08:36
<Lachy>
OMG! It really doesn't help when John responds to a difference of opinion with hostility, and then wonders where the attitude problem comes from :-(
12:23
<zcorpan>
wow...
12:23
zcorpan
is skipping the print thread
12:25
annevk
did that too
12:26
<annevk>
seemed like a waste of time
12:37
<krijnh>
annevk: what have you done wrong on GoT?
12:38
<annevk>
nothing lately
12:38
<krijnh>
And before?
12:39
<annevk>
before that nothing much either as far as I can tell
12:39
<annevk>
pointer?
12:40
<krijnh>
http://gathering.tweakers.net/forum/list_message/28361582#28361582
12:40
<annevk>
Ik denk dat hij bedoelt dat ik er tegenwoordig niet meer langskom
12:40
<krijnh>
Ah, dat kan ook ja
12:41
<krijnh>
Ik dacht al :)
12:41
<krijnh>
Verwart 'ie je met Faruk ofzo? ;p
14:59
<annevk>
http://quuz.org/xml5/play
15:06
<Philip`>
Did something just change with DOCTYPE handling, or am I imagining things?
15:07
<annevk>
Where?
15:08
<Philip`>
http://quuz.org/xml5/play
15:08
<annevk>
Changed as opposed to what, in that case?
15:09
<Philip`>
It appeared to be giving an empty document if I had a <!DOCTYPE > anywhere, but now it doesn't do that, but I can't remember if I was just doing something wrong at first :-)
15:09
<annevk>
a start tag is required, fwiw
15:09
<krijnh>
annevk: text-align:justifyx ?
15:09
<annevk>
or empty tag
15:10
<annevk>
krijnh, ah, I'll drop that line
15:10
<krijnh>
Perhaps an Opera extension, to fix justify :)
15:11
<Philip`>
Is it intentional that <!DOCTYPE HTML PUBLIC "foo"'> isn't parsed the same as in HTML5?
15:11
<Philip`>
(I have no idea how XML would parse that)
15:11
<hasather>
Philip`: you need a system id
15:12
<annevk>
doctypes are currently not ending up in the tree
15:12
<annevk>
you should be able to use the internal subset to add entities and default attribute values though
15:15
<annevk>
maybe when I have some more time I'll do it through oninput and XMLHttpRequest which should safe everyone from pressing that silly button
15:16
<annevk>
and maybe also provide some other output serializations
15:24
<Philip`>
Oh, in XML it's just a parse error when it hits the ' instead of finding a space character
15:24
<Philip`>
(whereas HTML5 ignores the ' and ends the doctype on the >, and XML5 thinks it's the start of a single-quoted string and carries on reading until it finds a closing single-quote somewhere later)
15:25
<annevk>
I've yet to enable parse errors; error handling could be changed I suppose to make it more HTML5 like
15:27
<Philip`>
You could save everyone from pressing that button by just implementing it in client-side JavaScript instead of Python or whatever it is :-)
15:35
<annevk>
maybe I should reimplement it in ocaml
15:40
<Philip`>
That'd just be crazy :-p
15:45
<Philip`>
http://quuz.org/xml5/play?source=%3Chtml%3E%26lt%3B%26lt%3B%3C%2Fhtml%3E is weird
15:47
<annevk>
there are issues with entity handling, yeah, :(
15:47
<annevk>
I should fix that
15:50
<Philip`>
http://quuz.org/xml5/play?source=%3Chtml%3E%26amp%3C%2Fhtml%3E - that also looks like issues :-)
15:51
<annevk>
it shouldn't have emitted the final ;
15:52
<zcorpan>
"# [16:32] <Philip`> (whereas HTML5 ignores the ' and ends the doctype on the >, and XML5 thinks it's the start of a single-quoted string and carries on reading until it finds a closing single-quote somewhere later) " -- that's not how i read the html5 spec
15:53
<zcorpan>
<!DOCTYPE HTML PUBLIC "foo"'> is, afaict, parsed as if it was <!DOCTYPE HTML PUBLIC "foo" '>'>
15:55
<annevk>
yeah
15:58
<Philip`>
http://quuz.org/xml5/play?source=%7C+%3Ca%3E+%7C - "Service Temporarily Unavailable"
15:59
<Philip`>
zcorpan: Argh, I'm not sure how I missed that when actually testing it - maybe I didn't bother looking at the test output...
