atlan-cli-run-test-loop

Atlan CLI Run Test Loop

Execute run/test/fix loops that a developer would expect from a normal app request.

Workflow

Resolve target app path.
Verify CLI availability first:
Check command -v atlan .
If missing, invoke atlan-cli-install-configure before run/test.
Do not begin by searching for a local atlan-cli repository.
Re-verify with command -v atlan && atlan --help .
If network/install is blocked, stop and ask the user to enable installation or provide an existing CLI binary path.
Verify infra prerequisites before run/e2e:
uv , temporal , dapr are available.
Dapr runtime initialization is present (config path exists and sidecar can start).
If missing, run atlan app init tools first; if issue persists, apply manual Dapr recovery and record it.
Use CLI-first commands:
atlan app run -p <app_path>
atlan app test -p <app_path> -t unit
atlan app test -p <app_path> -t e2e
Use fallback commands only when CLI path is unavailable or mismatched:
uv run poe start-deps
uv run main.py
uv run pytest
Record each cycle in loop_report.md using ../_shared/assets/loop_report.md .
If command behavior is unclear or conflicting, verify against CLI docs/code and run atlan-fact-verification-gate .
If a CLI mismatch appears, append proposal to ../_shared/references/cli-change-proposals.md .

Loop Contract

Capture commands, failures, root cause, patch plan, and rerun result.
Prefer deterministic command sequences and explicit paths.
Treat ATLAN-CLI-APP-0012 / dependency startup failures as infra blockers; collect logs and apply the run-matrix recovery steps.
Do not imply or perform CLI repo edits.

References

Source Transparency