Special Characters in IMAP Folders

I was always wondering, why folders on an Exchange server, containing special characters in the name, show up that different in the IMAP protocol (e.g. using Mail::IMAPClient in Perl scripts).

But now I understand, this is just the result of the IMAP specification, that uses a modified UTF-7 encoding for non-ASCII characters (see RFC 2060, section 5.1.3).

If you ever need to automatically convert folder names into that modified UTF-7, the Perl module Encode::IMAPUTF7 may render useful.

The German Umlauts are encoded in the following way:

ISO-8859-15	IMAP-UTF-7
ä	&AOQ-
ö	&APY-
ü	&APw-
Ä	&AMQ-
Ö	&ARN-
Ü	&ANw-
ß	&AN8-

Review 2024

Revisiting this topic in 2024, I first tried to get a Perl with that Module running in a container. And it was easier than expected.

• Open Alpine Linux: docker run -it --rm docker.io/library/alpine

• Install the Perl module: apk install perl-encode-imaputf7

• To encode into IMAP-UTF-7 in a UTF-8 enabled shell:

  perl -MEncode::IMAPUTF7 \
    -e '
      print Encode::IMAPUTF7::encode(
        "IMAP-UTF-7",
        Encode::decode("UTF-8", "ä ö ü Ä Ö Ü ß")
      ), "\n"
    '
  
  # Output: &AOQ- &APY- &APw- &AMQ- &ANY- &ANw- &AN8-
  

• Have fun with even more diacritics.

  perl -MEncode::IMAPUTF7 \
    -e '
      print Encode::IMAPUTF7::encode(
        "IMAP-UTF-7",
        Encode::decode("UTF-8", "á à á à â é è ø")
      ), "\n"
    '
  
  # Output: &AOE- &AOA- &AOE- &AOA- &AOI- &AOk- &AOg- &APg-
  

• It’s interesting how the & is handled, just the prefix & and suffix - with nothing inbetween.

  perl -MEncode::IMAPUTF7 \
    -e '
      print Encode::IMAPUTF7::encode(
        "IMAP-UTF-7",
        Encode::decode("UTF-8", "& a&b")
      ), "\n"
    '
  
  # Output: &- a&-b