Unfortunately, that fix broke some of our functional tests (i.e. made Sat4j PB incorrect). We did not notice it because we ran only a subset of our functional tests for that last minute addition to the release (the whole functional tests take two days to run for Sat4j PB).
We noticed the bug and found the origin of that bug some weeks ago. The issue was limited to a few similar test cases, so we were not really worried about the quality of the 2.3.1 release.
However, we received yesterday the first report from a user that Sat4j PB was reporting an incorrect answer. So we decided to stop distributing Sat4j PB 2.3.1. Since Sat4j Maxsat is depending on Sat4j PB, we also stopped distributing it.
Release 2.3.2 is on its way, and should hit the forge by mid March.
In the meanwhile, we strongly encourage Sat4j PB and MAXSAT users to switch back to release 2.3.0.