#10056: Inconsistent precedence of ~ -------------------------------------+------------------------------------- Reporter: crockeea | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.8.4 (Parser) | Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Type of failure: GHC rejects | Unknown/Multiple valid program | Test Case: Blocked By: | Blocking: Related Tickets: #10059 | Differential Revisions: -------------------------------------+------------------------------------- Comment (by RyanGlScott): What exactly ''should'' the precedence of `(~)` be, anyway? It definitely seems like it should be lower than most things, but how low? For example, should this: {{{#!hs f :: (Int -> Int ~ Int -> Int) => String }}} parse as this? {{{#!hs f :: ((Int -> Int) ~ (Int -> Int)) => String }}} Also, I'm not sure that I understand this concern:
Should `(~)` be imported as part of the Prelude? If no, then a lot of code breaks. If yes, that implies that hiding the `Prelude` also hides `(~)`, breaking less code, but still breaking code.
Does hiding `Prelude` necessarily mean that `(~)` won't be visible? I was under the impression that certain types would still be visible even if `Prelude` was hidden, e.g., `(->)`. Couldn't we make `(~)` another such type and sidestep that issue? As for the `LANGUAGE` pragma question, I think it would make sense for `TypeOperators` to enable use of `(~)`. I would keep `TypeFamilies` and `GADTs`' ability to enable use of `(~)` to avoid breaking code unnecessarily—perhaps a warning can be emitted if `(~)` is used in the presence of `TypeFamilies` or `GADTs` but not `TypeOperators`? -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/10056#comment:10 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler