08:31
<virtuelv>
http://politics.reddit.com/info/6181z/comments/c02ijhx <-- can't we morph to april 1st and add <sicko>
08:35
Lachy
notes that he's currently living in "The World's Best Country to live in" (Norway), but actually disagrees with it.
08:35
<Lachy>
Australia is actually the best country to live in ;-)
08:37
<othermaciej>
my favorite country to live in is California
08:37
<othermaciej>
unfortunately we remain under US occupation
08:37
<othermaciej>
but even so it's quite nice
08:42
<hsivonen>
othermaciej: I figured the "Designed by Apple in California" thing was a not so subtle way of distancing California from the rest of the US
08:44
<othermaciej>
hsivonen: nah, it's just because the idea of California has semiotic resonance with Apple's brand identity
08:45
mpt
is momentarily distracted by Wikipedia's "Californian independence" article
08:45
<hsivonen>
othermaciej: my Leopard box puts California and Ireland on the same depth in the implied taxonomy tree and doesn't mention the US or the EU at all
08:45
<othermaciej>
I don't think the EU considers itself a country yet
08:47
<hsivonen>
(by box I didn't mean computer but the packaging the Leopard DVD came in)
08:48
<hsivonen>
"Designed by Apple in California" "Assembled in Ireland"
08:50
<hsivonen>
back to RFCs: do I understand correctly that the generic scheme-unaware IRI restrictions merely ban some code points after the scheme and that's it?
08:55
<othermaciej>
I don't know
08:55
<othermaciej>
so far IRIs have not been relevant to actual web content so I haven't learned what that RFC says yet
08:57
<hsivonen>
hmm. interesting PUA characters are banned in the non-query part of the IRI
09:27
<hsivonen>
yay! javascript: IRIs have different character rules before and after a question mark in the script
09:27
<zcorpan>
hsivonen: ?
09:30
<hsivonen>
zcorpan: now I'm not sure if I'm reading the RFC in the wrong way of if the Jena IRI lib has a bug
09:31
<hsivonen>
zcorpan: but it seems to me that private use characters are allowed in the query part of the IRI per grammar but not in the path part
09:32
<hsivonen>
zcorpan: but the Jena IRI lib gives an error in both cases anyway...
09:44
<hsivonen>
Why does draft-hoehrmann-javascript-scheme-00 say: 'literal use of the character "/" should be avoided'?
09:49
<othermaciej>
hsivonen: probably to avoid looking like a hierarchical URL
09:49
<othermaciej>
but I don't think it is a problem in practice
09:52
<hsivonen>
othermaciej: I don't see why looking like a hierarchical URL would be a problem even in theory
09:53
<othermaciej>
hsivonen: you'd have to ask Bjoern
09:53
<hsivonen>
yeah
09:53
<othermaciej>
he usually has reasons for what he does
10:06
<hsivonen>
hmm. are there generic IRI processors that take a base IRI and a relative IRI and process them to produce a new absolute IRI that represents the relative IRI absolutized relative to the base IRI?
10:06
<hsivonen>
is there any non-contrived scenario where a javascript: IRI might participate in such processing?
10:08
<othermaciej>
I dunno, in what cases are things actually processed as IRIs?
10:09
<hsivonen>
I don't know.
10:20
<hsivonen>
will a browser ever consider an HTML document to have a javascript: URI as its base URI?
10:26
<mitsuhiko>
hsivonen: i have no idea what you're talking about but "/" is discouraged in javascript in an html context because it could erroneous close a script tag
10:27
zcorpan
added date and datetime text to the wiki
10:27
<hsivonen>
mitsuhiko: I am talking about javascript: IRIs
10:28
<mitsuhiko>
hmm. makes no sense then :)
10:28
<hsivonen>
mitsuhiko: that is, whether javascript:alert(4/2) is valid without percent escaping the /
10:48
<virtuelv>
hsivonen: works in at least two browsers
11:08
<hsivonen>
virtuelv: ok. (I expected unescaped form to work in browsers)
11:09
<virtuelv>
javascript:4/2 both outputs 2 in Opera and FF
11:14
<hsivonen>
I emailed Björn to ask for the rationale of the restriction
18:37
<Hixie>
hsivonen: i got a prince license. what should i do to set up prince as part of the spec creation pipeline any proposals?
18:45
<hsivonen>
Hixie: http://hsivonen.iki.fi/printing-wa10/
18:46
<hsivonen>
Hixie: first, you should get DejaVu working with Prince
18:46
<hsivonen>
Hixie: then perhaps diff this http://hsivonen.iki.fi/printing-wa10/specification-a4.css with your current style sheet and add my stuff in @media print
18:47
<hsivonen>
Hixie: might be better to leave the paper size to command line
18:47
<hsivonen>
Hixie: and run it once with A4 and once with Letter
18:48
<hsivonen>
Hixie: of and the entity table needs a something to select outside it unless Prince has added ::outside to tables while I wasn't looking
18:48
<hsivonen>
s/of/oh/
18:53
<Hixie>
DejaVu?
