- From: Ben Adida <ben@adida.net>
- Date: Tue, 10 Jun 2008 21:16:39 -0700
- To: RDFa <public-rdf-in-xhtml-tf@w3.org>
Hi all, I finally found the time to finish my Ruby cleanroom implementation, and I can now say with great certainty that the spec is fully correct. Well okay, I didn't quite finish in the sense that I have *not* tried to get the XMLLiterals right, in terms of namespaces and such. All tests other than the XMLLiterals pass, and I've inspected the failed tests by hand, and the only deviation is related to proper namespaces within XMLLiterals. (Specifically, test 11,92,94,100,101,102,103.) I wrote this cleanroom implementation straight from the spec, without trying to optimize anything. I made a few stupid Ruby mistakes, and I missed two points in the spec that I then had to fix: (A) I forgot to *not* set skip=true when @property was present, Step 4. (B) I forgot to clone the evaluation context when skip=true, Step 11. Interestingly, issue (A) was very quickly visible with multiple test failures. Issue (B), on the other hand, I didn't notice until the very end. As it turns out, only test 84 failed as a result. In any case, I can say now with great certainty that the spec is fully correct. If you're interested, the code in all of its ugliness is at: http://www.w3.org/2006/07/SWD/RDFa/impl/ruby/ (ignore the html5lib, which I'm deleting. You need it, but just get the gem.) and I have it running at: http://ben.adida.net/rdfa/ruby/rdfa.cgi?uri= Note that, to run the tests, I used triplr.org for conversion: http://triplr.org/ben.adida.net/rdfa/ruby/rdfa.cgi?uri= -Ben PS: and as it turns out, the HTML5 parser *does* pick up @rev, the bug was in *my* code.
Received on Wednesday, 11 June 2008 04:17:18 UTC