Question

I have this formula:

negationExpr
: NEGATION^* atom
;

atom
: ‘a’..’z’ | ‘A’..’Z’;

With grammer rules above, if I input formula ¬¬a, I would get this tree structure:

¬ being the root node,
¬ being left child; a being right child

However, What I would like to have is:
¬ being the root node,
second ¬ being the only child of the above node
a being the only child of the second ¬

Basically, I wat all the NEGATION sign have only have one child, it is possible? I know we could probabely use “rewrite rule” to restructure the tree, but I dont know how to do this.

Any help or advises is appreciated! Thanks!

Answer 1

Rewrite rules can follow each alternative for your parser rule.

rule :
   alt1 -> rewriteRule1
 | alt2 -> rewriteRule2
 ...
 | altN -> rewriteRuleN;

You’ll find that even after your parser grammar is working, you may need to restructure it to generate the correct tree. To address your specific problem, I suggest the following:

negationExpr :
   NEGATION negationExpr 
     -> ^(NEGATION negationExpr)
 | atom 
     -> atom;

This will add a level in the tree for each negation operator. The ^ will create a root for the token immediately following the parentheses and add the result of the next negationExpr rule as the child.