I'll commit soon, and port to trunk too.
New patch, also moving the start/endTrackingSearchers to before/afterClass in SolrTestCaseJ4, instead of in init/deleteCore.
this looks good, maybe this was causing the test meddling that it didnt happen in before/afterClass?
Yes, I believe it was!
In fact there were two separate things – the first test didn't fail when it should have, and the 2nd test failed to call startTrackingSearchers itself (it inherited the start from the first test).
Bulk close for 3.2