ActionMailer decoding multi-part w/ charset

[digger69]

I am using ActionMailer to parse incoming email from gmail. I found
that multi-part emails coming from Outlook clients are encoded with
charset='Windows-1252'. After creating a mail object (see
Receiver.receive() below) each part.body has charset='UTF-8' and
defaulted_charset=true. When I grab part.body.to_s the characters
aren't encoded correctly (hyphen encoded as =96 results in non-
printable char in UTF-8 string from to_s). I was assuming it would
grab the charset off the part header, but didn't see anything in
initialize that would do that (seems biased for composing email rather
than parsing existing).

Is this a limitation, defect, or just my ignorance? Thanks for any
help. As a workaround, I'm calling force_encoding() on the string and
then explicitly encoding to UTF-8.

Mark

Rails 3.0.10

class Receiver < ActionMailer::Base
def receive(email)
email
end
end

[Andrew France]

Hi Mark,

I recently came across a similar encoding problem with email and used #force_encoding as you describe. The issue is with the Mail gem that ActionMailer uses.

The gem currently has many encoding issues filed and it's hard to see if this exact MIME part encoding problem has been raised.

Regards,

Andrew

[digger69]

Hi Andrew, thanks for the note. I found a google group for mail and
found a related thread:
http://groups.google.com/group/mail-ruby/browse_thread/thread/fdf0e9ef275cf327

Looks like stepchud has a few patches for this:
https://github.com/stepchud/mail/commits/actionmailer-2.3.8_compatible

but hasn't gotten around to submitting pull request.

Best,
Mark