I thought the integrals for 4-component Hamiltonians are complex, but Dirac outputs only real integrals in FCIDUMP. Why is this so?
For certain symmetries, two-electron integrals become real anyway. What is your particular system ?
All the best,
Trond