chore(template): rebuild from templates 57e7024 #2

Merged
gravityfargo merged 1 commit from template-rebuild into main 2026-07-31 05:06:20 +00:00
Owner

Full template rebuild against templates@57e7024 (base python, no variants).

Every template-owned file was deleted and re-applied in layer order. Four files changed:

  • Makefile — new clean target (build/venv/cache artifacts plus a __pycache__ sweep), added to .PHONY. clean is now part of the shared-layer target contract.
  • pyproject.toml — replaced from the baseline; see the two divergences below.
  • uv.lock — regenerated last, after every layer landed.
  • .coveragerc.toml — see below.

The real tests (cli_test.py, sort_keys_test.py, smoke_test.py) were kept and no starter test was applied; smoke_test.py here is a genuine CLI test, not the template's starter. Both workflows came back byte-identical. No placeholders survive.

Divergences

1. [project.scripts] entry point — preserved, not substituted to the template's convention.

The template ships "<project-name>" = "<package>:main". Substituting that literally yields local_tools:main, but src/local_tools/__init__.py exports only __version__ — there is no main at package root. The real target is the typer app at local_tools.main:app, so that is what the rebuild kept. A literal substitution would have shipped a broken console script.

2. .coveragerc.toml — the project's omit was dropped, per the baseline-stands rule.

The displaced file carried:

# `python -m local_tools` shim; it only runs as a subprocess, never under pytest.
omit = ["src/local_tools/__main__.py"]

That reads as a legit patch rather than a config hack — it describes a real property of the project (a __main__.py shim that genuinely cannot be exercised in-process by pytest) rather than concealing untested code. It is reported here rather than reapplied; restoring it is your call.

It is not load-bearing for the gate: with the omit gone, __main__.py reports 0% and total coverage is 93.30% against a fail_under of 80.

3. Dev-group version floors droppedmypy>=2.3.0, pytest>=9.1.1, pytest-cov>=7.1.0, python-dotenv>=1.2.2, ruff>=0.16.1 become bare names, matching the baseline. Runtime dependencies were restored through uv add, not hand-edited, and resolved back to the same floors.

Local gate (make pre-commit) passes: ruff clean, ruff format no changes, mypy clean on 8 source files, 34 tests pass.

Full template rebuild against `templates@57e7024` (base `python`, no variants). Every template-owned file was deleted and re-applied in layer order. Four files changed: - `Makefile` — new `clean` target (build/venv/cache artifacts plus a `__pycache__` sweep), added to `.PHONY`. `clean` is now part of the shared-layer target contract. - `pyproject.toml` — replaced from the baseline; see the two divergences below. - `uv.lock` — regenerated last, after every layer landed. - `.coveragerc.toml` — see below. The real tests (`cli_test.py`, `sort_keys_test.py`, `smoke_test.py`) were kept and no starter test was applied; `smoke_test.py` here is a genuine CLI test, not the template's starter. Both workflows came back byte-identical. No placeholders survive. ## Divergences **1. `[project.scripts]` entry point — preserved, not substituted to the template's convention.** The template ships `"<project-name>" = "<package>:main"`. Substituting that literally yields `local_tools:main`, but `src/local_tools/__init__.py` exports only `__version__` — there is no `main` at package root. The real target is the typer app at `local_tools.main:app`, so that is what the rebuild kept. A literal substitution would have shipped a broken console script. **2. `.coveragerc.toml` — the project's `omit` was dropped, per the baseline-stands rule.** The displaced file carried: ``` # `python -m local_tools` shim; it only runs as a subprocess, never under pytest. omit = ["src/local_tools/__main__.py"] ``` That reads as a legit patch rather than a config hack — it describes a real property of the project (a `__main__.py` shim that genuinely cannot be exercised in-process by pytest) rather than concealing untested code. It is reported here rather than reapplied; restoring it is your call. It is not load-bearing for the gate: with the omit gone, `__main__.py` reports 0% and total coverage is 93.30% against a `fail_under` of 80. **3. Dev-group version floors dropped** — `mypy>=2.3.0`, `pytest>=9.1.1`, `pytest-cov>=7.1.0`, `python-dotenv>=1.2.2`, `ruff>=0.16.1` become bare names, matching the baseline. Runtime dependencies were restored through `uv add`, not hand-edited, and resolved back to the same floors. Local gate (`make pre-commit`) passes: ruff clean, ruff format no changes, mypy clean on 8 source files, 34 tests pass.
chore(template): rebuild from templates 57e7024
All checks were successful
Test / test (pull_request) Successful in 10s
Test / checks (pull_request) Successful in 13s
Test / code-quality (pull_request) Successful in 0s
376f6b5bb4
Sign in to join this conversation.
No reviewers
No labels
No milestone
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference
gravityfargo/local-tools!2
No description provided.