Hi silvas. Apologies for the hiatus for this review. I'll start to try to find examples of where semantics change when the risk level is set too high. The comments in the code (in LoopActions.cpp) do not show any such examples, so we'll probably have to "reverse-engineer" the loop-convert logic and find something that produces incorrect transformations that changes semantics. The other review comments are easily addressable though.
http://llvm-reviews.chandlerc.com/D339 _______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
