] <gsnedders> jgraham: ping
# [15:09] * Joins: aroben (n=aroben@unaffiliated/aroben)
# [15:13] * Quits: MikeSmith (n=MikeSmit@71-218-59-55.hlrn.qwest.net) ("Tomorrow to fresh woods, and pastures new.")
# [15:20] * Joins: sverrej (n=sverrej@59.94.248.196)
# [15:20] <yecril71> MSIE supports objStyle.text
# [15:23] <jgraham> gsnedders: I have a patch for anolis. Also I have updated pms.net to allow passing in options in some cases (specifically: in the case where one GETs a URL rather than POSTs a file) but I have no idea if it works right
# [15:24] <gsnedders> jgraham: email it to me
# [15:24] <gsnedders> (Sorry, now really isn't a good time for me.)
# [15:25] <jgraham> gsnedders: OK
# [15:30] * Joins: eric_carlson (n=ericc@nat/apple/x-4c87cc268a1bab97)
# [15:34] <jgraham> gsnedders: Sent
# [15:35] <gsnedders> jgraham: I'll reply later (where later means sometime in the future :))
# [15:39] <jgraham> gsnedders: No panic
# [15:39] <Philip`> gsnedders: That's usually what "later" means...
# [15:39] <jgraham> Philip`: Except on a closed-timelike-loop when it also means sometime in the past
# [15:40] <Philip`> jgraham: Good point
# [15:40] <gsnedders> Philip`: Go read up on emphasis by repetition.
# [16:02] * Quits: ap (n=ap@194.154.88.32)
# [16:06] * Quits: mstange (n=markus@aixd3.rhrk.uni-kl.de) (Remote closed the connection)
# [16:10] * Quits: virtuelv (n=virtuelv@pat-tdc.opera.com) ("Leaving")
# [16:12] <Lachy> sicking, re http://lists.w3.org/Archives/Public/public-html/2009Feb/0249.html - any existing filter that lets through unknown elements like <handler>, would be just as likely to let through new elements like <video> and new event attributes.
# [16:12] <Lachy> so, e.g. <video onloadstart="xssAttack();"> would be just as problematic
# [16:14] <Philip`> Lachy: They could block any attributes starting with "on"
# [16:16] <jgraham> Die, Spellchecking thread, die
# [16:16] <Lachy> Philip`, it's possible that they could use a proper whitelist, which would eliminate all the problems. But that doesn't mean there aren't systems out there using black lists that let unknown stuff slip through.
# [16:18] <Philip`> Lachy: There might be systems that use blacklists that block some unknown stuff (like attributes starting with "on") but not all other unknown stuff (like <handler> elements)
# [16:20] * Philip` doesn't know if there really are any, but it doesn't seem too preposterous
# [16:20] * Joins: robburns (n=robburns@adsl-75-34-51-221.dsl.chcgil.sbcglobal.net)
# [16:22] * Joins: robburns_ (n=robburns@adsl-69-215-34-30.dsl.chcgil.ameritech.net)
# [16:24] <Lachy> It would be irresponsible to assume that there aren't any such vulnerable systems. It's safer to accept their existence as a possibility
# [16:24] * Joins: mstange (n=markus@aixd3.rhrk.uni-kl.de)
# [16:25] * Quits: robburns_ (n=robburns@adsl-69-215-34-30.dsl.chcgil.ameritech.net) (Client Quit)
# [16:31] * Joins: billmason (n=bmason@ip102.unival.com)
# [16:35] * Quits: sverrej (n=sverrej@59.94.248.196) (Read error: 104 (Connection reset by peer))
# [16:36] * Joins: sverrej (n=sverrej@59.94.248.196)
# [16:38] * Quits: sverrej (n=sverrej@59.94.248.196) (Remote closed the connection)
# [16:42] * Joins: sverrej (n=sverrej@59.94.248.196)
# [16:43] * Quits: maikmerten (n=merten@ls5dhcp196.cs.uni-dortmund.de) (Remote closed the connection)
# [16:45] * Quits: robburns (n=robburns@adsl-75-34-51-221.dsl.chcgil.sbcglobal.net) (Read error: 110 (Connection timed out))
