Re: gnubol: [leclay@ibm.net: Can I help?]

[Ralph Jones <leclay@ibm.net>]

Tim,

One more thing.

I have a pcode interpreter, or virtual machine or virtual machine if you are
Javan, that we could use for testing the grammar.  However, it does not have an
index access method.

I might suggest that we first produce code for Java virtual machine, then we can
go back and produce C code after the virtual machine version is debugged.  We
could translate the byte code into C as the final step.  I'm not all that
familiar with Java, but what do you think?

Regards,
Ralph

Tim Josling wrote:

> Ralph,
>
> Your help is welcome. You can have a look at recent discussion -
> follow the links to the mailing list archive from
> <http://www.gnu.org/software/cobol/>. You will see we are looking
> at two models for the grammar, one using bison (available from
> <http://timjosling.homepage.com>) and another using pccts.
>
> The bison grammar which I wrote has not been fully debugged, and
> only includes the cobol nucleus plus a few other odds and ends.
> It started as a test of concept - can you do a cobol grammar with
> bison? The answer appears to be "yes, provided you don't mind
> coding quite a few hacks, which are all documented at the start
> of the grammar.
>
> Any comments on the grammar would be welcome, and if you want to
> write some code for the project, we can help you pick a task. If
> you can give us some more material on where your skills lie
> (languages you know for example, would you be interested in doing
> some bison work).
>
> (A minor point but in my understanding the bison hairy parser
> supports the same grammars as bison simple, but as reentrant
> code. If you are just relying on the parser you either need
> hacks, lexer feedback, predicates or backtracking to parse cobol
> IMO.)
>
> We have a good amount of COBOL expertise, although more is always
> useful as there are so many platforms etc. I think the critical
> resource is people to write and test code. To that end I am
> looking at building a subset of COBOL in phase I so that people
> can write some of the runtime and maybe even the compiler (though
> not the core parser) in COBOL. This is discussed in the mailing
> list archives.
>
> Tim Josling
>
> > X-Accept-Language: en
> > To: tweedy@lusars.net
> > Subject: Can I help?
> >
> > Dear Ms Tweedy,
> >
> > Can I help?
> >
> > I've been doing COBOL, on & off, for 34 years.  Starting with COBOL B on
> > a 16k Honeywell H120 & 3 tape drives.  My COBOL class was on a Burroughs
> > B280 improved, with 4k & 4 tape drives.  I started writing COBOL when
> > you started your compile & hoped it'd be finished when you got back from
> > lunch.
> >
> > How's the BNF definition of the grammar coming along?  Are you using
> > Bison Simple, 1 look ahead token, or Bison Hairy, 2 look ahead tokens?
> > I've never attempted Hairy.
> >
> > I've also used yacc for around 6 years.  The primary project was a
> > compiler for the Advantis proprietary Call Management Language (CAMEL).
> > I've used it for various other smaller tools development projects.
> >
> > Regards,
> > Ralph Jones
> >
> > ----- End forwarded message -----
>
> --
> This message was sent through the gnu-cobol mailing list.  To remove yourself
> from this mailing list, send a message to majordomo@lusars.net with the
> words "unsubscribe gnu-cobol" in the message body.  For more information on
> the GNU COBOL project, send mail to gnu-cobol-owner@lusars.net.


--
This message was sent through the gnu-cobol mailing list.  To remove yourself
from this mailing list, send a message to majordomo@lusars.net with the
words "unsubscribe gnu-cobol" in the message body.  For more information on
the GNU COBOL project, send mail to gnu-cobol-owner@lusars.net.