It seems that fmpq_mat_can_solve_multi_mod has a bug which is masked by its use of 60-bit primes which are unlikely to fail for random matrices.
If one sets it to start with $p = 2$, it actually hangs for some input (make check MOD=fmpq_mat with FLINT_TEST_MULTIPLIER=10).
It seems that
fmpq_mat_can_solve_multi_modhas a bug which is masked by its use of 60-bit primes which are unlikely to fail for random matrices.If one sets it to start with$p = 2$ , it actually hangs for some input (
make check MOD=fmpq_matwithFLINT_TEST_MULTIPLIER=10).