// Create a new TFE_TensorHandle with the same contents as 'h' but placed
// in the memory of the device name 'device_name'.
// If source and destination are the same device, then this creates a new handle
// that shares the underlying buffer. Otherwise, it currently requires at least
// one of the source or destination devices to be CPU (i.e., for the source or
// destination tensor to be placed in host memory).

   * overkill for most Atomics because they are usually irregularly
   * scattered in memory and thus don't interfere much with each
   * other. But Atomic objects residing in arrays will tend to be
   * placed adjacent to each other, and so will most often share
   * cache lines (with a huge negative performance impact) without
   * this precaution.
   *
   * In part because Cells are relatively large, we avoid creating

@Experimental
@Immutable
public interface Dependency extends Artifact {
    /**
     * {@return the type of the dependency}.
     * A dependency can be a <abbr>JAR</abbr> file,
     * a modular-<abbr>JAR</abbr> if it is intended to be placed on the module path,
     * a <abbr>JAR</abbr> containing test classes, <i>etc.</i>
     *
     * @see DependencyCoordinates#getType()
     */
    @Nonnull
    Type getType();

    /**

     * we did list an exclusion on c, it was only from within the context of
     * project b.  We will pick up project c in this case because no
     * restrictions were placed on d.  This demonstrates that a, b, c, & d will
     * all be collected.
     *
     * @throws Exception
     */
    @Test

| xlVersion | [4]byte     | `'1', ' ', ' ', ' '`
| xlMetaV2  | msgp object | All versions as single messagepack object
| [EOF] | |

### v1.1+

Version 1.1 added inline data, which will be placed after the metadata.

Therefore, the metadata is wrapped as a binary array for easy skipping.

| Entry          | Encoding       | Content
| ---------------|----------------|----------------------------------------

   * overkill for most Atomics because they are usually irregularly
   * scattered in memory and thus don't interfere much with each
   * other. But Atomic objects residing in arrays will tend to be
   * placed adjacent to each other, and so will most often share
   * cache lines (with a huge negative performance impact) without
   * this precaution.
   *
   * In part because Cells are relatively large, we avoid creating

  virtual absl::Status Dim(int dim_index, int64_t* dim) const = 0;

  // Returns the device which created the handle.
  virtual const char* DeviceName(absl::Status* status) const = 0;
  // Returns the device where the tensor was placed.
  virtual const char* BackingDeviceName(absl::Status* status) const = 0;
  // Returns the device type which created the handle.
  virtual const char* DeviceType(absl::Status* status) const = 0;

2. **Disclose significant AI involvement.** If AI tools played a major role in producing your contribution, say so in the description. We are not interested in incidental use, such as asking a chatbot for advice or accepting an autocomplete suggestion—but rather in cases where a tool generated substantial portions...

   * http://code.google.com/p/smhasher/source/browse/trunk/MurmurHash3.cpp, which contained the
   * following header:
   *
   * MurmurHash3 was written by Austin Appleby, and is placed in the public domain. The author
   * hereby disclaims copyright to this source code.
   */
  static int smear(int hashCode) {
    return C2 * Integer.rotateLeft(hashCode * C1, 15);
  }

   * http://code.google.com/p/smhasher/source/browse/trunk/MurmurHash3.cpp, which contained the
   * following header:
   *
   * MurmurHash3 was written by Austin Appleby, and is placed in the public domain. The author
   * hereby disclaims copyright to this source code.
   */
  static int smear(int hashCode) {
    return C2 * Integer.rotateLeft(hashCode * C1, 15);
  }