#14779: Compiling with -g fails -lint-core checks -------------------------------------+------------------------------------- Reporter: niteria | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: 8.4.2 Component: Compiler | Version: 8.3 (Debugging) | Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Type of failure: Compile-time | Unknown/Multiple crash or panic | Test Case: Blocked By: | Blocking: Related Tickets: #14122, #14123, | Differential Rev(s): phab:D4423 #8472, #14406 | Wiki Page: | -------------------------------------+------------------------------------- Comment (by niteria):
I don't get it. You've added a case to tryForCSE that looks a the original expression expr, not the stripped one expr''. How does that differ from the existing code where the test is in cse_bind?
Sorry, perhaps I was imprecise. I pointed to code before the change, the code after the change is: {{{ tryForCSE :: Bool -> CSEnv -> InExpr -> OutExpr tryForCSE toplevel env expr | toplevel && exprIsMbTickedLitString expr = expr -- See Note [Take care with literal strings] | Just e <- lookupCSEnv env expr'' = mkTicks ticks e | otherwise }}} Where `exprIsMbTickedLitString` looks through allowed ticks. I'm not married to the name by the way, I have a hard time coming up with a better one. For completeness: {{{ exprIsMbTickedLitString :: CoreExpr -> Bool exprIsMbTickedLitString = isJust . exprIsMbTickedLitString_maybe exprIsMbTickedLitString_maybe :: CoreExpr -> Maybe CoreExpr exprIsMbTickedLitString_maybe e@(Lit (MachStr _)) = Just e exprIsMbTickedLitString_maybe (Tick t e) | tickishPlace t == PlaceCostCentre = Nothing | otherwise = exprIsMbTickedLitString_maybe e exprIsMbTickedLitString_maybe _ = Nothing }}} I should be able to put up a complete phab patch soon and discussing the details should be easier. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/14779#comment:23 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler