The use of context-free grammars to define the syntax of programming languages is complicated by the phenomenon of ambiguity. Ambiguity can be resolved by resolution rules which specify a unique canonical parse[A&J75].
In this thesis, rejections of parsing actions by resolution rule have been studied theoretically. That method transforms parser by generating useless items and got relations. This study have started to trace changes in LR (k) automata as a result of a parser transform, and represent the result of transform by the coressponding grammar. For this purpose, two models are proposed, state item graph and parser description grammar. And grammar generating algorithm with those models is founded.
This method can be used in parser generating system to test the role of resolution rules.