Re: ODE solver

[Brian Gough <bjg at network-theory dot co dot uk>]

Ivo Alxneit writes:
 > so you mean
 > gsl_odeiv_system sys_without = {func2, NULL, 2, &mu2};
 > in my code?

Yes

 > in this case i get the expected segfault. (question: shouldn't 
 > gsl_odeiv_evolve_apply and related functions return an error and abort 
 > gracefully?).

Normally we don't check for null pointers because there would be so
many places to test.  I think you can replace the jacobian with a
function that causes a graceful failure.

 > i assume that i was correct on that only bsimp uses the jacobian 
 > matrix/function (maybe state in the manual that only algorithms where it is 
 > explicity mentionned make use of the jacobian).

I think this is in the manual somewhere, but probably not very prominent.

-- 
Brian Gough
(GSL Maintainer)

Network Theory Ltd,
Commercial support for GSL --- http://www.network-theory.co.uk/gsl/