> Apache 1.3.22 introduced a rather troubling bug in the listeners. > > Since we allow the child to start immediately, without waiting on > the prior child to finish, we have introduced a serious flaw. Both > processes are listening on the same socket, already primed to accept > connections. The client with an existing socket's data will be > randomly accepted by either process.
I don't see how that can happen. The thread that does the accepting on Windows (a single thread) does not release the start_mutex (which enables the other process to begin accepting connections) until after it has stopped accepting connections. In other words, unless there is a bug, only one process should ever be accepting connections. Bill
