On May 30, 2006, at 17:32, Andy Dougherty (via RT) wrote:
The second patch changes the initial value to a different non-zero value. This is useful for ferreting out hidden assumptions.
Applied now both (slightly modified: C<if> is pulled out of the loops). I and N registers are never set to 0 / 0.0 on subroutine entry. Thanks for the patch, leo
