On Sat, Apr 10, 2010 at 3:48 PM, Kim Gräsman <[email protected]> wrote: > Hello, > > On Sat, Apr 10, 2010 at 07:20, Dean Michael Berris > <[email protected]> wrote: >> >> If the system is POSIX compliant, this should not be an issue because >> SO_REUSEADDR applies to whether the listening application can re-use >> the socket in the TIME_WAIT state (that which has been closed >> already). This is important for efficiency reasons, although can >> definitely be dangerous in systems where the networking stack doesn't >> clean up properly. >> >> If this is a problem only in Windows then we should have a special >> compiler flag that turns it off. >> >> http://www.unixguide.net/network/socketfaq/4.5.shtml > > I'm not even sure it's a problem, really. It's just a bit confusing, > but I'm not sure I see a breaking scenario if listening to the same > port twice _doesn't_ throw, which seems to be scenario here. Thoughts? >
If the problem is really that in Windows, then it might be a bug in Asio. The thing is, in Linux at least, the expectation when you try to listen/bind to an address/port pair that is already bound to by some other process, the second attempt should fail. I am not sure if this is the same in Windows, and if removing the SO_REUSEADDR is the solution for that platform to keep the expectations consistent on both Linux and Windows, then it might be worth addressing. HTH -- Dean Michael Berris deanberris.com ------------------------------------------------------------------------------ Download Intel® Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev _______________________________________________ Cpp-netlib-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/cpp-netlib-devel
