Correct handling of seeds for MC on the fly trains Adding AliGenerator::SetSeed which AliFatal by default It has to be overwritten in the deriving classes Implemented in PYTHIA6, 8, Hijing, AMPT Change of signature in TUHKMgen which already had SetSeed function
Chiara O. implemented a way to bypass HIJING internal calculation of the sigmaNN so that one can set the value from outside, and to switch off the elastic collisions. By default everything should be as before. Constantin L. then improved the calculation of the geometrical Npart, which now is very close to the MC Glauber one. It still differs because HIJING normalizes internally to collisions with the eikonal approach for which the geometrical Npart can be zero (if we would accept events for which the geometrical Npart > 0 then the result would be equal to the one of the MC Glauber).
Introducing the interaction time into the aliroot generators. In case of gaussian beam profiles the interaction time is generated with gaussian distribution with sigma=sigma_z/c. In case one wants for shift the mean time (for example to simulate satellite collisions or pileup from another bunch crossing) it is enough to call SetTimeOrigin method and provide the time origin value in seconds. More details in bug#73877.