-
Jiri (George) Lebl authored
* src/eval.c: when forloop (sum/prod/for) is with floats then check for overrun, and if within 2^-20 times the step size, then still execute the last body (assume things were because of roundoff error), this makes things like for n=0 to 1 by 0.1 actually work as expected. * src/mpwrap.[ch], src/eval.c, src/funclib.c, src/matop.c: minor optimization in make_float and new macro to check real part for being floating point * src/geniustests.txt, src/testfourier.gel: update testsuite
4dea9c70