Fwd: [sage-devel] elliptic_e(0.5, 0.1) test failure on Solaris SPARC (error about 0.04%)



I got a separate message about this problem, here's what I said about it.
In summary, the problem is not in Maxima from what I can tell.

Robert Dodier

---------- Forwarded message ----------
From: Robert Dodier <robert.dodier at gmail.com>
Date: Sun, 9 Aug 2009 14:34:36 -0600
Subject: Re: [sage-devel] elliptic_e(0.5, 0.1) test failure on Solaris
SPARC (error about 0.04%)
To: William Stein <wstein at gmail.com>

On 8/9/09, William Stein <wstein at gmail.com> wrote:

>     sage: elliptic_e(0.5, 0.1)
> Expected:
>     0.498011394499
> Got:
>     0.497801100392

w/ 5.19.0 (official release + Clisp on Linux) I get .4980113944988315
which I guess is the same as the expected result to
as many digits as were given.

CVS Maxima from last week + ECL + Linux yields the same result.

My first guess is that some basic functions in ECL punt
to C libraries which give different results on Solaris.

Bug reports and what-not are probably better directed to the Maxima
mailing list than to me personally. If you're not a member of the list
it just goes into a moderation queue, which is emptied on a regular
basis if I'm not mistaken.

HTH

Robert