mirror of
https://github.com/kevinveenbirkenbach/computer-playbook.git
synced 2026-11-06 10:59:59 +00:00
kevinveenbirkenbachandClaude Opus 4.7
02da33b0a5
Under tests/integration/roles/ introduce four sub-clusters (meta/,
dependencies/, applications/, run_once/) plus nested dependencies/run_after/
and applications/id/. Fold tests/integration/application_id/ into
roles/applications/id/ and tests/integration/run_once/ into roles/run_once/.
Pull four application-related tests out of tests/integration/ into
roles/applications/, and test_run_after_references.py into the run_after/
sub-cluster.
Consolidate tests/integration/linting/ into tests/lint/{ansible,filesystem,
repository}/, so repository-wide lint checks now run under make test-lint
rather than make test-integration.
Each new cluster ships a README.md and __init__.py; cross-referencing
READMEs (categories, ports, services, roles, applications) are updated to
point at the new paths. Path depths inside moved tests and one hardcoded
ignore path in test_domain.py are adjusted accordingly. No behavior change;
all 32 lint tests and every moved cluster pass locally.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
586 B
586 B
Run-Once Flag Tests 🏁
Integration tests for the run_once_<role_name> fact pattern: flag suffixes MUST match their role folder name, flags MUST be defined before use, and inclusion of utils/once/flag.yml MUST follow the schema.
Tests in this directory MUST only cover run_once_* flags and their lifecycle. Tests for generic fact overrides, task includes, or block structure MUST live elsewhere under tests/integration/.
For framework, directory layout, and make test-integration usage see integration.md.