[commit: ghc] master: Improve kind-checking of type splices slightly (70b0f16)

[Simon Peyton Jones]

Repository : ssh://darcs.haskell.org//srv/darcs/ghc

On branch  : master

http://hackage.haskell.org/trac/ghc/changeset/70b0f1608784cccaf02fbbc6227c96be20ca7c7a

>---------------------------------------------------------------

commit 70b0f1608784cccaf02fbbc6227c96be20ca7c7a
Author: Simon Peyton Jones <simonpj at microsoft.com>
Date:   Thu Apr 5 11:39:33 2012 +0100

    Improve kind-checking of type splices slightly
    
    It is still Not Right, but can't readily be fixed until
    the major overhaul of TH described in Simon PJ's 2011
    GHC blog post.  (Then type splices will be done in
    the renamer.)

>---------------------------------------------------------------

 compiler/typecheck/TcHsType.lhs |   11 ++++++-----
 1 files changed, 6 insertions(+), 5 deletions(-)

diff --git a/compiler/typecheck/TcHsType.lhs b/compiler/typecheck/TcHsType.lhs
index c29f1a4..93649ac 100644
--- a/compiler/typecheck/TcHsType.lhs
+++ b/compiler/typecheck/TcHsType.lhs
@@ -444,13 +444,14 @@ tc_hs_type (HsCoreTy ty) exp_kind
 
 
 #ifdef GHCI	/* Only if bootstrapped */
--- This looks highly bogus to me
+-- This looks highly suspect to me
+-- It will really only be fixed properly when we do the TH
+-- reorganisation so that type splices happen in the renamer
 tc_hs_type hs_ty@(HsSpliceTy sp fvs _) exp_kind 
-  = do { (ty, kind) <- tcSpliceType sp fvs
+  = do { s <- getStage
+       ; traceTc "tc_hs_type: splice" (ppr sp $$ ppr s) 
+       ; (ty, kind) <- tcSpliceType sp fvs
        ; checkExpectedKind hs_ty kind exp_kind
-
---        ; kind' <- zonkType (mkZonkTcTyVar (\ _ -> return liftedTypeKind) mkTyVarTy) 
---                            kind
 --                     -- See Note [Kind of a type splice]
        ; return ty }
 #else