--
Tom Goetze
At 09:04 PM 8/5/2004, you wrote:
>Anyone know why the SocketAppender doesn't have a "setter" for the
>"address" member variable.
Address is the internal representation of the remoteHost variable.
Tom: But it only gets used once (to be checked for null). In the connect
method, the parameter named address (not the local member variable) is
actually
used for work.
>This appears to make it impossible to extend the class (since the connect
>method checks this.address for null and returns if it is true--even
though
>the method takes an InetAddress as a parameter, and the parameter value
is
>the one that is used in the method).
connect is also a default access (package protected) method.
Tom: You're right, this also needs to be changed.
>It seems to me that the "address" member variable could be removed
>entirely, and the check in connect could be removed, then sub-classing
>would be possible and the current logic would remain.
This does not makes sense.
>Any comments? Any log4j developers that want to implement this trivial
>change?
At this stage, there is clearly a misunderstanding. As far as I can tell,
there is no trivial change.
>--
>Tom Goetze
--
Ceki G�lc�
For log4j documentation consider "The complete log4j manual"
ISBN: 2970036908 http://www.qos.ch/shop/products/clm_t.jsp
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
