Bug Fixes for Elements

• Fixed some problems with radiation integral calculations for dipoles under some conditions:
  • If a dipole was split into several pieces, edge contributions to the radiation integrals were inappropriately added at the interior ``edges.'' This might happen if element division was used, or if the lattice definition contained a series of adjacent dipole elements with the same name.
  • In dipoles with gradients, if $k l = \sqrt{\left\vert K_1+1/\rho^2\right\vert} l \ll 1$, the radiation integral calculations were unreliable. Fixing this necessitated using a power series expansion for $k l<0.01$, which may result in some very small changes in radiation integrals (relative to the previous version) for weak dipoles, even if there is no gradient.
• Made a correction to the expression for $I_5$ radiation integral for the WIGGLER element, having to do with propagation of the beta function.
• Fixed problem with ZTRANSVERSE, which used the wrong value for $Z(0)$ for a broad-band impedance.
• Fixed bugs in RFDF implementation for N_KICKS$>1$ and LENGTH$>0$.