The source project of this merge request has been removed.
support for double precision added, errmsg on wild precision values
Constraint set on input values and error messages thrown if the precision values is neither 32 nor 64, this was not possible with the DPD_FLOATTYPE=double flag.