Julian Mehnle wrote:
Why can't your users' mail clients (or whatever produces those misformatted
e-mail addresses)
just conform to the standards?
My previous attempts to get Microsoft to conform to standards haven't been
successful. Do you have any thoughts on how I can be more effective?
When the sending mail client says "RCPT TO: <'fo...@bar.org'>", and
Courier replies with an error message, the client could just skip this
address and continue with the next "RCPT TO:" command. Of course, many
clients are broken and stop the sending process at the first error
message, even if the error message in question is definitely non-fatal.
But why does
*Courier* have to be fixed to accommodate those clients?
I wasn't aware that there was an option to reject the address at this point.
This might be ok, even if the entire mail was rejected. At least the user
would look for errors in their side, rather than blaming the server, and
causing panic. What would be the downside to adding checks for invalid
domain name characters? (this would address all of the potential quoting,
bracketing, etc)
Another alternative to help user education would be the error message
returned.
Currently what happens is that a little while later a non-delivery notice
appears such as....
---------------------------------------------------------------------------
UNDELIVERABLE MAIL
Your message to the following recipients cannot be delivered:
<'fo...@bar.com'>:
<<< No such domain.
----------------------------------------------------------------------------
In some cases, this triggers a panic about how the entire eMail system must
be broken because it couldn't find itself. (It is not as funny when it
happens to you)
As an alternative, the above message could be a bit more helpful to point
out the characters that were found that were suspect.
17 messages in net.sourceforge.lists.courier-usersRE: [courier-users] RE: How to remove...
