On Tue, Jul 19, 2016 at 10:34:18AM -0400, Doug Torrance wrote:
> I was wondering if you had any updates about this bug.
>
> Is violating RFC3501 a dealbreaker? Should wmbiff use search, or should we
> go ahead and release a version with your original patch to fix the bug?
Naw, I don't think using status breaks anything, as long as we make sure the
mailbox is closed before we call status on it.
Attached is my latest patch.
--- a/wmbiff/Imap4Client.c
+++ b/wmbiff/Imap4Client.c
@@ -294,6 +294,7 @@
/* recover connection state from the cache */
struct connection_state *scs = state_for_pcu(pc);
char buf[BUF_SIZE];
+ char examine_expect[BUF_SIZE];
static int command_id;
/* if it's not in the cache, try to open */
@@ -312,6 +313,22 @@
return -1;
}
+ command_id++;
+ tlscomm_printf(scs, "a%03d EXAMINE %s\r\n", command_id, pc->path);
+ snprintf(examine_expect, BUF_SIZE, "a%03d OK", command_id);
+ if (tlscomm_expect(scs, examine_expect, buf, 127) == 0) {
+ tlscomm_close(unbind(scs));
+ return -1;
+ }
+
+ command_id++;
+ tlscomm_printf(scs, "a%03d CLOSE\r\n", command_id);
+ snprintf(examine_expect, BUF_SIZE, "a%03d OK", command_id);
+ if (tlscomm_expect(scs, examine_expect, buf, 127) == 0) {
+ tlscomm_close(unbind(scs));
+ return -1;
+ }
+
/* if we've got it by now, try the status query */
command_id++;
tlscomm_printf(scs, "a%03d STATUS %s (MESSAGES UNSEEN)\r\n",
