Bill Baxter wrote > so if it makes sense as an expression > then might as well allow it to be used that way Problem: what happens to an expression which has an `assert' as subexpression, if -release is given to the compiler? -manfred