Tony Finch writes:
It might confuse software that analyses Received: fields, causing it
to fail to find all of them.
Yes.
However there is plenty of software out there which already
interleaves other fields amongst Received: fields, so robust code
already has to cope with it.
Right now, it's sensible to consider only the received fields before the
first other field, and assume that any later fields have been processed
by a nonstandard randomizer.
The proposal seems to invalidate that simple heuristic.
Arnt