'deriving Data' compilation regression 6.10.1 -> 6.12.1

[ronwalf]

Compiling the attached file in GHC 6.12.1 gives the following error, when GHC 6.10.1 compiles it fine:

[1 of 1] Compiling Main             ( baddata.hs, baddata.o )

baddata.hs:33:14:
    No instances for (Data Const, Data Var)
      arising from the 'deriving' clause of a data type declaration
                   at baddata.hs:33:14-17
    Possible fix:
      add an instance declaration for (Data Const, Data Var)
      or use a standalone 'deriving instance' declaration instead,
         so you can specify the instance context yourself
    When deriving the instance for (Data (Domain e g))

[ronwalf]

baddata2.hs gives a different, possibly more informative error:

baddata2.hs:34:0:
    No instances for (Data Const, Data Var)
      arising from a use of `k' at baddata2.hs:34:0-28
    Possible fix:
      add an instance declaration for (Data Const, Data Var)
    In the expression: k (z Domain)
    In the definition of `gunfold': gunfold k z _ = k (z Domain)
    When typechecking a standalone-derived method for `Data Domain':
      gunfold k z _ = k (z Domain)
    In the instance declaration for `Data Domain'

[ronwalf]

Simpler version, different error

[igloo]

Great testcase, thanks!

[simonpj]

Great report, thanks. Fixed by

Fri Apr  9 19:44:20 BST 2010  simonpj@microsoft.com
  * Fix Trac #3965: tighten conditions when deriving Data
  
  It's tricky to set up the context for a Data instance.  I got it wrong
  once, and fixed it -- hence the "extra_constraints" in
  TcDeriv.inferConstraints.  
  
  But it still wasn't right!  The tricky bit is that dataCast1 is only
  generated when T :: *->*, and dataCast2 when T :: *->*->*. (See
  the code in TcGenDeriv for dataCastX.

    M ./compiler/typecheck/TcDeriv.lhs -8 +14

Perhaps merge to 6.12.3

Simon

[igloo]

Merged