# [17:01] * Joins: zdobersek (n=zan@cpe-92-37-66-250.dynamic.amis.net)
# [17:04] * Quits: Maurice (n=ano@a80-101-46-164.adsl.xs4all.nl) ("Disconnected...")
# [17:08] * Joins: MikeSmith (n=MikeSmit@198.202.202.21)
# [17:15] * Quits: pauld (n=pauld@217.41.236.36)
# [17:20] * Quits: myakura (n=myakura@p3020-ipbf505marunouchi.tokyo.ocn.ne.jp) ("Leaving...")
# [17:21] * Joins: dglazkov (n=dglazkov@nat/google/x-2a290d9ede3d48d5)
# [17:31] * Joins: pauld (n=pauld@92.40.228.102.sub.mbb.three.co.uk)
# [17:35] * Joins: rubys (n=rubys@cpe-075-182-092-038.nc.res.rr.com)
# [17:37] * Quits: mstange (n=markus@aixd3.rhrk.uni-kl.de) ("ChatZilla 0.9.84-2009010213 [Firefox 3.2a1pre/20090210022740]")
# [17:41] * Joins: pauld_ (n=pauld@92.40.66.77.sub.mbb.three.co.uk)
# [17:45] * Parts: erlehmann (n=erlehman@86.59.25.121)
# [17:45] * Joins: zalan (n=kvirc@dsl54007086.pool.t-online.hu)
# [17:49] <JohnResig> Lachy: it's running in IE8 now: http://ejohn.org/apps/selectortest/
# [17:49] * Quits: zalan_ (n=kvirc@dsl54007086.pool.t-online.hu) (Read error: 60 (Operation timed out))
# [17:49] <JohnResig> Lachy: I'm getting 45.9% passing
# [17:52] * Quits: pesla\work (n=retep@procurios.xs4all.nl) ("( www.nnscript.com :: NoNameScript 4.21 :: www.esnation.com )")
# [17:52] * Joins: mlpug (n=mlpug@a88-115-168-225.elisa-laajakaista.fi)
# [17:52] * Joins: pauld__ (n=pauld@92.40.97.75.sub.mbb.three.co.uk)
# [17:53] <Lachy> JohnResig, ok. Can you mail publc-webapps and let them know the result
# [17:53] <JohnResig> Lachy: sure
# [17:55] <Lachy> it looks like many of those failures are due to lack of support for CSS3 selectors. But there are still a few worrying ones with the API
# [17:55] <Lachy> what are the "Whitespace Trim" tests testing?
# [17:56] <JohnResig> Lachy: I add extra whitespace characters around the selector (spaces, tabs, etc.) all of which should be trimmed
# [17:56] <JohnResig> (according to the spec)
# [17:56] <JohnResig> I'm sure if they fixed that it would solve a lot of problems - along with the implementation of proper exceptions
# [17:56] <JohnResig> that seems to be the majority
# [17:57] * Joins: dave_levin (n=dave_lev@c-98-203-247-78.hsd1.wa.comcast.net)
# [17:58] <Lachy> ok, that's what I though. I was sure that issue was pointed out to them long ago. Their inconsistent handling of whitespace of one fo the reasons I discovered and specced that anyway.
