#13102: orphan family instances can leak through the EPS in --make mode -------------------------------------+------------------------------------- Reporter: rwbarton | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.1 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: | Unknown/Multiple Type of failure: None/Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Comment (by rwbarton):
Wired-in things are below everything
Right, so this becomes a constraint that we have to adhere to under this plan. And it's a constraint I was intending to violate with the large tuples Generic stuff (tuples are wired-in, and I was going to use `deriving Generic` in the module that defines them). I'm not so happy with that exact idea any more for other reasons, so probably it shouldn't stop us from this plan.
Checking family-instance consistency. Here I am not so clear, but we should write down the plan.
I started writing up the plan in d6fd7922332a16fb958d3bf2c21ed792d12c98a7. Note that there is another point related to your item 2 here, in that (at least in principle) we should only check consistency of the new instances with the instances we transitively import, not any old instances we happen to know about. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/13102#comment:6 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler