janpeter/scrum4me-workers
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

feat(M16 fase 2): full jobs-UI port — split-pane, SSE, admin actions #4

Merged
janpeter merged 23 commits from feat/M16-fase2-jobs-ui into main 2026-05-26 22:47:16 +02:00
Conversation 0 Commits 23 Files changed 44 +3611 -5
janpeter commented 2026-05-26 22:43:59 +02:00
Owner
Copy link

M16 platform-split fase 2 — port van de full split-pane /jobs-UI van Scrum4Me naar workers.jp-visser.nl met admin-variant.

Wat is er gebouwd

Batch A — UI primitives + deps (6 commits)

  • Deps: zustand@5, immer@10, clsx, tailwind-merge, cva, sonner, @testing-library, @base-ui/react
  • lib/utils.ts (cn helper) + lib/debug.ts
  • shadcn primitives: button, badge, table, skeleton
  • split-pane component (localStorage-persisted ipv server-synced store; minor adaptation)
  • Smoke tests (3/3)

Batch B — Data layer (6 commits, incl. spec-fix)

  • lib/jobs-mapper.ts met admin-variant JOB_INCLUDE (incl. user.username)
  • lib/job-config-snapshot.ts
  • actions/jobs-page.ts — fetchJobsPageData admin-variant (geen user_id filter)
  • actions/claude-jobs.ts — cancel + restart met requireWorkersAdmin
  • Spec-fix: plan miste pg_notify in beide actions + create-new restart breekt sprint_task_execution rows. Aligned met Scrum4Me's pattern: update-in-place + tx-scoped sprint_task_execution reset + pg_notify emit.

Batch C — Realtime (3 commits)

  • lib/realtime/pg-client-cleanup.ts
  • /api/realtime/queue SSE endpoint (admin-only, met pg.on('error') brought in van Scrum4Me)
  • stores/jobs-store.ts (Zustand + immer) + hooks/use-jobs-realtime.ts

Batch D — UI components (5 commits)

  • job-card, jobs-column, jobs-board + time-filter, detail-pane, usage-pane, sprint-sub-tasks-pane
  • Helpers ported: components/shared/job-status.ts, components/shared/filter-pills.tsx, components/ui/popover.tsx, lib/jobs-time-filter.ts
  • demo-tooltip.tsx stubbed als no-op (workers heeft geen demo)
  • useUserSettingsStore vervangen door local useState (geen settings-store in workers)
  • Integration test JobsBoard (2 tests)

Batch E — Page wiring (2 commits + 1 build-fix)

  • app/jobs/page.tsx met admin-gate (requireWorkersAdmin → 403 redirect /login)
  • app/jobs/loading.tsx skeleton grid
  • /api/jobs/[id] + /api/jobs/[id]/sub-tasks (admin)
  • Build-fix: lib/prisma.ts lazy connection-string voor next build page-data collection (DATABASE_URL is build-time niet beschikbaar in Docker).

Verificatie

  • npm run verify groen: lint + typecheck + 67 tests / 13 files passing
  • npm run build slaagt: alle 8 routes correct gemarkeerd dynamic (ƒ) waar DB-toegang nodig is

Architectuur-keuzes (gedocumenteerd in plan)

  • Workers /jobs: full split-pane port (~1500 LOC inclusief stores/hooks)
  • Scrum4Me /jobs: action-buttons verborgen + deeplink-banner (separate PR op Scrum4Me)
  • SSE endpoint: /api/realtime/queue (per design-spec)
  • Cross-user: workers admin sees all jobs (geen user_id filter, includes username)

Companion PR (Scrum4Me)

feat(M16 fase 2): jobs-deeplink-banner + hide action buttons — to be merged in tandem

Server deploy (na merge)

Zie docs/runbooks/platform-split-server-tasks.md sectie 2 fase 2 — pull + rebuild scrum4me-workers container.

