test_bessel.c: bessel_Jn_e(0, 20000)

test_bessel.c: bessel_Jn_e(0, 20000)

[Yoshiki TSUNESADA]

 I encountered a glitch: A test code fails, in "specfunc/test_bessel.c", 
line 61,

  TEST_SF(s, gsl_sf_bessel_Jn_e, (0, 20000, &r), 
          -7.676508175684157103e-06, TEST_TOL4, GSL_SUCCESS);
          
The "expected" value -7.67651e-06 is exactly same as the one for
the test of line 60, the preceding line,

  TEST_SF(s, gsl_sf_bessel_Jn_e, (1, 1.0e+10, &r), 
          -7.676508175684157103e-06, TEST_TOL4, GSL_SUCCESS);

and this test passes. I guess the expected value in the line 61 
is mis-pasted. The following is an example session using Ruby/GSL:

   yoshiki@poirot[71] irb
   irb(main):001:0> require("rbgsl")
   => true
   irb(main):002:0> GSL::Sf::bessel_Jn(0, 20000)
   => 0.00556597490495494
   irb(main):003:0> GSL::Sf::bessel_Jn(1, 1.0e10)
   => -7.67650817568416e-06
 
 I don't have any other tools to compare and verify these results.
 
 Best regards,
 Yoshiki

Re: test_bessel.c: bessel_Jn_e(0, 20000)

[Brian Gough]

At Mon, 24 Sep 2007 22:58:10 -0400,
Yoshiki TSUNESADA wrote:
> 
>  I encountered a glitch: A test code fails, in "specfunc/test_bessel.c", 
> line 61,
> 
>   TEST_SF(s, gsl_sf_bessel_Jn_e, (0, 20000, &r), 
>           -7.676508175684157103e-06, TEST_TOL4, GSL_SUCCESS);

Thanks, I will correct the values.

In the GSL source these lines are commented out with #ifdef so they
shouldn't normally cause a failure when running 'make check'.

-- 
Brian Gough