Seven stats fields (games_lost, win_streak_best, lifetime_score,
draw_one_wins, draw_three_wins, avg_time_seconds on both branches)
had no isolated test coverage in the merge test suite. Added boundary
tests for each, including a concrete arithmetic verification of the
weighted-average recomputation.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Both public APIs in solitaire_sync had no test coverage:
- win_rate(): None before any game, 100/50/0% cases
- AchievementRecord::locked(), unlock(), idempotency preserving earliest date
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Previously, same-week progress from the remote device was discarded
entirely — local counts were always preferred. Now each goal's count
is merged with max() so progress earned on any device is preserved.
Adds two regression tests covering same-week and newer-week cases.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- solitaire_server: Axum auth, sync push/pull, leaderboard, daily
challenge, account deletion, JWT middleware, rate limiting via
tower_governor, SQLite migrations, health endpoint
- solitaire_server: expose build_test_router (no rate limiting) so
integration tests work without a peer IP in oneshot requests
- solitaire_sync: SyncPayload, merge logic, shared API types
- solitaire_data: SyncProvider trait, LocalOnlyProvider,
SolitaireServerClient, auth_tokens keyring integration, blanket
Box<dyn SyncProvider> impl
- solitaire_data/settings: derive Default on SyncBackend (clippy fix)
- .sqlx/: offline query cache so server compiles without a live DB
- sqlx: removed non-existent "offline" feature flag
- keyring v2: fixed Entry::new() returning Result<Entry>
- sqlx 0.8: all SQLite TEXT columns wrapped in Option<T>
- Integration tests: max_connections(1) on in-memory pool so all
connections share the same schema
All 191 tests pass; cargo clippy -D warnings clean.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
funman300