UI Gap Implementation

UI gap implementation and verification

Current amendment persistence: 19-record QBrain readback confirms the current cap10 requirement amendment, stable IDs/native mappings and 28 changed criteria still unsatisfied. Original design provenance, case synchronization receipts, v2.7 before copies and execution history remain retained; current case/export reconciliation and exact-source execution are separate obligations. CAP15-LEGACY retains its historical ID with current10-player text. Documentation does not satisfy criteria or establish a passing game.

Historical implementation snapshot. Its original 2 through 25/service/provider/install scope and results below are preserved. Current rules v2.8 (31 October 2026) instead require 2 through 10 players, 25 observers unchanged, guest-only public play and the owner GitHub admin-only PWA. See the current release contract and completion plan; this old approval does not certify changed source or the current 9-game/54-participation/432-solve matrix.

This implementation realizes UI-DEC-01..08 and the linked QBrain records. The implementation snapshot is identified by the SHA256 manifest in implementation-manifest.json; the Git base remains3273c121de6517aec34503d8c769b0159e9e52df with preserved pre-existing local work. The final execution receipt distinguishes source, browser/port fixtures, real SQL and remaining operational dependencies.

Moderators can request changes to a suggestion, hold a nominated candidate and resume its review with a reason. Proposers alone can revise a changes-requested category and its three pairs. Both legacy pending and new pending_review records support ordinary independent review. Deferrals never publish, consume an appeal or change scores. Atomic content audits retain actor, reason, case version, prior/new status and prior/new proposal. Migration011 extends the status constraint and one-open unique index without rewriting historical evidence.

stateDiagram-v2
    pending --> changes_requested: independent request + reason (suggestion)
    changes_requested --> pending_review: proposer resubmits three valid pairs
    pending --> on_hold: independent hold + reason (nomination)
    on_hold --> pending_review: independent resume + reason + threshold check
    pending_review --> approved: independent review
    pending_review --> rejected: independent review
    rejected --> appealed: existing one-appeal policy
    approved --> approved: complete curation makes future content playable

Loading

Drafts and uncertain commands live only in tab memory. Internal Next links preserve them across ordinary navigation. A purpose-separated opaque session-generation value accompanies verified session reads/guest entry; a replacement session for the same profile also clears old drafts, commands and component state. This value is neither the raw cookie nor its credential-store hash and cannot authenticate a request. Cancel/Escape in the native profile dialog makes no write and restores its opener; controls inside the dialog are draft-only until Save. Sign-out, reload, identity change and authorization loss clear private state.

A shared lock prevents duplicate submissions. Uncertain retries retain the exact endpoint, method, serialized payload and command ID across route changes. Successful results stay scoped and locked until a current mounted consumer acknowledges them, so a completed private-room creation cannot lose its destination while its original page is unmounted. Content/admin/game/room/social denial invalidates older read generations; late bodies cannot restore private projections or another session's success. No raw server/network exception is rendered. Room GET reads drain the body and handle status/ownership before parsing successful JSON; room commands use controlled failure text. Malformed GET/POST sentinel regressions verify the room alert cannot display parser body content.

No-solver summaries now schedule automatic continuation three seconds after the authoritative end. Three seconds is an automatic-advance delay, not a guaranteed minimum summary display; an optional manual Continue may advance sooner. A delayed worker starts a full20-second initial answer stage at its current time. Duplicate/manual races create one next round, and the eighth round schedules no ninth. A catalog revision race returns retryable unavailability for system deadlines so a valid job is not acknowledged as obsolete. The automatic transition is audited as system:deadline.

Discovery links carry UX-only provenance. A current authoritative last-seat loss returns a non-member to a fresh first-page board (fresh=1), with visible unavailable/retry feedback if that read fails. Existing members and uncertain winning joins are preserved; the pending command is resolved canonically. Direct room/code/invitation links retain read-only observation. Navigation alone sends no readiness or Leave command.

Styles enforce wrapping, minimum inline sizing, visible focus, contrasting control borders and44x44primary actions. Browser measurements cover320x568,390x844,844x390, the320CSS-pixel/4x-device-scale equivalent of1280at400%zoom, and200%root-text enlargement. Score/stage live-region text and error alerts are exercised with focus unchanged. These are DOM/layout and headless Chromium assertions, not physical screen-reader, Safari/iPhone or OS-installation certification.

Repeatable checks and evidence boundaries

• npm test -- --maxWorkers=2 --testTimeout=60000 --hookTimeout=60000: unit/domain/application/contract tests, including command timing, independent roles, exact retries and session/result lifecycle.
• pwsh -NoProfile -File tests/integration/run-phase1.ps1: owned disposable PostgreSQL, all migrations, real locking/projections/atomic audit/idempotency, role denial, catalog race and the2..25service matrix. Nomination-count setup and the race-test queue selector are disclosed fixtures; their SQL command transactions are real.
• WORDPAIRS_BROWSER_PATH=/usr/bin/chromium npm exec playwright -- test --config playwright.phase6.config.ts: actual React pages and browser interaction. HTTP and persistence/time ports are isolated fixtures using real application/domain services. This suite proves UI behavior, not real Realtime delivery.
• npm run typecheck, npm run lint, npm run build: required compile/lint/production checks. The build runs after the browser server releases .next.

See implementation-receipt.json, evidence, the updated case execution sections, criterion audit and QBrain readback. The separate full private-push2..25UI run, real provider credentials and physical phone installation remain open execution dependencies; these tests do not close TEST-ROOMSYNC-139 or LIVE-07. No shared/live requirements are invented from fixtures, and no deployment, commit or push is implied.

Independent hostile validation approved the final234-file source inventory at99%accuracy/99%completeness. All seven findings and earlier rejected candidates remain in the append-only review archive. Final evidence is542unit/61browser/68SQL tests, plus successful lint, typecheck and production build. The reviewer independently reran9focused domain/timing/private-state tests. Approval is scoped to this UI extension and does not certify contest entry or outstanding real transport/device/provider obligations.