Extended context-free parsing of regular right part grammars

Abstract

Regular right part grammar is an alternative to context-free grammar, in which right parts of productions are nondeterministic finite state machines to extend the descriptive power of context-free grammar by including notations for describing repetitions and alternations. However, on its LR parsing, extra work is required to identify the left end of a handle at reduction time because a nonterminal can derive potentially infinite number of strings via a single production. Properties of LALR(k) regular right part grammars and their extended LR(0) automata are examined. An improved method for building efficient extended LALR(k) parsers for such grammars is given, which uses only kernel items of extended LR(0) automata. Neither grammar transformation nor extra readback state of a parser is needed, which is used to construct readback machine to recognize the reverse of the state sequences leading to a reduction. For a reduction by particular production in a state, parser may refer to lookback states in which the parser may be restarted after a reduction. An optimizing algorithm to reduce these references is presented. An efficient method to compute lookahead sets by using only kernel items of extended LR(0) automation is also presented.