On 8/2/10 11:48 PM, Valery Pipin wrote:
> On ubuntu maverick, with sbcl-1.40
> I met as follows,
>
> running tests in rtest16:
> ********************** Problem 383 ***************
> Input:
> block([numer : true], is(abs(zeta(%i + 3)
> - (1.10721440843141 - .1482908671781754 %i)) <
> 1.e-15))
>
>
> Result:
> false
Known problem. The issue is that with sbcl (expt 2d0 3) returns a
double-float number that only has single-float accuracy. This also
affects ecl and ccl but not cmucl and clisp. Ccl has this fixed
recently, and I talked to someone who said they want to fix this in sbcl.
Ray