I have (I hope!) fixed the logging in tests.
The Fuseki test support code was manipulating the logging directly so
that whether logging messages appeared could depend if the test class
ran before or after firing up a background Fuseki test server. Also,
there was a stray log4j.properties in the Fuseki main jar.
The jena-jdbc tests now use log4j.properties from test/resources for
setting logging.
A this point, things should be stable and I hope to call the VOTE later
this week.
If you want to start doing checking early, then you could do so now and
then just verify there have been no relevant changes.
Andy
Jenkins is having an off moment - I've done a build/deploy directly
timestamp around 20130909.203300 and later which for Fuseki is build #21.