Hallo Kees!

 KvE> It looks good in Thunderbird, but it will probably get mangled
 KvE> on the way back. ;((

It looks perfect here.  I'll repost the parts that matter from a utf8 dump of 3
lines from your reply "MSGID: 2:280/5003.4 5b200566";

-={ ':read dump.utf8' starts }=-
 MK> in my pants", "Jag har ekorrar i mina byxor", "У меня в штанах
белка",
 MK> "Έχω σκίουροι στο παντελόνι μου" and "Ich habe
Eichhörnchen in meiner
 MK> Hose", within a single message which covers five languages thus vastly
-={ ':read dump.utf8' ends }=-

Note that is an exact replication from your reply (ie zero jiggery-pokery).  I
notice that Thunderbird does a perfect wordwrap given that it obviously counts
multibyte characters the same as 7-bit ascii characters.  Other web based apps
I've seen don't as they count characters as bytes.  Kudoes to your
implimentation.  :-)

Does the "CHRS: UTF-8 4" kludge make any difference at all?  If so I'll turn it
on here but just so you know it is a total waste of bytes at this end since
utf-8 characters work 100% at this end without any interference or help
required.

Het leven is goed,
Maurice

... Huil niet om mij, ik heb vi.
--- GNU bash, version 4.4.23(1)-release (x86_64-bonnell-linux-gnu)