[GHC] #15703: Significant compilation time blowup when refactoring singletons-heavy code

[GHC]

#15703: Significant compilation time blowup when refactoring singletons-heavy code
-------------------------------------+-------------------------------------
        Reporter:  RyanGlScott       |                Owner:  (none)
            Type:  bug               |               Status:  new
        Priority:  normal            |            Milestone:  8.8.1
       Component:  Compiler          |              Version:  8.6.1
      Resolution:                    |             Keywords:
Operating System:  Unknown/Multiple  |         Architecture:
 Type of failure:  Compile-time      |  Unknown/Multiple
  performance bug                    |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:                    |  Differential Rev(s):
       Wiki Page:                    |
-------------------------------------+-------------------------------------

Comment (by simonpj):

 Those are staggeringly large coercions decorating relatively tiny types.

 I wonder: could any of them be very large representations of `Refl`?  I
 note that `OptCoercion` doesn't use `isReflexiveCo` (which tests for
 equality of the two types) because doing so at every node would be
 expensive.  But you could try testing for `isReflexiveCo`, and spit out
 the smallest coercion which says `True` to `isReflexiveCo` but is not
 actually `Refl`.  You'd want to make this test for every sub-tree in the
 coercion.

 That's a guess.  But it's very hard to imagine that you ''really'' need
 such large proofs.  The trick is to find where the redundancy is.

-- 
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/15703#comment:6>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler