From: Matthijs Kooijman <m.kooijman@student.utwente.nl>
Date: Tue, 23 Jun 2009 10:09:22 +0000 (+0200)
Subject: Remove dummy mkConcSm clause for r = r bindings.
X-Git-Url: https://git.stderr.nl/gitweb?a=commitdiff_plain;h=e9b66ed3b5b661eccb08e8955455b1b2e6e95154;hp=f48b311f77a6d1cb8b35376c655dff771a42fc5b;p=matthijs%2Fmaster-project%2Fc%CE%BBash.git

Remove dummy mkConcSm clause for r = r bindings.

These bindings are generated by GHC, but are now removed by normalization
as they are not part of the normal form.
---

diff --git a/VHDL.hs b/VHDL.hs
index f0d423d..3eddd8b 100644
--- a/VHDL.hs
+++ b/VHDL.hs
@@ -319,11 +319,6 @@ mkConcSm (bndr, app@(CoreSyn.App _ _))= do
           return [AST.CSISm $ AST.CompInsSm (mkVHDLExtId label) (AST.IUEntity (AST.NSimple entity_id)) (AST.PMapAspect portmaps)]
     details -> error $ "Calling unsupported function " ++ pprString f ++ " with GlobalIdDetails " ++ pprString details
 
--- GHC generates some funny "r = r" bindings in let statements before
--- simplification. This outputs some dummy ConcSM for these, so things will at
--- least compile for now.
-mkConcSm (bndr, CoreSyn.Var _) = return [AST.CSPSm $ AST.ProcSm (mkVHDLBasicId "unused") [] []]
-
 -- A single alt case must be a selector. This means thee scrutinee is a simple
 -- variable, the alternative is a dataalt with a single non-wild binder that
 -- is also returned.