The new code looks clean and very well written. I made a test program comparing the gsl_bspline_eval_deriv output directly with PPPACK's bvalue() routine, which is an independent implementation of bspline derivatives (gsl_bspline_eval_deriv is based on bsplvd()).
The results look fine. I also ran it through valgrind with good results. I think it can be put into the next release. Patrick On Mon, Nov 24, 2008 at 03:37:57PM +0000, Brian Gough wrote: > At Wed, 19 Nov 2008 16:50:47 -0700, > Patrick Alken wrote: > > > > I'm moving this discussion to the gsl-discuss list. Good work! > > This is something several people have been asking for and I haven't > > had any time to work on. > > > > I will take a look at the patch in the next few days. > > Patrick, I'm happy for you to take the decision on this wrt the 1.12 > release. The main thing we want to avoid at this stage is breaking > the build in ways that we might not detect in our own testing. > > -- > Brian Gough
