Re: unmaintained package?

[Akim Demaille]

>>> "Bob" == Bob Rossi <address@hidden> writes:

 > I would contact the author, if I could figure out how to get the
 > email address. Also, I'm not solely reling on the original
 > poster. I would be willing to do the work myself with a little
 > guidance. I've already contributed to GDB's MI interface, and I've
 > worked with the readline author to get readline working correctly
 > for me in an asynchronous mode.

For the records, I have at hand a printout of a message from Odd Arild
Olsen that details his implementation of a push parser, but I don't
have his address...

 > I would really appreciate some guidance from someone on the project,
 > that could steer me in the correct direction.

Ahem... Read skeletons...

 > The good thing about the asynchronous mode, is that if it's easy to
 > modify the skeleton to work in an asynchronous manner, than you
 > wouldn't need another skeleton that worked in a non-asynchronous
 > manner, it could just call the ascync function.

 > For instance, there is now yyparse(), we could also make a yyaparse (),
 > which would be an asynchronous parse. yyparse () could be changed in the
 > skeleton to be
 >   while (1)
 >     yyaparse ();

 > and then we could export the new yyaparse () function to users, and
 > docuement it. Anyways this was off the top of my head, I'll have to
 > spend some time debugging a simple bison generated parser to understand
 > how it works.

That's also what I had in mind, but the cost of the change in terms of
parse speed, should be evaluated.  Also, take into account the fact
that by default parsers are non pure, which also possibly speeds them
up.  In your case, this is no longer doable, parsers have to be pure,
and maintain their state outside the core function.