Re: NEW IMPLEMENTATION OF BISON

[Hans Aberg]

At 10:29 +0100 2001/04/08, CJ wrote:
>I am developing a project where I have to  change the Bison algorithms in
>order to use a DR automata instead of LALR  one.  A DR ( discriminiting
>reverse) is a method  developt by  Jose Fortes
> . To be an idea of this method we can  read this in his thesis ..." 

Do have a URL to an electronic copy of this thesis?

-- Otherwise, your idea sounds great. What do you intend to call it? (I'm
not GNU, so you will have to wait for Akim or somebody replying about your
plans.)

The original Bison was written by Corbett; there used to be pixel-copy of
his thesis at:
  http://www.lrde.epita.fr/people/akim/download/

When you are speaking about a "new implementation of Bison", what do you
intend to keep of the original Bison? -- The Bison engine itself produces
some matrices, which can be used to jump around in a bottom-up parsing.
This is then merged with a skeleton file, which results in code that be
compiled with another language (like C).

Now, your proposal sounds as though that you want to replace the engine
that produces the matrices. And the skeleton part is too primitive to
produce much flexibility (that is, one tends to get stuck with one output
language). So it sounds as though everything needs to be rewritten.

  Hans Aberg