Add new codgen processor to throw ModelInvalidException on naming case and type collisions instead of NPE

[RanVaknin]

Adds a codegen processor that throws a typed exception when a model has two shapes whose names differ only in the case of their first letter AND have different types.

Example:

    "shapes": {
      "Endpoint": {
        "type": "structure",
        "members": {}
      },
      "endpoint": {
        "type": "string"
      }

Codegen capitalizes the first letter of each shape name to derive its Java class name, so both Endpoint and endpoint map to the class Endpoint.java. Codegen keys its shape map by that class name, so one shape silently overwrites the other, and a member still referencing the lost shape later can cause a codegen time error (observed as an NPE).

The processor only flags collisions where the two shapes have different types. Same type collisions produce equivalent output, so the overwrite is harmless.

The new processor detects when two shapes of different types map to the same generated class name, and throws a single ModelInvalidException listing every collision found.

[RanVaknin]

Closing the PR because tests keep failing. We are unable to reliably reproduce this issue because the original model that was submitted and broke codegen is no longer available.

[github-actions]

This pull request has been closed and the conversation has been locked. Comments on closed PRs are hard for our team to see. If you need more assistance, please open a new issue that references this one.