Mechanical, semantics-preserving cleanup so the codebase passes the new lint stage:
- ruff format: all 32 Python files (wraps long signatures, normalizes quotes/blank lines).
- nixpkgs-fmt: modules/drone-runner.nix.
- shfmt (-i 2 -ci): scripts/*.sh.
Lint fixes (reviewed, behavior-preserving — no test weakened):
- ruff SIM105: try/except-pass -> contextlib.suppress (abra.py app_config rm; lifecycle.py janitor).
- ruff SIM115: open().read() -> with open() (run_recipe_ci.py redaction-values + gitea-token).
- statix: merge repeated sops `secrets.*` keys into one `secrets = { ... }` (comments kept);
empty fn pattern `{ ... }:` -> `_:` (packages.nix).
- deadnix: drop unused lambda args (flake `self`; configuration.nix `lib`; overlay `final` -> `_`).
Verified on cc-ci: `scripts/lint.sh` -> lint: PASS; nixosConfigurations.cc-ci evaluates;
all Python byte-compiles. The deployed bridge/dashboard/runner source changes hash (reformat),
so cc-ci will be rebuilt to the new closure in W2 before the cold D1-D10 re-verification.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
7 lines
370 B
Python
7 lines
370 B
Python
# Per-recipe harness config for keycloak (DB-backed: keycloak + mariadb). Read by the shared
|
|
# conftest — enrolling this recipe needs NO change to runner/harness code (D5).
|
|
HEALTH_PATH = "/realms/master" # 200 JSON once keycloak is up (not "/", which redirects)
|
|
HEALTH_OK = (200,)
|
|
DEPLOY_TIMEOUT = 600 # JVM + DB migration are slow on a 2-vCPU VM
|
|
HTTP_TIMEOUT = 600
|