So as we were finishing the robot and testing it out, we encountered an issue with the Victor 884s for our drive train. Two of them weren’t working. I swapped the PWMs for them since we’re using tank drive, and now only one works. (At competition, I may be swapping a bunch of PWMs as that had seemed to have worked in the past.)
To be prepared for our competition next month, we might purchase a Victor Sp to replace it, but would it be safe to have one Victor Sp connected to a Cim motor and a Victor 884 connected to another Cim motor, both motors attached to the same gearbox?
If this was the Victor 888, I’d personally have to hesitation in doing so. But with the Victor 884s non-linearity, and the SPs linearity, I personally would not do it as both motors are mechanically connected, but not necessarily moving at the same speed with the same input.
You could do something fancy by artificially making your Victor SP non-linear through code, but that’s a hassle.
Yes it would be safe, but it would not be ideal. The 884 has a much different response curve than the SP, so the motors would not be sharing the load equally and one might heat up more than the other.
If they are connected to the same gearbox they will indeed be moving at the same speed. But they will not be sharing the load equally.
So I also assume it’s not very recommended to have one side of the drivetrain to be powered by two Victor Sps and the other side powered by two Victor 884s?
That would probably annoy your drivers no end.
No, it’s not.
Again, safe but not ideal. One side would have a different response curve than the other, resulting in one side moving at a different speed while you’re applying partial power.
Right, that’s what I meant, thanks. 
This statement bothered me - what do you mean by “in the past”? If you mean “last month”, OK. On the other hand, if you mean “last year”, you do realize that the control system has been completely replaced this year, and last year’s control system is no longer legal?