Improve selection logic in EmailGrammar::encodeHeader()

[TimWolla]

Since the rewritten quoted-printable encoder in commit 0e39d263df22ebe50cf99aa2c6e8290b9d09b8e8 some inputs are correctly encoded as quoted-printable that previously were not, seeing that quoted-printable encoding did something and selecting it as the desired encoding. However in many cases a simple quoted-string would be sufficient and more readable in the email's source code.

Quoted-printable encoding should only be necessary for:

• Overlong inputs.
• Inputs containing characters that are not in the printable ASCII range (i.e. outside 0x20 - 0x7F).
• Inputs containing ".

But this should be checked in details against the proper RFCs.