ResumeSite/PROJECT_STATUS.md

PROJECT_STATUS.md

Current phase: Phase 3 — implementation complete (all milestones ✅) Branch strategy: feature branches → main (never committed directly to main) Last updated: 2026-07-04

---

Milestones

#	Milestone	Status
0	Repo, tooling, tracking	✅
1	Astro base config + token/style layer	✅
2	i18n layer (locales, slug map, dictionaries)	✅
3	Content schemas + EN/NO data	✅
4	UI atoms + core layout (Header, Footer, SEO)	✅
5	Home page sections	✅
6	Case-study + capability templates	✅
7	Remaining pages (about, experience, contact, cv, colophon, 404)	✅
8	Behaviour modules (theme, nav, observer, lightbox, form)	✅
9	SEO layer (meta, hreflang, JSON-LD, OG images, sitemap)	✅
10	Contact relay (.NET 9)	✅
11	Tests (vitest + playwright)	✅
12	Docker + CI	✅
13	Verification pass	✅

Verification results

• Build: 21 pages + 20 OG images + sitemap/robots, static, clean.
• Lint / types: ESLint + Prettier clean; tsc --noEmit clean; astro check wired.
• Unit tests (vitest): 23 passing — dictionary parity, slug-map bijection, content-schema validation, JSON-LD builders.
• E2E (playwright): 23 passing — both locales render, language switch maps page↔page (incl. JS-disabled), CV downloads resolve, contact form happy-path + honeypot, and axe a11y with zero serious/critical violations across 5 templates × 2 themes.
• Docker: both images build; relay /healthz → 200; site container serves EN + NO with CSP header and the /Linkedin 301 redirect.
• Contrast: all body/label text ≥ 4.5:1 in both themes (axe-verified; ink-faint and the light accent were raised to meet AA).

Environment

• Node v22.23.1 · pnpm 9.15.9 · .NET SDK 10.0.201 (relay targets net9.0) · Docker 29.3.1.

Deviations from spec (with rationale)

1. Content as typed TS modules + zod, not Astro content-collection entries. The nested {en,no} model serves the spec's "facts once, prose twice" goal better than parallel per-locale entries; same build-time zod validation and CI-testable getters.
2. Sitemap hand-rolled from the slug map (localised slugs don't fit @astrojs/sitemap's parallel-slug assumption; keeps one route contract).
3. OG images via resvg + hand-built SVG, not satori (satori's parser can't read the vendored variable fonts). Deterministic offline builds; fonts vendored in-repo.
4. Relay targets net9.0 built with the .NET 10 SDK (only SDK available); runs on the aspnet:9.0 runtime in Docker.
5. Site nginx config co-located in site/ (not deploy/) so the image build context stays site/. deploy/ holds the compose stacks + env template as specified.
6. Theme-init is an external script (/theme-init.js) so the CSP can use script-src 'self' with no inline hash.

Pre-launch checklist (carried from Phase 1/2 — owner action)

• Replace the /Linkedin redirect target in site/nginx.conf with the real LinkedIn profile URL before cutover (currently a placeholder slug).
• Screenshot capture pass on seeded demo data for JobTrack + InboxIntel; drop real captures into the placeholder: media refs (src/data/projects/*).
• Drop the real outdoor headshot into the portrait ref (replaces placeholder:portrait).
• ATS-safe CV rework (EN/NO) — replace the PDFs at site/public/cv/ (same URLs).
• Native-Norwegian review of all NO content (launch gate).
• Legacy WordPress URL inventory → confirm the 301/410 map in nginx.conf.
• Configure CI secrets (REGISTRY_USER, REGISTRY_TOKEN) + the relay .env on the server.

Open questions / blocked-on-user

• Are the JobTrack / InboxIntel repos public? If so, add repo links in src/data/projects/* (currently omitted to avoid dead links).