Bun.color — color parsing / conversion utility
Moved deferred → out-of-scope. Re-checked against the now-shipped Patties UI
Out-of-scope (2026-05-31)
Moved deferred → out-of-scope. Re-checked against the now-shipped Patties UI catalog: theming ([[rfc-patties-ui-theming]] → ui/62-theming-presets) ships pre-authored tokens.css payloads merged at stamp time — the CLI never parses, converts, or generates color values, and persists no baseColor key. So patties-ui is not a Bun.color consumer. The only triggers the deferral named (a first-party design-token system, or a built-in component generating gradients/palettes) would each reverse the "framework, not design system" stance — a permanent no, not a wait-and-see. Kept as a discovered-and-dismissed record. Original deferral verdict preserved below.
Review verdict (2026-05-27)
Reject for v1; keep as future-work RFC. Patties is a meta-framework, not a design system. Styling is delegated to the user — Tailwind, CSS Modules, Vanilla Extract, whatever they choose. There is no framework slot for color utilities and the current project stance is "framework, not design system."
Revisit when:
- A first-party
patties/styleor design-token system enters the roadmap (currently not planned), OR - A built-in component (e.g.
<Image>with placeholder gradient generation) needs color parsing internally.
Re-evaluation trigger: This RFC is borderline-out-of-scope — if no design-tokens / styling surface ever opens, it likely moves to out-of-scope on next refresh.
No spec changes. File preserved as status: deferred.
Summary
Bun.color parses / converts color strings (hex, RGB, HSL, OKLCH) without a userland lib.
Motivation
Useful in design-token / styling-helper contexts. Patties has none of those today.
Trade-offs
- Adopting would commit to a styling opinion. Patties' current stance is "framework, not design system."
Open questions
- Does the framework actually want to ship design opinions? Current stance: no. Revisit if this changes.
- If yes, what's the API contract — runtime helper, build-time token generation, or both?