Feature roadmap: ecosystem-inspired improvements for ruff-sync

[Kilo59]

Feature Roadmap: Ecosystem-Inspired Improvements

After surveying how 8+ ecosystems handle shared linter/formatter configuration, here is a prioritized set of features and improvements to consider for ruff-sync.

Ecosystems Surveyed

Ecosystem	Tool	Mechanism	Key Insight
JavaScript	ESLint	extends: ["my-org-config"] via npm; flat config imports arrays	Composable config arrays with explicit precedence
JavaScript	Prettier	"prettier": "@my-org/prettier-config" in package.json	npm package as single source of truth
JavaScript	Stylelint	extends: "stylelint-config-standard" via npm	Same extends + overrides pattern as ESLint
Ruby	RuboCop	inherit_from: https://raw.githubusercontent.com/.../rubocop.yml	Native remote URL inheritance — closest analog to ruff-sync
Go	golangci-lint	--config path/to/file; no native remote extends	Teams resort to CI scripts that fetch configs — the exact gap ruff-sync fills
Rust	rustfmt/clippy	Parent-directory search; [workspace.lints] inheritance	Only works within monorepos; cross-repo sharing needs external tooling
Python	Ruff (extend)	extend = "../parent/ruff.toml" — local paths only	astral-sh/ruff#12352 requests remote URL support
Python	pre-commit	No include; each repo needs its own .pre-commit-config.yaml	Same config centralization gap
Multi-lang	Projen	Generates project files from typed definitions	"Config-as-code" generator model

RuboCop's inherit_from with remote URLs is the closest native peer to what ruff-sync does. If Ruff ever adopts remote extend (ruff#12352), ruff-sync would still add value through selective exclusions, format-preserving merges, check/diff commands, and multi-tool support.

---

🟢 Tier 1: High Impact, Natural Extensions

1. ✅ Config inheritance / layering (like ESLint flat config)

Allow composing config from multiple upstream sources with explicit precedence.

[tool.ruff-sync]
upstream = [
  "https://github.com/my-org/base-standards",         # base layer
  "https://github.com/my-org/python-web-standards",    # domain layer (overrides base)
]

Why: Organizations often have a base config + domain-specific overlays (web, data science, CLI tools). ESLint's flat config is an ordered array; RuboCop's inherit_from accepts a list.

2. Drift detection with richer CI output

Enhance check command for CI consumption.

• --output-format json|sarif|github for machine-readable output
• GitHub Actions annotations (::error file=...) to surface drift directly on PRs
• Exit code differentiation (e.g., 1 = out of sync, 2 = upstream unreachable)

Inspiration: golangci-lint outputs SARIF, JSON, and GitHub annotations; ESLint supports --format json and custom formatters.

3. Lock/pin upstream versions

Pin to a specific commit SHA or tag for deterministic, reproducible syncing.

• Generate a ruff-sync.lock (or section in pyproject.toml) recording the resolved commit SHA after pull
• check verifies against the locked version
• ruff-sync update to bump to latest and update the lock

Inspiration: npm package-lock.json, Go go.sum, pre-commit rev: pinning.

Why: Without pinning, a pull in CI could silently pick up breaking upstream changes.

---

🟡 Tier 2: Valuable Additions

4. Dry-run / preview mode

pull --dry-run that shows the full merged output or writes to stdout without modifying files. check --diff partially covers this, but a dedicated dry-run for pull would complete the picture.

Inspiration: terraform plan, npm pack --dry-run.

5. Merge strategy options per key

Instead of binary sync-or-skip, support different merge behaviors per key:

[tool.ruff-sync]
# Append upstream values to local list instead of replacing
merge-strategy.lint.extend-select = "append"
# Never touch this key
merge-strategy.target-version = "local-only"

Inspiration: Git merge strategies, ESLint flat config composition.

6. status command — richer sync reporting

$ ruff-sync status
ruff-sync v0.2.0
upstream: https://github.com/my-org/standards (main @ abc1234)
target:   pyproject.toml

  ruff   ✅ in sync (last pulled: 2026-03-10)
  mypy   ❌ out of sync (3 keys differ)

7. Monorepo support

Sync multiple sub-projects within a single repository, each with different exclusions:

[[tool.ruff-sync.targets]]
to = "packages/api"
exclude = ["lint.per-file-ignores"]

[[tool.ruff-sync.targets]]
to = "packages/worker"
exclude = ["target-version"]

---

🔵 Tier 3: Forward-Looking / Exploratory

8. Bidirectional sync / push command

Push local config changes back upstream by opening a PR. Use case: a team member discovers a useful rule downstream and wants to propose it upstream.

9. Webhook / GitHub App for automatic PRs

A hosted service watching the upstream repo and automatically opening PRs on downstream repos when config changes. (Renovate / Dependabot / pre-commit.ci model.)

10. Config validation

Validate the merged config before applying:

• Warn if upstream references deprecated Ruff rules
• Warn if upstream target-version conflicts with local requires-python
• Optionally run ruff check against the merged config to validate syntax

---

README: Additional Ecosystems to Document

The README's "How Other Ecosystems Solve This" section currently mentions ESLint, Prettier, and Ruff's extend. Consider adding:

• RuboCop (Ruby) — native inherit_from: <URL> is the closest peer
• Stylelint (CSS) — extends via npm packages
• golangci-lint (Go) — no native remote config; teams use CI workarounds
• Rust — [workspace.lints] for monorepos only
• pre-commit — no include/extends; repos must duplicate config

---

Related Issues

• Support syncing other pyproject.toml tool configurations (e.g., mypy, pytest) #69 — Multi-tool config syncing (overlaps with Tier 1 scope expansion)
• Add short, non‑ruff CLI aliases (e.g., tms, cfgsync, syncup) for occasional use #74 — Short CLI aliases (makes more sense if multi-tool lands)
• Add example Ruff configurations for various domains #83, Example Config: Data Science / Data Engineering #87 — Example config recipes (Tier 2)