sleep 30

./mc ls -r --versions minio1/newbucket >/tmp/minio1.txt
./mc ls -r --versions minio2/newbucket >/tmp/minio2.txt

out=$(diff -qpruN /tmp/minio1.txt /tmp/minio2.txt)
ret=$?
if [ $ret -ne 0 ]; then
	echo "BUG: expected no missing entries after replication resync: $out"
	exit 1
fi

**FastAPI** 调用的就是 `RequestValidationError` 类，因此，如果在 `response_model` 中使用 Pydantic 模型，且数据有错误时，在日志中就会看到这个错误。

但客户端或用户看不到这个错误。反之，客户端接收到的是 HTTP 状态码为 `500` 的「内部服务器错误」。

这是因为在*响应*或代码（不是在客户端的请求里）中出现的 Pydantic `ValidationError` 是代码的 bug。

修复错误时，客户端或用户不能访问错误的内部信息，否则会造成安全隐患。

### 覆盖 `HTTPException` 错误处理器

同理，也可以覆盖 `HTTPException` 处理器。

例如，只为错误返回纯文本响应，而不是返回 JSON 格式的内容：

```Python hl_lines="3-4  9-11  22"

          # Tests can do anything
          - "!$test"
          # Main code should only be used by appropriate binaries
          - "!**/operator/**"
          - "!**/istioctl/**"
          - "!**/tools/bug-report/**"
          # This should only really import operator API, but that is hard to express without a larger refactoring
          - "!**/pkg/kube/**"
          - "!**/pkg/url/**"
          - "!**/pkg/test/framework/**"

        MinMaxPriorityQueue<Integer> q = MinMaxPriorityQueue.create(perm);
        q.removeAt(i);
        assertIntactUsingStartedWith(perm, q);
      }
    }
  }

  /** Regression test for bug found. */
  public void testCorrectOrdering_regression() {
    MinMaxPriorityQueue<Integer> q = MinMaxPriorityQueue.create(ImmutableList.of(3, 5, 1, 4, 7));
    List<Integer> expected = ImmutableList.of(1, 3, 4, 5, 7);

It should be this way because if you have a Pydantic `ValidationError` in your *response* or anywhere in your code (not in the client's *request*), it's actually a bug in your code.

And while you fix it, your clients/users shouldn't have access to internal information about the error, as that could expose a security vulnerability.

### Override the `HTTPException` error handler

    assertThat(newArray).isEqualTo(new char[] {(char) 0, (char) 1, (char) 2});
    newArray[1] = (char) 5;
    assertThat((char) list.get(1)).isEqualTo((char) 1);
  }

  // This test stems from a real bug found by andrewk
  public void testAsList_subList_toArray_roundTrip() {
    char[] array = {(char) 0, (char) 1, (char) 2, (char) 3};
    List<Character> list = Chars.asList(array);

    LocalClassWithSeeminglyHiddenThisInStaticInitializer unused =
        new LocalClassWithSeeminglyHiddenThisInStaticInitializer();
  }

  /**
   * This class demonstrates a bug in getParameters() when the local class is inside static
   * initializer.
   */
  private static class LocalClassWithSeeminglyHiddenThisInStaticInitializer {
    static {
      class Local {

   * thereof.
   *
   * @param formal The type whose type variables or itself is mapped to other type(s). It's almost
   *     always a bug if {@code formal} isn't a type variable and contains no type variable. Make
   *     sure you are passing the two parameters in the right order.

    LocalClassWithSeeminglyHiddenThisInStaticInitializer unused =
        new LocalClassWithSeeminglyHiddenThisInStaticInitializer();
  }

  /**
   * This class demonstrates a bug in getParameters() when the local class is inside static
   * initializer.
   */
  private static class LocalClassWithSeeminglyHiddenThisInStaticInitializer {
    static {
      class Local {

    new OwnerTypeSubtypingTests().testAllDeclarations();
  }

  public void testWildcardSubtypes() throws Exception {
    new WildcardSubtypingTests().testAllDeclarations();
  }

  /**
   * This test reproduces the bug in canonicalizeWildcardType() when the type variable is
   * recursively bounded.
   */
  public void testRecursiveWildcardSubtypeBug() throws Exception {
    Exception e =
        assertThrows(