cleanup: drop 11 years old performance hack

The 'return () >>' hack was added in commit
    commit ac88f113
    Date:   Tue Jul 26 12:14:03 2005 +0000

Nowadays it has no effect on generated Core on -O1/-O2
and slightly bloats Core on -O0.
Signed-off-by: Sergei Trofimovich <siarheit@google.com>

compiler/utils/Pretty.hs

@@ -1018,9 +1018,6 @@ hPutLitString handle a l = if l == 0
 --     and async exception-safe.  We only have a single thread and don't
 --     care about exceptions, so we add a layer of fast buffering
 --     over the Handle interface.
---
--- (3) a few hacks in layLeft below to convince GHC to generate the right
---     code.
 
 printLeftRender :: Handle -> Doc -> IO ()
 printLeftRender hdl doc = do
@@ -1031,14 +1028,11 @@ printLeftRender hdl doc = do
 bufLeftRender :: BufHandle -> Doc -> IO ()
 bufLeftRender b doc = layLeft b (reduceDoc doc)
 
--- HACK ALERT!  the "return () >>" below convinces GHC to eta-expand
--- this function with the IO state lambda.  Otherwise we end up with
--- closures in all the case branches.
 layLeft :: BufHandle -> Doc -> IO ()
 layLeft b _ | b `seq` False  = undefined -- make it strict in b
 layLeft _ NoDoc              = error "layLeft: NoDoc"
-layLeft b (Union p q)        = return () >> layLeft b (first p q)
-layLeft b (Nest _ p)         = return () >> layLeft b p
+layLeft b (Union p q)        = layLeft b (first p q)
+layLeft b (Nest _ p)         = layLeft b p
 layLeft b Empty              = bPutChar b '\n'
 layLeft b (NilAbove p)       = bPutChar b '\n' >> layLeft b p
 layLeft b (TextBeside s _ p) = put b s >> layLeft b p