On Tue, 3 Feb 2004, Mark Crispin wrote:
>On Tue, 3 Feb 2004, ml wrote:
>> I suspect, the problem is that upon seeing the "+OK <blah blah blah>"
>> greetings from the pop3 server, c-client will issue a "CAPA" command. In
>> response to that, the server would say:
>> +OK capability list follows
>> PIPELINING
>> TOP
>> UIDL
>> APOP
>> .
>
>This server is saying that it does not support plaintext password
>USER/PASS authentication, but it does support APOP authentication.
>
>The c-client library does not support APOP, as APOP has been deprecated in
>favor of SASL based authentication. The SASL equivalent to APOP would be
>"AUTH CRAM-MD5".
>
>Are you certain that the server will accept a USER/PASS authentication?
>The CAPA response says that it will not. Your patch will work to force
>c-client to send USER/PASS even if the server says that it does not
>support it. This would be a bad thing if it turned out that the server
>didn't really support USER/PASS.
Mark,
Thank you very much for the response.
Yes, I'm sure it supports USER/PASS because when I continued on and say
"USER someusername" it responded with "+OK password please" and I logged
on fine. So, it looks like it didn't advertise the "USER" capa because it
preferred users to use APOP? Hmmm...
Anyway, if I were to force c-client to send USER/PASS, would it break
other "good" (e.g. SASL capable) servers?
For your information, I have decided on the following shorter patch.
Thanks again for your help.
Regards,
N.
----patch follows---
diff -r1.1 pop3.c
407a408
> LOCAL->cap.user = T; /* guess worst-case old server */
409,412c410
< else {
< LOCAL->cap.user = T; /* guess worst-case old server */
< return NIL; /* no CAPA on this server */
< }
---
> else return NIL; /* no CAPA on this server */
