> they are complaining about optimizing compilers which do things "behind your back" and are becoming an increasing nuisance. Actually no. I think there was ma

> they are complaining about optimizing compilers which do things "behind your back" and are becoming an increasing nuisance. Actually no. I think there was maybe one case where you could argue it was the optimizer that is producing different results than you'd expect for unoptimized output, but in most cases you have a problem that exists because of poor assumptions on the part of programmers/error prone language definitions.