Hi Assume that rpm in its original form was mal-functioning with the kernel -> the problem disappears with a patch applied to rpm. Then comes 'the Stanford bug succer' and removes that type of code that was originally causing the rpm problem. With the now perfected type of code in the kernel, the patch is creating a new problem. It is possibly to go in a circle to apply a new patch instead one might try to compile rpm without patches and only apply them one by one if the original problem can be recreated. regards guran