I have an origin repo that I have full access to and I don't want code to be pushed successfully to this repo unless the code being pushed successfully passes my unit tests. I saw many examples of .git/hooks/update scripts and they seem to break down into a few categories that do not fit my use case.
For example, in (http://git-scm.com/book/en/Customizing-Git-An-Example-Git-Enforced-Policy), the script accesses files locally in the .git/hooks/
directory (ACL example) or individual files in the new or old SHAs.
I need to do an entire build process on all the new files as if I were in the directory of the new commit and were running
#!/usr/bin/bash
mvn test
exit $?
I would recommend a guarded commit approach, where you are pushing to an intermediate repo, which: