System tests 2.0

Our system tests are unstable, slow, hard to maintain, not debuggable, and don't work under macOS. The system tests are powerful utility, and we need them in the Stork. We need to develop a new, improved solution.

Design

ToDo:

• Refactor the Rakefile (!415 (merged))
• Write the multistage Dockerfile to build and run the Stork. (!423 (merged))
• Write the first system test based on pytest (!428 (merged))
• Write a framework for system tests (!428 (merged))
• Write or rewrite the current tests using the framework (!428 (merged))
• Update the developer guide

Moved to separate issues:

• Rewrite the update package system tests (#746 (closed))

Abandon

• Improve the system test failure diagnostic
• Write the vscode configuration to integrate with Docker containers
• Write the performance counters for build and system tests commands