I don't agree that the FATAL priority level is useless. > A FATAL problem is typically just > an uncaught > exception -- and if the exception is uncaught there > really seems no > appropriate to actually *log* the problem.
Anything that causes the app to shut down should be logged, so that the reason (e.g., exception stack trace) may coupled with the log data that describes what led up to the problem. > The only appropriate senario that I see is having one > exception catch > block around my entire application which will log a FATAL > error for any > unexpected exception. Depending on what you're doing, there are other reasons to shut down the application. Key data may be corrupt or missing, with no graceful way to recover. .......................... yugen http://gvf.sourceforge.net .......................... -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>