[GHC] #14307: NamedFieldPuns should allow "ambiguous" field names

[GHC]

#14307: NamedFieldPuns should allow "ambiguous" field names
-------------------------------------+-------------------------------------
        Reporter:  mgsloan           |                Owner:  (none)
            Type:  feature request   |               Status:  new
        Priority:  low               |            Milestone:
       Component:  Compiler          |              Version:  8.2.1
      Resolution:                    |             Keywords:
Operating System:  Unknown/Multiple  |         Architecture:
                                     |  Unknown/Multiple
 Type of failure:  None/Unknown      |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:                    |  Differential Rev(s):
       Wiki Page:                    |
-------------------------------------+-------------------------------------

Comment (by simonpj):

 > This seems like poor behavior, because since a particular constructor is
 used, it is unambiguous which field is intended.

 Yes, but you need `-XDisambiguateRecordFields` for that
 ([http://downloads.haskell.org/~ghc/master/users-guide/glasgow_exts.html
 #record-field-disambiguation user manual entry]).  Haskell 2010 specifies
 that the code is should be rejected.  So I think GHC is behaving right
 here.

 Incidentally, `-XDisambiguateRecordFields` is also implied by
 `-XDuplicateRecordFields`.

 > I actually encountered this issue in a bit of a different usecase.

 Yes, here's the code
 {{{
 {-# LANGUAGE DuplicateRecordFields #-}
 {-# LANGUAGE NamedFieldPuns #-}

 module T14307 where

 data A = A { field :: Int }
 data B = B { field :: Int }

 f (C { field }) = field
 }}}
 You get two errors with 8.2
 {{{
 T14307.hs:10:4: error: Not in scope: data constructor ‘C’
    |
 10 | f (C { field }) = field
    |    ^

 T14307.hs:10:8: error:
     Ambiguous occurrence ‘field’
     It could refer to either the field ‘field’,
                              defined at T14307.hs:7:14
                           or the field ‘field’, defined at T14307.hs:6:14
    |
 10 | f (C { field }) = field
    |        ^^^^^
 }}}
 I think your point is that you'd like the second to be suppressed.  I see
 the point.  Are you also ok with getting just one error message from
 {{{
 f (A { fld = x }) = ...
 }}}
 namely `'A' is not in scope`; but no `fld is not in scope`?  That is: if
 the data constructor is not in scope, suppress out-of-scope or ambiguity
 messages for the fields.  I think that'd be fine.

-- 
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/14307#comment:3>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler