DevMojoIT: Remove superfluous first mvn install to avoid race condition

[famod]

As already mentioned on zulip, I think the first mvn install call was a c&p mistake or something. But I might be wrong.

One can clearly see the two "racy" executions in integration-tests/maven/target/test-classes/projects/multimodule-resources-classpath/build-multimodule-resources-classpath.log:

[INFO] Scanning for projects...
[INFO] Scanning for projects...
[INFO] ------------------------------------------------------------------------
[INFO] Reactor Build Order:
[INFO] 
[INFO] ------------------------------------------------------------------------
[INFO] Reactor Build Order:
[INFO] 
[INFO] multimodule-cp-resources                                           [pom]
[INFO] multimodule-cp-resources-rest                                      [jar]
[INFO] multimodule-cp-resources-html                                      [jar]
[INFO] multimodule-cp-resources-main                                      [jar]
[INFO] 
[INFO] ------------------< cp.acme:multimodule-cp-resources >------------------
[INFO] Building multimodule-cp-resources 1.0-SNAPSHOT                     [1/4]
[INFO] --------------------------------[ pom ]---------------------------------
[INFO] multimodule-cp-resources                                           [pom]
[INFO] multimodule-cp-resources-rest                                      [jar]
[INFO] multimodule-cp-resources-html                                      [jar]
[INFO] multimodule-cp-resources-main                                      [jar]
[INFO] 
[INFO] ------------------< cp.acme:multimodule-cp-resources >------------------
[INFO] Building multimodule-cp-resources 1.0-SNAPSHOT                     [1/4]
[INFO] --------------------------------[ pom ]---------------------------------

I added another assert before the existing return code assert as we might see more helpful output in case the test is still flaky.

[geoand]

Wow, great catch!