Hello! We’ve found a small issue that we think might be a bug in mime4j (or
maybe we’re just missing some setting somewhere)…
We are trying to parse an eml file where one of the recipients has a name
containing Korean characters. If the `To:` header is base64-encoded, it’s fine;
if it’s not base64-encoded, it fails to be read the name correctly.
Say we have an eml file with this header:
To: =?UTF-8?B?7Iuc7ZeY?= <koreant...@example.com>
This parses fine.
But this header the name is parsed incorrectly:
To: "시험" <koreant...@example.com>
I am using org.apache.james:apache-mime4j:0.8.10, and this is some sample Java
code to that produces the issue:
var is =
Main.class.getClassLoader().getResourceAsStream("not-base64-encoded.eml");
var parsed = new DefaultMessageBuilder().parseMessage(is);
var to = ((Mailbox) parsed.getTo().get(0)).getName();
assertEquals("시험", to);
Note that to generate an eml with the two different possible headers above we
are using org.eclipse.angus:angus-mail:2.0.2. The default behaviour there is to
generate the base64-encoded header. However if you set the mail.mime.allowutf8
property in the SMTP session to true then it creates the not-base64-encoded
header, which causes the problem for mime4j.
Does this look like a bug in mime4j? Or is there something we’re not setting
correctly?
Thanks in advance,
Yoni Gibbs.
