they have (45,67,3330), we have (45,67,3375) (p. 123)
Is this a bug?
A probable reason is that we currently do not do analysis based on first/follow sets, instead we add a descriptor for each possibility. This matters for Г1 since it contains different terminals. What are results for Г2/Г2*?