fix(submit): aggregate device time totals

[IvGolovach]

Summary

• Derive submissions.totalActiveTimeMs from persisted daily_breakdown.active_time_ms rows across all submitted devices.
• Preserve per-device replace semantics so resubmitting one device updates that device's row without double-counting prior active time.
• Keep session-shape metrics (longestContinuousMs, maxConcurrentSessions, sessionCount) tied to incoming timeMetrics because they cannot be safely recomputed from daily active-time buckets.
• Add focused submit route regressions for multi-device active-time sums, same-device replacement, and submissions without incoming timeMetrics.

Why

The submit path already recalculates token and cost totals from all persisted daily rows, but totalActiveTimeMs was still copied from the latest incoming payload. With multiple devices, all-time profile and leaderboard time totals could represent only the latest submitting device instead of the user's aggregate submitted history. This PR makes active-time totals follow the same persisted-row aggregate model as token totals.

Diff scope

• packages/frontend/src/app/api/submit/route.ts
• packages/frontend/__tests__/api/submitAuth.test.ts

Branch integrity

• Base branch: main
• Validated base SHA: c21f0f533205ad63e9154666c294bb789d2415c9
• Head branch: IvGolovach:codex/submit-device-time-metrics-20260528
• Head SHA: fdd4e0f191b434e47601e089d8d59d3d2fa00b70
• Ahead/behind vs fetched origin/main: 1 ahead / 0 behind
• Commit: fdd4e0f191b434e47601e089d8d59d3d2fa00b70 fix(submit): aggregate device time totals

Validation

• bun run --cwd packages/frontend test __tests__/api/submitAuth.test.ts __tests__/api/submit.test.ts: PASS, 60 tests passed.
• bun run --cwd packages/frontend lint -- src/app/api/submit/route.ts __tests__/api/submitAuth.test.ts: PASS, no output.
• git diff --check origin/main...HEAD: PASS, no output.
• git status --short --untracked-files=all: PASS, clean worktree.

Runtime safety

• No new external services, queues, timers, or background jobs.
• The submit transaction already recalculates persisted totals after daily rows are inserted or updated; this PR extends that aggregate query to active_time_ms.
• Same-device replacement still updates the existing daily row before totals are recomputed.
• No invariant regression introduced.

Migration notes

Not applicable - no database migration or schema change.

Rollback plan

Rollback: revert this PR. DB downgrade: not applicable. Data repair: not applicable. Operational caveats: reverting restores latest-payload totalActiveTimeMs behavior.

Known residual risks

• longestContinuousMs, maxConcurrentSessions, and sessionCount remain sourced from incoming timeMetrics when present because daily active-time buckets do not contain enough information to recompute those session-shape metrics correctly.
• Remote CI should provide the final full-suite proof for the PR SHA.

---

Summary by cubic

Aggregates all-time active time from persisted daily device rows in the api/submit route so multi-device totals are correct. Preserves same-device replacement and keeps session metrics from incoming timeMetrics.

• Bug Fixes
  • Compute submissions.totalActiveTimeMs from SUM of daily_breakdown.active_time_ms across all submitted devices.
  • Recompute totals after same-device row replacement to avoid double-counting; session-shape metrics (longestContinuousMs, maxConcurrentSessions, sessionCount) use incoming timeMetrics when provided.

^{Written for commit 7d06744. Summary will update on new commits.}

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
tokscale	Ready	Preview, Comment	May 30, 2026 1:31pm

[cubic-dev-ai]

No issues found across 2 files

_{Re-trigger cubic}