Invalid mail header encoding
Hi,
today, I stumbled across an issue, that I haven't seen before. It's related to a discussion on the Linux Kernel mailing list with subject [PATCH 1/2] mm/page_alloc: free order-0 pages through PCP in page_frag_free()
, where Mull mangled a header encoding in a non expected way.
Aaron Lu CCed, as usual, the involved parties, with Paweł Staszewski as one of them (note the special "l" in his name). This specific header reads =?UTF-8?q?Pawe=C5=82=20Staszewski?= <pstaszewski@itcare.pl>
at this point, which appears to be well encoded to my humble knowledge..
Mel Gorman briefly acknowledged this mail using Mutt 1.10.1, but now something interesting happened to Paweł's address, as it reads as "=?utf-8:iso-8859-1?B?UGF3ZcWC?= Staszewski" <pstaszewski@itcare.pl>
now. Note the charset portion: utf-8:iso-8859-1
. For some reason, Mutt added another charset to the existing charset. Since it correctly encodes this header as "utf-8" with BASE64 still, I have no idea, what the :iso-8859-1
part is about. You can have multiple encodings in a header, but not in this way.
You might want to check with https://tools.ietf.org/html/rfc2047, if you have any doubts.
I can provide the verbatim mails on request.