-=> Quoting mark lewis to Leonard Erickson <=-

 LE> "Handled" in this context  essentially means "we
know it's wrong,
 LE> but write your software to accept it anyway". In other words, it's
 LE> considered necessary to accept messages with this error for the
 LE> sake of proper network operation.
 
 LE> Also note that if you get a message with ^AAREA:, it will have to
 LE> leave your system with that fixed. :-)

 ml> i can't believe that you are advocating fixing someone else's bugs.
 ml> :-( 

That's the way the posted spec excerpt reads.

 LE> I think that's reasonable behavior. We can identify the error
 LE> *and* there's no ambiguity about what was *meant* so it is
 LE> better to correct the message than to throw it away. Sending
 LE> an error report back to the originator is optional. :-)

 ml> my understanding is that we are supposed to pass messages on,
 ml> unmodified other than what's needed for proper network operation. this
 ml> has been generally interpreted as adding info to the seenby and path
 ml> lines and really nothing more than that. 

Fixing ^AAREA seems to fall under the same "needed for technical
reasons" clause.

 ml> to send a message that arrives with ^aAREA: on to another site with
 ml> the ^a removed would be going against this, wouldn't it? 

Sending it *with* the ^A makes it likely that it'll fall on the floor
and be lost. So we either bounce back the ^AAREA containing messages,
or we correct them, *and* bounce back a copy of the original.

 ml> i think that an error report back to the originator would be
 ml> "required" so that they could fix it or at least notify the author
 ml> that he interpreted the spec incorrectly and that it should be fixed. 

If we are going to that much trouble, we can "fix the message" at the
same time. that way we aren't holding users responsible for the stupid
things their sysop does.

... [KABOOM]  Intel Outside!
--- Blue Wave/DOS v2.30