mirror of
https://github.com/kuhyx/testsAndMisc.git
synced 2026-11-11 20:54:00 +01:00
48 lines
2.9 KiB
JSON
48 lines
2.9 KiB
JSON
{
|
|
"intent": "Retint kcd2_dice_solver/src/index.css to the frozen unified-design-system palette and fix every violation in the repo's own DESIGN_AUDIT_TODO.md entry for this app.",
|
|
"scope": ["kcd2_dice_solver/src/index.css"],
|
|
"changes": [
|
|
"Root font-size 15px -> 16px (fixes body-text floor and straightens the rem-based spacing grid onto clean 4px multiples in one change, per the audit's own root-cause note)",
|
|
"New --tier-tin/--tier-silver custom properties replace hardcoded #a8a196/#cfd4da hex; tier-tin also darkened (was clustering within 2 L* points of --accent/--good) and tier-silver re-tinted warm (was the one cool-leaning hex in an otherwise all-warm palette)",
|
|
"--line darkened-neutral -> #736750 (was ~1.4:1 against bg/panel, now >=3:1 against both, computed)",
|
|
"header h1 letter-spacing 0.02em -> -0.01em (largest text was widening tracking instead of tightening it)",
|
|
"New --radius-sm(3px)/--radius-md(8px) tokens replace the ad-hoc 3/4/6/8px radii; nested elements (.search, .die-list, .die-main, .die-tag, .stepper-value, .stats div) now use the smaller radius since they sit well inside an 8px-radius, 16px-padding outer panel (nesting should trend toward 0, not stay equal to the outer radius)",
|
|
".search and .clear/.stepper-button padding corrected to an exact 2x horizontal:vertical ratio (0.5rem 1rem); was 1.5x and 1.75x respectively",
|
|
".hint font-size 0.85rem -> 1rem (flagged reading-content text under the 16px floor)"
|
|
],
|
|
"verification": [
|
|
{
|
|
"command": "pnpm run lint",
|
|
"result": "pass",
|
|
"evidence": "tsc --noEmit && eslint src clean"
|
|
},
|
|
{
|
|
"command": "pnpm run coverage",
|
|
"result": "pass",
|
|
"evidence": "14 test files, 210 tests passed; 100% statements/branches/functions/lines"
|
|
},
|
|
{
|
|
"command": "pnpm run build",
|
|
"result": "pass",
|
|
"evidence": "vite build succeeded, dist/ regenerated"
|
|
},
|
|
{
|
|
"command": "isolated git worktree re-run of lint+coverage+build at the commit's parent HEAD",
|
|
"result": "pass",
|
|
"evidence": "identical results to the dirty-tree run"
|
|
},
|
|
{
|
|
"command": "pnpm run preview + headless-Chrome screenshot of the real production build",
|
|
"result": "pass",
|
|
"evidence": "visually reviewed: crisp visible borders on search/die-list/panels, tighter title tracking, distinct tin/silver/gold tier colors, no layout breakage"
|
|
}
|
|
],
|
|
"risks": [
|
|
"No deploy target for this repo (local dev tool) — verification ceiling is the production build + preview, not a live check",
|
|
"Exact L*/contrast values were hand-computed with the standard sRGB->linear approximation (not the WCAG piecewise formula), same caveat the original audit itself carries; margins were kept generous (>=3.1:1 vs the 3:1 floor) to absorb the approximation error"
|
|
],
|
|
"rollback": [
|
|
"git revert this commit — the change is a single CSS file, no other files coupled to it"
|
|
]
|
|
}
|