#13064: Incorrect redudant imports warning -------------------------------------+------------------------------------- Reporter: phadej | Owner: (none) Type: bug | Status: new Priority: low | Milestone: 8.8.1 Component: Compiler | Version: 8.0.1 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Type of failure: Incorrect | Unknown/Multiple error/warning at compile-time | Test Case: Blocked By: | Blocking: Related Tickets: #15393 | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Changes (by simonpj): * cc: ekmett@… (added) Comment: I had forgotten all about [wiki:Commentary/Compiler/RelaxedUnusedImports]. Thanks: you are right to point to it as a possible design. We quite often fix a bug that forces libraries to change in some minor way; library authors are (perhaps reluctantly) used to this. In this case, the change affects only warnings, perhaps the least bad way to break a library. (Worse ways are: the library is now type incorrect; or (worse still) compiles cleanly but gives the wrong answers.) I think the awkward thing here is that library authors like to produce an updated version of the library that will work with many versions of GHC. For that they resort to CPP. Thus in `containers:Data.Map.Internal.hs` we see {{{ #if __GLASGOW_HASKELL__ >= 708 import qualified GHC.Exts as GHCExts #endif }}} The [https://prime.haskell.org/wiki/Libraries/3-Release-Policy three release policy says] ''Changes '''to basic libraries''' are planned and implemented so that, at any time, it is possible to write code that works with the latest three releases of GHC and base, without resorting to CPP, and without causing warnings even when compiled with -Wall.'' Note "to basic libraries". I don't see how it's possible to apply this policy to the compiler itself. Suppose a library compiles with GHC 12.0, and we want to make a change in the warnings the compiler produces. We could defer the change to 12.6; but then the author could just apply the policy starting at 12.6, so we'd have to defer to 12.12, and so on. I feel I must be missing something. Also as comment:24 points out, there is a non-CPP way to adapt too. Let's see what the CLC (cc'd) says. I'm adding Edward to cc as well. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/13064#comment:27 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler