Closed malb closed 7 years ago
We should wait with merging until the possible unallocated memory issue is resolved. set_random_seed
doesn't allow to reproduce it?
Its actually not such a huge bug. It is just a rather awkward behaviour. When nothing is found two things can happen: 1/ (a, b) = (0.0, (0.0, ... 0.0)) 2/ (a, b) = (random, ())
So I guess initialized memory is involved in case 2, yet nothing too deep I'd guess. Not sure if this should be fixed in fplll or fpylll.
For any practical purpose, the attached pickle files contains a triple
(B, radius, pruning)
where case 2/ happened (pruned enumeration on the whole basis B).
buugy_inst.pkl.zip