I've updated the API compatibility report for log4j/trunk against
log4j 1.2.8 at http://people.apache.org/~carnold/compatibility.html.
The recent modifications added the compatibility errors regarding
org.apache.log4j.jmx.HierarchyDynamicMBean.
I've also updated the coverage report at http://people.apache.org/
~carnold/coverage. None of the recently added tests
(o.a.l.xml.LoggerFactoryTest, o.a.l.xml.ResetTest,
o.a.l.net.TelnetAppenderTest, o.a.l.config.PropertySetterTest,
o.a.l.jmx,HierarchyMBeanTest, and the org.apache.log4j.concurrent
tests) were added to the runAll target and are not run as part of the
regular unit tests.
The new org.apache.log4j.jmx test did not compile since compiling the
JMX code had not been a prerequisite of running the unit tests. I
committed rev 502375 to fix that. After fixing that,
HierarchyMBeanTest failed to compile on JDK 1.3 since it used several
classes that were introduced in JDK 1.5, so I've suppressed it from
compiling with rev 502381. The test should be reworked to eliminate
JDK 1.5 dependencies and to be run as part of "runAll".
org.apache.log4j.concurrent.FileAppender used Boolean.valueOf(append)
in a diagnostic message which causes it to fail to compile or run on
JDK 1.3. I'll change just that occurrence to "append ?
Boolean.TRUE : Boolean.FALSE" which will compile on JDK 1.3 and
hopefully avoids the findbugs complaint about using new Boolean
(append). Fixed in rev 502386 (against misattributed to bug 19004,
should have been 24159).
o.a.l.helpers.DateLayoutTest.testSetOptionDateFormat started failing
after the recent commits due to svn commit 500441 on bug 40937.
Those changes were reverted in rev 502338 to get Gump running again.
I'd like to investigate this one myself and more tests are needed to
establish that the existing behavior is not changed except when
clearly beneficial.
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
