fix(core): retry locked SQLite statements cooperatively

[cms19859230182-lang]

Issue for this PR

Closes #42120

Type of change

• Bug fix
• New feature
• Refactor / code improvement
• Documentation

What does this PR do?

Interrupted turns can take ~11s to settle because locked SQLite statements spin in the native busy handler (busy_timeout = 5000), freezing the event loop while the interrupted step waits for the shared transaction permit inside its settlement mask.

• Statements that hit SQLITE_BUSY/LOCKED now fail fast and retry cooperatively in Sqlite.makeConnection (bounded ~1.6s, only LockTimeoutError, never re-runs a transaction).
• busy_timeout is set to 0 (node driver timeout too), so waits happen in Effect fibers and the event loop stays schedulable.

How did you verify your code works?

• New tests: a locked write retries and succeeds after the lock releases; an exhausted retry still reports LockTimeoutError; interrupted settlement queues behind an unrelated transaction (deterministic Deferred gates, no sleeps).
• bun test test/session-runner.test.ts test/database-drizzle.test.ts from packages/core: 158 passed.
• bun typecheck: passed.

Screenshots / recordings

Not a UI change.

Checklist

• I have tested my changes locally
• I have not included unrelated changes in this PR

[github-actions]

Thanks for your contribution!

This PR doesn't have a linked issue. All PRs must reference an existing issue.

Please:

1. Open an issue describing the bug/feature (if one doesn't exist)
2. Add Fixes #<number> or Closes #<number> to this PR description

See CONTRIBUTING.md for details.

[cms19859230182-lang]

FYI: this PR links #42120 (Closes #42120 in the description, cross-referenced on the issue timeline). The needs:issue label looks like a stale cross-reference index — happy to fix if I am missing something.