Paul J Stevens wrote:
Anton,
I've tested with the message you include below. I see the same strange
wrapping of the from line, but I stil don't get duplicate headers.
Dublicate heades i see if send mail via exim, but via command promt all
rigth.
I create second copy exim configuration and collect debug info about
problem.
Paul J Stevens wrote:
Anton,
Could you please send me the testmessage below as a gzipped
attachment. I'll try to reproduce the problem you're seeing. I
suspect there may be a problem with gmime's parsing of your koi8-r
encoded from address which triggers a bug in the old mimeparser still
in use in the message retrieval chain.
I'd also like to see a trace level 5 of a dbmail-smtp insertion
sequence.
more below...
Anton Nekhoroshikh wrote:
cat test.eml3 | ./dbmail-smtp -f
/usr/local/dbmail.new/bin/dbmail.conf -d [EMAIL PROTECTED]
cat test.eml3:
From:
=?koi8-r?Q?=E1=CE=D4=CF=CE=20=EE=C5=C8=CF=D2=CF=DB=C9=C8=20?=<[EMAIL PROTECTED]>
To: [EMAIL PROTECTED]
Subject: test
Content-Type: text/plain
test mail
cat /tmp/msg:
From:
=?koi8-r?Q?=E1=CE=D4=CF=CE=20=EE=C5=C8=CF=D2=CF=DB=C9=C8=20?=<[EMAIL PROTECTED]
l.ru>
To: [EMAIL PROTECTED]
Subject: test
MIME-Version: 1.0
Content-Type: text/plain
test mail
----------------------------------
why differed messages?
The message insertion chain parses messages through gmime. This is
done for splitting the message header from the body, converting
messages CRLF->LF, calculating rfcsize, building a hashtable of all
headers, etc.
The gmime representation of the header is what explains the changes
in the message header. These change entail: rewrapping of long
headerlines, cleaning up of mime-attribute quoting, and adding a
MIME-Version header when missing. All these changes should be well
within the bounds set by relevant RFC's. If they are not this means
we hit a bug in gmime.
I'll get on this asap.
