fix: correct default mapper docs and preserve sample greetings

[gabisonia]

Description

The default mapper is JsonMessageMapper<>, but 13 XML documentation sites describe a CloudEvents JSON envelope. Correct those descriptions and resolve the generic mapper references. Add 12 regression cases asserting plain JSON round trips through the sync and async defaults registered with AddBrighter and AddConsumers.

The RMQTaskQueue sample also loses Greeting during deserialization because its parameterless constructor leaves a getter-only property unpopulated. Add the public setter proposed in the issue so the greeting survives the round trip.

Related Issues

Fixes #4277

Type of Change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)
• Documentation update

Checklist

• I have read the Contributing Guide
• I have checked the documentation for relevant guidance (repository core guide and FirstMessage tutorial)
• I have added/updated XML documentation for any public API changes
• I have added/updated tests as appropriate
• My changes follow the existing code style and conventions

Additional Notes

Validation:

• Core suite: 1,040 passed and 7 skipped on each of .NET 9 and .NET 10.
• Extensions suite: 194 passed on .NET 9; 191 passed on .NET 10, including all 12 new cases.
• DI library builds across netstandard2.0 and .NET 8/9/10; all 13 corrected XML documentation sites resolve correctly.
• RMQTaskQueue sender and receiver builds pass.
• A temporary harness outside the repository exercised the actual sample type with its mapper and the default sync/async mappers. Normal, empty and null greetings preserve their values and request/correlation IDs in all 9 cases. A tenth check covers reassignment through the sample mapper. The original getter-only source reproduced greeting loss for normal and empty values across all three mapper paths.

The permanent regression tests use the existing TestDoubles convention. Sample-specific checks were local only; no live RabbitMQ broker was used.

[iancooper]

Thanks @gabisonia