[GHC] #12549: Panic on ":t datatypeName"

[GHC]

#12549: Panic on ":t datatypeName"
---------------------------------+--------------------------------------
        Reporter:  johnleo       |                Owner:  johnleo
            Type:  bug           |               Status:  new
        Priority:  normal        |            Milestone:
       Component:  Compiler      |              Version:  8.1
      Resolution:                |             Keywords:
Operating System:  MacOS X       |         Architecture:  x86_64 (amd64)
 Type of failure:  None/Unknown  |            Test Case:
      Blocked By:                |             Blocking:
 Related Tickets:                |  Differential Rev(s):
       Wiki Page:                |
---------------------------------+--------------------------------------

Comment (by johnleo):

 Thank you, everything makes sense except "does not substitute in `rho`".
 Perhaps I misunderstood something, but it seems it is essential to
 substitute into `rho` at some point and nowhere else is it done, so if I
 leave that out the original skolem variables are never replaced with taus
 and unification fails later on.  Indeed I tried this out and it failed: I
 get the error `No instance for (C k0 k1 a0) arising from a use of ‘f’`
 when I try to do `:t f` for `class C a where f :: a b c`.  If I leave in
 the substitution for `rho` everything seems to work fine.  Note that I
 keep the unchecked substitutions in `deeply_instantiate` and only revert
 the one in `new_meta_tv_x` (which was the one causing the bug in the first
 place), since I'm worried making the other ones checked will cause other
 problems.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12549#comment:22>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler