Re: Problems with an unrecognised Token

[Luca]

Il 22/04/2011 23.57, Martin Haug ha scritto:
> Hello,
> I'm new to Bison and tried to write my first bison Parser.
> Unfortunately, the Parser only gives my a Syntaxerror.
> It should parse Dates in the Form 'DD.MM.[YY]YY
> hh:mm[:ss[.sss]]' (without the '').
> I use bison version 2.4.3 and gcc version 4.2.4 under Ubuntu 8.04 or
> Fedora 14.
> The Output of Bison and my parser file are attached.
> I'm not sure what other Information you could need, so please help me
> out if you miss something.
> Kind regards
> icehawk
>
>
> _______________________________________________
> address@hidden https://lists.gnu.org/mailman/listinfo/help-bison
Hallo icehawk,
I quikly read your grammar and I noted you defined three tokens without 
semantic values (TWONUM, WHITESPACE and THREENUM, i.e. 258,259 and 260)
but when your lexer reads a date (datum) it just returns a single 'C' 
char (zeichen and zeichenold) rather then the right token value, so 
bison immediately generates a syntax error.
I think your lexer is bugged, please check the lexer output, it should 
returns the right token when the input is a date or a time.

Try to give a look to the following example (Reverse Polish Notation 
Calculator), taken from bison help:
http://www.gnu.org/software/bison/manual/html_node/RPN-Calc.html
your lexer should be close to this one:
http://www.gnu.org/software/bison/manual/html_node/Rpcalc-Lexer.html#Rpcalc-Lexer

Bye,
Luca