#13227: Loss of full laziness in mapFB -------------------------------------+------------------------------------- Reporter: simonpj | Owner: nomeata Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.0.1 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: | Unknown/Multiple Type of failure: None/Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Comment (by nomeata): Looks like a clear win. According to perf.haskell.org (which does run nofib!), the only significant changes are time improvements of 3.5% for cacheprof and k-nucleotide: https://perf.haskell.org/ghc/#revision/4c5af213a9fff60ca8de2ed9f5ae28955d21b... These might be spurious, as allocations do not change. But even if really nothing changes, as this is a simplifying commit, from this point of view there is no reason not do it. And I’d be happy to get rid of this rather squishy notion of “probably” one-shot. I will prepare a DR for review. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/13227#comment:7 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler