Remove UTF-7 support for email bodies

[pdg137]

In the last year or two I have started to receive a few legitimate emails encoded as UTF-7 (maybe 1 or 2 per 100,000 emails). I noticed that despite attempting to support UTF-7 email bodies, the Mail gem does not decode these emails properly.

Note that standard UTF-7 uses a + character to initiate a special sequence (see Wikipedia). The version using &, which the Mail gem supports, appears to be the "modified UTF-7" described in RFC 9051 and is only intended for email headers/addresses. It is certainly not a valid encoding for email bodies.

Note that

• Ruby does not support conversion to/from UTF-7:

  irb(main):021> "Hi!".encode(Encoding::UTF_7)
  (irb):21:in `encode': code converter not found (UTF-8 to UTF-7) (Encoding::ConverterNotFoundError)
          from (irb):21:in `<main>'
  
  irb(main):022> "Hi+ACE-".force_encoding(Encoding::UTF_7).encode(Encoding::UTF_8)
  (irb):22:in `encode': code converter not found (UTF-7 to UTF-8) (Encoding::ConverterNotFoundError)
          from (irb):22:in `<main>'

• The Net::IMAP package only supports the "modified UTF-7" version in its decode_utf7 method.
• Microsoft recommends against it.
• The HTML5 standard disallows UTF-7 support.
• We have issues like this related to attempting the decoding: Invalid utf-7 data in body causes exceptions #1404.

Given that we do not support the correct version of UTF-7 now, that Ruby itself does not support it, and there are many standards recommending against it, I think we should completely remove support for modified UTF-7 in email bodies.