query['Version'] = '2011-06-15'

            query_components = []
            for key in query:
                if query[key] is not None:
                    query_components.append("%s=%s" % (key, query[key]))

            query_string = '&'.join(query_components)
            sts_ep_url = self.sts_ep
            if query_string:
                sts_ep_url = self.sts_ep + '?' + query_string

		}
		return n, nil
	}

	// 2 characters...
	for idx := 0; idx < len(p); {
		i := bytes.Index(p[idx:], rr.recordDelimiter)
		if i < 0 {
			break
		}
		idx += i

		p[idx] = '\n'
		p = append(p[:idx+1], p[idx+2:]...)
		n--
	}

	if p[n-1] != rr.recordDelimiter[0] {
		return n, nil
	}

	if _, err = rr.reader.Read(rr.oneByte); err != nil {
		return n, err
	}

One example is when one of the most popular plugins started requiring Java 17 because the underlying framework behind the plugin began requiring 17, and code from the framework was used in the plugin itself.
What would have happened if Gradle was limited to _only Java 11_ at the time?

## Decision

1. The Gradle distribution will never include a Java runtime.
   Instead, Gradle will leverage Java toolchains for the Daemon and worker processes.

   * cached value, which will be recalculated when {@code get()} is called on the reserialized
   * instance. The actual memoization does not happen when the underlying delegate throws an
   * exception.
   *
   * <p>When the underlying delegate throws an exception then this memoizing supplier will keep
   * delegating calls until it returns valid data.
   *

    void testJoinWithEmptyStringElements() {
        String result = StringUtil.join(",", "", "middle", "");
        // First element is empty string (sb becomes "")
        // Second: sb.length() is 0, no delimiter, append "middle" (sb becomes "middle")
        // Third: sb.length() > 0, add delimiter + "" (sb becomes "middle,")
        assertEquals("middle,", result);
    }

    @Test

{* ../../docs_src/app_testing/tutorial001_py310.py hl[2,12,15:18] *}

/// tip | Astuce

Remarquez que les fonctions de test sont des `def` normales, pas des `async def`.

Et les appels au client sont aussi des appels normaux, sans utiliser `await`.

Cela vous permet d’utiliser `pytest` directement sans complications.

///

/// note | Détails techniques

        secondBarrier.await();
      }
    }

    @Override
    protected ScheduledExecutorService executor() {
      // use a bunch of threads so that weird overlapping schedules are more likely to happen.
      return newScheduledThreadPool(10);
    }

    @Override
    protected Scheduler scheduler() {
      return new CustomScheduler() {
        @Override

        secondBarrier.await();
      }
    }

    @Override
    protected ScheduledExecutorService executor() {
      // use a bunch of threads so that weird overlapping schedules are more likely to happen.
      return newScheduledThreadPool(10);
    }

    @Override
    protected Scheduler scheduler() {
      return new CustomScheduler() {
        @Override

        // Test with a very long URL
        StringBuilder longUrl = new StringBuilder("http://");
        for (int i = 0; i < 1000; i++) {
            longUrl.append("a");
        }
        longUrl.append(".example.com:9200");

        String originalValue = System.getProperty(Constants.FESS_SEARCH_ENGINE_HTTP_ADDRESS);

        try {

 *       returns {@code true}.
 *   <li>A class which implements {@code Comparable<UnrelatedType>} is very broken, and will cause
 *       undefined horrible things to happen in {@code Range}. For now, the Range API does not
 *       prevent its use, because this would also rule out all ungenerified (pre-JDK1.5) data types.
 *       <b>This may change in the future.</b>
 * </ul>
 *