fix really pathological case where io_timeouted would never

start over from the beginning because always new accept()ed
    connections came in and got newer, higher descriptors since the last
    io_timeouted loop.  (Dirk Engling)

CHANGES

@@ -1,4 +1,8 @@
 0.26:
+  fix really pathological case where io_timeouted would never
+    start over from the beginning because always new accept()ed
+    connections came in and got newer, higher descriptors since the last
+    io_timeouted loop.  (Dirk Engling)
 
 0.25:
   array_allocate no longer truncates the array

io/io_timeouted.c

@@ -14,5 +14,8 @@ int64 io_timeouted() {
     if (e->inuse && e->timeout.sec.x && taia_less(&e->timeout,&now))
       return ptr;
   }
+  ptr=0;	/* this is for really pathological cases, where more
+		   connections come in all the time and so the timeout
+		   handling does not trigger initially */
   return -1;
 }