RE: -ddump-parsed and infix operators

You're right: they are passed left-associatively, and fixed up later. That way the parser does not need to know about associativity and precedence. So it's not a bug. Just something you need to know about -ddump-parsed! Simon | -----Original Message----- | From: glasgow-haskell-users-bounces@haskell.org [mailto:glasgow-haskell-users- | bounces@haskell.org] On Behalf Of Roberto Zunino | Sent: 14 December 2004 19:37 | To: Glasgow-haskell-users@haskell.org | Subject: -ddump-parsed and infix operators | | I noticed that GHC 6.2.2 always parse infix operators as they were | left-associative. For example: | | %%%%% | Prelude> :info : | -- : is a data constructor | infixr 5 : | (:) :: forall a. a -> [a] -> [a] | Prelude> 1:2:3:[] | | ==================== Parser ==================== | ((1 GHC.Base.: 2) GHC.Base.: 3) GHC.Base.: GHC.Base.[] | | | [1,2,3] | %%%%% | | I suppose that the parse tree is then "fixed" in a later stage. | | (Or maybe the parse tree is fine but it gets unparsed in the wrong way?) | | However, the output above is a bit puzzling to me. Should I report this | as a minor bug or is this the intended output? | | Thanks to your attention, | Zun.
participants (1)
-
Simon Peyton-Jones