# [17:58] <Lachy> s/of one fo/was one of/
# [18:02] * Quits: pauld (n=pauld@92.40.228.102.sub.mbb.three.co.uk) (Connection timed out)
# [18:04] * Quits: MikeSmith (n=MikeSmit@198.202.202.21) ("Tomorrow to fresh woods, and pastures new.")
# [18:05] * pauld__ is now known as pauld
# [18:05] <JohnResig> Lachy: email sent
# [18:05] <Lachy> thanks
# [18:06] * Joins: smedero (n=smedero@mdp-nat251.mdp.com)
# [18:10] * Joins: Maurice (i=copyman@5ED548D4.cable.ziggo.nl)
# [18:17] * Quits: pauld_ (n=pauld@92.40.66.77.sub.mbb.three.co.uk) (Connection timed out)
# [18:17] * Joins: dglazkov_ (n=dglazkov@nat/google/x-33a59ad24f650167)
# [18:17] * Quits: dglazkov (n=dglazkov@nat/google/x-2a290d9ede3d48d5) (Read error: 60 (Operation timed out))
# [18:25] * Quits: pauld (n=pauld@92.40.97.75.sub.mbb.three.co.uk) (Read error: 60 (Operation timed out))
# [18:26] * dglazkov_ is now known as dglazkov
# [18:30] * Quits: Lachy (n=Lachlan@pat-tdc.opera.com) ("This computer has gone to sleep")
# [18:34] * Quits: weinig (n=weinig@c-67-180-35-124.hsd1.ca.comcast.net)
# [18:40] * Joins: pauld (n=pauld@host86-134-106-105.range86-134.btcentralplus.com)
# [18:49] * Joins: virtuelv (n=virtuelv@95.34.27.22.customer.cdi.no)
# [18:50] * Joins: svl (n=me@ip565744a7.direct-adsl.nl)
# [18:51] * Quits: pauld (n=pauld@host86-134-106-105.range86-134.btcentralplus.com)
# [18:59] * Joins: dbaron (n=dbaron@c-98-234-51-190.hsd1.ca.comcast.net)
# [18:59] * Joins: pauld (n=pauld@host86-134-106-105.range86-134.btcentralplus.com)
# [19:09] * Joins: Lachy (n=Lachlan@85.196.122.246)
# [19:12] * Joins: tantek (n=tantek@c-67-161-5-143.hsd1.ca.comcast.net)
# [19:14] * Quits: pauld (n=pauld@host86-134-106-105.range86-134.btcentralplus.com) ("Gone for a burton")
# [19:17] <yecril71> I feel quite comfortable writing XML, especially in an editor that supports autoclose.
# [19:17] <yecril71> And reading XML is much easier then reading something purportedly more readable,
# [19:18] <yecril71> like bash script or wikitext, or even C++.
# [19:18] <yecril71> The languages that rely on parentheses for nesting are unreadable.
# [19:19] <jcranmer> LISP?
# [19:19] <jcranmer> sorry
# [19:19] <jcranmer> LISP *patooie*
# [19:20] <yecril71> Python tries not to, but only for statements, not for expressions.
# [19:20] <yecril71> OTOH, Python is unwritable (at least for me).
# [19:24] <yecril71> So XML is a jolly good fellow, in spite of the rumours that some people are trying to spread :-)
# [19:27] * Joins: dimich (n=dimich@72.14.227.1)
# [19:32] * Quits: dave_levin (n=dave_lev@c-98-203-247-78.hsd1.wa.comcast.net)
# [19:38] * Quits: mpt (n=mpt@canonical/launchpad/mpt) ("Leaving")
# [19:43] * Quits: jwalden_ (n=waldo@c-76-102-160-171.hsd1.ca.comcast.net) ("inbound")
# [19:47] * Joins: maikmerten (n=maikmert@Lab49.l.pppool.de)
# [19:48] * Joins: zalan_ (n=kvirc@dsl54007086.pool.t-online.hu)
# [19:56] <dimich> Is XmlHttpRequest specification a part of HTML5 or some other specification? I can't find it in http://www.whatwg.org/specs/web-apps/current-work ...
