notFollowedBy and parsers which don't consume input.

[neongreen]

Currently notFollowedBy always succeeds with parsers that don't consume input:

-- This parser succeeds.
> parseTest (lookAhead (string "a")) "abc"
"a"

-- Therefore this parser should fail – but it doesn't.
> parseTest (notFollowedBy (lookAhead (string "a"))) "abc"
()

Is this bug old enough to be considered a feature? (Even if so, this behavior should probably be documented.) If not, here's a version that works (but no idea how much slower, if at all):

notFollowedBy' :: (Stream s m t, Show a) => ParsecT s u m a -> ParsecT s u m ()
notFollowedBy' p = try $ join $
      do {a <- try p; return (unexpected (show a));}
  <|> return (return ())

[aslatter]

I guess I'm ambivalent, here. It's definitely surprising behavior.

It might be in the "so old it's not a bug" bucket, however if this would break any parser I'm not even sure what they would look like.

[abooij]

I recently wasted some time trying to understand why notFollowedBy . notFollowedBy is not equal to lookAhead (particularly, I had a line of code that said "notFollowedBy eof" which IMHO makes a lot of sense). I think this behavior, while unfortunate, is acceptable, but should at least be documented.

[mrkkrp]

@tulcod, FYI this is fixed in Megaparsec. See here how it works. The examples (and upcoming tests) include notFollowedBy . notFollowedBy = lookAhead property and notFollowedBy eof works just like you would expect.

This particular combinator can be borrowed by Parsec if it's placed in Text.Parsec.Prim module, since I've written it as a primitive combinator, Parsec version would be:

-- | @notFollowedBy p@ only succeeds when parser @p@ fails. This parser
-- does not consume any input and can be used to implement the “longest
-- match” rule.

notFollowedBy :: Stream s m t => ParsecT s u m a -> ParsecT s u m ()
notFollowedBy p = ParsecT $ \s@(State input pos _) _ _ eok eerr -> do
  let cok' _ _ _ = eerr $ newErrorUnknown pos
      cerr'    _ = eok () s $ newErrorUnknown pos
      eok' _ _ _ = eerr $ newErrorUnknown pos
      eerr'    _ = eok () s $ newErrorUnknown pos
  unParser p s cok' cerr' eok' eerr'

Well, this is a bit more primitive than Megaparsec version… To match quality of Megaparsec error messages some additional work is required, so error messages will differ from those I posted.