Re: [Bug-gsl] Re: Interpolation

[Evgeny Kurbatov]

Hi Brian

> What do you think about returning NAN (instead of 0 and y0,yn) for
> values outside the range?  It's a more obvious failure.

You're right but I oriented on gsl_interp_eval_..._e functions and did
like that. Returning NaN is a good idea, though. It makes possible to
write a workaround to have any number as output instead of NaN.  I
attached the patch to original interp.c.

Regards,
Evgeny Kurbatov

interp.c.patch
Description: Text Data