This lab cannot be done with our current knowledge.

What is needed is this:

A lexer rule consumes the first line. Then,
the lexer starts using a different set of rules
for the remaining lines. This is accomplished
using modes. See Lab10.