Should the eq_file docs mention utf8?

With the code that is there now, a mismatch results in a somewhat opaque list of byte values. While it is not too hard to find utf8, mentioning it outright would save the extra step.

I can send a PR if you agree.