Refactoring how tests are run and coverage reports are generated

Problems trying to solve:

• Previously test is run with a "run_test.in" generated script.
• It is not obvious how to run the tests without meson.
• Extra add-in files just to make test working with meson.
• Unit tests are not very readable.

Obstacles:

• Test files should not be copied to the build folder because they are not to be distributed.
• Test should run against the files either in the source folder (src), or files to be distributed.
• Debian have different executable path for running Python3's coverage package.

Solution:

• "coverage" can simply be run as a python module (i.e. python3 -m coverage) and even support testing another package (i.e. `python3 -m
• meson's test() accepts a workdir parameter. So we can simply run command in the root directory without an extra script in builddir.
• Rewrite the "run_test.in" to be a "main.py" in the test folder to make it easier to run (python3 -m tests to run it) coverage run -m PACKAGE_NAME`). This will remove the need for an extra fix-up script for path.
• Running with colour_runner will group the tests in a more humanly readable way.