Re: [GHC] #8763: forM_ [1..N] does not get fused (allocates 50% more) (was: forM_ [1..N] does not get fused (10 times slower than go function))