[concurrency-interest] How has java.util.concurrent.locks been tested?

Doug Lea dl@cs.oswego.edu
Thu, 30 Dec 2004 10:04:03 -0500


> Certainly not ReentrantLock's, but why not using others based on
> ordinardy synchronized/wait/nodify stuff?

Some do. Some others could, but just happen not to.

> I am just trying
> to find out what would be the best way to test concurrent stuff apart
> from the stress tests you mentioned.

Most concurrency errors tend to either be simple enough to be caught
by simple unit tests, or so rare as to require stress testing.  The
enormous effort it takes to pre-wire other fixed interleavings doesn't
usually pay off.

> Any chance to see the code of your other tests?

I'll send some to you separately. Most take quirky undocumented
program arguments and don't explain what they are testing, so need
some minor cleanup before posting more widely.

-Doug