[FLINK-40108] Adding snapshot capability with split enumerator

[eskabetxe]

This pull request introduces a new connection management layer for the Flink JDBC connector, adding support for pluggable connection providers and improved connection configuration. The changes include new classes for connection options, a builder for those options, a minimal DataSource implementation for integration with connection pools, and interfaces for connection and statement handling. Additionally, HikariCP is added as a dependency (with SLF4J excluded), and a new async enumerator is provided for snapshot splitting.

Connection Management and Configuration:

• Added ConnectionOptions and ConnectionOptionsBuilder classes to encapsulate and build JDBC connection configuration, including timeouts and properties. [1] [2]
• Introduced ConnectionDataSource, a minimal DataSource implementation that delegates driver resolution, enabling proper classloader handling with connection pools like HikariCP.
• Added ConnectionException for consistent error handling around connection issues.
• Added ConnectionProvider interface, extending the base JDBC connection provider with methods for table/column metadata discovery and query building for snapshot splitters.

Statement and Result Handling:

• Introduced ResultSetMapper and StatementPreparer functional interfaces for mapping JDBC results and preparing statements, respectively. [1] [2]

Async Snapshot Splitting:

• Added AsyncSnapshotSplitterEnumerator, a background-threaded enumerator for computing and emitting snapshot splits asynchronously, improving scalability and responsiveness during snapshotting.

Dependency Management:

• Added HikariCP as a dependency in pom.xml, with SLF4J excluded to avoid conflicts.

[eskabetxe]

@RocMarshal @snuyanzin can you check please