[GHC] #10079: Coercible solver regression: Couldn't match rep of () with Const () b

[GHC]

#10079: Coercible solver regression: Couldn't match rep of () with Const () b
-------------------------------------+-------------------------------------
        Reporter:  glguy             |                   Owner:  goldfire
            Type:  bug               |                  Status:  new
        Priority:  high              |               Milestone:  7.10.1
       Component:  Compiler (Type    |                 Version:  7.10.1-rc1
  checker)                           |                Keywords:
      Resolution:                    |            Architecture:
Operating System:  Unknown/Multiple  |  Unknown/Multiple
 Type of failure:  GHC rejects       |               Test Case:  indexed-
  valid program                      |  types/should_compile/T10079
      Blocked By:                    |                Blocking:
 Related Tickets:  #7788, #8550      |  Differential Revisions:  Phab:D653
-------------------------------------+-------------------------------------

Comment (by goldfire):

 Just to clarify, the original report with this ticket ''is'' actually
 already fixed in the 7.10 branch (see comment:12). It's just that the fix
 in 7.10 is very incomplete, as Simon points out in comment:9. That said,
 the fix in 7.10 doesn't introduce ''new'' bad behavior.

 So, if we're happy to live with #7788, #8550, #9554, and #10139, then we
 can avoid merging this to 7.10.

 Personally, this is a hard call. If we don't merge, then various (ill-
 typed) programs will cause GHC to hang, sometimes mysteriously. In
 particular, #10139 has a program without `UndecidableInstances` but that
 hangs 7.10RC2. The patch at Phab:D653 allows the program to compile. But I
 see how this change could rock the boat! I certainly welcome other
 opinions about how to proceed.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10079#comment:22>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler