I've resurrected the XSLT conformance status page (Julian Scheid used to host it but needed to make some space).
http://www.zen60362.zen.co.uk/xslt/ Green is a pass, yellow is a fail, red is an error (sometimes fatal e.g. StackOverflowError). Stats are at the bottom of the results page. These are only the Xalan conformance tests, the Microsoft ones are much more complicated to run and diff against. For comparison, here are the overall results for Xalan: Total: 1670 Successes: 1624 Failures: 38 Errors: 8 Average transform time: 57ms Conformance: 97.2455084323883% As you can see we are a LOT faster, but there are still lots of failing tests. Most of them are because we don't produce exactly the same output as Xalan. I'm using the Python xmldiff tool to compute the diffs, and it's not quite as intelligent as it could be. Some examples: - differences in namespace prefixes (not significant in XML) - unused declared namespaces (we don't output namespace declarations that aren't used) - differences in character entity representations (e.g. 
 != ) Also we follow the XSLT spec rigorously by outputting a meta http-equiv element with the charset for all html output, which Xalan doesn't do under some circumstances (notably all-ASCII content). If anyone is interested in helping me hack on this, please let me know. -- Chris Burdess _______________________________________________ Classpath mailing list [email protected] http://lists.gnu.org/mailman/listinfo/classpath
