[antlr-interest] How do you implement conditional blocks in tree grammars?

[Kurt Otte]

That is pretty slick.  I wonder if I have run across a bug in the C target.

When I try this.

conditional
  : ^(CONDITIONAL exp=expression ^(IF ^(BLOCK .) ) )
  ;

Feed it this AST

(CONDITIONAL (DECIMAL 0) (IF (BLOCK (foo (NAME bar)))))

I get this error

ERROR: -Imaginary-:0  Unexpected node,syntax error, near DO

Where 'DO' is Down token.

Does this seem to point a bug in C Runtime or am I missing something else?

Thanks,

Kurt


On Wed, Sep 17, 2008 at 5:44 PM, Terence Parr <parrt at cs.usfca.edu> wrote:
> Hi Kurt,Here are two simple rules:
>
> ifstat
>     :   ^('if' c=expr s=. e=.?) // ^('if' expr stat stat?)
>         {interp.ifstat(this, (Boolean)$c.value, $s, $e);}
>     ;
>
> whilestat
>     :   ^('while' c=. s=.) // ^('while' expr stat)
>         {interp.whilestat(this, $c, $s);}
>     ;
>
> and here is the ifstat in the interpreter core:
>
>     public void ifstat(Eval evaluator, Boolean condition, SLAST stat,
> SLAST elsestat)
>         throws RecognitionException
>     {
>         nodes.push(nodes.index());
>         if ( condition.booleanValue() ) {
>             nodes.seek(stat.streamIndex);
>             evaluator.stat();
>         }
>         else if ( elsestat!=null ) {
>             nodes.seek(elsestat.streamIndex);
>             evaluator.stat();
>         }
>         nodes.pop(); // start after IF
>     }
>
> Slick, no? :)
>
> Ter
> On Sep 17, 2008, at 4:38 PM, Kurt Otte wrote:
>
>> So would the expected model be:
>>
>> ^((CONDITIONAL exp=expression
>>    {($exp.bool)}?=>  ^(IF ^(BLOCK statement+) )
>>    {(!$exp.bool)}?=>  ^(IF ^(BLOCK .) )
>>  )
>>
>> So if the expression is true, the statements are processed as
>> statements, and if the expression is false they are slurped up through
>> a wildcard?
>>
>> Do you recommend using a predicate (like shown above) to do the
>> switching or what is the best plan?  Is there a code snippet or
>> example you could point me to?
>>
>> Thanks so much for your help and quick responses,
>>
>> Kurt
>>
>>
>> On Wed, Sep 17, 2008 at 5:18 PM, Terence Parr <parrt at cs.usfca.edu>
>> wrote:
>>> Oh,I did that in a tree grammar so it should work for complete
>>> subtrees. At least, I did that in an interpreter of mine.
>>> T
>>>
>>> List: http://www.antlr.org/mailman/listinfo/antlr-interest
>>> Unsubscribe: http://www.antlr.org/mailman/options/antlr-interest/your-email-address
>>>
>>>
>>
>> List: http://www.antlr.org/mailman/listinfo/antlr-interest
>> Unsubscribe: http://www.antlr.org/mailman/options/antlr-interest/your-email-address
>>
>
>
> List: http://www.antlr.org/mailman/listinfo/antlr-interest
> Unsubscribe: http://www.antlr.org/mailman/options/antlr-interest/your-email-address
>
>