eddiecorrigall / database-revisions Goto Github PK

Given the current state of user feedback (printed to standard):

• what is missing that the user needs to know?
• what needs to change to help make it easier to read / interpret?
• etc.

Currently the service assumes that the lexicographically ordered file list is correctly ordered - it may not be.

Ensure that there is a path from the initial revision to the final revision, but back tracking from the final revision. Every IRevision has a IRevision::previousVersion property that should match with another IRevision::version otherwise it will be undefined. Make sure to implement a solution to verify a path from the final revision to the initial revision. If no such path exists, then an error should be thrown and the change set should not be applied.

• Setup in npm project
• Verify the tool works using npm install ...
• Document a release process

Write units tests for the project, especially for DatabaseMigrationService.

Current project state requires that "plugins" are implemented using the existing project dependencies and source code. Ideally plugins should be written by users of the tool to extend the capability and reduce the unused dependencies of the project.

For example, a Redis plugin should standalone so that a Redis client is not a dependency of this project, but a dependency of the system running the migration tool. This will also reduce the project size and overall security, while also enabling anyone to extend the tools capability.