On 25/06/2008, at 7:30 PM, Oliver Frietsch wrote:
Nothing changed. Total runtime is above one hour.
************
Paul,
thanks for this hint. I believed that the trunk actually contained the
log4j 2.0 code (where is that?).
I got log4j to compile. However, I had to grab the
NTEventLogAppender.dll from the prebuilt log4j-1.2.15 to my Windows
system directory to pass all unit tests. Can this "hack" be further
improved?
Ooh, that's something I haven't seen. Because I'm using a Mac I don't
think it tries that. More on unit tests shortly.
Now, while trying to build zeroconf, I'm already waiting several
minutes
to complete the "Running
org.apache.log4j.net.ZeroConfSocketHubAppenderTest" step - it seems to
do plain nothing, but the build process is waiting for it. I'll give
it
another hour, but I'm not very hopeful... CPU usage is low, and
there is
no network traffic either...
I have seen this happen on some computers, and appears to relate to
the local network. The test is sending out ZeroConf packets on the
network and testing that the jmDNS properly gets notifications. I
should probably add a default wait time to that test and give up after
a reasonable amount of time.
You can skip the running of unit tests for both log4j, and log4j-
zeroconf by using:
mvn -Dmaven.test.skip=true package
Obviously it would be nice to have all the unit tests running locally,
but I can say that for log4j-zeronconf they pass locally. When things
are working ok locally that test takes about a second to complete.
BTW, log4j 2.0 is only in experimental design phase, it has it's own
branch at the moment. Once the design gains traction it'll more than
likely move to trunk, with trunk->1.2 branch (again.. :) ).
cheers,
Paul
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
