In the current version of the docker-compose file, configuration options are specified in .env and then passed to the container using ENV_VAR=${ENV_VAR} under the environment section of the docker-compose file. This PR instead passes .env to the container directly, so any changes/additions made to .env are applied to the container without needing to add a corresponding entry in environment.

The motivation for this change is that in configuring LDAP on my installation of Davis, I was adding the LDAP env vars to .env but forgetting to pass them in the compose file. All the examples in the README use env-file syntax, so I just copied those over to my .env file without thinking. It took a few hours debugging my LDAP provider before realizing that Davis was never seeing any of my changes in the first place!

Setting aside my personal oversights, I prefer this method because it reduces redundant code (i.e. adding an entry in .env and a corresponding entry in environment per variable) which means less maintenance when additional configuration variables are added. It also reduces duplicated code across the four different compose files.

This PR also propagates c29dc78 to the .env file in the docker configuration directory.