// (in which case we must actually determine it) or the object is a
	// TypeName and we also want a type (in which case we might detect
	// a cycle which needs to be reported). Otherwise we can skip the
	// call and avoid a possible cycle error in favor of the more
	// informative "not a type/value" error that this function's caller
	// will issue (see go.dev/issue/25790).
	typ := obj.Type()

		return nil
	}
	if finfo != nil && finfo.flags&fOmitEmpty != 0 && isEmptyValue(val) {
		return nil
	}

	// Drill into interfaces and pointers.
	// This can turn into an infinite loop given a cyclic chain,
	// but it matches the Go 1 behavior.
	for val.Kind() == reflect.Interface || val.Kind() == reflect.Pointer {
		if val.IsNil() {
			return nil
		}
		val = val.Elem()
	}

	kind := val.Kind()

        then:
        task.taskDependencies.getDependencies(task)*.path as Set == [':middle:build', ':app:buildDependents'] as Set
    }

    def "build other projects with cyclic dependencies"() {
        given:
        def commonProject = TestUtil.createChildProject(project, "common")
        def middleProject = TestUtil.createChildProject(project, "middle")

// First, it's much faster (up to 50% speedup for real Go programs).
// Second, it eliminates race-related special cases from cgocall and scheduler.
// Third, in long-term it will allow to remove cyclic runtime/race dependency on cmd/go.

// A brief recap of the arm64 calling convention.
// Arguments are passed in R0...R7, the rest is on stack.
// Callee-saved registers are: R19...R28.
// Temporary registers are: R9...R15

    // Under that assumption, the remaining references can be safely marked as @Weak.
    // Using @Weak is necessary to avoid retain-cycles between BiEntry instances on iOS,
    // which would cause memory leaks when non-empty HashBiMap with cyclic BiEntry
    // instances is deallocated.
    @CheckForNull BiEntry<K, V> nextInKToVBucket;
    @Weak @CheckForNull BiEntry<K, V> nextInVToKBucket;

// Then we can simultaneously issue rsc.io/qr v0.3.0 (dropping the gf256 subdirectory)
// and rsc.io/qr/gf256 v0.1.0, including in their respective go.mod
// cyclic requirements pointing at each other: rsc.io/qr v0.3.0 requires
// rsc.io/qr/gf256 v0.1.0 and vice versa. Then a build can be
// using an older rsc.io/qr module that includes the gf256 package, but if

 * </ul>
 *
 * <h2>{@code ClassPath} and symlinks</h2>
 *
 * <p>In the case of directory classloaders, symlinks are supported but cycles are not traversed.
 * This guarantees discovery of each unique loadable resource. However, not all possible
 * aliases for resources on cyclic paths will be listed.
 *
 * @author Ben Yu
 * @since 14.0
 */
@ElementTypesAreNonnullByDefault
public final class ClassPath {

// finalizers, and they are otherwise unreachable, only the finalizer
// for A runs; once A is freed, the finalizer for B can run.
// If a cyclic structure includes a block with a finalizer, that
// cycle is not guaranteed to be garbage collected and the finalizer
// is not guaranteed to run, because there is no ordering that
// respects the dependencies.
//

// First, it's much faster (up to 50% speedup for real Go programs).
// Second, it eliminates race-related special cases from cgocall and scheduler.
// Third, in long-term it will allow to remove cyclic runtime/race dependency on cmd/go.

// A brief recap of the ppc64le calling convention.
// Arguments are passed in R3, R4, R5 ...
// SP must be 16-byte aligned.

// Note that for ppc64x, LLVM follows the standard ABI and

 * </ul>
 *
 * <h2>{@code ClassPath} and symlinks</h2>
 *
 * <p>In the case of directory classloaders, symlinks are supported but cycles are not traversed.
 * This guarantees discovery of each unique loadable resource. However, not all possible
 * aliases for resources on cyclic paths will be listed.
 *
 * @author Ben Yu
 * @since 14.0
 */
@ElementTypesAreNonnullByDefault
public final class ClassPath {