Re: LR Grammars not in LALR(1) or LR(1)

thp@cs.ucr.edu wrote:
>
> ... Almost every parser-
> >generation tool has some form of disambiguation rules, so nobody in
> >their right mind submits an LALR(1) grammar to, say, YACC. Rather,
> >they submit a more intuitive and compact ambiguous grammar, one that
> >reflects the semantics of the language, and then use the disambuation
> >rules to make it parsable.
>
> Even though precedence rules can be made a part of the grammar:
>
> I got the references:
> http://www.cs.uu.nl/people/visser/ftp/BSVV02.pdf
> http://citeseer.nj.nec.com/cache/papers/cs/6488/http:zSzzSzftp.cs.uu.nlzSzpeoplezSzvisserzSzpublicationszSz..zSzftpzSzP9707.pdf/visser97scannerles.pdf
>
> But I haven't looked it fully up yet.


This paper describes how disambiguation rules (e.g., priority and
associativity) can be used to filter parse trees produced by a GLR
parser. In fact, this paper is based on our experiences of using a
Scannerless GLR parser in an environment for language prototyping. In
order to facilitate the prototyping of language definitions you do not
want to be hindered by massaging your context-free grammar into some
(LA)LR form. We use a powerful formalism to define both the lexical
and context-free grammar of a programming language in a modular and
declarative manner.


-- Mark


----------------------------------------------------------------
M.G.J. van den Brand,
Department of Software Engineering
CWI
Kruislaan 413, NL-1098 SJ AMSTERDAM, The Netherlands.


Tel___(+31) 20 5924213 WWW____http://www.cwi.nl/~markvdb/
Fax___(+31) 20 5924199 Email__Mark.van.den.Brand@cwi.nl
----------------------------------------------------------------

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