integrate((1+tan(x)^2)/tan(x),x);



>>>>> "Gilles" == Gilles Schintgen  writes:

    Gilles> On Wednesday 23 November 2005 13:07, Nikos Apostolakis wrote:
    >> I get
    >> 
    >> log(sin(x)) - ((log(sin^2(x) - 1))/2
    >> 
    >> both with maxima 5.9.2 and with 5.9.1 in an older machine.

    Gilles> I did some testing:
    Gilles> maxima 5.9.1 cmucl:	log(tan(x))
    Gilles> maxima 5.9.1 gcl:	log(sin(x))-log(sin^2(x)-1)/2
    Gilles> maxima 5.9.1 clisp:	log(tan(x))
    Gilles> maxima 5.9.2 sbcl:	log(tan(x))
    Gilles> maxima 5.9.2 gcl:	log(sin(x))-log(sin^2(x)-1)/2

maxima cvs with clisp and cmucl gives the same answer as gcl.

This is totally bizarre.  Why should sbcl and gcl on the same version
and machine produce different answers?

Ray