Re: [Caml-list] What's wrong with my parser ?

[John Prevost <j.prevost@cs.cmu.edu>]

>>>>> "nf" == Nicolas FRANCOIS (AKA El Bofo) <nicolas.francois@free.fr> writes:

    nf> I define a polynom parser like this : let rec parse = let

  let rec parse =
    let parse_puissance = parser
      | [< ''^'; e = parse_int >] -> e
      | [< >] -> 1
    in
    let parse_x = parser
      | [< ''X'; e = parse_puissance >] -> e
      | [< >] -> 0
    in
    let parse_coeff = parser
      | [< n = R.parse >] -> n
      | [< >] -> R.one
    in
    let parse_monome = parser
      | [< n = R.parse; e = parse_x >] ->
	  monome n e
      | [< e = parse_x >] ->
	  monome R.one e
    in
      parser
	| [< ''+'; m = parse_monome; p = parse >] ->
	    m ++ p
	| [< ''-'; m = parse_monome; p = parse >] ->
	    (opp m) ++ p
	| [< m = parse_monome; p = parse >] ->
	    m ++ p
	| [< >] -> zero


    nf> parse_int and the R.parse functions work OK, and I'll add all
    nf> the "_=parse_space" after. But :

# let p1 = P.parse (Stream.of_string "1+X");;
Stack overflow during evaluation (looping recursion?).

    nf> What's the problem ?

Ahh.  I didn't look closely enough.  The problem can be seen by
looking at what happens when you try to parse the empty stream:

in parse:
  [< >] -> cases 1 and 2 fail, tries [< m = parse_monome; p = parse >]
    in m = parse_monome:
      [< >] -> I assume R.parse fails, so it tries [< e = parse_x >]
        in parse_x
          [< >] -> case 1 fails, case 2 [< >] succeeds
          parse_x returns 0
      parse_monome returns monome R.one 0
  [< m = parse_monome; ... >] succeeded, move to [< ...; p = parse >]

And at this point, parse is called on [< >] again and it loops.
Unfortunately, I can't remember how to explicitly check for the end of
the stream--and it's been removed from the O'Caml manual.  That's what
you have to do to avoid this.  A quick and dirty hack is to have a
special symbol (";", for instance) at the end of every phrase.

John.

-------------------
To unsubscribe, mail caml-list-request@inria.fr Archives: http://caml.inria.fr
Bug reports: http://caml.inria.fr/bin/caml-bugs FAQ: http://caml.inria.fr/FAQ/
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners