[Gcl-devel] Re: gcl: acosh issues

[Camm Maguire]

Greetings!

"Stavros Macrakis" <address@hidden> writes:

> The system acosh does not agree on branch choice with the Maxima acosh
> for some arguments.  I don't know if anyone cares or which is better....
> 
>  (lisp::acosh -2) =>  1.3169578969248168 + 3.1415926535897931 i
>  (acosh -2)       => -1.3169578969248164 + 3.1415926535897931 i
> 
> gcl is following the Common Lisp definition here.
> 

OK, as you state, I think gcl is right here.

> There is also a bug in the system acosh:
> 
>  (lisp::acosh -1.0) => Zero divisor (should be i pi)
> 

Just committed a fix for this into cvs.  Thanks!

> Nearby values do just fine.
> 
> Macsyma gets this one right.
> 
> 
> 

-- 
Camm Maguire                                            address@hidden
==========================================================================
"The earth is but one country, and mankind its citizens."  --  Baha'u'llah