#9590: AMP breaks `haskell2010` package -------------------------------------+------------------------------------- Reporter: hvr | Owner: Type: bug | Status: new Priority: high | Milestone: 7.10.1 Component: | Version: 7.9 libraries/haskell2010 | Keywords: AMP Resolution: | Architecture: Unknown/Multiple Operating System: | Difficulty: Unknown Unknown/Multiple | Blocked By: Type of failure: GHC | Related Tickets: rejects valid program | Test Case: | Blocking: | Differential Revisions: | -------------------------------------+------------------------------------- Comment (by ekmett): It is a bit more complex than just pasting the report. GHC internals actually know where `Monad` comes from for instance, so you'd have knock- on consequences. If we `do` adopt a `haskell2010` fully standards compliant package: 1.) We'd need to make some modifications to GHC itself to tell it where to desugar to for the `Monad` syntax and internal hackery. We'd need to figure out how to tell tools like cabal to do the right thing. 2.) Code that previously worked with the `haskell2010` or `haskell98` packages would cease to work with `base` and vice versa where those instances are concerned. You'd be giving up, say, `containers`, not just the rest of `base`. This actually comes at a price to a dozen or two packages that are built right now with the `haskell98` package and with these other dependencies as well, possibly breaking in non-obvious ways. It may be that to fully support both of these two very different usecases for `haskell2010` that you'd get two `haskell2010`-like packages to support both user pools. This makes for a very complex story to explain to end users, and a lot of engineering. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/9590#comment:7 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler