Lexer tables are big for any non trivial grammar

[mingodad]

It seems that the lexer generator doesn't do a great job in reducing/minimizing the generated tables and for any non trivial grammar the lexer tables are big like for sql the final parser is around 800KB without managing case insensitive tokens if we make it case insensitive by adding rules like <CREATE: <['C','c']['R','r']['E','e']['A','a']['T','t']['E','e']> > then it becomes even bigger fast.