[GHC] #14880: GHC panic: updateRole

[GHC]

#14880: GHC panic: updateRole
-------------------------------------+-------------------------------------
        Reporter:  RyanGlScott       |                Owner:  goldfire
            Type:  bug               |               Status:  new
        Priority:  normal            |            Milestone:
       Component:  Compiler (Type    |              Version:  8.2.2
  checker)                           |
      Resolution:                    |             Keywords:  TypeInType
Operating System:  Unknown/Multiple  |         Architecture:
 Type of failure:  Compile-time      |  Unknown/Multiple
  crash or panic                     |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:  #15076            |  Differential Rev(s):
       Wiki Page:                    |
-------------------------------------+-------------------------------------

Comment (by simonpj):

 In a call today Richard and Simon decided:

 * Kill 1decideKindGeneralisationPlan`; instead always generalise. (Explain
 why we don't need to worry about local generalisation as we do with
 terms.)

 * And hence kill `tc_hs_sig_type`, leaving `tc_hs_sig_type_and_gen`

 * `tc_hs_sig_type_and_gen` should not call `solveEqualities` (which fails
 if there are unsolved equalities)

 * Instead call `solveLocalEqualities`, gather unsolved equalities, and do
 not quantify over their free vars.   NB: `tcImplicitTKBndrs` already calls
 `solveLocalEqualities`, and must do so.  Needs a Note to explain why: it's
 so that we can top-sort the bound variables

 * Promote all the free vars of the unsolved constraints. '''Principle''':
 all free vars should either be generalised or promoted.

 * Get rid of `zonkPromote` and friends altogether

-- 
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/14880#comment:17>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler