There is an issue with the file sql-2003-2.marpa in proof_of_conceot.  In 
line 6 it says:
<SQL_Start_many> ::= <SQL_Start>+ rank => 0
instead of what seems that it should have been 
<SQL_Start_many> ::= <SQL_Start_Sequence>+ rank => 0

That generated the error:

Uncaught exception from user code:
        Could not resolve rule action named 'nonTerminalSemantic'
          Rule was SQL_Start_many ::= SQL_Start +
          Could not fully qualify "nonTerminalSemantic": no resolve package

When fixed it, I've got 

Uncaught exception from user code:
        Unproductive start symbol
        Marpa::R2 exception at 
C:\projects\MarpaX-Languages-SQL-AST-master\tests
ql.pl line 9.

my $g = Marpa::R2::Scanless::G->new({source => \$G});

Thanks
ZA

On Wednesday, December 10, 2014 11:17:12 PM UTC-5, Zeev Atlas wrote:
>
> Thank you
> I am less interested in the lex and grammar (I am less of theoretician) 
> but in the result structure (value) that should now allow me to extract 
>  things like all tables used in a SELECT statement, JOIN conditions, WHERE 
> conditions, etc.  In the past I wrote some local analyzer, using regular 
> expressions and contemplated better ones available on CPAN (all are partial 
> and/or dealing with SQL dialects that are NOT used by my real life 
> clients,) but once I saw the MARPA based tool, I knew that I have to move 
> there.
> I guess that I should, being that I am working on real production code, I 
> should find a way to have a full Oracle BNF (theoretically available if I 
> analyze their documentation and extract that BNF) and SQL Server BNF 
> (appears to be kept in secret Microsoft vaults :(
> In any case, I appreciate the Marpa project and the  SQL/AST package
> ZA.
>
> On Sunday, December 7, 2014 4:38:03 AM UTC-5, Durand Jean-Damien wrote:
>>
>> ps: I'll fix the fact that the XML is invalid - the concept does not 
>> change -;
>>
>>

-- 
You received this message because you are subscribed to the Google Groups 
"marpa parser" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to