Re: push parser

[Joel E. Denny]

On Sun, 31 Dec 2006, Bob Rossi wrote:

> Sure. You've recently made bison much more configurable in terms of a
> parser. So, have you had the chance to add tests that use all the
> different variations?

Some are tested, and some are not.  I believe the most important variation 
that isn't very well tested is the original: pure push parsing without 
pull parsing.  In other words:

  %pure-parser
  %push-parser

Since you're actually using the push parser at the moment (whereas I'm 
just fantasizing about using it later), I figured you might be better able 
to come up with practical use cases to test.

> Should the example just use the push parser or should it switch between
> push and pull mode?

Either or both would be great.

> What exactly is the example for?

More test cases.

> Take a look at the examples Odd originally had, what do you think of
> them?

They look interesting.  Do you have time to convert one or both to test 
cases?