Repository : ssh://darcs.haskell.org//srv/darcs/ghc On branch : supercompiler
http://hackage.haskell.org/trac/ghc/changeset/596159a3fdf211b15a0bb1e3c0beb1458c01e191 >--------------------------------------------------------------- commit 596159a3fdf211b15a0bb1e3c0beb1458c01e191 Author: Max Bolingbroke <batterseapo...@hotmail.com> Date: Mon Apr 23 17:00:58 2012 +0100 Comment only >--------------------------------------------------------------- compiler/supercompile/Supercompile/Core/Syntax.hs | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/compiler/supercompile/Supercompile/Core/Syntax.hs b/compiler/supercompile/Supercompile/Core/Syntax.hs index 0cdeb19..ccb21be 100644 --- a/compiler/supercompile/Supercompile/Core/Syntax.hs +++ b/compiler/supercompile/Supercompile/Core/Syntax.hs @@ -128,7 +128,7 @@ data TermF ann = Var Id | TyApp (ann (TermF ann)) Type | CoApp (ann (TermF ann)) Coercion | App (ann (TermF ann)) Id - | PrimOp PrimOp [Type] [ann (TermF ann)] -- FIXME: arguably we have just Vars as arguments for better Tag behaviour (otherwise improving the arguments is hidden by the Tag on the whole PrimOp stack frames). FIXME: in fact, we need to change this because *NOT ALL PRIMOP ARGUMENTS ARE STRICT* (e.g. the lazy polymorphic arguments to newMutVar#, newArray#) + | PrimOp PrimOp [Type] [ann (TermF ann)] -- FIXME: arguably we have just Vars as arguments for better Tag behaviour (otherwise improving the arguments is hidden by the Tag on the whole PrimOp stack frames). FIXME: in fact, we need to change this because *NOT ALL PRIMOP ARGUMENTS ARE STRICT* (e.g. the lazy polymorphic arguments to newMutVar#, newArray#). FIXME: the reason I haven't done this is because it means I should remove the PrimApply frame, which breaks the "question or answer" evaluator normalisation property. | Case (ann (TermF ann)) Id Type [AltF ann] -- NB: unlike GHC, for convenience we allow the list of alternatives to be empty | Let Id (ann (TermF ann)) (ann (TermF ann)) -- NB: might bind an unlifted thing, in which case evaluation changes. Unlike GHC, we do NOT assume the RHSes of unlifted bindings are ok-for-speculation. | LetRec [(Id, ann (TermF ann))] (ann (TermF ann)) _______________________________________________ Cvs-ghc mailing list Cvs-ghc@haskell.org http://www.haskell.org/mailman/listinfo/cvs-ghc
