type iamCache struct {
	updatedAt time.Time

	// map of policy names to policy definitions
	iamPolicyDocsMap map[string]PolicyDoc

	// map of regular username to credentials
	iamUsersMap map[string]UserIdentity
	// map of regular username to policy names
	iamUserPolicyMap *xsync.MapOf[string, MappedPolicy]

	// STS accounts are loaded on demand and not via the periodic IAM reload.

The code following the `yield` statement is executed after the response:

{* ../../docs_src/dependencies/tutorial007_py39.py hl[5:6] *}

/// tip

You can use `async` or regular functions.

**FastAPI** will do the right thing with each, the same as with normal dependencies.

///

## A dependency with `yield` and `try` { #a-dependency-with-yield-and-try }

Brands"}@font-face{font-family:"Font Awesome 5 Free";font-style:normal;font-weight:400;font-display:auto;src:url(./fonts/fa-regular-400.eot);src:url(./fonts/fa-regular-400.eot?#iefix) format("embedded-opentype"),url(./fonts/fa-regular-400.woff2) format("woff2"),url(./fonts/fa-regular-400.woff) format("woff"),url(./fonts/fa-regular-400.ttf) format("truetype"),url(./fonts/fa-regular-400.svg#fontawesome) format("svg")}.far{font-weight:400}@font-face{font-family:"Font Awesome 5 Free";font-style:normal;f...

	builder.WriteString(" <= ")
	builder.AddVar(builder, lte.Value)
}

func (lte Lte) NegationBuild(builder Builder) {
	Gt(lte).Build(builder)
}

// Like whether string matches regular expression
type Like Eq

func (like Like) Build(builder Builder) {
	builder.WriteQuoted(like.Column)
	builder.WriteString(" LIKE ")
	builder.AddVar(builder, like.Value)
}

    /** Current page number being displayed. */
    private int currentPageNumber;

    /** Search/filter parameter for duplicate host configuration ID. */
    public String id;

    /** Search/filter parameter for regular hostname pattern. */
    public String regularName;

    /** Search/filter parameter for duplicate hostname pattern. */
    public String duplicateHostName;

    assertThat(formEncode(255)).isEqualTo("%C3%BF")
  }

  @Throws(IOException::class)
  private fun formEncode(codePoint: Int): String {
    // Wrap the codepoint with regular printable characters to prevent trimming.
    val body =
      FormBody
        .Builder()
        .add("a", String(intArrayOf('b'.code, codePoint, 'c'.code), 0, 3))
        .build()
    val buffer = Buffer()

The unified KMS-based approach has several advantages:

- Key management is now centralized. There is one way to change or rotate encryption keys.
   There used to be two different mechanisms - one for regular S3 objects and one for IAM data.
- Reduced server startup time. For IAM encryption with the root credentials, MinIO had
   to use a memory-hard function (Argon2) that (on purpose) consumes a lot of memory and CPU.

    /**
     * Builds a complete QueryContext for search operations, applying all necessary filters,
     * boosts, and role-based access controls.
     *
     * @param searchRequestType the type of search request (e.g., regular search, admin search)
     * @param query the user's search query string
     * @param context a consumer that allows additional customization of the query context

This way, you can do the upgrade and make sure everything is still working as expected.

## `bump-pydantic` { #bump-pydantic }

In many cases, when you use regular Pydantic models without customizations, you will be able to automate most of the process of migrating from Pydantic v1 to Pydantic v2.

            activeThreads--
            startNextTask()
          }
        }
      }
    }
  }

  /**
   * This blocking queue hooks into a fake clock rather than using regular JVM timing for functions
   * like [poll]. It is only usable within task faker tasks.
   */
  private inner class TaskFakerBlockingQueue<T>(
    val delegate: BlockingQueue<T>,
  ) : AbstractQueue<T>(),