# [19:56] * Parts: annevk5 (n=annevk@pat-tdc.opera.com)
# [19:56] * Joins: jruderman (n=jruderma@c-76-102-160-171.hsd1.ca.comcast.net)
# [19:56] * Joins: dave_levin (n=dave_lev@72.14.227.1)
# [19:57] <svl> dimich: it started out there, but is now at http://www.w3.org/TR/XMLHttpRequest/
# [19:57] <dimich> svl: thx!
# [20:06] * Quits: zalan (n=kvirc@dsl54007086.pool.t-online.hu) (Read error: 110 (Connection timed out))
# [20:12] * Joins: jwalden_ (n=waldo@corp-246.mountainview.mozilla.com)
# [20:12] * jwalden_ is now known as jwalden
# [20:13] * Joins: annevk5 (n=annevk@85.196.122.246)
# [20:26] * Joins: blooberry (n=brian@c-67-188-123-89.hsd1.ca.comcast.net)
# [20:29] * Joins: mstange (n=markus@aixd3.rhrk.uni-kl.de)
# [20:32] * Joins: shepazu (n=schepers@m140436d0.tmodns.net)
# [20:39] * Quits: jruderman (n=jruderma@c-76-102-160-171.hsd1.ca.comcast.net)
# [20:42] * Parts: rubys (n=rubys@cpe-075-182-092-038.nc.res.rr.com)
# [20:48] * Joins: ojan (n=ojan@72.14.229.81)
# [21:06] * Joins: shepazutoo (n=schepers@m140436d0.tmodns.net)
# [21:07] * Quits: shepazu (n=schepers@m140436d0.tmodns.net) (Read error: 104 (Connection reset by peer))
# [21:08] * Quits: mlpug (n=mlpug@a88-115-168-225.elisa-laajakaista.fi) (Remote closed the connection)
# [21:09] <annevk5> dimich, you want http://dev.w3.org/2006/webapi/XMLHttpRequest/ and http://dev.w3.org/2006/webapi/XMLHttpRequest-2/ probably
# [21:09] * Quits: tantek (n=tantek@c-67-161-5-143.hsd1.ca.comcast.net) (Read error: 60 (Operation timed out))
# [21:13] * Quits: smedero (n=smedero@mdp-nat251.mdp.com) (Read error: 104 (Connection reset by peer))
# [21:14] * Joins: tantek (n=tantek@c-67-161-5-143.hsd1.ca.comcast.net)
# [21:14] <dimich> annevk5: thanks! I see, those are the latest ones :-)
# [21:15] <annevk5> yup
# [21:34] * ojan is now known as ojanLunch
# [21:35] <tantek> re: hit-testing and transparency, since this effects event handling, is this something HTML5 says (should say or does say) something about? real world case today, the "Don't Click" viral attack that occurred on Twitter this morning (PST). exploit source here for your inspection: http://pastie.org/387315
# [21:36] * Quits: dbaron (n=dbaron@c-98-234-51-190.hsd1.ca.comcast.net) ("8403864 bytes have been tenured, next gc will be global.")
# [21:36] * Joins: ap (n=ap@194.154.88.32)
# [21:37] <Philip`> Hixie: Why does rowspan=0 make cells very tall, when IE just ignores rowspan=0 entirely and makes it 1 row tall?
# [21:38] <Philip`> (Presumably it can't be needed for compatibility, and so it's unnecessary complexity)
# [21:38] * Quits: mstange (n=markus@aixd3.rhrk.uni-kl.de) ("ChatZilla 0.9.84-2009010213 [Firefox 3.2a1pre/20090210022740]")
