[antlr-interest] How to suppress tokens with rewrite rules

[Simon Woods]

OK. Sorted  ... perhaps not the best way (as my lack of knowledge means 
that I'm unaware whether there are better ways to do this) but it seems 
to work...

forstatement
    : ^(FOR ID)
      -> {%{$ID.text}}
    ;

getstatement
    : ^(GET (s+=ID)+)
      -> listcolumns(columns={$s})
//      -> template(ids={$s}) "<ids; separator=\",\">" in-line version

The definitive ref arrived today!


> Hi again
>
> In trying to learn antlr, I've been trying to find out the best way to 
> remove certain tokens through rewrite, but I'm not sure of the best way.
>
> The tree grammar will receive a forstatement "FOR tableX" and a 
> getstatement "GET columnX,columnY ..."
>
> I'm simply trying to suppress the FOR and GET keywords
>
> I have tried with a tree grammar/string template
>
> ===Tree Grammar ===
>
> mainstatement
>     : forstatement getstatement
>       -> 
> translate(forClause={$forstatement.text},getClause={$getstatement.text})
>     ;
>    
> forstatement
>     : ^(FOR ID)
>       -> striptoken(text={$ID.text})
>     ;
>
> getstatement
>     : ^(GET ID+)
>       -> striptoken(text={$ID.text})
>     ;   
>
>
> ==== STG ====
>
> group QL2SQL;
>
> translate(forClause, getClause) ::= <<
> SELECT <getClause; separator=","> FROM <forClause>
>  >>
>
> striptoken(text) ::= <<
> <text>
>  >>
>
> Now I'm *sure* this isn't the best way, nor does it work for the GET ID+ 
> rule as I only get the last get entry in the comma separate value list 
> of columns ...
>
> Thanks again
>
> Simon