Generation of efficient LALR parsers for regular right part grammars

Yet another generation of LALR parsers for regular right part grammars

“…Note that self conflicts are quite rare in real RRPGs. For instance, no self conflicts arise in the regular right part grammar of PASCAL-S in [11], and examples of stacking conflicts in previous papers do not contain self conflicts except for the example in [13], which was given as a counter example for the method of [7]. Therefore, almost all RRPGs can be parsed by the first method as simply as usual LALR grammars.…”

“…2, (4,4) In this paper we deal with LALR parser for RRPGs. As in [7], the corresponding grammar, which is called ELALR grammar, is defined as follows: For q∈Q, p∈Reduce(q), the set of lookahead symbols of p at q is denoted as lookahead(p,q).…”

“…(A) Transform the ELR(1) grammar to an equivalent LR grammar and apply standard techniques for constructing the LR parser [3], [6]. (B) Build the ELR parser directly from the ELR(1) grammar [1], [4], [7], [8], [9], [10], [12], [13].…”

“…(B1) Add readback states to the parser [1] [4] (B2) Transform to an equivalent RRPG if stacking conflicts occur [8] (B3) Use lookback states at reduction if stacking conflicts occur [7], [10] (B4) Use counters at reduction if stacking conflicts occur [9] [12] (B5) Use path numbers at reduction if stacking conflicts occur [13] Approach (B1) adds inefficiency to the parser. Approach (B2) has some of the same problems as the approach (A) if stacking conflicts occur.…”

How to cover a grammar

“…~, where ~ is a regular expression of terminals and non-terminals. Such grammars go under the names of RTN grammars [8], or extended contextfree grammars [9], or regular right part grammars [10]. Without loss of generality we may restrict the format of the fufite state automata, and stipulate that it have one initial •tale I~' and one final state/~, and only the following type of rules:…”