Register the broker wait before the first poll (closes #157)

GET /api/v1/messages read the client output queue and only then
registered with the broker. Notify wakes only the channels registered
at that moment, so a message queued between the empty read and the
registration waited out the whole long-poll timeout. The handler now
registers first and removes its channel on every exit path; the IRC
relay loop registers before each drain in the same way.

The handlers read the queue through a function field that defaults to
Database.PollMessages, so tests can act just after a read: one checks
that a message queued there comes back at once, others that no wait
channel stays registered however a request ends.

Model: opus-5-5

internal/ircserver/relay.go

@@ -27,17 +27,22 @@ func (c *Conn) relayMessages(ctx context.Context) {
 		default:
 		}
 
+		// Register with the broker before draining, so that a
+		// message queued after the drain has read the queue
+		// still signals waitCh.
+		waitCh := c.brk.Wait(c.sessionID)
+
 		// Drain any available messages.
 		delivered := c.drainQueue(ctx)
 
 		if delivered {
+			c.brk.Remove(c.sessionID, waitCh)
+
 			// Tight loop while there are messages.
 			continue
 		}
 
 		// Wait for notification or timeout.
-		waitCh := c.brk.Wait(c.sessionID)
-
 		select {
 		case <-waitCh:
 			// New message notification — loop back.