>>> Some MUAs are sloppy and don't tell me who the bad recipient is when
>>> they get bad news after RCPT TO:.  Thus, I don't know which of
>>>
>>>       
>> For exactly that reason, I don't do any rejection at RCPT TO level for 
>> MUAs.
>>     
> Yep, what he said.  Essentially, for MUA submissions, don't reject at ACL
> time and instead let the routers generate a bounce instead.
>   

I appreciate your thinking, but I don't agree with it.  It seems a bit 
extreme just for overcoming this one little problem in Exim (which is 
easily worked around 98% of the time with the method I described).  If 
you accept anything from an MUA, you get some undesirable usability 
problems:


1.  It can take a short time for the user to see the bounce (for 
example, if they poll for new messages every 10 minutes).  That's too 
bad for someone who sends a message and then heads out the door.


2.  If I send a message to 5 recipients and 1 of them is bad, 4 people 
will get a message with a bad recipient in the headers.  When they 
reply, they'll get rejections or bounces.  It's subjective, but I think 
most people would like the chance to correct a bad address before the 
message goes to the others.


Granted, these things can still occur even with recipient checking, but 
why not handled the cases that can be caught?


-- 
## List details at http://lists.exim.org/mailman/listinfo/exim-users 
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

Reply via email to