Bug fix for the data.align_tensor.calc_R() function - divide by zeros are now avoided.
When Aa is zero, Ar is now set to NaN.
Authored by: bugman 2011-08-11
Parent: [r14362]
Child: [r14364]
