NaturalKey on self-aggregating aggregates (closes #4277)

[jeremydmiller]

Closes #4277.

Summary

Adds an end-to-end regression test to `Bug_4197_fetch_for_writing_natural_key.cs` matching the exact self-aggregating shape from the issue:

```csharp
public sealed record Bug4277SelfAggregate(Guid Id, [property: NaturalKey] Bug4197AggregateKey Key)
{
[NaturalKeySource]
public static Bug4277SelfAggregate Create(Bug4197AggregateCreatedEvent e)
=> new(e.Id, new Bug4197AggregateKey(e.Key));
}
```

Before the fix, calling `FetchForWriting<Bug4277SelfAggregate, Bug4197AggregateKey>(key)` returned a null aggregate because the natural-key lookup row was never written — the discovery scan in JasperFx.Events silently skipped the static `Create` method on the aggregate type.

Upstream fix

Companion PR: JasperFx/jasperfx#187, shipping as JasperFx.Events 1.29.1. That PR:

• Widens `discoverNaturalKey` to scan static methods on the aggregate type, not just instance methods.
• Teaches `buildExtractor` to call a static factory on `docType` that returns `docType` and read the natural-key property off the returned aggregate.
• Adds three JasperFx-side unit tests that exercise the three `[NaturalKeySource]` discovery pathways directly.

Dependency

This PR bumps `Directory.Packages.props` `JasperFx.Events` from `1.29.0` to `1.29.1`. CI on this branch will fail with `NU1102` until `JasperFx.Events 1.29.1` is published to NuGet.org via the upstream PR. Local verification performed against a packed dev build of `1.29.1`:

• `Bug_4197_fetch_for_writing_natural_key` — 3/3 pass (including the new test)
• Broader natural-key + aggregation sweep — 283/283 pass on net9.0, no regressions

🤖 Generated with Claude Code