Re: type families + GADT = type unsafety?

(Simon, does this mean that non-~ discharging will become subject to GADT-style type annotation rules?) No, it does not. No type annotations required in non-GADT-related code, even if equalities are involved. Tom -- Tom Schrijvers Department of Computer Science K.U. Leuven Celestijnenlaan

Re: ANNOUNCE: GHC 6.8.1 Release Candidate

Ian Lynagh wrote: Hi Adrian, On Fri, Sep 14, 2007 at 07:50:47AM +0100, Adrian Hey wrote: [29 of 53] Compiling Data.Tree.AVL.Join ( Data.Tree.AVL/Data/Tree/AVL/Join.hs, dist\build/Data/Tree/AVL/Join.o ) ghc.exe: panic! (the 'impossible' happened) (GHC version 6.8.20070912 for

Re: module containing GADTs no longer compiles in ghc 6.8.0

Hi Daniel, On Thu, Sep 27, 2007 at 07:45:58PM -0300, Daniel Gorín wrote: After replacing them with case statements everything compiles fine as long as I don't turn on -O2 optimizations :( This boiled-down example illustrates my problem: Thanks for the testcase! I've trac'ed it

Re[2]: type families + GADT = type unsafety?

Hello Stefan, Friday, September 28, 2007, 1:10:09 AM, you wrote: data Foo a where A :: Foo Int B :: Foo Bool becomes data Foo a = (a ~ Int) = A | (a ~ Bool) = B hm :) this looks like my quasi-proposal of unifying data and function definitions still has some meaning. i proposed to