Wire de SSE-events uit /api/realtime/solo door naar de Zustand-store
zodat het Solo Paneel zonder refresh meebeweegt met DB-mutaties uit
welke bron dan ook (web, REST, MCP).
Migratie 20260427000216_extend_realtime_payload: voegt new-state
velden aan de pg_notify-payload toe (task_status, task_sort_order,
task_title, story_status, story_sort_order, story_title, story_code)
zodat de client geen extra fetch nodig heeft per event.
Store-uitbreiding (stores/solo-store.ts):
- pendingOps: Set<task-id> die optimistic-writes markeert; realtime
echos voor die ids worden onderdrukt zodat eigen UI-mutaties niet
twee keer toegepast worden of door een latere echo overschreven
- handleRealtimeEvent: dispatch op entity + op
- task UPDATE/INSERT: bestaande tasks krijgen status/title/sort_order
bijgewerkt; onbekende tasks worden genegeerd (story-context
ontbreekt — gebruiker ziet ze pas na refresh)
- task DELETE: verwijdert uit store
- story UPDATE: werkt story_title/story_code bij op alle child-tasks
in de store
- story DELETE: verwijdert alle child-tasks (cascade reflectie)
Unit-test: 7 scenario's (status update, pendingOps echo-suppression,
DELETE, story-rename cascade, story-delete cascade, unknown task
no-op).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Add notify_task_change() and notify_story_change() PL/pgSQL functions
plus AFTER INSERT/UPDATE/DELETE triggers on tasks and stories. Each
write emits a JSONB payload on the 'scrum4me_changes' channel with
op, entity, id, product_id, sprint_id, assignee_id and (for UPDATE)
the list of changed columns. Tasks resolve product/sprint/assignee
via their parent story so the SSE handler can filter without an extra
DB roundtrip.
The migration is a side-effect-only change (no Prisma model/schema
diff) so the Prisma Client and TypeScript types are unaffected.
Verified locally with a node-pg LISTEN client: both task and story
mutations produce the expected payload within milliseconds.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Een herbruikbaar script dat één milestone (PBI + stories + tasks) uit
docs/scrum4me-backlog.md leest en idempotent toevoegt aan de DB,
zonder bestaande data of sprint-toewijzingen te raken.
npm run db:insert-milestone -- M8 # standaard SCRUM4ME
npm run db:insert-milestone -- M8 --product CODE # ander product
npm run db:insert-milestone -- M8 --dry-run # niets schrijven
Parser-uitbreiding (parse-backlog.ts):
- M7 en M8 toegevoegd aan priority/goal/sprint_status maps
- Onbekende milestones krijgen defaults (priority 4, sprint COMPLETED,
goal = header-titel) zodat M9..M∞ vanzelf werken
- Optionele { strict: false } skipt de "≥ 8 milestones / ≥ 60 stories"
asserts; de bestaande seed-flow behoudt strict: true (default)
Idempotency:
- Pbi upsert by (product_id, code = milestone.key)
- Story upsert by (product_id, code = ST-XXX)
- Tasks alleen aangemaakt als de story op dit moment 0 tasks heeft
Sprint-toewijzing wordt expliciet niet meegenomen — stories landen in
de backlog en moeten via Sprint Planning in een sprint geplaatst worden.
Geverifieerd: M8 inserted (PBI + 6 stories + 6 tasks), tweede run laat
zien "PBI reused, 0 created" — geen duplicaten.
