> > > > Well, not exactly of the case mentioned above, but still think that this is > > safe (ugly, too). An alternative is to keep around the bodies until after > > inlining. I have infrastructure for that in my tree, but it is hard to > > tune to > > do: first the alternative function body may have different symbol references > > (as a result of different early inlinin) which may not be resolved in > > current > > binary so we can not use it at all. Second keepin many alternatives of every > > body around makes code size estimates in inliner go crazy. > > /me inserts his usual "partitioning to the rescue" comment... > > ;)
Hmm, OK. I fail to see how that would save the day here :) It is not a problem about the fact that we can not switch flags according the function but about the linking that happens prior partitioning. Honza > > Richard.