* <p>Whenever possible, it is preferable to test directly for some observable change resulting
   * from GC, as with {@link #awaitClear}. Because there are no guarantees for the order of GC
   * finalization processing, there may still be some unfinished work for the GC to do after this
   * method returns.
   *
   * <p>This method does not create any memory pressure as would be required to cause soft
   * references to be processed.
   *

 *
 * Be advised that short/NetBIOS name usage is not supported with this authenticator. Always specify full FQDNs.
 * This can be a problem if using DFS in it's default configuration as they still return referrals in short form.
 * See <a href="https://support.microsoft.com/en-us/kb/244380">KB-244380</a> for compatible server configuration.
 * See {@link jcifs.Configuration#isDfsConvertToFQDN()} for a workaround.
 */

* API usage monitoring systems
* response data injection systems
* etc.

## Simple and Powerful { #simple-and-powerful }

Although the hierarchical dependency injection system is very simple to define and use, it's still very powerful.

You can define dependencies that in turn can define dependencies themselves.

        it.close();

        // Assert
        assertFalse(it.hasNext());
        verify(tree, times(1)).release();
    }

    @Test
    @DisplayName("doCloseInternal throwing still releases exactly once and next() recovers")
    void closeInternalThrowsButReleases() throws Exception {
        // Arrange
        stubAcquireReturnsSelf();
        FileEntry initial = entry("first");

        // The normalization removes all duplicate backslashes, so \\\\server becomes \server
        // This test validates that paths which should be UNC but aren't detected due to normalization
        // still pass validation (which is the current behavior, though not ideal)

        // These paths normalize to \server\share which is NOT detected as UNC due to normalization

        // Verify we get a copy, not the original
        rawKey[0] = (byte) ~rawKey[0];
        byte[] rawKey2 = keyManager.getRawKey(sessionId);
        assertArrayEquals(testKey, rawKey2, "Should still match original");
    }

    @Test
    public void testRemoveSessionKey() {
        String sessionId = "test-session-3";
        keyManager.storeSessionKey(sessionId, testKey, "AES");

    `NullPointerException` or `OutOfMemoryError`. We now call `Callback.onFailure()` with an
    `IOException` that reports the call as canceled. The triggering exception is still delivered to
    the thread's `UncaughtExceptionHandler`.
 *  Fix: Don't evict incomplete entries when iterating the cache. We had a bug where iterating
    `Cache.urls()` would prevent in-flight entries from being written.

    // ... trigger server failover ...
    
    // Verify that witness notification was received and handled
    assertTrue(listener.waitForNotification(30000));  // 30 second timeout
    
    // Verify file is still accessible after failover
    assertTrue(file.exists());
}

@Test
public void testWitnessServiceDiscovery() throws Exception {
    CIFSContext context = getTestContext();

	if err != nil {
		t.Fatal(err)
	}

	// Object was uploaded with 4 known bad drives, so we should still be able to lose 3 drives and still write to the object.
	erasureDisks = xl.getDisks()
	z.serverPools[0].erasureDisksMu.Lock()
	xl.getDisks = func() []StorageAPI {
		erasureDisks[7] = nil
		erasureDisks[8] = nil

    assertEquals(MAX_SIZE, removalListener.getCount());
    CacheTesting.checkValidState(cache);
  }

  /**
   * With an unlimited-size cache with maxWeight of 0, entries weighing 0 should still be cached.
   * Entries with positive weight should not be cached (nor dump existing cache).
   */
  public void testEviction_maxWeight_zero() {