Here's another version that removes a lot of the unused imports. I have to confess I've just skimmed it, I'll take a closer look at the base classes (where all the real action is) tomorrow.
I guess my question for everyone is what the consensus is here? My reservation is mostly that this touches a lot of files, and we're trying to cut 5.0 next week. I'm reluctant to change this much code this close to setting the tag for 5.0. This seems more like style/cleanup than fixing outstanding problems, so I'm thinking that waiting until the label is set for 5.0 and putting this in 5.1 might be the right thing to do here, assuming there are no objections. Waiting will make resolving conflicts even more "interesting" since Alan is going to be in the test code for removing old-style solr.xml....
That said, if some of the heavy cloud hitters (e.g. Mark Miller Alan Woodward and Noble Paul) think differently I'll be happy to commit this after the precommit problem is fixed (see below).
The precommit target fails with:
/test-framework/src/java/org/apache/solr/BaseDistributedSearchTestCase.java:87: warning - @ShardsFixed(num is an unknown tag.
That's not a big deal though, I can ask someone who knows how to fix that where to look. LuceneTestCase? Anyway, if we go forward we can fix that I'm sure.
BTW, I'm "Git challenged", and last time I tried to pull down a patch I somehow got the wrong stuff, thanks for providing it! Generally we like to have the patch rather than go to a repo so the JIRA is self-contained, we never know when a repo will go away.