Exmh, nmh & decoding of headers.

Hi all.

For a while now, I've been using the nmh 'decode' feature in my
format-files, instead of my former solution which was to pipe
all mail through mimedecode, and store it as 8-bit.

It's working rather well, but I've noticed that when I compose
messages in exmh (sedit) with headers that contain multiple words with
characters that need to be encoded, the resulting header-string will be
decoded by nmh scans as one (rather longish) word.

So, does anyone know if I should start tweaking at exmh or nmh?
I'm guessing that exmh uses an older RFC as base for how it encodes
headers that need encoding, but I'm not deep enough into this yet...

I'll include my format-file, and a few examples.

/Mattias

scan.basic
Description: scan.basic

scan.latin1
Description: scan.latin1

> --- Begin Message ---
>
>
> --- End Message ---
> --- Begin Message ---
> Testing again...
>
> --- End Message ---