- Email message with japanese characters shown as garbage) where the bottom issue is the same as our issue; also

creating our issue in the site to hopefully higher the priority of this issue at Sun.


here is our code:

1 subject = msg.getHeader("Subject", null);
2 if (subject != null) {
3 try {
4 subject = MimeUtility.decodeText(subject);
5 } catch (UnsupportedEncodingException ioEx) {


here is how we get string from email subject and decode it.

- from outlook 2000, send out and email that has subject of :

abc?@?W??def

to our browser based application and outlook2000.

- after the line 1, we get an encoded string like:

String subject= "=?iso-2022-jp?B?YWJjGyRCLSEtOC1qGyhCZGVm?="

and this MIME Base64 encoded chars using iso-2022-jp seems to be fine.

- after decodeText() in line 4, we get:

String subject= "abc���def"

where we expect:

abc?@?W??def

- here is the source of the email received by outlook 2000:

  Subject: =?iso-2022-jp?B?YWJjGyRCLSEtOC1qGyhCZGVm?=

and outlook 2000 can decode it and display "abc?@?W??def" correctly by looking at the email received.
(this confirms that the encoded form of "=?iso-2022-jp?B?YWJjGyRCLSEtOC1qGyhCZGVm?="
has been encoded correctly.)

Thus, decodeText() seems to have garbled the characters of ?@?W?? (Maru1, RomanNumber4, Kabushiki-gaisha) when it converts from iso-2022-jp to Unicode.
we have also confirmed that regular Japanese characters are converted correctly.

found a similar issue in https://javapartner.sun.com/partner/bugs/data/bugs/6173388.html (6173388 JavaMail 1.3.2