On 8/8/08, Michael Lasarev <lasarevm at gmail.com> wrote:
> Running tests in rtest15:
> ********************** Problem 174 ***************
> Input:
> exp(1.0b-1)
>
> Result:
> 1.10517091807564763b0
>
> This differed from the expected result:
> 1.105170918075648b0
> It turns out not to be dependent on the processor speed, but
> only on *when* the test is run: If Maxima is re-started then
> all tests pass with no problems, but subsequent runs of the
> test produce the error, even if I have Maxima clear the
> memory (e.g., Maxima > clear memory > kill all).
I don't see that behavior in current CVS builds.
My first guess is that fpprec (precision for bigfloat computations)
was changed in some test script and not reset.
That doesn't seem to be an issue at the moment.
Incidentally reset(fpprec) changes fpprec to its initial value;
kill(all) doesn't have that effect.
You might try to encourage the package maintainers to
update to a recent version; there is a 5.16 release in the works.
HTH, & thanks for your interest in Maxima.
Robert Dodier