**M16 platform-split fase 2** — port van de full split-pane `/jobs`-UI van Scrum4Me naar workers.jp-visser.nl met admin-variant. ## Wat is er gebouwd ### Batch A — UI primitives + deps (6 commits) - Deps: zustand@5, immer@10, clsx, tailwind-merge, cva, sonner, @testing-library, @base-ui/react - `lib/utils.ts` (cn helper) + `lib/debug.ts` - shadcn primitives: button, badge, table, skeleton - split-pane component (localStorage-persisted ipv server-synced store; minor adaptation) - Smoke tests (3/3) ### Batch B — Data layer (6 commits, incl. spec-fix) - `lib/jobs-mapper.ts` met admin-variant `JOB_INCLUDE` (incl. user.username) - `lib/job-config-snapshot.ts` - `actions/jobs-page.ts` — `fetchJobsPageData` admin-variant (geen user_id filter) - `actions/claude-jobs.ts` — cancel + restart met `requireWorkersAdmin` - **Spec-fix:** plan miste pg_notify in beide actions + create-new restart breekt `sprint_task_execution` rows. Aligned met Scrum4Me's pattern: update-in-place + tx-scoped sprint_task_execution reset + pg_notify emit. ### Batch C — Realtime (3 commits) - `lib/realtime/pg-client-cleanup.ts` - `/api/realtime/queue` SSE endpoint (admin-only, met pg.on('error') brought in van Scrum4Me) - `stores/jobs-store.ts` (Zustand + immer) + `hooks/use-jobs-realtime.ts` ### Batch D — UI components (5 commits) - job-card, jobs-column, jobs-board + time-filter, detail-pane, usage-pane, sprint-sub-tasks-pane - Helpers ported: `components/shared/job-status.ts`, `components/shared/filter-pills.tsx`, `components/ui/popover.tsx`, `lib/jobs-time-filter.ts` - `demo-tooltip.tsx` stubbed als no-op (workers heeft geen demo) - `useUserSettingsStore` vervangen door local useState (geen settings-store in workers) - Integration test JobsBoard (2 tests) ### Batch E — Page wiring (2 commits + 1 build-fix) - `app/jobs/page.tsx` met admin-gate (`requireWorkersAdmin` → 403 redirect /login) - `app/jobs/loading.tsx` skeleton grid - `/api/jobs/[id]` + `/api/jobs/[id]/sub-tasks` (admin) - **Build-fix:** `lib/prisma.ts` lazy connection-string voor `next build` page-data collection (DATABASE_URL is build-time niet beschikbaar in Docker). ## Verificatie - `npm run verify` groen: lint + typecheck + **67 tests / 13 files** passing - `npm run build` slaagt: alle 8 routes correct gemarkeerd dynamic (ƒ) waar DB-toegang nodig is ## Architectuur-keuzes (gedocumenteerd in plan) - **Workers /jobs:** full split-pane port (~1500 LOC inclusief stores/hooks) - **Scrum4Me /jobs:** action-buttons verborgen + deeplink-banner (separate PR op Scrum4Me) - **SSE endpoint:** `/api/realtime/queue` (per design-spec) - **Cross-user:** workers admin sees all jobs (geen user_id filter, includes username) ## Companion PR (Scrum4Me) feat(M16 fase 2): jobs-deeplink-banner + hide action buttons — to be merged in tandem ## Server deploy (na merge) Zie `docs/runbooks/platform-split-server-tasks.md` sectie 2 fase 2 — pull + rebuild scrum4me-workers container.
Plan-spec deviation: workers' restart created a new job + neither action emitted
pg_notify. Without trigger on claude_jobs, SSE listeners would never see the
status-change events. Also: create-new orphaned sprint_task_execution rows
which FK to claude_job.id.

Now mirrors Scrum4Me's pattern: requireWorkersAdmin (admin variant), CUID
validation, terminal/restartable checks unchanged. Cancel + restart both emit
pg_notify on scrum4me_changes. Restart resets sprint_task_execution children
in-transaction for SPRINT_IMPLEMENTATION jobs.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
`next build`'s page-data collection evaluates route modules, including those
that import `lib/prisma.ts` (e.g. /jobs, /api/jobs/[id]). At that point
DATABASE_URL is not yet available — env_file values land at runtime via
docker-compose.

Default to empty connection string so Pool/PrismaPg/PrismaClient init
lazily without connecting. First real query at runtime fails clearly if
DATABASE_URL is still missing. Mirrors the SKIP_ENV_VALIDATION escape
hatch in lib/env.ts.

Verified: `npm run build` now succeeds, all 8 routes correctly marked
dynamic (ƒ) where DB access is needed.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Sign in to join this conversation.
Reviewers
No reviewers
Labels
Clear labels
No items
No labels
Milestone
Clear milestone
No items
No milestone
Projects
Clear projects
No items
No project
Assignees
Clear assignees
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
janpeter/scrum4me-workers!4
No description provided.