cvs commit: fptools/ghc/compiler/basicTypes Module.lhs fptools/ghc/compiler/hsSyn HsImpExp.lhs fptools/ghc/compiler/parser Parser.y fptools/ghc/compiler/rename RnEnv.lhs RnHiFiles.hi-boot-5 RnHiFiles.hi-boot-6 RnHiFiles.lhs RnIfaces.lhs RnNames.lhs ... Date: Thu, 12 Sep 2002 07:19:01 -0400 (EDT)

Simon Peyton Jones simonpj@glass.cse.ogi.edu 

simonpj     2002/09/12 04:18:25 PDT
Sender: cvs-ghc-admin@haskell.org
Errors-To: cvs-ghc-admin@haskell.org
X-BeenThere: cvs-ghc@haskell.org
X-Mailman-Version: 2.0.8
Precedence: bulk
List-Help: <mailto:cvs-ghc-request@haskell.org?subject=help>
List-Post: <mailto:cvs-ghc@haskell.org>
List-Subscribe: <http://www.haskell.org/mailman/listinfo/cvs-ghc>,
	<mailto:cvs-ghc-request@haskell.org?subject=subscribe>
List-Id: GHC/CVS discussion and fptools/ghc CVS commit messages <cvs-ghc.haskell.org>
List-Unsubscribe: <http://www.haskell.org/mailman/listinfo/cvs-ghc>,
	<mailto:cvs-ghc-request@haskell.org?subject=unsubscribe>
List-Archive: <http://www.haskell.org/pipermail/cvs-ghc/>
Date: Thu, 12 Sep 2002 04:18:25 -0700

  Modified files:        (Branch: meta-haskell-branch)
    ghc/compiler/basicTypes Module.lhs 
    ghc/compiler/hsSyn   HsImpExp.lhs 
    ghc/compiler/parser  Parser.y 
    ghc/compiler/rename  RnEnv.lhs RnHiFiles.hi-boot-5 
                         RnHiFiles.hi-boot-6 RnHiFiles.lhs 
                         RnIfaces.lhs RnNames.lhs 
    ghc/compiler/typecheck TcRnDriver.lhs TcRnTypes.lhs 
  Log:
  	--------------------------------
  	Improve error recovery when an
  	interface file doesn't exist
  	--------------------------------
  
  Up to now, when compiling M, GHC looks in M.hi to see if
  its usages are up to date.  But M.hi might mention Foo.hi, which
  no longer exists; that is not an error, so no error is emitted.
  
  The problem was that if M *did* still import Foo, the error-suppression
  carried on, and we didn't get the error "no Foo.hi" instead we got
  lots of "x is not in scope".
  
  Solution is to make loadInterface a bit cleverer, including knowing
  whether the reason for the load is for the purpose of usage tracking.
  Actually that means that fewer functions are exported from RnHiFiles
  which is good.
  
  Also WhereFrom(..) moves from basicTypes/Module to typecheck/TcRnTypes,
  which is satisfactory (less global).
  
  Revision  Changes    Path
  1.51.2.8  +0 -28     fptools/ghc/compiler/basicTypes/Module.lhs
  1.19.14.5 +6 -3      fptools/ghc/compiler/hsSyn/HsImpExp.lhs
  1.95.2.20 +5 -5      fptools/ghc/compiler/parser/Parser.y
  1.147.2.22 +2 -2      fptools/ghc/compiler/rename/RnEnv.lhs
  1.2.14.3  +1 -1      fptools/ghc/compiler/rename/RnHiFiles.hi-boot-5
  1.2.2.2   +1 -1      fptools/ghc/compiler/rename/RnHiFiles.hi-boot-6
  1.64.2.17 +47 -38    fptools/ghc/compiler/rename/RnHiFiles.lhs
  1.152.2.15 +34 -29    fptools/ghc/compiler/rename/RnIfaces.lhs
  1.131.2.12 +11 -12    fptools/ghc/compiler/rename/RnNames.lhs
  1.1.2.5   +4 -4      fptools/ghc/compiler/typecheck/Attic/TcRnDriver.lhs
  1.1.2.5   +31 -2     fptools/ghc/compiler/typecheck/Attic/TcRnTypes.lhs
_______________________________________________
Cvs-ghc mailing list
Cvs-ghc@haskell.org
http://www.haskell.org/mailman/listinfo/cvs-ghc