The test mentioned in the synopsis fails on Windows ones per 4 times approximately.
Posted Date : 2006-07-07 16:12:14.0

N/A

This is a bug in the test itself. The point is that it triggers key events before
dialog.setVisible(true) gets called. It then expects that the events will be queued
in the type-ahead queue. However it may not happen, because the dialog sets type-ahead
marker in the process of dialog.setVisible(true) execution. Till that moment the key events
could already be dispatched to the current focus owner. This is timing dependant.
Posted Date : 2006-07-07 16:12:14.0

But at the same time we can't generate the key events after dialog.setVisible(true)
because this breaks the condition being tested. So, the only way is to try to trigger
the events in the time of processing the setVisible() call. To be precize we have to do
this just at the moment when the marker is set but the dialog is not yet shown.
Posted Date : 2006-07-07 16:15:58.0

The bug was fixed into mustang b93 but the test still fails on some Windows machine
(reported by SQE). Howeverv we can't reproduce the failure on our machines.
Posted Date : 2006-07-28 09:57:54.0