Krzysztof kuhy Rudnicki
a10d190b65
Some checks failed
file length / file-length (push) Waiting to run
md-naming / markdown naming convention (push) Waiting to run
palette drift / palette (push) Waiting to run
design_system tests / test (push) Has been cancelled
gatelock pre-commit / pre-commit (push) Has been cancelled
gatelock tests / test (3.10) (push) Has been cancelled
gatelock tests / test (3.11) (push) Has been cancelled
gatelock tests / test (3.12) (push) Has been cancelled
web_ui tests / test (push) Has been cancelled
Markdown across every repo had accumulated a dozen ad-hoc names (refactor_claude_todo.md x23, DESIGN_AUDIT_TODO.md, NEXT_SESSION, HANDOFF, PLAN, SPEC) with nothing distinguishing a live task from a finished record. An audit found docs asserting work was pending that shipped months earlier, and one asserting completion while the work was open. Four namespaces now: README.md, CLAUDE*.md, DOCS*.md, TODO*.md. Task files carry the literal "REMOVE ME AFTER FINISH" and are deleted when the work lands. Enforced bidirectionally -- a TODO without the marker fails, AND any file carrying the marker that is not named TODO*.md fails. Without the reverse rule a task can be renamed into the permanent namespace and quietly live forever, which is the failure being fixed. md_naming/ mirrors the file_length topology and imports its tables (EXCLUDED_DIRS, THIRD_PARTY_REPOS, VENDORED_ANYWHERE, GENERATED_PATTERN) rather than restating them, so the two gates cannot disagree about the same file. 71 tests, 100% branch coverage. stale.py warns on TODOs past 90 days but never fails: no machine can prove a task is finished, so the honest mechanism is a nudge. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01A47aX5DJwmA3QFn3pkEWi7 |
||
|---|---|---|
| .. | ||
| lib | ||
| scripts | ||
| test | ||
| .gitignore | ||
| .pre-commit-config.yaml | ||
| analysis_options.yaml | ||
| pubspec.yaml | ||
| README.md | ||
design_system
The shared Flutter component layer for kuhy's apps.
~/utils/unified-design-system/ froze the token layer — one palette, one
spacing scale, one type scale — but it is documentation only: 1220 lines of
prose and annotated HTML, zero importable code. Every repo transcribed the
token table by hand, and then drifted from it.
This package is that same token set as Dart, plus the handful of widgets that had each been reimplemented in three or more repos. The components own the tokens: a consumer that depends on this package has nothing local left to edit, so the palette is enforced by the compiler and the package manager instead of by prose plus discipline.
Consuming it
Tag-pinned git dependency, matching the convention sync_settings_ui and
crdt_sync_dart already use:
dependencies:
design_system:
git:
url: https://github.com/kuhyx/utils
ref: design_system-v0.1.0
path: design_system
Then delete the repo's local lib/ui/theme.dart and import the package:
import 'package:design_system/design_system.dart';
MaterialApp(
theme: buildLightTheme(),
darkTheme: buildDarkTheme(),
// ...
);
What's in it
| Export | Replaces |
|---|---|
AppPalette |
~318 raw Color(0x…) literals, almost all inside hand-copied theme files |
AppSpacing / AppRadius / AppTextSize |
the six-repo consensus scales, transcribed per repo |
buildLightTheme() / buildDarkTheme() |
10 divergent theme.dart copies (2 byte-identical, 8 drifted) |
AppStatusColors |
4 hand-rolled ThemeExtensions with inconsistent field sets |
confirmDestructive() |
~35 inlined AlertDialogs |
showToast() / showError() |
~38 raw SnackBar sites and zero helpers |
EmptyState |
4 variants, only one of them reusable |
SectionHeader |
3 near-identical private widgets |
Why the scales are what they are
The values are the six-repo consensus — todo, home_inventory, dufs-cloud,
habit_stack, and untools all shipped byte-identical copies — which is also
exactly what unified-design-system/DOCS-tokens.md specifies. Repos that had
drifted (diet-guard, macro-cam) are corrected to this package, not the
other way round.
Why explicit ColorSchemes, never fromSeed
The palette is hand-picked to hit specific contrast ratios. A seeded scheme
silently replaces those values with algorithmically derived ones. The
secondary/tertiary slots are set explicitly for the same reason: without
them, widgets that reach for secondaryContainer (e.g. SegmentedButton's
selected segment) fall back to stock Material teal — confirmed on-device.
Development
flutter analyze # must be clean
dart format lib test
flutter test --coverage
scripts/check_coverage.sh coverage/lcov.info 100 # 100% line coverage, gated
Coverage is a gate, not a convention: this ships as a 100%-covered shared package by requirement. The gate runs on pre-push.