From 67e9d89ebd27300d68ce7a11dc10d5315e3da1f7 Mon Sep 17 00:00:00 2001
From: Jeff Hill <johill@lanl.gov>
Date: Tue, 30 Sep 2008 16:31:09 +0000
Subject: [PATCH] extended comment

---
 src/libCom/osi/os/WIN32/osdThread.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/src/libCom/osi/os/WIN32/osdThread.c b/src/libCom/osi/os/WIN32/osdThread.c
index be6d35c5d..1d9cdd4b1 100644
--- a/src/libCom/osi/os/WIN32/osdThread.c
+++ b/src/libCom/osi/os/WIN32/osdThread.c
@@ -793,13 +793,17 @@ epicsShareFunc void epicsShareAPI epicsThreadSleep ( double seconds )
  * epicsThreadSleepQuantum ()
  */
 double epicsShareAPI epicsThreadSleepQuantum ()
-{
+{ 
     /*
      * Its worth noting here that the sleep quantum on windows can
      * mysteriously get better. I eventually tracked this down to 
      * codes that call timeBeginPeriod(1). Calling timeBeginPeriod()
      * specifying a better timer resolution also increases the interrupt
      * load. This appears to be related to java applet activity.
+     * The function timeGetDevCaps can tell us the range of periods
+     * that can be specified to timeBeginPeriod, but alas there
+     * appears to be no way to find out what the value of the global 
+     * minimum of all timeBeginPeriod calls for all processes is.
      */
     static const double secPerTick = 100e-9;
     DWORD adjustment;