|[11 earlier articles]|
|Re: Backtracking yacc firstname.lastname@example.org (1992-09-21)|
|Re: Backtracking yacc email@example.com (1992-09-21)|
|Re: Backtracking yacc firstname.lastname@example.org (1992-09-23)|
|Re: Backtracking yacc email@example.com (1992-09-23)|
|Re: Backtracking yacc firstname.lastname@example.org (1992-09-25)|
|Re: Backtracking yacc email@example.com (1992-09-25)|
|Re: Backtracking yacc firstname.lastname@example.org (1992-09-25)|
|From:||email@example.com (Martin Neitzel)|
|Organization:||Inst. f. Informatik, TU Braunschweig, FRG|
|Date:||Fri, 25 Sep 1992 20:47:09 GMT|
|Keywords:||LL(1), errors, parse|
> ... The stack of a deterministic left
> parser contains the single viable suffix of the sentential
> form that is begun by what is parsed, while a right parser
> knows at best a set of possible continuations.
Is it possible that there is [again] some invalid intermixing of
"parsers" and "grammars" and their respective properties going on?
As long as the grammar at hand has the LL(1) property, shouldn't both
LL and LR parsers be able to make the same consise predictions?
Wouldn't the "set of possible continuations" just have one element?
Return to the
Search the comp.compilers archives again.