docs(readme): reorder sections for better flow

## Starting point

The "When to use this package" section appeared immediately after
"Installation", before the comparison table, which interrupted the
natural flow from requirements to alternatives comparison.

## Changes

Moved "When to use this package" section in README.md:
- From: lines 46-60 (after Installation section)
- To: lines 71-84 (after Comparison with alternatives table)

Also cleaned up minor whitespace:
- Removed trailing spaces in setup example code
- Removed trailing newline at end of file

## Result

Documentation now flows more logically:
1. Installation
2. Comparison table (detailed feature breakdown)
3. When to use (guidance based on that comparison)
4. Scope

The detailed comparison table now precedes the decision guidance,
allowing readers to understand capabilities before making choices.

Moved using safe-line-mover MCP for deterministic copy/paste.

Author: FirstUnicorn

README.md

@@ -43,21 +43,6 @@ from messaging.infrastructure import SqlAlchemyOutboxRepository
 - PostgreSQL (outbox persistence)
 - Kafka Connect with Debezium CDC (publishing infrastructure)
 
-## When to use this package
-
-**Use `python-eventing` if you need:**
-- Guaranteed event delivery via transactional outbox pattern
-- Kafka-based microservice messaging with CDC publishing
-- Dead letter queue handling with database bookkeeping
-- Idempotent consumer patterns with durable deduplication
-- Native broker integration (FastStream, Debezium CDC, RabbitMQ DLX)
-
-**Consider alternatives if:**
-- Simple in-process events only → [`pyventus`](https://github.com/mdapena/pyventus)
-- FastAPI request-scoped events → [`fastapi-events`](https://github.com/melvinkcx/fastapi-events)
-- Non-Kafka message brokers without CDC support
-- No need for durable outbox persistence
-
 Support scale: `❌` none, `✅` basic, `✅✅` strong, `✅✅✅` first-class
 
 ## Comparison with alternatives
@@ -83,6 +68,21 @@ Support scale: `❌` none, `✅` basic, `✅✅` strong, `✅✅✅` first-class
 | Consumer batch handling | ❌ | ❌ | ✅✅✅ | `fastapi-events` supports `handle_many(...)`; this package stays one-message-per-consume today |
 | FastAPI-local event flow | ❌ | ✅ | ✅✅✅ | This package intentionally avoids request-lifecycle middleware eventing |
 
+## When to use this package
+
+**Use `python-eventing` if you need:**
+- Guaranteed event delivery via transactional outbox pattern
+- Kafka-based microservice messaging with CDC publishing
+- Dead letter queue handling with database bookkeeping
+- Idempotent consumer patterns with durable deduplication
+- Native broker integration (FastStream, Debezium CDC, RabbitMQ DLX)
+
+**Consider alternatives if:**
+- Simple in-process events only → [`pyventus`](https://github.com/mdapena/pyventus)
+- FastAPI request-scoped events → [`fastapi-events`](https://github.com/melvinkcx/fastapi-events)
+- Non-Kafka message brokers without CDC support
+- No need for durable outbox persistence
+
 ## Scope
 
 **Included:**
@@ -455,11 +455,11 @@ app = FastAPI()
 async def startup():
     # Database engine
     engine = create_async_engine("postgresql+asyncpg://...")
-    
+
     # Outbox repository
     outbox_repo = SqlAlchemyOutboxRepository(engine)
     app.state.outbox_repository = outbox_repo
-    
+
     # Optional: EventBus for advanced patterns
     event_bus = build_event_bus()
     app.state.event_bus = event_bus
@@ -585,4 +585,3 @@ poetry run pytest
 - See `.github/workflows/README.md` for details
 
 **For complete local test coverage on Windows:** See [Windows Testing with WSL2](docs/WINDOWS_TESTING_WITH_WSL2.md) (requires Docker Desktop WSL integration setup).
-