http://bugzilla.spamassassin.org/show_bug.cgi?id=4260
------- Additional Comments From [EMAIL PROTECTED] 2005-04-26 16:02 ------- Theo, I'm concerned that the factor of 3 slowdown you saw in mass check had to do with the listener getting the old replies and then rejecting them based on the ID not matching. What is the overhead of closing a UDP socket and creating a new one? Doing that once per message will make it unlikely that a listener will be available when an old reply shows up. As for the chance that the OS will assign the same port again, 1) when I've looked at port assignments they seem to usually be sequential; 2) even if they are random rather than sequential (or when the port numbers eventually wrap), the IDs are not likely to match when the port numbers happen to. That said, I don't see a collision problem with your scheme or Justin's variation of it, and I agree that it is simpler than changing the socket with each message, as the code change stays inside DnsResolver. It is worth a try, but if your mass check is still three times slower, we may have to go to changing the socket with each message. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