18:54
<hsivonen>
Hixie: http://dejavu.sourceforge.net/wiki/index.php/Main_Page
18:55
<hsivonen>
If you are running Prince on Linux, you might already have it out of the box if Prince finds the fonts that came with the distro
18:55
<hsivonen>
If Prince fails to find glyphs for all the characters you put in the spec, it'll tell you
18:56
<Hixie>
no fonts came with this distro
18:56
<Hixie>
i wonder how i install this dejavu thing into my local user
18:57
<hsivonen>
Hixie: you can install it in Prince's own font directory if you like
18:59
<Hixie>
wow, dejavu has good coverage of the pages i need
18:59
<hsivonen>
Hixie: I spent some time googling for a font that had WARNING SIGN
19:00
<gsnedders>
html5lib's repo has an external to itself, but it uses the https:// URL which only works for project members
19:00
<hsivonen>
(Code200* crashed Cocoa or something)
19:00
<Hixie>
i don't have mscorefonts either
19:04
<Hixie>
good lord these corefonts are hard to get easily
19:05
<Falen>
Is there any open-source HTML parser in C? To convert HTML->Text, and read tags.. etc..
19:05
<hsivonen>
Hixie: both Ubuntu and suse make it easy
19:06
<Hixie>
i don't have root on this debian box
19:11
<Hixie>
man the lengths i go to sometimes
19:12
Hixie
downloaded the .EXEs, downloaded the source to cabextract and compiled it, extracted the TTFs...
19:13
<Hixie>
btw the fonts changed names since you did the patch
19:18
<Philip`>
Falen: There's a tokeniser in C++, which might not be quite what you want but at least it exists :-)
19:22
<webben_>
Falen: libtidy is in C ... you might be able to adapt that
19:24
<Philip`>
Oh, I forgot about non-HTML5 HTML parsers - that would presumably make it easier
19:58
<Hixie>
hsivonen: http://damowmow.com/a.pdf
19:59
<Hixie>
hsivonen: my stylesheet had changed quite a bit so diffing it didn't work well, but i tried to take what i thought was important
19:59
<Hixie>
hsivonen: any advice? (prince in particular seems to have a problem with indenting with <dd>s?)
20:07
<hsivonen>
Hixie: the space between the left page edge and the body box should probably be much narrower
20:08
<hsivonen>
Hixie: also, I'd get rid of the colors and link underlines
20:09
<hsivonen>
whoa page 78 is not good
20:10
<hsivonen>
Hixie: I didn't have the dd problems, but I serialized from firefox
20:10
<jgraham_>
gsnedders: Yeah I should just fix that (there was a reason for it but the side effect is unacceptable)
20:11
<hsivonen>
Hixie: in case Prince doesn't do end tag inference right, parsing and serializing with html5lib would probably help
20:12
<hsivonen>
Hixie: it appears you omitted the bit that made the entity table multicolumn and thus more compact
20:14
<hsivonen>
Hixie: table in section 5.3.2 needs special font size attention
21:30
<ddfreyne>
Hm, question.
21:30
<ddfreyne>
The 'b' and 'i' elements... should these be interpreted as purely presentational, or do they have some semantic value as well?
21:31
<Lachy>
ddfreyne, see http://lachy.id.au/log/2007/05/b-and-i
21:31
<ddfreyne>
I'm thinking purely presentational; you'd use them where you would usually use bold and italic in non-hyper text, am I right?
21:31
<ddfreyne>
thanks
21:32
<ddfreyne>
sweet, that's exactly what I was thinking
21:34
<Philip`>
You'd use them where you would usually use bold and italic in non-hyper text if otherwise some meaning would be lost, but not where you'd use them just because they're pretty
21:34
<ddfreyne>
yeah, exactly
21:34
<ddfreyne>
I was just thinking of blog posts marked up like "<b><i>hello, blah blah blah</i></b>"
21:35
<Philip`>
(Woah, I only just noticed that X3D doesn't use a namespace)
21:37
<Hixie>
hsivonen: the left side gap is intentional (for issue notes and handwritten margin notes)
21:38
<Hixie>
hsivonen: the colours are also intentional (for colour printing) and the link underlines are useful to know exactly what is being hyperlinked
21:38
<Hixie>
though i suppose i could remove them
21:39
<Hixie>
hsivonen: fixed the column thing
21:39
<Hixie>
kinda
21:40
<ddfreyne>
I was just thinking of blog posts marked up like "<b><i>hello, blah blah blah</i></b>"
21:40
<ddfreyne>
err.
21:40
<ddfreyne>
wrong window :)
21:41
<Hixie>
i'm missing a bunch of glyphs still
21:43
<Hixie>
hsivonen: regenning a.pdf, take a look in a few minutes
21:44
<Philip`>
(Actually, the specification has no namespace, and a non-zero number of implementations produce non-namespaced content, but a non-zero number has "taken a few liberties until a firm specification is defined" and accepts content in the http://www.web3d.org/specifications/x3d-namespace namespace)
21:44
<Philip`>
(which seems kind of a suboptimal situation)
21:57
<Lachy>
http://html5.lachy.id.au/ has finally been restored. I believe I have found the issue and that it was actually caused by other php scripts on my server, not these ones.
22:59
<jgraham_>
gsnedders: I fixed the svn:externals
23:01
gsnedders
hugs jgraham_ for doing so (randomly)
23:03
jgraham_
feel more appreciated than normal for typing a few simple commands :)