Subj : OPT MPWD interpretation...
To   : Ozz Nixon
From : Joachim Probst
Date : Sat Mar 07 2020 11:19 am

Hello Ozz!

06 Mar 20 22:42, you wrote to all:

 ON> Okay, while working with another BinkP mailer developer ~ we had
 ON> different opinions of the wording and functionality of MPWD support.

 ON> If I have presented (as the answering system) Zone1 Zone99, Zone103
 ON> And the originator presented Zone1 Zone103 Zone200

 ON> Any we have passwords that are different for 1 and 103 - as we are
 ON> each others uplink.

 ON> MPWD in my understanding would be expenting the passwords presented in
 ON> the order the answering system 1, 99, 103
 ON> So I would be expecting to receive MsgHdr Len M_PWD pwd1 - pwd2
 ON> In his mind, it is the originators order, pwd1 pwd2 -

 ON> I know most people just match their passwords, but, a couple
 ON> situations, the password was out of my control ... and I would like to
 ON> implement M_PWD support ... (this may be a BinkD question in some
 ON> heads, but, it goes wider to other mailers)... so I ask here.

Here I am completly with you, not with the way to do it, but that we have a 
valid point one should takle. The problem I see with the one password is: how 
to deal with the password matching not all but only a few akas correctly. 
Ignoring the session? Ignoring the akas not fitting to the password? The FTS is 
only working on with a correctly password protected session or not. What's with 
'partly' password protected sessions?

So I really like the idea of having an authentication per aka.

But when extending the protocol on this point I would recomend doing it in a 
better downward compatible way.

What about introducing a new keyword, either as M_OPT MPWD <aka> 
<password/CRAM> or as M_MPWD <aka> <password>? The old M_PWD would still be 
used as with no password.

This would make a matching of password to aka fairly easy and a system could 
easily support old and new style in paralell, like if there is no new password 
line, I take a look in the old style protocol and if I see the new line I just 
ignore the old style password.

I would think this way would make it much clearer for interpretation.

 ON> Thank you,
 ON> Ozz

Joachim


--- GoldED+/LNX 1.1.5--b20170303
 * Origin:  ----> FidoPI  (2:240/6309)

.