Skip to content

site: the family's coloured stock, with lea's accent (lockfamily#72) - #502

Merged
DavidGodefroid merged 2 commits into
mainfrom
site/lockfamily-72-colour
Nov 18, 2026
Merged

DavidGodefroid merged 2 commits into
mainfrom
site/lockfamily-72-colour

Conversation

@DavidGodefroid

@DavidGodefroid DavidGodefroid commented Nov 18, 2026 •

Copy link
Copy Markdown
Owner

Refs DavidGodefroid/lockfamily#72

Applies the family's colour direction (lockfamily site-kit/COLOUR.md, "the ticket is printed on coloured stock") to leadeck.dev, with lea's accent. No look of lea's own: the kit does the work.

What changed

  • site/assets/tokens.css, lock-site.css and logo.svg: copied byte for byte from lockfamily site/72-brighten (e8bf7b6), steps 1 to 3 of the resync list in site-kit/KIT.md.
  • Every page (index.html, install/, how-it-works/, security/, faq/, 404.html): <html lang="en" data-accent="lea">, and both theme-color metas moved to the new --paper (#fbf7ee, #1f1d26). No page writes an accent colour; no page-specific CSS was needed.
  • site/assets/og.png and apple-touch-icon.png: rendered again from the kit's tools/og.html (e8bf7b6), light scheme: the card and the icon sit on lea's stock.
  • tests/site/test_site.py: the three kit hashes, the <html> line and the theme-color values follow.
  • assets/BRAND.md: says where the colour rule lives.

To check before merging

  • The kit is lockfamily site/72-brighten at e8bf7b6 (lockfamily#76), not yet on lockfamily main. tokens.css, lock-site.css, logo.svg, copy.js and the fonts were compared with that commit with cmp: identical.
  • Step 6 of the resync list (README banners docs/assets/banner-light.svg, banner-dark.svg) does not apply here: lea has no such files, its README shows lea's own wordmark (assets/logo.svg, logo-light.svg), which this PR leaves alone.

How it was verified

  • make test-site: the unit test passes and the end-to-end test in headless Chrome ends on all 112 passed (every page, light and dark, 1440 and 390 px: no overflow, --paper background, display face, Copy button, no console error).
  • Every page looked at in both schemes at 360, 768 and 1280 px (prefers-reduced-motion: reduce): no horizontal overflow.
  • Contrast: a script walked every visible text run of the six pages in both schemes at 360 and 1280 px (4780 runs) and compared its colour with the nearest opaque background: none under 4.5:1 (3:1 for large text). The demo SVG keeps the deck's own terminal palette and is not part of that check.

Before / after (home page)

Full pages; click one to open it at full size. The files are on the throwaway branch pr-assets/502 (delete it after the merge).

Before After
light, 1280 px before, light, 1280 px after, light, 1280 px
light, 768 px before, light, 768 px after, light, 768 px
light, 360 px before, light, 360 px after, light, 360 px
dark, 1280 px before, dark, 1280 px after, dark, 1280 px
dark, 768 px before, dark, 768 px after, dark, 768 px
dark, 360 px before, dark, 360 px after, dark, 360 px

🤖 Generated with Claude Code

DavidGodefroid and others added 2 commits November 18, 2026 12:13
Copies the brighter kit from lockfamily 94661f1 (tokens.css, lock-site.css,
logo.svg, byte for byte), names lea's accent once per page
(<html data-accent="lea">), moves theme-color to the new --paper and
re-renders og.png and the touch icon from the kit's og.html. The test's
kit hashes and head assertions follow.

Refs DavidGodefroid/lockfamily#72

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
…ly#72)

Rendered from lockfamily e8bf7b6 site-kit/tools/og.html: the card and the
icon sit on lea's coloured stock. tokens.css, lock-site.css and logo.svg
are unchanged at that commit (checked with cmp).

Refs DavidGodefroid/lockfamily#72

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
@DavidGodefroid
DavidGodefroid merged commit f965555 into main Nov 18, 2026
26 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant