#8763: forM_ [1..N] does not get fused (allocates 50% more) -------------------------------------+------------------------------------- Reporter: nh2 | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: 8.8.1 Component: Compiler | Version: 7.6.3 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Type of failure: Runtime | Unknown/Multiple performance bug | Test Case: Blocked By: | Blocking: Related Tickets: #7206 | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Comment (by sgraf): Right, actually I was aware of INLINE unfoldings being captured as the unoptimized variant, but have never seen it to actively prevent (or rather postpone) inlining of another part of the code. Quite a neat trick! I've identified some other functions that should probably rewritten the same way (`traverse_`, `foldrM`, etc.). No regressions, no improvements according to NoFib. Phab:D5131 passes `./validate.sh`. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/8763#comment:79 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler