The LR parser of Fig. 4.53 handles the four situations in which the top state is 4 or 5 (which occur when + and * are on top of the stack. respectively) and the next input is + or * in exactly the same way: by calling the routine et , which inserts an id between them. We could easily envision an LR parser for expressions involving the full , set of arithmetic operators behaving in the same fashion: insert id between the adjacent operators. In certain languages (such as PUI or C but not Fortran or Pascal), it would be wise to treat, in a special way. the case in which I is on top of the stack. and * is the next input. Why? What would be a reasonable course of action for the error corrector to take?
