Avoid redundant checks for validating custom SQL queries

[rydnr]

Information

Currently, the biggest bottleneck is the validation of the SQL queries, which is O(n). QueryJ could calculate a hash for each query which gets validated correctly, so that afterwards such hashes could be cached and used to avoid unnecessary database access.

Actions required

• Write the code to calculate a hash for each SQL query. The hash can be used as hashCode() implementation, surviving restarts.
• The hash is cached on disk if the validation succeeds.
• Before validating the query, the hash is checked and used to skip the validation.

Steps to reproduce

• There's no way to reliable obtain a value identifying each query uniquely.
• Consecutive executions take the same time.

Acceptance criteria

• The Sql API provides a method to obtain its hash.
• QueryJ runs much faster after the hashes have been computed and cached.