Hallo Rob!

 RS> I think that's a completely different issue.

Not really.  The fact that nothing is ever done about these "bugs" when raised 
with suitable evidence isn't going to change anything if my observations have 
been correct over the decades.

 RS> I agree: tossers probably should not be stripping 0x8d's.

Amen.  There ought to be a law against it.

 RS> I don't think that has anything to do with "CHRS: UTF-8 2" vs
 RS> "CHRS: UTF-8 4".

It doesn't matter.  4, 2 or any other number will change nothing in the case of 
utf8.  The 0x8d training bytes will all be stripped, as theu will for all the 
8-bit characters.  I posted a CP866 as an example and lo and behold the single 
16-bit charater with the trailing bit 0x8d was deleted.  The only differece is 
that in utf8 those bytes matter way more than they do in purely 8-bit land.

Het leven is goed,
Maurice

... Huil niet om mij, ik heb vi.
--- GNU bash, version 5.0.7(1)-release (x86_64-pc-linux-gnu)