Build p6spy before integration test

[felixbarny]

The integration tests won't be dependent on a snapshot repository
but will rather install the latest master into the local maven repo

[quintonm]

Great idea! However, the build is failing...

[typekpb]

@quintonm build is failing due to problems with tomcat+websphere liberty profile, which we're aware of => should not be a blocker here

@felixbarny the JDK switching might not work properly
See: https://travis-ci.org/p6spy/p6spy-it/jobs/159510803

  "jdk": "oraclejdk7",
  "env": "AS=glassfish31x",

however in the log:

$ jdk_switcher use oraclejdk8
Switching to Oracle JDK8 (java-8-oracle), JAVA_HOME will be set to /usr/lib/jvm/java-8-oracle
before_install.3
13.64s$ git clone https://github.com/p6spy/p6spy.git

...

$ popd
~/build/p6spy/p6spy-it
before_install.7
0.00s$ export JAVA_HOME=$JAVA_HOME_TEMP
before_install.8
0.00s$ echo $JAVA_HOME
/usr/lib/jvm/java-7-oracle
before_install.9
0.06s$ java -version
java version "1.8.0_101"
Java(TM) SE Runtime Environment (build 1.8.0_101-b13)
Java HotSpot(TM) 64-Bit Server VM (build 25.101-b13, mixed mode)
before_install.10
0.00s$ export ORG_GRADLE_PROJECT_container=${AS}
...

So to me it seems like java was not switched back to 7 after the p6spy build.

[typekpb]

please see the discussion in the: p6spy/p6spy#350

Moreover, I believe we should follow suggestion on: http://stackoverflow.com/a/27891518/1581069 and disable gradle caching here via:

configurations.all {
    resolutionStrategy.cacheChangingModulesFor 0, 'seconds'
}