Add support for __network_test_flags to workaround CI problems

[richm]

On some CI systems the tests that create routes can create them in such a way that they override the "default" default routes, interrupting network connectivity and causing Ansible to hang. The parameter __network_test_flags gives network role developers a way to ensure that the newly created route is given a lower priority and does not override the "default" default route. A dict was chosen as the type of this variable so that it could be easily extended in the future for future use cases.

[coveralls]

Coverage increased (+0.9%) to 43.481% when pulling 268fcf0551a7c8d9b52249d2fdfc2e5421d29a75 on richm:large-metric into cfbd14cd8ab4df0525c9c2aecb47b8b2fe3f01a2 on linux-system-roles:main.

[liangwen12year]

lgtm, but it seems that the tox -e black failed.

[richm]

lgtm, but it seems that the tox -e black failed.

yes - I don't want to merge this yet - I'm anticipating other reviewers may have suggested changes, or an entirely different approach - this is here just as a PoC - I'm using this PR to test in baseos CI

[liangwen12year]

This pr can be closed as we prefer the solution in https://github.com/linux-system-roles/network/pull/532.