Re: any lr(1) wizards still there?

[Neil Jerram]

>>>>> "Martin" == Martin Grabmueller <address@hidden> writes:

    Martin> Hello list, today I had a look at the old stuff in the
    Martin> guile-rgx-ctax module, especially the lr(1)-parser
    Martin> generator contained in the lang subdirectory.  At first it
    Martin> seemed easy to port it to Guile 1.5, but then I found out
    Martin> that the source is from times when #f and '() was both
    Martin> false.  In quite some places it was easy to fix, but in
    Martin> others...

    Martin> Deep in the guts of the LR(1)-generator-phase there is a
    Martin> lot of code which tries to take the CAR of #t, and I
    Martin> managed to localize where, but when I tried to fix it, I
    Martin> felt the brain-slugs attacking and decided to C-c k <RET>.

    Martin> So the question is: Is there anybody around who eats
    Martin> sophisticated optimized fancy generalized high-level
    Martin> Algorithms for breakfast, or maybe even one of the
    Martin> original authors?

ISTR that the version I looked at (from the guile-lang-allover
package) used the modern syntax for module names.  Extrapolating
wildly, this suggests that Andrew Archibald did some work on this
code, and that he might also have resolved the nil/#f issues.

So, if you can still find guile-lang-allover, it might hold some
clues.

        Neil