08:14
<hsivonen>
I turned comments off on the feed autodiscovery entry on the blog because it attracts spam
10:25
<annevk>
zcorpan_, I decided not to parse end tags in the same crazy way
10:25
<annevk>
zcorpan_, but more like how XML handles them
10:26
<annevk>
zcorpan_, I suppose XML5 will be backwards incompatible with XML 1.0, although in theory we could handle <:> maybe...
10:29
<annevk>
Philip`, fixed one of your bugs with entity handling (&lt;)
10:34
<zcorpan_>
annevk: how can it be more like XML when it is drocanian?
10:35
<annevk>
it's not draconian...
10:36
<zcorpan_>
1.0
10:36
<annevk>
oh, more like what would be natural handling for end tags given the production for end tags in XML
10:36
<zcorpan_>
ah. ok
10:38
<annevk>
zcorpan_, you mentioned something about encoding problems yesterday?
10:39
<zcorpan_>
annevk: that was re http://krijnhoetmer.nl/irc-logs/whatwg/20070729#l-145
10:39
<annevk>
oh, too bad
10:40
<annevk>
I was hoping that would explain the mysterious failure messages on xml5/play
11:01
<zcorpan_>
why isn't <figure><object data=...>foo</object><legend>...</legend></figure> conforming?
11:07
<zcorpan_>
or actually... why are block-level children of object even allowed when child of figure
11:09
<zcorpan_>
aha. a <figure> doesn't always represent a paragraph
12:13
annevk
can't figure out why "| |" makes the server act strangely
13:26
<zcorpan_>
http://www.search-this.com/2007/07/30/html5-tables/
14:25
<annevk>
hmm, you can never render tables in one pass
14:27
<annevk>
also, didn't <col> imply <colgroup> in HTML4 or is that just browsers?
14:28
<Philip`>
"The COLGROUP element ... Start tag: required, End tag: optional" - sounds like it's not meant to be implied
14:31
<zcorpan_>
just internet explorer, actually, iirc
14:31
<zcorpan_>
but that's not relevant, because <tbody> is also implied and it is not required
14:33
<annevk>
prolly makes sense to make it optional, yes
20:09
<zcorpan_>
there is a Mobile Web Initiative Test Suites Working Group?
20:10
<zcorpan_>
isn't developing test suites pretty central for a WG?
20:17
<zcorpan_>
their tests use <![CDATA[... wonder what mobile browsers that don't use xml parsers do with that
21:14
<zcorpan_>
how does one check the presence of a getter in JS?
21:15
<zcorpan_>
i.e. an Element.prototype.__defineGetter__()
21:23
<Philip`>
__lookupGetter__?
21:27
<zcorpan_>
ah! cheers
21:31
<jgraham>
Philip`: Do you have all your <canvas> tests somewhere?
21:31
jgraham
wants to point out "tests available" as a feature of <canvas>
21:32
<Philip`>
http://canvex.lazyilluminati.com/tests/tests/ is the most recent online version
21:33
<Philip`>
(I have a few more I haven't got around to uploading yet)
21:34
<Philip`>
(Also, that particular page hits two CSS bugs in WebKit which is a bit irritating and I suppose I ought to work around that at some point)
21:35
<jgraham>
Thanks
21:39
<zcorpan_>
http://simon.html5.org/sandbox/js/elementtraversal.js
21:42
<Philip`>
Is it just me, or is <canvas> much more fun to use than SVG?
21:42
<Philip`>
I don't really know much about SVG since I've not used it much, but I've not used it much because it never looked like any fun :-)
21:43
<zcorpan_>
dunno, i have played very little with svg and practically nothing with canvas
21:43
<jgraham>
I know basically nothing about either SVG or <canvas>
21:44
<jgraham>
But it's pretty clear that <canvas> should be in the HTML 5 spec because it's not going away
21:44
<jgraham>
and it's a simpler solution to a narrower problem than SVG
21:45
<zcorpan_>
who can find bugs in my ElementTraversal implementation? :)
21:47
<Philip`>
The if in "if (e) { while (e) { ... } }" is redundant
21:47
<zcorpan_>
ha, indeed
21:48
<zcorpan_>
fixed
21:56
<Philip`>
zcorpan_: Is that script meant to work if the UA already implements firstElementChild/etc?
21:57
<zcorpan_>
Philip`: the intent is that if the UA already has implemented it then the script is ignored
22:04
<Philip`>
Ah, I didn't know FF returned stuff from __lookupGetter__ on native properties
22:04
<Philip`>
(Does anything else implement __lookupGetter__ at all?)
22:04
<Philip`>
"For the purpose of ElementTraversal, an entity reference node which represents an element must be treated as an element node." - your code doesn't do that
22:05
<zcorpan_>
indeed
22:05
<Philip`>
(as far as I can tell)
22:06
<zcorpan_>
mozilla doesn't put entity reference nodes in the dom though
22:06
<zcorpan_>
perhaps you can insert them manually
22:12
<Philip`>
Setting things like e.childElementCount=1 isn't handled the same as other settings of readonly properties
22:12
Philip`
doesn't see any more serious issues
22:13
<Philip`>
(That is, any issues which are more serious; not any more issues which are serious)
22:15
<othermaciej>
I know a lot about both <canvas> and SVG
22:15
<othermaciej>
and I think they are both kinda neat
22:15
<othermaciej>
there is a big overlapping range of use cases, as well as some problems where one or the other is much better suited
22:29
<jgraham>
othermaciej++ (re: public-html)
22:54
<Philip`>
Ooh, I never knew Safari did <input type=slider>
22:54
<Philip`>
Uh
22:54
<Philip`>
Ooh, I never knew Safari did <input type=range>
22:55
<Philip`>
but it appears to not handle non-integer steps, which is annoying
23:22
<Philip`>
Why is &#13; a parse error? It sounds like a fairly legitimate thing to do
23:32
<zcorpan_>
Philip`: good question
23:33
<zcorpan_>
wonder how &#10;&#13; is treated in browsers
23:34
<zcorpan_>
(or is it the other way around?)
23:44
<zcorpan_>
http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C%21DOCTYPE%20html%3E%0D%0A%3Cpre%3Ea%26%2313%3B%26%2310%3Bb
23:49
<zcorpan_>
http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C%21DOCTYPE%20html%3E%0D%0A%3Cpre%3Ea%26%2313%3B%26%2310%3Bb%3Cscript%3E%0D%0A%20var%20node%20%3D%20document.getElementsByTagName%28%22pre%22%29%5B0%5D.firstChild%3B%0D%0A%20node.data%20%3D%20node.data.replace%28/%5Cr/g%2C%20%22%5C%5Cr%22%29.replace%28/%5Cn/g%2C%20%22%5C%5Cn%22%29%3B%0D%0A%3C/script%3E
23:51
<zcorpan_>
ie: &#13;&#10; becomes \r. opera: it becomes \r\n but only one linebreak is rendered. safari, firefox: it becomes \n\n (as per html5)
23:51
<zcorpan_>
(unless i'm misreading the spec)
23:54
<PlayPause>
!seen nickshanks