IMP3.0 encoding of the Subject: line

[Toomas Aas]

Hello!

Our webmail site is currently running IMP 2.2.5, and I am in the 
process of hauling it over to IMP 3.0, along with translating 
IMP3 into Estonian. I feel like I'm 90% 'there', but I have one 
nasty problem.

The problem is with the encoding of Subject: line. When sender uses 
umlauts (üõöä) in the Subject: line, the actual Subject line sent 
looks something like this (example 1):

Subject: =?ISO-8859-4?B?/PX25CBtaXMgamFtYQ==?= see on !?

(As you can see, I've configured IMP to use ISO-8859-4 charset for 
Estonian.)

The recipients who also use IMP 3.0 can see this subject line OK. 
But for some 'external' recipients, the subject line appears 
garbled.

If the message with umlauts in subject is sent using IMP 2.2.5, the 
subject line looks something like this (example 2, line wrapping 
courtesy of my MUA):

Subject:
=?ISO-8859-4?Q?IMP2:_t=E4pit=E4h?==?ISO-8859-4?Q?tedega_kiri_=FC=F5=
F6=E4?=

The same recipients who see the IMP3.0 subject line as 'garbage', 
can read this IMP2.2.5 subject line just fine.

The visible difference is that example 1 starts with 
=?ISO-8859-4?B? and example 2 starts with =?ISO-8859-4?Q?. I 
ass*u*me that ?B? means Binary and ?Q? means Quoted-printable. 
Thus, IMP3 encodes the subject line as Binary which causes it being 
seen as 'garbage' by some mail readers, while IMP2 encodes it as 
Quoted-printable, which seems to be OK.

How/where does IMP decide which encoding to use? How can I change 
it? Will it help?

Thanks in advance,
--
Toomas Aas | toomas.aas@raad.tartu.ee | http://www.raad.tartu.ee/~toomas/
* Tell me what you need, and I'll tell you how to get along without it.