On Nov 6, 2007, at 2:12 AM, Amnon Lahav wrote:

thanks for your detailed answer curt , that was my mistake i was under the impression that treshold specifies the only level which will be printed out
.
if i understand right from what i've read in order to use filters i must
switch to a log4j xml configurations .

I believe that is correct.

i avoided the ("message",e) type because i tough i'll only be getting
exceptions there (error.log) so i won't need the "explanation" it'll appear in the stacktrace , but as i allso found out the exception gets printed out as a stacktrace only in the info log and not the error log whre is only
printed out as a e.getMessage() type , why is that happening ?
thanks .

The message (first) parameter is declared as an Object and somewhere later it is expected that its toString() method will be called to get the text actually used as the message. The message parameter is commonly a String (and String.toString() returns itself). So the following are all reasonable:

logger.info("Hello, World");
logger.info(new StringBuffer("Hello, World"));
logger.info(new Double(Math.PI));

In each case, somewhere in processing the object's toString() method is called to get the text message. There is no special treatment for an exception. At the same place, Double.toString() would be called, Exception.toString() is called. There is no code in log4j that's like:

if (msg instanceof Throwable) {
    //
// I know that the user really wanted to see the output of printStackTrace
   renderedMessage = ...
} else if (msg instanceof SomeOtherSpecialClass) {
   renderedMessage = ...
} else {
   //
   //    if not any special case that we likely never documented
   //       then call toString() to get the rendered message
   renderedMessage = msg.toString();
}



---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to