[commit: ghc] master: Comments only (d1295da)

[git at git.haskell.org]

Repository : ssh://git@git.haskell.org/ghc

On branch  : master
Link       : http://ghc.haskell.org/trac/ghc/changeset/d1295da3a4031cd102de77ab65d6d5b9b452213c/ghc

>---------------------------------------------------------------

commit d1295da3a4031cd102de77ab65d6d5b9b452213c
Author: Simon Peyton Jones <simonpj at microsoft.com>
Date:   Fri May 1 15:51:10 2015 +0100

    Comments only


>---------------------------------------------------------------

d1295da3a4031cd102de77ab65d6d5b9b452213c
 compiler/typecheck/TcFlatten.hs | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/compiler/typecheck/TcFlatten.hs b/compiler/typecheck/TcFlatten.hs
index 4c74ba9..09ed340 100644
--- a/compiler/typecheck/TcFlatten.hs
+++ b/compiler/typecheck/TcFlatten.hs
@@ -1576,6 +1576,11 @@ Consider f1 = (Given, ReprEq)
          f2 = (Derived, NomEq)
           f = (Derived, ReprEq)
 
+I thought maybe we could never get Derived ReprEq constraints, but
+we can; straight from the Wanteds during improvment. And from a Derived
+ReprEq we could conceivably get a Derived NomEq improvment (by decomposing
+a type constructor with Nomninal role), and hence unify.
+
 Note [canRewriteOrSame]
 ~~~~~~~~~~~~~~~~~~~~~~~
 canRewriteOrSame is similar but