Each <code>DATA</code> directive initializes a section of the
corresponding memory.
The memory not explicitly initialized is zeroed.
The general form of the <code>DATA</code> directive is

<pre>
DATA	symbol+offset(SB)/width, value
</pre>

<p>
which initializes the symbol memory at the given offset and width with the given value.

     */
    public StringSource(CharSequence content, String location) {
        this.content = (content != null) ? content.toString() : "";
        this.location = (location != null) ? location : "(memory)";
        this.hashCode = this.content.hashCode();
    }

    @Override
    public InputStream getInputStream() throws IOException {
        return new ByteArrayInputStream(content.getBytes(StandardCharsets.UTF_8));

import com.google.common.collect.CollectionBenchmarkSampleData.Element;
import java.util.HashMap;
import java.util.Map;
import org.jspecify.annotations.NullUnmarked;

/** Benchmarks for memory consumption of map implementations. */
@NullUnmarked
public class MapsMemoryBenchmark {
  static final Map<String, MapsImplEnum> mapEnums =
      uniqueIndex(
          Iterables.<MapsImplEnum>concat(

 * proportional to the highest valued character that has a replacement. For example a replacement
 * map containing the single character '{@literal \}u1000' will require approximately 16K of memory.
 * As such sharing this data structure between escaper instances is the primary goal of this class.
 *
 * @author David Beaumont
 * @since 15.0
 */
@GwtCompatible
public final class ArrayBasedEscaperMap {
  /**

        // We can't directly test if password is wiped since auth is out of scope,
        // but the close() method should have been called
    }

    /**
     * Test secure memory clearing in close()
     */
    @Test
    @DisplayName("Test secure memory clearing on close")
    public void testCloseSecurelyClearsMemory() {
        char[] testPassword = "CloseTestPass123!".toCharArray();

 *       <i>deeply</i> immutable.
 * </ul>
 *
 * <h4>Performance notes</h4>
 *
 * <ul>
 *   <li>Implementations can be generally assumed to prioritize memory efficiency, then speed of
 *       access, and lastly speed of creation.
 *   <li>The {@code copyOf} methods will sometimes recognize that the actual copy operation is

        connection.connect();
        return connection;
    }

    @Override
    public RdmaMemoryRegion registerMemory(ByteBuffer buffer, EnumSet<RdmaAccess> access) throws IOException {
        // TCP doesn't need real memory registration
        return new TcpMemoryRegion(buffer, access);
    }

    @Override
    public String getProviderName() {
        return "TCP Fallback";
    }

    @Override

   * {@snippet :
   * Stream.of("foo", "quux", "banana", "elephant")
   *     .collect(least(2, comparingInt(String::length)))
   * // returns {"foo", "quux"}
   * }
   *
   * <p>This {@code Collector} uses O(k) memory and takes expected time O(n) (worst-case O(n log
   * k)), as opposed to e.g. {@code Stream.sorted(comparator).limit(k)}, which currently takes O(n
   * log n) time and O(n) space.
   *

        /** File system errors (not found, access denied, etc.) */
        FILE_SYSTEM,
        /** Protocol errors (invalid message, unsupported operation, etc.) */
        PROTOCOL,
        /** Resource errors (out of memory, disk space, etc.) */
        RESOURCE,
        /** Configuration errors */
        CONFIGURATION,
        /** Transient errors that may succeed on retry */
        TRANSIENT,
        /** Unknown or unclassified errors */

models or graphs is equivalent to running untrusted code.

If you need to run untrusted models, execute them inside a
[**sandbox**](https://developers.google.com/code-sandboxing). Memory corruptions
in TensorFlow ops can be recognized as security issues only if they are
reachable and exploitable through production-grade, benign models.

### Compilation