Verification change for charset

[ssmith-techempower]

Previously, we would flag responses that set a charset as warnings as this extra header data was unnecessary and may impact performance. After some internal discussions, it's been decided that the charset should be considered mandatory for text/html, which is the content type for the fortunes test. This PR changes the verification script to reflect this. Failure to include charset for text/html or including charset for other content types is flagged as a warning.

[NateBrady23]

   WARN for http://127.0.0.1:8080/plaintext

     Content encoding found "text/plain; charset=utf-8" where "text/plain" is acceptable.
     Additional response bytes may negatively affect benchmark performance.
     See http://frameworkbenchmarks.readthedocs.org/en/latest/Project-Information/Framework-Tests/#specific-test-requirements

@ssmith-techempower then this should not be a warning, correct?

Edit: Ah, maybe I have misunderstood. I assumed we were taking off the warning for all text responses, including text/plain as it has the same implications on browser security as text/html.

[msmith-techempower]

@nbrady-techempower I do not remember the specifics, but the comment in the code is what was decided upon: # "text/html" requires charset to be set. The others do not I think you're right.

@ssmith-techempower I think you need to remove this warning as it is now a requirement and should therefore be an error.

[ssmith-techempower]

After some internal discussion, we've decided that charset should be optional for plaintext, so we will consider both "text/plain" and "text/plain; charset=utf-8" as being correct.

[msmith-techempower]

Hard to know who this would affect, but since most of the tests we expect to pass are, I say LGTM!