# [21:42] <Philip`> Oh, I suppose it's there because HTML4 said that's what rowspan=0 means
# [21:43] <Philip`> Firefox already ignores it in quirks mode
# [21:44] * Quits: dolske (n=dolske@firefox/developer/dolske)
# [21:44] <Philip`> Safari 3.2 and IE6 and IE8 always ignore it (in quirks and standards)
# [21:44] <Philip`> Opera 9.6 never ignores it
# [21:49] * Quits: jwalden (n=waldo@corp-246.mountainview.mozilla.com) ("reboot time")
# [21:54] * Quits: shepazutoo (n=schepers@m140436d0.tmodns.net) (Connection timed out)
# [21:55] * Quits: yecril71 (n=giecrilj@piekna-gts.2a.pl)
# [21:55] * Joins: jruderman (n=jruderma@corp-246.mountainview.mozilla.com)
# [21:55] * Quits: ROBOd (n=robod@89.122.216.38) ("http://www.robodesign.ro")
# [22:00] * Joins: dbaron (n=dbaron@corp-246.mountainview.mozilla.com)
# [22:04] * Joins: dolske (n=dolske@corp-246.mountainview.mozilla.com)
# [22:07] * Joins: zalan (n=kvirc@dsl540070CA.pool.t-online.hu)
# [22:08] * Joins: jwalden (n=waldo@corp-246.mountainview.mozilla.com)
# [22:09] * Quits: zalan_ (n=kvirc@dsl54007086.pool.t-online.hu) (Read error: 110 (Connection timed out))
# [22:13] * Quits: zdobersek (n=zan@cpe-92-37-66-250.dynamic.amis.net) ("Leaving.")
# [22:15] * Quits: maikmerten (n=maikmert@Lab49.l.pppool.de) (Remote closed the connection)
# [22:20] <Hixie> lol someone on reddit told me to look up the acronym "css"
# [22:21] * Quits: tantek (n=tantek@c-67-161-5-143.hsd1.ca.comcast.net) (Read error: 113 (No route to host))
# [22:24] * Joins: weinig (n=weinig@17.203.15.158)
# [22:27] * Joins: roc (n=roc@202.0.36.64)
# [22:28] * Joins: nessy (n=nessy@203-158-43-142.dyn.iinet.net.au)
# [22:30] * Joins: doublec (n=chris@202.0.36.64)
# [22:30] <gsnedders> Hixie: where?
# [22:32] <Hixie> http://www.reddit.com/r/worldnews/comments/7wr4i/new_eu_rule_requires_all_web_servers_to_log_ip/c07mfrb?context=3
# [22:33] * Quits: heycam (n=cam@210-84-15-128.dyn.iinet.net.au) ("bye")
# [22:33] <gsnedders> Hixie: Yeah, I just found that by finding your profile by googling for hixie reddit :P
# [22:34] <gsnedders> Hixie: Also, re:top gear, I too have see things about filming shots for the races after the actual race
# [22:34] <Hixie> yeah i found a blog entry about it later
# [22:34] <Hixie> see a comment i made
# [22:36] <Philip`> Top Gear faking shots? That'd never happen :-(
# [22:36] <gsnedders> Well, they already faked a death.
# [22:36] <gsnedders> And then revealed that the dead lived.
# [22:37] <Hixie> it's not really faking shots
# [22:37] <Hixie> i mean the shots are real
# [22:37] <Hixie> and what they depict did happen
# [22:37] <Philip`> Well, faking the context of shots
# [22:38] <Hixie> that's what tv is pretty much all about :-)
# [22:38] <Philip`> I remember some people commenting on http://www.dailymail.co.uk/tvshowbiz/article-471541/BBC-admit-Top-Gear-caravan-blaze-fake.html actually having faked shots (simulating big fires by putting little fires in front of the camera)
# [22:38] <gsnedders> This is why live TV is interesting.
# [22:38] <gsnedders> A challenge of how well you can do it.
# [22:38] <Philip`> (Not commenting on that article, just on the event which that article describes)
# [22:38] <annevk5> Yet another series of e-mail with someone from QA about namespaces and attributes; I've enough of this fricking RDF tax
# [22:38] * Quits: nessy (n=nessy@203-158-43-142.dyn.iinet.net.au) ("This computer has gone to sleep")
# [22:39] <annevk5> Someone please take Namespaces in XML around the back and shoot it.
# [22:40] <Hixie> jesus, svg 1.2 tiny makes elements focusable based on _whether there is an event listener for DOMFocusIn_
# [22:40] <Hixie> not even on whether the event handler doesn't cancel the event or anything
# [22:42] * Philip` saw a recent remake of The Quatermass Experiment which was broadcast live, and it's probably the first (and only) live-broadcast drama show he's ever seen
# [22:43] <Philip`> (It worked pretty well, except for some occasional rubbish sound balancing and a couple of wrong lines)
