[
https://issues.apache.org/jira/browse/LOG4J2-2484?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16659894#comment-16659894
]
Robert Ross commented on LOG4J2-2484:
-------------------------------------
Personally, I think you're misinterpreting "for writing diagnostic output" in
the POSIX requirements. They're referring to diagnostics output for an error.
Not any and all trace data output. Most data you're outputting in the
StatusLogger are not errror messages at all. If that's the case, and this
should be followed to the letter, than about 70% of Log4J routing configuration
should go away since most logging messages would be considered "diagnostics"
and everything should be forced to STDERR for all use cases. I thought one of
the primary goals of Log4J was to be able to configure output location to user
preferences. Why would StatusLogger obey a different set of rules?
Either way, what exactly are you requesting in terms of a "valid use case".
And who determines it is "valid"?
> Allow setting of output stream on StatusLogger
> ----------------------------------------------
>
> Key: LOG4J2-2484
> URL: https://issues.apache.org/jira/browse/LOG4J2-2484
> Project: Log4j 2
> Issue Type: Improvement
> Components: Core
> Reporter: Robert Ross
> Priority: Minor
> Labels: StatusLogger, debugging
>
> The StatusLogger internally uses a SimpleLogger. SimpleLogger offers the
> ability to control the output stream, but StatusLogger encapsulates and hides
> this capability, always using System.err as the output stream.
> It would be useful to expose this option so that the API offers some control
> over the output stream used by StatusLogger.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
