The terminals the input may continue with on stateStack (bottom first), sorted by name, with the end of the input ahead of a token named $.
The terminals the input may continue with on stateStack (bottom first), sorted by name, with the end of the input ahead of a token named $.
LALR(1) merges lookaheads of states with the same core, so a state may reduce on a terminal that cannot follow there. Each candidate is kept only if replaying the reductions on it reaches a shift or the accept.