[GitHub] commons-pool pull request #17: POOL-359: prevent NPE closing multiple Generi...

[mswintermeyer]

GitHub user mswintermeyer opened a pull request:

https://github.com/apache/commons-pool/pull/17

POOL-359: prevent NPE closing multiple GenericObjectPool's



You can merge this pull request into a Git repository by running:

$ git pull https://github.com/mswintermeyer/commons-pool fixNPEClosePools

Alternatively you can review and apply these changes as the patch at:

https://github.com/apache/commons-pool/pull/17.patch

To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:

This closes #17

----
commit 0fdde72475b378cbd0fb707d52385196d928a1da
Author: Michael Wintermeyer <***@...>
Date: 2018-11-16T20:23:51Z

POOL-359: prevent NPE closing multiple GenericObjectPool's

----


---

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-***@commons.apache.org
For additional commands, e-mail: dev-***@commons.apache.org

[mswintermeyer]

Github user mswintermeyer commented on the issue:

https://github.com/apache/commons-pool/pull/17

If `cancel` should not be called more than once, then I believe that would require a greater re-architecture. `GenericObjectPool`'s would need to keep track of whether another object pool called `cancel`: to rely on something other than non-static state (`closed` variable on `BaseGenericObjectPool`) to call a static method. To me, it seems cleanest for static methods to be designed defensively. So I'm not aware of a reason to not allow this, but I acknowledge I'm new to the codebase.

Side note that the risk of an NPE in `cancel` seems somewhat new. [This commit](https://github.com/apache/commons-pool/commit/867614829ba6b041ca6becbad0f43cb180003404#diff-38e254894b87bdf9a1758778c7ffd50fL96) refactored `EvictionTimer` to depend on the executor's queue rather than an integer `usageCount`.


---

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-***@commons.apache.org
For additional commands, e-mail: dev-***@commons.apache.org

[grimreaper]

Github user grimreaper commented on the issue:

https://github.com/apache/commons-pool/pull/17

Works for me. Blocked on INFRA-17267
travis fails for unrelated reasons, but tests pass locally:
```
[INFO] ------------------------------------------------------------------------
[INFO] BUILD SUCCESS
[INFO] ------------------------------------------------------------------------
[INFO] Total time: 05:45 min
[INFO] Finished at: 2018-11-18T20:54:44-08:00
[INFO] ------------------------------------------------------------------------
```


---

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-***@commons.apache.org
For additional commands, e-mail: dev-***@commons.apache.org

[graben]

Github user graben commented on the issue:

https://github.com/apache/commons-pool/pull/17

#18 fixes actual failing travis build!


---

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-***@commons.apache.org
For additional commands, e-mail: dev-***@commons.apache.org

[mswintermeyer]

Github user mswintermeyer commented on the issue:

https://github.com/apache/commons-pool/pull/17

@garydgregory the test was failing on my machine without the main code change, so thanks for flagging that it was not consistent! I've updated the test, and I believe this time it's a consistent failure :)


---

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-***@commons.apache.org
For additional commands, e-mail: dev-***@commons.apache.org

[mswintermeyer]

Github user mswintermeyer commented on the issue:

https://github.com/apache/commons-pool/pull/17

Also, thank you for merging in the meantime, as well as for contributing your time to maintain the project!


---

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-***@commons.apache.org
For additional commands, e-mail: dev-***@commons.apache.org

[grimreaper]

Github user grimreaper commented on the issue:

https://github.com/apache/commons-pool/pull/17

The new test doesn't fail for me if I reverse the original change:
```
[INFO] -------------------------------------------------------
[INFO] T E S T S
[INFO] -------------------------------------------------------
[INFO] Running org.apache.commons.pool2.impl.TestGenericObjectPool
[INFO] Tests run: 84, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 106.528 s - in org.apache.commons.pool2.impl.TestGenericObjectPool
[INFO]
[INFO] Results:
[INFO]
[INFO] Tests run: 84, Failures: 0, Errors: 0, Skipped: 0
[INFO]
[INFO] ------------------------------------------------------------------------
[INFO] BUILD SUCCESS
[INFO] ------------------------------------------------------------------------
[INFO] Total time: 01:57 min
[INFO] Finished at: 2018-11-20T22:58:11-08:00
[INFO] ------------------------------------------------------------------------
mvn test -Dtest=TestGenericObjectPool 34.10s user 7.66s system 34% cpu 1:59.76 total; max RSS 305800Ki
[45682 22:58:11.583 ***@FlyingEagle ...ache/commons/commons-pool]∮git diff (git:commons-pool)-[master●]
diff --git i/src/main/java/org/apache/commons/pool2/impl/EvictionTimer.java w/src/main/java/org/apache/commons/pool2/impl/EvictionTimer.java
index b92b87a9..2fc9ffd8 100644
--- i/src/main/java/org/apache/commons/pool2/impl/EvictionTimer.java
+++ w/src/main/java/org/apache/commons/pool2/impl/EvictionTimer.java
@@ -94,7 +94,7 @@ static synchronized void schedule(
static synchronized void cancel(
final BaseGenericObjectPool<?>.Evictor task, final long timeout, final TimeUnit unit) {
task.cancel();
- if (executor != null && executor.getQueue().size() == 0) {
+ if (executor.getQueue().size() == 0) {
executor.shutdown();
try {
executor.awaitTermination(timeout, unit);
```


---

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-***@commons.apache.org
For additional commands, e-mail: dev-***@commons.apache.org

[grimreaper]

Github user grimreaper commented on the issue:

https://github.com/apache/commons-pool/pull/17

hrm... or at least its not consistent. I just ran it twice more and failed only once.
∎mvn test -Dtest=TestGenericObjectPool#testCloseMultiplePools

I'm not generally the biggest fan of tests that don't rely on a fixed clock: you're guaranteed to eventually get a failure.


---

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-***@commons.apache.org
For additional commands, e-mail: dev-***@commons.apache.org

[garydgregory]

Github user garydgregory commented on the issue:

https://github.com/apache/commons-pool/pull/17

I think this PR can be closed.


---

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-***@commons.apache.org
For additional commands, e-mail: dev-***@commons.apache.org

[mswintermeyer]

Github user mswintermeyer closed the pull request at:

https://github.com/apache/commons-pool/pull/17


---

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-***@commons.apache.org
For additional commands, e-mail: dev-***@commons.apache.org