[Top] [All Lists]

Re: Using RSET with the EHLO/HELO fallback logic

2008-02-25 20:47:39

<Valdis(_dot_)Kletnieks(_at_)vt(_dot_)edu> wrote:

Am I the only one who thinks that if the remote end
doesn't understand EHLO, it's highly unlikely that
any revision of 2821 is going to improve matters?

LOL, you certainly have a point.  I always understood
the RSET blurb as a recipe how to trick "old" servers
into doing their job without new connection.

Some "old" servers might accept HELO after rejecting
EHLO.  Another set might be confused after rejecting
EHLO, and RSET fixes that before trying HELO.  And a
third set could require a new connection without EHLO.

Where "old" could mean really 821, or only pretending
to be old for reasons of their own, e.g., if they know
a client where accepting EHLO results in some trouble.
Anything that can go wrong will go wrong, but I have
no example where not accepting EHLO is a good idea.