Add end-to-end (Playwright) coverage for core user flows

[chiptus]

This was generated by AI during triage.

Context

Captured from stale PR #10 ("comprehensive e2e test suite") so the coverage intent isn't lost. That PR can't be merged as-is (see below).

Goal

Extend the Playwright e2e suite beyond the current artists / auth / navigation specs to cover the core user flows (from PR #10):

• Voting (must-go / interested / won't-go)
• Groups (create / manage / join)
• Schedule / sets viewing + navigation (timeline & list; reveal levels)
• Festival & edition selection
• Admin dashboard / management
• Set detail page (voting, notes, social links)

Why PR #10 can't just be merged

PR #10 (Aug 2025) targets the pre-TanStack-migration app. Routing, the data layer, the Schedule (now reveal levels + festival timezone), and the src/api structure have all changed since, so its proposed specs would need to be rewritten against current selectors/flows rather than rebased. None of its proposed specs (voting/groups/schedule/festival-selection/admin/sets.spec.ts) ever landed on main (which only has artists/auth/navigation.spec.ts).

Open questions (triage before agent-ready)

• Scope: one large suite vs. slice per flow (recommend per-flow, incremental).
• Test data / setup: rely on the existing local-Supabase e2e harness (scripts/setup-test-env.sh); confirm the seed covers these flows.
• Relationship to Adopt boundary-level mocking (MSW / Supabase) for query-hook tests instead of mocking internal hooks #64 (unit-test boundary mocking): separate surface (e2e vs. unit) — no overlap, but both touch test infrastructure.

References

• Stale PR: feat(tests): add comprehensive e2e test suite #10
• Related: Adopt boundary-level mocking (MSW / Supabase) for query-hook tests instead of mocking internal hooks #64 (unit-test boundary mocking), tests/README.md, tests/e2e/*

[chiptus]

This was generated by AI during triage.

Triage: tracking epic — split into per-flow slices

Grilled during triage. The Playwright harness already exists and works (3 specs: artists / auth / navigation, plus TestHelpers and the local-Supabase seed), and tests/README.md already lists these same flows as "planned" — so this is executing a documented intent. Decision: do it as per-flow incremental slices, each its own independently-landable ticket, rather than one large flaky suite. This issue becomes the tracking epic (stays open until the slices land, like #27).

Slices (ordered by value)

• e2e: voting flow spec (must-go / interested / won't-go) #111 — voting flow (ready-for-agent) — must-go / interested / won't-go, the core loop
• groups flow — create / manage / join (incl. invite codes)
• schedule / reveal-level viewing — timeline + list, gated by reveal level
• admin dashboard / management
• set-detail page — voting, notes, social links

(Festival/edition selection is already partly covered by navigation.spec.)

Cross-cutting notes

• Seed coverage is the gating dependency for every slice: each needs deterministic seeded state (a voter with votable timed sets, a group with a known invite, an admin user, an edition at full reveal). Slices should extend supabase/seed.sql as needed rather than depend on prod-synced data.
• Harness docs are stale — tests/README.md still calls the app "Boom Voter" and cites base URL :5173 (the app runs on :8080). A light doc refresh should ride along with the first slice.
• Adopt boundary-level mocking (MSW / Supabase) for query-hook tests instead of mocking internal hooks #64 (unit-test boundary mocking) is a separate surface (unit vs e2e) — no overlap.

Close this epic once the slices above are done.

---

Generated by Claude Code

[chiptus]

This was generated by AI during triage.

Labeled ready-for-human — this is a tracking epic, not a single delegate-safe unit of work: continuing to split off and prioritize per-flow slices (groups, schedule, admin, set-detail) requires human judgment about ordering and scope, same as #27. #111 (voting flow) is already spun off and ready-for-agent.

---

Generated by Claude Code