Following up to myself...
On 16 Feb 2005 at 17:56, Fred Viles wrote about
"Re: [exim] auth_hack_outlook4 for E":
| On 17 Feb 2005 at 0:35, Piotr Roszatycki wrote about
| "Re: [exim] auth_hack_outlook4 for E":
|...
| | I'm sorry. It shouldn't so offensive as it was. I've just spent some time
| | on searching Google and mailing list archives to find a solution for
| | broken OE4.xx SMTP AUTH support. I've found an explanation of Philip Hazel
| | and he was clear that he won't include such support in any case. I'll try
| | to find the URL.
|
| It's the "for some anti-Microsoft reasons" that needs to be
| substantiated.
I should add that while I agree with Philip's rationale for not
accepting the patch, it is a moot point. The fact that it isn't
needed is an even better reason.
| And, FWIW, exim 4.x *does* support such broken MUAs. No code patch
| neeeded. Just add this to the end of your list of authenticators:
|
| bogus:
| driver = plaintext
| public_name = "PRIVATE\n250-AUTH=LOGIN PLAIN"
| server_prompts = User Name : Password
| server_condition = no
Not quite right. With one fix and a couple cosmetic improvements,
that should be:
bogus:
driver = plaintext
public_name = "\r\n250-AUTH=PLAIN LOGIN"
server_prompts = User Name : Password
server_condition = no
The '\r' is needed for a correct <CR><LF> line ending, but there's no
need for a fake mechanism name. This example also assumes you've
implemented both PLAIN and LOGIN authenticators, at least. CRAM-MD5
doesn't need to be listed on the broken AUTH line, even if you
implement it, since the broken MUAs don't support it anyway (right?).
- Fred
