This PR aims to fix compatibility with this breaking change. It would be possible for us to implement a proper GobApron.Environment.compare, but I don't know if we really need it.
I'm more worried about our comparison of Apron.Abstract1.t-s via Stdlib.compare, which won't work now. It doesn't seem to break our tests, but it might still be possible in some case we end up needing to call such comparison and crashing (which will explicitly break instead of being implicitly wrong sometimes).
https://github.com/antoinemine/apron/issues/99 is mitigated by https://github.com/antoinemine/apron/pull/108, which should be released as Apron v0.9.15 soon.
This PR aims to fix compatibility with this breaking change. It would be possible for us to implement a proper
GobApron.Environment.compare, but I don't know if we really need it. I'm more worried about our comparison ofApron.Abstract1.t-s viaStdlib.compare, which won't work now. It doesn't seem to break our tests, but it might still be possible in some case we end up needing to call such comparison and crashing (which will explicitly break instead of being implicitly wrong sometimes).