I ran it with default logging and it failed, as expected. I've run it three times with FINEST logging, and it passed every time. Presumably, it really is bringing out some concurrency issue, and the extra logging changes the timings enough to prevent the bug from happening.

Interesting.

Incidentally, it does not take long when it works, less than a minute. When it fails, it spends most of its time doing 120 iterations of "Wait about 15 sec." before reporting "Some of RandomStressTask tasks not yet finished". When it passed, it only did one wait. Before really getting into debug, I'm going to reduce the maximum number of waits to save time when it fails.

Patricia


On 9/25/2010 7:17 AM, Jonathan Costers wrote:
I actually did come across a test that may be interesting to run for these
changes ...

com/sun/jini/test/impl/mahalo/RandomStressTest.td

I have not been able to get this test to pass on any of my machines or VMs
...
It seems to make heavy use of TaskManager and RetryTask.
It can take a while to finish as well, so beware.

You think this may be a good one?

you can run it if you change to the<river>/qa directory and execute:
ant -Drun.tests=com/sun/jini/test/impl/mahalo/RandomStressTest.td run-tests

or similarly in an IDE, set property
run.tests=com/sun/jini/test/impl/mahalo/RandomStressTest.td
and run the run-tests target

make sure everything is built first, of course.

2010/9/21 Patricia Shanahan<[email protected]>

I'm testing my new TaskManager the , but I have some anomalies. It would
help me to get some more testing of
https://svn.apache.org/repos/asf/incubator/river/jtsk/skunk/patsTaskManagerdone 
in other WindowsXP environments.

Both the head revision and revision  998737 need to be tested. Revision
998737 is the one I plan to merge into the trunk. It changes the interface
between TaskManager and its callers, with minimal changes to TaskManager.

It is important that it be tested widely, because it affects a lot of
critical classes, and would be difficult to back out.

The head revision drops in a revised TaskManager. It should be easy to back
out if necessary.

Patricia



Reply via email to