GitHub becomes the primary repository host; the site keeps its VPS + nginx + Dockge host, the three-job pipeline and the same concurrency, so what moves is the CI host and the write-path target. .github/workflows/{deploy,deploy-infra}.yml are ports of the Gitea files with permissions: contents: read on every job. .gitea/workflows/ stays until the first green GitHub run (D9); its infra test step carries the new template variable names because the template it validates is the one this repository now ships.
The write-path values are renamed GITEA_TOKEN / GITEA_CONTENTS_API -> CONTENT_API_TOKEN / CONTENT_API_BASE and the upstream header becomes Bearer — in nginx.conf, in the compose envsubst filter and in the dev proxy, so dev and production send the identical header (D4). draft.ts learns GitHub's answer to a stale blob sha (409) while keeping the 422 "does not match" and 412 legs; smoke-draft.ts check 14 drives a stubbed proxy that answers 409 with GitHub's body and asserts the conflict refreshes the base, keeps the draft and never stops the save sequence.
Both workflows read the stack directory and the compose service / container name from the STACK_DIR and CONTAINER_NAME repository variables instead of hard-coding them, and each job refuses a missing, relative, trailing-slash or compose-disagreeing value in its first step, before any artifact download or rsync reaches the VPS (D15).
Verified: tsc -b and oxlint pass with no new warnings; validate:content, smoke-draft (14 checks) and smoke-generated-ids (7 checks) pass; both workflows parse, every run/script block passes bash -n with the variables substituted, and the guard refuses every bad value against the real compose.yaml. Owner-side variables, secrets, the write token and the VPS config rewrite stay open in spec-v6 §6.
Edit-mode document scope comes from the page manifests (src/i18n/edit-scope.ts): an entity's structure document and its two text documents are read as one unit, the category labels only where a project renders. A route change fetches only the still-missing documents and adopts them through DraftStore.adoptBase, which refuses a document the draft has already changed. A bundle-seeded baseline is never PUT over the branch: saveDraft reads the branch first and reports a conflict. Every manifest now claims profile.name, which the footer renders on every route.
scripts/generate-content-types.ts writes the gitignored src/types/content-ids.ts (id unions + project category values) from the four structure documents; wired as generate:ids / predev / prebuild / prepare. src/types/index.ts re-exports them type-only and keeps only the code-owned unions.
Edit mode now writes JSON only: the v4 union leg is deleted (saveIdUnions, diffIdsWithUnions, applyUnionChanges, readUnionBlock, the EditBar union card, edit.union.* labels). Validator rule 7 requires a label for every category a project uses, in both ui/projects.json documents.
Tests: smoke-draft.ts 11/11 (rename re-sluging + category-label rule replace checks 9-12); new scripts/smoke-generated-ids.ts 7/7. Verified: pnpm build (chunk names unchanged, no runtime reference to the generated module), lint with no new warnings, validate:content, resume HTML byte-identical before/after with identical documents.
Docs: spec-v5-generated-content-ids.md added; spec-v4-live-edit.md §3.1 marked superseded + §12 row 23; AGENTS.md and README.md updated.
Deleting a record in edit mode removed the structure row and its text but left the id union member in src/types/index.ts, so Record<SkillGroupId, SkillGroupText> rejected the shortened text document (TS2741) and pnpm build failed. draft.ts: readUnionBlock reads a union as its declaration span (CRLF/LF safe, never round-trips), diffIdsWithUnions reports missing and stale members against committedStructures(), applyUnionChanges rewrites only that member list in one PUT; newIdsInCommitted/patchUnionSource removed. types/index.ts drops logistics-technology. Verified: smoke-draft 12/12 (new checks 9-12), validate:content, lint (no new warnings), build.
Repo side of spec-v4-deploy-infra: deploy-infra.yml (paths-filtered nginx redeploy, test-before-touch, health probe); compose.yaml rename + secrets/site_config entrypoint override; site identity externalized - SITE_URL/SITE_NAME read from env (site.ts, vite site-identity plugin, index.html tokens, generated robots.txt); nginx.conf/vite dev proxy read GITEA_CONTENTS_API; .env.example; AGENTS.md + specs updated. Rev.7: site identity delivered as Gitea Actions variables (vars. context, public values) instead of secrets; upload-artifact unified to v7 across jobs.
A first tag commit on a text document that does not hold the array yet
(an older target branch, e.g. profile.currentInterests still living in
the structure document) built an object {"0": …}; the join then overrode
the structure array with it and items.map crashed on the re-render.
setPath now creates [] for [n] segments under an undefined root.
smoke-draft check 8 reproduces the old-baseline tag edit (8/8 pass).
Row 18: ui.about.subtitle/experience.tldrText/philosophy.quotes -> profile.subtitle/.tldr/.philosophyQuotes in content.{vi,en}.json; call sites repointed with EditableText wrap; SectionHeading subtitle prop becomes ReactNode; edit.fields labels added. Row 19: experience.company/.duration, projects.role/.metrics, profile.currentInterests/.roles, philosophy keywords move to text documents; VI carries localized dates/company labels/metrics, en overrides with originals, proper nouns and term lists identical (en omits); draft templates, JsonPanel hints, loader comments, resume TL;DR follow the new shapes. Verified: build, validate:content, lint (no new warnings), smoke-draft 7/7, per-language SSR marker smoke on affected routes, build:resume-all.
