new ArrayList<T>(1) sets an initial capacity, but the size is still zero. Clearly something was thrown from run.run() which EventQueue.invokeAndWait did not rethrow; I can only guess that some category of error (OOME?) was thrown in EQ and logged by other means. The complete log file might provide a clue, but that was not supplied.

At any rate, an AtomicReference<T> would be both clearer and safer here; get() would just be null in case a strange error was thrown, which could be returned as is from doEventAccess, though it would be better to keep track of the fact that run.run() did _not_ return null and throw some kind of exception:

final AtomicBoolean wasSet = new AtomicBoolean();
final AR<T> res = new AR<T>;
...
try {
  res.add(run.run());
  wasSet.set(true);
} catch (...) {...}
...
if (wasSet.get()) {
  return res.get();
} else if (exc.get() != null) {
  throw exc.get(); // wrapping as needed
} else {
  throw new IllegalStateException("???");
}

(Could use an AtomicReference<Throwable> as well to hold the exception; strange to be using two different idioms for the same purpose here!)

I guess I will take care of it.

core-main #ff3de058c18b

Integrated into 'main-golden', will be available in build *201204120400* on http://bits.netbeans.org/dev/nightly/ (upload may still be in progress)
Changeset: http://hg.netbeans.org/main-golden/rev/ff3de058c18b
User: Jesse Glick <jglick@netbeans.org>
Log: #210991: IOOBE in Mutex.doEventAccess

*** Bug 211067 has been marked as a duplicate of this bug. ***

you are right. in 201204120400 build I see such exception in log.

SEVERE [org.openide.util.RequestProcessor]: Error in RequestProcessor org.netbeans.modules.project.ui.OpenProjectList$4
java.lang.IllegalStateException: #210991: got neither a result nor an exception
	at org.openide.util.Mutex.doEventAccess(Mutex.java:1400)
	at org.openide.util.Mutex.readAccess(Mutex.java:272)
	at org.netbeans.modules.project.ui.OpenProjectList.open(OpenProjectList.java:748)
	at org.netbeans.modules.project.ui.OpenProjectList$4.run(OpenProjectList.java:584)
	at org.openide.util.RequestProcessor$Task.run(RequestProcessor.java:1452)
[catch] at org.openide.util.RequestProcessor$Processor.run(RequestProcessor.java:2032)

Not sure what this means. EventQueue.invokeAndWait was called but its Runnable does not seem to have been run. Duplicate was also on a Mac, so perhaps it is a bug in the Mac JDK. If you can consistently reproduce I can add more logging for diagnosis.

(In reply to comment #7)
> Not sure what this means. EventQueue.invokeAndWait was called but its Runnable
> does not seem to have been run. Duplicate was also on a Mac, so perhaps it is a
> bug in the Mac JDK. If you can consistently reproduce I can add more logging
> for diagnosis.
Yes I can (5-10 tests per day catches this exception)

P.S. I use jdk 1.6.0_31 too (MacOSX update appeared few days ago)

Integrated into 'main-golden', will be available in build *201204150400* on http://bits.netbeans.org/dev/nightly/ (upload may still be in progress)
Changeset: http://hg.netbeans.org/main-golden/rev/ea5cb379cc9f
User: Jesse Glick <jglick@netbeans.org>
Log: Diagnostics for #210991.

SEVERE [org.openide.util.RequestProcessor]: Error in RequestProcessor org.netbeans.modules.project.ui.OpenProjectList$4
java.lang.IllegalStateException: #210991: got neither a result nor an exception; started=false finished=false invoked=true
	at org.openide.util.Mutex.doEventAccess(Mutex.java:1406)
	at org.openide.util.Mutex.readAccess(Mutex.java:273)
	at org.netbeans.modules.project.ui.OpenProjectList.open(OpenProjectList.java:748)
	at org.netbeans.modules.project.ui.OpenProjectList$4.run(OpenProjectList.java:584)
	at org.openide.util.RequestProcessor$Task.run(RequestProcessor.java:1452)
[catch] at org.openide.util.RequestProcessor$Processor.run(RequestProcessor.java:2032)

*** Bug 211569 has been marked as a duplicate of this bug. ***

*** Bug 212178 has been marked as a duplicate of this bug. ***

*** Bug 212734 has been marked as a duplicate of this bug. ***

This bug already has 5 duplicates 
see http://statistics.netbeans.org/exceptions/detail.do?id=186935

It seems like this may be already fixed in 6u32 builds on Mac. Tomas will check with Apple engineers.

Please, try 6u32 and reopen if you encounter this problem again.

Created attachment 120080 [details]
stacktrace

*** Bug 213508 has been marked as a duplicate of this bug. ***

Nice days!
After Russia Day I updated my MacOSX machines (to Java 1.6.0_33), and I have stable test results!