Re: [Bug-gsl] bug in periodic csplines?

[Brian Gough]

Stefan Jahn writes:
 > i think i found a small bug in periodic cubic splines calculation.  for
 > the 3-point calculation you need to apply this patch:
 > 
 > cspline.c: cspline_init_periodic patch
 > 
 >     - const double h2 = xa[3] - xa[2];
 > 
 >     - g[1] = 3.0 * ((ya[1] - ya[2]) / h2 - (ya[2] - ya[1]) / h1);
 >     + g[1] = 3.0 * ((ya[1] - ya[2]) / h0 - (ya[2] - ya[1]) / h1);
 > 
 > can anyone verify that?  please answer privately as well, because i am
 > not subscribed to the list.

Thanks for the patch!  

I've verified the problem and that the patch corrected it.  I will
include it in the next release.  Sorry for the delay in replying.

-- 
best regards,

Brian Gough