Regression: Detection of unused imports is imprecise (since GHC 8.0.1)
GHCs detection of unused imports used to be very precise. But apparently this changed with GHC 8.0.1. I already noticed for a while that there is something fishy going on and finally came up with a minimal example.
Steps to reproduce:
-- Bar.hs
module Bar (bar, module Control.Monad) where
import Control.Monad
bar :: Integer
bar = 23
-- Foo.hs
module Foo where
import Bar
import Control.Monad (forM_)
foo :: Monad m => [a] -> (a -> m ()) -> m ()
foo = forM_
baz :: Integer
baz = bar
$ ghci -Wall -Werror Foo.hs
- *Expected result:** The program is reject.
- *Actual result:** The program was reject by GHC < 8.0.1, however later versions of GHC accept the program.
Trac metadata
| Trac field | Value |
|---|---|
| Version | 8.4.1 |
| Type | Bug |
| TypeOfFailure | OtherFailure |
| Priority | normal |
| Resolution | Unresolved |
| Component | Compiler |
| Test case | |
| Differential revisions | |
| BlockedBy | |
| Related | |
| Blocking | |
| CC | |
| Operating system | |
| Architecture |