Log Message:
Update to 3.013

Upstream changes:
version 3.013: Sat 24 Jun 17:25:11 CEST 2023

        Changes:
        - In header lines, leave the actual white-space related to FWS
          intact.  You may want to add s/\t/ / before you print a header.
          GitHub#7 [Jacob Balazer]
        - When the comment in an address has no closing ')', do consume
          the whole remaining line, not fail parsing.
          GitHub#6 [Jacob Balazer]
        - When the charset of a ::Body is not specified, a real attempt is
          made to autodetect the best setting.  For reading, constructing,
          and writing bodies.  GitHub#12 [Jacob Balazer]
          ** This may change the output of your regression tests: previously,
          the default interpretation was 'us-ascii' for reading and 'utf-8'
          for writing.

        Fixes:
        - ::Field::Full consumePhrase() should take minimal encoded
          phrase [Jonas Ryssel]
        - ::Field::Address overloaded cmp() used 'eq' where it should
          have been 'cmp'
        - partNumber when the multipart parent is a nested [Geoffrey Bennett]
        - ::Body::Multipart, even when there is only a single \n after the
          trailing boundery, then there is a epilogue.
          github issue #16 [Jacob Balazer]

        Improvements:
        - run regression-tests with MIME::Entity on devel system.
        - ::Body->decode autodects missing charset.
        - The ability to set the charset detector via ::Body::Encode
          method charsetDetectAlgorithm()
        - New release of User::Identity
        - When the message/rfc822 and has Content-Transfer-Encoding, then do
          not parse it as "nested" but as normal part.
          github issue #17 [Todd Richmond]