[antlr-interest] Nubie: How to approach multiple "required" subrules in any order

[Jim Idle]

The trick is not to try and enforce the semantics by encoding them in the grammar, then you have to put the rules together carefully to ensure that recovery happens. I need to write a wiki article about recovery I think.

 

So, you allow any number of those rules in your parser and then when you walk the AST, count the occurrences and issue a good semantic based error about not being allowed to duplicate them. Without testing it, you would have:

 

big_rule : token_big rset -> ^(ROOT token_big rset) ;

rset  : ( r1 | r2 | r3)* ;

r1 : t1 EQ^ STRING;

r2...

 

While you are getting to know this stuff, do not use 'literals' in your parser, make lexer tokens and use their names:

 

EQ : '=';

// Other keywords/operators

STRING ...

ID ...

COMMENT ...

WS :

ANY : . {issue error - unknown character} ; // Always the last rule in the lexer

 

Jim

 

 

From: antlr-interest-bounces at antlr.org [mailto:antlr-interest-bounces at antlr.org] On Behalf Of gary mazzaferro
Sent: Friday, November 06, 2009 7:41 AM
To: antlr-interest at antlr.org
Subject: [antlr-interest] Nubie: How to approach multiple "required" subrules in any order

 

H,

I'm new to antlr and learning as I'm writing a parser to an ast tree. I'm not sure how to approach this problem ( I think its simple) and was hoping someone could help me out. 

I have a input stream with a token followed by 3 other rules. Each of the 3 "rules" occurs only once for the parse to be correct, but they can arrive in any order. If it fails, it should continue on the next characters after token_big. How do I write the "big_rule" ?
                            

big_rule : token_big  ( rule1 | rule2 | rule3 )  -> " also, what do I put here ?"
                not right >>>  ^      ^        ^

rule1: token1 '=' STRING -> ^('=' token1 STRING);
rule2: token2 '=' STRING -> ^('=' token2 STRING);
rule3: token3 '=' STRING -> ^('=' token3 STRING);

cheers,
gary



-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.antlr.org/pipermail/antlr-interest/attachments/20091106/af7fda1f/attachment.html