I think the point here is to provide an example showing how different compilers implement certain aspects of source code. There is no implication that you shoul

I think the point here is to provide an example showing how different compilers implement certain aspects of source code. There is no implication that you should be able to depend on exact floating-point comparisons in programs. We avoid depending on exact floating-point comparisons, not because of any inherent indeterminism in that comparison, but because it does not generally yield exactly what our algorithm needs.