But what is wanted here is not logging, but a way of notifying the application that something is happening inside the server.
If that notification then translates to log output being produced then that is not the concern of the AltRMI code.
ah. the penny dropped; ignore my prior post :D
Lemme see if I understand:
you have low-level utility code (altrmi) which for some reason crosses what would normally be an exception boundary, yet you do not want this code to handle any exception itself, but rather have the code that uses the utility code to do that?
this makes sense for altrmi, indeed, as it does for BCEL-related code, or stuff that deals with connection/thread/io management, IOW the real actual low-level stuff.
For lots of other stuff, there is no exception boundary to cross and you just want to rethrow the exception. Like A JDBC driver: it doesn't log, it just throws excpetions. Right?
For yet more stuff (the active part of an api), it does make sense to log. In the avalon world, there is an exception boundary between component and container, so you *do* want your components to do logging. Like JAMES doing logging instead of talking to a monitor.
Finally, stuff which is closely coupled to the active part of an api can also do logging as long as the active api is doing logging. If you have a top-level component using utility code in other classes or beans which are not to be reused outside that component, it makes sense to do logging.
I think the issue here is: Is monitoring a server always equivalent to logging?
OTOH: Is making log calls always equivalent to actual logging, or is it sometimes translated into monitoring calls?
IOW: can you (do you want to) connect an altrmi server to an A-F logging implementation?
- LSD
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