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* docs(ST-511): add backlog entry for entity codes feature
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix(ST-511): add createWithCodeRetry helper to handle P2002 race on auto codes
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix(ST-511): retry on auto-code unique conflict in story and pbi create
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix(ST-511): surface field errors for code and title in PBI dialog
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix(ST-511): read create-state errors in Story dialog fieldError
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* docs(ST-512): add backlog entry for REST API code/description/implementation_plan extensions; mark ST-511 done
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* feat(ST-512): extend REST API with code, description and implementation_plan
- GET /api/products returns code, description and definition_of_done
- GET /api/products/:id/next-story returns story.code and per-task code + implementation_plan
- GET /api/sprints/:id/tasks returns description, implementation_plan, story_code and derived per-task code
- POST /api/todos accepts and returns optional description (max 2000)
All changes are additive — existing clients ignore unknown keys.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* docs(ST-512): mark ST-512 as done
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* docs(ST-513): add backlog entry for API hardening for Claude Code
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* feat(ST-513): add task and story status mappers for API boundary
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* feat(ST-513): expose lowercase status on API and accept lowercase in PATCH /api/tasks
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* feat(ST-513): add metadata JSONB column to StoryLog
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* feat(ST-513): accept optional metadata in story log and switch validation errors to 422
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* feat(ST-513): add GET /api/health endpoint with optional db ping
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* feat(ST-513): add GET /api/products/:id/claude-context bundled endpoint
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* docs(ST-513): add docs/API.md and link from CLAUDE.md specs table
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* docs(ST-513): mark ST-513 as done
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix(ST-513): split 400 (malformed JSON) from 422 (validation), reject 'review'
Codex review on PR #2:
- P2.1: routes treated JSON parse failures as 422 instead of 400, breaking
the contract in docs/API.md. Replace `request.json().catch(() => null)`
with try/catch in 4 routes (tasks, reorder, todos, story-log) so a
malformed body returns 400 and only well-formed-but-invalid bodies
return 422.
- P2.2: PATCH /api/tasks/:id accepted `status: "review"`, but the sprint
task list UI does not render REVIEW (no label/color, the cycle helper
falls back to TO_DO). Reject `review` at the API until the sprint UI
is extended; document the subset in docs/API.md.
Tests in __tests__/api updated for the new contract (29 assertions:
zod-failures now expect 422, status payloads use lowercase API values,
sprint-tasks mocks include the new story relation).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds a Solo Demo PBI under the active sprint with 4 stories:
- 3 claimed by demo user, tasks covering TO_DO / IN_PROGRESS / DONE columns
- 1 unassigned story so the sheet demo shows a claimable item
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Todos can now exist without a product link. Changed relation from
Cascade to SetNull so deleting a product doesn't delete unlinked todos.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Aligns schema with the API and server action, both of which already
require product_id. Changes onDelete from SetNull to Cascade — deleting
a product now also removes its todos.
Run: npx prisma db push
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Schema: bio (160), bio_detail (2000) en avatar_data (bytea) op User
- POST /api/profile/avatar: validatie MIME-type + max 12 MB vóór verwerking,
Sharp resize naar max 700x700 (fit inside), output WebP q85, opgeslagen als bytea in Neon
- GET /api/profile/avatar: serveert avatar met Cache-Control private 1u
- updateProfileAction: slaat bio en bio_detail op via Server Action + Zod
- ProfileEditor client component: avatar preview, upload met client-side validatie,
bio-velden met tekenlimieten
- Settings page: profiel-sectie bovenaan, uitgeschakeld voor demo-gebruiker
- next.config: sharp als serverExternalPackage
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add ProductMember model (many-to-many User ↔ Product)
- Add productAccessFilter helper (owner OR member OR clause)
- Replace all ownership checks across actions and API routes
- Add addProductMemberAction / removeProductMemberAction / leaveProductAction
- Add TeamManager component in product settings (owner adds/removes Developers)
- Add LeaveProductButton in user settings (member leaves a product team)
- Regenerate Prisma Client after schema migration
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Seed maakt een actieve sprint aan voor de demo gebruiker met 3 stories.
Statusbalk aangepast naar h-14 zodat deze overeenkomt met de navigatiebalk.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Prisma schema: sqlite → postgresql provider met directUrl
- prisma.config.ts: directUrl toegevoegd
- seed.ts: dynamisch SQLite of PostgreSQL adapter op basis van DATABASE_URL
- ci.yml: preview deploy op PR, productie deploy op push naar main
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
