py(deps[dev]): Require ruff>=0.16.0

[tony]

Summary

• Pin a ruff>=0.16.0 floor in the dev and lint dependency groups so contributors and CI resolve the same linter and see the same diagnostics.
• Relock uv.lock onto ruff 0.16.0.
• Adopt ruff's default rule set. select becomes extend-select, so this project's linters layer on top of ruff's curated default set instead of replacing it. Enabled rules go from 351 to 565.
• One new finding, scoped as an ignore. The wider rule set surfaced a single diagnostic — S102 on the exec in docs/conf.py — and it is intentional.

Changes

pyproject.toml — ruff raised to >=0.16.0 in [dependency-groups] dev and lint; [tool.ruff.lint] select renamed to extend-select with a comment recording why select stays unset; S102 per-file-ignored for docs/conf.py.

uv.lock — ruff 0.15.22 -> 0.16.0.

CHANGES — entries for the floor bump and the rule-set adoption.

Design decisions

Why extend-select rather than a wider select

ruff 0.16 ships a curated default rule set as its recommended baseline. An explicit select replaces that set rather than extending it, so this project was silently opting out of everything ruff considers a sensible default. There is no token that names the default set from inside select — ruff's selector grammar has no DEFAULT — so the only way to get defaults-plus-extras is to leave select unset and layer the project's linters through extend-select.

The alternative of naming whole prefixes (PL, S, PYI, ...) in select was rejected: it drags in the full contents of each family rather than the curated subset, which is far noisier for no added signal.

Every extend-select entry stays on its own line with a trailing comment naming the linter it selects. FA100's comment now names flake8-future-annotations rather than describing the rule.

The one finding

S102 (exec-builtin) in docs/conf.py. Sphinx collects the project's title, copyright, and URLs by exec'ing the package's __about__.py into a dict, because g/__init__.py re-exports only __version__. The exec'd input is a tracked file in this repository, not user data, so the flake8-bandit warning has nothing to warn about here. Scoped as a per-file-ignore on docs/conf.py rather than repo-wide, with the reason recorded in the config.

No code changes were needed: nothing else in src/, tests/, or docs/ trips a default rule.

ruff 0.16 rule stabilizations

ruff 0.16.0 stabilizes several rules that live inside prefixes this project already selects, notably RUF036, RUF063, and RUF068, plus behavior changes to UP019 and the B/SIM/TRY families. None of them fire here.

ruff 0.16.0 also formats Python and pycon code blocks inside Markdown by default. All Markdown under the repo (including docs/) is picked up by ruff format and came back already formatted, so there is no churn. CHANGES has no .md extension and is not in scope for ruff format; it stays on the prettier path in the format_markdown make target.

See https://astral.sh/blog/ruff-v0.16.0 for the full release notes.

Test plan

• uv run ruff check . — all checks passed
• uv run ruff format --check . — every file already formatted
• uv run mypy . — no issues found
• uv run py.test — passes in a normal checkout

Note on running the suite from a linked worktree: tests/test_cli.py::test_command_line[g-cmd-inside-git-dir] and [g-cmd-help-inside-git-dir] fail there because find_repo_type requires .git to be a directory, and a linked worktree's .git is a file. Both fail identically at the base commit, so this is not a regression from this branch, and CI's plain checkout is unaffected.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 96.89%. Comparing base (c3da8ed) to head (2fc33bf).

Additional details and impacted files

@@           Coverage Diff           @@
##           master      #63   +/-   ##
=======================================
  Coverage   96.89%   96.89%           
=======================================
  Files           5        5           
  Lines         129      129           
=======================================
  Hits          125      125           
  Misses          4        4           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.