Raymond Toy wrote (on Fri, 20 May 2005 at 16:46 -0400):
> >>>>> "Raymond" == Raymond Toy writes:
>
> Raymond> In any case, I see that it's failing on this test:
>
> Raymond> (c2:3^(-1/3)/gamma(1/3),closeto(-airy_dai(0)-c2,1.0e-15));
>
> Raymond> And airy_dai(0),numer causes an error about NIL not being a
> Raymond> double-float.
>
> Raymond> At least with cmucl.
>
> These and a few other minor problems preventing the test suite from
> finishing have been fixed. They all stem from the simple problem that
> CMUCL is fairly picky about things like
>
>
> (let (z)
> (declare (double-float z))
> (setf z 1d0)
> ...)
>
> because Z is initialized to NIL.
>
> Ray
Oh yes, sbcl and cmucl are a bit picky, thats why it is good to test
with those lisps :-)
With the new code in cvs there is no more error thrown, but the
teststuite still is not happy and give sthe following output:
Running tests in rtest14.mac:
********************** Problem 111 ***************
Input:
closeto(AS_10_4_10(%i - 1), 1.0e-15)
Result:
1
airy_ai(%i - 1) airy_dbi(%i - 1) - airy_bi(%i - 1) airy_dai(%i - 1) - ---
%pi
This differed from the expected result:
true
********************** Problem 113 ***************
Input:
closeto(AS_10_4_10(- %i - 1), 1.0e-15)
Result:
airy_ai(- %i - 1) airy_dbi(- %i - 1) - airy_bi(- %i - 1) airy_dai(- %i - 1)
1
- ---
%pi
This differed from the expected result:
true
141/143 tests passed.
The following 2 problems failed: (113 111)
I'm too tired to looko what the reason is for that - I nedd a cup of
sleep :-(
Good Night,
'Andreas
--
Wherever I lay my .emacs, there's my $HOME.