[slf4j-user] Fwd: jul-to-slf4j bridge not working properly with JUL Logger#isLoggable()

Guus der Kinderen guus.der.kinderen at gmail.com
Tue Sep 10 17:55:56 CEST 2013


Hello,

Today, I've ran into a snag using the jul-to-slf4j SLF4JBridgeHandler
(distribution version 1.7.5). Although I worked around the issue, I wonder
if I'm doing something wrong, or if I hit a bug in the implementation.

The problem popped up during a refactoring of a project. In this project, I
switched from using version 1.x of the Jersey framework to version 2.x.
Both versions use JUL, which is why the SLF4JBridge was already in place
(and doing its magic without incident), using the following two lines in a
ContextListener

SLF4JBridgeHandler.removeHandlersForRootLogger();
SLF4JBridgeHandler.install();

After the move was complete, I noticed that a lot less information was
logged. When looking into the source code of the newly added framework, I
noticed the following:

A log statement that was generated using the snippet below got logged just
fine:

LOGGER.info(LocalizationMessages.INIT_MSG(Version.getBuildId()));

However, in the same method, a different log statement was generated
conditionally based on the 'isLoggable' method response:

if (LOGGER.isLoggable(Level.CONFIG)) {
  final StringBuilder sb;
  /* lots of string generation */
  LOGGER.log(Level.CONFIG, sb.toString());
}

The condition in the if statement evaluated to false, no matter what.

I remedied the problem by adding this last line to my jul-to-slf4j bridge
invocation:

SLF4JBridgeHandler.removeHandlersForRootLogger();
SLF4JBridgeHandler.install();
java.util.logging.LogManager.getLogManager().getLogger( "" ).setLevel(
Level.ALL );

Immediately, all of the missing log statements popped up in my log files
again. Changing the desired log level in my log configuration (which
happens to be log4j2) also has the desired effect (as in: lower level JUL
statements do not get logged).

I'm not familiar enough with all of the implementations involved to
evaluate a) why my fix was needed, and b) if this fix is a proper
work-around, rather than something that's likely to fail in upcoming
releases of either product. I would love to get some feedback.

Kind regards,

  Guus
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.qos.ch/pipermail/slf4j-user/attachments/20130910/ae0b9178/attachment.html>


More information about the slf4j-user mailing list