devops: Use the same entrypoint.sh for all Docker stages

[gustavovalverde]

Describe the issue or request

We're using a different entrypoint for tests and for our user image, and both have different behaviors and approaches when running Zebra. This causes some tests to behave or be called differently, not allowing a single standard way to interact with the resulting artifact.

Expected Behavior

Handle production and non-production use-cases with the same parameters/variables.

Current Behavior

Some tests fail if we use the same approach as with the production image, like passing the NETWORK environment variable, for example:

• https://github.com/ZcashFoundation/zebra/actions/runs/6323068078/job/17170285661

This is caused by the entrypoint, as one considers this changes, and the other doesn't.

Possible Solution

Have a single entrypoint, with all the scenarios combined.

Additional Information/Context

No response

Is this happening on PRs?

Just on #7616

Is this happening on the main branch?

No