Re: [Nmh-workers] More fun with charset functions

[Ken Hornstein]

>> Isn't that just punting the problem downstream, forcing others to
>> investigate unknown-8bit, probably with suspicion; "Is it trying to
>> trigger some flaw?"?  Better to plug foulness at the source and refuse
>> to produce it IMO.  I can't see why it's a valid thing to attempt in
>> the modern day.
>
>I'm with Ralph on this.  After two decades of MIME, it is inexcusable to
>get this wrong.

I can understand this thinking (although it seems like we would be the only
MUA who actually does that), but architecturally it's a little hard now,
since write_charset_8bit() currently is always expected to return something
valid.  We could call adios() from inside of that, but doing that seems
wrong to me as well.

--Ken