Fix STARTTLS cap withouth validating if it's mandatory.

[jbiosca78]

Describe your changes

testssl doesn't check if STARTTLS is mandatory and issues a message to warn users.
But instead of raising a warning message, it caps grade to T, which sets the score to 0. This is incorrect and unfair to users with security concerns.

What is your pull request about?

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• Typo fix
• Documentation update
• Update of other files

If it's a code change please check the boxes which are applicable

• For the main program: My edits contain no tabs and the indentation is five spaces
• I've read CONTRIBUTING.md and Coding_Convention.md
• I have tested this fix against >=2 hosts and I couldn't spot a problem
• I have tested this new feature against >=2 hosts which show this feature and >=2 host which does not (in order to avoid side effects) . I couldn't spot a problem
• For the new feature I have made corresponding changes to the documentation and / or to help()
• If it's a bigger change: I added myself to CREDITS.md (alphabetical order) and the change to CHANGELOG.md

[drwetter]

But instead of raising a warning message, it caps grade to T, which sets the score to 0. This is incorrect and unfair to users with security concerns.

Au contraire. This is fair to users as STARTTLS involves an upgrade from clear text to TLS which can be intercepted by a man in the middle. A server or a client has no or minimal means to detect this. There are a few tickets about this.

Another thing is the rating. We basically use the one from SSLlabs (with permission). STARTTLS wasn't in scope of that, it's our own addition.

I'll let this hang for a while to reconsider the rating and discussion.

[drwetter]

I'll close this in favor of #2564. Please see comment in the code.