/** Returns the original request that initiated this web socket. */
  fun request(): Request

  /**
   * Returns the size in bytes of all messages enqueued to be transmitted to the server. This
   * doesn't include framing overhead. If compression is enabled, uncompressed messages size
   * is used to calculate this value. It also doesn't include any bytes buffered by the operating

import org.slf4j.Logger;
import org.slf4j.LoggerFactory;

/**
 * RDMA buffer manager for efficient memory region pooling.
 *
 * Manages pools of pre-registered memory regions to avoid the overhead
 * of frequent registration/deregistration during RDMA operations.
 */
public class RdmaBufferManager implements AutoCloseable {

    private static final Logger log = LoggerFactory.getLogger(RdmaBufferManager.class);

- **Memory efficient** - Specialized collections like `LruHashMap` and `ArrayMap` for memory-conscious applications
- **Java 21 features** - Switch expressions and pattern matching for reduced overhead

## 🧪 Building and Testing

### Development Setup
```bash
# Clone the repository
git clone https://github.com/codelibs/corelib.git
cd corelib

# Compile the project
mvn clean compile

// in tests only.)
TF_CAPI_EXPORT int TF_PickUnusedPortOrDie(void);

// Fast path method that makes constructing a single scalar tensor require less
// overhead and copies.
TF_CAPI_EXPORT extern TFE_TensorHandle* TFE_NewTensorHandleFromScalar(
    TF_DataType data_type, void* data, size_t len, TF_Status* status);

// Specify the server_def that enables collective ops.

The *Secure Channel* splits the object content into chunks of a fixed size of `65536` bytes. The last chunk may be smaller to avoid adding additional overhead and is treated specially to prevent truncation attacks. The nonce value is 96 bits long and generated randomly per object / multi-part part. The *Secure Channel* supports plaintexts up to `65536 * 2^32 = 256 TiB`.

#### Randomness

But if you are certain that the content that you are returning is **serializable with JSON**, you can pass it directly to the response class and avoid the extra overhead that FastAPI would have by passing your return content through the `jsonable_encoder` before passing it to the response class.

{* ../../docs_src/custom_response/tutorial001b.py hl[2,7] *}

/// info

### Multi destination replication

The replication design for multiple sites works in a similar manner as described above for two site scenario. However there are some
important exceptions.

 * form}), the memory requirement is 8*N bytes for the copy of the dataset plus an overhead which is
 * independent of N (but depends on the quantiles being computed). When calling {@link
 * ScaleAndIndex#computeInPlace computeInPlace} (in {@linkplain ScaleAndIndexes#computeInPlace
 * either form}), only the overhead is required. The number of object allocations is independent of
 * N in both cases.
 *

        byte[] decompressed = searchHelper.gzipDecompress(compressed);

        assertEquals(testData, new String(decompressed));
        // For small strings, gzip overhead might make compressed data larger
        // So we just verify that compression/decompression works correctly
        assertNotNull(compressed);
        assertTrue(compressed.length > 0);
    }

 *       allocating garbage).
 * </ul>
 *
 * <p>Disadvantages compared to {@code long[]}:
 *
 * <ul>
 *   <li>Memory footprint has a fixed overhead (about 24 bytes per instance).
 *   <li><i>Some</i> construction use cases force the data to be copied (though several construction
 *       APIs are offered that don't).