Correct which discovered Chronicle artifacts withChronicle picks up

Documentation/chronicle/add-event-sourcing.md

@@ -89,7 +89,7 @@ const app = await builder.build();
 await app.run({ port: Number(process.env.PORT ?? 3000) });
 ```
 
-Importing `@cratis/arc.chronicle` adds `withChronicle` to the builder. Call it before or after `discover`: the integration records each discovered event type and projection, including those discovered earlier, and hands them to Chronicle when it connects. The application connects when a command or query first needs Chronicle, and Chronicle creates the `MyArcApp` event store then. Events and read models go to the namespace of the request's tenant; this application resolves no tenant, so they go to the `Default` namespace.
+Importing `@cratis/arc.chronicle` adds `withChronicle` to the builder. Call it before `discover`: the integration records each discovered event type and projection and hands them to Chronicle when it connects. Classes with Arc metadata are recorded either way, but a Chronicle-only class discovered before `withChronicle` is dropped unless `activateArtifactsInScopes` is on. The application connects when a command or query first needs Chronicle, and Chronicle creates the `MyArcApp` event store then. Events and read models go to the namespace of the request's tenant; this application resolves no tenant, so they go to the `Default` namespace.
 
 Call `withChronicle` before `add()` for Chronicle-only artifacts: without an Arc decorator, `add()` rejects them until Chronicle is registered. `chronicle://localhost:35000` without credentials uses the SDK's development client and accepts the kernel's self-signed certificate. That fits a local kernel only. [Registration options](registration-options.md) shows how to read the connection from `appsettings.json` or environment variables, and how to pass a client you create yourself.
 

Documentation/chronicle/reactors/index.md

@@ -24,7 +24,7 @@ export class ShelfBuilder {
 }
 ```
 
-`CreateShelf` is an ordinary Arc `@command()` in your application. Register the reactor the way you register every other artifact, with `builder.add(...)` after [`withChronicle`](../add-event-sourcing.md), or with `builder.discover(...)` in either order. The Arc-owned Chronicle client starts observing it when the application builds.
+`CreateShelf` is an ordinary Arc `@command()` in your application. Register the reactor the way you register every other artifact, with `builder.add(...)` or `builder.discover(...)` after [`withChronicle`](../add-event-sourcing.md). A reactor has no Arc decorator, so one discovered before `withChronicle` is dropped unless [`activateArtifactsInScopes`](scoped-activation.md) is on. The Arc-owned Chronicle client starts observing it when the application builds.
 
 ## How a handler is found
 

Documentation/chronicle/registration-options.md

@@ -15,7 +15,7 @@ builder.withChronicle({ connectionString: 'chronicle://localhost:35000', eventSt
 
 Importing `@cratis/arc.chronicle` adds the `withChronicle` method to the Node builder and to the Fetch API builder from `@cratis/arc.core/fetch`. The package also exports the function `withChronicle(builder, options)`, which does the same; the [Library sample](https://github.com/Cratis/Arc.TypeScript/blob/main/Samples/Library/main.ts) uses that form.
 
-Call it before or after `discover(...)`. The integration records the event types, projections, reducers, reactors, and constraints discovered by Arc, including artifacts discovered earlier. For Chronicle-only artifacts passed to `add(...)`, call `withChronicle` first: Arc otherwise rejects them without an Arc decorator. Arc-owned clients pass the registered artifacts to the SDK and set `discoveryPatterns: []`, so they do not depend on SDK file scanning. For a caller-owned client, import and register your artifacts explicitly or configure its `discoveryPatterns`. Since SDK 6.10.0, compiled JavaScript entry points no longer scan `.ts` files by default; explicit patterns still apply.
+The integration records the event types, projections, reducers, reactors, and constraints Arc discovers. A class with Arc metadata is recorded whether `discover(...)` runs before or after `withChronicle`. A Chronicle-only class, one without an Arc decorator, is picked up only when `withChronicle` runs first, so call it before `discover(...)`. The same goes for `add(...)`, which otherwise rejects such a class. With [`activateArtifactsInScopes`](reactors/scoped-activation.md) on, Chronicle-only classes discovered earlier are picked up too. Arc-owned clients pass the registered artifacts to the SDK and set `discoveryPatterns: []`, so they do not depend on SDK file scanning. For a caller-owned client, import and register your artifacts explicitly or configure its `discoveryPatterns`. Since SDK 6.10.0, compiled JavaScript entry points no longer scan `.ts` files by default; explicit patterns still apply.
 
 Each recorded reactor and reducer also gets a scoped service registration in Arc, but only when nothing else registers that class. A registration in `builder.services` (before or after `withChronicle`), an entry in `options.services`, or an Arc lifetime decorator such as `@singleton()` always wins, and discovering an artifact twice does not add a second registration. With a caller-supplied `ServiceRegistry` in `options.services`, Arc adds no such registrations; register the artifacts yourself. By default Arc does not construct reactors or reducers through these registrations; the Chronicle SDK creates them. To opt in, see [Scoped activation](reactors/scoped-activation.md).