fix(conn): Handle remote hangup

Not handling this was an issue for keep-alive connections because
requests would get assigned to a closed connection and then immediately
error. Handling the HUP event makes this situation much less likely. It
is still possible however; consider the situation where a HUP arrives
while the event loop is busy processing new requests to add. The
connection is disconnected, but the HUP hasn't been processed, and a
request could be assigned to it. This case is, however, unlikely.

src/http/conn.rs

@@ -553,6 +553,13 @@ impl<K: Key, T: Transport, H: MessageHandler<T>> Conn<K, T, H> {
             }
         }
 
+        if events.is_hup() {
+            trace!("Conn::ready got hangup");
+            let _ = scope.deregister(&self.0.transport);
+            self.on_remove();
+            return ReadyResult::Done(None);
+        }
+
         // if the user had an io interest, but the transport was blocked differently,
         // the event needs to be translated to what the user was actually expecting.
         //
@@ -593,7 +600,7 @@ impl<K: Key, T: Transport, H: MessageHandler<T>> Conn<K, T, H> {
             self.0.on_writable(scope);
         }
 
-        let events = match self.0.register() {
+        let mut events = match self.0.register() {
             Reg::Read => EventSet::readable(),
             Reg::Write => EventSet::writable(),
             Reg::ReadWrite => EventSet::readable() | EventSet::writable(),
@@ -610,6 +617,8 @@ impl<K: Key, T: Transport, H: MessageHandler<T>> Conn<K, T, H> {
             return ReadyResult::Continue(self);
         }
 
+        events = events | EventSet::hup();
+
         trace!("scope.reregister({:?})", events);
         match scope.reregister(&self.0.transport, events, PollOpt::level()) {
             Ok(..) => {