Eric Giguère

I see it as `v4` being wrong instead of `dev.major`. It should not truncate small values with an hidden tolerance. Here it's the sparse matrix operations which drop small values...

We have a similar issue for a while in #1800. It felt through the cracks since we are focusing on v5.

Fixed in version 5

No, not yet, it will be supported in v5. You could try to use `photocurrent_mesolve` which is very similar, but with a slower algorithm. ps. There is an error in...

V5 is now out. It support using a Liouvillian instead of Hamiltonian in mcsolve.

Code in `__init__.py` is executed on import, including printing and raising error. But I don't know if we can differ between import methods (`import qutip` vs `from qutip import *`)