[GHC] #16261: [] interpreted as an overloaded list in rule LHS (was: Mysterious rule shadowing warning)

[GHC]

#16261: [] interpreted as an overloaded list in rule LHS
-------------------------------------+-------------------------------------
        Reporter:  dfeuer            |                Owner:  (none)
            Type:  bug               |               Status:  new
        Priority:  normal            |            Milestone:  8.10.1
       Component:  Compiler          |              Version:  8.6.3
      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:                    |  Differential Rev(s):
       Wiki Page:                    |
-------------------------------------+-------------------------------------
Description changed by dfeuer:

Old description:

> Compiling `Data.Set.Internal` produces the following:
>
> {{{
> Data/Set/Internal.hs:958:11: warning: [-Winline-rule-shadowing]
>     Rule "Set.toAscListBack" may never fire
>       because rule "Class op fromListN" for ‘GHCExts.fromListN’ might
> fire first
>     Probable fix: add phase [n] or [~n] to the competing rule
>     |
> 958 | {-# RULES "Set.toAscListBack" [1] foldrFB (:) [] = toAscList #-}
>     |           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>
> Data/Set/Internal.hs:960:11: warning: [-Winline-rule-shadowing]
>     Rule "Set.toDescListBack" may never fire
>       because rule "Class op fromListN" for ‘GHCExts.fromListN’ might
> fire first
>     Probable fix: add phase [n] or [~n] to the competing rule
>     |
> 960 | {-# RULES "Set.toDescListBack" [1] foldlFB (\xs x -> x : xs) [] =
> toDescList #-}
>     |
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> }}}
>
> What the heck does the class op rule for `fromListN` have to do with
> anything anywhere near here? Note in particular that this module neither
> defines `fromListN` (it takes its default definition) nor uses it in code
> or `RULES`.

New description:

 Compiling `Data.Set.Internal` produces the following:

 {{{
 Data/Set/Internal.hs:958:11: warning: [-Winline-rule-shadowing]
     Rule "Set.toAscListBack" may never fire
       because rule "Class op fromListN" for ‘GHCExts.fromListN’ might fire
 first
     Probable fix: add phase [n] or [~n] to the competing rule
     |
 958 | {-# RULES "Set.toAscListBack" [1] foldrFB (:) [] = toAscList #-}
     |           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

 Data/Set/Internal.hs:960:11: warning: [-Winline-rule-shadowing]
     Rule "Set.toDescListBack" may never fire
       because rule "Class op fromListN" for ‘GHCExts.fromListN’ might fire
 first
     Probable fix: add phase [n] or [~n] to the competing rule
     |
 960 | {-# RULES "Set.toDescListBack" [1] foldlFB (\xs x -> x : xs) [] =
 toDescList #-}
     |
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 }}}

 As mpickering figured out, the trouble is that `[]` in the rule LHS is
 being interpreted as an overloaded list. The module does ''not'' enable
 `OverloadedLists`. Furthermore, even if it ''did'', I would argue that it
 is ''never'' correct to put one in a rule LHS, so we should always
 interpret `[]` as the empty list constructor in `RULES` left-hand sides.

--

-- 
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/16261#comment:6>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler