From: Peter Vanroose <Peter.V<anroose@es...>  20030311 09:42:12

> On Borland and Mac OSX, this kills the stack: > > bool vnl_rnpoly_solve::compute() > { > int i,j; > int ideg[M], terms[M], polyn[M][T][M]; > vnl_rnpoly_solve_cmplx ans[LEN][M]; Any idea why? Could it have to do with external #defines for e.g. M or T ? Or just because T is too large? (2500) I've changed these to M_ and T_ now. Could you try with a smaller value of T_? Peter. 