[antlr-interest] Please help a newbie!

[Peter Ashford]

Richard Clark wrote:

>On Aug 3, 2004, at 15:06, Peter Ashford wrote:
>
>  
>
>>Ah... I did wonder if I might have been using the AST gratutiously just
>>because I didn't understand the basics properly :o)  If I can bug you
>>some more - how would I go about parsing the expressions without using
>>the tree?
>>
>>    
>>
>
>Look at Terrence's "beginner's tutorial" -- it shows an expression 
>parser w/o a tree walker.
><http://www.cs.usfca.edu/~parrt/course/652/lectures/antlr.html>
>
>...Richard
>  
>

Thanks for that - I've gone through an refactored my grammar to only use 
the one parser and it works fine (a bit simpler too for my simple 
purposes :o))

And so now I 've reached my next problem :o)

I've added variable assignments to the grammar but they don't work.  The 
scanner reports unexpected char on the first character of an identifier 
(e.g. "myval = 123" => "line 1:1: unexpected char 'm'")

I noticed the documentation talks about the testLiterals options, but 
I've used that and it doesn 't seem to help.

- here's the parser rule:

    assign
    { float f; }
            : i:ID EQ f=f_expr                 { assignTo(i->getText(),
    f); }
            ;


And here's the lexer rule:

    protected
    ALPHA  : ('a'..'z' 'A'..'Z');

    ID
    options {
      paraphrase = "an identifier";
      testLiterals = true;
    }
           : ALPHA (ALPHA | DIGIT)*;


The whole grammar is printed below:

---------------------------------------------------------------------------

header {
    extern void assignTo(std::string id, float value);   
}

options {
    language="Cpp";   
    genHashLines = true;  
}


//=======================================================================
class MyParser extends Parser;

script returns [std::string value]
{std::string s; }
        : (assign)* s=result               { value = s; }     
        ;
     
assign
{ float f; }
        : i:ID EQ f=f_expr                 { assignTo(i->getText(), f); }
        ;

result returns [std::string value]
{float f;}
        : RESULT EQ f=f_expr               { char buf[127];
                                             sprintf(buf,"%f", f);
                                             value = std::string(buf); }
        ;

f_expr returns [float value=0]
{float f;}
        : value=f_mexpr
          (   PLUS  f=f_mexpr              { value += f; }
            | MINUS f=f_mexpr              { value -= f; }
          )*
        ; 


f_mexpr returns [float value=0]
{ float f; }
        : value=f_atom (   MULT f=f_atom   { value *= f;} 
                         | DIV  f=f_atom   { value /= f;}
                       )*
        ;
       
f_atom returns [float value=0]  
{ float f; }
        : n:NUMBER                         { value = (float) 
atof(n->getText().c_str()); }
        | LPAREN f=f_expr RPAREN           { value = f; }
        ;

 
//=======================================================================
class MyLexer extends Lexer;

options {
    k=2;                               // needed for newline junk
    charVocabulary='\u0000'..'\u007F'; // allow ascii
    testLiterals=false;
}

LPAREN : '(' ;
RPAREN : ')' ;
PLUS   : '+' ;
MINUS  : '-' ;
MULT   : '*' ;
DIV    : '/' ;
EQ     : '=' ;

RESULT : "result";

protected
DIGIT  : ('0'..'9');

protected
ALPHA  : ('a'..'z' 'A'..'Z');

NUMBER
options {
  paraphrase = "a number";
}
       : (DIGIT)+ ( | ('.' (DIGIT)*));

ID
options {
  paraphrase = "an identifier";
  testLiterals = true;
}
       : ALPHA (ALPHA | DIGIT)*;

STRING
options {
  paraphrase = "a quoted string";
}
       : '"'! (~'"') '"'!;

COMMENT: "//" (~'\n')* '\n' {_ttype = 
ANTLR_USE_NAMESPACE(antlr)Token::SKIP;}
         ;

WS     : ( ' '
         | '\r' '\n'
         | '\n'
         | '\t'
         )
         {_ttype = ANTLR_USE_NAMESPACE(antlr)Token::SKIP;}
         ;   


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.antlr.org/pipermail/antlr-interest/attachments/20040804/7f0ff038/attachment.html