#10011: The Data instance for Ratio violates internal invariants. -------------------------------------+------------------------------------- Reporter: ekmett | Owner: Type: bug | Status: new Priority: normal | Milestone: 7.10.1 Component: Core Libraries | Version: 7.10.1-rc1 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Type of failure: Incorrect result | Unknown/Multiple at runtime | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Revisions: Phab:D625 -------------------------------------+------------------------------------- Comment (by ekmett): This work scoped in this ticket is done, but you are right, we should definitely add a regression test to ensure this doesn't backslide in the future. Along the way we found a much less critical, but related issue (ok, really, feature request), which I'll file a separate ticket for -- with an eye towards fixing it on a longer time horizon: I'd eventually like to fix up the "virtual" data constructor we talk about in the Data instance to use `(%)`, since we're using a virtual constructor here, and that would let `gshow` and the like do the right thing, but when we tried that as part of this patch, the fact that `dataToExpQ` currently filters out things that are illegal constructors caused a test failure. Fixing _that_ is a big enough issue to punt to 7.12. Fixing that right might also let us fix things like dataToExpQ for things like Maps, Sets, and other containers that also have to use the virtual constructor model to preserve internal invariants, which would be a pretty big boon to heavy users of template haskell. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/10011#comment:11 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler