[GHC] #10182: lookupIfaceGlobal crash with SOURCE import

[GHC]

#10182: lookupIfaceGlobal crash with SOURCE import
-------------------------------------+-------------------------------------
        Reporter:  simonpj           |                   Owner:  ezyang
            Type:  bug               |                  Status:  patch
        Priority:  normal            |               Milestone:  7.12.1
       Component:  Compiler          |                 Version:  7.8.4
      Resolution:                    |                Keywords:
Operating System:  Unknown/Multiple  |            Architecture:
 Type of failure:  None/Unknown      |  Unknown/Multiple
      Blocked By:                    |               Test Case:
 Related Tickets:                    |                Blocking:
                                     |  Differential Revisions:
                                     |  https://phabricator.haskell.org/D860
-------------------------------------+-------------------------------------

Comment (by nomeata):

 With Edwards’s change to the test case to not do a self-SOURCE-import, but
 rather indirected through another module, we get:

 {{{
 *** Core Lint errors : in result of Simplifier ***
 T8743.hs-boot:3:10: warning:
     [RHS of $fxToRowMaybe :: forall a_anp. ToRow (Maybe a_anp)]
     idArity 1 exceeds typeArity 0: $fxToRowMaybe
 }}}

 The error message is misleading: It occurs when compiling `T8743.hs`
 proper. The desugared code before optimization is fine, after it occurs.
 My guess is that still from compiling `T8740.hs-boot`, GHC has `ToRow` in
 its environment as an abstract data type, without the information that it
 is just a newtype around `a -> [()]`. This is used by the renamer when
 parsing `T8740.hs`, so `typeArity`, which looks at the `DataTypCon`,
 returns `0` instead of `1`.

 Rough guess: The renamer should use the `T8743.hs`’s definition of
 `ToRow`, even if a (less useful) definition is already known from a boot
 file.

 Edward, does this explanation help you hunt down the problem? I know
 little about the plumbing of the various environments and such.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10182#comment:9>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler