fix: use configured SDK logger

[DYSfu]

Summary:

• Wire SDKConfig.logger into the Address, Blockchain, Mempool, and Transaction wrappers.
• Add request-failure coverage for the configured logger path.
• Remove the stray describe.only so the full Jest suite runs.

Validation:

• npm ci --legacy-peer-deps
• npx jest src/__tests__/logger.test.ts --runInBand -> 4 tests passed
• npm run build:main
• npm test -- --runInBand -> 26 tests passed
• git diff --check

Related:

• BGL (Bitgesell) Bounty/Improvement Program ($50000 budget) bitgesell#81

[MyTH-zyxeon]

Review assist for maintainers:

This looks like a useful SDK correctness/testability fix: the Address, Blockchain, Mempool, and Transactions wrappers now use SDKConfig.logger instead of falling back to a no-op instance field, and the new Jest coverage mocks axios failures across all four wrappers. Removing describe.only from address.test.ts also restores full-suite visibility.

Suggested merge checks:

• Run npm ci --legacy-peer-deps, the focused logger Jest test, npm run build:main, and the full Jest suite from the PR body.
• Confirm the SDKConfig.logger type matches the value passed by the wrappers on failures. If the catch path passes an Error object, either type the logger accordingly or normalize to a string/message before calling it.
• Verify the logger tests remain fully mocked and do not hit the public Bitaps API or require live RPC credentials.
• Check overlap with the separate full-suite/describe.only cleanup PRs so maintainers merge in an order that avoids duplicate or stale diffs.

I do not see wallet/key, signing, broadcast, live RPC, or live-chain behavior changes in this patch as written; the acceptance criterion is that configured logger callbacks are actually invoked on request failures and the whole SDK test suite is no longer accidentally focused.