Re: More Test 15 [CGF please comment]

["Pierre A. Humblet" <Pierre dot Humblet at ieee dot org>]

At 10:45 PM 4/11/2001 -0400, Christopher Faylor wrote:
>On Wed, Apr 11, 2001 at 08:47:09PM -0400, Pierre A. Humblet wrote:
>Thank you for the analysis.  This is exactly what I was looking for.
>Would you care to submit a patch incorporating your setsockopt idea?

Sure, see attachment. I have tested it on Win98 and WinME. 
No more connections in TIME_WAIT...

>I wonder if this should also be employed for UNIX domain sockets.

What do you mean? The pipe close problem on Win9X or something else?

>Hopefully the connections that you are seeing will go away within
>a few seconds.  I don't see any reason for them to linger for a
>long time.

With the Test 15 server calling select() 100x/sec, more than 2000 
connections accumulated on WinME. Thus the timeout is at least 20 sec...
That's long. Anyway it's fixed now.

>>Why do you use PeekMessage( PM_NOREMOVE) in peek_windows, and not
>>GetQueueStatus()?
>
>I assume that it was used because the person who wrote the original
>select code didn't know about the side effect and I just implemented his
>idea when I rewrote select().
>
>In any event, it looks like GetMessageStatus is a simpler interface.
>Would you like to submit a patch?

I tried but it doesn't work. Apparently GetQueueStatus() doesn't 
consider pending NonQueuedMessages. 

Pierre
--- select.cc.in	Tue Mar 20 14:50:28 2001
+++ select.cc	Fri Apr 13 19:41:14 2001
@@ -1329,6 +1329,11 @@
     {
       select_printf ("connection to si->exitsock %p", si->exitsock);
       SOCKET s = socket (AF_INET, SOCK_STREAM, 0);
+
+	  /* Set LINGER with 0 timeout for hard close */
+	  struct linger tmp = {1, 0}; /* On, 0 delay */
+	  (void) setsockopt (s, SOL_SOCKET, SO_LINGER, (char *)&tmp, sizeof(tmp));
+
       /* Connecting to si->exitsock will cause any executing select to wake
 	 up.  When this happens then the exitsock condition will cause the
 	 thread to terminate. */
@@ -1338,12 +1343,12 @@
 	  select_printf ("connect failed");
 	  /* FIXME: now what? */
 	}
-      shutdown (s, 2);
+      shutdown (s, SD_BOTH);
       closesocket (s);

       /* Wait for thread to go away */
       WaitForSingleObject (si->thread, INFINITE);
-      shutdown (si->exitsock, 2);
+      shutdown (si->exitsock, SD_BOTH);
       closesocket (si->exitsock);
       CloseHandle (si->thread);
       stuff->device_specific[FHDEVN(FH_SOCKET)] = NULL;