Handling of $gammalim and $factlim in simpgamma

I have committed the suggested changes.

I think with this last revision and all the other changes we have a much more
complete gamma function.

Here the points which have changed:

1. Numerically evaluation for complex bigfloat added.
2. More accurate for bigfloat and complex bigfloat.
3. Check for negative integers and float or bigfloat values
   which represent a negative integer.
4. Check for an overflow in gamma-lanczos.
3. Simplifying expressions like gamma(z+1)/gamma(z-1)
4. $factlim and $gammalim work independently and have max values
   which will work. This prevends unintentional system hangs.
5. Cut off efficiency hacks which bypass the simplifier.
6. Get realpart and imagpart when we know it is a number.
   Thus we no longer get unnecessary questions from asksign.

I think the discussion on the mailing list related to the double float precision
of gamma has clarified the quality of the gamma-lanczos algorithm.

At last the work on the documentation has to be continued.

Dieter Kaiser

-----Urspr?ngliche Nachricht-----
Von: robert.dodier at gmail.com [mailto:robert.dodier at gmail.com] 
Gesendet: Mittwoch, 22. Oktober 2008 06:19
An: Dieter Kaiser
Cc: maxima at math.utexas.edu
Betreff: Re: [Maxima] Handling of $gammalim and $factlim in simpgamma

On 10/21/08, Dieter Kaiser <drdieterkaiser at web.de> wrote:

>  So what do you think? Should we commit these changes?

OK by me.

Disclaimer: I am not an expert on this topic.

best

Robert Dodier