> > Chris, is there a tgamma on your system?
> Yes, there is. I am not entirely sure how to test it for correctness, but I
> managed to cobble it into the configure script to at least see if it was
OK, thanks. You could attempt a patch but it probably is a bit tricky because it appears gnuplot's gamma is based off the log gamma function, not the true gamma function. Although not too difficult, proper fix would require some extensive rearranging of defines and is better left to Ethan or Hans.
Attached is a quick hack to see if the tgamma() function might work for you. This works for me. By experimenting and changing the tgamma() to gamma() I see that, yes, gamma() on my system is actually log gamma.