Instead of relying on examples, implement unit tests and update GitHub Actions to exercise tests.

Offsets and counts for comparing line output should be allowed (e.g. compare only the first 10 lines of output, or 10 lines after the 1st, etc.)

Filtering tests does not output the correct number of tests being run/detected.

When the return code is explicitly specified to be 0, the return-code check is skipped

The problem is at contest/TestCase.py#L142:

            if self.return_code and int(self.return_code) != proc.returncode:

should be:

            if self.return_code is not None and int(self.return_code) != proc.returncode:

Switch to using loguru for logging instead of configuring/managing our own logger

When specifying a resource to copy to the test folder, one must specify the src and dst, even if the name is going to be preserved. A resource should be allowed to be specified as a string, and contest should figure out the src and dst from that. i.e. allow:

# see https://github.com/Lnk2past/contest
executable: ./bin/main.exe
test-cases:
    - name: test-01
      resources:
        - data

The output_file and python_console_app tests both fail right after loading the yaml with the error:

##[error]Process completed with exit code 1.