"Eric S. Raymond" <esr(_at_)thyrsus(_dot_)com> writes:
Todd Sabin <tas(_at_)webspan(_dot_)net>:
One one side, I can use my fetchmail 5.6.2 to go to one of our UW imap
server (running IMAP4rev1 v12.264), but when I try to go to the 2nd
server (running IMAP4rev1 2000.283), it fails. Details :
[...]
You need the patch below. I sent this a while back when someone else
had the same problem, but it doesn't seem to have been applied to
the source, yet. The patch is against an older version of fetchmail,
probably 5.5.5...
Assuming this works, something similar needs to be done to the rest
of the SASL mechanisms as well...
Argh. Todd, your fix is not going to apply as-is. The problem is that
I refactored the way CRAM is handled so that I could support CRAM-MD5
authentication in ODMR as well as IMAP. Take a look at the new cram.c
and odmr.c modules to see what I mean...
Would you please download the 5.6.3 sources and do what's necessary
to make CRAM work? I'd do this myself, but the IMAP server I have
available to test with doesn't support CRAM.
You may have to pull something odd, like passing a pointer to the
appropriate ok_* function in through do_cram_md5()'s arguments.
Well, here's what I came up with. This patch is against 5.6.4 and
make CRAM-MD5 work again, for me at least. It was broken in the
general case, not just the uw-imap2000 case. The problem was the
usage of gen_transact, which caused part of the CRAM dialog to be
tagged, which is wrong. Note that I haven't tested the odmr stuff.
BTW, I'm not so happy with the ok function. The char* argument seems
to be completely unnecessary most of the places *_ok is used. Note
the otherwise unnecessary dummy variable in do_cram_md5 now. Also,
there's no size argument for the char*, which makes it an accident
waiting to happen. Would you accept a patch to eliminate that arg?
I.e., make things int foo_ok (int sock);
Also, feel free to ignore the driver.c change. It's unrelated, but
something that I've been adding to my builds for a long time. I fetch
mail from a number of mailboxes and find that having a header that
tells which mailbox fetchmail pulled it from makes sorting it properly
that much simpler. Figured someone else might find it useful.
Todd
p.s. If you want to use a free mail server that support CRAM-MD5, try
www.mailandnews.com. Their IMAP server has other issues, but it does
support CRAM-MD5.
diff -u fetchmail-5.6.4 fetchmail-5.6.4-tas/
diff -u fetchmail-5.6.4/cram.c fetchmail-5.6.4-tas/cram.c
--- fetchmail-5.6.4/cram.c Sun Feb 11 00:41:10 2001
+++ fetchmail-5.6.4-tas/cram.c Sun Feb 11 16:47:07 2001
@@ -60,7 +60,8 @@
MD5Final (response, &ctx);
}
-int do_cram_md5 (int sock, struct query *ctl)
+int do_cram_md5 (int sock, struct query *ctl,
+ int (*ok_method)(int, char*))
/* authenticate as per RFC2195 */
{
int result;
@@ -69,6 +70,7 @@
unsigned char msg_id[768];
unsigned char response[16];
unsigned char reply[1024];
+ unsigned char dummy[MSGBUFSIZE+1];
gen_send (sock, "AUTHENTICATE CRAM-MD5");
@@ -131,7 +133,10 @@
/* ship the authentication back, accept the server's responses */
/* PMDF5.2 IMAP has a bug that requires this to be a single write */
- result = gen_transact(sock, buf1, sizeof(buf1));
+ strcat (buf1, "\r\n");
+ SockWrite (sock, buf1, strlen (buf1));
+
+ result = ok_method (sock, dummy);
if (result)
return(result);
else
diff -u fetchmail-5.6.4/driver.c fetchmail-5.6.4-tas/driver.c
--- fetchmail-5.6.4/driver.c Wed Feb 7 01:59:22 2001
+++ fetchmail-5.6.4-tas/driver.c Sun Feb 11 16:15:59 2001
@@ -1184,6 +1184,13 @@
stuffline(ctl, errmsg);
}
+ {
+ char buf[200];
+ snprintf (buf, sizeof (buf), "X-fetched-from: %s(_at_)%s\r\n",
+ ctl->remotename, ctl->server.truename);
+ stuffline (ctl, buf);
+ }
+
/* issue the delimiter line */
cp = buf;
*cp++ = '\r';
diff -u fetchmail-5.6.4/fetchmail.h fetchmail-5.6.4-tas/fetchmail.h
--- fetchmail-5.6.4/fetchmail.h Sat Feb 10 16:14:57 2001
+++ fetchmail-5.6.4-tas/fetchmail.h Sun Feb 11 16:41:42 2001
@@ -533,7 +533,7 @@
char *host_fqdn(void);
char *rfc822timestamp(void);
flag isafile(int);
-int do_cram_md5 (int sock, struct query *ctl);
+int do_cram_md5 (int sock, struct query *ctl, int (*ok_method)(int, char*));
void yyerror(const char *);
int yylex(void);
diff -u fetchmail-5.6.4/imap.c fetchmail-5.6.4-tas/imap.c
--- fetchmail-5.6.4/imap.c Wed Feb 7 00:49:28 2001
+++ fetchmail-5.6.4-tas/imap.c Sun Feb 11 16:42:01 2001
@@ -852,7 +852,7 @@
report (stdout, _("CRAM-MD5 authentication is supported\n"));
if (ctl->server.protocol != P_IMAP_LOGIN)
{
- if ((ok = do_cram_md5 (sock, ctl)))
+ if ((ok = do_cram_md5 (sock, ctl, imap_ok)))
{
if (outlevel >= O_MONITOR)
report (stdout, "IMAP> *\n");
diff -u fetchmail-5.6.4/odmr.c fetchmail-5.6.4-tas/odmr.c
--- fetchmail-5.6.4/odmr.c Thu Feb 8 15:25:15 2001
+++ fetchmail-5.6.4-tas/odmr.c Sun Feb 11 16:42:10 2001
@@ -62,7 +62,7 @@
*bytes = *countp = *newp = -1;
/* authenticate via CRAM-MD5 */
- ok = do_cram_md5(sock, ctl);
+ ok = do_cram_md5(sock, ctl, odmr_ok);
if (ok)
return(ok);