test failure with 5.16.3 / ecl (Bessel)



I have observed this problem working on the abs function too. But I have not
understand the reason.

I have formulated the function closeto() more simple as:

(%i2) func(f):=block([numer:true],abs(f))$

I think the above function should numerically evaluate the expression f because
of setting the flag $numer. But it does not:

(%i3) func(bessel_j(3,10));
(%o3) bessel_j(3, 10)

But it works for the function cabs:

(%i6) func(f) :=block([numer:true],cabs(f))$

(%i7) func(bessel_j(3,10));
(%o7) 0.058379379305187

Then I have tried the following:

(%i8) func(f) :=block([numer:true],f)$

This expression don't numerically evaluate too:

(%i9) func(sin(10));
(%o9) sin(10);

I don't understand why it works within the testsuite (GCL and CLISP) and for
cabs but not in other cases.

Dieter Kaiser

-----Urspr?ngliche Nachricht-----
Von: maxima-bounces at math.utexas.edu [mailto:maxima-bounces at math.utexas.edu] Im
Auftrag von Raymond Toy
Gesendet: Freitag, 29. August 2008 20:30
An: Oliver Kullmann
Cc: maxima at math.utexas.edu
Betreff: Re: [Maxima] test failure with 5.16.3 / ecl (Bessel)

Oliver Kullmann wrote:
> Hi,
> 
> when building maxima (5.16.3) with ecl I get
> 
> Running tests in rtest14:
> ********************** Problem 135 ***************
> Input:
> closeto(AS_10_4_17(- 10), 1.e-15)
> 
> 
> Result:
>                2  20 sqrt(10)             2  20 sqrt(10)
> 10 (bessel_j(- -, -----------) - bessel_j(-, -----------))
>                3       3                  3       3
> ---------------------------------------------------------- + airy_dai(- 10)
>                             3
> 

I see this on my Mac OS X build.  I don't see it on my Solaris build.  I 
do not know why.

Ray
_______________________________________________
Maxima mailing list
Maxima at math.utexas.edu
http://www.math.utexas.edu/mailman/listinfo/maxima