[antlr-interest] V3 Rule Action allowing inline tree building?

[Lucien Antlr]

Hi Sam/ and list,

That's a great response, and pretty clean too.  Thank you for your time in helping me.

I was hoping to have the same readability as the original grammar that was available in V2.  So I concocted the following nasty alternative for those interested...

// Create what ever tree you want here...
fragment dummyAST
     : i=ID -> ^( LITERAL_enum ID $i )
 ;
 
enumerator    :       
                'int' /*i=ID*/      z=dummyAST

         { // do something with internal tree from dummyAST
              System.out.println(((CommonTree)($z.tree)).toStringTree() +" ..............................\n");
          }
                   (a=EQ b=expr)?  // = e
                    (c=EQ)?  // =
                    ';'

                -> ^('int' ($a $b)? ($c)? )  // rewrite the output, dropping the dummyAST output
        ;


 Cheers,
Luc




________________________________
From: Sam Harwell <sharwell at pixelminegames.com>
To: Lucien Antlr <lucien_antlr at yahoo.com>; antlr-interest at antlr.org
Sent: Thursday, March 19, 2009 3:15:04 AM
Subject: RE: [antlr-interest] V3 Rule Action allowing inline tree building?


Once you get the hang of it, you can do it by hand without too much trouble. For example, in converting a v2 grammar to v3, I encountered this in the members section:
 
stringAlias =
        #(#[BLOCK], #(#[ALT], #[STRING_LITERAL], #[EOA]), #[EOB]);
 
Which became this (you can use CommonTreeinstead of GrammarAST). Note that this uses the CSharp3 target so the function names are capitalized, unlike the Java target.
 
/*
 * stringAlias = ^(BLOCK[] ^(ALT[] STRING_LITERAL[] EOA[]) EOB[])
 */
stringAlias = (GrammarAST)adaptor.Create( BLOCK, "BLOCK" );
{
    GrammarAST alt = (GrammarAST)adaptor.Create( ALT, "ALT" );
    adaptor.AddChild( alt, adaptor.Create( STRING_LITERAL, "STRING_LITERAL" ) );
    adaptor.AddChild( alt, adaptor.Create( EOA, "EOA" ) );
    adaptor.AddChild( stringAlias, alt );
}
adaptor.AddChild( stringAlias, adaptor.Create( EOB, "EOB" ) );
 
In your case, you’d use something like this:
 
CommonTreetree = (CommonTree)adaptor.Nil();
adaptor.AddChild( tree, adaptor.Create( LITERAL_enum, "enum" ) );
adaptor.AddChild( tree, adaptor.Create( ID, enumName ) );
 
Sam
 
From:antlr-interest-bounces at antlr.org [mailto:antlr-interest-bounces at antlr.org] On Behalf Of Lucien Antlr
Sent: Wednesday, March 18, 2009 11:53 AM
To: antlr-interest at antlr.org
Subject: [antlr-interest] V3 Rule Action allowing inline tree building?
 
Hi folks,
 
I saw back in 2005 there was a bit of discussion around inline tree building.   ( linked here 
http://markmail.org/message/7zigjknpakugflje ).  For practice I am converting the freely available C v2 ANTLR .g  and I came over a rule that has me stuck...
 
It keeps the internal symbol-table by adding part of a tree into the symbol table along with the C symbol (in this case enum).
 
Normally a v2 rewrite would look like below (from stdcparser.g)
 
enumerator[String enumName]
        :       i:ID                { symbolTable.add(  i.getText(),
                                                        #(   null,
                                                            #[LITERAL_enum, "enum"],
                                                            #[ ID, enumName]
                                                         )
                                                     );
                                    }
                (ASSIGN constExpr)?
        ;
#( ) under v2 calls ASTFactory.create(), but under v3 it just appears to pass straight through into the Java target?  
 
This is true even when re-written, as I guessed it should be below.
 
enumerator[String enumName]  
        :       i=ID         { symbolTable.add(  $i.text,
                                                        ^(   NULL,
                                                            ^(LITERAL_enum),
                                                            ^(ID $enumName.text)
                                                         )
                                                     );
                                    }
                (ASSIGN constExpr)?
        ;
The only thing I came up with is "rewriting" the rule, and creating a dummy rule inside and using $dummy.tree inside the action.
 
Any thoughts from the more experienced community?
 
Kind Regards,
 
Luc


      
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.antlr.org/pipermail/antlr-interest/attachments/20090324/80072486/attachment.html