Back to Questions
1.7k
views3
comments
Problem with FParsec grammar ors <|>
f#
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25

open System
open System.Collections.Generic
open FParsec
open FParsec.Primitives
open FParsec.CharParsers

type Parser(verbs : string seq, targets : string seq) =   //integer
    let int = pint32 .>> ws |>> Expr.Int

    //double precision float
    let double = pfloat .>> ws |>> Expr.Double

    //single precision float
    let float = pfloat .>> (pchar 'F' <|> pchar 'f') .>> ws |>> fun f -> Convert.ToSingle(f) |> Expr.Float

    //number (int, double or float)
    let number = (int <|> double <|> float)

    member this.ParseNumber(input : string) =
        let result = run parseNumber input
        match result with
        | Success(v,_,_) -> v
        | Failure(msg,_,_) ->
            System.Console.WriteLine(msg);
            exit 1
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23

class Program
	{
		static void Main(string[] args)
		{
			try
			{
				Parser parser = new Parser(null, null);
				var rint = parser.ParseInt("1");			// <-- Int 1
				var rfloat = parser.ParseFloat("1.0f");		// <-- Float 1.0f
				var rdouble = parser.ParseDouble("1.0");	// <-- Double 1.0
				var rnumber = parser.ParseNumber("1.0");	// <-- Error about expected input on any other that the first in the list
				var result = parser.Parse("crm cnn");
				Console.WriteLine(result.ToString());
			}

			catch (Exception ex)
			{
				Console.WriteLine(ex.Message);
			}

		}

	}

I am getting an error when trying to match on my 'number' parser that ors together the int, float and double parsers. Separately they work fine, but in together I get an error for which ever ones are not first in the list. ie int <|> float <|> double would error on input to match a float or double...

Thanks,

Scott

.

let parseNumber = ws >>. number .>> eof

By on 8/7/2011 1:25 AM ()

When you run the parser 

1

 (pint32 <|> pfloat)

on "1.0", the integer parser will succeed after having parsed "1" and the pfloat will never be tried.

The <|> operator doest not implement a longest match rule, and it only tries the second parser if the first parser hasn't consumed input before failing. This is explained in more detail here and here. So, if you want to parse alternative number formats, you either have to make sure that the parser for each individual number format can only parse numbers of its specific format and doesn't fail after consuming input, or -- when the number format allows for it -- build your parser with the help of FParsec's numberLiteral parser.

By on 8/7/2011 4:24 AM ()

Thanks Stephan. That makes sense, I will see if I can rewrite the rules to account for that. ;)

By on 8/7/2011 9:47 AM ()

    Topic tags

    IntelliFactory Offices Copyright (c) 2011-2012 IntelliFactory. All rights reserved.
    Home | Products | Consulting | Trainings | Blogs | Jobs | Contact Us | Terms of Use | Privacy Policy | Cookie Policy     		Built with WebSharper