feat(transactions): add single-node transaction support with session-aware reads, commits, aborts, and transaction metrics

readme.md

@@ -290,10 +290,12 @@ await client.connect();
 
 TLS is available for TCP listeners. `getConnectionUri()` includes `?tls=true` when TLS is enabled; pass the trusted CA to the MongoDB driver with `tlsCAFile`, `ca`, or `secureContext`.
 
-Authentication verifies SCRAM credentials, denies unauthenticated commands, and enforces command-level built-in roles for supported operations.
+Authentication verifies SCRAM credentials, denies unauthenticated commands, and enforces command-level built-in roles for supported operations. `connectionStatus` reports the authenticated users and roles for the current socket.
 
 Supported built-in role names are `root`, `read`, `readWrite`, `dbAdmin`, `userAdmin`, `clusterMonitor`, plus `readAnyDatabase`, `readWriteAnyDatabase`, `dbAdminAnyDatabase`, and `userAdminAnyDatabase`. When `usersPath` is set, SmartDB persists SCRAM credential material atomically and does not store plaintext passwords.
 
+Single-node transactions are supported through official MongoDB driver sessions. Writes with `startTransaction` and `autocommit: false` are buffered per logical session, reads inside the transaction see the buffered overlay, `commitTransaction` applies the write set with conflict checks, and `abortTransaction` discards it.
+
 Basic user management commands are available for authenticated users with `root` or `userAdmin` privileges:
 
 ```typescript
@@ -317,7 +319,7 @@ await client.db('admin').command({ usersInfo: 'reader' });
 | `port` | `number` | Configured port (TCP mode) |
 | `host` | `string` | Configured host (TCP mode) |
 | `socketPath` | `string \| undefined` | Socket path (socket mode) |
-| `getMetrics()` | `Promise<ISmartDbMetrics>` | Server metrics (db/collection counts, uptime) |
+| `getMetrics()` | `Promise<ISmartDbMetrics>` | Server metrics (db/collection counts, sessions, transactions, auth, uptime) |
 | `getOpLog(params?)` | `Promise<IOpLogResult>` | Query oplog entries with optional filters |
 | `getOpLogStats()` | `Promise<IOpLogStats>` | Aggregate oplog statistics |
 | `revertToSeq(seq, dryRun?)` | `Promise<IRevertResult>` | Revert to a specific oplog sequence |
@@ -531,7 +533,7 @@ const names = await collection.distinct('name');
 | **Aggregation** | `aggregate`, `count`, `distinct` |
 | **Indexes** | `createIndexes`, `dropIndexes`, `listIndexes` |
 | **Sessions** | `startSession`, `endSessions` |
-| **Transactions** | `commitTransaction`, `abortTransaction` |
+| **Transactions** | `startTransaction`, `commitTransaction`, `abortTransaction` through driver sessions |
 | **Admin** | `ping`, `listDatabases`, `listCollections`, `drop`, `dropDatabase`, `create`, `serverStatus`, `buildInfo`, `dbStats`, `collStats`, `connectionStatus`, `currentOp`, `renameCollection` |
 
 Compatible with wire protocol versions 0–21 (driver versions 3.6 through 7.0).
@@ -540,7 +542,7 @@ Compatible with wire protocol versions 0–21 (driver versions 3.6 through 7.0).
 
 ## Rust Crate Architecture 🦀
 
-The Rust engine is organized as a Cargo workspace with 8 focused crates:
+The Rust engine is organized as a Cargo workspace with 9 focused crates:
 
 | Crate | Purpose |
 |---|---|
@@ -551,6 +553,7 @@ The Rust engine is organized as a Cargo workspace with 8 focused crates:
 | `rustdb-storage` | Storage backends (memory, file), OpLog with point-in-time replay |
 | `rustdb-index` | B-tree/hash indexes, query planner (IXSCAN/COLLSCAN) |
 | `rustdb-txn` | Transaction + session management with snapshot isolation |
+| `rustdb-auth` | SCRAM-SHA-256 credential handling, user metadata persistence, RBAC checks |
 | `rustdb-commands` | 40+ command handlers wiring everything together |
 
 Cross-compiled for `linux_amd64` and `linux_arm64` via [@git.zone/tsrust](https://www.npmjs.com/package/@git.zone/tsrust).
@@ -563,6 +566,7 @@ The Bitcask-style file storage engine includes several reliability features:
 - **CRC32 checksums** — every record is integrity-checked on read
 - **Automatic compaction** — dead records are reclaimed when they exceed 50% of file size, runs on startup and after every write
 - **Hint file staleness detection** — the hint file records the data file size at write time; if data.rdb changed since (e.g. crash after a delete), the engine falls back to a full scan to ensure tombstones are not lost
+- **Torn-tail repair** — startup scans `data.rdb` to the last valid record, truncates invalid trailing bytes, and preserves all verified records after interrupted writes
 - **Stale socket cleanup** — orphaned `/tmp/smartdb-*.sock` files from crashed instances are automatically cleaned up on startup
 
 ### Data Integrity CLI 🔍