|LALR(1) Lookahead calculation firstname.lastname@example.org (1998-09-18)|
|Re: LALR(1) Lookahead calculation email@example.com (Chris F Clark) (1998-09-22)|
|Re: LALR(1) Lookahead calculation firstname.lastname@example.org (1998-09-22)|
|Re: LALR(1) Lookahead calculation corbett@lupa.Eng.Sun.COM (1998-09-24)|
|Re: LALR(1) Lookahead calculation email@example.com (1998-09-26)|
|Re: LALR(1) Lookahead calculation firstname.lastname@example.org (Chris F Clark) (1998-09-26)|
|From:||corbett@lupa.Eng.Sun.COM (Robert Corbett)|
|Date:||24 Sep 1998 00:19:07 -0400|
|Organization:||Sun Microsystems Computer Corporation|
Chris F Clark <email@example.com> wrote:
>> Right now the generator is LR(0) . . . adapting it to calculate the
>> lookahead tokens required for LALR(1) . . . is there a more simplistic
>> alternative to these algorithms?
>Yes, find David Spector's paper on "state splitting".
State splitting is not a way of computing LALR(1) lookahead sets.
State splitting is a way of producing an LR(1) parser from a parser
based on the LR(0) state set.
State splitting is very hard. I have read several papers purporting
to describe state splitting algorithms. The only one I have seen that
I believe to be correct is the paper "A Practical State Splitting
Algorithm for Constructing LR-Parsers" by Kristensen and Madsen
published by Aarhus University.
Return to the
Search the comp.compilers archives again.