XREQ: Correct behaviour on hitting ZMQ_HWM

This reverts part of commit 84e0c7991a to get
correct ZMQ_HWM semantics with XREQ sockets:

When sending a message to an XREQ socket, the underlying pipe is selected in
a round-robin fashion. If an underlying pipe is full it is skipped. If there
are no underlying pipes, or all underlying pipes are full then zmq_send()
shall block or return EAGAIN, depending on whether or not the call is blocking.
Messages are never dropped.

src/xreq.cpp

@@ -23,8 +23,7 @@
 #include "err.hpp"
 
 zmq::xreq_t::xreq_t (class app_thread_t *parent_) :
-    socket_base_t (parent_),
-    dropping (false)
+    socket_base_t (parent_)
 {
     options.requires_in = true;
     options.requires_out = true;
@@ -78,25 +77,7 @@ int zmq::xreq_t::xsetsockopt (int option_, const void *optval_,
 
 int zmq::xreq_t::xsend (zmq_msg_t *msg_, int flags_)
 {
-    while (true) {
-
-        //  If we are ignoring the current message, just drop it and return.
-        if (dropping) {
-            if (!(msg_->flags & ZMQ_MSG_MORE))
-                dropping = false;
-            int rc = zmq_msg_close (msg_);
-            zmq_assert (rc == 0);
-            rc = zmq_msg_init (msg_);
-            zmq_assert (rc == 0);
-            return 0;
-        }
-
-        int rc = lb.send (msg_, flags_);
-        if (rc != 0 && errno == EAGAIN)
-            dropping = true;
-        else
-            return rc;
-    }
+    return lb.send (msg_, flags_);
 }
 
 int zmq::xreq_t::xrecv (zmq_msg_t *msg_, int flags_)
@@ -111,8 +92,6 @@ bool zmq::xreq_t::xhas_in ()
 
 bool zmq::xreq_t::xhas_out ()
 {
-    //  Socket is always ready for writing. When the queue is full, message
-    //  will be silently dropped.
-    return true;
+    return lb.has_out ();
 }
 

src/xreq.hpp

@@ -55,9 +55,6 @@ namespace zmq
         fq_t fq;
         lb_t lb;
 
-        //  If true, curently sent message is being dropped.
-        bool dropping;
-
         xreq_t (const xreq_t&);
         void operator = (const xreq_t&);
     };