Re: [Bug-gsl] GSL gsl_sf_hyperg_2F1_e

[Patrick Alken]

Brian, this is a known bug, but I forgot to comment out that test case 
in test_hyperg.c when I was working on this a while back. I just did a 
commit which should fix the issue for you

Patrick

On 04/30/2015 08:23 AM, Brian Gladman wrote:
> When building with MSVC, I am now seing a single test failure:
>
> GSL_IEEE_MODE="double-precision,round-to-nearest,mask-all"
> FAIL: gsl_sf_hyperg_2F1_e(3.23191, -4.0229, 8.02291, 0.5, &r) [2115]
>    unexpected return code: 11
>    4.300243900348178494e-01  2.655871889646902615e-06
> FAIL: Hypergeometric Functions [2144]
>
> with an unexpected return code of 11, indicating that the maximum number
> of iterations have been exceeded.
>
> Changing the number of iterations in the loop that compiles the series
> sum makes no difference, nor does changing the floating point model used
> to compile the library.
>
> The value returned doesn't look too bad when compared with the value
> Mathematica gives:
>
> gsl: 0.430024390034817|8494
> mm:  0.430024390034817|064619266095678
>
> so I am wondering if anyone has any bright ideas on what I might be able
> to do about this.
>
>     best regards,
>
>       Brian Gladman