15:59
<Philip`>
But at least IE does ignore the ' if you do that
16:00
<Philip`>
(which is possibly what I was thinking of)
16:00
<zcorpan>
ie is possibly reparsing
16:01
<Philip`>
http://software.hixie.ch/utilities/js/live-dom-viewer/?a%3C%21DOCTYPE%20%22%22%22%3Eb%22%3Ec%0D%0Ad%3C%21DOCTYPE%20%22%22%20%22%3Ee%22%3Ef
16:02
<annevk>
Philip`, weird...
16:03
<annevk>
works fine if I try it directly on the server...
16:03
<zcorpan>
Philip`: oh. wow.
16:04
<Philip`>
IE never really parses the doctype - it just tries to jump through the quotes and work out where the end of the doctype is, then does string matches through the entire doctype content to decide if it should be quirks mode
16:05
<Philip`>
(See quirkiness of http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C%21DOCTYPE%20fooDOCTYPE%20NETSCfoo%3E vs http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C%21DOCTYPE%20fooDOCTYPE%20NETSfoo%3E )
16:12
<annevk>
Philip`, thanks for the bug reports btw, I'll look into them somewhere next week
16:14
<Philip`>
(Also http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C%21DOCTYPE%20DTD%20HTML%204%20http%3A//%3E works peculiarly in IE, when you remove a / or change the 4 into a 3 or 5 first)
16:25
<zcorpan>
can someone read hungarian? http://weblabor.hu/forumok/temak/18561
16:28
<Philip`>
http://www.tranexp.com:2000/InterTran?type=url&url=http%3A%2F%2Fweblabor.hu%2Fforumok%2Ftemak%2F18561&text=&from=hun&to=eng can't read it especially well
16:32
<zcorpan_>
http://quuz.org/xml5/play?source=%3C%21DOCTYPE+a%5B%3C%21ENTITY+a+%22%3Ca%2F%3E%22%3E%5D%3E%26a%3B
16:32
<zcorpan_>
that was not allowed in xml 1.0
16:34
<Philip`>
http://quuz.org/xml5/play?source=%EF%BF%BD - internal server error
16:37
<zcorpan_>
Philip`: no wonder, it decodes as windows-1252 instead of utf-8...
16:37
<zcorpan_>
gotta love tools that assume that there is a mapping between language and encoding
16:57
<zcorpan_>
annevk: the xml declaration is not part of the infoset in xml 1.0 i think
17:18
<zcorpan_>
annevk: <:> is interesting. per xml 1.0 without namespaces it is just an element with the name ":"
17:22
<zcorpan_>
annevk: http://quuz.org/xml5/play?source=%3Ca%3E%3C%2Fb+x%3D%22%3E%22%3E is different from html5
17:25
<Philip`>
http://quuz.org/xml5/play?source=%3Ca%3E%3Cb%3E%0D%0A%3Ca%2F%3E%3Cb%3E%0D%0A%3Ca%2F%2F%3E%3Cb%3E%0D%0A%3Ca%2F%2F%2F%3E%3Cb%3E%0D%0A%3Ca%2F%2F%2F%2F%3E%3Cb%3E seems odd
17:27
<zcorpan_>
<??>
20:13
<gavin>
hrm, are the list archives going to be fixed some time soon?
20:16
<hsivonen>
gavin: I wouldn't expect fixes to the whatwg infrastructure while Hixie is on vacation
20:17
<gavin>
ok
20:17
<gavin>
when is he back, again?
20:19
<hsivonen>
gavin: IIRC, he went away for three weeks, but I've lost track of the starting point of the three weeks
20:20
<gavin>
ok