Having a hard time with my first grammar

[Joe Bryant]

I am creating my first grammar, and I thought it was pretty simple. It is a line oriented data file I need to parse, and I am stuck on getting the first line to be recognized. Assuming the first line contains....

KEYWORD:18.2

And my grammar has:

grammar PCG;

character       :   (line? EOL) + ;

line            :   id ':' DECIMAL
                |   id ':' ANY
                |   BLANK_LINE
                |   COMMENT_LINE
                |   id '|' DECIMAL
                ;

id              :   ID ;

ANY             :   ~[\r\n]+;

BLANK_LINE      :   [ \t]+;

TAG             :   [a-zA-Z]+;

COMMENT_LINE    :   '#' ~[\r\n]*;

EOL             :   [\r\n]+ ->skip;

DECIMAL         :   ([0-9]+ ('.' [0-9]+)?)
                    |       ('.' [0-9]+)
                    ;

ID              :   [a-zA-Z]+ ;

I always get the message :

line 1:0 mismatched input 'KEYWORD:18.2' expecting {BLANK_LINE, COMMENT_LINE, EOL, ID}

Any help would be appreciated to get me over my startup hump.

[Jeff Saremi]

Joe,

I too am a starter. I am going to give you an advice that someone gave me on this forum.

Forget about all your parsing rules.

Write a grammar for lexer only. (lexer grammar MyGrammar;) which contains only the uppercase rules above

Inside your test runner, add a block that would print out all your tokens after scanning.

Example in C#:

            var lexer = new MyPreProcessorLexer(new AntlrInputStream(input));
            var tokenStream = new CommonTokenStream(lexer);
            tokenStream.Fill();
            foreach (var token in tokenStream.GetTokens())
            {
                Console.WriteLine(token);
            }

Then examine your token types/values with your input letter by letter and word by word. See if they come out in the order expected and the type you wanted them.

Until you are 100% satisfied do not move to the parsing stage.

Jeff

[Joe Bryant]

That was amazing advice, and it is indeed helping. Thank you very much.