Resolve "Problem in Command DUMPEMFIELDS and DUMPFIELDS"

added Bug label

assigned to @gsell

added 3 commits

d13a0b7c - type of num_gridp{x,y,z}_m changed to long unsigned int
5a1644c7 - throw exception if denominator is close to zero
3a23a452 - FM3DH5BlockBase class: ensure that index not out of range

@ext-calvo_p @ext-rogers_c we still have the division by zero error in Cyclotron::apply(). Any idea to solve this?

I fear that the code and tracking is written with the assumption that particles don't have zero radius. There might be serious errors in the tracking in this case. I think the proposed solution of throwing an exception is valid. Perhaps @adelmann can comment.

One way to possibly fix it on the user side could be to change the input from

Z_START=-0.02, Z_STEPS=41, DZ=0.001

to

Z_START=-0.0201, Z_STEPS=41, DZ=0.001 ?

I do not understand "particles don't have zero radius".

We mostly deal with macro particles and indeed for tracking and field calculations thee particles are viewed as delta functions i.e. they have no spatial extend but indeed maybe a macro charge k*e.

Does that make sense?

Sorry for the misunderstanding and the unclarity. What I meant is that in the issue, particles tracked in a cyclotron with OPAL-cycl have R = 0, causing divisions by zero. I wonder if the tracking is properly handled even for small R, or if certain assumptions are made.