One work around is to fool the compiler by reordering the source code case statements so it cannot build its jump table.

I am very familiar with this since Kail in their infinite wisdom has moved the jump table feature from the compiler (which I use) to the optimizer (which I will not use). Why a feature for which the compiler 'knows' all that is required has to be an optimization I do not understand. Is it not the 'duty' of a C compiler to produce the most effiecient code possible?

Now, someone is going to ask "why not?" well NASA states "we fly what we test" and so do I. Due to the (way too) many outside products we interface with (since another division of my company compete with them they are, of course, most willing to tell me all I want to know) I occasinally have to debug 'on the fly' and doing that with a version that has different timing than the one that show the problem - you've got to be kidding.

Erik