Andre Poenitz wrote: > It's nasty when optimizers 'improve' ;-} iirc from the POV of C++ Standard such optimizations are not permitted unless you guarantee that I/O remain the same, right?
> Could you try to add a 'volatile' in front of the 'support::PathChanger p(pp)' > and see whether it makes a difference? it makes no difference > [I think it should not, but then, the optimizer seems overly eager here] pavel
