>>>>> In <9eek9bqq1j.fsf@gowron.rz-berlin.mpg.de> Jochen Küpper wrote:

> In my .gnus I have 
> ,----
>| (add-to-list 'mm-charset-synonym-alist (if (and (not (mm-coding-system-p 'unicode-1-1-utf-7))
>|                                                 (mm-coding-system-p 'utf-7))
>|                                            '(unicode-1-1-utf-7 . utf-7)))
> `----

I loaded utf-7.el and tried it.

> However, when sending (S o m) messages with the unicode-1-1-utf-7
> coding system, Gnus gives the following error (and stops):
> ,----
>| mm-encode-body: Invalid coding system: unicode-1-1-utf-7
> `----

Though I couldn't find the way to reproduce it using `S o m', it
seems to be the same to encode the part like the following:

<# part type="text/plain" charset=unicode-1-1-utf-7 disposition=inline>
jugemujugemugokonosurikirekaijarisuigyonosuigyomatsuungyomatsu...

Gnus still confuses `charset' with `coding system' in several
places I think, and the problem is caused by one of them.  Does
the following patch help?