*
 * <p>If your iterator supports modification through {@code remove()}, you may wish to override the
 * verify() method, which is called after each sequence and is guaranteed to be called
 * using the latest values obtained from {@link IteratorTester#newTargetIterator()}.
 *
 * <p>The value you pass to the parameter {@code steps} should be greater than the length of your

  std::string DebugString() const override;

  // Returns a Boolean hint indicating whether callers should prefer
  // `SummarizeValue` to resolving this handle and formatting the tensor.
  //
  // For example some tensor handles may represent distributed values, in which
  // case placement information is lost when resolving the handle.
  //
  // If false, a caller might implement pretty-printing by resolving and

//
// If the device is successfully registered, `status` is set to TF_OK. Otherwise
// the device is not usable. In case of a bad status, `device.delete_device` is
// still called on `device_info` (i.e. the caller does not retain ownership).
//
// This API is highly experimental, and in particular is expected to change when
// it starts supporting operations with attributes and when tf.function support
// is added.

            }
        }
        // out is closed by the caller and should never be closed by the handler.
        return nil
    }

    err := manager.RegisterStreamingHandler(grid.HandlerDiskInfo, StreamHandler{
        Handle: handler,
        Subroute: "asubroute",
        OutCapacity: 1,
        InCapacity: 1,
    })
```

Sample call:
```go
    // Get a connection to the remote host

/**
 * An extension of {@code DataInput} for reading from in-memory byte arrays; its methods offer
 * identical functionality but do not throw {@link IOException}.
 *
 * <p><b>Warning:</b> The caller is responsible for not attempting to read past the end of the
 * array. If any method encounters the end of the array prematurely, it throws {@link

 *
 * <p>{@code ByteSource} provides two kinds of methods:
 *
 * <ul>
 *   <li><b>Methods that return a stream:</b> These methods should return a <i>new</i>, independent
 *       instance each time they are called. The caller is responsible for ensuring that the
 *       returned stream is closed.
 *   <li><b>Convenience methods:</b> These are implementations of common operations that are

        SmbInfoAllocation info = new SmbInfoAllocation();
        
        // Read total allocation units.
        info.alloc = readInt8( buffer, bufferIndex );
        bufferIndex += 8;
        
        // read caller available allocation units 
        info.free = readInt8( buffer, bufferIndex );
        bufferIndex += 8;

        // skip actual free units
        bufferIndex += 8;
        

	if r.throttle == nil {
		return r.r.Read(buf)
	}
	if r.lastErr != nil {
		err = r.lastErr
		return
	}
	b := r.throttle.Burst()  // maximum available tokens
	need := len(buf)         // number of bytes requested by caller
	hdr := r.opts.HeaderSize // remaining header bytes
	var tokens int           // number of tokens to request

	if hdr > 0 { // available tokens go towards header first
		if hdr < b { // all of header can be accommodated

   * {@code null}!) before returning it to callers. Depending on how the code is structured, a
   * nullness analysis might not understand that the field has been populated. To avoid that problem
   * without having to add {@code @SuppressWarnings}, the code can call this method.
   *
   * <p>Why <i>not</i> just add {@code SuppressWarnings}? The problem is that this method is

    // 3 options here for ordering
    // 1) Legacy Platform - based on the Platform/Provider existing ordering in
    // sslSocket.enabledCipherSuites
    // 2) OkHttp Client - based on MODERN_TLS source code ordering
    // 3) Caller specified but assuming the visible defaults in MODERN_CIPHER_SUITES are rejigged
    // to match legacy i.e. the platform/provider
    //
    // Opting for 2 here and keeping MODERN_TLS in line with secure browsers.