Re: Looking for shift/reduce conflict analyser

grosch@cocolab.sub.com (Josef Grosch)
Tue, 27 Dec 1994 22:21:00 GMT

          From comp.compilers
Related articles
Looking for shift/reduce conflict analyser rvdpol@inter.nl.net (1994-12-23)
Re: Looking for shift/reduce conflict analyser grosch@cocolab.sub.com (1994-12-27)
Re: Looking for shift/reduce conflict analyser withagen@prl.philips.nl (1995-01-11)
| List of all articles for this month |
Newsgroups: comp.compilers
From: grosch@cocolab.sub.com (Josef Grosch)
Keywords: parse, LALR, tools
Organization: Compilers Central
Date: Tue, 27 Dec 1994 22:21:00 GMT

Denis Kieft (rvdpol@solair1.inter.nl.net) writes:


> Looking for a shift/reduce conflict analyser.
> ...
> In fact, the GMD-Cocktail project LALR-parser generator prints such a tree when
> constructing the parser. But when using the LALR-parser generator on a large
> grammar (4000 lines,definitions, programs and semantic actions not included)
> it dumps core. For grammars of other languages it works fine.


Newer versions of the Cocktail Toolbox contain the LR(1) parser generator
LARK which is a successor of Lalr. It is compatible with Lalr and it
incorporates many improvements. With respect to the analysis of
shift/reduce conflicts LARK has the following advantages:


- the conflict information is computed much faster
- the conflict information is shorter without any loss of details
- the conflict information is correct (as far as I can tell, Lalr
    sometimes computes nonsense)
- the conflict information is computed for huge grammars such as COBOL,
    PL/I, or C++ without any problems


For more information about Lark or Cocktail please contact:


Josef Grosch


Hagsfelder Allee 16
D-76131 Karlsruhe
Germany


Tel.: +49-721-697061
Fax : +49-721-661966
Mail: grosch@cocolab.sub.com
--

Post a followup to this message

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