For a strategic customer it's absolutely critical that @PlanningVariable(nullable=true) works in OptaPlanner 6, even with multiple variables. This hasn't been tested in an integration test or an example yet. We need to do that.
I believe this bug will break CH's for that strategic customer: https://issues.jboss.org/browse/PLANNER-197 Note: they aren't using CH at the moment (so it's not a regression), but they are expecting to be able to use it in BRMS 6.0.
Integration test for single nullable variable added. https://github.com/droolsjbpm/optaplanner/commit/9a2584d9b0871b05885900d52e81cc290fb5d232 CH problem reproduced on that test and fixed with this commit: https://github.com/droolsjbpm/optaplanner/commit/3a27feb384aed7d546167c1bd6313a87aae5b94b The isMoveDoable() method still has the wrong name, so most of PLANNER-197 is still open.
The isMoveDoable() naming will be handled for 6.1. It does not prevent nullable=true from working correctly. So given this issue is fixed (see comment above that links the fix and the integration test to prove it works).
Verified with BRMS-6.0.0.ER7
qe_test_coverage: covered by community unit tests