Je n'ai jamais pu l'utiliser dans un projet complet, car il n'avait pas d'intégration de sécurité, et je ne pouvais donc pas remplacer toutes les fonctionnalités que j'avais avec les générateurs complets basés sur Flask-apispec. J'avais dans mon backlog de projets de créer une pull request pour ajouter cette fonctionnalité.

            logger.info("[RAG:EVAL] Evaluation completed. hasRelevant={}, relevantCount={}, totalResults={}, elapsedTime={}ms",
                    result.isHasRelevantResults(), result.getRelevantDocIds().size(), searchResults.size(),
                    System.currentTimeMillis() - startTime);
            if (logger.isDebugEnabled()) {
                logger.debug("[RAG:EVAL] Result evaluation completed. hasRelevant={}, relevantDocIds={}, elapsedTime={}ms",

java.net.Inet4Address#getHostAddress()
java.net.Inet6Address#getHostAddress()
java.net.InetSocketAddress#toString()

@defaultMessage avoid DNS lookups by accident: if you have a valid reason, then @SuppressWarnings with that reason so its completely clear
java.net.InetAddress#getHostName()
java.net.InetAddress#getCanonicalHostName()

java.net.InetSocketAddress#getHostName() @ Use getHostString() instead, which avoids a DNS lookup

                return true;
            });
            resultBuf.append("Exported ").append(count).append(" documents.");
            if (logger.isInfoEnabled()) {
                logger.info("[EXPORT] Completed: exportedCount={}, elapsedTime={}ms", count, System.currentTimeMillis() - startTime);
            }
        } catch (final Exception e) {
            logger.warn("Failed to export documents.", e);

Les conteneurs Linux s'exécutent en utilisant le même noyau Linux que l'hôte (machine, machine virtuelle, serveur cloud, etc.). Cela signifie simplement qu'ils sont très légers (comparés à des machines virtuelles complètes émulant un système d'exploitation entier).

Ainsi, les conteneurs consomment **peu de ressources**, une quantité comparable à l'exécution directe des processus (alors qu'une machine virtuelle consommerait beaucoup plus).

        assertEquals(newState, entry.getLeaseState());
        assertEquals(initialEpoch + 1, entry.getEpoch());
        assertFalse(entry.isBreaking()); // Should be false after break handling completes
    }

    @Test
    @DisplayName("Should release lease")
    void testReleaseLease() {
        String path = "/share/file.txt";
        int requestedState = Smb2LeaseState.SMB2_LEASE_READ_HANDLE;

You can add more than one event handler function.

And your application won't start receiving requests until all the `startup` event handlers have completed.

### `shutdown` event { #shutdown-event }

To add a function that should be run when the application is shutting down, declare it with the event `"shutdown"`:

// for the local lock mutex.
// We do not block unlocking or maintenance, but they add to the count.
// The limit is set to allow for bursty behavior,
// but prevent requests to overload the server completely.
// Rejected clients are expected to retry.
const lockMutexWaitLimit = 1000

// lockRequesterInfo stores various info from the client for each lock that is requested.
type lockRequesterInfo struct {

   * Regression test for an edge case where closing response body in the HTTP engine doesn't release
   * the corresponding stream allocation. This test keeps those response bodies alive and reads
   * them after the redirect has completed. This forces a connection to not be reused where it would
   * be otherwise.
   *
   *
   * This test leaks a response body by not closing it.
   *
   * https://github.com/square/okhttp/issues/2409
   */

     */
    void recordMisses(int count);

    /**
     * Records the successful load of a new entry. This should be called when a cache request causes
     * an entry to be loaded, and the loading completes successfully. In contrast to {@link
     * #recordMisses}, this method should only be called by the loading thread.
     *
     * @param loadTime the number of nanoseconds the cache spent computing or retrieving the new