#10746: No non-exhaustive pattern match warning given for empty case analysis -------------------------------------+------------------------------------- Reporter: bgamari | Owner: Type: bug | Status: patch Priority: normal | Milestone: Component: Compiler | Version: 7.10.2 Resolution: | Keywords: | PatternMatchWarnings Operating System: Unknown/Multiple | Architecture: Type of failure: Incorrect | Unknown/Multiple warning at compile-time | Test Case: Blocked By: | Blocking: Related Tickets: #7669, #11806 | Differential Rev(s): Phab:D2105 Wiki Page: | -------------------------------------+------------------------------------- Comment (by dfeuer): I don't know about the other stuff, but I think I disagree on the bounded bit. In that situation, I would expect the compiler to go into an infinite loop (unless it has pity on us in that particular case and happens to detect the loop, which is not the pattern checker's problem in any case). If I tell the type checker to go into an infinite loop and heat up my computer, it's really okay if it does that. Again, I can always write {{{#!hs foo :: F Int -> a foo !_ = error "Unreachable, but that's not provable within Haskell." }}} if I want to impose an ad hoc postulate about program behavior. Empty case is for what GHC *can* prove. The case of terminally stuck closed type families (ones that no additional information could possibly reduce) is a bit weird; I think for now at least you should probably treat them the same as stuck open type families and assume them inhabited. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/10746#comment:27 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler