Which of the following features cannot be captured by context-free grammars?

Which of the following conversions is not possible (algorithmically)?

Given that language $${L_1}$$ is regular and that the language $${L_1} \cap {L_2}$$ is regular is the language $${L_2}$$ is always regular?

The number of sub-strings (of all lengths inclusive) that can be formed from a character string of length $$n$$ is