
#14558: Unable to parse integer-gmp's Cabal file -------------------------------------+------------------------------------- Reporter: taylorfausak | Owner: hvr Type: task | Status: new Priority: normal | Milestone: Component: Core Libraries | Version: 8.2.2 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 taylorfausak): (1) is correct. As far as I can tell, the Cabal files on Hackage match the Cabal files in the GHC repository. This is good and I'm happy that it happened, but it's not why I opened this ticket. (2) is also correct. GHC and its built-in libraries require Cabal 2. This is obviously fine for GHC itself, but any downstream clients using GHC < 8.2 or Cabal < 2 are broken by this. (For example: https://github.com/tfausak/grawlix/commit/861f0f9d3f19ae9d94ce11b52311459e5a....) I don't think we are stuck with this problem for 8.2. As far as Stack < 1.6.1 is concerned, revising the `integer-gmp-1.0.1.0` release or releasing a new version of `integer-gmp` would fix things. Specifically, removing the caret constraint (and relaxing the `cabal-version`) would allow older versions of Cabal and Stack to successfully parse `integer- gmp.cabal`. The patch is pretty small: https://github.com/ghc/ghc/commit/82aae8c81f5380af5f5ab9c381b1d296104696f4 For what it's worth, I ran into this problem because Stack 1.5.1 ships with Cabal 1.24. I have fixed this problem for myself by upgrading to Stack 1.6.1. Anybody still using Stack < 1.6.1 will be broken if they try to use a snapshot with GHC >= 8.2. This includes anyone using the Haskell Platform, which bundles Stack 1.5.1. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/14558#comment:18 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler