David Ronis wrote:
> Just updated from cvs/trunk on a i686-linux box, running clisp-2.47. I
> had the following testsuite failure:
>
>
> Running tests in rtest_gamma:
> ********************** Problem 134 ***************
> Input:
> closeto(double_factorial(10.0), 3840.0, 1.0E-12)
>
>
> Result:
> 1.364242052659392E-12
>
> This differed from the expected result:
Interesting. This works for me with clisp 2.47 (on Solaris, though).
A peek at double-factorial doesn't show anything out of the ordinary to
cause this loss of precision.
Ray