// and transforms the module in place to cluster the given ops for compilation
// that is compatible with the given device_type. The MLIR should be in the TF
// Executor Dialect for graph nodes and edges or TF Functional. It will convert
// to TF Functional internally. Individual Op inside a node should be the
// Tensorflow Dialect. The output MLIR is in the TF Functional Dialect.  The

  // We need to introduce this version to adapt to the output of gpu inference
  // converter. The single argument overload version calls this function.
  //
  // When add_edges_to_output_of_downstream_nodes is true, the output edges of
  // the xla_launch_node's immediate downstream nodes would be attached to the
  // generated xla node. For example, if the original graph is
  // StatefulPartitionedCall{_xla_compile_id=1} -> XlaClusterOutput -> NodeA

// and transforms the module in place to cluster the given ops for compilation
// that is compatible with the given device_type. The MLIR should be in the TF
// Executor Dialect for graph nodes and edges or be in TF Functional already.
// Individual Op inside a node should be the Tensorflow Functional Dialect. The
// output MLIR is in the TF Functional Dialect. Returns OkStatus if passed,
// otherwise an error.
//
// Inputs:

            }
            renderer.completeChildren();
        }
    }

    private void doRender(final RenderableDependency node, boolean last, Set<Object> visited) {
        // Do a shallow render of any constraint edges, and do not mark the node as visited.
        if (node.getResolutionState() == RenderableDependency.ResolutionState.RESOLVED_CONSTRAINT) {
            renderNode(node, last, false, dependenciesRenderer);

     * to resolve the platform itself. If the platform was declared as a dependency,
     * then the engine thinks that the platform module is unresolved. We need to
     * remember such edges, because in case a virtual platform gets defined, the error
     * is no longer valid and we can attach the target revision.
     *
     * @param edge the orphan edge
     */
    void addOrphanEdge(EdgeState edge) {

 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */

package org.gradle.execution.plan.edges;

import com.google.common.collect.ImmutableSortedSet;
import org.gradle.execution.plan.Node;

import java.util.HashSet;
import java.util.Iterator;
import java.util.NavigableSet;
import java.util.Set;

	reachable := ReachableBlocks(f)

	// Get rid of edges from dead to live code.
	for _, b := range f.Blocks {
		if reachable[b.ID] {
			continue
		}
		for i := 0; i < len(b.Succs); {
			e := b.Succs[i]
			if reachable[e.b.ID] {
				b.removeEdge(i)
			} else {
				i++
			}
		}
	}

	// Get rid of dead edges from live code.
	for _, b := range f.Blocks {
		if !reachable[b.ID] {

    if (connections == null) {
      return false;
    }

    if (allowsSelfLoops()) {
      // Remove self-loop (if any) first, so we don't get CME while removing incident edges.
      if (connections.removeSuccessor(node) != null) {
        connections.removePredecessor(node);
        --edgeCount;
      }
    }

    for (N successor : ImmutableList.copyOf(connections.successors())) {

  private final boolean allowsSelfLoops;
  private final ElementOrder<N> nodeOrder;

  final MapIteratorCache<N, GraphConnections<N, V>> nodeConnections;

  long edgeCount; // must be updated when edges are added or removed

  /** Constructs a graph with the properties specified in {@code builder}. */
  StandardValueGraph(AbstractGraphBuilder<? super N> builder) {
    this(
        builder,

    if (connections == null) {
      return false;
    }

    if (allowsSelfLoops()) {
      // Remove self-loop (if any) first, so we don't get CME while removing incident edges.
      if (connections.removeSuccessor(node) != null) {
        connections.removePredecessor(node);
        --edgeCount;
      }
    }

    for (N successor : ImmutableList.copyOf(connections.successors())) {