This same phenomena occurs on NetWare. Timing values of 1987 and 1977 (should be 2000) occasionally occur. I think it has to do with the resolution of the timers in the JVM.
I was going to bring this up, but wanted to explore OS issues first to see if it is all on our side. The sleep delays being innaccurate change sounds good to me!(or your other change suggestion) Either that or I will just ignore the test every time it fails on NetWare... Jeff Tulley ([EMAIL PROTECTED]) (801)861-5322 Novell, Inc., the leading provider of Net services software. >>> [EMAIL PROTECTED] 10/18/01 11:01:37 AM >>> ----- Original Message ----- From: "Stefan Bodewig" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Thursday, October 18, 2001 05:53 Subject: Re: cvs commit: jakarta-ant/src/testcases/org/apache/tools/ant ProjectTest.java > On Wed, 8 Aug 2001, Nico Seessle <[EMAIL PROTECTED]> wrote: > > > But org.apache.tools.ant.taskdefs.SleepTest fails sometimes for > > me... Don't know why, sorry. hey, that is v.wierd. All that test does is verifies that the duration of the test exceeds that specified delay. Now maybe either the OS doesnt sleep that long, or the timer for measuring sleep delays is inaccurate. We could add in more support for approximate errors, or just take the actual timing out of the test altogether and just retain the argument checking. -steve
