Re: [Help-gsl] Incomplete elliptic integral E(phi, k) decreasing with phi?

[Frank Reininghaus]

Hi Liam,

thanks for the source code.

On Saturday 22 March 2008 22:15:55 Liam Healy wrote:
> I'm not sure what you mean by "seem to be correct also";
> do you mean you got my results, or you got the correct results?

I meant that I get the correct results.

Indeed, I also get the correct results with your program (using GSL 1.9 on 
Kubuntu 7.10):

phi = 0.6*M_PI
answer = 1.740987787999229264

phi = M_PI
answer = 2.934924418678854785

phi = -0.5 * M_PI
answer = -1.467462209339427393

phi = -0.49999999*M_PI
answer = -1.467462182132591275


By looking at the mailing list archives, I found out that this is indeed a bug 
that was fixed in version 1.9, see

http://lists.gnu.org/archive/html/info-gsl/2007-02/msg00000.html

You should update to the current version (1.10), then everything should work 
fine. The source code is available at

ftp://ftp.gnu.org/gnu/gsl/gsl-1.10.tar.gz

Regards,
Frank