It appears that frameworks use different status codes depending on what they receive and we need to ensure that a request produces identical responses across every framework.
Instead of bloating the project with many tests there should be a runnable program that sends our expected suite of fixture request to each service as it is running in the container and gather the results in some structured format so it's easy to see which framework is inconsistent for which request.
We can start the generator in a state where it crashes (or perhaps we may have some env var to disable it or just not start it at all) and the program can run its suite and print its results.
This extends beyond status codes as response bodies must be consistent.
It appears that frameworks use different status codes depending on what they receive and we need to ensure that a request produces identical responses across every framework.
Instead of bloating the project with many tests there should be a runnable program that sends our expected suite of fixture request to each service as it is running in the container and gather the results in some structured format so it's easy to see which framework is inconsistent for which request.
We can start the generator in a state where it crashes (or perhaps we may have some env var to disable it or just not start it at all) and the program can run its suite and print its results.
This extends beyond status codes as response bodies must be consistent.