[antlr-interest] accessing the current node
mzukowski at bco.com
mzukowski at bco.com
Tue Feb 19 13:34:51 PST 2002
I think you can only make Tokens into root nodes, not rules like you are
trying to do with "binaryPointcutOp^".
Also I find it weird that you have LOR and LAND at the same precedence. Try
pointcut:
: pointcuts ((LAND^ {##.setType(AND_POINTCUT);}| LOR^
##.setType(LOR_POINTCUT);) pointcuts)*
;
Also, I'm guessing 'cause you didn't say what was going wrong with the rules
below. Any error messages or wrong tree structures?
Monty
> -----Original Message-----
> From: Greg Smolyn [mailto:smolyn at cs.ubc.ca]
> Sent: Tuesday, February 19, 2002 1:26 PM
> To: antlr-interest at yahoogroups.com
> Subject: Re: [antlr-interest] accessing the current node
>
>
>
> Thanks for the help, Monty.
>
> I just have one other dumb question that is causing me
> difficulty at the
> moment, can anyone help?
>
> Say I have some structure like this:
>
> pointcut1 && pointcut2 && pointcut3
> (yadda yadda)
>
> Currently my grammar is something like (yes, what I'm showing here is
> broken... it's somewhere between what I had and what I want!):
>
> pointcut
> : pointcuts (binaryPointcut)*
> ;
>
> pointcuts
> : "execution" LPAREN! functionSig RPAREN! body
> { ## = #(NPointcut, ##); }
> ;
>
> binaryPointcut
> : binaryPointcutOp^ pointcuts
> { ## = #(NPointcut, ##); }
> ;
>
> binaryPointcutOp
> : LAND
> | LOR
> ;
>
> for reasons of simplicity in my aspect weaver, I want to turn
> the "&&" into a
> type of pointcut itself-- basically I want to create
>
> andPointcut(andPointcut(pointcut1, pointcut2), pointcut3) (I
> *think* that's
> the right order of operations?)./
>
> I'm really stuck in a LALR world right now in my head, and I
> can easily come
> up with a yacc solution. :) Does anyone know the ANTLR way
> to do this?
>
>
> thanks!
>
> --
> Greg Smolyn (smolyn at cs.ubc.ca)
> Software Practices Lab, UBC
> http://www.cs.ubc.ca/labs/spl/
>
>
>
> Your use of Yahoo! Groups is subject to
http://docs.yahoo.com/info/terms/
Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
More information about the antlr-interest
mailing list