?edit entry with Basic-auth probe; DraftStore copy-on-write over live
branch documents (?ref GET / branch PUT, last_commit_sha optimistic
concurrency, three-way rebase on 412); EditableText inline editors;
JsonPanel with templates and orphan-free delete; EditBar sequential-PUT
save with conflict actions and merge hint on non-master branches.
scripts/smoke-draft.ts regression check.
Phase 1 of spec-v4-live-edit: data/i18n foundation, no visible feature.
- src/data/{profile,skills,experience,projects}.json: language-neutral
structure documents with thin typed loaders; all prose moved to
src/i18n/locales/{vi,en}/content.json; chrome strings to ui.json
- src/i18n/: dict-types, cross-document validator (runs at module load and
via pnpm validate:content), loader joining structure+text once (UI/CONTENT),
I18nProvider/useI18n mounted in App under LanguageProvider
- all 22 consumers read joined records via useI18n(); field.ts and
strings.ts deleted; *En field pairs removed from the domain types
- hero tagline renders as plain-text segments (D9); philosophy quotes as
paragraph arrays; footer display name now localizes per language (D6)
- resume generator consumes UI/CONTENT; generated resume HTML verified
byte-identical (EN+VI) before/after the migration
- AGENTS.md and v3/v3-i18n + v4 specs updated; deploy.yml concurrency group
index.html advertised nothing: no title, no description, no Open Graph, no canonical. The head now carries both theme-color values, an OG block with an absolute 1200x1600 image, a Twitter card and a Person JSON-LD block with sameAs and a PostalAddress. src/data/site.ts holds the canonical origin; generate-sitemap.ts writes dist/sitemap.xml from navPaths plus the project list (10 urls) and fails the build if dist/index.html no longer contains siteUrl, because an HTML head cannot import TypeScript and the origin is therefore written twice by hand. Chained into pnpm build so dist/ is complete before the deploy workflow uploads it.
profile-photo.webp~ (1.2 MB) was a stray editor backup sitting in public/ and served as a real asset; icons.svg had no reference in any file; the components barrel exported 14 names with 0 importers. ProjectSection.number (16 entries), FeaturedAchievement with ExperienceEntry.achievements, and Profile.resumeUrl were never rendered by any surface. fadeInUp/.animate-in and the download-resume selectors had no consumers - motion owns entrance animation. Four unreferenced string keys removed.
Timeline structure
- Spine drawn per node (.timeline-entry::before) instead of one line on .timeline, so it ends at
the TLDR diamond; marker and spine centres coincide at x=6 (gutter 36 desktop / 28 mobile)
- TLDR terminal node: markup in ExperiencePage + Experience section, copy in i18n/strings.ts
(experience.tldr / experience.tldrText), CSS .timeline-tldr* (stub, accent link, hollow
diamond, chip, text)
- ExperienceEntry gains variant?: 'current' | 'gap'; the class is composed in JSX rather than
hard-coding entry ids
Node decoration
- current node: TLDR-family panel (--bg-surface, 1px --border, radius-lg, whisper shadow) with a
2px accent left edge, real entry padding (16/18/20, mobile 14/16/18) and a pulsing marker;
mobile pulse capped at 6px spread so the ring never touches the panel border
- gap node: dashed spine (repeating-linear-gradient), 10px muted marker, no divider,
italic-muted title and company
Data
- New relax-gap entry (Aug 2026 – Oct 2026) between open-to-work and lazada-senior-swe; unique
entry ids; Nexts and SAP bullets aligned across VI/EN
Typography
- Add --measure-page (680px); .prose-wide / .lede-wide; .page-body 700 -> 680; philosophy grid
3 columns, 2 at <=1024px
Cleanup
- Achievements dropped from the SPA timeline and their dead .achievement* rules deleted
(.achievement-metric kept: projects pages still use it)
Resume generator
- "${exp.achievements?.map(...).join('')}" short-circuited the whole chain and printed the
literal "undefined" in every PDF once an entry had no achievements; now (exp.achievements ?? [])
Docs: spec-ui-experience.md, spec-v3-typography.md, AGENTS.md
Verified: tsc -b clean; vite build green; oxlint 11 errors / 9 warnings (HEAD baseline was 15/9 —
the four that disappeared were the achievements markup findings); pnpm build:resume-all renders
both PDFs with the new ordering and no "undefined".
- Add version stamp footer 'Last updated: YYYY-MM-DD' to PDF
- Implement bilingual support (English + Vietnamese) for resume PDF
- Add Language type and *En optional fields to types
- Add Vietnamese translations to profile, experience, projects data
- resolveField utility for language-aware field resolution
- Language-specific section labels (Experience/Kinh nghiệm, etc.)
- --lang CLI arg: 'en' (default) or 'vi'
- Output: resume.pdf (EN), resume-vi.pdf (VI)
- Update CI workflow to generate and deploy both PDF versions
- Add build:resume-all npm script
- Update UI components (Hero, DownloadResume) for language selection
- Add spec-v2-resume-future.md for deferred enhancements
- Refactor spec-v2-resume-playwright.md (condense implementation notes)