[GHC] #11796: Warn about unwanted instances in a modular way

[GHC]

#11796: Warn about unwanted instances in a modular way
-------------------------------------+-------------------------------------
        Reporter:  Lemming           |                Owner:
            Type:  feature request   |               Status:  new
        Priority:  normal            |            Milestone:
       Component:  Compiler          |              Version:  8.1
      Resolution:                    |             Keywords:
Operating System:  Unknown/Multiple  |         Architecture:
                                     |  Unknown/Multiple
 Type of failure:  None/Unknown      |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:  #11219            |  Differential Rev(s):
       Wiki Page:                    |
-------------------------------------+-------------------------------------

Comment (by Lemming):

 Replying to [comment:3 RyanGlScott]:

 > * We currently have `WARNING` and `DEPRECATED` pragmas
 ([http://downloads.haskell.org/~ghc/8.0.1/docs/html/users_guide/glasgow_exts.html
 #warning-and-deprecated-pragmas link]) that serve a very similar role to
 what you're proposing.

 A unification with the WARNING pragma looks like a good idea to me. I
 think I did not consider it, because currently warnings cannot be grouped.
 But if WARNINGs can be grouped we could re-use this capability for
 `instance` warnings.

 >   This always emits a warning. But we could envision a flags like
 `-Wpragma-error="unsafePerformIO"` or `-Wpragma-
 suppress="unsafePerformIO"` that would either elevate a `WARNING` pragma
 to an error or suppress its warning altogether, respectively. (We'd also
 need `-Wpragma-warn` for when we want to explicitly request a warning,
 rather than an error). So implementing `-Wpragma-error`, `-Wpragma-warn`,
 and `-Wpragma-suppress` seems like an important step.

 The machinery sketched in #11219 seems like a good way to switch between
 the effect of hitting a warning.


 > * Lastly, we'd need some form of conflict resolution for unique strings.
 I could certainly see a scenario where a package has multiple `FooInt`
 strings floating around. Perhaps we could use a representation of
 `-Wpragma-warn=<package>:<module>:<uniq_string>`, where `<package>` and
 `<module>` are optional.

 For `instance`s the module name should not be part of the identifier. The
 definition of the instance can be anywhere in the package in a private
 module. This module can be renamed without even a minor version bump. But
 when I think about it, the same is true for function identifiers.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11796#comment:4>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler