Re: [Bug-gsl] missing "long_double" implementation for complex math

[Brian Gough]

Johannes Rauh writes:
 > I am using version 1.4 of GSL. When I recently tried to increase
 > the accuracy of my program by changing all complex variables from
 > simple gsl_complex to gsl_complex_long_double, I noticed that there
 > is indeed a type called gsl_complex_long_double, but on the other
 > hand none of the functions from gsl_complex_math.h can work with
 > this type (for example there is no function like
 > gsl_complex_long_double_add_real). The macros (e.g. GSL_REAL,
 > GSL_IM), however, work well.  Another thing I noticed ist, that it
 > is currently not possible to use the gsl routines for eigenvalue
 > calculations with long double precision.  Regards,

The gsl_complex_long_double types are intended for user data.

The GSL library routines always work in double precision only (it is
sufficient for most scientific purposes).

-- 
Brian Gough