FieldRead timeslice issues

[rokuingh]

An ESMPy user found some issues with the timeslice parameter to FieldRead. He sent a reproducer, and used xarray to validate the results. The IO related concerns here are:

• there is no bounds check on timeslice, it fills extra memory with zeros

• the timeslice seems to be returned incorrectly

• there is no way to return a timeslice that does not start with time=0 (e.g. 2-4 or just time 5)

The reproducer can be run with:

mpirun -n 4 python vtu2ugrid.py

timeslicereproducer.tar.zip

[rokuingh]

@Chrismarsh @billsacks