How to avoid using lexer.dat & parser.dat

[Moonbiter]

Hello, I am using SableCC 3 in a project having the target to
implement Design by contract in java
(http://sourceforge.net/projects/jdefprog, refer to the SVN code when
interested in taking a look).

I would like to avoid using lexer.dat & parser.dat and instead have
the "big arrays" initialized directly in the code (I have seen the
generated files have the values but they are commented). So I wonder:
there is an option to generate parser.java & lexer.java in a way to
achieve this result?

Thank you in advance,

Federico Tomassetti

_______________________________________________
SableCC-Discussion mailing list
SableCC-Discussion@...
http://lists.sablecc.org/listinfo/sablecc-discussion

[Etienne M. Gagnon]

Hi Federico,

You could uncomment the tables (and comment-out the static initializer
code that loads the .dat files), but you will likely get either a
compiler error, or a virtual machine error. I have created the .dat
files because table initializers are transformed into code, in Java, and
the bytecode specification limits methods to 65K bytes. For bigger
grammars, SableCC generates tables that cause the <init> method (the
bytecode method for static initialization) to exceed that size.

Usually, you should not put generated *.java and *.dat class into your
SVN repository. You should simply put the *.sablecc3 grammars and use an
Ant script (ant.apache.org) to build your application, and copy the
generated *.dat files along the *.class files.

Etienne

Federico Tomassetti wrote:

> Hello, I am using SableCC 3 in a project having the target to
> implement Design by contract in java
> (http://sourceforge.net/projects/jdefprog, refer to the SVN code when
> interested in taking a look).
>
> I would like to avoid using lexer.dat & parser.dat and instead have
> the "big arrays" initialized directly in the code (I have seen the
> generated files have the values but they are commented). So I wonder:
> there is an option to generate parser.java & lexer.java in a way to
> achieve this result

--
Etienne M. Gagnon, Ph.D.
SableCC:                                            http://sablecc.org




_______________________________________________
SableCC-Discussion mailing list
SableCC-Discussion@...
http://lists.sablecc.org/listinfo/sablecc-discussion

[Niklas Matthies]

Even disregarding the 64K issue, it makes sense to have the *.dat
files, as the initialization bytecode is not only slower¹, but also
consumes more space, both on disc and in memory. (Some people actually
use string literals as a workaround.)

-- Niklas Matthies

¹) At least potentially. The SableCC code isn't the most efficient
either.

On Mon 2009-06-01 at 20:51h, Etienne M. Gagnon wrote on sablecc-discussion:

> Hi Federico,
>
> You could uncomment the tables (and comment-out the static initializer
> code that loads the .dat files), but you will likely get either a
> compiler error, or a virtual machine error. I have created the .dat
> files because table initializers are transformed into code, in Java, and
> the bytecode specification limits methods to 65K bytes. For bigger
> grammars, SableCC generates tables that cause the <init> method (the
> bytecode method for static initialization) to exceed that size.
>
> Usually, you should not put generated *.java and *.dat class into your
> SVN repository. You should simply put the *.sablecc3 grammars and use an
> Ant script (ant.apache.org) to build your application, and copy the
> generated *.dat files along the *.class files.
>
> Etienne
>
> Federico Tomassetti wrote:
> > Hello, I am using SableCC 3 in a project having the target to
> > implement Design by contract in java
> > (http://sourceforge.net/projects/jdefprog, refer to the SVN code when
> > interested in taking a look).
> >
> > I would like to avoid using lexer.dat & parser.dat and instead have
> > the "big arrays" initialized directly in the code (I have seen the
> > generated files have the values but they are commented). So I wonder:
> > there is an option to generate parser.java & lexer.java in a way to
> > achieve this result
>
> --
> Etienne M. Gagnon, Ph.D.
> SableCC:                                            http://sablecc.org
>
>

> _______________________________________________
> SableCC-Discussion mailing list
> SableCC-Discussion@...
> http://lists.sablecc.org/listinfo/sablecc-discussion



_______________________________________________
SableCC-Discussion mailing list
SableCC-Discussion@...
http://lists.sablecc.org/listinfo/sablecc-discussion