// is std::nullopt, all signatures will be imported. Otherwise, only names
// in `exported_names` are imported.
//
// Note that the word `Lite` means it is a lighter version compared to
// ConvertSavedModelV1ToMlir(), and is not related to TFLite.
//
// TODO(b/179683149): Rename this class to avoid confusion with TFLite.
absl::StatusOr<mlir::OwningOpRef<mlir::ModuleOp>> ConvertSavedModelV1ToMlirLite(

    fi
done < <(find_files | sed 's|/.readonly||')

if [ ${#conflicts[@]} -gt 0 ]; then
    exec 1>&2
    for dir in "${conflicts[@]}"; do
        echo "Found ${dir}/.readonly, but files changed compared to \"${branch}\" branch."
    done
    exit 1
else
    echo "Readonly packages verified."
fi

// Kubernetes, as specified in requests and limits, describing each pod in the
// current scale target (e.g. CPU or memory).  The values will be averaged
// together before being compared to the target.  Such metrics are built in to
// Kubernetes, and have special scaling options on top of those available to
// normal per-pod metrics using the "pods" source.  Only one "target" type
// should be set.

// Kubernetes, as specified in requests and limits, describing each pod in the
// current scale target (e.g. CPU or memory).  The values will be averaged
// together before being compared to the target.  Such metrics are built in to
// Kubernetes, and have special scaling options on top of those available to
// normal per-pod metrics using the "pods" source.  Only one "target" type
// should be set.

// Kubernetes, as specified in requests and limits, describing each pod in the
// current scale target (e.g. CPU or memory).  The values will be averaged
// together before being compared to the target.  Such metrics are built in to
// Kubernetes, and have special scaling options on top of those available to
// normal per-pod metrics using the "pods" source.  Only one "target" type
// should be set.

Java serialization is one way to implement this, however, despite its simplicity of implementation, it has several drawbacks:

- **Performance:**
Java's built-in serialization mechanism is often slower compared to other serialization solutions.
This is due to Java's use of reflection and the need to maintain a lot of metadata.

- **Size of Serialized Data:**

		return name
	}
	return FromToolchain(name)
}

// ToolchainMax returns the maximum of x and y interpreted as toolchain names,
// compared using Compare(FromToolchain(x), FromToolchain(y)).
// If x and y compare equal, Max returns x.
func ToolchainMax(x, y string) string {
	if Compare(FromToolchain(x), FromToolchain(y)) < 0 {
		return y
	}
	return x
}

        // If there are different output properties compared to the previous execution, then we do have output changes
        if (!previous.keySet().equals(current.keySet())) {
            return true;
        }

        // Otherwise, do deep compare of outputs
        ChangeDetectorVisitor visitor = new ChangeDetectorVisitor();

        def compared = PathUtil.getPathComparator(CASE_SENSITIVE).compare(path1, path2)
        def entry1 = new ChildMap.Entry<FileSystemNode>(path1, child1)
        def entry2 = new ChildMap.Entry<FileSystemNode>(path2, child2)
        return compared < 0 ? ChildMapFactory.childMapFromSorted([entry1, entry2]) : ChildMapFactory.childMapFromSorted([entry2, entry1])

 */

package org.jetbrains.kotlin.analysis.api.platform

/**
 * A marker interface for a provider that can be composed, i.e. multiple instances of the same provider can be composed into a single
 * provider.
 *
 * Composable providers share certain traits: There is a notion of a sequentially composed [KotlinCompositeProvider] of that kind, and there