29 Aug
2008
29 Aug
'08
4:38 p.m.
Tim,
Looking at the code, it comes down to a call to unsafe_atan_bf(). We need to make a unit test and find out if the function always fails or just in certain quadrants.
Definitely a DUH! moment. I needed to call init_big_pi(). There are still some idiosyncrasies to work out. And, the AP math doesn't seem to like the high bailout used by the dividebrot5 fractal type. Haven't had a chance to look at the Xfractint problem. Jonathan