This is a public chat log generated from the #semsol IRC channel.
09:54:59
bengee ping
09:55:26
pong
09:55:32
;)
09:55:54
yeah, the problem is that XML Literals are coming out different from how they go in
09:56:43
but isn't that ok in UTF-8 contexts?
09:57:03
I dunno
09:57:26
it's not ok for me, because I can't roundtrip xmlliterals
09:57:31
keeping the ents might break rdf/xml, now?
09:57:40
i'm not sure what the ideal behaviour would be
09:57:54
s/now/no/
09:57:55
but this looks wwrong: I put in: <div><br/> <p>hello world</p></div>
09:58:12
and get out <div><br></br> <p>hello world</p></div>
09:58:58
don't think I can do anything about that, I'm just passing through what the XML parser gives me, I think
09:59:13
yeah, that's what i was afriad of :(
10:03:17
if you don't type the literal, but make it a cdata section, it should be kept as is
10:04:07
rss1-style
10:06:45
yeah
10:07:15
it's shame you have to know to do that though
10:21:35
what if ARC were to manually seek out xmlliterals, check for well-formedness, and then encase in a cdata section
10:22:35
I think the markup is entity-decoded at this stage already
10:22:58
ah - you're streaming the parsing eh?
10:23:08
yes
10:23:56
why did they make it so f**kin difficult :p
10:25:10
well, the xml parser *has* to decode ents, otherwise it'd have to remember doctypes, too
10:25:31
yeah
10:26:08
i mean rdf/xml seems ridiculously difficult to write a perfect parser for
10:26:23
oh, that one. yeah, true
10:26:44
and the xmlliterals is the hardest part
10:27:15
well - maybe not!
10:27:34
collections are painful
14:13:31
bengee: you got mail - you can also reply here if you like
14:42:53
bengee: I think I've fixed it
14:43:49
i just added $d = htmlspecialchars($d); to the start of function h6Cdata($d)
14:45:01
hey danbri
14:45:07
did you hear back from peter mika?
14:58:50
bengee: that should of course be: $d = htmlspecialchars($d, ENT_NOQUOTES);