Detect CodeGen gaps by leveraging both server and client generation for TypeSpec-ToDo

[lirenhe]

1. Generate the server code for TypeSpec-ToDo with some basic implementation.
2. Generate client code for Java, Python, JS & .NET and try to run them against service code.

Please help to detect gap by scanning the code and run some manual testing.
Areas to verify:
a. The behavior of different error code handling.
b. Paging behavior.
c. Basic CRUD operation could work E2E.

[lirenhe]

cc @ArcturusZhang , could you help with server generation and work with language owner to verify the client code E2E

[ArcturusZhang]

For dotnet generated library:

1. NoAuth is not working. This spec has @useAuth(NoAuth) on the Users namespace. But the generated Users subclient is actually not affected at all.
2. ApiKeyAuth<cache, "session-id"> is not quite working in the cache way.
3. The @visibility("none") property Validated is exposed in the client generated result (known issue for generator)
4. Pageable operations are not generated as pageable because pageable is not implemented for unbranded libraries.

[weidongxu-microsoft]

Java

• Emitter only support service specify Auth configure in one place. The configure on root namespace takes effect. So NoAuth on Users does not work.
• Error model is not generated. PR in progress.
• Pageable operation is not generated as pageable. Design with architects pending.

[msyyc]

Python issues summary: Azure/autorest.python#2934