# [22:44] <gsnedders> Philip`: Does "The Bill" count as drama? AFIAK they've done special live versions of that
# [22:44] <Philip`> gsnedders: It probably does, but I've not watched it since about ten years ago :-)
# [22:45] <gsnedders> I've never watched it :)
# [22:47] <Hixie> so in svg... if you hook up an event listener to catch events for a group of elements (e.g. on a <g>)
# [22:47] <Hixie> the element suddenly becomes focusable itself
# [22:47] <Hixie> good lord
# [22:47] <Hixie> that's gotta make debugging svg a pain in the ass
# [22:47] * Joins: shepazu (n=schepers@m1e0436d0.tmodns.net)
# [22:48] <gsnedders> and then shepazu joins, right on cue.
# [22:49] <Hixie> heh
# [22:49] <Hixie> shepazu: any idea where in svg 1.2 tiny it says that focus is lost when the element is removed?
# [22:53] * Quits: svl (n=me@ip565744a7.direct-adsl.nl) ("And back he spurred like a madman, shrieking a curse to the sky.")
# [22:53] * Joins: roc_ (n=roc@202.0.36.64)
# [22:53] * Quits: shepazu (n=schepers@m1e0436d0.tmodns.net) ("Core Breach")
# [22:55] * weinig is now known as weinig|bbl
# [22:55] * Joins: doublec_ (n=chris@202.0.36.64)
# [22:58] * Quits: roc (n=roc@202.0.36.64) (Read error: 110 (Connection timed out))
# [22:58] * Quits: doublec (n=chris@202.0.36.64) (Read error: 110 (Connection timed out))
# [23:06] * Quits: zalan (n=kvirc@dsl540070CA.pool.t-online.hu) ("KVIrc 3.4.0 Virgo http://www.kvirc.net/")
# [23:14] * Quits: Maurice (i=copyman@5ED548D4.cable.ziggo.nl) ("Disconnected...")
# [23:14] * Joins: roc (n=roc@202.0.36.64)
# [23:23] <Hixie> aaaah!
# [23:23] <Hixie> pimpmyspec.net broke back compat!
# [23:24] * Quits: doublec_ (n=chris@202.0.36.64) (Read error: 110 (Connection timed out))
# [23:24] <annevk5> ooh, it broke the WHATWG credo?
# [23:25] <Hixie> jesus, this is going to be a pain
# [23:25] * Hixie tries to work out what options he wants based on the html source he's looking at
# [23:25] * Joins: heycam (n=cam@clm-laptop.infotech.monash.edu.au)
# [23:26] <Hixie> jesus what a lot of options
# [23:26] * Joins: doublec_ (n=chris@202.0.36.64)
# [23:26] <Hixie> i hope there are good defaults
# [23:27] <Hixie> it never ends!
# [23:27] <Hixie> bad jgraham
# [23:27] <Hixie> bad!
# [23:27] <Hixie> :-P
# [23:28] <Hixie> well whatever options i picked were the wrong options, clearly
# [23:28] <Hixie> this is a 64000 line diff
# [23:29] <Hixie> i wonder what i need to change
# [23:29] <Hixie> let's try w3c_compat_xref_a_placement
# [23:30] * Joins: slightlyoff (n=slightly@nat/google/x-adb15e441420d900)
# [23:30] <Hixie> ok that helped, 44000 lines now
# [23:31] <Hixie> it stopped omitting tags, hmm
# [23:31] * Hixie looks
# [23:31] * Quits: roc_ (n=roc@202.0.36.64) (Read error: 110 (Connection timed out))
# [23:31] <Hixie> aha, omit_optional_tags
# [23:32] * Joins: weinig (n=weinig@17.244.18.151)
# [23:32] <Hixie> maybe i want lxml_html
# [23:33] <Hixie> oh yes, much faster
# [23:35] * ojanLunch is now known as ojan
# [23:36] <Hixie> no that didn't work either
# [23:36] <Hixie> wtf
# [23:36] <Hixie> man this is a pain
# [23:36] <Hixie> i wonder how to get back to what it was like before
# [23:37] <Philip`> Try using http://www.google.com/search?q=cache:mEF8fH15-joJ:pimpmyspec.net/
# [23:38] <Hixie> uh huh
# [23:38] <Hixie> i can't get it to act as before!
# [23:38] * Hixie cries
# [23:38] <Philip`> What kinds of differences does it have?
