Determine testing methodology #14

Merged

tepichord merged 9 commits from feat/11-determine-testing-methodology into main

2026-08-19 08:56:13 -07:00

tepichord commented

2026-08-19 00:09:02 -07:00

Owner

Closes #11

Automated PR created by forgejo-done adapter.

Closes #11 Automated PR created by forgejo-done adapter.

tepichord added 9 commits

2026-08-19 00:09:02 -07:00

fix(control-plane): correct typo on line 22 of cleanup.sh 0f254d19a4

|| truec → || true

Refs: #11

feat(worker): add source guard to create-worker-vm.sh 1e494001f2

Wrap main "$@" in BASH_SOURCE check so BATS can source the file
and test individual functions without triggering main.

Refs: #11

feat(testing): add test dependencies to flake.nix 112833f4ce

Add python3, pytest, shellcheck, hadolint, yamllint,
bats-with-libraries, and ruff to buildInputs.

Refs: #11

feat(testing): create pytest config and test suites 1b902661bf

- pyproject.toml with addopts = "-v"
- tests/test_static.py: terraform fmt, shellcheck, hadolint, yamllint, ruff
- tests/test_terraform.py: terraform test wrapper
- tests/test_shell.py: BATS wrapper

Refs: #11

feat(testing): add terraform mock provider tests for control-plane 580b8fff6e

Mock libvirt and talos providers to test offline:
- Verify network subnet is 10.5.0.1
- Verify control plane disk is 40GB

Refs: #11

feat(testing): add BATS tests for worker shell helpers 41684a7e08

Test generate_mac produces valid MAC addresses and
create_vm_dir creates directories correctly.

Refs: #11

docs(repo): add testing section to AGENTS.md 27f5191d46

Refs: #11

fix(tests): skip tests when required tools are unavailable 3f91bae0e8

Tests that depend on external tools (terraform, shellcheck, hadolint,
yamllint, ruff, bats) now gracefully skip via pytest.mark.skipif when
the tool is not found on PATH, instead of raising FileNotFoundError.

Also adds .gitignore for __pycache__ and .pytest_cache.

fix(tests): use stdout+stderr in assertion messages for all linters 8c3e1234a5

All lint tools (terraform fmt, shellcheck, hadolint, yamllint, ruff)
write diagnostics to stdout, not stderr. Using only r.stderr.decode()
produced empty assertion messages on failure, making test failures
undebuggable. Switched to (r.stdout + r.stderr).decode() to capture
all output regardless of which stream the tool uses.