Check for incosistent use of betas in track

[rdemaria]

Differently from twiss, track does not support the delta_s parameter:

https://github.com/MethodicalAcceleratorDesign/MAD-X/blob/8695bd422dc403a01aa185e9fea16603bbd5b3e1/src/trrun.f90#L3348

where deltae is pt from the start command

https://github.com/MethodicalAcceleratorDesign/MAD-X/blob/8695bd422dc403a01aa185e9fea16603bbd5b3e1/src/mad_track.c#L295

while deltap is taken from the delta parameter in track:

https://github.com/MethodicalAcceleratorDesign/MAD-X/blob/8695bd422dc403a01aa185e9fea16603bbd5b3e1/src/trrun.f90#L93

  deltap = get_value('probe ','deltap ')
  betas  = get_value('probe ','beta ')
  gammas = get_value('probe ','gamma ')
  dtbyds = get_value('probe ','dtbyds ')
  arad   = get_value('probe ','arad ')
  radiate  = get_value('probe ','radiate ') .ne. zero

  thin_cf =  get_option('thin_cf ').ne.zero
  bet0    =  get_value('beam ','beta ')

  bet0i   = one / bet0
  beti    = one / betas

  deltas  = get_variable('track_deltap ')

In this example the tracking with deltap=0 and deltap!=0 should give the same results, but they don't

length=2.1;

ss: sequence, l=length;
    b: drift, at=length / 2, l=length;
endsequence;


beam, particle=proton, pc=0.1, sequence=ss, radiate=FALSE;
use, sequence=ss;
dpp=0.1;one=1;
bet0i=1/beam->beta;

track, onepass, onetable,deltap=dpp;
start, pt=0.1;
run, turns=1;
endtrack;

track, onepass, onetable,deltap=0;
start, pt=0.1+sqrt((one+dpp)^2+bet0i^2-one)-bet0i;
run, turns=1;
endtrack;


track, onepass, onetable,deltap=dpp;
start, px=0.1;
run, turns=1;
endtrack;

track, onepass, onetable,deltap=0;
start, px=0.1,pt=0.0+sqrt((one+dpp)^2+bet0i^2-one)-bet0i;
run, turns=1;
endtrack;