ChangeLogTag: Thu Apr 17 17:07:36 2003 Irfan Pyarali <irfan@cs.wustl.edu>

2 files changed, 27 insertions, 0 deletions

diff --git a/ChangeLog b/ChangeLog
index 870e110cc99..a4db4b40205 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,23 @@
+Thu Apr 17 17:07:36 2003  Irfan Pyarali  <irfan@cs.wustl.edu>
+
+	* tests/Process_Mutex_Test.cpp: Reverted my change:
+
+	  Thu Apr 17 16:04:54 2003  Irfan Pyarali  <irfan@cs.wustl.edu>
+
+	  David Levine informed me that he had placed that sleep()
+	  purposely:
+
+	  Thu Jul 22 11:54:58 1999  David L. Levine  <levine@cs.wustl.edu>
+
+	  and said that:
+
+	  "I think the problem has to do with calling ::waitpid () before
+	  a fork'ed child has actually been created.  It's not enough for
+	  fork () to have returned to the parent.  This really shouldn't
+	  be a problem, but it sure looked like it was on that platform."
+
+	  I have still closed bug 91 since this is a exceptional case.
+
 Thu Apr 17 16:04:54 2003  Irfan Pyarali  <irfan@cs.wustl.edu>
 
 	* tests/Process_Mutex_Test.cpp: Removed call to sleep(). It was
diff --git a/tests/Process_Mutex_Test.cpp b/tests/Process_Mutex_Test.cpp
index b6fec466b2b..3392540c052 100644
--- a/tests/Process_Mutex_Test.cpp
+++ b/tests/Process_Mutex_Test.cpp
@@ -178,6 +178,13 @@ ACE_TMAIN (int argc, ACE_TCHAR *argv[])
           ACE_DEBUG ((LM_DEBUG,
                       ACE_TEXT ("Parent spawned child process with pid = %d.\n"),
                       children[i].getpid ()));
+
+          // Give the newly spawned child process a chance to start...
+          // David Levine thinks this sleep() is required because
+          // calling ::waitpid () before a fork'ed child has actually
+          // been created may be a problem on some platforms.  It's
+          // not enough for fork() to have returned to the parent.
+          ACE_OS::sleep (1);
         }
 
       for (i = 0; i < n_processes; i++)