divide($rfs-base-value, divide($rfs-base-value * 0 + 1, $rfs-rem-value));\n}\n\n// Cache $rfs-breakpoint unit to prevent multiple calls\n$rfs-breakpoint-unit-cache: unit($rfs-breakpoint);\n\n// Remove unit from $rfs-breakpoint for calculations\n@if $rfs-breakpoint-unit-cache == px {\n  $rfs-breakpoint: divide($rfs-breakpoint, $rfs-breakpoint * 0 + 1);\n}\n@else if $rfs-breakpoint-unit-cache == rem or $rfs-breakpoint-unit-cache == \"em\" {\n  $rfs-breakpoint: divide($rfs-breakpoint, divide($rfs-breakpoint...

  The incorrect loop logic might lead to a plugin being loaded multiple times, consequently preventing any Pod from being scheduled, which is unexpected. ([#122370](https://github.com/kubernetes/kubernetes/pull/122370), [@caohe](https://github.com/caohe)) [SIG Scheduling]

### Other (Cleanup or Flake)

          + ")";
    }
  }

  /**
   * Returns a supplier that caches the instance supplied by the delegate and removes the cached
   * value after the specified time has passed. Subsequent calls to {@code get()} return the cached
   * value if the expiration time has not passed. After the expiration time, a new value is
   * retrieved, cached, and returned. See: <a
   * href="http://en.wikipedia.org/wiki/Memoization">memoization</a>

  - Regenerate list of EC2 instances
  - Fix pricing endpoint in AWS China Region
  
  ### Azure
  - Add optional jitter on initial VMSS VM cache refresh, keep the refreshes spread over time
  - Serve from cache for the whole period of ongoing throttling
  - Fix unwanted VMSS VMs cache invalidations
  - Enforce setting the number of retries if cloud provider backoff is enabled
  - Don't update capacity if VMSS provisioning state is updating

    /** The request attribute key for storing cached user agent type */
    protected static final String USER_AGENT_TYPE = "ViewHelper.UserAgent";

    /**
     * Determines the user agent type from the current HTTP request.
     * The method analyzes the User-Agent header to categorize the browser type
     * and caches the result in the request attribute for subsequent calls.
     *

  The incorrect loop logic might lead to a plugin being loaded multiple times, consequently preventing any Pod from being scheduled, which is unexpected. ([#122371](https://github.com/kubernetes/kubernetes/pull/122371), [@caohe](https://github.com/caohe)) [SIG Scheduling]

## Dependencies

### Added
_Nothing has changed._

### Changed
_Nothing has changed._

### Removed
_Nothing has changed._



# v1.26.12

IAGvCvG,EAAKsD,KAAK7B,GAAK,SAAUuC,EAAIpK,GAC5B,GAAuC,oBAA3BA,EAAQmK,gBAAkC3D,EAAiB,CACtE,IAAIlG,EAAON,EAAQmK,eAAgBC,GACnC,OAAO9J,EAAO,CAAEA,GAAS,OAI3B8F,EAAKsG,OAAO7E,GAAM,SAAUuC,GAC3B,IAAIuC,EAASvC,EAAGlH,QAASuF,EAAWC,GACpC,OAAO,SAAUpI,GAChB,IAAIxB,EAAwC,oBAA1BwB,EAAKsM,kBACtBtM,EAAKsM,iBAAkB,MACxB,OAAO9N,GAAQA,EAAK8F,QAAU+H,IAMhCvG,EAAKsD,KAAK7B,GAAK,SAAUuC,EAAIpK,GAC5B,GAAuC,oBAA3BA,EAAQmK,gBAAkC3D,EAAiB,CACtE,IAAI1H,EAAME,EAAGgC,EACZV,EAAON,EAAQmK,eAAgBC,GAEhC,GAAK9J,EAAO,CAIX,IADAxB,EAAOwB,...

    same implementation as their `tf.losses` equivalent.
*   For Keras model, the individual call of `Model.evaluate` uses no cached data
    for evaluation, while `Model.fit` uses cached data when `validation_data`
    arg is provided for better performance.
*   Adds a `save_traces` argument to `model.save`/ `tf.keras.models.save_model`

		return replicateMetadata
	}

	// Compare only necessary headers
	compareKeys := []string{
		"Expires",
		"Cache-Control",
		"Content-Language",
		"Content-Disposition",
		"X-Amz-Object-Lock-Mode",
		"X-Amz-Object-Lock-Retain-Until-Date",
		"X-Amz-Object-Lock-Legal-Hold",
		"X-Amz-Website-Redirect-Location",

                else:
                    yield {**(maybe_nested_state or {}), **(maybe_original_state or {})}

    return merged_lifespan  # type: ignore[return-value]


# Cache for endpoint context to avoid re-extracting on every request
_endpoint_context_cache: dict[int, EndpointContext] = {}


def _extract_endpoint_context(func: Any) -> EndpointContext: