prisma-driver-adapter-implementation

Implement SqlDriverAdapterFactory with a connect() method returning a SqlDriverAdapter that provides queryRaw, executeRaw, executeScript, startTransaction, and dispose. Match the exact @prisma/driver-adapter-utils version of your target Prisma release and follow the transaction and error-mapping contracts.

startTransaction must acquire one dedicated connection, begin the transaction, apply the isolation level, and return a Transaction bound to that connection. The commit() and rollback() methods are lifecycle cleanup hooks that release the connection once; Prisma issues the actual SQL COMMIT or ROLLBACK through executeRaw.

Yes, through optional createSavepoint, rollbackToSavepoint, and releaseSavepoint methods on the Transaction object. Implement them only where the provider supports savepoints, validate identifiers, and never track transaction depth on the shared adapter since parallel transactions make global depth incorrect.

P2039 surfaces when an unmapped driver error reaches Prisma. Preserve originalCode and originalMessage in your DriverAdapterError conversion so P2039 carries useful details, map known conditions to structured MappedError kinds, and rethrow genuinely unexpected non-driver errors instead of fabricating GenericJs ids.

A shadow database is an isolated database used by Prisma Migrate, created via connectToShadowDb() on SqlMigrationAwareDriverAdapterFactory. It must use cryptographically unique quoted names, never point at the primary database, and be dropped during disposal or failure cleanup.