Paul Eggert wrote: > I suppose we [could] test both with volatile function pointers, to test the > actual Gnulib code, and without the pointers, to test possible compiler > misoptimizations. Though that would generate more of the false-positive > warnings that motivated the change....
Yes. But it wouldn't solve our dilemma: If we enable the compiler warnings, they are annoying. If we silence them, we lose our showcase. I don't know which is better... We can leave it as-is. Bruno
