On 5/4/08, Marc Lehmann <[EMAIL PROTECTED]> wrote: > On Sat, May 03, 2008 at 05:58:53PM -0700, Cameron Dale <[EMAIL PROTECTED]> > wrote: > > I know how to fix this, but it doesn't seem necessary as I can't get > > it to generate an error. I set the max and min port options to the > > same port, start the downloader, stop it, then restart it again, all > > without errors. Perhaps the OS or python have changed, because > > BitTornado still doesn't set SO_REUSEADDR. > > Then your test is flawed - the os hasn't changed, when a conenction is > active on that port no os will let you re-bind. Of coruse it works when no > connection is active.
Then I don't understand your problem. The idea is I stop bittornado, then restart it before the timeout on the address has expired, and the address comes up as still in use, right? If that's not the bug, then describe what it really is. > > If you can reproduce an error with the current BitTornado, let me know > > how and I'll look into fixing it, otherwise I'll eventually close > > this. > > I can reproduce it any time, and its a bug in any case. As I said, please tell me how, because I can not. > I can not understand why you can even consider closing the bug report while > you understand that it really is a bug. If you don't have the time/are too > lazy (very valid reasons) to fix it, leave it open. But closing it despite > knowing it is valid and still buggy makes no sense to me at all. I did not close it. I can not reproduce it, so I don't think it's a bug. I know how to use SO_REUSEADDR, but I'm not going to patch the code for something that isn't a bug. Show me how to reproduce it, and I'll fix it. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
