doLast {
                        println "files: " + files.collect { it.name }
                    }
                }
            }
"""

        and: 'preconditions for cleanup of untracked files'
        gcFile.createFile().lastModified = daysAgo(2)
        writeJournalInceptionTimestamp(daysAgo(8))

        when: 'cleaning build is started'

      (filter_ty && filter_ty.hasRank() && filter_ty.getRank() != 4)) {
    return emitOptionalError(location, "Invalid ranks");
  }

  // If either input or filter is unranked, we will just return unranked output
  // shape.
  if (!input_ty || !filter_ty || !input_ty.hasRank() || !filter_ty.hasRank()) {
    Type result_type;
    result_type = UnrankedTensorType::get(

  return success();
}

namespace {

// Hoist coefficient-wise unary operation out of the Unpack op:
//
//   %unpacked:N = "tf.Unpack"(%0)
//   %neg0 = "tf.Neg"(%unpacked#0)
//   %neg1 = "tf.Neg"(%unpacked#1)
//   ...
//   %negN-1 = "tf.Neg"(%unpacked:N-1)
//
// Rewrite it to:
//
//   %neg = "tf.Neg"(%0)
//   %unpacked:N = "tf.Unpack"(%neg)

  }

  // incompatible_shape_error=false implies that the op will either return a
  // valid result or a scalar boolean indicating the error. For unranked outputs
  // we don't know which one it is. TF shape inference turns unranked outputs
  // into ranked ones if it can statically evaluate the broadcast, see the shape
  // function of tf.Equal.
  auto ty = mlir::dyn_cast<RankedTensorType>(op.getType());

  }

  // Returns the inferred function signature of the given function body. Input
  // types are unranked tensor of the respective datatype in the function and
  // result types are inferred by the shape_refiner_. Result types need not be
  // unranked tensors and could be ranked tensors in cases where result type
  // depends on an op with static output shape like tf.Const.

  // CHECK: %[[UNPACK:.*]]:2 = "tf.Unpack"(%[[NEG]]) <{axis = 1 : i64}> {device = ""}
  %unpacked:2 = "tf.Unpack"(%arg0) {axis = 1 : i64, device = ""}
                : (tensor<?x2xf32>) -> (tensor<?xf32>, tensor<?xf32>)
  %0 = "tf.Neg"(%unpacked#0): (tensor<?xf32>) -> tensor<?xf32>
  %1 = "tf.Neg"(%unpacked#1): (tensor<?xf32>) -> tensor<?xf32>

  // CHECK: return %[[UNPACK]]#0, %[[UNPACK]]#1

	WNOHANG                                     = 0x1
	WNOTHREAD                                   = 0x20000000
	WNOWAIT                                     = 0x1000000
	WSTOPPED                                    = 0x2
	WUNTRACED                                   = 0x2
	XATTR_CREATE                                = 0x1
	XATTR_REPLACE                               = 0x2
	XDP_COPY                                    = 0x2

    DCOMMENT("Unexpected rank mismatch: " << lhs << " vs " << rhs);
    return lhs;
  }

  SmallVector<int64_t> shape;
  bool refined_shape = false;
  // Build the shape of the refined type, if lhs is unranked it
  // will be directly the shape of the refined type, otherwise we merged by
  // taking the most specialized. This combines `10x?x?` and `?x?x8` into
  // `10x?x8`.
  if (!lhs_shape_type.hasRank()) {

	for _, v := range job.Status.Conditions {
		if v.Status == v1.ConditionTrue {
			return &v.Type
		}
	}
	return nil
}

// TestPastDeadlineJobFinished ensures that a Job is correctly tracked until
// reaching the active deadline, at which point it is marked as Failed.
func TestPastDeadlineJobFinished(t *testing.T) {
	_, ctx := ktesting.NewTestContext(t)
	clientset := fake.NewSimpleClientset()

	// running, and not all running pods are in the podManager - for instance, force deleting a pod
	// from the apiserver will remove it from the podManager, but the pod may still be terminating and
	// tracked by the podWorkers. Components that need to know the actual consumed resources of the
	// node or are driven by podWorkers and the sync*Pod methods (status, volume, stats) should also
	// consult the podWorkers when reconciling.