Re: Bug #13587 Updated: Recurence makes crash

On Sun, 7 Oct 2001, Jeroen van Wolffelaar wrote: > <snip> > > Read the archives before responding with this kind of stuff. It WILL give > > a big performance it. > > There's no more information than in this bug report as to why it is > such a big performance hit. Maybe you was referring to what Zeev said in reaction to bug 1901. But Zeev was talking about a big performance hit when you change from a resizable array (zeev calls it stack, but stack is an other word for LIFO's) to heap implementation for the execution stack. This has nothing to do with the way of recovery (if any) from resizeable-array c.q. heap overflow. In other words, that's irrelevant. By the way, a linked list implementation would in general be faster, while requiring size_t*stack_depth more memory. If the stack remains small all the time however, the current implementation is very slightly faster. Just curious: Why has there been chosen for resizeble array implementation and not a linked list? --Jeroen Jeroen van Wolffelaar Jeroen@A-Eskwadraat.nl http://www.A-Eskwadraat.nl/~jeroen