fixes calculation of beta from beta0 issue

[martinschwinzerl]

In the tracking maps for the spacecharge elements,
beta has been calculated as

beta = beta0 / rvv

However, since rvv is defined as rvv = beta / beta0,

beta = beta0 * rvv

would be the correct way to calculate beta. This fixes inconsistencies between
SixTrackLib >= 0.6 and pysixtrack

[rdemaria]

The argument is correct. For simplecticity reasons we should only use beta0, although is less accurate.
I propose to change both pysixtrack and sixtracklib in this way.

[giadarol]

Indeed, beta should disappear and we should use beta0 everywhere in the space charge lenses.

As Riccardo says it is the only way of preserving symplecticity with a 4D lens.

Moreover keeping beta does not really improve the accuracy: the interaction time depends on the beta of the particle but depends also on the path length (which in turn depends on px, py). These two effects fight against each other in typical cases (positive dispersion), therefore it does not really make much sense to keep one while neglecting the other.