#3263: Warnings for monadic values not used
------------------------------+---------------------------------------------
Reporter: NeilMitchell | Owner:
Type: feature request | Status: new
Priority: normal | Milestone:
Component: Compiler | Version: 6.10.2
Severity: normal | Resolution:
Keywords: | Testcase:
Os: Unknown/Multiple | Architecture: Unknown/Multiple
------------------------------+---------------------------------------------
Comment (by isaacdupree):
Quite an interesting idea!
-fwarn-unused-monad-bind: as well as (), it should exclude returns that
are polymorphic in any value (this idiom is sometimes used for functions
that do not return, such as `exitWith :: ExitCode -> IO a` and possibly
[parts of?] `callCC :: MonadCont m => ((a -> m b) -> m a) -> m a`)
-fwarn-wrong-bind: good idea... and you'd get around the warning, should
it prove necessary to, by explicitly binding to `_` (either `do _ <- ...`
or `... >>= \_ ->`)?
If you want to avoid `mapM` space leaks in real code among unsuspecting
programmers, we'll need a well-designed warning-message that will convince
them not to just write
{{{
--avoid silly warning: we *don't* use the result
_ <- mapM ...
}}}
Possibly some pragma action could help GHC give more precise suggestions
(like "use mapM_")
I guess the warnings would hardcode Monad, and probably hardcode () as
well, or would it be better (easier? harder?) to apply to any zero-
information data (aka `data Zero` and `data One = One` and perhaps some
newtype and/or GADT magic)?
-Isaac Dupree
--
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/3263#comment:1>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler_______________________________________________
Glasgow-haskell-bugs mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs
