Hi,
I know this is very presumptuous of me to arrive on the list and request a
change, but that's what I am going to do.
I installed courier-imap-1.5.3.20020921 to work with postfix and was
trying to do some testing to verify that having creating the userdb for
authentication, I had in fact followed the instructions correctly. I was
testing using telnet to the POP3 port (since I don't speak IMAP that well
:-). Anyway, I could not under any circumstances authenticate
successfully. Eventually I relented and tried an IMAP login and to my
delight/confusion succeeded. I did a bit of searching through the mailing
list archives and saw people with similar (pop and telnet) problems but no
resolution. That got me digging through strace output from all sorts of
daemons and reading code and eventually I worked out that the problem was,
in part, caused by the telnet client.
During the POP3 conversation, eg
# telnet 0 110
Trying 0.0.0.0...
Connected to 0.
Escape character is '^]'.
+OK Hello there.
USER fred
+OK Password required.
PASS blahblah
....
the telnet client sends its input literally as follows:
USER fred\r\n
and
PASS blahblah\r\n
When pop3login passed this to authdaemon it sends something like:
AUTH 41\npop3\nlogin\nfred\r\nblahblah\r\n\n
so when authdaemon splits it up it gets
user = fred\r
pass = blahblah\r
This of course fails the lookup.
I started hunting around in the code for the "best" place to "fix" this
problem and decided it was in pop3login.c. I would like to propose that
pop3login.c be changed to tokenise the input after "USER" and "PASS" using
delimiters "\r\n" instead of "\n". There is some precedent for this - the
lines are first tokenised using " \t\r".
The changes to imap/pop3login.c are as follows:
--- pop3login.c.orig Wed Oct 9 16:01:29 2002
+++ pop3login.c Wed Oct 9 16:05:00 2002
@@ -190,7 +190,7 @@
continue;
}
- p=strtok(0, "\n");
+ p=strtok(0, "\r\n");
if (p)
{
if (user) free(user);
@@ -257,7 +257,7 @@
}
} else if (strcmp(p, "PASS") == 0)
{
- p=strtok(0, "\n");
+ p=strtok(0, "\r\n");
if (!user || p == 0)
{
Would such changes break anything? I can't (from my uneducated standpoint)
see how they would. I know they certainly help in debugging POP.
Colin
--
Colin Campbell
Unix Support/Postmaster/Hostmaster
CITEC
+61 7 3227 6334
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
courier-users mailing list
[EMAIL PROTECTED]
Unsubscribe: https://lists.sourceforge.net/lists/listinfo/courier-users
