There have been cases with wallets upgrading the framework to a newer version and existing credentials would stop working. Is this something of interest to test in the interop tests at the framework level?
The tests would do something like:
Start the container(s) with a perviously released version of a framework (like Credo)
Issue creds and use them
Stop the agent(s)
Upgrade the installation of the framework in the container(s)
Restart the agent(s)
Use existing credentials
issue new creds and used them in combination with old creds
If this test always used the latest release -1, we may only have to go back one official release in the testing. We will have already proven that the last release -1 works, then we can assume we are working from a working valid state with the new release -1.
The test should be common among frameworks and the framework backchannel knows how to upgrade. Or a separate upgrade service that knows how to upgrade and start each framework.
There have been cases with wallets upgrading the framework to a newer version and existing credentials would stop working. Is this something of interest to test in the interop tests at the framework level?
The tests would do something like:
If this test always used the latest release -1, we may only have to go back one official release in the testing. We will have already proven that the last release -1 works, then we can assume we are working from a working valid state with the new release -1.
The test should be common among frameworks and the framework backchannel knows how to upgrade. Or a separate upgrade service that knows how to upgrade and start each framework.