dockerrun.sh messes with permissions

[tomascohen]

We need to do it better.

There shouldn't be need to on the mounted code (to say the least) as it will bork the permissions/owner on the host OS.

In KTD we pass the running user ID and make sure commands don't change ownership in a way that could break things.

[tomascohen]

I think the need to run setup_aspen_user_debian.sh is, on its own, a symptom of a design issue in the ADB stack. If you look closely to the actual Aspen image, you will notice that everything actually belongs to the www-data user.

In my opinion, the dockerrun.sh script should do what we do in KTD and just change the UID for www-data if needed. But not break the git clone with an otherwise unneeded chown.