rbend thick slicing issue

[rdemaria]

While investigating FCC lattice we found that with:

select, flag=makethin, class=rbend, slice = 1, thick = true;

An element declared with:

bc1: rbend,l:=lbc1 ,angle:=anglebc1 ,e1:=0 * anglebc1 ,e2:=0 * anglebc1 ;
bc1.1: bc1,ktap:= -0.000002307748712;

is transformed into

bc1: rbend,l:=lbc1 ,angle:=anglebc1 ,e1:=0 * anglebc1 ,e2:=0 * anglebc1 ,thick=true;
_bc1.1: bc1,l:=( lbc1 ) * ( 1.0 / sinc ( anglebc1 * 0.5 ) ) ,angle:=anglebc1 ,e1:=( 0 * anglebc1 ) + ( ( anglebc1 ) / ( 2 ) ) ,e2:=( 0 * anglebc1 ) + ( ( anglebc1 ) / ( 2 ) ) ,ktap:= -0.000002307748712,thick=true;

It looks like the code is transforming a thick rbend in a thick sbend without changing to an sbend.

In addition, the position of the dipole is changed from

bc1.1: bc1, at = 44.550043009893280,ktap:= -0.000002307748712;

to

_bc1.1, at = ( 44.550043009893 ) + ( ( ( lbc1 ) * ( 1.0 / sinc ( anglebc1 * 0.5 ) ) ) * ( 0 ) ) ;

This should not happen anyway because the center position should not be shifted regardless of the rbarc option.