Re: RFC-XXXX Boundary Marking

[Mark Crispin <mrc@xxxxxxxxxxxxxxxxxxxxxxxx>]

I agree with David Robinson that the boundary-marking scheme of RFC-XXXX would
be better off in the simpler mode of not including <EOL> as the start of the
sequence.

As I understand it, some implementation techniques involve examining the data
line at a time and this is to help those implementations.  I think this merely
pushes the <EOL> search to a lower level without impacting performance for
better or worse, but I admit to a platform/implementation technique bias here.