* and then a list is built. These exceptions are usually thrown in "fatal" cases (and usually prevent Maven
     * from starting), and these exceptions may end up very early on output.
     *
     * @param message the base message to enhance
     * @param problems the collection of problems to include in the message
     * @return a formatted message including details of all problems
     */

* The logic depends on external resources or could block in any other way
* Other dependencies need their results (it's a sub-dependency)
* The logic can be shared by multiple endpoints to do things like error early, authentication, etc.
* They need to handle cleanup (e.g., DB sessions, file handles), using dependencies with `yield`
* Their logic needs input data from the request, like headers, query parameters, etc.

     * Flag representing parser processing result: if there were some fatal errors during
     * {@link Parser#parseInvocation(ParserRequest)} this method will return {@code true} and invoker should
     * handle this request as "early failure". In these cases, {@link #options()} usually is absent.
     */
    boolean parsingFailed();

    /**
     * Returns {@code true} if this call happens in "embedded" mode.
     *

var globalMiddlewares = []mux.MiddlewareFunc{
	// set x-amz-request-id header and others
	addCustomHeadersMiddleware,
	// The generic tracer needs to be the first middleware to catch all requests
	// returned early by any other middleware (but after the middleware that
	// sets the amz request id).
	httpTracerMiddleware,
	// Auth middleware verifies incoming authorization headers and routes them

==============================================================================*/

#ifndef TENSORFLOW_C_EAGER_GRADIENTS_INTERNAL_H_
#define TENSORFLOW_C_EAGER_GRADIENTS_INTERNAL_H_

#include "tensorflow/c/eager/gradients.h"

namespace tensorflow {
namespace gradients {
namespace internal {

// Helper functions which delegate to `AbstractOperation`, update
// the state of the ForwardOperation and call the tape as appropriate.

 * Tasks execute with the same happens-before order that the function calls to {@link #submit} and
 * {@link #submitAsync} that submitted those tasks had.
 *
 * <p>This class has limited support for cancellation and other "early completions":
 *
 * <ul>
 *   <li>While calls to {@code submit} and {@code submitAsync} return a {@code Future} that can be
 *       cancelled, cancellation never propagates to a task that has started to run -- neither to

   developers and others in the free software community to coordinate our efforts.

 * **Be pragmatic**: Questions are encouraged and should be asked early in the process to avoid
   problems later. Be thoughtful and considerate when seeking out the appropriate forum for your
   questions. Those who are asked should be responsive and helpful.

그리고 `Depends(scope="request")`(기본값)를 사용하면, `yield` 이후의 종료 코드는 응답이 전송된 후에 실행됩니다.

자세한 내용은 [`yield`가 있는 의존성 - 조기 종료와 `scope`](../tutorial/dependencies/dependencies-with-yield.md#early-exit-and-scope) 문서를 참고하세요.

### `yield`가 있는 의존성과 `StreamingResponse`, 기술 세부사항 { #dependencies-with-yield-and-streamingresponse-technical-details }

  }

  private fun logWithTime(message: String) {
    val startNs = startNs
    val timeMs =
      if (startNs == null) {
        // Event occurred before start, for an example an early cancel.
        0L
      } else {
        TimeUnit.NANOSECONDS.toMillis(System.nanoTime() - startNs)
      }

    logger.invoke("[$timeMs ms] $message")
  }

 - CallStart
 - **CacheMiss**
 - ProxySelectStart
 - ... Standard Events ...
 - CallEnd

### Conditional Cache Hit

When cache flags require checking the cache results are still valid an early cacheConditionalHit event is
received followed by a cache hit or miss.  Critically in the cache hit scenario the server won’t send the response body.