Comments only (isIrrefutablePat)

compiler/hsSyn/HsPat.hs

@@ -540,8 +540,12 @@ isIrrefutableHsPat :: OutputableBndr id => LPat id -> Bool
 --      (NB: this is not quite the same as the (silly) defn
 --      in 3.17.2 of the Haskell 98 report.)
 --
--- isIrrefutableHsPat returns False if it's in doubt; specifically
--- on a ConPatIn it doesn't know the size of the constructor family
+-- WARNING: isIrrefutableHsPat returns False if it's in doubt.
+-- Specifically on a ConPatIn, which is what it sees for a
+-- (LPat Name) in the renamer, it doesn't know the size of the
+-- constructor family, so it returns False.  Result: only
+-- tuple patterns are considered irrefuable at the renamer stage.
+--
 -- But if it returns True, the pattern is definitely irrefutable
 isIrrefutableHsPat pat
   = go pat

compiler/rename/RnExpr.hs

@@ -1437,6 +1437,11 @@ ado _ctxt []        tail _ = return (tail, emptyNameSet)
 -- the bind form, which would give rise to a Monad constraint.
 ado ctxt [(L _ (BindStmt pat rhs _ _),_)] tail _
   | isIrrefutableHsPat pat, (False,tail') <- needJoin tail
+    -- WARNING: isIrrefutableHsPat on (HsPat Name) doesn't have enough info
+    --          to know which types have only one constructor.  So only
+    --          tuples come out as irrefutable; other single-constructor
+    --          types, and newtypes, will not.  See the code for
+    --          isIrrefuatableHsPat
   = mkApplicativeStmt ctxt [ApplicativeArgOne pat rhs] False tail'
 
 ado _ctxt [(one,_)] tail _ = return (one:tail, emptyNameSet)