On Tue, 2004-05-04 at 21:43, Mike Stump wrote: > After asking around here, our position is that while we would like to > see no regressions on compile time at -O0, we do _not_ want to hold up > the merge to mainline for it. > OK. Thanks. Diego.