http://bugzilla.spamassassin.org/show_bug.cgi?id=3917
------- Additional Comments From [EMAIL PROTECTED] 2004-11-02 04:16 -------
sleep(5) between the call to shutdown and the close doesn't help at all.
I added a -l option to the call to spamc -x to see the error. Tthe results are
interesting. I'm testing with two different email messages. Each test consists
of calling spamc with the message that I am testing with, printing out the
return ocde, sleeping 5 seconds, then repeating. The error output with -l is
about a mismatch in the number of bytes expected and the number received from
spamd. Each message is consistent in how many bytes are received when it
truncates. They are each different, but totally consistent for any one message.
It is not like a single packet loss. One shows 78130 bytes claimed, 33792 bytes
seen, the other 109053 bytes claimed, 41984 bytes seen.
But here's kicker. I tested using the first 39166 bytes of the message of the
first test. It also showed 33792 bytes seen, in other words shutting down the
connection at the exact same place. That indicates some kind of dependency on
the content of the message.
I tried adding immediately after $client = $server->accept()
setsockopt($client, SOL_SOCKET, SO_LINGER, pack("II", 1, 0)) or
warn "Can't set SO_LINGER on for socket: $!";
and it did not help. Neither did
setsockopt($client, SOL_SOCKET, SO_LINGER, pack("II", 1, 5)) or
warn "Can't set SO_LINGER on for socket: $!";
I still don't have a test case that is independent of spamd, but I have a start
at the server side of it.
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