# [23:40] <Hixie> e.g.
# [23:40] <Hixie> - </style><link href=status.css rel=stylesheet><script>
# [23:40] <Hixie> + </style>
# [23:40] <Hixie> + <link href=status.css rel=stylesheet>
# [23:40] <Hixie> + <script>
# [23:40] <Hixie> or:
# [23:40] <Hixie> - </script><body class=draft onload=init()>
# [23:40] <Hixie> + </script>
# [23:40] <Hixie> + </head>
# [23:40] <Hixie> + <body class=draft onload=init()>
# [23:41] <Hixie> and:
# [23:41] <Hixie> - <h2 class="no-num no-toc" id=draft-recommendation-—-date:-01-jan-1901>Draft Recommendation — 12 February 2009</h2>
# [23:41] <Hixie> + <h2 class="no-num no-toc" id=draft-recommendation-\342\200\224-date:-01-jan-1901>Draft Recommendation \342\200\224 12 February 2009</h2>
# [23:42] * Joins: shepazu (n=schepers@m140436d0.tmodns.net)
# [23:43] <Philip`> How odd
# [23:44] <annevk5> oops, clearly there should've been stable.pimpmyspec.net or something (or this should've been on unstable.pimpmyspec.net)
# [23:46] <Philip`> That sounds like a lot of effort compared to just waiting until jgraham comes back and fixes it
# [23:46] * Quits: weinig|bbl (n=weinig@17.203.15.158) (Connection timed out)
# [23:47] * Quits: virtuelv (n=virtuelv@95.34.27.22.customer.cdi.no) ("Leaving")
# [23:47] * Quits: ap (n=ap@194.154.88.32)
# [23:48] * Parts: billmason (n=bmason@ip102.unival.com)
# [23:49] * Joins: virtuelv (n=virtuelv@95.34.27.22.customer.cdi.no)
# [23:51] <heycam> Hixie, what's the difference between http://svn.whatwg.org/webapps/index and http://specs/web-apps/current-work/
# [23:51] <Hixie> the second one doesn't exist
# [23:51] <heycam> oops
# [23:51] <Hixie> or is that a trick question?
# [23:51] <Hixie> :-P
# [23:52] <heycam> insert an "www.whatwg.org/" in there
# [23:52] <Hixie> oh
# [23:52] <Hixie> the former is the svn repo of the latter
# [23:52] <Hixie> the latter is my working directory
# [23:52] <gsnedders> ergh. almost an hour between trains.
# [23:52] <heycam> ok, but i can be reasonably up to date by just "svn up"ing the former, and reading index?
# [23:52] <Hixie> i edit current-work/working-copy, then when i run my update script it gets copied over to current-work/source and regenned to current-work/index, and when i commit it goes to svn.whatwg.org and dev.w3.org
# [23:53] <Hixie> yes
# [23:53] <heycam> ok i see
# [23:53] <heycam> thanks
# [23:53] <Hixie> there are three levels; last stable checkin, what i last saved and generated, and what i literally am typing right now
# [23:53] <Hixie> svn is the first of those three
# [23:53] <Hixie> (as is w3c cvs)
# [23:53] <Hixie> if you're looking at the source of the doc, you want the /source file not /index
# [23:54] <Hixie> /index has all the cross-references, etc
# [23:54] <heycam> right. i just want an offline copy of the spec, so reading index is what i need, i think.
# [23:54] <Hixie> k
# [23:54] <Hixie> ok well since i can't work on the spec until jgraham gets back, i guess i'll go to work and get some food
# [23:54] <gsnedders> Hixie: pms.net down?
# [23:54] <Hixie> gsnedders: see above
# [23:54] * Philip` hopes jgraham hasn't been hit by a bus, because that'd be the end of the entire HTML5 endeavour
# [23:55] <gsnedders> ah
# [23:55] <Hixie> right, back in a bit
# [23:56] * gsnedders hopes he can find the right bus stop to get off at tomorrow
# [23:57] * gsnedders also he doesn't have such a bad cold
# Session Close: Fri Feb 13 00:00:00 2009

The end :)