I have implemented a simple Scanner-Generator, which runs properly in all situations.
It's not a code issue, but a question about how to optimize DFA. Minimization for one Accept Node using the Hopcroft's algorithm and Accept Nodes for accepting the same rule worked well, but the algorithm did not minimize the DFA with several different Accept Nodes. For example, if you create a DFA that accepts two "if", "[a-z][a-z0-9]*", and then run Hopcroft's algorithm, the unique Accept Node disappears. Of course, each node was grouped into different groups from the beginning.
I would like to know how to minimize this type of DFA. I would like you to answer even if there is no way to solve these minimization problems.