Produce the original stack trace

[gajus]

Current Behavior

    at executeQuery (/Users/gajus/Developer/gajus/slonik/src/routines/executeQuery.ts:127:11)
    at query (/Users/gajus/Developer/gajus/slonik/src/connectionMethods/query.ts:17:28)
    at any (/Users/gajus/Developer/gajus/slonik/src/connectionMethods/any.ts:17:31)
    at Object.any (/Users/gajus/Developer/gajus/slonik/src/binders/bindPoolConnection.ts:27:17)
    at /Users/gajus/Developer/gajus/slonik/src/binders/bindPool.ts:24:40
    at createConnection (/Users/gajus/Developer/gajus/slonik/src/factories/createConnection.ts:169:20)
    at processTicksAndRejections (node:internal/process/task_queues:95:5)
    at async Object.any (/Users/gajus/Developer/gajus/slonik/src/binders/bindPool.ts:18:14)

Desired Behavior

    at createPool (src/factories/createPool.ts:57:9)
    at test/helpers/createIntegrationTests.ts:54:35

Motivation

It would make it a lot easier to debug captured errors.

Implementation

The TLDR of why this is happening is because we have async steps between the query invocation location and query execution, e.g.

1. connecting to the database
2. await for type overrides
3. interceptors

The fix would require that we capture the stack trace at the earliest possible and then override error stack trace before re-throwing it.