[commit: ghc] : Comments only (a4c1c7b)

Repository : ssh://darcs.haskell.org//srv/darcs/ghc

On branch  : 

http://hackage.haskell.org/trac/ghc/changeset/a4c1c7b22b5523f2e38d02400ed155abdf2dedef

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

commit a4c1c7b22b5523f2e38d02400ed155abdf2dedef
Author: Max Bolingbroke <[email protected]>
Date:   Fri Apr 27 11:32:12 2012 +0100

    Comments only

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

 compiler/supercompile/Supercompile/Drive/MSG.hs    |    4 +++-
 .../supercompile/Supercompile/Drive/Process3.hs    |    6 +++---
 2 files changed, 6 insertions(+), 4 deletions(-)

diff --git a/compiler/supercompile/Supercompile/Drive/MSG.hs b/compiler/supercompile/Supercompile/Drive/MSG.hs
index 2e6fb5c..08fde0a 100644
--- a/compiler/supercompile/Supercompile/Drive/MSG.hs
+++ b/compiler/supercompile/Supercompile/Drive/MSG.hs
@@ -62,7 +62,9 @@ rnBndr2' rn2 x_l x_r = MSG $ \_ s -> Right (s, rnBndr2'' (uniqAway (msgInScopeSe
   -- BUT I don't want to just put ids_l/ids_r in the RnEnv2 because that will force common occurrence sites to rename (in msgPend) which
   -- which will make things drastically worse! Instead, rnBndr2' does the uniqAway manually.)
 
--- FIXME: rigid variable occurrences do not get correct type/info
+-- FIXME: rigid variable occurrences do not get correct type/info.
+-- This probably dosent' matter for internal binders since the supercompiler's normal renaming mechanism
+-- will propagate binding-site info down to the use sites, but it matters a TON for the stack binders!
 rnBndr2'' :: (Var -> Var)
           -> RnEnv2 -> Var -> Var -> (RnEnv2, Var)
 rnBndr2'' f rn2 x_l x_r = (rn2', x')
diff --git a/compiler/supercompile/Supercompile/Drive/Process3.hs b/compiler/supercompile/Supercompile/Drive/Process3.hs
index 3ee8967..dc8a3a9 100644
--- a/compiler/supercompile/Supercompile/Drive/Process3.hs
+++ b/compiler/supercompile/Supercompile/Drive/Process3.hs
@@ -485,13 +485,13 @@ data MemoHow = Skip | CheckOnly | CheckAndRemember
 -- If you do then you can start with this term:
 --   [1] let $dNum = ww3 in * a $dNum
 --
--- Looks like this after reduction+GC:
+-- Looks like this after reduction+GC (the update for $dNum is dead):
 --   [2] case ww3 of Num ...
 --
--- And if we reduce+split [1] instead we get:
+-- And if we reduce+split [1] instead we get (the update for $dNum is residualised):
 --   [3] case $dNum of Num ...
 --
--- Reducing+GCing [3] term gives us [3] again, and that is alpha equivalent to [2],
+-- Reducing+GCing [3] gives us [3] again, and that is alpha equivalent to [2],
 -- so we tie back to it rather than continuing. But that means our code is:
 --   let h @a ww3 = let $dNum = ww3
 --                  in h a $dNum