Use vectors instead of points to move_charges_inside_semiconductor

[hervasa2]

Unsupported operations (since #489) with CartesianPoint's were used in move_charges_inside_semiconductor causing errors when drifting charges (see #552)

I switched them to vectors by using point -> point - cartesian_zero.
This PR fixes these issues and correctly passes verbose when using move_charges_inside_semiconductor

simulate_waveforms( mcevents_table::AbstractVector{<:NamedTuple}, sim::Simulation{T}

[fhagemann]

Can we add a test for this, to make sure that things like this are also tested in the future and covered by the tests?
(Just put a very large charge cloud very close to a surface in any of the tests, maybe the diffusion tests?)

[fhagemann]

Downgrade tests failed because the current method of computing a weighted mean of CartesianPoints requires a change introduced in StatsBase@0.33.14 --> JuliaStats/StatsBase.jl@9916e14

[hervasa2]

Thanks for adding the tests. I think this is good to go.