Re: [Bug-gsl] Bug in gsl_blas_dnrm2

[Brian Gough]

At Fri, 12 Feb 2010 11:26:17 +0100,
Pierrick Bruneau wrote:
> Normally the norm should be 16.88, and gsl_blas_dnrm2 produces 10.00...
> Or maybe I missed something ?
> 
> Anyway, I attached a little source sample that reproduces this case, and 
> compares to what is obtained using gsl_blas_ddot.
> 
> [2 main.c <application/octet-stream (7bit)>]
> #include <gsl/gsl_vector.h>
> #include <gsl/gsl_cblas.h>
> #include <math.h>

Thanks for the email.  This is a problem with the program not GSL as
such -- you need #include <gsl/gsl_blas.h> because the function is not
declared in your program so will not be called correctly.  Compiling
with gcc -Wall will pick up this error.

See http://www.network-theory.co.uk/docs/gccintro/gccintro_19.html for
more information.

-- 
Brian Gough

GNU Scientific Library -
http://www.gnu.org/software/gsl/