Vladimir Sitnikov

Regarding the change, there should be a way to enable debug logging when running from both command line and from within IDE. Current approach is as follows: 1) `build.properties` (e.g....

@grzm , I'm on vacation till the 27th and I can't review patches (un)fortunately.

@grzm, it looks like (see Travis results) the code is not Java 6 / Java 7 compatible. Can you please adjust that?

The following works for me: 1) Install Java 6 somehow (https://support.apple.com/kb/DL1572 might work) 2) Configure [toolchains.xml](https://github.com/pgjdbc/pgjdbc/blob/master/toolchains.xml) (it should be placed to `~/.m2/toolchains.xml`) I have the following for Java 6: ```xml...

> By the way, do your tests run in under 5 seconds? The output was from `mvn compile` just to show how a compiler from toolchains is used

@grzm , I do remember this exists. Thanks for your patience.

@davecramer , not (?) so funny failure is https://travis-ci.org/pgjdbc/pgjdbc/jobs/359002027#L1044 ``` testPGbox[binary = REGULAR](org.postgresql.test.jdbc2.GeometricTest) Time elapsed: 0.016 sec

@davecramer , frankly speaking, I don't think this PR does makes code cleaner. The PR makes subtle changes. For instance, existing implementation of `setCleanupRef` did `unprepare` the reference if there...

@ahachete , I see there is a good point to abstract `Cleaner` pattern. However, 1) `ServerResourcesCleaner#processDeadParsedObjects sendCloseStatementClearReference` does look too tightly coupled to the specific case. Do you think cleaner...

>However, when you say previous code had a single field, in reality is not true: it had three, but they were in the QueryExecutorImpl class (parsedQueryMap, parsedQueryCleanupQueue and cleanupRef). Exactly....