There are a couple differences between CE and Fuse S2I processes, we should work out the common approach so the customers get the same experience from using xPaaS images.

These are the differences I have found (I went through all S2I assemble scripts):

• name difference: $MAVEN_BUILD_DIR (Fuse) vs. $ARTIFACT_DIR (CE)
• semantics: settings.xml
  • CE uses settings.xml baked in image with placeholders for proxy and miror settings
  • Fuse uses settings.xml located in .m2/settings.xml in git repo
• Fuse uses two variables for maven args ($MAVEN_ARGS and $MAVEN_EXTRA_ARGS), CE uses only $MAVEN_ARGS
• CE uses deployments dir in git repo for binary deployments, Fuse uses target (java) or root (karaf) directory in git
• CE copies the git sources to home directory (unnecessary)
• Fuse deletes maven artifacts from /tmp if $MAVEN_CLEAR_REPO is set to true (unnecessary)
• Fuse (karaf) has $BUILD_LOG_LEVEL environment variable not used anywhere else