http://issues.apache.org/SpamAssassin/show_bug.cgi?id=3710





------- Additional Comments From [EMAIL PROTECTED]  2006-03-08 14:11 -------
I'm not against a winsock2 implementation. If someone can submit a patch, I can
resurrect my remaining  Windows machine long enough to verify and check in a
patch. I'm not enough of an expert to write this myself from scratch, but I
could understand and verify something that someone else got working.

I don't see right now the benefit of winsock2 for this. We don't need to use
socket_t when we can use size_t. The compatibility problem we have is with
signals not working with Windows, and that doesn't change with winsock2. Even
having use of non-blocking calls for the read and writes doesn't help us with
timeouts on connect, as far as I can tell.

Here's an article discussing timeouts with connect: 
http://www.codeguru.com/forum/showthread.php?t=312668
It doesn't seem that winsock2 helps with that. The solution in that article
requires at least XP.

The Windows equivalent of using signals would be to spin a timer thread. That's
more than I was ready to do for spamc. Again, I would consider a patch for that.

In the meantime, the patch I uploaded solves the immediate problem and does not
restrict the WIN32 version of spamc any more than it already was. So I'm still
looking for two votes for checking it in to 3.1.




------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

Reply via email to