Compare Berkeley yacc to Sun's yacc

Bryan Heil writes


> I would be interested in comments from other users on their
> experience with Berkeley yacc and how they feel it compares to other
> implementations of yacc, particularly Sun's yacc.


I recently was a part of a project where we developed a parser for
Burroughs COBOL. Anyone who has ever taken up a similar exercise will know
how big the grammar is. The development took place on Sun, and predictably
the yacc provided by Sun gave up very soon. We had no choice but to resort
to byacc. As John commented, it turned out to be very fast.


Incidentally, it failed all of a sudden once. Since we had the source we
found that the problem was due to a table being of 'short's and the values
were overflowing. We '#define'd 'short' to 'int' and it worked smoothly
thereafter.


In short, my experience says, go for byacc.


viren.
--

Return to the comp.compilers page.
Search the comp.compilers archives again.