11 Mar
2010
11 Mar
'10
8:27 p.m.
On 9 Mar 2010, at 20:40, Story Henry wrote:
A very basic implementation of foaf+ssl is up here:
http://webid.myxwiki.org/
While fixing Chrome [0], we found a bug with my previous implementation of
webid.myxwiki.org, that made any services using http://foafssl.org/srv/idp unable to login
users.
Luckily Damian Steer, the author of the Java RDFa library [1] we are using, was here to
help me out of this. He found that the cause of the problem was the invalid XHTML I had
written on the xwiki homepages. I had placed a <div> inside a table, which the java
html5 parsers - which the rdfa parser feeds on - saw as an error, and corrected in such a
way as to make the RDFa graph invalid. I now replaced this with a <tbody> and it
works fine.
This goes to show the importance of validating ones XHTML early. Validating XHTML, which
is much cleaner than html, is going to be much easier, than working out the details that
html5 parsers have to deal with.
Of course when popping something in a validator it helps to do the work correctly.
It's easier to test for 0 bugs, than to decide which ones are legitimate and which
ones not. So for XHTML+RDFa development, it would help if one could put the following at
the top of the document
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML+RDFa 1.0//EN"
"http://www.w3.org/MarkUp/DTD/xhtml-rdfa-1.dtd">
Then the rdfa tags would no longer be flagged by http://validator.w3.org/
What is the best way to do this in XWiki?
Henry
PS. If you had trouble logging into http://foaf.me or http://nanoblog.me/ previously, try
again. It should work now.
[0] http://code.google.com/p/chromium/issues/detail?id=37765
[1] http://github.com/shellac/java-rdfa