Should we have `HeaderViolation` primitive?

BadRequest may only have field violations and no header violations. Why? Shouldn't we add support for header violations as well?

So the following error:

$ curl -H 'Authorization: Bearer <REDACTED>' https://chat.googleapis.com/v1/spaces
{
  "error": {
    "code": 401,
    "message": "Request had invalid authentication credentials. Expected OAuth 2 access token, login cookie or other valid authentication credential. See https://developers.google.com/identity/sign-in/web/devconsole-project.",
    "status": "UNAUTHENTICATED"
  }
}

could look like this:

{
  "error": {
    "code": 401,
    "message": "Request had invalid authentication credentials. Expected OAuth 2 access token, login cookie or other valid authentication credential. See https://developers.google.com/identity/sign-in/web/devconsole-project.",
    "status": "UNAUTHENTICATED",
    "details": [
      {
        "@type": "type.googleapis.com/google.rpc.BadRequest",
        "headerViolations": [
          {
            "header": "Authorization",
            "description": "Missing or invalid OAuth 2 access token in Authorization header."
          },
          {
            "header": "Cookie",
            "description": "Missing or invalid login cookie."
          }
        ]
      }
    ]
  }
}

here's another usage example:

{
  "error": {
    "code": 400,
    "message": "Failed to parse 'Content-Type' header as UTF-8.",
    "status": "INVALID_ARGUMENT",
    "details": [
      {
        "@type": "type.googleapis.com/google.rpc.BadRequest",
        "headerViolations": [
          {
            "header": "Content-Type",
            "description": "The value of the 'Content-Type' header could not be decoded as UTF-8. Please ensure it is properly encoded."
          }
        ]
      }
    ]
  }
}

Drawbacks:

1. There's an overhead of N bytes for storing an array of header violations for each BadRequest response, in Rust it's 24 bytes even when the array is empty.